# Csiro Contract Ligo

Document Sample

					T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

Activities during 2006, plans for 2007
Concatenation of all Instrument Science LSC/Lab reports and MOU plans
14 August 2006      D. Shoemaker, compiler
T060198-00-Z

Aciga

Activities Related to Attachment ACF:
a) Interferometer Configurations
(i) Benchtop Variable Reflectivity Mirror (VRM) with no arm cavities
Whilst this experiment has not been completed, due mainly to funding issues, good progress
has been made. The system has been characterized. Error signals for all the longitudinal dof s
have been examined and half of these are now understood and operational. We estimate that
another 3 months is required to complete the set up and extract the data. The status of the
experiment was reported at the Elba meeting [1].
(ii) Squeezed state injection
No progress as (i) not yet completed.
(iii) Benchtop RSE with VRM
Whilst no experimental progress has been made, due to the status reported in (i), a control
strategy has been worked out. It is based on extending the technique used in (i) of injecting a
PM modulated sub carrier into the interferometer via through the signal/VRM and operating
he interferometer on a dark fringe. Our proposal has been published in CQG [2].
b) Squeezed Light Generation
(i) Travelling wave, doubly resonant squeezer
After initial results with the ring cavity, doubly resonant squeezer, reported at Amaldi 6, the
system was tweaked to produce 3 dB of squeezing down to 30 Hz [3]. It was found that the
maximum pump power was limited to 200 mW thereby limiting the amount of squeezing
producing. Beyond this level an oscillation, believed to be photothermal in origin, set in. In
an effort to control this oscillation a piezo with a much higher resonant frequency (250kHz)
was designed and built enabling the construction of a fast control, servo (bandwidth expected
to exceed 100 kHz). We are in the process of installing this new piezo and re optimising the
system. The causes limiting squeezing to above 30Hz were investigated (see (iii)).
(ii) Control of the squeeze angle No activity
(iii) Classical limitations to reaching very low frequencies
A thorough investigation of mechanisms was undertaken. We were able to build a homodyne
detector with 70 dB common mode rejection. We found that beam pointing fluctuations on
the surface on the non uniform photodiode surface is a major factor. Results have been
submitted for publication [4].
(iv) 3rd generation squeezer.
In collaboration with MIT and Hannover we have begun an investigation of new crystals and
different materials, in particular, PPKTP. Early results with PPKTP in a different application
at ANU (at a different wavelength) are very encouraging. We developed and patented a novel
optical readout technique for locking the phase matching condition in OPA [5].
People
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

ANU: McKenzie, Mow Lowry, Rabeling, Gossler, Lam, Gray, McClelland.
[1] D. Rabeling, M.B. Gray and D.E. McClelland, "Control of Advanced GW Interferometer
Configurations" Elba Meeting 2006.
[2] D. Rabeling, S. Gossler, M.B Gray and D.E. McClelland, " A new topology for the
control of complex interferometers", Class. Quant. Grav.23, S267-S275 (2006).
[3] K. McKenzie, M.B. Gray, S.Gossler, P.K. Lam and D.E. McClelland, "Squeezed state
generation for interferometric gravitational wave detection", Classical and Quantum
Gravity.23, S245-S250.
[4] K. McKenzie, M.B. Gray, P.K.Lam and D.E. McClelland, "Technical limitations to
homodyne detection at audio frequencies", Applied Optics (2006) submitted
[5] K. McKenzie, M.B. Gray, P.K.Lam and D.E. McClelland, "Nonlinear phase matching
locking via optical readout", Optics Express (2006) submitted.

Plans

a) Interferometer Configurations
• Complete bench top Variable Reflectivity Mirror (VRM) with no arm cavities using an
Output Mode Cleaner for control signal injection.
• Complete a bench top RSE with VRM experiment
• Commence an experimental investigation of the use of squeezing in the above
configurations
• interact with the AdvLIGO controls group to explore issues around the use of the ANU

b) Squeezed Light Generation
• explore the limitations on squeezing with the current travelling wave doubly resonant
MgO:LNiO3 system and optimise performance
• employ techniques to lock all dofs without using noise locking
• investigate new non linear materials
• design and build a 3rd generation squeezer with the goal to achieve 6 dB across the LIGO
band
• participate in the effort at Caltech to integrate squeezing with the 40m interferometer

c) Other Contributions
Standard Quantum Limit
• develop a low frequency torsional pendulum
• measure displacement noise of a mirror mounted on the torsional suspension

CaRT

a.    Interferometer Configurations for LIGO-III - The exploration of candidate optical
configurations for LIGO-III and beyond was put on hold this year; no progress to report. We
expect to return to this at a low level during the coming year.
b. Testing Quantum Theory for Macroscopic Systems Using Future LIGO
Interferometers - This project, initiated in CaRT, has become a collaboration between CaRT
and Yanbei Chen‘s group at the AEI, and Thorne has turned the leadership of it over to Chen.
CaRT participants are Chao Li, Yasushi Mino and at a lower level Mike Boyle, Curt Cutler,
Ilya Mandel and Kip Thorne. AEI participants are Yanbei Chen, HelgeMuller-Ebhardt,
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

Henning Rehbein, and Kentaro Somiya. Much of this research is focusing on a specific class
of experiments in which the LIGO mirrors‘ position variable x is prepared in a coherent state
(or a state close to coherent) using a feedback system plus observations, then this state is
manipulated by altering LIGO‘s optical spring (via motion of the signal recycling mirror),
resulting in a highly squeezed state of x, and then a measurement of x is made. The new,
highly squeezed state is mapped out by a series of identical experiments of this type.
Mathematical descriptions of the various stages of such an experiment are being developed.
One goal of this work is to develop a much deeper quantum mechanical, measurement-theory
understanding of LIGO interferometers.

Plans

a. Interferometer Configurations for LIGO-III - The exploration of candidate optical
configurations for LIGO-III and beyond was put on hold this year; no progress to report. We
expect to return to this at a low level during the coming year.

b. Testing Quantum Theory for Macroscopic Systems Using Future LIGO
Interferometers - This project, initiated in CaRT, has become a collaboration between CaRT
and Yanbei Chen's group at the AEI, and Thorne has turned the leadership of it over to Chen.
CaRT participants are Chao Li, Yasushi Mino and at a lower level Mike Boyle, Curt Cutler,
Ilya Mandel and Kip Thorne. AEI participants are Yanbei Chen, Helge Muller-Ebhardt,
Henning Rehbein, and Kentaro Somiya. Much of this research is focusing on a specific class
of experiments in which the LIGO mirrors' position variable x is prepared in a coherent state
(or a state close to coherent) using a feedback system plus observations, then this state is
manipulated by altering LIGO's optical spring (via motion of the signal recycling mirror),
resulting in a highly squeezed state of x, and then a measurement of x is made. The new,
highly squeezed state is mapped out by a series of identical experiments of this type.
Mathematical descriptions of the various stages of such an experiment are being developed.
One goal of this work is to develop a much deeper quantum mechanical, measurement-theory
understanding of LIGO interferometers.

Columbia
>> We contributed to the AdLIGO effort.
GECo joined the AdLIGO effort and it is taking the lead in design, development, test,
installation and commissioning of the AdLIGO timing and timing diagnostic system. GECo
students are presently testing and validating the next generation of the DuoTone based timing
monitoring system, which will also be part of the AdLIGO conceptual design.

Plans

Florida
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

Stable recycling cavities/TCS and ASC issues.
(Guido Mueller, Muzammil Arain, Charles Perry, Wan Wu, V. Quetschke)
The work on stable recycling cavities has shifted to core optics issues. These issues need to
be solved before any meaningful analysis of stable vs. unstable recycling cavities can
continue. Our group started calculating thermal effects in the main interferometer (mainly the
ITM mirrors) and studied thermal compensation schemes. We calculated the thermal lensing
and mode degradation in the interferometer. Our results agree well with results obtained by
Phil Willems et al. boosting the confidence in our modeling abilities. We started developing
our own strategy for the thermal compensation system parallel to Phil Willems proposed
strategy. Both strategies might continue to evolve in parallel to provide alternative solutions
to this problem or they merge taking the best of both strategies. We also developed new ideas
like the use of negative dn/dT material in the compensation plate and started evaluating its
potential.
We started using Finesse to calculate alignment sensing signals. Together with Andreas
Freise (Birmingham) we wrote the main Finesse file for the baseline design which will
enable us to calculate WFS signals. So far it has been used to calculate 10-mode amplitudes
and phases generated by different tilted mirrors. However, there are still some inconsistencies
which need to be resolved. This first step should lead to a better understanding of the
complex alignment issue and provided us with a good starting point for the ongoing WFS
signal calculations.
We have started adapting the Finesse code to model stable recycling cavities A strawman
layout in the LI GO vacuum envelope is in progress.
Cryogenic interferometers
(Ramsey Lundock, Stacy Wise, David Tanner)
If a future interferometer is built to operate at cryogenic temperatures, novel methods of
removing the head deposited in the mirror by the residual absorption of the laser light will be
needed. We have completed a theoretical and numeric study of the radiative heat transfer due
to photon tunneling between two surfaces held close together compared to thermal
wavelengths. In this case, evanescent coupling between the warmer and cooler objects can
enhance the heat transfer by many orders of magnitude.
We have designed and have nearly finished construction of a cryogenic apparatus to
sapphire test pieces and capacitive readout for separation and angular adjustments. We will
use this to measure the heat conduction between an object at 4.2 K and one at 20—-50 K.

Plans
a) Interferometer Configurations
Stable recycling cavities/TCS and ASC issues. (Muzammil Arain, Guido Mueller)
We will calculate alignment signals for Advanced LIGO for the stable and unstable recycling
case. The current Finesse file will be modified to solve the discrepancies which are likely
related with some mode mismatches in the file. We will also study the TCS system and
together with Phil Willems we will develop a realistic strategy for to handle the thermal
beam deformations which we expect to see in Advanced LIGO. We will also work on a
realistic design which fits into the LIGO vacuum envelope.

b) Symmetric RF-sidebands/Length Sensing and Control (Muzammil Arain, Guido
Mueller, Volker Quetchke) Once we have a good handle on the TCS system we will revisit
the symmetric RF sideband design and study its applicability although we still expect that the
losses are probably too high. We will also collaborate with Rana Adhikari and Peter Fritschel
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

on the development of a new length sensing scheme for Adv. LIGO. This will directly impact
the ASC sensing scheme.

c) Other Contributions
Near-field enhancement of heat transfer
As research aimed at some future cryogenic detector, we will continue to investigate near-
field radiative cooling of mirror substrates, including the effect of coatings, the requirements
on separation of hot/cold objects, and the transfer of mechanical motion between the two
objects. An experiment to demonstrate the effect will start taking data during the fall.

GEO

GEO has continued to support the 40m work, mainly through membership of the technical
advisory committee. The initial program of work at Glasgow on 3 mirror cavities is nearing
completion. The control system has been fully characterised and papers are in preparation.
At Hannover, squeezed optical fields in the full GW audio band were demonstrated. A new
control scheme were developed and utilized for stable phase locking of the squeezed field in
respect to a table-top Michelson interferometer. A 1/f role-up of homodyne detector noise at
frequencies below 100 Hz was observed and is under investigation.

The optical Kerr effect in interferometers was investigated theoretically, and showed that
Kerr media in the arms of a GW detector with resonant sideband extraction can be useful.
Code has been developed that can be used to calculate the optical signal- and noise-transfer-
functions of multiple-mirror cavities including the effect of radiation pressure. The results
match independent results from the MIT code.
An all-reflective arm cavity based on a high efficiency reflection grating in 1st order Littrow
arrangement was studied experimentally. A grating with 99.6% diffraction efficiency was
fabricated and cavity finesse of about 1500 demonstrated.
A diffractive coupler has been designed and is being manufactured for incorporation into a
suspended cavity at Glasgow.
At Birmingham (Freise) Simulation code (FINESSE) has been upgraded the numeric
interferometer simulation FINESSE to be used for non-classical interferometry. The
components needed to handle gratings and the structures needed to simulate quantum noise
have been added (the latter in collaboration with Hannover). A finesse model of Advanced LI
GO has been released, several input files have been prepared to enable work on both LSC
and ASC.
Kentaro Somiya is currently working on Advanced LIGO ISC.
At Golm/Hannover Yanbei Chen and Kentaro Somiya have completed the following work:
Analysis of the contribution of laser frequency and intensity noise in high-power RSE
interferometers, with RF and DC readout schemes (in collaboration with Kawamura from
NAOJ and Mio from Tokyo University).
Helge Müller-Ebhardt, with YC and KS, analyzed the quantum-noise-limited performance of
Sagnac interferometers with detuned signal recycling, and their possible application to intra-
Archana Pai, with YC and KS, in collaboration with Shuichi Sato and Seiji Kawamura from
NAOJ, Keiko Kokeyama from Ochanomizu University, and Robert Ward from Caltech,
designed an "multi-interferometer" configuration that does not sense displacement noises of
mirrors (up to linear order), but has non-vanishing sensitivity to gravitational waves.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

Helge Müller-Ebhardt, with YC and KS, have been studying the possibility of using
Advanced LI GO interferometers to test quantum mechanics (also in collaboration with Kip
Thorne's research group at Caltech, and Sam Waldman from LIGO-Caltech).
In collaboration with Farid Khalili, Sergey Vyatchanin, and Stefan Danilishin, studied the
discrete variational readout scheme, and realized that it is equivalent to a heterodyne
detection with generic, but still periodic, modulation demodulation functions. The gain in
sensitivity previously reported by Danilishin et al. originated from ponderomotive squeezing
at sideband frequencies around integer multiples of the modulation frequency.

Plans
a) Interferometer Configurations
At Birmingham it is proposed to provide, distribute and maintain the FINESSE code with
suitable input files for Advanced LIGO and to obtain simulation results with respect to LSC
and ISC.
At Glasgow and Hannover it is proposed to continue to support configurations research at the
Caltech 40m prototype, mainly through participation on the advisory committee and
occasional visits. At Glasgow a suspended, diffractively-coupled cavity will be prepared and
constructed, in cooperation with the Hannover group. The results from the previously
reported "3 mirror cavity" tests will be written up for publication. At Hannover, the
squeezing experiment will continue with an investigation of excess homodyne noise below
100\,Hz. The feasibility of incorporating Kerr media in interferometers will be considered An
all-reflective low-loss 50/50 grating beam splitter will be fabricated. This beam splitter will
be used in a Michelson interferometer. Power-recycling will be used to measure the loss. At
Golm and Hannover, theoretical and modelling work will continue. The main topics will be
simulations and calculations relating to the development of Advanced LIGO ISC and
continued consideration of potential QND schemes for 3rd generation detectors.

Maryland
Plans
a) Interferometer Configurations
Pan will study the influence of the degeneracy of the signal-recycling cavity (SRC) on the
alignment and sensing control system, by numerically computing and optimizing the control
matrix through a modal model simulation.
Buonanno and Pan (in collaboration with Chen (AEI)) will carry out theoretical
investigations of designs of a 4kmlong SRC for Advanced LIGO, focusing on the possibility
to reduce the loss of SNR due to optical losses in the SRC, the change of detector response to
GW signals for various choices of long-SRC parameters, especially the optimization of the
detector noise curve for narrowband GW signals (e.g., LMXBs), the possibility to collect
signal power in both sidebands, and practical issues including the light scattering in a beam
tube containing both an arm cavity and a SRC, and length and alignment control systems in a
long-SRC interferometer. Buonanno and Pan will also investigate the control of the
dynamical instability in signal-recycled interferometers and possible optical configurations
for third generation interferometers.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

Moscow

2.2 Improvements of the Advanced LIGO topology via optical rigidity
The sensitivity of the Advanced LIGO gravitational-wave detector sensitivity can be
improved by means of relatively modest modifications of the detector topology. The best
known method to do this is based on the use of the optical rigidity, which turns the
gravitational-wave detector end masses into harmonic oscillators providing better sensitivity
in the narrow band close to its resonance frequency than a free mass does [6, 7, 8]. F.Ya.
Khalili, S.P. Vyatchanin and graduated student V.I. Lazebny has explored the narrowband
regime of the optical rigidity in detail, taking into account the optical losses in the detector
arms. It was shown that the optical losses limit the gain in sensitivity and achievable signal-
to-noise ratio. Nevertheless, for parameter values planned for the Advanced LIGO
gravitational wave detector, this limitation is insignificant [9, 10].

2.3 Development of the optical bars/optical lever topology of gravitational wave
detectors
The intracavity topologies [11, 12, 13, 14, 15] allow to obtain sensitivity better than the
Standard Quantum Limit while keeping the optical pumping energy in the antenna at
relatively low level. In the optical bars/optical lever scheme the optical field acts as two rigid
springs located between the end mirrors of the antenna and additional local mirror placed
between them. When gravitational waves change the end mirrors position, these optical
springs change position of the local mirror. In order to detect the local mirror displacement
(relative to some external reference mass), an additional meter • the local meter • has to be
used. For example, it can be implemented as a small-scale Fabry-Perot cavity. In the paper
[16] F.Ya.Khalili and his former graduate student S.L.Danilishin continued to develop the
optical bars/optical lever intracavity topology of the laser gravitational-wave detectors and
have analyzed the following important issues which had not been considered in the previous
papers.
1.      It is evident that the local meter is the key element of the intracavity topologies which
dene the sensitivity. F.Ya.Khalili and S.L.Danilishin analyzed the combination of the optical
lever intracavity topology with the local meter based on the discrete sampling variation
measurement (DSVM). They have shown that for the realistic parameters of
the local meter, the calculated sensitivity can be about one order of magnitude better that the
Standard Quantum Limit (providing that all other noise sources common for all topologies
are substantially reduced). This sensitivity can be provided with the power circulating in the
main interferometer about ten times smaller than proposed for the Advanced LIGO.
2.      The optical losses in the mirrors may be the most important sensitivity limitation for the
intracavity topologies. In the previous papers their influence was calculated only for the
simplified model based on two coupled harmonic oscillators. Now it is calculated rigorously.
It was shown, that the optical rigidity suppresses at low frequencies mechanical fluctuations
of the test masses created by the optical power fluctuations.
3.      Gravitational-wave detector based on the intracavity topology is, in essence, a quantum
measuring device which has to be coupled with the environment as weakly as possible
• i.e. its mirrors have to have as small transmittance as possible. At the same time, some
means for injection of the optical power inside have to be provided. F.Ya.Khalili and
S.L.Danilishin proposed the scheme which allows to preserve high quality factor for the anti-
symmetric (signal) mode, while providing the optimal coupling for the symmetric (power)
mode. This scheme is similar to one which is used in the standard topology and consists of a
beamsplitter and three recycling mirrors. In addition, this scheme preserves the symmetry of
the topology, thus reducing the influence of the pumping power amplitude and phase
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

fluctuations. In the paper [17], F.Ya.Khalili further investigated the possibility of use the
spectral quantum variational measurement in the optical bars/optical lever intracavity
topology. He analyzed the sensitivity limitations of this scheme imposed by the optical losses
both in the local meter itself and in the filter cavity. He showed that the main sensitivity
limitation stems from the filter cavity losses. In order to overcome it, it is necessary to
increase the filter cavity length. Estimates for two possible experimental scenarios was done.
In the full scale laser gravitational wave detector, the sensitivity about ten times better than
the Standard Quantum Limit can be achieved using high, but realistic value of the optical
power circulating in the local meter cavity, about several kilowatts, small (m ~ 1g) local
mirror, and long (kilometer-scale) filter cavity!

A prototype experiment with the goal to overcome the Standard Quantum Limit by factor 23
can be performed using much smaller circulating power (about tens of watts) and relatively
short filter cavity (with length ~ 10m).

NAOJ

V. Leonhardt, F. Kawazoe, K. Kokeyama, S. Sato, and S. Kawamura started an experiment
to demonstrate the new signal extraction method in the 4-m suspended RSE interferometer at
NAOJ. This signal extraction method, which would be a backup method for Advanced LIGO
interferometer, uses two modulations: one phase modulation and one amplitude modulation.
The phase modulation sidebands will transmit through the Michelson interferometer because
of the Michelson asymmetry, while the amplitude modulation sidebands will be completely
reflected back because of the Michelson asymmetry. This contrast of the behavior between
the two sets of the sidebands makes it possible to extract most effectively the length signal of
the signal extraction cavity.
We started the experiment with building the Mach-Zehnder interferometer for the double
modulation. This is necessary to produce double modulation without the sidebands of
sidebands. We succeeded in locking the Mach-Zehnder interferometer with one phase
modulation. We should still add the amplitude modulation.
With this phase-modulated light coming from the Mach-Zehnder interferometer, we
succeeded in locking the Fabry-Perot Michelson interferometer. We still have to add the
power recycling mirror and the signal extraction mirror.
We also improved the miniature suspension system so that the residual Q-factor was reduced
to the reasonable level.
As for the squeezing experiment, V. Leonhardt, S. Sakata, S. Sato, and S. Kawamura
investigated the possibility of observing the radiation pressure noise using very light mirrors
without employing an extremely high power laser nor a cavity of extremely high finesse. We
calculated all the noise sources including the thermal noise and concluded that it is quite
possible to observe the radiation pressure noise with the following specifications: mass of the
mirror of 20 mg, fiber with diameter of 10 micron and length of 10 mm, and finesse of the
cavity of 7000.
For this experiment we successfully made the 10 micron fiber, and we established the design
of the set-up of the experiment.
N. Atsushi and S. Kawamura have been working on the new configuration for Advanced
LIGO. We focused on the locked Fabry-Perot type RSE, and found that it could give the
better sensitivity than the ordinary RSE. We need more investigation for this.
S. Kawamura visited Caltech for 40 days in summer of 2006 to work on the 40m prototype.
He has been working on the lock acquisition of the 40m detuned RSE interferometer.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

Plans

a) Interferometer Configurations
V. Leonhardt, F. Kawazoe, K. Kokeyama, S. Sato, and S. Kawamura will continue an
experiment to demonstrate the new signal extraction method in the 4-m suspended RSE
interferometer at NAOJ. This signal extraction method, which would be a backup method for
Advanced LIGO interferometer, uses two modulations: one phase modulation and one
amplitude modulation. The phase modulation sidebands will transmit through the Michelson
interferometer because of the Michelson asymmetry, while the amplitude modulation
sidebands will be completely reflected back because of the Michelson asymmetry. This
contrast of the behavior between the two sets of the sidebands makes it possible to extract
most effectively the length signal of the signal extraction cavity.

During this period we will first add the amplitude modulation to the existing Mach-Zehnder
interferometer. With this double-modulated light, we will try to lock the power recycled
Michelson interferometer, then the tuned RSE without the arm cavities. After we lock the
central part of the interferometer, we will try to lock the whole RSE interferometer.

N. Atsushi and S. Kawamura will continue to work on the new configuration for Advanced
LIGO. During this period we will focus on the realistic comparison between the locked
Fabry-Perot type RSE and the ordinary RSE in terms of sensitivity. S. Kawamura will visit
Caltech in summer of 2007 to work on the 40m prototype. He will focus on whatever is the
most important issue at that time.

b) Squeezed Light Generation
V. Leonhardt, S. Sakata, S. Sato, and S. Kawamura will start the experiment of the quantum
non-demolition. Our first goal is to observe the radiation pressure noise in the Fabry-Perot
Michelson interferometer. And then we will try to reduce it using the homodyne detection.
We have already found that it is quite possible to observe the radiation pressure noise with
the following specifications: mass of the mirror of 20 mg, fiber with diameter of 10 micron
and length of 10 mm, and finesse of the cavity of 7000. We will build the whole set-up as
well as making the super-small mirror attached to the super-thin fiber. We will then measure
Q of the mirrors. We will measure the sensitivity curve and do noise hunting to improve the
noise so that we can see the radiation pressure noise.

Stanford

(R. Byer, M. Fejer, E. Gustafson, P. Beyersdorf, K-X. Sun, A. Bullington, P. Lu, S.
Sinha)
a) Wavefront distortions due to thermoelastic deformations
Our manuscript originally intended for submission to Classical and Quantum Gravity has
now been submitted to the Journal of the Optical Society of America. We have received
P.P. Lu, A.L. Bullington, P. Beyersdorf, S. Traeger, J. Mansell, R. Beausoleil, E. Gustafson,
R.L. Byer and M.M. Fejer,"Wavefront distortion of the specular and diffracted beams
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

produced by thermo-elastic deformation of a diffraction grating heated by a Gaussian laser
beam", in submission to JOSA-A (2006).

Melody is an object-oriented MATLAB model for the simulation of interferometer response
to thermal loading from a high power laser source and is currently being supported and
updated by A. Bullington to meet the needs of the LIGO community. The latest revisions to
Melody are posted online at www.stanford.edu/~abull/.
A method for studying thermal compensation has been added to Melody by modifying the
finite element simulation code Numerical Optic, originally written by Ryan Lawrence at
MIT. The Numerical Optic model now includes heating by an auxiliary laser beam in the
form of either a Gaussian or annular beam. The thermal operators generated by Numerical
Optic can be subsequently used in Melody to determine how the input power that gives the
maximum sideband recycled power changes with auxiliary heat. Since Melody also has the
capability to simulate the various length sensing error signals, the thermal compensation
version of Numerical Optic was used to study the affect of changing auxiliary laser power on
these error signals. However, the amount of computation time needed to generate all of the
necessary temperature profiles in Numerical Optic and subsequently compute the
corresponding thermal operators proved too great to provide any useful simulation data.
Several minor revisions have been made to Melody, including updating the MATLAB
version checking code and correcting an incorrect function call. The latest version of Melody
to date is v2.1c.

As power levels are increased in interferometric detectors, thermally-induced wavefront
distortions become of increasing concern. We are in the process of characterizing thermal
using moderate input optical power (up to 20 W) by inducing thermal distortions via
absorption in the underlying substrate of the folding mirror. This is expected to approximate
the effect of thermal loading of low loss optical coatings under much higher optical powers.
Mirrors have been custom coated by Research Electro-Optics (REO) using a variety of
substrates such as fused silica, BK7, undoped YAG and sapphire in order to compare the
properties of one substrate material against another. Other materials which may be
investigated are silicon or KG5, an IR absorbing glass.
A 30 W beam has been mode-matched into a low finesse cavity to provide a clean, spatially
filtered higher power beam. This beam is then mode-matched to the mode-cleaner cavity
being studied under thermal load. The test cavity currently being studied was constructed
with low-loss fused silica input/output optics and a folding mirror with a low loss coating but
a lossy substrate made from KG5. This cavity was then thermally loaded while being locked
to resonance in its polarization-dependent high finesse resonant mode (a finesse of
approximately 8000). The cavity began showing thermally induced behavior at a lower than
expected threshold near 100 mW of input laser power instead of a few watts. Starting at 100
mW of input power, the transmitted spatial mode was distorted by coupling to various higher
order modes. At 600 mW and higher, the transmitted power no longer increased linearly with
input power and the transmitted spatial mode became severely distorted by strong coupling to
higher order modes. The behavior of the cavity was totally repeatable - for every given input
power, the same transmitted power and coupling to higher order modes were observed.
The same mode-cleaner test cavity has also been thermally loaded in its polarization-
dependent low finesse resonant mode (a finesse of approximately 300). In this case, the
transmitted power increased linearly with incident power until the input power reached 5 W,
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

where the coupling began to decline and higher order modes also appeared, superposed with
the fundamental mode. This behavior continued up to the onset of thermal cycling at 6.5 W,
where the longitudinal cavity length adjustment was no longer able to keep the cavity stably
locked given the curvature distortion of the rear optic (lossy folding mirror). Thermal cycling
continued at higher input powers, indicating that 6.5 W is the maximum amount of thermal
loading this particular mode-cleaner can handle. We are currently attempting to fit the
observed behavior with theoretical models.

d) Active control of cavity mode shape
Last year, we reported on our experiment to test an optical cavity with alternative mirror
shape that could support optical modes that were less affected by thermal noise in the test
mass substrate. The adaptive optic was a gold-coated electrostatically-actuated membrane.

