Transcript LISA 7th Symposium Poster 05-20

LISA STUDIES AT THE UNIVERSITY OF COLORADO
Michael J. Nickerson, Ellery B. Ames, John L. Hall, and Peter L. Bender
JILA, University of Colorado and NIST, Boulder, CO
Current LISA Projects at JILA
Dual-Cylinder Laser Reference Cavity Design
Laser Reference Cavity
Cavity Construction
Our current primary project is a flight-compatible laser stabilization reference cavity. It is a
dual-cylinder ULE cavity with reduced sensitivity to thermal fluctuations and vibrations in
the millihertz frequency regime of interest to LISA.
Voltage Reference Stability
A recently concluded project verified the stability of commercial voltage references at
millihertz frequencies.
Our laser reference cavity consists of two concentric cylinders. The inner
cylinder is the Fabry-Perot spacer, having a 30mm outer diameter and a 100mm
length, and held in the center by a 5mm thick disk. The outer cylinder,
providing support and a stable environment for the inner cylinder, has a 70mm
OD, 50mm ID, and a 100mm length. A cutaway view of the system is shown at
right.
Both the cylinders and the connecting disk are constructed of a single piece of
ULE. ULE endcaps are attached to the outer cylinder with epoxy.
This design evolved from experiments by
Nottcutt et al. [Opt. Lett. 30, 1815 (2005)]
demonstrating high performance by vertical
cavities supported by center-mounted disks. Its
purpose is to provide high thermal, thermomechanical, and acceleration isolation. Such a
design may allow for possible use in space
without the need for clamping during launch.
Recent Work on Cavity-Stabilized Lasers
5 * 10-15/rtHz frequency stability at 10-3 Hz:
A.D. Lindlow et al., Science 319, 1805 (2008)
A cutaway view of the laser reference cavity. The inner
cylinder forms the Fabry-Perot spacer.
T. Rosenband et al., Science 319, 1808 (2008)
Cavity Mounting
The laser reference cavity is mounted on a 55mm high isolation platform, designed to provide
thermal isolation. This platform consists of three stages of 10mm dia. ULE legs, separated by
10mm thick Zerodur disks. The lowermost legs rest on the floor of the vacuum chamber.
Understanding of thermal noise limitations:
K. Numata et al., Phys. Rev. Lett. 93, 250602 (2004)
Two aluminum thermal shields rest on the disks, shielding the reference cavity from thermal
radiation. A diagram of the mount is shown at left.
M. Nottcutt et al., Phys. Rev. A 73, 031804 (2006)
Importance of symmetrical resonance cavity support:
A diagram of the reference cavity and
isolation platform. The lowermost legs
rest on the vacuum chamber floor.
M. Nottcutt et al., Optics Lett. 30, 1815 (2005)
Voltage Reference Stability
Dual-Cylinder Laser Reference Cavity Performance Estimates
One limit to the stability of electrically applied forces to the LISA proof masses is the stability of
onboard voltage references in the relevant frequency range. Commercial voltage references do
not have well characterized noise at low frequencies, and thus we have investigated the
performance of a variety of commercial references, to somewhat extend the results obtained
earlier by Heinzel et al. in Hannover.
Noise spectra were obtained by running a pair of voltage references through a differential
amplifier, and recording this output for analysis in software. Moderate temperature control
(~10mK/rtHz) was applied to the references and amplifiers.
Many voltage references were tested, and from these the four best performers were selected for
more detailed study. These were the AD587LN, LT1021, MAX6162, and the LT1236.
Of the external noise sources, the differential amplifier was found to have under a 7%
contribution to the overall noise. The measurement circuit and power supply were found to have
similarly negligible noise. Temperature sensitivity was also determined to be negligible in this
case, with a correlation to
the voltage reference of
under 10-6 V/K.
Below is a table containing
the results of this analysis,
giving the stability at
0.1mHz of the four
references. A typical noise
spectrum is also presented.
The best performer was the
AD587LN, with a relative
noise of (2.1 ± 0.6) *
10-6/rtHz at 0.1mHz
[Ellery Ames, Honors
Thesis, Univ. of Colorado
(2008)].
These results are consistent
with those obtained by
Heinzel et al. [CP873,
Laser Interferometer Space
Antenna – 6th Int. LISA
Symps. (2006), p. 291].
In addition to thermal protection, the staged system minimizes stress felt by the cavity due to
thermally induced changes in the dimensions of any metal parts, such as the vacuum chamber
and thermal shields.
Thermal Isolation
Model calculations show that the changes in the Fabry-Perot cavity temperature will be less than 10-7 of those of the
surrounding vacuum chamber at 10-3 Hz, which may be regarded as negligible.
Thermo-mechanical Isolation
Finite element analysis by M. Nickerson indicates that the lengthwise strain dL/L of
the cavity will less than 2 * 10-8 that of any thermally induced changes in the
stainless steel vacuum chamber dimensions. To the right is a slice of the cavity
system showing the relative displacement for a radial expansion of the lowermost
legs of the isolation system, as would be
expected for thermal expansion of the vacuum
chamber.
Vibration Isolation
Finite element analyses by T. Rosenband
(NIST) and M. Nickerson indicate that the
lengthwise strain dL/L of the cavity will be
under 2 * 10-12 for a 1 m/s2 acceleration
along the cavity axis. The figure to the left
shows the relative displacement of the laser
cavity due to a vertical acceleration, with the
mount held fixed.
Typical noise spectrum of the AD587LN, LT1021, MAX6126, and LT1236 voltage
references. The vertical axis is fractional noise level per rtHz.
Voltage Reference Output Voltage [V] Data Runs Noise Level [ppm/rtHz]
AD587LN
10
20
2.1 ± 0.6
LT1021
5
5
7.5 ± 1.1
MAX6162
5
4
5.3 ± 1.5
LT1236
10
4
6.6 ± 2.2
Relative noise levels of several commercial voltage references at 0.1 mHz.
A slice of the laser cavity and upper stage mount,
showing the relative displacement for a 1g vertical
acceleration.
06/13/2008
Rosenband’s analysis indicated that the
minimum length strain occurred if the
supporting disk was displaced by a small
amount. Thus in the design, it was offset
from the center by 190 microns. Nickerson’s
analysis agrees to within manufacturing
tolerances. As shown to the left with a
support offset of 0.21mm, the maximum
displacement under a 1g acceleration was 2 *
10-8 m, with a cavity strain of 1.4 * 10-11.
A slice of the laser cavity and upper stage mount,
showing the relative displacement for a thermomechanical radial expansion of the mount base.