INSTITUTE OF PHYSICS PUBLISHING CLASSICAL AND QUANTUM GRAVITY
Class. Quantum Grav. 22 (2005) S149–S154 doi:10.1088/0264-9381/22/10/003
Successful testing of the LISA Technology Package
(LTP) interferometer engineering model
G Heinzel
1
, C Braxmaier
2
, M Caldwell
3
, K Danzmann
1,4
, F Draaisma
5
,
AGarc
´
ıa
1
, J Hough
6
, O Jennrich
7
, U Johann
2
, C Killow
6
, K Middleton
3
,
MtePlate
7
, D Robertson
6
,AR
¨
udiger
1
, R Schilling
1
, F Steier
1
, V Wand
1
and H Ward
6
1
Max-Planck-Institut f
¨
ur Gravitationsphysik (Albert-Einstein-Institut), Callinstrasse 38,
D-30167 Hannover, Germany
2
EADS Astrium GmbH, 88039 Friedrichshafen, Germany
3
Rutherford Appleton Laboratories, Chilton, UK
4
Universit
¨
at Hannover, Institut f
¨
ur Atom- und Molek
¨
ulphysik, Callinstr. 38, D-30167 Hannover,
Germany
5
The Netherlands Organisation for Applied Scientific Research (TNO), Technisch Physische
Dienst (TPD), Delft, The Netherlands
6
Department of Physics and Astronomy, University of Glasgow, Glasgow, UK
7
ESTEC, Noordwijk, The Netherlands
E-mail: gerhard.heinzel@aei.mpg.de
Received 8 November 2004, in final form 10 January 2005
Published 21 April 2005
Online at stacks.iop.org/CQG/22/S149
Abstract
The LISA Technology Package (LTP), to be launched by ESA in 2008,
is a technology demonstration mission in preparation for the LISA space-
borne gravitational wave detector. A central part of the LTP is the optical
metrology package (heterodyne interferometer with phasemeter) that measures
the distance between two test masses with a noise level of 10 pm Hz
−1/2
between 3 mHz and 30 mHz and also the test mass alignment with a noise level
of <10 nrad Hz
−1/2
. An engineering model of the interferometer has been built
and environmentally tested. Extensive functionality and performance tests
were conducted. This paper reports on the successful test results.
PACS numbers: 06.30.Bp, 06.30.Gv, 07.60.Ly, 07.87.+v, 42.30.Rx, 95.55.Ym
(Some figures in this article are in colour only in the electronic version)
1. Introduction
The LISA Technology Package (LTP) aboard the LISA Pathfinder (LPF) mission is an
important milestone towards the LISA mission. Its purpose is to demonstrate the operation of
free-floating test masses as mirrors of an interferometer and the performance of the drag-free
0264-9381/05/100149+06$30.00 © 2005 IOP Publishing Ltd Printed in the UK S149
S150 G Heinzel et al
operation. The sensitivity goal is relaxed by a factor of 10 with respect to the LISA
specifications and thus is 3 × 10
−14
ms
−2
Hz
−1/2
at 1 mHz for LPF.
The interferometer of the LTP continuously monitors the test masses in all operating
modes by measuring
• the distance between the two test masses (called x
1
− x
2
),
• the position of one test mass with respect to the optical bench (called x
1
),
• the differential alignment of the two test masses (with two sets of measurements: dc and
differential wavefront sensing [2]),
• the alignment of one test mass with respect to the optical bench.
These signals not only serve as diagnostic monitor, but also as feedback signals for test
mass stabilization in some operational modes.
The principle of operation and general design of the interferometer and phasemeter are
described in [1, 2]. An engineering model (EM) of the interferometer has been constructed
using the same optical design, materials and bonding techniques that are planned for the flight
model (FM). The EM was subjected to comprehensive environmental tests, which it survived
without any measurable degradation or change. The functionality, performance and noise
level of the interferometric measurement system were characterized. This paper describes the
construction techniques of the optical bench (OB), the environmental tests and the measured
performance which is very satisfactory overall and within the specifications at almost all
frequencies.