We investigated the effect of misalignment of the adaptive optic on the performance of the
optical cavity. It was determined that misalignment would cause the sensitivity to degrade for
two reasons: a relative increase in the shot-noise when the power coupling to the cavity
decreases, and/or an increase of thermo-elastic noise as the mode shape changes to a shape
that less effectively averages over the spatial variations of the noisy mirror surface. We found
that the first of these effects was more significant. In addition we observed that the flat-top
cavity is about twice as sensitive to misalignment as a conventional cavity supporting
Gaussian modes would be if it had the same full-width half-maximum mode shape.

Our manuscript describing this work, "Cavity with a deformable mirror for tailoring the
shape of the eigenmode", has been accepted for publication in the September 10th issue of
Applied Optics. It describes the experiment to generate and study flat-top modes in a cavity
that could be used to reduce the thermo-elastic noise in Advanced LIGO.

We have evaluated a collection of MEMS mirrors for studying the coupling of shot noise and
radiation pressure noise. The MEMS mirrors were thin flexures that were stressed to the
point of buckling. This "Euler spring" was intended to allow the mirror to respond freely to
forces, however the alignment of the mirror on the buckled flexure proved problematic and
has led us to pursue a method of simulating this effect by measuring the (technical) radiation
pressure incident on a mirror and using a feed-forward circuit to drive the mirror via a
piezoelectric actuator. This provides an analogous coupling of amplitude-to-phase noise that
would be present in a free mirror, but does not require the low mass mirrors or high laser
power that makes the direct measurement problematic.

We are hosting an undergraduate student from San Jose State for the summer (Charlotte Nix)
who is developing a source of technical noise that simulates quantum noise allowing the
coupling between phase and amplitude noise in table-top interferometers to be investigated in
a manner analogous to the coupling of radiation pressure and shot noise that will allow
advanced LIGO to exceed the standard quantum limit. The electronic servos have been
designed and fabricated and the optical layout of the experiment has been determined. We
are currently aligning the optics.

f)    Control of quantum noise in heterodyne detection
Peter Beyersdorf has developed a framework for analyzing quantum noise in heterodyne
detection. Preliminary calculations done in this framework suggest that with an appropriate
laser spectrum, heterodyne detection can be made to have many of the same advantages that
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

homodyne detection provides, namely identical quantum noise limited sensitivity and
suppression of modulator noise via arm cavity filtering. A manuscript describing these results
is in preparation.

g) All-reflective topologies
Our investigation of all reflective topologies has been delayed as we concentrate our efforts
on quantum noise and grating characterization
Last year, we supported an undergraduate summer research student (Jon O'Bryan) to
construct an interferometer for measuring the phase noise produced by the transverse
displacement noise of gratings. This Mach-Zender interferometer was designed with a
flexible geometry that would allow us to test the phase noise introduced by a grating that is
used in different configurations including both single- and double-pass configurations. This
project is currently on hold while the student is on leave.

This summer, we are hosting a master's student from San Jose State (Adnan Alam) who is
measuring the loss in gratings that are exposed to a dirty vacuum system and high-intensity
radiation to qualify the gratings for use in Advanced LIGO. He has designed an ellipsometer
to monitor the contamination of the grating surface and is currently constructing an
experiment that will allow this contamination to be monitored while the grating is exposed to
a dirty vacuum system.

h) Multiport grating interferometry
Our recent studies on grating interferometry have extended previous investigations at
Stanford on two-port grating interferometry to multi-port grating interferometry. We have
considered several concepts that could be of benefit to future ground-based detectors. We
have investigated the use of multi-port gratings as beam-splitters and cavity couplers,
focusing on two issues; understanding how the additional output ports of the grating may be
used to provide sensing information on auxiliary degrees-of-freedom of the diffractive optic,
and using the unique properties of such gratings to combine the function of the beam-splitter
and the input coupler test masses in a single optic.

The flexibility built into the multiport grating can help to simplify two-arm resonant Fabry-
Perot interferometers such as LIGO. This idea was presented by K.X. Sun over a decade ago
in a LIGO research note. The recent effort is intended to reduce the number of optics in order
to reduce thermal noise. An experimental effort is underway to verify a design that would
significantly simplify advanced gravity wave interferometer detector architectures.

j) Angular sensing
We developed an optical lever that reads out the angle of a diffracted beam at a near-grazing
angle from an incident beam at a near-normal angle. The sensitivity of this angle to the
relative angle of the grating surface is much larger than that of the zero order reflected beam.
The compression of the spatial profile of the beam by the grating further enhances the
sensitivity.
We have experimentally demonstrated a grating-based optical lever with a 1200 lines/mm
grating and a (first order) diffraction angle of 81 degrees. Over a 5 cm baseline, we measured
an angular sensitivity better than 10 nrad/rHz. The construction is extremely simple with no
angular magnification telescopes needed. The entire sensor could fit on the suspension cage
and measure the relative angle of a grating attached to the penultimate mass, or one etched
onto the core optics of Advanced LIGO with better common mode noise rejection than the
optical levers of initial LIGO.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

We have also demonstrated a symmetric grating sensor that utilizes a custom, ~935
lines/mm, optical grating. Both the +1 and -1 diffraction orders, which were symmetrically
distributed on either side of the grating surface normal, were used for angular sensing. With a
symmetric grating sensor, laser frequency variations will induce common-mode motions of
the two diffraction orders which can be cancelled out by proper combination of the two
detector outputs. The symmetric grating sensor is therefore robust against laser frequency
noise and thus lowers the laser frequency stability requirement.

Using the symmetric grating sensor, we demonstrated an angular sensitivity better than 1
nrad/rHz. For LIGO with 4 km arms, a 1-nrad/rHz sensitivity implies that local angular
sensing can align the remote beam spot with a resolution of 25 micrometers, well below the
current 1 mm upper limit. Further, the grating-based angular measurement does not introduce
any additional optical elements between the sensed surface and the photodiode. Potential
noise sources due to motion or instability of in-path optics such as an optical telescope would
thus be eliminated.
Current efforts on component technology involve specialty grating design and experimental
verification of gratings with an appropriate diffraction efficiency at large diffraction angles.
Our experimental efforts are focused on improving the symmetrical grating angular sensor.
In addition, we are setting up to demonstrate the application of a grating angular sensor in a
rotational pendulum.

Stanford has filed an invention disclosure for the symmetric grating angular sensor due to its
potential application in precision instrumentation. Scientists from National Ignition Facility
(NIF) have expressed interest in using the grating angular sensor as an alignment tool.

[1] Ke-Xun Sun, and Robert Byer, "Compact, High Precision Grating Angular Sensor for
LIGO Interferometer
Control," LIGO Science Collaboration, LIGO Hanford Observatory, Washington, Aug. 14-
17, 2005. LIGO Document G050451-00-Z.
[2] Ke-Xun Sun, Patrick Lu, and Robert Byer, "Progress in Grating Optical Sensors for
Gravitational Wave
Detection", LIGO Science Collaboration, Joint Meeting of Optics and Suspension Working
Groups, LIGO Hanford
Observatory, Washington, March 22, 2006. LIGO Document G060142-00-Z.

k) Collaboration with Lawrence Livermore National Laboratory (LLNL) Lawrence
Livermore National Laboratory has been a center of design and fabrication of high quality
dielectric gratings thanks to the need from NIF. We have established a collaboration with
LLNL on grating design and fabrication. LLNL is partially supporting one of our graduate
students and making available high-value dielectric gratings for testing purposes. Patrick Lu
has spent most of the past year training at LLNL to carry out the development of optical
grating technology. In turn, Stanford will characterize the gratings for LLNL. The LLNL
collaboration is thus an important component of Stanford's LIGO in both a scientific and a
financial sense.

[1] Patrick Lu, Ke-Xun Sun, and Robert Byer, "Grating Fabrication for Gravitational Wave
Interferometers and LISA
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

GRS ", 6th International LISA Symposium, Goddard Space Flight Center, Greenbelt,
Maryland, June 19-23, 2006. To appear in the AIP Proceedings of the 6th International LISA
Symposium.

l) Grating scattering loss measurement
LIGO requires that the scattering loss of transmissive optical elements be less than 10-5 to
10-6 for minimization of stray light noises. The major portion of the scattering losses
originates from the top few layers in the dielectric coating stack, although there may be some
contribution from the interface between the dielectric stack and the substrate as well.
Commercial gratings have considerably higher scattering losses than are acceptable for
LIGO. Therefore, it will be important to design, fabricate, and test gratings with low
scattering losses. It is possible that gratings will turn out to have comparable, or even lower
scattering losses than transmissive optics, as the reflective gratings has only one surface to
interact with the laser beam, and this would be an additional benefit of an all-reflective
configuration.
We are in the process of setting up an existing scatterometer with in-line optics to see if it can
be used to measure grating scattering losses.

m) Grating mode cleaner
A grating mode-cleaner should have considerably higher power handling capability than the
transmissive version discussed in section (c). We are working on the design and fabrication
of a high efficiency (greater than 99%) dielectric grating through our collaboration with
LLNL. Our preliminary design studies suggest that it will be possible to achieve the grating
efficiencies required. An initial trial fabrication has already shown a diffraction efficiency of
96% at 1064 nm.

Plans

a) Interferometer Configurations

Advanced Configurations (R. Byer, M. Fejer, E. Gustafson, A. Bullington, S. Sinha)
Study all-reflective topologies and gratings for application to future ground-based detectors;
Make the following updates to Melody;
for all-reflective configurations.

b) Squeezed Light Generation
(R. Byer, M. Fejer, E. Gustafson, P. Beyersdorf)
Investigate the relationship between transverse displacement noise of gratings and the phase
noise it produces for various interferometer configurations; This will be used to specify an
additional constraint on suspension system designs that would be used for suspending
diffractive optics.
Use the quantum-noise test-bed to simulate the effects of shot noise and radiation pressure
noise on suspended mass interferometers in the regime where both effects are of comparable
magnitude;
This test-bed will be used to explore the effects on the quantum noise caused by having
multiple carrier frequencies injected into an interferometer. Externally induced amplitude and
phase noise will simulate quantum noise, but with much larger amplitude to ease the
challenge of observation. The goal is to develop an RF readout scheme that can have a
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

frequency dependent readout phase to take advantage of squeezed light, while avoiding the
"nonstationary" shot noise that typically accompanies configurations with RF readout.

c) Other Contributions
(R. Byer, M. Fejer, E. Gustafson, K-X. Sun, A. Bullington, P. Lu, S. Sinha)
Continue measuring the thermal distortion of mode-cleaners with high circulating powers;
The mode structure and power coupled into the mode-cleaner are currently being determined
as a function of increasing incident power. The results are subsequently compared with
computer models, including theoretical predictions of the mode-cleaner‘s behavior from
MELODY.

using custom-coated optics for the study of cavity response under thermal load from high
power laser illumination, including cavities with optical substrates made from undoped YAG
and sapphire;
Contingent on their availability, we will design an all-reflective mode-cleaner using gratings

Refine requirements for gratings as core optics in novel configurations to allow high
circulating power in future upgrades to LIGO;

Collaborate with Lawrence Livermore National Laboratory, which routinely fabricates low-
loss gratings of a size suitable for applications in future LIGO detectors, to characterize their
gratings and understand the limitations of their fabrication process;

Extend the grating-based optical lever to use two symmetric diffracted beams for better noise
rejection, and develop a more general diffractive element to allow high-sensitivity sensing of
2 degrees of angular motion; Fabricate gratings for testing LIGO local optical sensor
concepts, using on-campus facilities in micro- and nano-scale research facilities.

U. Washington

For 2005-2006 the main progress of the University of Washington Quantum System
Engineering Group (UWQSEG) Group is the appearance of John Sidles' and Daniel Sigg's
article:
Article title: Optical Torques in Suspended Fabry-Perot Interferometers Journal title: Physics
Letters A First author: Dr. John A Sidles Corresponding author: Dr. Daniel Sigg Citation
information: Vol 354/3 pp 167-172 . The dynamical analysis developed in the Sidles/Sigg
article has greatly assisted the mission-critical mirror curvature design of AdvLIGO.

Also achieved in 2006 is public-domain access of the UW QSE Group's digital emulation
http://staff.washington.edu/jon/gr-osx/gr-osx.html

We anticipate that the UW QSE Group's open-source software/hardware will provide a
hardware-in-the-loop (HWIL) emulation capability that will substantially speed the
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

Plans

Our plan is to demonstrate real-time hardware-in-the-loop (HWIL) emulation of the dynamic
torsional instabilities of the Sidles/Sigg article. To this end, we will seek SBIR and STTR
funding from the NSF. Our longer-term broad objective is to demonstrate that the GNU
Radio/USRP platform is well-suited to provide voltage-level real-time emulation of the non-
Markovian dynamical instabilities, both torsional and longitudinal, that will be encountered
during the commissioning phase of advLIGO. As is generally true of HWIL engineering
dynamical systems will greatly improve the speed and reliability of advLIGO
commissioning.

LIGO Lab
40-Meter Accomplishments over Past Year
The 40 Meter Laboratory is developing a prototype of the Advanced LIGO optical
configuration and controls. The construction phase is complete and a first set of lock
acquisition milestones has been reached. Active participants presently are Ben Abbott, Rana
Adhikari, Rolf Bork, Dan Busby, Keisuke Goda, Jay Heefner, Alex Ivanov, Osamu
Miyakawa, Bob Taylor, Monica Varvella, Steve Vass, Sam Waldman, Rob Ward, and Alan
Weinstein, with many important contributions from visitors and the LIGO Laboratory staff.
Visitors include: Seiji Kawamura and Fumkio Kawazoe from NAOJ, Noriyasu Nakagawa
from TAMA, David Blair and Ju Li from UWA Perth, Shally Sharaf from University of
Rochester, Matt Evans and Valera Frolov from LIGO, Kentaro Somiya from AEI. Several
SURF/REU students participate each summer.
In the past year (since Fall 2005), accomplishments include the following:
     We achieved full lock of the interferometer in the Advanced LIGO configuration
(power-recycled Fabry-Perot Michelson with detuned signal cavity). Making use of both RF
and DC-derived length-sensing signals and analog and digital Common Mode servos, and
exploited the power and flexibility of our digital control system. We automated and
optimized the alignment and lock acquisition procedures, so that they became routine.
     With full lock, we have measured the optical response to differential and common-
mode arm motion at a variety of laser power levels and arm cavity offsets. The optical
responses are in excellent agreement with predictions based on the theory developed by
Buonanno and Chen and with detailed numerical calculations. The optical response is
dominated by a high frequency (~4 kHz) RSE peak and a low frequency (~40 Hz) optical
spring resonance, both due to the detuned resonant signal extraction cavity. The noise
spectrum, however, was dominated by technical noise sources so that the expected quantum-
limited sensing noise spectrum was not observed. These results are now published in Phys.
Rev. D74, 022001 (2006).
     The control scheme that we have employed makes use of signals derived from two RF
sidebands, at 33 MHz and 166 MHz, in order to cleanly separate the signals that sense the
arm length changes from those that sense the Michelson and recycling cavity length changes.
In order for this to work, we must avoid the generation of sidebands-on-sidebands; so the RF
sidebands must be applied using electro-optic modulators arranged in parallel, in a Mach-
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

Zehnder interferometer configuration. A paper describing this issue and our solution for it
has been accepted for publication in the journal Classical and Quantum Gravity.
     We have made significant improvements to our infrastructure, electronics, electro-
optics, and diagnostics systems. Continued development, optimization and automation of the
lock acquisition and control, calibration and characterization of the response, and noise
reduction are now proceeding rapidly. With the interferometer in various locked
configurations, we can measure fully calibrated displacement noise spectra and are
developing a comprehensive noise budget analysis, to guide the noise reduction efforts.
     To ensure that these procedures are directly relevant for Advanced LIGO, we continue
to pursue detailed simulations of the dynamics of both the 40-Meter and Advanced LIGO
interferometers. We are collaborating with the LIGO e2e group and the VIRGO Orsay group
to develop a detailed time-domain model simulation of Advanced LIGO and 40-Meter
interferometers, as well as of Advanced VIRGO.
     A major effort during calendar 2006 has been the implementation of a DC detection
system at the asymmetric port of the interferometer. We expect to install the in-vacuum
hardware by the end of summer 2006, implement and commission the new control
electronics at that time, and begin experiments with the new system.
     Beginning in Spring 2006, Keisuke Goda from MIT has been at the 40m lab,
implementing a vacuum squeezing apparatus for injection into the asymmetric port of the
interferometer. The goal is to observe a reduction in the quantum-limited displacement noise
in some carefully-selected frequency range. As of this summer, the in-air beamline has been
assembled. This includes a 2W pickoff beam from our main laser, feeding a second-harmonic
generator (SHG) and an optical parametric oscillator (OPO). The injection into the
asymmetric port will require one fixed and one movable in-vacuum mirror, to be installed
along with the DC readout beamline at the end of the summer.
     The 40-Meter team continues to work closely with the LIGO Controls group, the
Advanced LIGO Interferometer Sensing and Control group, the Intermediate LIGO group,
the LIGO e2e simulation group, and LIGO Laboratory engineers and management. Graduate
students, REU (Research Experiences for Undergraduates) summer students, visiting
students, and visiting scientists have contributed to all aspects of the project over the last
seven years. In particular, REU students have made major contributions to design of the main
interferometer optical plant and the length and alignment control systems, to the
configuration and commissioning of the pre-stabilized laser, digital suspension controllers,
suspended-mass input mode cleaner, and optical lever alignment sensing systems, the
simulation of lock acquisition and dynamics of the dual-recycled interferometers with e2e,
and the development of a displacement noise budget. We will continue to involve students
and visitors with all aspects of the project and its goals. The laboratory continues to be a
popular tour site for local students, journalists, scientific visitors, and dignitaries.

40-Meter Near-term plans
During FY 2007 in the Caltech 40-Meter Interferometer Laboratory we plan to:
    continue the development and optimization of a variety of robust and automated
procedures for acquiring lock and controlling our prototype power- and signal-recycled
Fabry-Perot Michelson interferometer;
    ensure through simulation that the procedures will work effectively for Advanced
LIGO;
    study the noise spectrum and the various contributions to the noise, and implement
methods to reduce the technical noise so as to expose the quantum-limited noise;
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

     install, commission, and fully test a prototype homodyne detection system, including
in-vacuum output mode cleaner, DC photodetector, and new-generation PCIX-based
controls;
     explore length sensing schemes which avoid the use of high-frequency RF sidebands
(since RF electronics at frequencies much above 100 MHz present a variety of technical
difficulties), and develop tests for these schemes using the 40m interferometer;
     inject squeezed vacuum into the asymmetric port of the interferometer and measure its
effect on the interferometer response;
     use our input mode cleaner to study the angular instabilities that may be important for
interferometer sensing and control technologies and architectures, suspension point
interferometers.

LASTI Facility (Ottaway and Mason)
In April 2006 the Quad Controls Prototype has been installed into the BSC on a stiff (no
isolation) table at LASTI. Since then the system identification(sys-id) has been completed,
simple damping loops constructed, the electrostatic drive installed, and eddy current damping
installed onto the reaction chain.
The two stage Internal Seismic Isolation System (ISI) is currently undergoing construction in
the LASTI high-bay. Except for the sensors, actuators and springs the entire structure is
assembled with only minor problems. The actuators and GS13 pods been fit checked. We are
currently working on installing the springs and suspending the system at a reduced weight
load to compensate for the softer than expected springs. At the same time we are redoing the
spring FEA analysis to try and understand the spring constant and to see what the stress will
be if we try these springs at the full load.
Significant progress has been made on understanding improved control strategies for the
triple pendulums. Experiment demonstrations of previously theoretically predicted results
have been achieved. These results were obtained by forming an optical cavity between the
bottom mass of the triple under study and another that acted as an inertial reference. In
addition significant changes to the control dynamics of triple pendulums were observed and
understood when the bottom mass of the test mass was locked to an inertial reference.
The LASTI team is preparing to receive the HAM SAS seismic isolation system in late
November. Preparation has been ongoing to ensure that the LASTI infrastructure is ready to
receive this and conduct the tests on this system in a timely manner. An undergraduate
summer student (Cassandra Hunt) has been developing a Finite Element Analysis model of
the HAM SAS and a triple pendulum based on the original work of Dennis Coyne. Once
completed, this model will be used to guide the testing program for HAM SAS at LASTI.
The MIT Quantum Controls group has been using the LASTI infrastructure to successfully
perform experiments on suspended cavities with high stored optical power to test mass ratios.
In this regime radiation pressure effects become dominant driver of the test-mass dynamics.
A more detailed description of these experiments appears in other parts of this report.

Plans for 2007:

In the following twelve months we plan to do the following things at LASTI:

1.    Continue to characterize the quadruple controls prototype, develop optimum control
strategies and provide feedback to the suspension group on suggested design improvements.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

2.    Install the fully vacuum compatible ISI platform and noise prototype into the LASTI
BSC chamber and perform a full system identification. In addition to this we will implement
the controls strategies that are currently being developed by Stanford and confirm the
isolation performance. These tests are critical to ensure the smooth implementation of
3.    Install the HAM SAS system into the LASTI vacuum system and fully characterize the
system with the aim of informing the HAM Chamber seismic isolation down select team by
mid March 2007.
4.    Continue with investigation of radiation pressure dominated suspended cavities. This
work will continue to done by the MIT Quantum Control group.

Modeling and Simulation (Yamamoto)
Time domain simulation
We completed the module simulating a dual recycled Michelson (DRM) cavity based on the
scalar field approximation. Using this module, we have investigated the 40-meter system in-
lock state length sensing and control system (LSC), comparing model results with the
performance of the actual system. The simulation properly demonstrated the optical
resonance and optical spring resonance observed in the 40-meter experiment.
We have calculated the sensitivity of the Advanced LIGO optical configuration focusing on
the quantum noise effect, and compared model results with theoretical calculations.
Although the simulation only approximates quantum noise, the two approaches agreed
reasonably well and can be used to address the feasibility of using an RF demodulation
scheme.
We have prepared a simulation of an arm of the Advanced LIGO configuration and used it to
study length (LSC) and alignment sensing and control (ASC). We demonstrated that the lock
acquisition is possible with actuation forces whose strength is limited by the noise induced by
LSC. We also investigated the alignment control and have concluded that pitch and yaw of
input and end test masses can be controlled with the current design of sensing and actuation
system. These studies do not include some details, and further investigation is needed with
those details included to quantify the performance and the Advanced LIGO arm with realistic
LSC and ASC

Static Interferometer Simulation
We began the development of a new model to calculate static fields in the core optics system
to study the effects of various configurations and parameters (finite aperture and thickness of
mirrors, surface aberration, tolerance of mirror curvatures, thermal lensing effects). The
primary goal is to provide a modular flexible system so that various different proposed
systems with different imperfections can be studied. The first application will be to define
the tolerance of the detector to the radius of curvature of the mirror, which will be used to
define the polishing requirement for the Advanced LIGO test mass.
A simulation of a Fabry-Perot arm is ready, and it is used to study the requirement of the
mirror surface aberration and the dynamics of the Parametric Instability.

Work Planned for FY2007
The following Time Domain Simulation effort is planned for FY 2007:
     Formulate and implement the code for the fast simulation of a dual recycled Michelson
(DRM) cavity using a modal model.
     Alignment control and stability study of the Advanced LIGO system with strong
     More rigorous implementation of the quantum noise in the model.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

The following Static Interferometer Simulation work is planned in FY 2007:
     Complete the first version of the static field calculation tool.
     Provide the tolerance of the radius of curvature requirement.
     Use the simulation to study the thermal lensing effect and the parametric instability.
     Detailed study of the Parametric Instability and investigate how to minimize the risk of
the problem.

Quantum Sensing (Mavalvala)
The Quantum Measurement (QM) group has identified generation of squeezed vacuum states
and their use in sub-quantum noise limited (QNL) interferometry as a leading candidate for
improvements to next-generation gravitational-wave (GW) interferometers. Squeezed states
of light reduce the noise in one quadrature at the expense of additional noise in the
orthogonal quadrature. When squeezed vacuum states are injected into the antisymmetric
port of an interferometer, sub-QNL performance can be achieved. We have developed two
experimental approaches to generation of squeezed vacuum states that are appropriate for use
in GW interferometers.
Nonlinear optical media
(2)      (3)
nonlinearities in optical materials can be used to correlate the quadratures
of the electromagnetic field. Optical parametric amplification (OPA) or oscillation (OPO)
gives rise to correlations between the quadratures of the optical field, hence the fields are
squeezed. When the input light (the seed field) is quantum noise limited, the correlated
photons created can have noise below the QNL in one quadrature, at the expense of increased
noise in the orthogonal quadrature. When the device is seeded with vacuum (replacing the
coherent seed field by a beam block), the vacuum modes are squeezed as well.
Following the techniques used by our collaborators at the Australian National University, we
are generating about 3 dB of squeezing in the kilohertz band, and we have also implemented
techniques to lock the OPO and homodyne angle simultaneously in our second-generation
squeezer. A setup using new 5 percent MgO-doped LiNbO3 crystals (with much lower losses
and photo refraction) was designed and tested at MIT and is now being implemented on the
40-meter interferometer. Testing squeezed state injection in the 40-meter prototype, which
uses the Advanced LIGO resonant sideband extraction technique, is the main activity on this
experiment over the next year.
The MIT group, in collaboration with the ANU and Hannover groups, has also procured new
OPO/A crystals made of PPKTP, which allows for much lower pump power, and has been
shown to give larger amounts of squeezing than MgO-LiNbO3. These are being tested both at
MIT and at the 40m.
An alternative and novel method for achieving squeezing is based on quantum correlations
created by back-action in a high-power, low-mass suspended mirror interferometer. This
technique relies on the fundamental quantum mechanics of the shot noise and radiation-
pressure noise correlations in an opto-mechanical oscillator system. We are carrying out an
experiment to directly measure squeezing that arises from these radiation-pressure and shot
noise correlations. Broadband squeezing from optical-mechanical correlations in the audio
frequency band will be a first of its kind measurement, if we succeed.
This experiment serves the dual purpose of generating low-frequency vacuum squeezing,
which is one of the major requirements for squeezed light injection in next-generation GW
interferometers, as well as providing a test bed for exploring the fundamental physics of
quantum correlations and quantum decoherence due to optical-mechanical couplings in a
macroscopic mechanical oscillator system.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

We have developed a complete theory and noise budget for this experiment. A paper on the
mathematical formulation of the theory appeared in Phys. Rev. A 72, 013818 (2005), and an
experimental description paper appeared in Phys. Rev. A 73, 023801 (2006).
The radiation pressure experiment is being conducted in one of the LASTI HAM chambers
using the existing PSL and isolation stacks. It is being carried out in three phases with
growing complexity. In the Phase I cavity, with modestly high finesse (~1000) and two 250 g
mirrors suspended as 1 Hz pendula, we achieved an optical spring with the largest ever
spring constant, and also observed a radiation pressure induced instability due to phonon-
photon coupling. These results are to appear in Phys. Rev. A 74 (in press). Phase II is now
underway, using a cavity with one 250 g and one 1 g mirror oscillator, and a finesse of
~10000. In this phase we have achieved a factor of 100 larger optical rigidity (making our
optical beam ‗stiffer‘ than any known material), and have indeed reached the rigidity needed
to observe radiation pressure induced squeezing. Installation of the third phase should begin
in Fall 2006, where two identical Phase II cavities will be used to make a differential
measurement (to remove classical noise).
In addition to serving as a source of squeezed vacuum that could be injected into future GW
interferometers, this experiment also has the potential to be an excellent test bed for reaching
the quantum ground state of the 1 g mirror, which is a mass regime that is many orders of
magnitude larger than other competitive experiments with the same goal – of seeing the
quantum superposition of states of a macroscopic object. In fact, in Phase II we have
achieved optical cooling of the 1 gm mirror that is promising for reaching the quantum
ground state in the future.
Installation of Phase III, observation of squeezing, and, eventually, of a quantum ground
state, are expected to be the main activities over the next two years.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

Lasers

Aciga

Installed an improved injection-locking servo electronics for 10W laser at Gingin HPTF.
Completed development of anauto-locker and installed it at Gingin HPTF.
Realigned 10W laser at Gingin HPTF.
Increased the multimode output power of the Adelaide high power laser to 104W.
People: David Hosken, Damien Mudge, Peter Veitch, Jesper Munch

Plans

a) Laser Development
• Investigate the use of an unstable resonator with the Adelaide 100W laser
• Injection lock the 100W laser.
• Develop injection-locked high power laser for Gingin HPTF, using laser diode stacks and
ceramic YAG gain medium.
• Install a high power laser at Gingin.
• Measure intensity and frequency noise of injection-locked high power laser.
• Investigate options for increasing the output power of the high power laser to 200W.
• Collaborate with LDG to develop a 180W high power laser for Advanced LIGO

GEO
The GEO group is taking the leading role in the Advanced LIGO pre-stabilized laser system
(PSL) development. This includes overseeing the laser development at the Laser Zentrum
Hannover (LZH) as well as the development of the power and frequency stabilization feed-
back control loops and the spatial control of the beam profile. The AdvLIGO PSL work was
presented at NSF reviews and the AEI worked together with LIGO to prepare the cost and
schedule information for the NSF AdvLIGO baseline review.
During the period August 15, 2005 to August 15, 2006, the GEO600 group worked on the
a) Work on the development of the Advanced LIGO laser system continued at the
LZH. During this MOU period there were a number of problems that slowed down the laser
development. After a first demonstration of the output power in the 180W range a new
oscillator design was chosen. This new design does not need a hard aperture in the laser
cavity but instead uses the degeneration of the stability regions of the different laser modes
caused by the astigmatism in thermal lens. During the change of the resonator topology new
laser crystals, lenses and mirrors were installed. After a long tiresome process many of these
components were found not to fulfill their specifications. The mirrors for example introduced
a different phase shift for s and p-polarized light. The undoped end caps of the crystals
showed too high absorption and the lenses were delivered with the wrong power. The lesson
learned is, that a careful acceptance test has to be passed by each component before
installation in the system.
Finally the 180W output power level could be achieved again and the concept of mode
selection due to unstable higher modes could be demonstrated. Due to these delays the spatial
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

beamshape could not be measured in time to design the AdvLIGO premodecleaner during
this period as it was planned. (report: LIGO-DCC P050057-00-R, LIGO-DCC P040053-00-
R)
The design of the engineering model of the diagnostic board was finished and the
engineering prototype is expected to be ready for test in fall this year. (LSC talk LIGO-DCC
G060042-00-Z)
A second change in the systems design was to move from a 12W injection locked oscillator
to a four head 35W Nd:YVO amplifier system as the front end. Such a system has the
advantage that it can be tested and used in the enhanced LIGO experiment. The amplifier
system was tested and it could be shown that the Nd:YAG line lies well in the amplification
region. (LSC talks LIGO-DCC G060113-00-Z)
b) Good progress was made in the power stabilization experiment. Shot noise limited
performance corresponding to a detected photocurrent of 80mA could be demonstrated down
to 100Hz. The source of the excess noise at a level of 5E-9/sqrt(Hz) at 10Hz was caused by
noise intrinsic in the photodiodes that can not be seen in the dark noise but rather seems to be
correlated to the large photo current that flows in the device. A balanced detection
experiment in vacuum was set up to measure this excess noise for diodes from different
vendors.
(Seifert F. et al, Opt. Lett. 31, No. 13 (2006) pp2000-2002 / LSC talks LIGO-DCC G060115-
00-Z)
c) Benno Willke chaired the Lasers Working Group of the LSC.
d) Coating thermal noise (see OPT)

In addition the GEO600 group continued the investigations of cross-coupling between the
different stabilization loops of the GEO600 PSL and long-term test of this system at the GEO
site. During this periode we found, that one fof the laser diodes of the 12W oscillator caused
frequent glitches in the laser power which coupled into the interferometer channels. After
replacing the laser diode and some subsequent system alignment and optimization the
glitches were gone.

Plans
a) Laser Development
The GEO group is taking the leading role in the Advanced LIGO pre-stabilized laser system
(PSL) development. This includes overseeing the laser development at the Laser Zentrum
Hannover (LZH) as well as the development of the power and frequency stabilization feed-
back control loops and spatial control of the laser beam. The GEO group will not only
provide the manpower but as well the money needed to buy equipment for the laser
development and fabrication.
In addition the GEO group will work in the LDG on techniques towards pre-stabilized laser
systems for third generation detectors. Specifically we will work on injection-locking studies
and high-sensitivity power-fluctuations measurements with photodiodes and other techniques
to allow measurements at the 1E-9/sqrt(Hz) level. The GEO group is responsible for the
operation and maintenance of the PSL of the one of the LSC interferometers (GEO600).
(i) Work on the development of the Advanced LIGO laser system will continue at the LZH.
During this MOU period the free-running noise and the spatial beam profile of a 200W
laboratory prototype system will be measured. Based on the results an AdvLIGO pre-
modecleaner will be designed. Stabilization experiments of this system will continue toward
a pre-stabilized laser system that comes close to the requirements for AdvLIGO. Based on
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

the results of the 200W laser characterization PSL systems simulation will start with the goal
to define requirements for sensors, actuators and the filters for the final control loop design.
(ii) The GEO group will provide the laser for Enhanced LIGO. During this MOU period we
will work with LZH and LIGO to bring the amplifier from the brass board to the engineering
prototype and to the reference system level. After a final check the first laser will be
delivered to LIGO.
(iii) One of the main challenges for the AdLIGO PSL will be the power stabilization.
Experiments will continue to analyze and reduce the excess noise sources between in-loop
and out-of-loop measurements.

LSU

Rupal Amin and Joe Giaime from LSU have been cooperating with MIT-LIGO to study a
method to increase interferometer injected power by roughly three fold. This method
involves insertion of an optical amplifier downstream of the interferometers' main lasers.
Although simple in concept, the consequence of increasing injected power may result in new
or at least enhanced noise sources on the pre-stabilized laser table. These effects need study
and schemes to mitigate them.
In the past year, LSU purchased a commercially available optical amplifier, and Rupal Amin
in collaboration with MIT staff tested this device at the MIT LIGO Advanced System Test
Interferometer (LASTI) lab. These tests focused on the preliminary characterization of the
amplifier including pointing stability, thermal dependence of gain, depolarization, and testing
of an in-house computer model of the amplifier output. Measurements at LASTI provided
satisfactory results for the latter three items. Results for pointing stability indicated poor
performance; a study is presently underway to determine source of these pointing
fluctuations.

Plans

a) Laser Development
Rupal Amin plans to carry out his Ph.D. research at LIGO Livingston on the topic of
reducing LIGO's noise floor by employing higher laser power, will collaborate with David
Ottaway and others at MIT, and in cooperation with the LSC lasers technical working group,
to test this amplifier as a candidate for an incremental power upgrade. The laser amplifier is
now installed in LLO's optics lab and is being used with the spare MOPA for continued
testing.

Stanford

Progress Report For the period August 2005 to August 2006 Lasers and Optics (R. Byer, M.
Fejer, K.X. Sun,, B. Lantz, P. Lu, S. Saraf, S. Sinha, K. Urbanek)

a) Scaling of the Nd:YAG slab MOPA laser system to higher powers In this period, we
have worked to make the Nd:YAG MOPA system more reliable. To this end, we have taken
the following steps: 1) We have removed one of the rod amplifiers to reduce the total number
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

of amplifier stages. 2) We are passing the 30 W from our last rod amplifier through a low-
finesse mode-cleaner so that pointing stability and beam quality can be guaranteed for the
end-pumped power amplifiers further downstream. These steps have resulted in a reduction
of incident power on the first end-pumped slab in our system from 40 W to 23 W. However,
we believe that the increased reliability is more important. Furthermore, the spot size of our
saturating beam indicates that we need only 10 W to fully saturate the gain of the slab. Thus,
23 W should be more than enough power to extract all the power available from the first slab,
the output of which is currently 84 W. The second end-pumped slab was initially pumped by
fiber-coupled 808 nm LIMO diodes. With the exception of the alteration of the spot size from
the fiber-coupled pump diodes, the pumping geometry for this second end-pumped slab was
identical to that of the first. During initial characterization of the second slab's extraction
efficiency (35 W from a single pass with 320 W of incident pump power), optical damage
occurred at the interface between the central gain-section of the slab and the non-doped input
end due to excess thermal dissipation. As this was the last slab from the most recent
production run in our inventory, it will become necessary to design and fabricate a new batch
of slabs should we decide to continue with this particular experiment. Currently, we are
pursuing an alternate strategy that takes advantage of a few remaining slabs which were
fabricated previously, without incorporation of features for parasitic suppression. Our near-
term plan is to evaluate one of these as an amplifier stage in a side-by-side comparison,
assuming that we can get a saturating beam of around 100 W from one of the fiber amplifier
stages that are described in the next section. By characterizing the mode quality produced by
the two types of slab amplifiers (one designed with parasitic suppression in mind and one
without parasitic suppression), we will determine the effect of the design modifications
employed. Our goals are to attain a high level of reliability and better knowledge of the
saturating beam's Gaussian parameters. This information will also be important for our
proposed work on the use of super-Gaussian beams to maximize power extraction. This last
task has been delayed for much of the past year because the laser has been used heavily for
the thermally-loaded table-top (mode-cleaner) interferometer experiments described in the
Advanced Detector Configurations Development Group section.
We have published two papers on this work.

1) Arun Kumar Sridharan, Shailendhar Saraf, Supriyo Sinha, and Robert L. Byer, "Zigzag
slabs for solid- state laser amplifiers: batch fabrication and parasitic oscillation suppression",
Applied Optics 45, 3340-3351 (2006).

2) Shally Saraf, Karel Urbanek, Robert L. Byer and Peter King, "Quantum noise
measurements in a continuous-wave laser-diode-pumped Nd:YAG saturated amplifier",
Optics Letters 30, 1195-1197 (2005).

b) Diffractive optical elements for wavefront shaping in slab amplifiers Last year, we
fabricated diffractive optical elements (DOEs) which convert round 1.06 um Gaussian
beams into square, 7th-order super-Gaussians. We planned to use these super-Gaussian
beams to maximize extraction from end- and edge-pumped slab amplifiers. Also, these
super-Gaussians will be useful for efficient coupling to optical cavities that resonate with
a flat-top mode. The DOEs were fabricated by lithographically patterning photoresist-
coated fused silica wafers, reflowing the photoresist with acetone vapor, and
subsequently transferring the patterns into the underlying fused silica wafer using
reactive ion etching. The elements measure approximately one millimeter laterally and
are one micron tall. They were designed to operate transmissively and to accommodate
700 um diameter beams at a wavelength of 1.06 um. We measured the profiles of the
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

DOEs with a Zygo white-light interferometer, and then simulated the diffraction of light
transmitted through these profiles. The DOEs were then illuminated with 30 W of
TEM00 light and the output beam profiles were measured at various propagation
distances. The experimental results were in very close agreement with the simulated
results, with slight differences being attributed to thermal lensing. The ripple on the flat
top beam was within 15% rms. We found that a collimated flat-top beam could be
produced by illuminating the DOEs with a slightly diverging Gaussian beam. The size of
the resulting super-Gaussian was compatible with propagation through slab amplifiers
measuring one mm by one mm in cross-section. Patrick Lu has spent most of the past
year training at LLNL to carry out the development of optical grating technology, as
discussed in the Advanced Configurations Section. He plans to continue the work on
DOEs once he completes the training phase of the LLNL collaboration.
c) Scaling of fiber MOPA to 160 W In addition to increasing the output power of the slab
MOPA system, we are also working to determine if the high efficiencies and good beam
quality promised by high power fiber amplifiers can meet the LIGO requirements. Our
goal is to demonstrate a 200-W class fiber amplifier with greater than 80% of the output
power is in the fundamental mode, and to determine if the amplified signal meets the
noise requirements for future gravitational wave detectors. Fiber amplifiers are very
attractive for gravitational wave detection because their small mode areas (and
consequently low saturation powers) could allow single stage amplification from 1 W of
input power to 200 W of output power, potentially simplifying and increasing the
reliability of the MOPA system. Furthermore, the high efficiency of fiber systems means
that there should be less heat generated (which would allow for more compact cooling
geometries) and less total pump power required (which would extend the lifetime of the
pump laser diodes). In the past six months, we have focused on increasing the stability
and reliability of our 140-W fiber MOPA system.

We are investigating several different approaches which, in conjunction, should produce a
source that meets the power and noise requirements required for future gravitational wave
detector laser sources. We are looking at ytterbium-doped double-clad gain fibers from
Nufern that will minimize mode-coupling between the guided core modes. This should result
in decreased intensity noise, improved polarization extinction ratio and better pointing
stability. The mode-coupling in a typical low NA core, large mode area (LMA) overmoded
fiber is a function of the smoothness of the refractive index profile in the core, the core-inner
cladding roughness, and induced micro-bending of the fiber. These parameters are highly
dependent on the manufacturing process; therefore, the fibers must be evaluated
experimentally to judge how easily the fundamental mode couples to the higher order modes.
We purchased sets of fiber from an alternate fiber vendor, Liekki, to test how their fibers
compare with those that we have been using for the past m core two years. Using a length of
one of the new vendor's fibers with a 20 m cladding (referred to as 20/400), we have obtained
promising results. and 400 We obtained over 50W of polarized output power at which point
our fiber cracked. The fiber is thought to have cracked due to mechanical stress and not due
to thermal problems since there was no evidence of overheating seen in the jacket. In
subsequent efforts, we will spool the fiber with a larger radius. Since, at this stage, we were
primarily interested in determining the mode coupling co-efficient of the gain fiber, we
purchased a significantly smaller length of fiber than the optimum to reduce costs. The
polarization extinction ratio (PER) of the output was excellent and was characterized to be
over 150:1 at the highest output power. Furthermore, the mode quality of the amplifier output
was very good and the mode profile was very stable. We are currently exploring the m
cladding. This core and a 250 use of another fiber from Liekki with a 30 fiber has
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

significantly higher pump absorption per unit length than the 20/400 which will allow shorter
fiber lengths. In addition to examining different gain fibers, we have also been looking at
ways to integrate the fiber amplifier system, which should enhance the overall system
reliability and could decrease its long term maintenance costs. To this end, we have been
looking at ways to launch the signal power into a single mode fiber and then tapering the
single mode fiber to the large mode area double clad fiber. In a future configuration, a fiber
oscillator could then excite the fiber amplifier without requiring free-space optics. To test the
tapered fibers, we launched light from a non-planar ring oscillator (NPRO) through a
polarizer into a polarization-maintaining (PM) single mode fiber. The amount of power
launched into the single mode PM fiber was controlled by a half-wave waveplate. The single
mode fiber ensures that the free-space beam is not launched into higher order modes. The
single mode fiber is then "adiabatically tapered" to a passive double clad PM fiber with the
same geometry as the gain fiber. We have expended considerable effort in developing this
technology during this last period using our fiber splicer to produce ~600 micron long,
tapered splices that approximate a true adiabatic taper. With this in-house splicing technique,
we have been able to taper the single mode fiber to an undoped 30 um large mode area fiber
in two stages and obtain a 23 dB polarization extinction ratio with good beam quality and
negligible power degradation. We are now working on applying this tapering technique to a
doped, large mode area fiber.
d) Investigation of phosphate fiber laser systems We have also begun to explore the use of
Yb-doped phosphate fibers as high power pre-amplifiers or power amplifiers for the LIGO
laser system. Phosphate glasses offer higher solubilities for rare earth oxides; therefore,
phosphate fibers can be more highly doped than silicate fibers, resulting in shorter fiber
lengths. During this reporting period, we demonstrated a 20-W phosphate laser at 1.06 um
with M2 < 1.1.
e) Investigation of silicate bonding for use in slab and fiber laser systems. We have
continued to investigate silicate bonding in this period. We have bonded nearly 100 fused
silica samples with various solution concentrations and volumes with a range of heat
treatments, and we have begun characterizing these bonds.

Photodetector Development (Z Rao, D. Jackrel, K.X. Sun, J. Harris, M. Fejer)
f) Photodiode development Advanced LIGO photodiode development has continued with
an effort to create high efficiency back-illuminated devices which can handle high power.
InGaAs devices and GaInNAs(Sb) devices were grown using MBE and fabricated in the
Stanford Nanofabrication Facility (SNF), and characterized using the 700 mW Nd:YAG laser
setup at Stanford. Both the InGaAs and GaInNAs(Sb) diodes efficiently absorbed laser
powers greater than 300 mW before saturating, and the external efficiencies of the back
illuminated devices at these power levels were found to be roughly 55% and 40%
respectively. Much of this loss was found to come from the fact that more than 30% of the
light was parasitically absorbed due to free-carrier absorption in the thick highly doped
substrate. In order to improve the external efficiency, the highly doped GaAs substrate must
be thinned to avoid large parasitic absorption in this layer. After thinning the substrates m the
efficiencies rose to 75% and 62% for the InGaAs and to roughly 150 GaInNAs devices,
respectively. However, parasitic absorption in the substrate is still limiting the quantum
efficiency of both devices. The response linearity of both types of devices was very good
(R2>0.995) prior to saturation, and similar devices showed a 3-dB bandwidth at megahertz
frequencies when tested in m were the RF laser set-up at MIT. Efforts to thin the substrate
to below 100 not successful due to a failure of the etch-stop layer, which protects the device
layers from the substrate removal etchant. The previous samples were grown with a AlGaAs
etching-stop layer with 50% Al, which doesn't provide high enough etching selectivity from
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

GaAs substrate. The epitaxial structure will be redesigned to incorporate a more effective
etch-stop layer in future devices, such as an AlGaAs layer with Al concentration up to 90%.
We have been in renewed discussion with LIGO scientists since the requirements on the
photodiodes for Advanced LIGO have changed. The new requirements on the maximum
detected power and frequency response pose significant challenges on the photodiodes. We
are developing the capability to do RF measurements to characterize our current devices and
to help identify potential improvements we can make on future device design, processing and
packaging.

g) Photodiode development for length sensing and wavefront sensing: For either the
LIGO enhancement or Advanced LIGO, it may be desirable to have high-bandwidth, high-
power photodiodes in both single and quad forms with low cross-talk for length and
alignment sensing and control. We are currently working on the development of back-
illuminated high power diode chipsets using the InGaAs and InGaAsSb family materials
studied by David Jackrel [1]. Our aim is to develop rf-capable detectors by using proper
package technology to enhance photodiode performance in the 100 MHz region. The
outcome could be the route to the fabrication of rf single- and quad-photodiodes in quantity
within a year. While we remain open to using the custom made back illuminated high power
photodiodes developed at Stanford, we have begun selecting and testing from among the
possible commercial products to accelerate the process. Initial attachment and lead-bonding
tests on our first chips from ROITHNER LASERTECHNIK GmbH in Vienna, Austria were
unsuccessful due to bonding pad failure during the lead-bonding step. We are working with
this manufacturer as well as others to obtain chips with more robust bonding pads. [1] David
Jackrel, "InGaAs and GaInNAs(Sb) 1064 nm Photodetectors and Solar Cells", Ph. D thesis,
Stanford University, December 2005.

h) Optical platform for testing of photodetectors at LHO: An on-site optical
photodetector test platform at one of the LIGO sites would facilitate commissioning,
operation, and calibration. It would also benefit continuing photodetector development. We
have been in discussion with LHO staff on the design of an optics test platform that would
simulate all side-bands at the LIGO interferometer dark port, and excitations for AS_Q and
AS_I components, of which the former is the signal, and the latter needs to be suppressed.
We are in the process of preparing a formal proposal to LIGO. We have also pointed out the
possibility of using multiplexed polarization and current subtraction scheme for AS_I
mitigation.

Plans

a) Laser Development
Power-scaling of MOPA laser systems (R. Byer, K. Urbanek, P. Lu)
Scale the Nd:YAG slab MOPA laser system to higher powers; We will work to compare the
mode quality produced by two different types of zig-zag slab amplifiers, one incorporating
parasitic suppression design features and one not. Our near-term plan is to compare side-by-
side, the two design types. The parasitic suppressing slab is currently being fed by the first
stage of our existing zig-zag slab amplifier at 84 W. The non-parasitic suppressing slab will
be fed by a fiber amplifier stage of the type described in the next section. This task
anticipates that we will be able to obtain a saturation power level close to 100 W from a fiber
amplifier. By characterizing the mode quality produced by the two types of slab amplifiers,
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

we will be able to determine the effects of the parasitic design features and determine the
influence of saturating beam parameters under Gaussian conditions.

We will continue measuring the thermal distortion of bench-top mode-cleaners with high
circulating powers. The mode structure and power coupled into the mode-cleaner are
currently being determined as a function of (increasing) incident power. The results are
subsequently compared with computer models, including theoretical predictions of the mode-
cleaner‘s behavior from MELODY.

We will work on the use of super-Gaussian beams to maximize power extraction from
saturated amplifier stages. Our goal will be todetermine the fidelity with which we can
convert a super-Gaussian back into a TEM00 Gaussian beam.

We will investigate the use of silicate bonding to bond dielectric coated YAG pieces for use
in parasitic mode suppression to improve the small signal gain of an end-pumped slab.

Development of fiber laser systems (R.L. Byer, S. Sinha)
Scale fiber MOPA to 180 W; We will continue work to scale the fiber MOPA output to 180
W with a single stage of amplification while maintaining good beam quality and acceptable
polarization extinction ratio. We expect to demonstrate over 50% optical-to-optical
efficiency. We will continue studies aimed at improving the launching efficiency of the
signal power into the fiber MOPA amplifier.

Characterize fiber MOPA output; We will characterize the intensity noise, phase noise and
pointing stability of this system at the maximum operating power. We will measure the
TEM00 mode content by locking the output to a mode cleaner and we will characterize mode
content fluctuations over time. We also plan to characterize the reliability of the system by
running it continuously for a week to determine if either catastrophic failure or a gradual
reduction in performance is observed (for example from fiber photo-darkening). We will also
determine if thermal fluctuations at the fiber fixtures create pointing instability.

Characterize fused silica silicate bonded samples; We will characterize the Fresnel reflection
off the silicate bond using a modelocked laser. In addition, we will measure the optical
damage threshold and mechanical strength of the bonded samples and observe how these
properties depend on silicate bonding solution concentration and solution volume.

Investigate phosphate fibers for possible applications in Advanced LIGO; Using primarily
ARO (Army Research Office) support, we are planning to build a 25-W single frequency
fiber MOPA using highly doped double clad phosphate fiber. We will investigate which
cooling geometry is most suitable for these fibers with high thermal loading. We will
characterize this system to determine if these fibers can be used for creating compact LIGO
pre-amplifiers and/or scaled to over 100 W for use as power amplifiers.

b) Other Contributions

Development of high-power photodiodes (Z Rao, J. Harris, M. Fejer, K.X. Sun, B. Lantz)
Continue the development of high power photodiodes; The requirements on the photodiodes
for Advanced LIGO have evolved significantly since David Jackrel began working on the
development of rear-illuminated, high current photodiodes in 1999. The current requirements
of 1 Watt detected power and 100 MHz frequency response pose serious challenges. We are
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

setting up the capability to make RF measurements to characterize our current devices and to
help identify potential improvements we can make on future device design, processing and
packaging. We calculate that the maximum capacitance the required photodiode can have is
more than an order of magnitude smaller than the measured capacitance of our current
devices. To reduce the capacitance, we plan to try reducing the doping in the intrinsic layer to
increase the depletion width, which could potentially allow us to increase the thickness of the
intrinsic layer. Also, to further improve the external efficiency of the photodiodes, the
substrate would have to be thinned
below 100 •m. Attempts to achieve such thin structures have failed, however, due to an
ineffective etch-stop layer (the function of which is to protect the device layers from the
etchant used for substrate removal) , i.e., an AlGaAs layer with 50% Al. In the future, the
epitaxial growth procedure could be redesigned to accommodate a more effective etch-stop
layer, such as an AlGaAs layer with Al concentration up to 90%.

We could also attempt to increase the operating bias for these devices. Increased bias would
improve the performance in two ways; first, by increasing the thickness of the depletion
region which lowers the capacitance, and second, by increasing the allowed power per unit
area of diode (by sweeping the free carriers out of the depletion region more quickly, thus
keeping the device in the linear regime up to higher power densities). Increasing the power
density would mean that we could use a smaller device for a given power, and smaller
devices have smaller capacitance. It is not clear yet how far LIGO is prepared to pursue the
development of advanced photodiodes for Advanced LIGO, so we will continue our
discussions with the LIGO lab to help determine the appropriate level of effort for this task.
They could eventually become integrated into Advanced LIGO test systems for length
sensing of auxiliary DOF‘s.

Continue the development of photodiodes for wavefront sensing. We will continue to
develop rf-capable detectors by using appropriate packaging technologies to enhance
photodiode performance up to 100 MHz. We have begun selecting and testing from among
the possible commercial chip suppliers. Our initial goal will be to create and test a 4-quadrant
device by mounting commercial InGaAs chips in our own package. Initial attachment and
lead-bonding tests on the first set of test chips (from ROITHNER LASERTECHNIK GmbH
in Vienna, Austria) were unsuccessful due to bonding pad failure during the lead-bonding
step. We are working with this manufacturer, as well as with other manufacturers, to obtain
chips with more robust bonding pads. Packaging and testing studies will be continued.

LIGO Lab
LASTI Facility (Ottaway)
The feasibility of a commercially available optical amplifier to increase the available power
of the Initial LIGO PSLs for use in Enhanced LIGO was tested at the LASTI facility in late
winter. This approach has become the fall back option in case Laser Zentrum Hanover cannot
deliver its 30 Watt front end to the Advanced LIGO laser on a time scale that is compatible
with Enhanced LIGO. The saturation properties of this optical amplifier were used to validate
the output of a laser modeling code. This will enable us to design with confidence a system
quickly in the future that will meet Enhanced LIGO requirements. In addition to this we
verified the utility of a birefringence compensation technique using a Faraday Rotator and
double passing of a single laser rod. In addition to this, mechanical stability issues of the
laser gain medium were also identified.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

Optics

Aciga

Progress
•     Gingin facility
o Pre-mode cleaner and frequency stabilization installed
The NPRO has been frequency stabilized to a reference cavity supplied by ANU. The pre-
mode cleaner has been added to the optical path and locked to the 10W laser. We can lock
the 80m cavity to the 10W laser together with pre-mode cleaner and reference cavity locked
for further high power experiment.
o Hartmann sensor
To overcome the interference problem encountered in the Gingin facility, the light source of
the Hartmann sensor was change from laser to LED. The new setup for the Hartmann sensor
was installed (by University of Adelaide). Preliminary thermal lensing effect with close to ~1
kW optical power inside the 80m cavity, giving a focal length of 6 km, and the long term
drift were tested. We confirmed that cavity mode change is consistent with the observed
thermal lens. Thermal lensing due to residual absorption in the fused silica compensation was
observed.
o Study of optimising the heating pattern on the compensation plate.
We created numerical tools using finite element modeling to find the optimal heating pattern
on the compensation plate. This is essential to correctly compensate thermal lensing and limit
the generation of higher order optical modes. We detailed a practical way to generate the
heating pattern using a CO2 laser with a binary mask. FFT optical code has been used to
simulate the transmission of the CO2 heating beam and showed the success of this method.
o Modeling of the effect of non-uniform absorption for Gingin cavity
Although there is no absorption map for the sapphire test masses used in Gingin experiment,
we used the data of another sapphire test mass which has highly non uniform absorption (half
of the central part of the TM with ~ 100ppm/cm and the other half with ~ 40ppm/ cm) to
simulate thermal lensing and investigate the beam distortion. Using LIGO FemLab code we
calculated the thermal lensing effect and FFT code to simulate the optical mode shape. We
found that there is a small astigmatism of the thermal lensing with similar pattern to the
absorption map, but the distortion is small (<0.1%). This confirms the experimental results
obtained in Gingin where we did not find that the beam quality was degraded due to thermal
lensing within the measurement sensitivity. .
o Investigating the negative dn/dT material for thermal compensation
We are investigating negative dn/dt material for thermal compensation, first with CaF2. The
problem with CaF2 is the high thermal expansion coefficient that leads to the large surface
thermal deformation (positive lens) and stress birefringence (negative lens but astigmatic).
We also investigated quartz that has negative dn/dt and relative small thermal expansion
coefficient comparing to CaF2. The surface thermal deformation and stress birefringence still
exit but are smaller than that in CaF2. A possible solution to compensate such effect has been
proposed. A detailed report on such issue is under preparation.
o Stage 1 is now close to completion. Due in main to a lack of resources, the high power
laser has not yet been installed, Stage 2 commissioning has not begun and the digital control
system has not been procured.
•     Parametric instability study
o Investigating the high order modes diffraction loss in a cavity
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

In collaboration with Phil Willems and Erika D'Ambrosio in LIGO it was determined a
tensor nature of the diffraction losses in an Advanced LIGO type of cavity. As a consequence
it was shown that the cavity eigenvalues will overestimate the power loss per round trip due
to diffraction losses. It was also shown that the finite size of the test masses that induce the
diffraction losses will have a strong impact in the final shape of the resonant higher order
modes inside the cavity, varying at the same time the frequency and Q values of the resonant
optical modes.
o Modeling of ring damper on the test mass Q-factor and thermal noise performance, taking
into account the coating loss.
o Again due in the main to lack of resources experimental progress has been delayed.
However as listed above excellent progress has
People
UWA: Chunnong Zhao, Li Ju, David Blair, Jerome Degallaaix, Slawek Gras, Pablo Barriga,
Zewu Yan, Yaohui Fan
UA: Aidan Brooks, Peter Veitch, Jesper Munch.