2. Construction
The optical bench was manufactured under EADS Astrium lead at Rutherford Appleton
Laboratories (RAL) according to the optical design given in [1, 2] with support from Glasgow
University and AEI Hannover. The baseplate is made from Zerodur
TM
, and the optical
components (of typical size 7 mm × 15 mm × 20 mm) were attached by hydroxy-catalysis
bonding [3, 4] to form a very rugged quasi-monolithic structure. Most of the components
were aligned using a template manufactured from brass. The template was manufactured
to approximately 10 µm accuracy, and the resulting absolute positioning accuracy of the
components was estimated to be approximately 0.5 mm. The final recombination beamsplitters
were aligned with light beams present and photodiodes installed by observing and optimizing
the photodiode signals (interference contrast and mutual beam alignment) in real time. The
beamsplitters were pre-aligned using a mechanical manipulator and fine aligned during the
few minutes before bonding sets in after application of the bonding agent.
A handling accident in a late assembly stage caused four components that were already
bonded to break. Those bonds older than a few weeks broke in the glass, not at the bond. This
involuntarily demonstrated the strength of the quasi-monolithic bond. Furthermore, it forced
us to develop a repair strategy which consisted in bonding small interface plates (formed like
a bridge) over the damaged part of the baseplate and then bonding a new component onto the
bridge plate (see figure 1, right).
In the next assembly stage, sideplates and stiffening rods made from Zerodur
TM
(see
figure 1, left) were attached by EADS Astrium Immenstaad using the ‘insert’ technology to
reliably connect Zerodur
TM
pieces to each other.
3. Environmental testing
The optical bench EM was subjected to thermo-vacuum and vibrational testing at the facilities
of TNO/TPD Delft with personnel from TNO/TPD, AEI Hannover and EADS Astrium
Successful testing of the LTP interferometer engineering model S151
Figure 1. Picture of the optical bench (OB) engineering model (EM) and detail of the bridge plates
used to repair broken bonds.
Immenstaad during the first half of 2004 [5]. Dummy masses having the same mass and
moments of inertia as the real test mass enclosures were fixed to the Zerodur
TM
structure.
Thermal cycling included several cycles between 0
◦
C and 40
◦
C, and the vibrational tests
included sine-wave and random excitations in three axes reaching up to 25 g at the struts.
A comparison of the alignment data of each photodiode recorded before and after the
environmental tests yielded the following upper limits:
• No beam position has moved more than 5 µm on any photodiode.
• No beam injector has changed in alignment by more than 20 µrad.
These changes, if at all real, are of no significance to the functionality and performance
of the optical bench. They also imply that none of the bonded components have moved.
4. Functional tests
A comprehensive set of functional and performance tests was carried out while the optical
bench EM was mounted in a large vacuum chamber at TNO/TPD. The tests were carried out
with personnel and equipment from AEI Hannover, in particular the laser, modulation bench
and phasemeter that were used in the prototype development. Functionally, the interferometer
behaved completely according to the predictions, in particular:
• The test mass displacement fluctuations x
1
and x
1
− x
2
can be measured with constant
high sensitivity at any arbitrary displacement of the test masses along the sensitive x-axes
within several 100 µm.
• The differential wavefront sensing (DWS) alignment measurements yield very sensitive
alignment information for both test masses if each test mass is within ±500 µrad of its
nominal position.
• For misalignments of a test mass larger than ±500 µrad, the interferometer contrast is
lost. The (less sensitive) dc alignment signals, however, still yield useful information
about the direction of the misalignment.
S152 G Heinzel et al
5.7
5.8
5.9
6
6.1
6.2
6.3
6.4
6.5
6.6
0 1 2 3 4 5 6 7 8 9
0
20
40
60
80
100
120
140
phase x
1
- x
2
[rad]
optical pathlength fluctuation
between mirrors [nm]
time [10
3
sec]
-0.01
-0.005
0
0.005
0.01
0 1 2 3 4 5 6 7 8 9
-1.5
-1
-0.5
0
0.5
1
1.5
phase x
1
- x
2
[rad]
optical pathlength fluctuation
between mirrors [nm]
time [10
3
sec]
54nm/ h drift subtracted
Figure 2. Time series of the x
1
− x
2
measurement. In the right-hand graph, a best-fit linear
function was subtracted.