Plans
High optical power tests
At the HOPTF in Gingin W.A., we will
1. Maintain the operation of the Adelaide 10W laser
2. Complete installation of the Hartmann sensor
3. Re Integrate the auto-alignment system
4. Redo test 1 (single 80m cavity; 10 W laser; ITM substrate inside cavity, no compensation
plate
(CP)) using the Hartmann sensor
5. Complete test 2a (single 80m cavity, 10W laser, ITM substrate outside the cavity, no CP)
6. Begin test 2b (single 80m cavity, 50W laser, ITM substrate outside the cavity, no CP)
7. Experimentally explore negative dn/dt material compensation.
ii) Parametric instabilities (UWA)
1. Start testing parametric instability experimentally in Gingin facility
2. Detailed modeling of ring damper on test mass to determine design for reducing
parametric
gain while maintain reasonable noise performance.
iii) Output Modecleaner (ANU)
Work with the mLIGO/AdvLIGO detection group to design an OMC for AdvLIGO. In
particular:
1. estimate noise couplings introduced by an OMC
2. collate with other noise coupling to set the requirements on an OMC
3. commence work on a preliminary design for the AdvLIGO OMC
4. build an experimental in vacuum test rig to shakedown unforeseen issues associated with
propagating the beam from the last MMT through to the photodetectors.
Success in the above programs depends on funding.

Coating Losses (ANU)
(i) Develop a coating free beamsplitter and demonstrate its use in a coating free benchtop
interferometer
(ii) Investigate applicability of the technique with suspended optics
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

CaRT

a.    Light scattering in the LIGO beam tubes and design of the baffles that control it:
Thorne, working with input from Albert Lazzarini (LIGO-T060013-02-E), computed noise
for Advanced LIGO due to light scattering in the LIGO beam tube, assuming the light-
scattering BRDF of initial LIGO mirrors and using the as-built baffle configurations and
beam-tube vibration spectra from Rai Weiss. The mirror BRDF has two components: sharply
forward scattering due to macroscopic undulations of the mirror figure, and isotropic
scattering due to micron-sized particles and defects on the mirror surface. Thorne's
calculations predict that the sharply forward scattering will produce acceptably small noise:
about 1/100 of the Advanced LIGO noise specification at all frequencies, or lower. However
the isotropic scattering is very dangerous for Advanced LIGO at the initial LIGO scattering
level of 70ppm: the danger arises from scattering at large angles, into the mirrors' vacuum
chambers and near beam tube (before the gate valve), and then backscatter off the chamber
walls, tube walls, and objects in the chamber. This noise is so large that it might contribute to
the noise bump in initial LIGO spectra at ~30 Hz. Thorne's calculations and results are
summarized in a 27 May 2006 memo to Lazzarini (available by request), which will be
combined with Lazzarini's input (LIGO-T060013-02-E) in a future LIGO technical report.
Thorne is revising his in-preparation Phys Rev paper on light-scattering noise (co-authored
with Eanna Flanagan) to incorporate the two-component BRDF of the initial LIGO mirrors;
the original draft was confined to the BRDF ~ 1/(scattering angle)A2 exhibited by earlier
mirrors.
b. Optimization of the Degeneracy of the Signal Recycling (SR) Cavity for Advanced
LIGO: Pan completed his modal analysis of the optimal degeneracy for advanced LIGO's
SR cavity, and is writing a paper on it for publication; preliminary results are in LIGO
T060004-00-R, and more complete results in Pan's PhD thesis (May 2006). Pan verified
quantitatively Thorne's rough estimates of the severely unacceptable sensitivity of the highly
degenerate baseline SR cavity design to mirror deformations, and he found that the minimum
(and acceptable) loss of signal to
noise is achieved when the cavity is moderately nondegenerate, with Gouy phase is between
0.2 and 1.3 radians Pan also discovered, in this range, the danger of a forest of resonances of
parasitic light modes, which must be dealt with in the cavity redesign, and he worked on the
development of an FFT simulation code for Advanced LIGO (in collaboration with
HiroYamamoto of LIGO Lab-Caltech), aimed at providing more reliable modeling of the
influence of figure errors on a redesigned SR cavity.
b1. 4km SR Cavity [New project, not in 2005 MOU]: Pan has initiated a modal-analysis
study of a 4km-long SR cavity, as a possible alternative to the unacceptably degenerate
baseline design. His analysis shows that such a cavity can provide the optimal degeneracy.
He is currently investigating alignment sensing and control issues for such a cavity. Pan has
completed his PhD and will continue this work as a postdoc in Buonanno's group at the
University of Maryland.
c.    Influence of SR-Cavity Optical Losses on Noise Performance of Advanced LIGO:
No progress to report.
d. Duality Relations for Optical Cavities with Nearly Concentric and Nearly Flat
Mirrors: Savov, with his collaborators
(Chen at AEI, D'Ambrosio and Agresti at LIGO Lab-Caltech) has written and submitted for
publication a paper deriving the duality relation that Savov discovered numerically: gr-
qc/0511062.
e.    Parametric Instability in Advanced LIGO: Savov found that Mexican Hat mirrors
have somewhat worse parametric-instability performance than spherical mirrors. Thorne
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

showed that for spherical mirrors as well as Mexican Hat mirrors, low-finesse Stokes modes
may be capable of triggering the parametric instability (though low-finesse AntiStokes
modes might compensate the effect, and produce stability). The Braginsky-Vyatchanin
theory of the instability is invalid for low-finesse Stokes and Anti-Stokes modes, so Thorne
devised a computational method for dealing with them and Vyatchanin (MSURG) devised an
alternative method. Savov is in the late stage of
implementing Thorne's method so as to explore the influence of low-finesse modes on the
parametric instability in Advanced LIGO. For a brief summary of these issues and ideas, see
LIGO-G050441-00-Z.

Plans

(seems to be a repetition of the report on work undertaken)

a. Light scattering in the LIGO beam tubes and design of the baffles that control it:
Thorne, working with input from Albert Lazzarini (LIGO-T060013-02-E), computed noise
for Advanced LIGO due to light scattering in the LIGO beam tube, assuming the light-
scattering BRDF of initial LIGO mirrors and using the as-built baffle configurations and
beam-tube vibration spectra from Rai Weiss. The mirror BRDF has two components: sharply
forward scattering due to macroscopic undulations of the mirror figure, and isotropic
scattering due to micron-sized particles and defects on the mirror surface. Thorne's
calculations predict that the sharply forward scattering will produce acceptably small noise:
about 1/100 of the Advanced LIGO noise specification at all frequencies, or lower.

However the isotropic scattering is very dangerous for Advanced LIGO at the initial LIGO
scattering level of 70ppm: the danger arises from scattering at large angles, into the mirrors'
vacuum chambers and near beam tube (before the gate valve), and then backscatter off the
chamber walls, tube walls, and objects in the chamber. This noise is so large that it might
contribute to the noise bump in initial LIGO spectra at ~30 Hz. Thorne's calculations and
results are summarized in a 27 May 2006 memo to Lazzarini (available by request), which
will be combined with Lazzarini's input (LIGOT060013-02-E) in a future LIGO technical
report. Thorne is revising his in-preparation Phys Rev paper on lightscattering noise (co-
authored with Eanna Flanagan) to incorporate the two-component BRDF of the initial LIGO
mirrors; the original draft was confined to the BRDF ~ 1/(scattering angle)^2 exhibited by
earlier mirrors.

b. Optimization of the Degeneracy of the Signal Recycling (SR) Cavity for Advanced
LIGO: Pan completed his modal analysis of the optimal degeneracy for advanced LIGO's
SR cavity, and is writing a paper on it for publication; preliminary results are in LIGO
T060004-00-R, and more complete results in Pan's PhD thesis (May 2006). Pan verified
quantitatively Thorne's rough estimates of the severely unacceptable sensitivity of the highly
degenerate baseline SR cavity design to mirror deformations, and he found that the minimum
(and acceptable) loss of signal to
noise is achieved when the cavity is moderately nondegenerate, with Gouy phase is between

Pan also discovered, in this range, the danger of a forest of resonances of parasitic light
modes, which must be dealt with in the cavity redesign, and he worked on the development
of an FFT simulation code for Advanced LIGO (in collaboration with Hiro Yamamoto of
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

LIGO Lab-Caltech), aimed at providing more reliable modeling of the influence of figure
errors on a redesigned SR cavity.
b1. 4 km SR Cavity [New project, not in 2005 MOU]: Pan has initiated a modal-analysis
study of a 4km-long SR cavity, as a possible alternative to the unacceptably degenerate
baseline design. His analysis shows that such a cavity can provide the optimal degeneracy.
He is currently investigating alignment sensing and control issues for such a cavity. Pan has
completed his PhD and will continue this work as a postdoc in Buonanno's group at the
University of Maryland.

c. Influence of SR-Cavity Optical Losses on Noise Performance of Advanced LIGO: No
progress to report.

d. Duality Relations for Optical Cavities with Nearly Concentric and Nearly Flat
Mirrors: Savov, with his collaborators (Chen at AEI, D'Ambrosio and Agresti at LIGO Lab-
Caltech) has written and submitted for publication a paper deriving the duality relation that
Savov discovered numerically: gr-qc/0511062 .

e. Parametric Instability in Advanced LIGO: Savov found that Mexican Hat mirrors have
somewhat worse parametric-instability performance than spherical mirrors. Thorne showed
that for spherical mirrors as well as Mexican Hat mirrors, low-finesse Stokes modes may be
capable of triggering the parametric instability (though lowfinesse AntiStokes modes might
compensate the effect, and produce stability). The Braginsky-Vyatchanin theory of the
instability is invalid for low-finesse Stokes and Anti-Stokes modes, so Thorne devised a
computational method for dealing with them and Vyatchanin (MSURG) devised an
alternative method. Savov is in the late stage of implementing Thorne's method so as to
explore the influence of low-finesse modes on the parametric instability in Advanced LIGO.
For a brief summary of these issues and ideas, see LIGO-G050441-00-Z.

Embry-Riddle
In this area the activities embarked on differ somewhat from what was originally planned.
Based on the conditions of the PI's successful NSF grant application, the work towards
composite test masses was put on hold while the PI embarked upon research towards the
characterization and improvement of coating thermal noise and coating
thermorefractive noise. Towards this end, the PI equipped a laboratory consisting of two
rooms with most of the standard laboratory test equipment (Optical tables, Optics hardware,
Fast data acquisition, Oscilloscopes, Lock-in amplifier, Spectrum Analyzers both RF and
Audio, Lasers, Power supplies, Electronics workstation, Vacuum system, etc.) The general
laboratory setup is now complete and two experiments are up and running. The first is a Q
measurement system to measure the mechanical loss in the coatings of test samples with the
aim of reducing the thermal noise from the coatings on the Advanced LIGO core optics. The
Q measurement system differs slightly from previous systems within the collaboration as the
readout for the sample's excitation is a Mach-Zender interferometer. The beams in both arms
of the interferometer are transmitted through the sample. Any differential strain in the
sample between the locations interrogated by the two beams, is read out as a differential
phase shift due to strain-induced birefringence in fused silica. Access to both output ports
allows for balanced detection, which reduces greatly reduces the common mode noise.
Undergraduates Steven Zech and Julian Horvath assisted with the setting up of this
experiment. The second experiment consists of a coating test-sample placed in a furnace,
wherein the reflectivity of the coating is measured as a function of temperature from 300K
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

and up to about 700K. The furnace and optics chain are functioning as intended. A dummy
sample is currently in the apparatus to be replaced by the real sample of interest later this
week. Preliminary results from both experiments are anticipated within the next few weeks.

Coordination of discussions on ring dampers to reduce the effects of the Parametric
Instability.

Plans
a) Optics Characterization
As part of the LIGO coatings research effort, ERGWAG will perform measurements of
mechanical loss in coatings under consideration for Advanced LIGO. These measurements
allow us to make estimates of the thermal noise contribution from core optic coatings which
currently dominate the projected noise spectrum of Advanced LIGO at mid-frequencies in
the region of lowest noise. The work will be performed in coordination with the LIGO
Laboratory.

ERGWAG will also make measurements to enable us to estimate the thermorefractive noise
contribution from Advanced LIGO optical coatings – a potentially significant source of noise
in Advanced LIGO. Estimates of the strength of this noise source depend on the index of
refraction variation of the high-index coating material with temperature (dn/dT). This
quantity is not well known, even for the baseline high-index coating material. By measuring
the reflectivity changes of test-coatings at particular wavelengths as the coatings are heated,
an estimate of dn/dT for the high-index material can be obtained.

Florida

The main OPT activities of the UF-LIGO group are focused on the Advanced LIGO
interferometer design and development, specifically i) responsibility for the AdvLIGO Input
Optics, ii) development of high power optical components for the initial LIGO upgrade
(Enhanced LIGO), and iii) development of stable recycling cavity topologies and associated
thermal compensation issues for AdvLIGO. Over the period covered by the prior MOU, the
OPT personnel were Muzammil Arain, Ken Franzen, Joe Gleason, Guido Mueller, Volker
Quetschke, Dave Reitze, Simon Stepuk, David Tanner, Luke Williams, Stacy Wise, and Wan
Wu. Specific OPT-related research for the Input Optics included:
1.    Modulation Development for Advanced LIGO
a.    High Power electro-optic modulator (EOMs) development (Wan Wu, Guido Mueller,
Volker Quetschke, Dave Reitze, David Tanner, Luke Williams) - RTP-based EOMs have
been designed and prototyped for high power operation. We tested electro-optic modulators
using RTP crystals and showed that no significant thermal lensing for powers of up to 100W
was seen. Casings and impedance matching circuits were developed to make these
modulators usable for Advanced LIGO. The EOM design underwent a successful
Preliminary Design Review (PDR) in April 2006.
b.    Complex modulation (Volker Quetschke, Dave Reitze, David Tanner) - We began
modeling and experimental efforts to explore the possibility of imposing several sidebands
on the carrier light without creating intermodulation cross-products based on complex
modulation, i.e. simultaneous modulation of phase and amplitude.
c.    Mach-Zehnder modulation (Stacy Wise, Volker Quetschke, Dave Reitze, David Tanner
- We experimentally demonstrated a stable, quasi-monolithic setup for generating clean
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

sideband spectra and have investigated the magnitude and the respective influence of the
noise sources on the Advanced LIGO configuration. A draft of technical note will be released
shortly. The experiment is undergoing an upgrade to high power operation.
d.    Sideband phase and amplitude noise measurements (Wan Wu, Volker Quetschke,
Guido Mueller) - an experiment to measure the amplitude and phase stability of individual
EOMs has been developed using 2 phase-locked and amplitude-stabilized NRPO lasers. This
experiment has not progressed as fast as we would like and is in the noise-hunting stage; we
are currently limited by beam jitter noise.
e.    We realized that our design which symmetrizes the RF-sidebands puts very stringent
requirements on the losses including modal losses. Its realization depends even more on a
very sophisticated TCS system to improve mode matching and reduce modal losses within
the interferometer. However, detailed analysis hasn't continued because of our involvement
in the TCS design.
2.Faraday isolator development ( Simon Stepuk, Volker Quetschke, Ken Franzen, Dave
Reitze, David Tanner, Luke Williams, and IAP group) – The first prototype of a Faraday
isolator for high power with low thermal lensing was developed and tested. Optical
performance of the isolation and thermal compensation was verified for conditions near
Design Review (PDR) in April 2006. Current efforts directed at designing a vacuum
compatible housing for the LIGO upgrade.
3.Adaptive control of laser modal properties for mode-matching (Muzammil Arain, Joe
Gleason, Guido Mueller, Volker Quetschke, Dave Reitze, David Tanner) - development,
modeling, and table-top testing of an adaptive optical system for the AdvLIGO mode
matching telescope has been developed, using create a dynamic lens using thermal lensing by
heating the optical element with an auxiliary laser. In the first stage of the experiment an
argon-ion laser was used and later the optics were changed to a CO2 laser. A theoretical
model was developed that agrees well with the measurements and it was shown that this
method can be used to create aberration free lenses for a wide dynamic range of focal
lengths.
4.Mode Cleaner Optical Design (Dave Reitze, Guido Mueller, Luke Williams) - We have
finalized the specification and design of the mode cleaner mirrors for AdvLIGO and LASTI.
In addition, an experiment is underway to measure the long time damage thresholds of high
laser powers on high reflective coatings.
5.Optical layout for AdvLIGO (Luke Williams, Dave Reitze, David Tanner, Guido
Mueller) - Detailed SolidWorks optical layouts for the Input Optics and power recycling
cavities have been completed to the current status of the AdvLIGO design
6.Development of spatial beamsplitter (Volker Quetschke, Guido Mueller, Luke Williams)
- The current Wavefront sensors use quadrant photodetectors to spatially resolve the beat
signal between the carrier and the modulation sidebands. These photo detectors have
significant cross talk
between the receiver elements which increases with the modulation frequency. It is unlikely
that these quad detectors can be used at frequencies approaching 100MHz. We realized that
the quadrant photo detectors can be replaced by spatial beam splitters and single element fast
photo receivers. We designed a compact housing to build a first breadboard model of a
wavefront sensor based on this spatial beam splitter. This design is currently under
construction in our machine shop.
7.Fast pointing actuator: (Volker Quetschke, Guido Mueller, Luke Williams) Correcting
laser beam pointing in the input beam requires a large bandwidth actuator but relatively small
tuning range. However, we are still using PZT actuated mirrors which have an unnecessary
large tuning range but a limited bandwidth due to their low resonance frequencies. Prisms,
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

build from electro-optic materials, have a much smaller tuning range but their bandwidth is
much larger. They should be much better adopted to the problem at hand. We just received a
first set of RTP prisms and will perform our first tests of these pointing actuators.
8. LSC Optics Working Group Chair (Dave Reitze) - Duties included maintaining
coordination and communication of the OWG, running OWG meetings, participation in
organizing LSC meetings, MOU reviews.

Plans
a) Optics Characterization
This section also includes planned development work on the input optics, for both Enhanced
(Muzammil Arain, Ken Franzen, Guido Mueller, Volker Quetschke, Dave Reitze, David
Tanner, Luke Williams, and Wan Wu)
We will continue to work on development of Input Optics, optical layouts, and thermal
compensation (TCS) for enhanced and Advanced LIGO.
Input Optics
1. Modulation
a. Enhanced LIGO modulators (Volker Quetschke, Luke Williams, Guido Mueller, Dave
Reitze) - work will include FEA modeling on electric fields in the EOM crystal, final design
of the ELIGO modulator housing, building and testing of a prototype single crystal, double
electrode ELIGO EOM.
b. Complex modulation (Volker Quetschke) - Finish complex modulation effort (experiment
and requirements for AdvLIGO) and compare with Mach-Zehnder modulation
c. Mach-Zehnder modulation (Volker Quetschke, Guido Mueller, Dave Reitze) - Finish
LIGO technical note on MZ noise couplings. Finish high power Mach-Zehnder modulation
and compare with complex modulation. Write a technical note comparing MZ and complex
modulation to choose between the two (unless a sub-carrier readout scheme is adopted by the
ISC group).
d. Sideband phase and amplitude noise measurements (Wan Wu, Volker Quetschke, Guido
Mueller) – successfully identify and mitigate noise sources in the experiment, achieve 10^-8
RIN on the NPROs, move the optics into vacuum, and try to measure amplitude and phase
noise to better than 10^-7.
e. A recently developed method to explore the feasibility of measuring the amplitude and
phase stability of individual sidebands will be used. The sideband properties are measured
with the help of a second laser, phase locked to the laser that is to be investigated.
2. Faraday isolation (Volker Quetschke, Ken Franzen, Dave Reitze, David Tanner, Luke
Williams, and IAP group)
a. complete design and testing of new vacuum housing
b. select the optimal high power polarizers; develop and test alternative REFL beam pick-off
method using either compensated wedges.
c. build and test a prototype of the ELIGO Faraday isolator
3. Mode Cleaner (Dave Reitze, David Tanner, Guido Mueller, Luke Williams) - begin
procurement of LASTI mode cleaner optics.
4. Adaptive control of laser modal properties for mode-matching (Muzammil Arain,
Guido Mueller, Volker Quetschke,
Dave Reitze, David Tanner) - develop sensing and actuation requirements and servo-loop
design for CO2-based adaptive telescope, explore dynamic focusing using bi-morphic
mirrors and ring heating.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

5. ELIGO Upgrade (UF-LIGO) - Build and deliver ELIGO EOMs and Faraday isolators to
sites. Depending on progress of S5, begin installation at LLO.
Thermal compensation and adaptive mode-matching system
(Muzammil Arain, Guido Mueller in collaboration with Phil Willems at the LIGO Lab) - To
investigate further the issue of thermal compensation, we will keep working closely with the
thermal compensation system group. We are planning next round of experiments to test the
viability of negative dn/dT material based compensation schemes.

b) Other Contributions
Dave Reitze will continue in the role of OWG chair.

GEO
Activities Related to Attachment OPT Coating Losses for Advanced LIGO and beyond
a.1.) Measurements of multi-layer coatings applied to fused silica substrates
(Crooks, Murray, Reid, Rowan, Hough + I. MacLaren)
Reduction of the mechanical losses associated with the addition of coatings to substrates and
associated thermal noise remains an important research area for Advanced LIGO and is vital
for the success of any future detectors that aim to have sensitivities better than Advanced
LIGO.
We have continued our investigations of the mechanical loss factors of dielectric coatings
applied to bulk fused silica substrates. This work is carried out in collaboration with Stanford
University, Syracuse University, MIT, and Hobart and William Smith Colleges. Previously
we studied the loss factors of coatings of SiO2/Ta2O5 applied to fused silica substrates in
which the Ta2O5 component of the coating had been doped with increasing amounts of
TiO2by the coater (LMA Lyons.)
LMA were initially unable to provide accurate values for the concentration of the dopant
added to the Ta2O5 component: We have thus worked with colleagues in Glasgow (Dr. Ian
Maclaren) to use Electron Energy Loss Spectroscopy (EELS) to:
a.    provide a more accurate estimation of the concentration of dopant in each of the
coatings.
EELS analysis of three of the five samples produced by LMA was carried out and the
percentage of cation dopant of the tantala component of the coatings was found to be:
Formula 1,      8.5±1.3% TiO2,
Formula 3,      22.5±2.9% TiO2, Formula 4,       54±5% TiO2.
b.    study the homogeneity of the distribution of dopant throughout the coating. The TiO2
appeared quantitatively to be evenly distributed throughout the Ta2O5.
LMA subsequently provided further independent estimates of the amount of dopant in
their coatings - as below:
Formula 1,      6.0±0.6% TiO2,
Formula 2,       13±1% TiO2,
Formula 3,      24±2% TiO2,
Formula 4,      54.5±5% TiO2,
Formula 5,       14.5±1%TiO2.
Thus our measurements and those of LMA appear to be in reasonable agreement.
c.    We have combined our measurements of the mechanical loss of these coatings with
results
from our collaborators on nominally identical coatings applied to thinner substrates and these
provide a relatively consistent outcome - each of the doped coatings has a lower mechanical
loss than an undoped coating (the doped coating losses are ~60% of the undoped coating) but
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

there appears no strong correlation between amount of dopant and level of loss. A paper
reporting the results of this work is in the final stages of preparation.
a. 2.) Measurements of multi-layer coatings applied to sapphire substrates
We have not carried out significant experimental work in this area over this period, instead
focussing on the properties of coatings applied to fused silica and silicon.
a.3.) Mechanical loss associated with coatings for diffractive optics
(Cumming, Heptonstall, Rowan & Hough)
This work has been carried out in collaboration with colleagues in Hanover and Jena.
Preliminary work has been carried out studying the mechanical loss of silica samples on
whose surface had been etched a diffraction grating. Initial tests indicated that the placement
of an etched diffraction grating (200nm deep) on a silica disk sample did not significantly
increase the mechanical loss of the sample. However we believe that there was excess loss
associated with the suspension technique used to obtain these results, so further work has
been undertaken to minimise this loss source.
Small variations in temperature (~a few degrees Celsius) around room temperature have been
shown to have an impact on the measured mechanical loss (of the order of 20%), indicating a
coupling of energy out of the system at certain mode frequencies.
Once these sources of excess loss have been identified and minimised, measurements will be
made on annealed blank disks, and a disk with a 50nm deep etched grating and a multilayer
optical dielectric coating, to simulate a test mass optic.
a. 4.) Coating loss measurements using thin cantilever substrates
(Martin, Reid, Heptonstall, Holt, Rowan & Hough)
We have carried out measurements of the mechanical loss of coatings applied to thin
cantilevers of fused silica and silicon to study the mechanical loss of single layers of coating
materials - that is, of single layers of ion-beam-sputtered SiO2, Ta2O5 and doped
1.     A selection of <110> oriented single-crystal silicon cantilevers have been coated with a
0.5μm thick single layer of doped tantala ("Formula 5", doped with ~14% titania) and
mechanical loss measurements have been carried out on one cantilever between 85 and 290
K.
The mechanical loss associated with the coating layer over this temperature range has thus
been calculated. The calculated coating loss decreases with temperature to ~200K, where the
coating loss reaches ~2xlO"4. The coating loss increases with further cooling, showing signs
of a possible dissipation peak around 90-100K.
The control (uncoated) sample whose mechanical loss is used in the analysis of the results
here was not however taken through the same annealing and preparation stages as the coated
sample prior to the coating being deposited. Therefore the mechanical loss of an uncoated
control sample which has been through the appropriate annealing cycle is now being
measured prior to the measurement of a second coated cantilever.
2.     In collaboration with colleagues from LMA, the mechanical loss of a series of thin
silica cantilevers with various coatings has been measured. The cantilevers (of dimensions
40mmx5mmxll0|j,m) were clamped directly in a vice by LMA and the mechanical loss of
several of the bending modes of the cantilevers was measured in both the Glasgow and LMA
labs.
The mechanical loss of a 0.5μm single layer coating of silica was calculated to be
approximately lxl0-4, with studies of undoped and doped tantala coatings yielding calculated
losses of ~4.5xl0-4 and ~3xl0-4 respectively. The doped coating tested here was "Formula 5*"
whose exact composition has not been disclosed by LMA.
3.     Some of the measurements of the samples above showed evidence of energy loss
associated with the clamping technique thus loss measurements have also been carried out on
silica ribbon cantilevers where laser welding was used to attach a section of ribbon to a silica
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

slide to form a clamping block. These measurements suggest that this technique can
effectively de-couple the ribbon cantilever from the clamp used to support it and thus
minimise excess loss into the clamping structure.
This laser welding technique has been used to attach larger silica cantilevers (60mm by 5mm
by 0.11 mm) to silica clamping blocks. The cantilevers, supplied by LMA, will be coated
with various types of doped tantala and the loss of the coated samples will be measured in
both Glasgow and Lyon.
These studies will allow a more accurate estimation of the losses of the silica and tantala
coating materials to be obtained, which is necessary for the successful implementation of
coating designs proposed by Pinto to minimise the amount of tantala in a coating thereby
reducing the coating thermal noise.

Plans
a) Optics Characterization
Coating Losses for Advanced LIGO and beyond
a.1.) Measurements of multi-layer coatings applied to fused silica substrates (Crooks,
Murray, Reid, Rowan, Hough + I. MacLaren)
Reduction of the mechanical loss associated with the addition of coatings to substrates and
associated thermal noise remains an important research area for Advanced LIGO and is vital
for the success of any future detectors that aim to have sensitivities better than Advanced
LIGO.
We thus propose to continue to work with our LSC colleagues- Stanford University,
Syracuse University, MIT, and Hobart and William Smith Colleges, on studies of the excess
mechanical losses associated with adding dielectric coatings to test mass substrates
a.2.) Measurements of multi-layer coatings applied to sapphire substrates (Chalkley,
Murray, Cumming, Faller, Hough & Rowan)
We aim to revisit our studies of multi-layer coatings applied to sapphire substrates once our
work on nodal supports has progressed further.

a.3.) Mechanical loss associated with coatings for diffractive optics (Cumming,
Heptonstall, Rowan and Hough)
Investigation of the mechanical loss associated with the coatings for diffractive optics will
continue in collaboration with colleagues in Hanover and Jena. Once the sources of excess
loss in the suspension have been identified and minimised, measurements will be carried out
on annealed blank disks, and a disk with a 50nm etched grating and a multilayer optical
dielectric coating, to simulate a complete test mass optic.

a.4.) Coating loss measurements using thin cantilever substrates (Martin, Reid,
Heptonstall, Holt, Cunningham, Rowan & Hough)
The loss of <110> oriented single-crystal silicon cantilevers has previously been studied
between 290 and 80 K. Currently, our small cryostat is being adapted for cooling to liquid
Helium temperature, which will allow measurement of the mechanical loss around 18 K,
where the thermal expansion co-efficient (and hence the thermoelastic dissipation) goes to
zero. We have recently recruited a new group member with expertise in semiconductor
nanofabrication and are working towards being able to fabricate in-house silicon cantilevers
for further study. In particular, the effects of crystal axis orientation, doping and surface to
volume ratio on the mechanical loss will be investigated. We intend to study further single
layer coatings of tantala and silica on both silica and silicon substrates and investigate the
effect of doping and coating thickness on the mechanical loss. We plan to be in a position to
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

carry out these measurements from room temperature down to (or close to) 4K. Over the next
six months a third cryostat facility capable of carrying out mechanical measurements down to
(or close to) 4K will be commissioned for measuring similar small cantilever-type samples.

b) Other Contributions
Studies of use of non-Gaussian beams
b.1 Numerical simulations of non-Gaussian mode resonators (Nelson, Strain)
We plan to use numerical methods to explore non-gaussian mode resonators for potential
application in GW detectors. A LabVIEW based Fox-Li simulation has been developed and
this will be used to explore the design parameters of cavities containing various flat-topped
(mesa) beam shapes. Results will be compared to analytical calculations and experiment
already carried out within the LSC.

Hobart

Research on thermal noise in Advanced LIGO substrates is focused in two areas: developing
an improved understanding of the loss mechanisms infused silica, and optimizing preparation
method for the Advanced LIGO optics.
Our group has taken the lead role in understanding and reducing loss in the fused silica
substrates for Advanced LIGO. Our paper modeling that loss has recently appeared in
Physics Letters A (PLA 352 (2006) 3-6). That model indicates that the loss decreases with
frequency and with surface-to-volume ratio (S/V). The Advanced LIGO test masses, with
their large volume and low frequency range of interest, benefit from both of these
dependencies and should have a loss well below the loss in the applied mirror coating and
below the loss arising from the laser. Unfortunately there are no measurements to test the
model at these conditions. In March 2005 I ordered a cantilever rod sample from Heraeus
with a S/V ratio similar to the Advanced LIGO optics and four resonances below 4 kHz. I
specified that the rod sample was to have a flame-polished surface. I also specified the
annealing curve so that the sample would cool at only 1 °C/hr from the annealing point to the
stress point. This slow cool down should minimize any residual stress in the sample.
During the intervening months while we waited for the delivery of the large cantilever
sample, we suspended a rod sample of Suprasil 312 horizontally by welded attachments at
the nodes of the lowest mode. This suspension proved very difficult to hang. The preannealed
Q was 86 million. When hanging the annealed sample, we broke a suspension fiber which
caused one of the suspension bobs to fall and strike the sample. The bob broke in half and the
sample was badly chipped. Examination using a polarimeter showed that the damaged area
was very stressed. Nevertheless we measured the Q and found 102 million. Then we repaired
the major damaged areas and annealed the sample again. We later discovered that we failed
to fix one crack in the surface. The sample, still slightly damaged, was found to have a Q of
161 million. At this point the new large cantilever sample arrived and we removed the
nodally suspended sample from the bell jar.
In February 2006, we took delivery of the sample. Unfortunately the surface of the sample
was found to contain several defects. It appears that the surface was not prepared as specified
and that leads one to speculate whether the sample was properly annealed. Despite these
defects we suspended the sample (http://web.mac.com/penn/
iWeb/Lab/LargeCantileverSample.html) but the measured Q's were a disappointing 50
million. The frequency dependence was flat, indicative of some excess loss mechanism,
rather than the f0.7 expected for Suprasil 312.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

Heraeus has agreed to repair the sample which will include preparing a new thick rod sample,
welding it to the large counterweight, and annealing the entire piece. I expect this process to
take a few months. See http://www.ligo.caltech.edu/docs/G/G060140-00.pdf

The HWSLG is responsible for performing the experiments to determine the optimal
annealing cycle for Advanced LIGO optics. The goal of these experiments is to anneal
each optics so that its mechanical loss is on or near the minimum loss surface (as described in
Penn et al., http://arxiv.org/abs/gr-qc/0507097 ) but not to alter the optical characteristics in
the process. These experiments will be performed on a set of optics that are logarithmically-
spaced in size (V/S) between previously measured small optics and Advanced LIGO optics.
Both the mechanical loss and the cooling rate of the annealing process will depend on the
size of the optic. The set of optics were ordered last summer from Heraeus. They have now
arrived at Caltech where they are being characterized. At HWS the vacuum chamber has
been fitted with a single wire loop suspension and the vacuum annealing oven has been
upgraded with a turbo pump and a residual gas analyzer. We expect delivery of the optics
from Caltech at any time. These measurements will require 1-1.5 years depending on the
number of annealing cycles required to optimize the process. Each annealing cycle can
require two weeks or more for large optics.

Plans
a) Optics Characterization
The HWSLG is involved in a series of experiments to determine the optimal annealing
cycle for Advanced LIGO optics. The goal of these experiments is to anneal each optics so
that its mechanical loss is on or near the mimimum loss surface (as described in Penn et al.,
http://arxiv.org/abs/gr-qc/0507097 ) but not to alter the optical characteristics in the process.
These experiments will be performed on a set of optics that are logarithmically-spaced in size
between previously measured small optics and Advanced LIGO optics. Both the mechanical
loss and the cooling rate of the annealing process will depend on the size of the optic. The set
of optics were ordered last summer from Heraeus. They have now arrived at Caltech where
they are being characterized. At HWS the vacuum chamber has been fitted with a single wire
loop suspension and the vacuum annealing oven has been upgraded with a turbo pump and a
residual gas analyzer. We expect delivery of the optics from Caltech at any time. These
measurements will require 1–1.5 years depending on the number of annealing cycles required
to optimize the process. Each annealing cycle can require two weeks or more for large optics.

The HWSLG is continuing its efforts to determine the minimal mechanical loss in fused
silica. Previously we had developed a model of the loss that depended on frequency and on
the volume-to-surface ratio. However that model lacked measurements in the region of low
frequency and large V/S, the regime in which Advanced LIGO optics reside. We seek to
directly measure the loss in this region by measuring the loss in a large cantilever rod. In
March 2005, we ordered this cantilever sample from Heraeus. One year later the sample
arrived. But the surface was damaged and not prepared as requested. In addition there was no
confirmation that the sample was annealed as specified. We are now in the process of having
the sample replaced and reannealed. I suspect that this process will require a few months.
Everything that Heraeus does requires a few months. Our hope is that this sample will fill in
the gap in data in our model and confirm the low loss that we expect for Advanced LIGO
optics.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

IAP

High aperture (20mm) high power version Faraday isolator for installing in LHO and as a
prototype of AdLIGO Faraday isolator was tested at high power laser in UF. The test
included i)measurements of isolation ratio as a function of laser power, beam position on the
aperture, angle of the beam direction, ii)measurements of thermal lens and its adaptive
compensation as a function of laser power, thickness and orientation of compensating DKDP
crystal, and iii)vacuum test. The first two measurements showed parameters which
completely fits AdLIGO requirements. Set of polarizers and compensating DKDP crystals
was manufactured and delivered to UF for installation in LLO. On the other hand the Faraday
isolator did not passed through vacuum test. Design of the Faraday isolator housing must be
changed to avoid any air holes. Also procedure of each individual magnet cleaning and the
magnet system assembling must be change in order to fit AdLIGO vacuum requirements.
We have finished to model bulk surface (by CO2 laser) heat deposition in AdLIGO Core
Optics and thermal compensation using scanning beam of CO2 laser. White light
interferometer and scanning Hartman sensor used in these measurements. The 2D version of
the scanning Hartman sensor was modified. As a result it provides us better sensitivity and
faster scanning simultaneously.

Plans

a) Optics Characterization
Design of the Faraday isolator housing will be changed to avoid any air holes. We will
design magnet system with reduced outside magnet field. Also procedure of each individual
magnet cleaning and the magnet system assembling will be developed in collaboration with
UF group. New version of magnet system will be manufactured, delivered and passed
through vacuum test. New set of DKDP crystals will be grown, polished and delivered to UF.
After vacuum test the full optical test of high aperture (20mm) Faraday isolator with dual
(depolarization and aberration) compensation will be done before installing in HLO.

We will continue to model bulk (by CW Yb:fiber) heat deposition in AdLIGO Core Optics
and thermal compensation using scanning beam of CO2 laser. The white light interferometer
and 2D version of the scanning Hartman sensor will be used in these measurements. The
white light interferometer will be modified to provide maximal time of sequence recording at
least one hour. Improved mathematical algorithm will be developed for the scanning
Hartman sensor. A white light interferometer for in situ measurements of ETM optical
thickness will be installed and tested on ETM in the LIGO Livingston detector (if invitation
will be issued).

Southern

Presentations:
E. E. Doomes, S. C. McGuire, R. Tittsworth, "Initial X-ray absorption spectroscopy
measurements in LIGO mirror coatings," paper presented at the LIGO Scientific
Collaboration (LSC) Meeting, Hanford Washington, August 13-17, 2005; LIGO-G050383-
00-Z.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

S. C. McGuire, E. A. Mackey and G. P. Lamaze, "Trace element content comparison for
high-loss and low-loss sapphire," paper presented at the LIGO Scientific Collaboration
(LSC) Meeting, Hanford Washington, August 13-17, 2005; LIGO-G050382-00-Z.
S. C. McGuire, "LIGO, materials research and physics education at Southern University,"
Invited colloquium, given to the department of physics, University of Colorado, Colorado
Springs, CO, April 7, 2006.
S. C. McGuire, "LIGO at Southern University," Leroy Posey Invited Memorial Seminar,
Department of Mathematics, Southern University and A&M College, April 20,2006.
S. C. McGuire, "What is different about HBCUs?" Invited talk presented at the 2006
Department Chairs Meeting, American Center for Physics, College Park, MD, June 11, 2006.
Reports:
S. C. McGuire and E. E. Doomes, "LIGO Mirror Coatings XRF Measurements," monthly
progress report made to the Coatings SubGroup of the Optics Working Group;
teleconference, May 20, 2006.
S. C. McGuire and E. E. Doomes, "LIGO Mirror Coatings XRF Measurements," monthly
progress report made to the Coatings SubGroup of the Optics Working Group;
teleconference, June 8, 2006.
Our research and educational outreach program will continue during the next 12-month
period as described in the updated (2006) MOU attachments with no major changes
anticipated. The research focus centers on understanding the possible role of Titania (TiO2)
on 1064 nm absorption within the mirror coatings. Our immediate technical objective is the
determination of Titania (TiO2) content and structure within the Tantala (Ta2O5) layers of
the multiplayer coatings on SiO2 substrates. For this purpose the four samples received from
the Caltech coatings group will undergo detailed XRF, XANES and EXAFS analysis at the
Double Crystal Monochromator (DCM) beamline of the CAMD/LSU Laboratory. The effort
is being led by Professor Edward E. Doomes of the Physics Department in collaboration with
CAMD senior scientists. In addition, we are employing a recently acquired atomic force
microscope (AFM) for the LIGO-SUBR Advanced Optical Materials Laboratory. This
addition will enable microscopic visual imaging of the coating surfaces and promote an
expansion our on-campus capabilities in the areas of morphology and chemical composition
of multilayers and surface microscopy.

Plans

a) Optics Characterization
Test mass multilayer coatings characterization:
The SUBR Physics Group will investigate the chemical and nano structural characteristics of
test mass multilayer coatings under consideration for use in advanced versions of the
interferometer. The goal of this work is to help identify and understand potential sources of
1064 nm absorption in the coatings. In pursuing this goal we will take advantage of
synchrotron radiation probes and data analysis systems available at the Center for Advanced
Microstructures and Devices (CAMD) in Baton Rouge, LA. Appropriate experiment
proposals for beam time are being submitted for evaluation by CAMD. Experiments will be
designed and carried out be SUBR Physics Group
personnel in collaboration with senior CAMD scientists at the Double Crystal
Monochromator (DCM) Beamline.

Continuing studies focus on the determination of the titania (TiO2) content within the
Tantala (Ta2O5) layers of coatings recently provided to SUBR by Caltech. Techniques to
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

be employed include (1) X-ray fluorescence (XRF) for chemical content determination, and
(2) X-ray absorption near edge spectroscopy (XANES) for information on atomic valence
state, three-dimensional geometry and coordination environment of the element under
investigation, and (3) Extended X-ray absorption fluorescence spectroscopy (EXAFS) for
information on coordination environment and nearest neighbor atom environment. During
the next performance period these measurements will be complemented with (a) grazing
incidence XRF experiments designed to avoid possible interferences from the substrate and
(b) nanometer-level atomic force microscope (AFM) imaging of the coatings to examine
surface morphology features.

Stanford

a) Optical Materials (V. Kondilenko, A. Markosyan, A. Alexandrovski, R. Route, M. Fejer)
V. Kondilenko accepted a permanent position at Coherent Crystal Associates in New Jersey
in April 2006. He has been replaced by Ashot Markosyan from Moscow State University
whose background is in solid-state physics. Ashot began working with the photo-thermal
common-path interferometry (PCI) apparatus in May 2006, and he has quickly come up to
speed on making optical loss measurements in both bulk and coated optical elements. a)
Optical absorption in sapphire crystals Prior to leaving the program, Volodymyr Kondilenko
carried out optical absorption measurements on Crystal Systems' (CSI) sapphire crystals and
coated fused silica optics using the PCI technique. Previous annealing studies on CSI
sapphire windows (25 mm dia by 10-12.5 mm thick) showed that the nominal optical
absorption of as-grown sapphire (40-60 ppm/cm) could be reduced by post-growth, vacuum
heat-treatment at >1800 C for periods of 100 hrs. Absorption losses were reduced in some
cases to as low as 12 ppm/cm under slow cooling conditions typical of those used for large-
scale sapphire crystals. With high temperature vacuum heat-treatment followed by slow
cooling, the improvements appear to be permanent as compared with intermediate
temperature heat-treatment followed by rapid cooling in which the improvements are
reversible. This is consistent with elimination of point defects through out-diffusion rather
than modification to their thermal equilibrium. The ability to cool slowly after high
temperature vacuum heat-treatment would be important in terms of minimizing stresses in
large size optical elements.

During the prior reporting interval, we completed an "engineering" series of four vacuum
heat-treatment experiments followed by commercial re-polishing on 25 mm dia. windows. It
purpose was to identify any lingering furnace control problems, as well as to gain some idea
of the experimental parameters needed to determine the kinetics of the out-diffusion process
using fully characterized windows of the largest size that can be accommodated in the
furnace. Following the engineering series, heat-treatment and optical testing was initiated on
a "science" series using four, 37 mm dia. by 25 mm high windows. Heat-treatment times
were intended to be varied up to 100 hrs in the first cycle, and after all four windows had
been re-polished, they were planned to be re-measured and analyzed. The goals of this work
were to determine the kinetics of the apparent out-diffusion process, to determine if there is a
lower limit to the optical absorption achievable by vacuum heat-treatment, and to determine
if the kinetics are adequate to reduce the optical losses in large-size sapphire crystals within a
reasonable time. (Diffusion times typically scale as the square of the distance.) The largest
specimens that we could accommodate initially were cylinders 50 mm dia. by 50 mm high.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

Aging of the furnace components during these comparatively lengthy periods at very high
temperatures, and problems with post-processing polishing, were encountered, however, and
we sought to come up with a more reliable furnace that would allow us to carry out extended
high temperature processing without drop-outs that would subject the test
samples to high cooling rates. [Billingsley LIGO-T020103] We worked therefore to
commission an rf-induction-heated vacuum furnace that could operate at higher temperatures
and with significantly higher pumping speeds.

The susceptor was a tungsten crucible large enough to accommodate crystals up to 37 mm
dia. by 25 mm high. Although this was somewhat smaller than the capacity of the tungsten
mesh resistance-heated furnace used for our initial high temperature vacuum heat-treatment
engineering experiments, it was hoped to be more reliable, and the slight penalty in sample
size was still thought sufficient to allow us to determine the kinetics of the heat-treatment
process. Extensive testing of this rf furnace showed, however, that the capacity of the re-
circulating cooling water system in the building was insufficient to permit reliable processing
at temperatures in the 1800 C range without periodic drop-outs. Since drop-outs would have
compromised the heat-treatment process that we were trying to characterize, and invalidated
a large amount of spectral optical loss measurements that we had accumulated on the as-
received sapphire cylinders, we shelved the idea of using the rf-induction furnace and
decided to await the arrival and installation of a new, resistance-heated vacuum furnace from
Oxy-Gon Industries which was purchased using other sources of funding. This furnace
arrived in January 2006 and since then, has been undergoing installation and testing to work
out bugs, etc. We have only recently completed necessary modifications to the re-circulating
cooling water system in our laboratory and developed confidence in the furnace's temperature
control system which is complicated and not as "bullet-proof‖ as one might expect from a
commercial instrument. The heated zone in this new furnace is 4" dia by 4" high, more than
adequate for the sapphire cylinders that we envision studying. We hope to resume heat-
treatment of the 37 mm dia. by 25 mm high "science" series of sapphire cylinders in the near
future.

b) Absorption in optical coatings The PCI technique has also been used to carry out
absorption studies on multi-layer optical coatings on single crystal sapphire and fused silica
windows. These studies are being carried out in an LSC collaboration through a systematic
comparison of multi-layer antireflection coatings composed of Nb2O5/SiO2, Ta2O5/SiO2,
ZrO2/SiO2, Ta2O5/Al2O3, and SiO2/Al2O3 on fused silica windows. The first two had
almost identical absorption at the level of 1 ppm or less, and the controlling issue turned out
to be not the optical absorption loss, but rather the effect on mechanical Q of the coated optic
and its thermal noise. We are currently supporting the latest efforts by the LSC and
LMA/VIRGO to evaluate TiO2-doping of Ta2O5/SiO2 coatings as a means to reduce
thermal noise. [Harry LIGO-P040023]. In addition, we have supported LIGO in evaluating
coated optic MMT14KO4-1, which was used to evaluate several assembly processes in the
recent ITM change at LHO and is still under study to evaluate cleaning and dust removal
procedures.

Our collaborative paper on mechanical dissipation associated with optical coatings has been
published: D R M Crooks, G Cagnoli, M M Fejer, G Harry, J Hough, B T Khuri-Yakub, S
Penn, R Route, S Rowan, P H Sneddon, I O Wygant and G G Yaralioglu, "Experimental
measurements of mechanical dissipation associated with dielectric coatings formed using
SiO2, Ta2O5 and Al2O3", Class. Quantum Grav. 23 (2006) 4953-4965
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

We have completed one manuscript on the study of mechanical and optical losses in titania-
doped tantala/silica coatings:
1) G.M. Harry et al., "Titania-doped tantala/silica coatings for gravitational-wave detection",
to be submitted. UV LED-based charge management system for test mass charging
mitigation (K.X. Sun, R.Route, R.L. Byer)

c) Test mass charging mitigation Test mass charging has become an increasing concern for
maintaining the sensitivity of Advanced LI GO. Test mass charging can be induced by
cosmic ray radiation and impacts by charged "debris" within the vacuum system. Electrical
charges on test mass surfaces have been found to reduce pendulum Q and may causing
random force couplings between test masses and the electrical environment [1-3]. We are
investigating ways to mitigate charging effects on test masses by using the photoelectric
effect to actively manage charge distribution, specifically the use of modulated UV LED
light at 257 nm [4]. (An LED source would be a significant improvement over the deep UV,
mercury lamp charge management system used on the Gravity Probe-B mission [5], and
adopted for the LISA design.) In these missions, photoelectron emission is accomplished by
deep UV illumination of conducting, metal-coated surfaces. To be useful for LIGO
applications, the surfaces of test masses and test mass coatings in particular would have to be
slightly conductive in order to allow surface charges to migrate. Technologies for making the
surfaces of dielectric materials such as Ta2O5 and SiO2 slightly electrically conductive
without affecting their optical and/or mechanical loss properties have not been addressed
experimentally, although we are discussing the possibility of making such surfaces with
colleagues in GEO. Further discussions are planned in this area. On the hardware side, we are
currently evaluating a pre-release, deep UV, 257 nm, light-emitting AlGaN diode (LED) [6]
that has the advantages of high performance, low electromagnetic interference, low cost,
small volume, and low EMI emission. With it, we have demonstrated MHz switching with
associated changes in photocurrent, and we have demonstrated AC charge management at
frequencies outside the gravitational wave signal band [5]. To determine whether UV LEDs
might be a practical charge management system, we have been carrying out a lifetime test.
Since the study started in March 2006, our UV LED has been running pulse-modulated with
a 10% duty cycle at 1 KHz for ~4 months with an average UV output of 20 (J.W. We have
not observed any appreciable decay in power or any significant spectral shift during this
period, and we plan to continue the study.

[1] S. Rowan, S. Twyford, R Hutchins and J Hough, "Investigations into the effects of
electrostatic charge on the Q factor of a prototype fused silica suspension for use in
gravitational wave detectors", Class. Quantum Grav. 14 (1997) 1537-1541.
[2] V. B. Braginsky, V. P. Mitrofanov, L. G. Prokhorov, K. V. Tokmakov, "Measurement of
electric charges on fused silica test mass", LSC Meeting, Livingstone, March 20-23, 2002,
LIGO-G020033-00-Z
[3] Gregory Harry, "Effect of Charging on Thermal Noise", LSC Meeting, Livingston March
2004, LIGO Document G040063-00-R
[4] Ke-Xun Sun, Sei Higuchi, Allex Goh, Dale Gill, Brett Allard, Saps Buchman, and Robert
Byer, "LIGO Test Mass Charging Mitigation Using Modulated LED UV Light," LIGO
Science Collaboration Meeting, March 19-22, 2006, LIGO Hanford Observatory, LIGO
Document LIGO-G050143-00-Z
[5] K. Sun, B. Allard, S. Williams, S. Buchman, and R. L. Byer, "LED Deep UV Source for
Charge Management of Gravitational Reference Sensors," presented at Amaldi 6
Conferences on Gravitational Waves, June 2005, published as an Amaldi 6 highlight at Class.
Quantum Grav. 23 (2006) S141-S150 [6] Ke-Xun Sun, Sei Higuchi, Allex Goh, Brett Allard,
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

Dale Gill, Saps Buchman, and Robert Byer, "Spectral and Power Stability Tests of Deep UV
LEDs for AC Charge Management", 6th International LISA Symposium, Goddard Space
Flight Center, Greenbelt, Maryland, June 19-23, 2006. To appear in: AIP Proceedings of the
6th International LISA Symposium

Plans

a) Optics Characterization (A. Markosyan, A. Alexandrovski, R. Route, M. Fejer)
Continue the measurement of optical absorption losses in coated fused silica optics in a
collaborative study with LIGO aimed at optimizing coating deposition, heat-treatment
parameters and associated cleaning and handling procedures.

b) Other Contributions (R. Route, K.X. Sun, M. Fejer)
Continue vacuum heat-treatment processing and photo-thermal common-path interferometry
(PCI) measurements of absorption losses in sapphire optical elements; With the adoption of
fused silica for the Advanced LIGO test mass material, the need for near-term study of large
size sapphire optics has diminished. However, for future gravitational wave detectors, it
remains desirable to continue heat-treatment studies on sapphire. The focus will be to
identify post-growth processing conditions that reduce optical absorption losses from as-
grown levels of 40-60 ppm/cm to the range of 10 ppm/cm. This will be done on sapphire
windows 37 mm dia. by 25 mm high, using newly acquired, tungsten-mesh resistance-heated
(~ 1800 °C) high-vacuum furnace a that was purchased with other sources of funding. The
study will elucidate the effects of high-temperature, high-vacuum heat-treatment on
absorption losses in Crystal Systems, Inc. (CSI) sapphire, aiming to determine if the levels of
12.5 ppm/cm achieved previously on smaller size optics can be demonstrated on larger size
optics. To do this, we will focus on investigating the kinetics of the process to predict if
large-size sapphire boules could be vacuum heat-treated at high temperatures in reasonable
amounts of time. We will also study the homogeneity of the heat-treated samples.

In collaboration with colleagues in GEO, we will extend our PCI apparatus to the mid-IR
range by adding pump and probe lasers so that we can evaluate both transmissive (silicon)
and all-reflective (optical grating) components in the 1.55 micron band.

Continue life testing our UV LED to determine if it can be a practical charge management
system for LIGO. Since the study started in March 2006, our UV LED has been running
pulse-modulated
with a 10% duty cycle at 1 kHz for ~4 months at an average output of 20 �� So far, we
W.
have not observed any appreciable decay in power or a significant spectral shift during this
period, and plan to continue the study.

LIGO Lab
Coatings - OPT (Harry)
In the coating research program at MIT, we have received and analyzed a number of coatings
in the last year for their mechanical loss. All of the coating samples we measured in the past
year have been coated at CSIRO in Australia. We looked at three separate samples of single
layer tantala, with different thicknesses and stresses. The results show that mechanical loss
does vary, and gets worse, with increased thickness. This is about a 25% effect when the
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

thickness is 1.8 microns, up to a 100% effect for 4.6 microns, compared to the usual quarter
wave layer. A sample of silica/silica doped titania was more promising, showing an
improved mechanical loss compared to initial LIGO silica/tantala coatings. Thermal noise
from silica/silica doped titania coatings is not predicted to be as good as the advanced LIGO
baseline silica/titania doped tantala coatings, however. A silica/tantala coating with
intentionally low oxygen content was found to have poor mechanical loss. It is unknown
whether this is because of the low oxygen content, or a result of not annealing the sample
post-coating for fear of introducing oxygen back in. This sample has now been annealed at
Stanford in a nitrogen atmosphere and is at Hobart and William Smith Colleges for further
mechanical loss measurements. A silica/lutetium doped tantala sample was disappointing,
with poorer mechanical loss than the initial LIGO coatings, as was a silica/tantala sample
where the bombardment ion was xenon rather than the usual argon. LMA/Virgo tried using
cobalt as a dopant into tantala, which reportedly gave significant mechanical loss
improvements, but the optical absorption was unacceptable and no samples were ever
delivered to LIGO. Some work was also done in the LIGO Lab on the effect of cleaning
methods on optical absorption of coatings, primarily in support of the mirror replacement at
Hanford, but the results are also of use for advanced LIGO research. It was found that
alconox cleaning did not contribute to increased absorption.