• All these measurements produce valid results as soon as properly modulated light arrives
at the optical bench and the phasemeter is operational. No acquisition or ‘locking’
procedure is necessary. The interferometer recovers immediately and automatically from
momentary dropouts of, e.g., the laser power or blocking of a beam.
5. Performance tests
The purpose of the performance tests was to demonstrate the noise behaviour of the
interferometer. Out of several stretches of data taken overnight, one stretch of 9000 s length
was selected that showed little external disturbance. The contrast of all interferometers was
stable between 80 and 90%.
Figure 2 shows a typical time series of the x
1
− x
2
measurement. It is dominated by a
linear thermal drift. If this drift is subtracted (right picture), it is still dominated by a slow
drifting motion. These drifts are probably due to thermal effects in the metal mirror mounts.
The interferometer noise is barely visible as the finite thickness of the line in the right-hand
graph.
Figure 3 shows a typical linear spectral density of the x
1
− x
2
measurement. The dark
solid curve shows the measurement described here. For comparison, several other curves are
also shown (from top to bottom at 2 × 10
−2
Hz):
• A curve measured at TNO in a comparable configuration, but without the fibre optical
pathlength (OPD) stabilization (see [6]). In all three measurements below, OPD
stabilization was used.
• The LPF mission goal. Note that in contrast to earlier inconsistent interpretations, the
y-axis is now consistently referring to optical pathlength, and the LPF mission goal of
85 pm Hz
−1/2
test mass displacement is hence plotted as 170 pm Hz
−1
optical pathlength
(at 30 mHz).
• The earlier ‘best curve’ with quadrant diodes [6], from the Glasgow prototype
interferometer (also developed during the pre-investigations), measured in a collaboration
between staff from Glasgow and AEI at Glasgow University using both Glasgow and AEI
equipment.
• The LPF interferometer goal. This is unchanged by the re-interpretation and remains at
9pmHz
−1/2
optical pathlength (at 30 mHz).
Successful testing of the LTP interferometer engineering model S153
10
-5
10
-4
10
-3
10
-2
0.1
10
-4
10
-3
10
-2
0.1 1 10
1
10
100
1000
10000
phase [rad/√Hz]
optical pathlength [pm/√Hz]
frequency [Hz]
LPF mission goal
interferometer goal
LPF EM x
1
- x
2
, no OPD-stab, QPD
Glasgow/AEI, QPD
LPF EM x
1
- x
2
, QPD
Glasgow/AEI, SED
Figure 3. Linear spectral density of the x
1
− x
2
measurement (see the text). ‘QPD’ and ‘SED’
refer to quadrant- and single-element diodes, respectively.
1
10
100
1000
10
-3
10
-2
0.1 1 10
Differential misalignment (nrad/√Hz)
referred to test mass
Frequency (Hz)
reference ifo (PDR) y
reference ifo (PDR) x
x
1
- x
2
ifo (PD12) y
x
1
- x
2
ifo (PD12) x
10 nrad/sqrt(Hz)
Figure 4. Linear spectral density of the DWS alignment measurements (see the text).
• The result discussed here, which represents a new ‘best curve’ as measured with quadrant
diodes.
• The ‘best curve’, from the Glasgow prototype interferometer (also developed during the
pre-investigations), measured in a collaboration between staff from Glasgow and AEI
at Glasgow University using both Glasgow and AEI equipment [6]. This curve was
measured with single-element diodes, which consistently show a better noise behaviour
than quadrant diodes, the reason for which is under investigation.
Figure 4 shows a linear spectral density of the DWS alignment measurements.
The predicted noise level of 10 nrad Hz
−1/2
(referred to the test mass misalignment) is
reached in the x
1
− x
2
measurement, where the static misalignment could be minimized. The
higher noise of the reference interferometer (photodiode PDR) alignment can be traced to the
fact that this diode shows a large static DWS misalignment (due to the imperfect beam injector
alignment).