In the next year, the coating program will continue to explore new coatings. Continuing
work trying to reduce optical absorption in titania-doped tantala will be pursued with
LMA/Virgo. This is to be sure absorption reduction can be achieved, and to have a coating
with improved thermal noise that is suitable for advanced LIGO use. Coating runs to
continue reducing mechanical loss will also be done. Single layer coatings will be a priority,
with samples coated with tantala, titania doped tantala, silica, and silica doped titania. In
addition to Q and optical absorption, these samples will be measured for Young‘s modulus
and damage threshold. Alumina as a dopant, other rare earths beyond lutetium as dopants,
hafnia with dopants, and tantala with multiple dopants possibly including cobalt will all be
pursued for mechanical loss reduction. As a lower priority, annealing studies and plasma
coating will be investigated and tried as time and budgets allow.

Materials - OPT (Harry)
Other materials work has been primarily on establishing new collaborations. Close work
with Trinity University on research into the effects of charge buildup on noise has lead to an
NSF grant proposal which is hoped will fund the research next year. Design and preliminary
construction of a Kelvin probe was done at MIT. Once the experiment is complete, charge
measurements will be pursued at Trinity to quantify noise from charge fluctuations.
Discussions with Hai Ping Cheng at University of Florida on modeling of physical properties
of amorphous materials may lead to theoretical insights into causes of mechanical loss in
silica and possibly coating materials.

Mesa-Beam Research (Willems)
We have upgraded our beam spatial mode sensor from the 2D BeamScan scanned-slit
detector to the 3D Beamview CCD instrument. This gave us much more detailed metrology,
and using it we could verify the gaussian mode shapes in our spherical mirror test cavity at
about the 7% level, ensuring that detector inhomogeneity, clipping, diffraction, and parasitic
interferometers did not corrupt the cavity leakage field. In order to maximize the ratio of
desired mode power over other transverse modes in the cavity, we have implemented a dither
lock scheme to lock the cavity on top of the cavity mode peak, rather than at the side in the
offset-lock scheme.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

This characterization done, we have replaced the spherical end mirror with the Mexican hat
end mirror, and replaced the input optics to provide an appropriately larger input beam to the
cavity. The locking became somewhat more difficult, because the transverse mode spectrum
is now more bunched up than before, but we find that it is easy to lock to the low-order
transverse modes of the mesa beam cavity, and their appearance is in good agreement with
theoretical expectations.
We found it necessary to remount the flat mirrors in less distorting mounts to make the
fundamental mesa beam mode robust, but having done that we get good agreement between
the measured mode shapes and the predictions. Having done this, we measured the distortion
of the fundamental mesa beam mode to misalignment of the cavity. Initial measurements
showed the mode distorting with the expected shape, but half as rapidly with mirror tilt than
predicted. We are currently re-measuring the misalignment sensitivity to be sure of this
result.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

Suspension and Isolation

Aciga

2a) Coating Losses (ANU) Whilst not listed in our August 2005 SUS Attachment The ANU
group pursued research on avoiding coating thermal noise. We designed a coating free mirror
(CFM) based on total internal reflection and Brewster angle coupling demonstrated operation
in a cavity in which one mirror was a CFM, achieving a finesse of 4000 with reflectivity of
the CFM of 99.89% [paper in preparation].

2b) Suspension Design (UWA)
Completed the assembly of the first isolator in clean environment. The isolator is under final
tuning before being put into vacuum system. Completed the clean room at the South end
station for the assembly of the second isolator and the assembly of the second isolator is
underway (pre-isolation stage assembled, all parts cleaned and baked).

Completed the design of the suspension and control system. This will be installed under the
isolator in the South arm. Assembly too longer than anticipated.

People:
ANU: Jeff Cumpston, Stefan Gossler, David McClelland
UWA: Jean-Charles Dumas, Ben Lee (2005), Eu-Jeen Chin (2005), Chunnong Zhao, Li Ju

Plans
b) Suspension Design for Advanced LIGO
b) Suspensions for Advanced LIGO (UWA)
Developing an alternative soft suspension for Advanced detectors
(i) Characterize the full UWA isolator
(ii) Perform a differential test of the high performance isolation system in an 80m arm cavity
with small optics in
dummy test masses
(iii) Pursue advanced digital remote control system.
Success in the above programs depends on funding.

GEO
b)      Suspension Design for Advanced LIGO and beyond

b. 1.) UK Advanced LIGO Project
(Glasgow/GEO600 + RAL + Birmingham (with Strathclyde))
We have continued to provide project management and oversight of all work packages within
the UK
Advanced LIGO Project. The project is proceeding well with progress being mapped with the
US program. The UK deliverables are on track and in principle we have agreed to meet the
site installation commitments that lie beyond the end of the UK grant in September 2009.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

The noise prototype design is well underway at Rutherford Appleton Laboratory with
significant
input/support from Glasgow and Birmingham (with Strathclyde) and Caltech.
The OSEM and electronics design/fabrication activities at Birmingham with the support of
Strathclyde, Glasgow, RAL and Caltech are also proceeding well.
Glasgow delivered the four silica test mass substrates that form WP5 of the UK project early
this year, on time and within budget. This forms a major part of the Glasgow deliverables.

b.2.) Control and noise prototype suspensions
(Glasgow/GEO600 + RAL + Birmingham (with Strathclyde) + Advanced LIGO SUS Project
Team) The controls prototype is installed and is currently being tested at LASTI. Design
work on the noise prototype is progressing very well. The SUS preliminary design review
(PDR #3) has recently been completed covering the whole design of the noise prototype
including electronics/OSEMs (previously reviewed in PDR #1 in 2005) and the silica
monolithic final stage design (previously reviewed in PDR #2 in 2005). The assembly
/installation and testing of the noise prototype at LASTI will commence in spring 2007.

b.3.) Suspension support structure and mass catcher/installation jig design
(RAL + Glasgow/GEO600 + Advanced LIGO SUS Project Team)
The design of the support structure/mass catcher/installation jig for the noise prototype is in a
reasonably mature state. Currently modal frequency investigations and FE analysis are being
conducted at RAL with assistance from Glasgow to check the lowest modal frequency of the
full structure (upper structure + lower structure/mass catcher). The target is 75 Hz and there
is reasonable confidence from the modeling work already conducted that the structure will
achieve this.

Earthquake stops for the lower structure have been mechanically designed and these will be
prototyped in the near future.

(RAL + Glasgow/GEO600 + Advanced LIGO SUS Project Team)
prototype currently being tested at LASTI. Significant knowledge was gained on the static
and dynamic behaviour of cantilever blade springs during the design and construction of the
controls prototype and its effect on the alignment of the quadruple pendulum suspension. A
pre-noise prototype "marionette" suspension has recently been constructed at RAL for
advanced testing of various aspects of the noise prototype suspension design, for example,
trialled in the marionette and demonstrated to work successfully.

b.5) Active damping of suspensions - optically sensed electromagnetic actuator (OSEM)
development
(Birmingham (with Strathclyde) + Glasgow/GEO600 + RAL + Advanced LIGO SUS Project
Team) Development of the noise prototype OSEM mechanical design has been carried out.
Noise prototype OSEM part and assembly drawings have been generated and will be
submitted to the Advanced LIGO approval process in due course. A manufacturing study has
been undertaken to determine the optimum division of production tasks. Orders are ready to
be placed with external contractors pending the formal approval of the design. A number of
prototype OSEMs have been fabricated and assembled allowing for further characterization
of the units. Fit & function tests have been carried out at Caltech using an Advanced LIGO
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

triple suspension and one of the prototype units. The OSEM electronics have been developed
and are reaching maturity. Explicit requirements have been generated by the US SUS team
and we are working to incorporate these into our design. Monitoring and diagnostics design
is also maturing at this stage. Documentation required for the noise prototype PDR#3 was
generated. This mainly involved updating Interface Control and Design documents. The
noise prototype electrostatic drive electronics for Advanced LIGO have been developed. A
design is in place and testing of key features is underway (for example the implementation of
an active water cooling system if required). We have continued the development and
characterization of the interferometric OSEM as a potential backup in the event of a change
in the OSEM sensing requirements. We have continued to perform the relevant management
activities in support of the project.

b.6.) Passive damping of suspensions — eddy current damping (ECD)
(Glasgow/GEO600 + RAL + Advanced LIGO SUS Project Team)
In the Advanced LIGO quadruple suspensions ECD will be used for selected degrees of
freedom in conjunction with active damping provided by refined hybrid OSEMs. The ECD
units have been optimised with respect to mass, performance and location within the
restricted space available. Lessons learned on ECD design from the controls prototype tests
at LASTI will be fed back to the noise prototype design.

b. 7.) Electrostatic drives
(Glasgow/GEO600 + Birmingham (with Strathclyde) + RAL + Advanced LIGO SUS Project
Team) Experimental and design work on ESD drive electronics has been continuing between
Birmingham and Strathclyde and other UK collaborators, specifically the low noise front-end
electronics and the revised PCB for the high-voltage driver. Glasgow have completed the
electrostatic gold coating mask design and will pass this to Caltech in the near future who
will arrange for the gold coating deposition on the reaction mass in the US.

b.8.) CO2 laser pulling and welding of silica fibres and ribbons
(Glasgow/GEO600 + Advanced LIGO SUS Project Team)
The basic development of the CO2 laser pulling and welding machine is complete at
Glasgow. The fibre/ribbon neck thinning problem has been solved using the variable
feed/pull design and precision drive system. The ribbons for Advanced LIGO will be welded
to the refined silica ears using a lateral overlap weld. The ribbons will be fabricated from 5
mm x 0.5 mm cross-section slides. The slide ends (tabs) remaining at the ends of the ribbons
after pulling will be welded directly to the horns on the ears. This minimises the total number
of welds in the system.

A package of information with machine design and procurement information for the parts for
the LASTI machine will be dispatched to RAL in the near future. Refinement and
optimization of machine design details and the pulling/welding processes for application to
Advanced LIGO will continue at Glasgow in the run-up to the noise prototype installation.
Development of the optical profiling device for silica fibres and ribbons has progressed well
and optimisation work will continue with a machine being supplied as part of the laser pulled
ribbon characterisation suite.

b.9.) Characterisation of silica ribbon suspensions
(Glasgow/GEO600 + Advanced LIGO SUS Project Team)
The solution to the thin neck region in the pulled fibres and ribbons was determined within
the last six months. The CO2 laser machine developed at Glasgow is now capable of pulling
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

ribbons and fibres of variable taper length ranging from a few mm to tens of mm. The
removal of the thinned neck region allows strength testing on ribbons to continue. In addition
to the one already in operation at Glasgow, a new strength testing jig is currently being
designed for testing of various components including silica fibres/ribbons, silicate bonds and
drum ended wires. A paper was submitted during this period characterising flame pulled
ribbons "Characterisation of mechanical loss in synthetic fused silica ribbons", A.
Heptonstall, G. Cagnoli, J. Hough and S. Rowan, Phys Lett A, 354 (2006) 353-359. Similar
work is being conducted on CO2 laser fabricated ribbons.

b.10.) Silica ear design for Advanced LIGO
(Glasgow/GEO600 + Advanced LIGO SUS Project Team)
Following initial testing the preliminary silica ear design has been refined to facilitate an
improved ribbon welding configuration (lateral overlap weld). This has the added benefit of
further improving the intrinsic strength of the ears by resulting in lower static stress levels in
the horn region. The refined ears will be delivered to Glasgow within the next three to four
months. In the meantime welding development continues using test horns of similar
dimensions to the welding horns on the ears. The bond strengths for the current ear footprint
have been shown to be high enough for the Advanced LIGO application.

b.11.) Fabrication of dummy final monolithic stage for quadruple pendulum
(Glasgow/GEO600 + Advanced LIGO SUS Project Team)
Welding development has continued in the build up to producing a dummy monolithic final
suspension stage from composite metal/silica masses. The design of the composite masses
has commenced. Silica plates will be embedded in the 40 kg metal masses. Silica ears will be
silicate bonded to the plates on the masses and the masses will be positioned within a dummy
lower structure. The full assembly/welding procedure will then be refined using this set-up.
Construction of this dummy stage will take place towards the end of 2006 as a pre-assembly
for the LASTI installation scheduled in spring 2007.

b.12.) Violin mode damping of silica ribbons
(Glasgow/GEO600 with Strathclyde + Advanced LIGO SUS Project Team)
A programme is well underway to investigate the requirement and method for violin mode
damping. The work is being performed fully integrating the GEO experience. It has been
shown analytically that active damping of the violin modes should work, but a design study
for a sensor is required to find the optimum approach. Passive damping does not seem to
offer a practical solution for damping of the most important lower order modes. This is
primarily due to the properties of the Advanced LIGO ribbons particularly the high tension in
them due to the larger masses of Advanced LIGO. Passive damping does not reduce the
damping time sufficiently (i.e. to less than of the order of days) without also introducing
unacceptable amounts of thermal noise in the bounce mode. Experimental development of a
suitable active technique will commence early in the forthcoming period.

c) Other Contributions
c.1.) Investigations of charge mitigation techniques
(Reid, Rowan, Hough & Cunningham)
An evaluation of the use of a Kelvin Probe for measuring the work function in semi-
conductors and charge distribution in dielectric materials is underway. A postdoc, Stuart
Reid, has undergone two and a half days of training at KP Technology based in Wick,
Scotland and taken preliminary measurements on silica, sapphire and silicon substrates in
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

addition to dielectric coatings applied to silica and sapphire. An in-air system has been
purchased to continue these investigations.

Preliminary measurements show that the charging effect on fused silica and dielectric
coatings on silica or sapphire is easily observable. Measurements of boron-doped and
undoped silicon samples have shown the measured work function to be very sensitive to
photoelectric effects. As yet we have not investigated further the use of stannous oxide to
increase the conductivity of fused silica.

c.2.) Improved suspension techniques development for bulk mechanical loss measurements
(Chalkley, Murray, Cumming, Faller, Hough & Rowan)
To reduce the effects of support/suspension losses seen in some previous loss measurements
on bulk substrates, a prototype nodal support was developed based on previous studies by
Numata et al. The prototype was tested using a previously measured 75mm diameter x 25mm
sapphire sample, resulting in a Q of 5.799x107 at 93759Hz, a mode which had previously
been inaccessible for measurement using a silk thread suspension technique. This value was
the highest Q measured for the sapphire substrate. Q values for other modes, however, were
lower than those measured using a silk thread suspension. We believe that the prototype
nodal support structure was not entirely rigid, causing excess loss in the nodal support. An
improved nodal support system has been designed, with spherical ruby contacts, a more rigid
structure, built-in centring of the contacts on the face of the test mass, and an improved
securing mechanism.

c.3.) Measurements of silicon ribbon flexures
Our recent loss measurements using silicon ribbon flexures are covered in the work described
in section a.4).

(Cantley, Heptonstall, Jones & Cagnoli)
Previously preliminary blade loss measurements were carried out but appeared to be limited
by recoil loss within the blade support system. Investigations into this problem have not
progressed significantly over the last period due to design effort being placed on continuing
development of the silica suspension technology directly applicable to the Advanced LIGO
noise prototype. It is planned in the forthcoming period that in addition to investigations of
blade strengthening techniques using flame polishing or coating we will also investigate CO2
laser polishing. This development will work will resume following installation of the noise
prototype at LASTI which is scheduled for spring 2007.

c. 5.) Thermal noise experiment in Hanover
(Ribichini, Lueck & Danzmann)
The efforts to purchase an etalon to serve as an ultra-short resonator for directly measuring
the thermal noise of an optical coating continued. In the past the Laser Zentrum Hanover
tried to produce such etalons without success. In the first months of 2006 the third attempt
failed due to too high scattering in a lambda/4 layer close to the lOum SiO2 spacer. All other
parameters matched the specifications. Both the absolute length, which needed to be correct
within 5 nm, and the parallelism (< 100nrad) fulfilled the design criteria. The enhanced
scattering losses in the layer adjacent to the spacer decreased the finesse and hence prohibited
achieving a throughput of more than 3%. This not only decreased the available signal size as
a consequence of less light power but due to increased linewidth also decreased the slope and
hence further decreased the signal obtained per unit of length change. The resulting linewidth
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

also makes it more difficult to use RF readout schemes. The sensitivity reached with this
etalon was in the 5E-17m/VHz region. This sensitivity is too poor to observe the predicted
thermal noise of the etalon. Requests at other coating companies were either turned down as
being too much research effort and therefore too costly (REO, ATF) or are still being
considered (SMA Virgo). Based on experience in the Laser Zentrum Hanover it is technically
feasible to produce such an etalon with the required specifications and this investigation will
continue ensuring we maintain strong contact with the coating companies.

Plans
b) Suspension Design for Advanced LIGO
Suspension Design for Advanced LIGO and beyond
b.1.) Activities forming WP1 of UK Advanced LIGO Project – Project Management
(RAL + Glasgow/GEO600 + UK Advanced LIGO Project Team)
Continuing project management and oversight of all of the work packages within the UK
Advanced LIGO Project. Further details of this work package can be found in the UK
proposal ―Exploring the Dark Side of the Universe: Proposal for UK Involvement in
Advanced LIGO‖, Issue 2, November 2002 and the website for Advanced LIGO UK (RAL)
both of which can be accessed via the Advanced LIGO UK (Glasgow) website available via:
additional information on this work package and can be accessed via the Glasgow page.

b.2.) Activities forming WP2 of UK Advanced LIGO Project – Main Suspension Science
(Glasgow/GEO600 + RAL + UK Advanced LIGO Project Team)
Continuing scientific input to the suspensions for Advanced LIGO based on the development
of the triple suspension systems for GEO 600 and development of the Advanced LIGO
quadruple suspension controls prototype which is currently being tested at LASTI. Provision
of the reaction masses, penultimate masses and silica ears for the main suspension systems of
Advanced LIGO. For further details on this work package refer to the UK proposal
―Exploring the Dark Side of the Universe: Proposal for UK Involvement in Advanced
LIGO‖, Issue 2, November 2002 online via the Advanced LIGO UK (Glasgow) website via:

b.3.) Activities forming WP3 of UK Advanced LIGO Project – Main Suspension
Systems
(RAL + Glasgow/GEO600 + UK Advanced LIGO Project Team)
Continuing development of the final mechanical designs for the main suspension systems
based on the controls prototype and earlier GEO600 suspensions. Finalise the design and
manufacture the noise prototype for delivery to LASTI. Commence with the process of
finalising the designs for the final article suspensions for delivery to the two LIGO sites. For
further details on this work package refer to the UK proposal ―Exploring the Dark Side of the
Universe: Proposal for UK Involvement in Advanced LIGO‖, Issue 2, November 2002 online
via the Advanced LIGO UK (Glasgow) website via:
additional information on this work package and can be accessed via the Glasgow page.

b.4.) Activities forming WP5 of UK Advanced LIGO Project – Optical Material
(Glasgow/GEO600 + UK Advanced LIGO Project Team)
This work package towards provision of four silica blanks each of 40 kg for the mirrors of
one interferometer was successfully completed in February 2006. Further details of this work
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

package can be found in the UK proposal ―Exploring the Dark Side of the Universe: Proposal
for UK Involvement in Advanced LIGO‖, Issue 2, November 2002 accessed via the

b.5.) Control and noise prototype suspensions
(Glasgow/GEO600 + RAL + Birmingham (with Strathclyde) + Advanced LIGO SUS
Project Team)
Design and experimental checks will continue during the ongoing development of the noise
prototype quadruple suspension. Results from the LASTI controls prototype tests which are
currently underway will continue to be reviewed and fed back into the design process. The
SUS preliminary design review (PDR #3) has recently been successfully completed and we
have now entered the final design phase. The assembly/installation and testing of the
quadruple noise prototype at LASTI is planned to commence in March 2007.

b.6.) Suspension support structure and mass catcher/installation jig design
(RAL + Glasgow/GEO600 + Advanced LIGO SUS Project Team)
Modal frequency investigations and FE analysis will be imminently completed at RAL with
Glasgow‘s input to check the lowest modal frequency and dynamic behaviour of the noise
prototype structure (upper structure + lower structure/mass catcher). The mechanical design
of the earthquake stops is relatively mature and these will be prototyped in the near future.
Work will continue in interfacing the functionality of the lower structure as an assembly tool
and a mass catcher for the final silica monolithic stage.

(RAL + Glasgow/GEO600 + Advanced LIGO SUS Project Team)
suspension constructed at RAL for advanced testing of various aspects of the noise prototype
suspension design. Any further lessons learned from both the controls and noise prototype
suspension assembly/tests at LASTI will be fed back into the final suspension designs.

b.8.) Active damping of suspensions – optically sensed electromagnetic actuator
(OSEM)
development (Birmingham (with Strathclyde) + Glasgow/GEO600 + RAL + Advanced
LIGO SUS Project Team)
The mechanical drawings for the OSEMs will be released and approved. Orders will be
placed with external contractors for fabrication of the mechanical parts. Assembly of the
noise prototype OSEMs will be carried out in-house at Birmingham in a controlled clean
environment. A number of complete OSEM units shall be shipped to RAL for a test assembly
on the quadruple suspensions. The OSEM electronics and monitoring and diagnostic will be
completed, tested and built for delivery. The tests and assembly of the electrostatic drive
electronics will be completed. The OSEM production units and all electronics shall be
shipped to the LASTI test facility Support will continue to be provided for assembly,
installation and testing of OSEMs, OSEM electronics and electrostatic drive electronics at
LASTI from spring 2007 onwards.

b.9.) Passive damping of suspensions – eddy current damping (ECD) (Glasgow/GEO600
+ RAL + Advanced LIGO SUS Project Team)
Feedback from the performance of the ECD units in the controls prototype at LASTI will
continue to be incorporated into the final noise prototype design. Feed back from the noise
prototype tests will subsequently be incorporated.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

b.10.) Electrostatic drives
(Glasgow/GEO600 + Birmingham (with Strathclyde) + RAL + Advanced LIGO SUS Project
Team)
Experimental and design work on ESD drive electronics will be completed between
Birmingham with Strathclyde and other UK collaborators, specifically the low noise front-
end electronics and the revised PCB for the high-voltage driver. We will continue to provide
our support in this area. This work will be completed for application to the installation of the
noise prototype at LASTI in spring 2007. The mask required for the gold coating of the noise
prototype reaction mass has been designed by Glasgow and will be manufactured by Caltech.
The reaction mass and its spare will be delivered to Caltech at the end of 2006 for application
of the gold coating.

b.11.) CO2 laser pulling and welding of silica fibres and ribbons (Glasgow/GEO600 +
The basic development phase of the CO2 fibre pulling and welding machine is complete.
Refinement and optimization of machine design details and the pulling/welding processes for
application to LASTI and Advanced LIGO will continue at Glasgow in the run-up to the
noise prototype installation at LASTI in spring 2007. Glasgow has started to transfer the
design and procurement information package for RAL to order the machine components to
be delivered to LASTI early next year. This transfer will be complete within the next few
months. Development of the optical profiling device for silica fibres and ribbons is already
mature and will continue to be refined for application during the LASTI noise prototype
installation.

b.12.) Characterisation of silica ribbon suspensions (Glasgow/GEO600 + Advanced LIGO
SUS Project Team)
The CO2 laser machine is capable of pulling ribbons and fibres of variable taper length
ranging from a few mm to tens of mm. The thermal noise and flexure point calculations for
the tapered ribbons will be completed for the noise prototype to determine the optimum taper
design. The Glasgow fibre/ribbon strength testing jig is currently being upgraded and this
will be complete within the next few months. Quality factor measurements in tapered silica
ribbons and fibres will continue in parallel with this work.

b.13.) Silica ear design for Advanced LIGO (Glasgow/GEO600 + Advanced LIGO SUS
Project Team)
The refined silica ears for the noise prototype will be delivered to the US towards the end of
the year. The ear bonding area has been shown to be sufficiently strong for application to
Advanced LIGO. The ear design work will be fully completed for the noise prototype
installation at LASTI in spring 2007.

b.14.) Fabrication of dummy final monolithic stage for quadruple pendulum
(Glasgow/GEO600 + Advanced LIGO SUS Project Team)
Construction of this will take place towards the end of the year as a pre-assembly for the
LASTI installation scheduled in spring 2007. Silica plates will be embedded in the 40 kg
metal masses. Silica ears will be silicate bonded to the plates on the masses and the masses
will be positioned within a dummy lower structure. The full assembly/welding procedure will
then be further refined using this set-up.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

b.15.) Violin mode damping of silica ribbons (Glasgow/GEO600 with Strathclyde +
Since passive damping has been shown to be unsuitable the most suitable active method for
violin mode damping of the final stage suspension ribbons will continue to be investigated.
Experimental development will commence early in the forthcoming period. It is ultimately
intended that the developed violin mode damping system can be tested in the noise prototype
at LASTI. Suitable interface design features are currently being incorporated in the lower
structure to accommodate such a system.

c) Other Contributions

c.1.) Investigations of charge mitigation techniques (Reid, Rowan, Hough & Cunningham)
An evaluation of the use of a Kelvin Probe for measuring the work function in semi-
conductors and charge distribution in dielectric materials will continue. Over the next period
a calibration process will be used to enable the Kelvin Probe signal to be related to the
magnitude of the charge on a surface. A study of the charging and discharging of fused silica
substrates can then be carried out.

c.2.) Improved suspension techniques development for bulk mechanical loss
measurements (Chalkley, Murray, Cumming, Faller, Hough & Rowan)
The development of improved suspension techniques for bulk mechanical loss measurements
will continue. Aspects of the performance of the nodal support will be tested by varying the
support force applied to the test substrate and use of materials with different hardness
characteristics as spherical contacts. The nodal support, with some adaptations, is also
planned for use with the measurement of coating loss in thin silica discs.

c.3.) Measurements of silicon ribbon flexures (Reid, Martin, Cunningham, Rowan &
Hough)
We will continue our studies of the mechanical loss of silicon cantilevers as per the
discussion in a.4) above.

c.4.) Silica cantilever blade development (Cantley, Heptonstall, Cagnoli)
Investigation and development of silica cantilever blades has not progressed over the
previous period so that research and development effort could be focussed on direct
requirements for the Advanced LIGO noise prototype. This work will be resumed in the
forthcoming period. In addition to blade strengthening by flame polishing, or coating we will
also investigate application of the CO2 laser polishing technique to silica blades. This work
will commence following installation of the noise prototype at LASTI in spring 2007.

c.5.) Thermal noise experiment in Hanover (Ribichini, Lueck & Danzmann)
Based on experience within the Laser Zentrum Hanover it is considered to be technically
feasible to produce an etalon with the required specification to serve as an ultra-short
resonator for directly measuring the thermal noise of an optical coating. However, the
production of such an etalon is proving to be a challenge. Communications will be
maintained with the etalon coating companies to continue to investigate a method of
obtaining a suitable low loss, high throughput etalon.

Hobart
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

The HWSLG continues to be a part of the Coating Program, which is attempting to find
a high reflectivity mirror coating for Advanced LIGO that has minimal thermal noise
contribution. The Coating Program is very important to the success of Advanced LIGO since
it is the coating thermal noise that is the limiting noise source in the minimum of the
Advanced LIGO noise curve. Our facility acts as a second test facility, along with MIT, for
thin coating samples and is the primary lab for testing the effects of annealing on coating
loss. We work in close coordination with the MIT Coating group.

The initial goal of the Coating Program was to understand the source of the noise in the
current LIGO coatings. We discovered that the loss was primarily in the high index, tantala
layers rather than the low index, silica layers (see Penn, et al., Classical and Quantum
Gravity 20 (2003) 2917-2928). While the proportion of loss ascribed to those two materials
has changed slightly due to the direct measurement of the Young's modulus by Sheila
Rowan, the essential conclusion stands. Thus our research has focused on finding a high
index coating material with low mechanical loss. We have explored using other high index
dielectrics instead of tantala, and doping the tantala in order to lower its loss. At present the
best candidate coating is tantala doped with titanium.

During the past year, our group has measured several samples including doped tantala/silica
coatings and annealed silica/alumina coatings. We have also compared the loss in ion beam
polished coatings with mechanically polished coatings. Finally we have initiated work with
Prof. Krystyn van Vliet, of the MIT Materials Science Department, about measuring the
Young's modulus of our coatings using a nanoindenter.

For the next year, although a baseline coating for Advanced LIGO has been approved, we
will continue to explore other coating options because the coating loss is still about a factor
of three above the quantum noise limit of the laser.

Our work with CSIRO has shown that we can reduce the mechanical loss in titania coatings
by doping them with silica. The new coating is stable, but does not have lower thermal noise
since the process also lowers the refractive index requiring thicker coatings. This result
suggests that we might be able to use other low index dielectrics, such as alumina, as stable
dopants, and that we might be able to have coatings with two dopants. Some of the coatings
being considered are tantala doped with silica and titania, tantala doped with alumina and
titania, and titania doped with alumina and silica. In addition, a cobalt-oxide and silica
coating was found to have a very low mechanical loss, but its optical loss was unacceptably
high. We are considering doping the cobalt layer to see if the optical loss can be lowered
without ruining the mechanical loss.

Finally, we have shown that a high temperature anneal of a silica-alumina coating resulted in
a significantly lower mechanical loss. Unfortunately the loss in that coating was still
unacceptably high. Currently, the coating manufacturers anneal our coatings at the highest
temperature that the coating can withstand. However, we are planning to test if we can lower
the loss by performing an annealing with a very slow cool down.

At present, Initial LIGO, while having nominally reached its design sensitivity, appears to be
limited in the 40-100 Hz band by an as-yet undetermined noise source. At its minimal level,
this noise has an fA-3 dependence similar to suspension thermal noise. Moreover, the
suspension thermal noise had not been directly measured for an isolated Initial LIGO
suspension. Therefore the HWSLG, along with Gregg Harry and Andri Gretarsson, set out to
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

measure the suspension thermal noise and locate possible sources of excess loss. In Fall
2005, Gregg Harry and I began a study of the suspension thermal noise in Initial LIGO.
At HWS we measured the Young's modulus and the mechanical loss for the free and
tensioned suspension wire. The structural (frequency independent) loss was found to be 1.7e-
4, which was about a factor 2 lower than previous measurements. The measurement of the
Young's modulus and the thermoelastic peak allowed us to determine the physical constants
of the wire to improved accuracy. These reduced uncertainties can be attributed to the
superior clamping and isolation method employed by our group.

At MIT, we constructed a large optic suspension that was identical to the system used at the
sites. We suspended a LIGO I pathfinder optic in a spare optics cage using suspension wire
from the same spool used at the sites. We monitored the suspension wire motion using an X-
Y shadow sensor at the top and bottom of both wires. Initial results from measurements of
the violin modes show that the loss in this suspension is about a factor 10 higher than the loss
due to the wires. This loss was also shown to vary with clamp condition. These results
suggest that the Initial LIGO wire loop suspension contains excess noise, most likely from
rubbing friction, and that this suspension thermal noise could make a significant contribution
to the excess noise seen in Initial LIGO.

We are currently investigating alternative wire clamping methods in order to better
understand and reduce this noise. Our goal is to understand this excess noise and if possible
find a means for reducing it. If feasible, we would hope to implement this suspension

The LIGO-Virgo Thermal Noise meeting: LIGO and Virgo have been working to develop
a full, formal collaboration. Thus far, the expression of that collaboration has been almost
entirely within the realm of data analysis. I was of the opinion that there should also be close
collaboration in the technical areas as well. Therefore I proposed the idea to Benoit Mours
that there be a joint LIGO-Virgo meeting on thermal noise. The purpose of the meeting is to
familiarize the attendees with the current research being performed in both collaborations, to
discuss areas of possible collaboration, and to outline avenues of research in preparation for
the next generation of detectors. Researchers in both collaborations have supported the idea
of the meeting. The meeting will occur on October 7 at the Virgo observatory.

Plans

a) Coating Losses
The HWSLG plans to continue its longstanding role in the Coating Program. Our facility acts
as a second test facility, along with MIT, for thin coating samples and is the primary lab for
testing the effects of annealing on coating loss. We work in close coordination with the MIT
Coating group. The Coating Program is very important to the success of Advanced LIGO
since it is the coating thermal noise that is the limiting noise source in the minimum of the
Advanced LIGO noise curve. The goal of our research is to develop a coating in which the
thermal noise will be less than the quantum noise from the laser. The intial goal of the
Coating Program was to understand the source of the noise in the current LIGO coatings. We
discovered that the loss was primarily in the high index, tantala layers rather than the low
index, silica layers (see Penn, et al., Classical and Quantum Gravity 20 (2003) 2917-2928).
While the proportion of loss ascribed to those two materials has changed slightly due to the
direct measurement of the Young's modulus by Sheila Rowan, the essential conclusion
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

stands. Thus our research has focused on finding a high index coating material with low
mechanical loss. We have explored using other high index dielectrics instead of tantala, and
doping the tantala in order to lower its loss. At present the best candidate coating is tantala
doped with titanium.

Our work with CSIRO has shown that we can reduce the mechanical loss in titania coatings
by doping them with silica. The new coating is stable, but does not have lower thermal noise
since the process also lowers the refractive index requiring thicker coatings. This result
suggests that we might be able to use other low index dielectrics, such as alumina, as stable
dopants, and that we might be able to have coatings with two dopants. Some of the coatings
being considered are tantala doped with silica and titania, tantala doped with alumina and
titania, and titania doped with alumina and silica. In addition, a cobalt-oxide and silica
coating was found to have a very low mechanical loss, but its optical loss was unacceptably
high. We are considering doping the cobalt layer to see if the optical loss can be lowered
without ruining the mechanical loss.

One major difficulty in understanding coating loss is in understanding the physical
parameters of the coating layer, which can differ significantly from the bulk material values.
While visiting MIT, the HWS PI initiated a program with Prof. Krystyn van Vliet of the MIT
Mechanical Engineering department. Prof. van Vliet uses nanoindenting to measure coating
parameters. She is currently in the process of measuring various tantala coatings so that we
can compare her results to measurements performed at Glasgow.

Finally, we have shown that a high temperature anneal of a silica-alumina coating resulted in
a significantly lower mechanical loss. Unfortunately the loss in that coating was still
unacceptably high. Currently, the coating manufacturers anneal our coatings at the highest
temperature that the coating can withstand. However, we are planning to test if we can lower
the loss by performing an annealing with a very slow cooldown.

c) Other Contributions

At present Initial LIGO appears to be limited in the 40-100 Hz band by an as-yet
undetermined noise source. This noise has an f^-3 dependence similar to suspension thermal
noise. In Fall 2005, Gregg Harry and I began a study of the suspension thermal noise in
Initial LIGO. We performed free and tensioned wire measurements to determine the inherent
loss in the wire material. We also setup a large optic suspension at MIT using a spare
Pathfinder optic in order to measure the loss in the full suspension. We found that while the
internal friction of the wire was about a factor two lower than previous measurments, but the
loss in the full suspension was a factor 10 higher than the internal loss. Moreover the loss in
the suspension was shown to vary with clamp condition. These
results indicate that Inital LIGO's wire loop suspension contains excess noise, most likely
from rubbing friction, and that this suspension thermal noise could make a major contribution
to the excess noise in Initial LIGO. We are currently investigating alternative wire clamping
methods in order to better understand and reduce this noise. Our goal is to understand this
excess noise and if possible find a means for reducing it. If feasible, we would hope to
implement this suspension upgrade in Intermediate LIGO.

The LIGO-Virgo Thermal Noise meeting: LIGO and Virgo have been working to develop a
full, formal collaboration. Thus far, the expression of that collaboration has been almost
entirely within the realm of data analysis. I was of the opinion that there should also be close
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

collaboration in the technical areas as well. Therefore I proposed the idea to Benoit Mours
that there be a joint LIGO-Virgo meeting on thermal noise. The purpose of the meeting is to
familiarize the attendees with the current research being performed in both collaborations, to
discuss areas of possible collaboration, and to avenue of research in preparation for the next
generation of detectors. Researchers in both collaborations have supported the idea of the
meeting. The meeting will occur on October 7 at the Virgo observatory.

LSU

Joe Giaime continued chairing of Suspensions and Seismic Isolation Working Group, and
continued his roles as a scientific lead of the Advanced LIGO seismic isolation development
team.

There is a seismic isolation BSC prototype now under assembly at MIT LASTI. LSU & LLO
are assembling and testing instrument pods.

Plans

b) Suspension Design for Advanced LIGO
Joe Giaime will continue leading the Suspensions and Seismic Isolation LSC working group.
Joe Giaime will continue scientific and technical (but remote) involvement in ETF and
LASTI. We expect a joint hire with LIGO of a postdoc to work full time on the LASTI
seismic (and other) development experiments.

Moscow

1.1Measurements of variations of electrical charge distribution on a fused silica sample
by means of the new electrometer V.P. Mitrofanov and his post graduate student L.G.
Prokhorov continued to search factors which determine the charge distribution and its
evolution on fused silica samples by means of the specially developed electrometer with
capacitive probe placed under the rotating sample. This electrometer has small influence on
the charge distribution situated on the test mass. The electrometer's sensitivity is about 104
electrons/cm2Hz1=2. It has allowed to find factors which determined formation of the charge
distribution when the fused silica test mass is in air before evacuation of the vacuum
chamber. They have carried out measurements of time variations of electrical charge located
on the fused silica sample. Rare (once per several hours) variations of the charge density of
about 104 electrons/cm2 were observed in the measurements which were carried out in air
[1]. These variations were likely produced by dust charging. The new system of calibration
of the probe electrometer with rotating sample has been also developed and realized. It has
allowed one to improve an accuracy of measurement of electrical charge variations on fused
silica sample. At present the electrometer is modified to carry out the measurements in
vacuum.

1.2 Development of the experimental setup for demonstration of low noise damping of
the suspension violin modes
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

High Q-factors of the mirror internal modes and high light power in the interferometer may
cause the parametric instability due to the interaction of mechanical and optical modes [2].
Also the high quality factor (Q ' 108) of violin modes of the fused silica suspension may
create the instability of the interferometer length control servo as well as the inconvenience
associated with the long ring-down time. Cold damping system can be used to exclude these
effects.

V.P. Mitrofanov andK.V. Tokmakov have proposed a system of cold damping of fused silica
suspension violin modes [3]. They started the implementation of experimental setup for
realization of the violin mode low noise damping using the model of fused silica fiber
suspension. To decrease effects of seismic noise on the damped fiber thermal noise
measurements both ends of fiber were welded to the massive (M = 2.4 kg) monolithic fused
silica frame. Usually, in high Q suspensions developed in previous experiments of MSU
group and other groups, only the upper end of the fiber has been welded to the support
structure while the bottom one was loaded by a free suspended mass. The preliminary
experimental measurements have shown that the system has provided the fiber violin mode Q
of 3107 and a high enough value of the coupling coefficient of fiber violin mode with
bending plate mode. The optical sensor designed for measurement of the fused silica fiber
displacement was developed and realized. In contrast to the shadow sensor it is based on the
refraction of a laser beam which passes through the fused silica fiber. The beam detection
angle depending on the position of the beam relatively to the fiber is measured by means of
photo detector. The maximum sensitivity of the sensor is investigated. Experimental
researches of low noise damping are temporary delayed because Dr. K.Tokmakov has quit
the MSU group and is employed by Glasgow group.

1.3 The investigation of thermal and excess noises in the fused silica plates with
multilayer reflective coating
This part of experimental researches is developed by I.A. Bilenko. It is aimed at the direct
measurement of the mechanical noise in the multilayer mirrors which can make a
contribution to the Advanced LIGO noise budget. At present 3 kinds of effects those, under
certain conditions, may take place in the mirror coatings are considered: •ripples• on the
mirror surface produced by equilibrium thermal fluctuations, relaxation of the stresses
produced by the film deposition process and structure reconstruction caused by high power
optical field (appearing of •precursors• of the laser breakdown). The projected sensitivity of
the sensor based on the 50 mW Nd:YAG laser and Pound-Drever scheme is ' 5x10^-
14cm/rHz at 500 Hz. In the Jan-June 2006 the reference cavity was replaced with the better
one (finesse value F = 1 103). The mode matching optic was improved and sufficient (about
10%) coupling factor is obtained. The feedback system was tested and partly redesigned.
Nevertheless, in order to reach the projected sensitivity more improvements have to be done.
The major efforts are concentrated on the preparation of the measurement cavity with small
movable mirror. The technique of welding and noiseless angle/position adjusting is under
development now.

2.1 The analysis of the actions of the cosmic rays on gravitational wave detectors
V.B. Braginsky, O.G. Ryazhskaya and S.P. Vyatchanin have analyzed the actions of cosmic
rays on LIGO mirrors. The most significant results of this analysis are:
1.    The rare but very powerful cosmic ray showers with energy 0:5 2TeV will leave in
Advanced LIGO mirror an energy 60 230GeV each. Such an event will occur once per' 3106
sec in each mirror. This energy will create a hot •spur• in the mirror bulk and will transfer
mechanical momentum to the mirror [4]. Due to thermoelastic effect hot •spurs• will distort
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

the mirror's shape. Both effects may be observed at the planned Advanced LIGO sensitivity,
but both effects may be •vetoed• in the measurements with two antennae.
2.    The same powerful cosmic rays showers will produce lunches of low energy electrons
(103 104 electrons in each shower) which will be deposited on the mirror (effect of AC
charging) [4]. As mentioned above the jump of charging at the presence of DC charge on the
mirror may also mimic gravitational wave signal.
3.    Low energy cosmic rays with energy 0:2 2GeV will hit the Advanced LIGO mirror
much more frequently: about 15 times per sec at sea level. Vetoing procedure is not possible
at such high rate of events. But the thermoelastic distortion of the mirror will create a small
displacements of the mirror's surface (approximately 103 times smaller than the SQL level)
[5].

Plans
Missing – no MOU

Sannio

1.    - We investigated optimal thickness configurations for multilayer dielectric mirror
coatings, yielding the lowest thermal noise for a prescribed reflectivity. This is an important
issue, as the limit sensitivity of LIGO and advLIGO is set by coating thermal noise. The
current design (alternating high/low index silica/tantala quarter-wavelength layers) yields the
maximum reflectivity for a fixed number of layers, but not the lowest noise for a prescribed
reflectivity.

We first attacked the problem using genetic optimization (GO). We wrote a dedicated
FORTRAN code using perhaps the best (and best documented) public domain GO engine
(PIKAIA). The code is very flexible as it allows to impose several concurrent objectives and
constraints (e.g., minimizing the thermal noise while keeping the optical losses below some
threshold, and the reflectivity at a prescribed level), to use several (more than two) different
dielectric media in the same design, and to accommodate continuous (e.g., the thicknesses) as
well as discrete (e.g., the possible values of the refraction indexes) optimization parameters.
Another set of codes (MATHEMATICA notebook functions) for characterizing the
frequency and angular (polarization-dependent) response of the optimised coatings, and for
computing the average and std. deviation in their nominal figures (reflectivity, noise, optical
losses) due to random errors in the sputtered layers thicknesses were also written.
For the simplest case of thermal noise minimization at some prescribed reflectivity level
using only two different dielectric media (e. g., silica and tantala), the genetically evolved
"optimal" syntheses show a stacked (non quarter-wavelength) doublet structure, except for
the first and last layer which are somewhat different. We made an exhaustive investigation of
these end-tweaked stacked-doublet configurations, and wrote a MATHEMATICA package
for finding the lowest-noise one for a given target reflectivity. Under typical design
conditions, we found that the coating thermal noise PSD could be reduced in this way by
more than 10%, corresponding to some 25% boost in the event rate, under the simplest
assumption of homogenously/isotropically distributed sources radiating most of their GW
energy in the coating-noise dominated frequency band.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

The work done has been documented at various stages in the LIGO literature [1]-[3], and is
the subject of a paper in the SPIE Proceedings [4]. On the basis of our results, four optimized
10cm-diam, 10cm-thick prototype mirrors are being made at the Laboratoire Materiaux
Avancees in Lyon, FR (substrates kindly provided by the Caltech LIGO-Lab). The
prototypes will be delivered to Caltech for testing on the TNI by the end of September.
This activity has been funded for f.y. 2006 by the INFN (Istituto Nazionale di Fisica
Nucleare) Group V, through the COAT project (30KEU funding; Innocenzo M. Pinto, PI).
More recently, the focus of our investigation started shifting toward the problem of getting
reliable and accurate measurements of the mechanical noise properties of the coating media.
Coating (noise) optimization at prescribed reflectivity depends critically on the accuracy
whereby the ratio $\frac{\phi_H Y_H} {\phi_L Y_L}$ is known, $\phi_{H,L}$ and
$Y_{H,L}$ being the mechanical loss-angle and Young modulus of the High-Low index
coating material. Technology related product feature dispersion, and measurement
uncertainties make the value of this ratio still rather uncertain. This ratio is furthermore
critically dependent on the accuracy of the numerical method one uses to compute the stored
energy ratio between the coating and the substrate. Quite recently, both LMA and Glasgow
(S. Rowan) started making mechanical-loss (ringdown) measurements on silica substrates
with various coatings based on a different (cantilever) geometry. For this geometry we
developed a fully analytic ringdown model [5].

[1] I.M. Pinto et al., "Optimized Coatings" - LSC Meeting, August 14th - 17th, Hanford WA,
LIGO G050363-00-R [2] J. Agresti, et al., "Optimized multilayer dielectric mirror coatings
for gravitational wave interferometers," Report LIGO-P060027-00-Z.
[3] J. Agresti et al., "Optimizing dielectric mirror coatings for best tradeoff between
reflectivity and noise," LIGO-T060172-00-R. [4] J. Agresti, et al., "Optimized multilayer
dielectric mirror coatings for gravitational wave interferometers," in Advances in Thin-Film
Coatings for Optical Applications III, M. J. Ellison, Ed., Proc. SPIE, vol. 6286, SPIE,
Bellingham, WA, USA, 2006 (ISBN 0-8194-6365-5), in print.
[5] V. Pierro, I.M. Pinto, "Measuring coating mechanical quality factors in a layered
cantilever geometry: a fully analytic model", Technical Report, LIGO-T060173-00-R.

2.    - We investigated the analytic structure of a continuous family of hyperboloidal
beams introduced by Bondarescu and Thorne which generalizes the nearly-flat and
nearly-concentric mesa beam configurations of interest for advanced LIGO. Capitalizing on
certain results from the optical engineering Literature on flat-top beams, an insight-providing
and computationally-effective representation of the Thorne-Bondarescu family has been
derived in terms of rapidly-converging Gauss-Laguerre wavefunctions. Also, a generalization
(involving fractional Fourier transform operators of complex order) of some recently
discovered duality relations between the nearly-flat and nearly-concentric mesa
configurations has been obtained. The work done is the subject of a LIGO report [1b], of a
paper in the SPIE Proceedings [2b]. A condensed (brief communication) version has been
published in the Phys. Rev. D [3b]. The analysis is now being extended to the higher order
modes, in the perspective of determining the optimal beam profiles from the viewpoint of
both tilt stability and thermal noise minimization.

[1b]V. Galdi et al., "On the analytic structure of a family of hyperboloidal beams of potential
interest for advanced LIGO," Report LIGO-P060003-01-R (also, e-print ArXiv:gr-
qc/0602074).
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

[2b] V. Galdi et al., "Analytic structure and generalized duality relations for a family of
hyperboloidal beams and supporting mirrors of potential interest for future gravitational
wave detection interferometers," in Laser Beam Shaping VII, F. M. Dickey and D. L.
Shealy, Eds., Proc. SPIE, vol. 6290, SPIE, Bellingham, USA, 2006 (ISBN 0-8194-6369-8),
in print.
[3b] V. Galdi, et al., "Analytic structure of a family of hyperboloidal beams of potential
interest for advanced LIGO," Phys. Rev. D73 (2006)127101.

Plans
Missing from MOU.

Stanford

Materials and Thermal Noise (M. Fejer, R. Route)
a) Investigation of dielectric mirror coating mechanical losses
In collaboration with Glasgow, Syracuse, MIT, and Caltech, we have continued our studies
of the Q factor of fused silica samples to investigate and reduce the mechanical losses of
coatings. A paper on this work is currently in preparation.
b) Factors that control the mechanical Q of silicon
We continued a collaborative study with GEO-Glasgow on determining the factors that
control the mechanical Q of bulk silicon resonators. Progress has remained limited on this
task during the current reporting period because of other priorities. However, we have
continued to work on a set of material specifications for the next round of acquisitions and
study of bulk silicon resonators.
c) Fabrication of silicon cantilevers for dissipation studies
To complement measurements of surface-dominated energy dissipation effects with
relatively thin cantilevers (thickness: approx. 45 μm), we have fabricated 500 μm thick
samples for dissipation measurements in the bulk-dominated regime. We purchased 3 mm
and 6 mm thick 4-inch silicon wafers (CZ-grown, boron-doped, 10-20 Ohm cm) from
Siltronix in France, for these studies. A first batch of 500 μm thick cantilevers with 3 mm
thick end-pieces, fabricated from a single wafer by KOH etching, was sent to GEO-Glasgow
group for evaluation. (Thick end-pieces are required for the reduction of losses, e.g. due to
stick-slip effects, at the clamped end of the cantilever.) Poor quality of the masking silicon
nitride layer in combination with the long processing times required led to the generation of a
relatively large number of defects (inverse pyramids) during the KOH etch of 3 mm thick
wafers. These defects may be a source of excess mechanical loss at the clamping interfaces.
As an alternative to KOH etching, cantilevers with thick end-pieces were produced by silicon
fusion bonding. Diced 3 mm thick end-pieces were bonded to both sides of one end of a
cantilever that was diced from a 500 μm thick wafer. In this initial set of bonded cantilevers,
IR light microscopy revealed large non-bonded areas at the interfaces, presumably due to
particle contamination from the dicing processes. Samples of these cantilevers were sent last
year to Glasgow for investigation, while improved fabrication methods are being developed.
A paper on first results from this work, "Mechanical Dissipation in Silicon Flexures", S.
Reid, G. Cagnoli, D.R.M. Crooks, J. Hough, P. Murray, S. Rowan, M.M. Fejer, R. Route and
S. Zappe, has now been published: Physics Letters A; vol.351, no.4-5, p.205-11,(6 March 6,
2006)

Active Alignment, Isolation, Control and Suspension Design
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

(D. DeBra, B. Lantz, N. Robertson, W. Hua, C. Hardham (now at SQUID Labs), W East,
T Casebolt, M DeGree)
Norna Robertson has continued to serve as Cognizant Scientist for the Suspensions
We have continued to work on the design and development of suspensions for Advanced
LIGO, and to work on general issues of integration of suspensions with other subsystems
including the isolation systems. This work has been done in collaboration with colleagues in
the LIGO Scientific Collaboration in the USA at Caltech, MIT, the two LIGO observatories
and LSU, and in the UK at the University of Glasgow, Rutherford Appleton Laboratory,
University of Birmingham and University of Strathclyde.
d) Testing and characterization of the ETM controls prototype
During this period, modeling, design work and interpretation of behavior of quadruple
pendulum suspensions for supporting the test masses in Advanced LIGO continued. In
particular the construction of an all-metal controls prototype quadruple suspension was
completed at Caltech, and preliminary investigations of its behaviour including pitch stability
and mode frequencies were carried out. Norna Robertson visited Caltech in October 2005 to
work in the lab with Caltech colleagues and review progress there. Several findings were
made at Caltech concerning the behaviour of the suspension that revealed limitations to our
earlier models and that have now been incorporated in fuller models. In particular it was
found that the finite stiffness of the cantilever blades in the direction transverse to their long
axis was such that, with the original design, the quadruple pendulum was unstable in pitch,
and how to compensate for it was subsequently understood by using finite element and
physical analysis. The controls prototype has since been disassembled, cleaned, reassembled
and shipped to the LASTI facility at MIT in early February 2006 where it is now undergoing
further testing and characterization including measurements of transfer functions and use of
active damping for local control.
e) Design of the noise prototype ETM/ITM
The design of the noise prototype ETM/ITM quadruple suspension is now under
development in the UK, and builds on the lessons learned from the controls prototype.
Work has also been carried out on theoretical considerations of gas damping in the small gap
between the test mass and its reaction mass in the quadruple suspension system, to look at
whether such damping could be a significant noise source. A squeezed-film damping effect
in small gaps has been modeled and observed by researchers using microbeam resonators.
Applying this model to the Advanced LIGO situation, we concluded that this effect alone
would not significantly increase the suspension thermal noise. However if there was
significant outgassing of the surfaces in the gap, raising the pressure by a factor of 10 above
the ambient, the displacement due to suspension thermal noise would be increased by ~ 35%
at 10 Hz. A paper on these findings has been written for the DCC. (Some Notes on Gas
Damping in Small Gaps as a Potential Noise Source for Advanced LIGO Suspensions, Norna
A Robertson, DCC # T050241-00-R).
The suspension thermal noise section of IFOMODEL used in BENCH was updated in May
2006 to reflect the current set of parameters for the quadruple suspension.
A document addressing suspension design requirements for the test masses for degrees of
freedom not covered in the conceptual design document was prepared and submitted to the
DCC. (T060155-00-R)
f)    Design of other BSC and HAM suspensions
1) Input modecleaner design
During a review of the ribbon/fibre and bonding aspects of the quadruple suspension design
(see section f below) questions were raised about a proposal to change from using silica
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

fibres to using steel wires for the final stage suspension of the modecleaner mirrors whose
noise requirements are significantly relaxed from those of the test masses. Norna Robertson
prepared a response for a second review, and a paper was written and submitted to the DCC
laying out the case for the use of steel wires. (References: T060014-00-R and T060008-00-
R.). Following this review, the decision was taken that the baseline be changed to steel wires.
A summer student, Chris Cueva, is setting up an experiment in the laboratory at Stanford to
look at how well violin mode frequencies for a modecleaner suspension on steel wires can be
matched, with a target of keeping the spread to 2%.
2) Recycling mirror design
Some work on modelling recycling mirror suspensions has been done during this period
while the recycling cavity design is being reconsidered. MATLAB modeling work was
carried out to consider reducing this suspension from a triple to a double pendulum. The
baseline remains a triple. Various parameters such as mirror size and overall length are still
under discussion.
3) Output modecleaner design (OMC)
Work has begun on the conceptual design of a double pendulum suspension for the output
modecleaner. Our summer student Chris Cueva has been carrying out modeling work, and a
document summarising an outline design incorporating this work was posted on the OMC
wiki site for discussion at an OMC design meeting. Further modeling resulting from these
discussions is underway.
4) Pick-off mirror for BSC chamber
Modelling of a conceptual design of double pendulum suspension for a pick-off mirror has
begun, and an initial set of parameters checked for resonant frequencies and mode coupling
to allow damping at the upper mass.
Norna Robertson visited Caltech in June 2006 for discussion with LIGO colleagues on
various suspensions including B C and D above.
g) SUS and SEI design reviews and planning meetings for BSC and HAM optics;
Two reviews pertinent to the suspension design have been held in the past year. In October
2005 a preliminary design review was held on the ribbon/fiber and bonding aspects of the
design, which also raised questions about the use of steel wires for the mode-cleaner (see iii
A above) Norna Robertson participated in this review and Roger Route served on the review
committee. In June 2006 a quad mechanical preliminary design review was held. Brian Lantz
and Norna Robertson served on the review committee. Norna Robertson also served on the
review committee in the HAM SAS passive isolation system review during Dec 05 - Jan 06.
Brian Lantz and Norna Robertson participated in the annual NSF LIGO review in Nov 2005
and the NSF Advanced LIGO review in May/June 2006.
h) Integration of suspension and isolation systems
We have been working with the Suspensions team at Caltech and the Rutherford Appleton
Laboratory to study the interactions between the support frame for the quadruple pendulum
and the second stage of the isolation system. The suspensions team shipped one of the full
scale frames to the Stanford ETF, and Brian Lantz conducted a series of measurements to
study the impact that the frame resonances have on the performance of the isolation system.
There were several important results which came from this work:
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

Figure 1 Tarmigan Casebolt adjusts the damping strut (diagonal element jutting down to the
left) as it attaches to the pendulum frame (large aluminum structure on the right side of the
picture).
1) The control system can work with the frame resonances exactly as they are, but the
performance is not as good. The 10 Hz isolation is about 30% worse, and there are
big transmission peaks at the flexible modes of the frame, which make the isolation much
worse at those frequencies.
2) The frame resonances can be actively damped to decrease the coupling into the
control loop by about almost a factor of 20. This damping can be accomplished with the
existing sensors and actuators, but it is tedious and makes the system less robust to
plant changes. The active damping performance could be improved by using signals from
small inertial sensors placed at the free end of the pendulum frame. However,
installing new sensors at this location is not a trivial undertaking, due to
the number of systems-level requirements such objects would need to adhere to since they
are very close to the main optic and the main beam.

At the urging of Prof. DeBra, we began an investigation of damping struts which could be
added to the system to help damp the frame resonances. We used a lightweight aluminum
strut installed at an angle between the main optical table and a point near the tip of the
pendulum frame. One end of the strut was rigidly attached to the optics table, and the other
end was attached to the pendulum frame with a layer of visco-elastic damping material. This
configuration created a constrained-layer damping mechanism, with the visco-elastic
damping material constrained between the pendulum frame and the strut. This work was
carried out by Tarmigan Casebolt. The constrained layer damping improved the performance
of the pendulum frame quite dramatically. We tried several different sizes of damping
material, and found that the performance of the damping strut improved as the constrained
layer got larger. This behavior is expected, but improvements will not continue forever,
because if the layer gets too stiff, very little of the vibrational energy will be stored in the
layer, which limits the amount of energy which the layer can dissipate. The measurements
are described in LIGO document G060007, "Results of the Frame Interaction Study at the
ETF".

Impact of Frame Damping on Stage 2 rotation mode rX
original
frame clamping 2.56 cm?
frame damping 5.11 em?
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and
2007 plans
~~ frame damping 10.2 cm?

1CD
60   70
(req(Hz)
Figure 2 Performance
of                                                            the constrained layer
damping strut. The
purple curve shows the
original coupling of the
61                                                            Hz frame resonance
into                                                          the tilt control loop.
The                                                           three green curves are
the                                                           mode coupling with
different sizes of
damping material. As
the                                                         stiffness of the damping
material increases, the frequency of the mode increases and the Q decreases. The
largest piece we tried showed the best performance.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

j) Improvements in the isolation performance of the ETF Technology Demonstrator at
1 Hz and at 10 Hz.
We improved the 1 Hz performance of the Technology Demonstrator by making two
changes. First, we improved the noise of the GS-13 inertial sensor used to control the second
stage. This is described below.
Second, we realized that the 1 Hz horizontal translational motion at the witness sensor is a
result of the rotational motion of stage 2 about its rotation center. The coupling results from
the fact that the witness sensor is placed on the optics table surface, and is therefore about 40
cm above the rotation center. By improving the pitch and yaw isolation of stage 2 at 1 Hz,
using the same blend filters designed for the single stage HAM platform, we reduced the
rotational motion, and improved the isolation performance, as seen in figure 3.

Horizontal FIR blending performance X

freq Hz
Figure 3 Horizontal performance of the ETF Technology Demonstrator at night. The 1 Hz
isolation performance is better than the requirement, and almost gets to the motion
requirement, despite the large motion of the floor.

k) Improvements to the sensor performance of the ETF Technology Demonstrator.
We replaced the preamp of the GS-13 with a simpler design that has lower noise, as shown
on the next page in figure 4. Several of these preamps have been built for us by Caltech, and
we have re-instrumented the 6 feedback and the 2 witness GS-13s on the Tech Demo. Using
W.S. Hua's coherent subtraction technique, we estimate that the noise floor at 1 Hz is about
6e-12 m/rtHz and about 2e-13 m/rtHz at 10 Hz. The performance shown in figure 3 indicates
a directly measured performance of 2e-11 m/rtHz at 1 Hz.
T060198-00-Z Compilation of LSC 2006 Instrument Science
achievements and 2007 plans
10'1
STS-2 wit X GS13witG GS13witW STS wit X noise GS13 noise G
GS13 noise W STS Input noise GS13 Input noise

.
.
Compare the witness
sensors, noise floor
statistically corrected
Freq (Hz)
Figure 4 Noise estimates of various sensors.
The purple and gold curves are the noise
estimates for the 2 witness GS-13
seismometers fitted with the new Advanced
LIGO preamplifiers. The spike at 1.2 Hz is
an artifact of the data collection, not the
preamp.

l) Techniques for efficient implementation
of control applications at LASTI, and for
We have written an easy-to-use Matlab tool
to design the banks of notch filters needed for the control loops. The system was written by
William East. The program is designed to simplify one of the most tedious parts of the
control loop design process: the implementation of notch filters to reduce the loop gain at
frequencies near mechanical resonances of the plant. The resonances could destabilize the
system if they are not notched, so this design step needs to be repeated for each degree-of-
freedom of every platform in Advanced LIGO. This works out to be 270 banks of filters. It is
clear that this work should dramatically improve the commissioning time for Advanced
LIGO. A screen-shot of the design tool is shown in figure 5.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

Figure 5 Screenshot from the automatic notch filter design tool. The red curve is the filter set
that has been generated. The dark blue curve is the open loop gain before filtering, and the
light blue curve is the gain after the filters have been applied.

Matt DeGree has implemented a system which allows us to check the status of all the sensors
and actuators for the two-stage Technology Demonstrator isolation and alignment system in
the Stanford Engineering Test Facility. It is called the "Tickle Tester" because it applies a
small signal to each of the 12 actuators in turn, and then measures the system response at the
local sensors. By comparing the sensors near each actuator to each other, to the sensors at
other actuators, and to the recorded values from previous tests, we can automatically confirm
that all the sensors and actuators are working correctly, or quickly isolate the malfunctioning
component. This system will be used as a regular health check at the Stanford ETF, and we
plan to evolve its function into an easy-to-use diagnostic system for Advanced LIGO.
m) Work on the prototype isolation system
We worked in collaboration with the LIGO Lab and LASTI to finish out the ASI contract.
We will help the LASTI lab develop implementation and testing plans for the next prototype
isolation system.
n) Support for the installation and testing of the BSC isolation and alignment system
to be installed at LASTI
The next prototype for Advanced LIGO is now in production. Near the end of the previous
reporting period, we requested a design change from the design contractors. The design has
now been modified, the system has moved into production. The production is being managed
by Ken Mason at MIT. Parts of that isolation stage have begun to arrive at MIT, and the team
there is beginning to assemble the system.
o) Isolation system for the HAM chamber
There has been significant work on the isolation system for the HAM chamber.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

We studied the possibility of simplifying the design and reducing the performance of the
system, in response to a significant relaxation of the isolation requirements for the optics in
the HAM chambers. In the summer of 2005, there was a meeting at Caltech to discuss the
possibility of relaxing the requirements for the auxiliary optics for Advanced LIGO (not the
main interferometer sensitivity), and the implications for the seismic system. At that time,
simple estimates indicated that a single-stage isolation platform in the HAM tank could
provide the performance required for Advanced LIGO.
We have developed a new conceptual design for a single stage HAM system. The results of
that study can be found in LIGO-G060190 "Single Stage HAM for Advanced LIGO:
Performance Modeling" by Brian Lantz, et al. This conceptual design was presented to LIGO
in April 2006, and was subsequently chosen as the new baseline for Advanced LIGO,
representing a substantial cost and complexity savings for Advanced LIGO.

Figure 6 Conceptual design of a single stage HAM isolation system by Corwin Hardham of
Squid Labs.
Our modeling shows that this system, which has passive isolation resonance frequencies
between 1 and 2 Hz and a control system like that in use on the Stanford ETF Technology
Demonstrator, can meet the proposed requirements. The modeled system can achieve a 10 Hz
motion of about 1.4e-11 m/rtHz, and an rms motion between 0.6 Hz and 100 Hz of about
1.4e-10 meters, as shown in figure 7. When the triple pendulum supporting the mode cleaner
optic is included, the motion of the mode cleaner optic at 10 Hz is about 2.2e-16 m/rtHz, as
shown in figure 8, nearly a factor of 2 below the motion requirement of 4e-16 m/rtHz.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and
2007 plans
-total
-stg1 disp -stg 1 GS-13 -HEPlK -HEPIry - HEPI rz ■ rms total
proposed req
acceptible motion curve by Peter F.
• rms proposed req
Noise contributions to Horizontal Suspension
Point Motion, with RMS
10°                                 freq (Hz)

Figure 7 Performance estimate of the single
stage HAM for Advanced LIGO, showing the
expected horizontal
motion at the
suspension point of
the                                                         triple pendulum. This
motion is dominated
by                                                          the translation and the
pitch of HEPI system.
Longitudinal Motion
of                                                          the mode clearer
triple
10
10
to
10
freq                                                          (Hi)
Figure 8
Performance estimate
of                                                            a mode cleaner triple
pendulum on the
proposed single stage
HAM isolation
system.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

p) Design for a coaxial double-pendulum
The coaxial double pendulum is a suggested improvement to the mirror suspension system
for the optics in the LIGO Interim Upgrade. The interim upgrade is a proposed series of
upgrades to the Initial LIGO detector which can give modest…

Figure 9: Front view of coaxial double pendulum. The test mass is blue, and the intermediate
mass is grey. The physical length of the system (optics table to test mass center) is the same
as Initial LIGO.

…improvements to the sensitivity before we can install the Advanced LIGO detector. The
coaxial double pendulum looks a bit different than the multiple pendulums envisioned for
Advanced LIGO. Instead of locating the intermediate mass part way down between the
support point and the final mass, the intermediate mass looks like a hollow tube, and it is
placed around the final mass. Thus, the two stages are coaxial. A retrofit for Initial LIGO
could use this feature to improve the suspension system, yet leave the test mass (the
interferometer optic) in its current location in space. To apply forces to the test mass, one can
use actuators which push between the intermediate mass and the test mass. Similar
configurations have been studied before. In particular, our system bears some resemblance to
the system studied by M. Stephens at MIT. The work at Stanford was pursued by an
undergraduate, Mike Thielvoldt. We modeled a system which would have better thermal
noise than the optics in Initial LIGO, and have improved seismic isolation in the gravitational
wave band. By using two sets of actuators - larger force actuators on the intermediate stage,
and low-force, low-noise actuators on the test mass, one should be able to both improve the
range and lower the noise on the test mass. We made some preliminary designs, and modeled
the performance of such a system. Preliminary results look promising but two issues have
arisen which give us some concern. First, the coaxial nature of the system leads to large
cross-coupling between the longitudinal drive on the test mass and the pitch of the test mass.
This generates modes in the longitudinal motion which are at several hertz - in the band
which we are trying to lower the rms motion. The second issue is that such an improvement
will be both disruptive and expensive, issues of great concern to the group which is planning
the Interim Upgrades. Never-the-less, we feel that having pushed the plans ahead far enough
to see where the real issues lie is a critical step in making the correct choices for the upgrade.
Due to the limited scope of the current plans for the upgrade, this work has been ended for
now.
q) Temperature monitoring of the actuators in the ETF Technology Demonstrator.
William East has also completed a test of the thermal response of the in vacuum actuators.
One of the vertical actuators for the Technology Demonstrator, the two-stage active isolation
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

platform in the Stanford ETF, was instrumented with sensitive thermal readouts. The
temperature rise of the actuator was measured under a variety of loading conditions. Using
these measurements and scaling them to the platform designed for the BSC tanks in
Advanced LIGO, we calculated that in normal operation the temperature rise of the
Advanced LIGO actuators should only be a few microdegrees Celsius. This result indicated
that the set of glues we plan to use in the Advanced LIGO actuators should not pose a risk to
the vacuum cleanliness requirements. This work is described in detail in the LIGO document
T060076.
r) Improved vertical seismometer design
We have begun work on a new vertical seismometer. The goal of this work is to improve our
ability to detect tilt between 10 mHz and 100 mHz by using a differential pair of spatial
separated instruments, and to improve our sensitivity to vertical motion at 10 Hz to 100 Hz.
The idea for the instrument is to build a rotational device with a pivot at the center of
buoyancy of the proof-mass, but with a center of mass displaced to one side. An off-center
offload spring supports the load of the proof-mass. Thus, vertical ground motion induces tilt
in the proof-mass. This configuration allows us to cancel the impact of air pressure
fluctuations on the proof-mass location, one of the two largest sources of noise in this type of
instrument. Aaron Streets, a rotation student in Applied Physics, began the design and
construction of a bench-top instrument of this type. We have begun to assemble that
instrument, see figure 10. The goal of the bench instrument is to allow us to get experience
with one of these devices, and try out different configurations of springs, readouts, actuators,
offload springs, etc. Ke-Xun Sun has begun investigating the use of an optical lever with a
grating mirror to use for the readout. Work on the instrument is now being conducted by
Tarmigan Casebolt.

Figure 10: Bench-top vertical seismometer under assembly.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

Plans

a) Coating Losses
Coatings, Materials and Thermal Noise (M. Fejer, R. Route)
Investigations into the level of excess loss introduced by dielectric mirror coatings applied to
test-mass substrates; Reduction of the mechanical loss associated with coatings applied to
substrates and associated thermal noise remains an important research area for
Advanced LIGO and is vital for the success of any future detectors with sensitivities better

We will evaluate in collaboration with Glasgow the usefulness of silicon cantilevers with
deposited dielectric thin films for the estimation of dielectric mirror coating mechanical
losses.

In collaboration with Glasgow, MIT, Syracuse and Hobart and William Smith Colleges, we
will continue our studies of coated substrate materials to investigate and reduce the
mechanical losses of coatings. Participation as required in modeling efforts on the effects of
inhomogeneous mechanical losses on the expected thermal noise from a finite sized test
mass;

Design and fabrication, using Stanford‘s extensive MEMS technology, of custom flexures
from doped and non-doped single crystal silicon for evaluation as suspension elements for
test masses; This work will be carried out in collaboration with Glasgow.

Continued optical and Q measurements on crystalline materials; In collaboration with
Glasgow we will continue our investigations of the optical and Q factors of samples of single
crystal silicon with varying dopants and cut along different crystallographic axes.
Contingent on measurements currently underway at Glasgow, we may fabricate additional
cantilevers for the measurement of bulk-dominated loss effects. The quality of the silicon
nitride masking layer and the cleanliness of surfaces in the case of silicon fusion bonding,
respectively, will be improved.

b) Suspension Design for Advanced LIGO
Active Alignment, Isolation, Control and Suspension Design (D. DeBra, M. DeGree, B.
Lantz, N. Robertson, T Casebolt, G Allen)

Norna Robertson will continue to serve as Cognizant Scientist for the Suspensions subsystem

We will carry out further design, development and characterisation of suspension systems for
LASTI, for LIGO post S5 and for Advanced LIGO, and we will carry out further design,
development and characterisation of the seismic isolation systems at the ETF, at LASTI, and
for Advanced LIGO in general. We will:
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

Work with other members of the SUS team on general design issues as they arise, and in
particular on aspects of
i) the characterization and interpretation of results from the ETM controls prototype
quadruple suspension currently being studied at LASTI,
ii) the development and interpretation of results from the noise prototype ETM/ITM
quadruple suspension, due to delivered to LASTI in 2007,
iii) the modeling and design of other BSC and HAM suspensions as required;
This includes suspensions for Advanced LIGO such as the beamsplitter and recycling mirror
triple suspensions, and the output modecleaner suspension for post S5 LIGO.

Start to look at aspects of suspensions for future detectors in collaboration with the GEO-
Glasgow group; The use of large (few hundred kg) composite silicon test masses with
suitable low thermal noise suspensions for operation at cryogenic temperatures is under
consideration for application in future detectors. We will consider the design of seismic and
mechanical isolation and control systems for such suspensions

Work with SUS and SEI colleagues on issues of integration of suspension and isolation
systems, and on issues of integration with other subsystems; This will include collaborating
on tests of the ETM/ITM noise prototype in conjunction with the BSC active isolation
platform at LASTI.

Continue efforts to improve the isolation performance of the ETF Technology Demonstrator
at 1 Hz and at 10 Hz;

Implement a MIMO controller in addition to the SISO controllers now in use on the ETF
Technology Demonstrator to see if the performance can be improved, and evaluate the utility
of such a control layer for Advanced LIGO;

Continue to develop methods and transfer knowledge to speed up the implementation of
control application at LASTI, and for Advanced LIGO;

Support the installation and testing of the BSC isolation and alignment system currently
being installed at LASTI;

Work with the LIGO lab to develop a final design for the single stage HAM system and have
that design built;

Finish the construction and begin testing the new vertical seismometer. We plan to build the
mechanism, implement closed loop control, and implement at least one form of optical

Participate in upcoming SUS and SEI design reviews and other reviews and planning
meetings associated with the Advanced LIGO activities.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

There are a number of tasks which we would like to pursue, in consultation with the SWG,
given sufficient time and additional manpower. These include:
- Control re-allocation and global control studies of the ETF technology demonstrator, using
a capacitive displacement sensor located between the second stage and the support structure.
- Feed-forward isolation of the first stage of the ETF Technology Demonstrator to reduce the
motion around 10 Hz from tip and tilt, using vertical seismometers.

All of the above will be carried out in collaboration with colleagues in the LSC including the
LIGO lab and the GEO group.

Trinity

A great deal of progress has been achieved by Dennis Ugolini and undergraduate Robert
McKinney in the development of a probe to measure surface charge on LIGO optics, as
detailed in the collaboration note T060140-00-R. To summarize: Kelvin and capacitive
probes work by modulating the capacitance between a charged sample and the probe, causing
a current flow to and from the probe element which can be read as a voltage. Most
commercial probes modulate the capacitance by vibrating the probe tip, and can cost $7,000 or more. Our goal, based on an idea from Rai Weiss, was to modulate with an optical chopper, alternately exposing and occluding the probe from the sample. Both a rotary chopper from Stanford Research Systems and a tuning-fork chopper from Boston Electronics were used, both costing$1,000 or less. The tuning-fork chopper has the
advantages of smaller size and ultra-high vacuum compatibility.

- Both probe-chopper systems have been proven effective and calibrated by using a
conducting sample with a known applied voltage. The tuning-fork chopper system has
proven more difficult to work with. The chopping action is created by an alternating
magnetic field generated by a current coil, which pushes and pulls the two chopper arms.
This coil unfortunately blares noise like a radio antenna at the signal frequency, resulting in a
significant, fluctuating signal even when no sample is present. Since the collaboration note
was prepared, however, we have found that we can measure a signal at *twice* the chopping
frequency with far less noise. The reason appears to be related to the chopper's 50% duty
cycle--during the "closed" part of the cycle, the blades pinch together and overlap, allowing
the electric field to fringe around the outer edges of the chopping blades and generate a
voltage at the probe tip. Measuring at twice the chopping frequency has improved the
sensitivity of this probe system by an order of magnitude.
-
- We can deposit charge on a dielectric surface, either through "stamping" with a metal bolt
or Allen wrench, or by rubbing with a wig or piece of felt, and watch the magnitude of the
surface charge decay over time as the charge spreads out across the dielectric surface. The
decay curve can be approximately but not well fit with an exponential shape (in fact, a log
shape appears to work better in most cases), and the time constant varies by nearly an order
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

of magnitude from one day to the next for the same sample. We have ruled out changes in
temperature and humidity, but still expect that the differences in decay times are related to
charge exchange in the air. An all-metal vacuum-compatible probe has been assembled, and
a chamber and pumps have been loaned to Trinity University from MIT, arriving just within
the last few days. Thus the next steps will be (a) to move the apparatus into vacuum, and (b)
start taking correlation time measurements with real LIGO optical material.

Plans

b) Other Contributions
Charging Research
The group will continue to study surface charging of optics, to better understand their
potential noise contribution to Initial and Advanced LIGO. The measurements over the
coming year will involve the tuning-fork chopper probe developed at Trinity University, as
well as a commercial Kelvin probe from Besocke Delta Phi GmbH, both operating in
vacuum to better simulate the LIGO optical environment. Goals over the coming year will
include:
- Commission and calibrate both probes in high vacuum with a conducting sample and
known applied voltage.
- Deposit charge on a dielectric surface and measure the decay of surface charge magnitude
over time in vacuum. Understand the shape of the decay curve, and extract a correlation time
constant.
- Measure the magnitude of charge buildup on the surface of a LIGO optic in vacuum over
time.
- Deposit charge on the LIGO optic and repeat the correlation time measurement.
- Repeat the charge level and correlation time measurements for different optical materials
(sapphire) and coatings (silica/tantala, silica/titania-doped tantala, others under
development).
- Repeat the charge level and correlation time measurements for optics subjected to different
cleaning and handling procedures, including contact with liquinox, methanol, isopropyl
alcohol, de-ionized water, carbon dioxide snow, ionized dry nitrogen, First Contact cleaning
solution, laboratory gloves, and low-lint wipes.

LIGO Lab
SUSPENSIONS (Romie)
Helena Armandula, Mark Barton, Doug Cook, Dennis Coyne, Jay Heefner, Mohana
Mageswaran, Ken Mailand, Rich Mittleman, Dave Ottaway, Janeen Romie, Laurent Ruet,
Brett Shapiro, Oddvar Spjeld, Bob Taylor, Calum Torrie all are active in the suspensions
work.
Status and Accomplishments
LIGO Lab SUS personnel delivered a quadruple pendulum controls prototype to LASTI in
the beginning of 2006. Prior to this, the suspension with its full reaction chain, had been
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

UHV cleaned, fully assembled, suspended and characterized at Caltech before being shipped
in two assembled sections to LASTI. While at Caltech, extensive suspension modeling was
done to address two anomalies that showed up in the LASTI prototype. An instability in pitch
turned out to be due to lateral compliance in the blades. The vertical frequencies were lower
than expected due to an anti-spring effect from the angled wires. A great deal of finite
element modeling and experimental work was done on the suspension structure design and
prototype to provide modal information to the SEI group and optimization paths for the UK
group for the noise prototype structure. Mathematica models for the controls and noise
prototypes were updated to include these new findings and posted on this web site
http://www.ligo.caltech.edu/~mbarton/SUSmodels/index.html. A detailed assembly
procedure was created for the controls prototype with recommendations for the noise
prototype, http://www.ligo.caltech.edu/docs/T/T060039-00/T060039-00.pdf.
At LASTI the suspension was cartridge-loaded into the BSC chamber. SUS personnel from
Caltech, LLO and the UK groups came to support LASTI personnel in assembly, test,
installation and alignment. Quad testing includes functional electronics test, pendulum
frequency measurements, active damping test and eddy current damping tests. SUS also
delivered installation tooling to move the suspension about the lab, mount its sections to a
test stand and install it into the chamber. LIGO Lab is continuing to design quad installation
fixtures in the form of an articulating arm fixture to be attached to the BSC chamber door
flange. Fabrication of this arm for LASTI is ongoing.

In October 2005, there was a Ribbon/Fiber/Ear/Bonding Preliminary Design Review, for
which the bulk of the preparation work was done by the Glasgow SUS group. In March 2006,
most of the SUS team attended the LSC meeting and a SUS Workshop. Many topics at the
workshop were ribbon/fiber/ear/bonding related and the actions out of that workshop are
tracked at the weekly SUS meetings on Tuesdays. There is a weekly SUS Design meeting on
Tuesdays as well for the designers and engineers. Subsequent to the UK Electronics Review
last summer, LIGO Lab electronics engineers have been coordinating with their UK
colleagues on electronics interface control issues and refinement of electronics requirements.
In July, all SUS personnel participated in the Quadruple Pendulum Suspension Preliminary
Design #3, for which the primary preparation work was carried out by the RAL UK group.
Design of the quad noise prototype is ongoing in the UK with support from the US SUS
members. Work is continuing on a number of HAM Cavity suspensions as discussed in the
next section.

Plans for the Near Future
LIGO Lab SUS personnel will continue to support LASTI in testing the quad controls
prototype and the UK in designing the noise prototype. We are currently working on
conceptual designs for the AOS subsystem‘s ETM double pendulum telescope suspension
and the pick-off double pendulum suspension. SUS folks are working with IO folks in
refining the recycling mirror triple pendulum suspension requirements. In addition to this, we
are developing an output mode cleaner double pendulum suspension for Enhanced and
Advanced LIGO. Plans to prototype an Advanced LIGO output mode cleaner suspension are
in place and this prototype is scheduled to be installed in the LIGO interferometers under the
Enhanced LIGO plan.
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

In the fall, we will be holding a Preliminary Design Review of the HAM Cavity suspensions,
including the recycling mirrors and the input and output mode cleaners. This review will
include test results from LASTI. This will kick off the final design effort for the HAM
Cavity suspensions. There will also be a US suspension electronics preliminary design
review this fall.

Suspensions in Initial and Enhanced LIGO – SUS (Harry)
Work has been done at MIT on better understanding mechanical loss in the wire suspensions
used in initial LIGO in collaboration with LIGO visitor Steve Penn from Hobart and William
Smith Colleges. This is both to better predict the suspension thermal noise for use in initial
LIGO commissioning and to explore ways to improve this noise for enhancements to initial
LIGO. It has been found that the torque on the screws holding the wire clamps does affect
the mechanical loss, but no level of torque has been found which allows the material limits to
be reached. Fluctuations in room temperature have been ruled out as an explanation for the
changing violin modes seen at the sites. Preliminary work has begun on a wire collet based
alternative clamping mechanism to reach the material limits. Early results from this were not
an improvement on the clamps, but modifications to the collet are being explored.
Redesigning the clamp block to use larger clamps that can take more torque is also being
explored.

The fabrication and procurement of the mechanical components is complete for the BSC
seismic isolation, with all parts delivered to LASTI. The pod subassemblies have been
shipped directly to LLO with instruments for assembly and filling of the pods with a
traceable inert gas. The assembly at LASTI has been completed ―in air‖ in preparation for
both functional testing and tuning of the structure.
A decision has been made to design the HAM seismic isolation in a single-stage instead of
two-stage like the BSC seismic isolation. The mechanical design has been started by Corwin
Hardham.
In the next 12 months we plan on doing the following:
1.    Complete the testing of the BSC seismic isolation. Following testing the structure will
be disassembled for cleaning to vacuum requirements and reassembled onto the test stand.
2.    Install the noise prototype quad to the optics table and insert both the seismic isolation
and quad assembly into the BSC chamber
3.    Characterize both the seismic and quad and implement control strategies. Write
assembly procedures and complete all fabrication drawings.
4.    Finish the design of the single stage HAM seismic isolation and place contracts for the
fabrication of the prototype
The seismic isolation development team consists of Joe Giaime, Brian Lantz, Ken Mason,
Richard Mittleman, Ben Abbott, Jay Heefner, Myron MacInnis, and Brian O‘Reilly.

Isolation/Suspension R&D (DeSalvo)
A passive Seismic Attenuation System was designed and is being prototyped, based on the
SAS technology. The single-stage, in-vacuum, passive pre-isolation system presently being
assembled is expected to deliver the similar seismic attenuation of the three stages (external
hydraulic, and the two in-vacuum electromagnetically driven) of the present ad-LIGO
T060198-00-Z Compilation of LSC 2006 Instrument Science achievements and 2007 plans

baseline active pre-isolation system. All components of the system have been separately
prototyped and tested, and delivered attenuation performance exceeding the requirements.
The fabrication of a full HAM-sized prototype has been completed and assembly has started.
Commissioning of the test unit at LASTI is scheduled for the fall. The project advancement
can be viewed in the web pages below
http://www.galliemorelli.com/eng/contacts.html
http://www.ligo.caltech.edu/~coyne/AL/SEI/HAM_SAS/HAM-SAS_development.htm
http://www.ligo.caltech.edu/~citsas/HAM-SAS/
http://www.ligo.caltech.edu/~desalvo/HAMSAS/

Alberto Stochino (undergrad student) Studied compensation mechanisms for inertial
saturation of the 1/f^2 spring attenuation mechanisms. Found a solution that improves the
GAS spring performance from 60 to 80 dB.

Maria Paola Clarizia (undergrad student) Studied the possibility of remote, wire-less static
tuning and positioning of suspensions.

Marie Giron (undergrad student), Michael Floyd (high school student), Michael Koyfman
(high school student) Studying the loss behavior of glassy metal springs. The glassy metal
GAS springs show an unusual loss mechanism, its understanding could make glassy metals
useful for advanced seismic attenuation and mirror suspensions.

Juri Agresti Made a PhD thesis on thermal noise analysis of mirror for Advanced LIGO,
including comparative analysis of standard Gaussian and Mesa beam performances.

TNI (Black)