A Φ 3.
5 M SiC telescope for HERSCHEL Mission
Emmanuel SEINa*, Yves TOULEMONTa, Frédéric SAFAa, Michel DURANa,
Pierre DENYb, Daniel DE CHAMBUREc**, Thomas PASSVOGELc, Göran PILBRATTc
a
ASTRIUM(Toulouse, France), bBOOSTEC(Tarbes, France),
c
ESTEC-ESA (Noordwijk,Netherlands)
Since ten years ASTRIUM has developed sintered Silicon Carbide (SiC) technology for space applications. Its
unique thermo-mechanical properties, associated with its polishing capability, make SiC an ideal material for
building ultra-stable lightweight space based telescopes or mirrors. SiC is a cost effective alternative to
Beryllium and the ultra-lighweighted ULE. In Complement to the material manufacturing process, ASTRIUM
has developed several assembly techniques (bolting, brazing, bonding) for manufacturing large and complex SiC
assemblies. This technology is now perfectly mature and mastered. SiC is baselined for most of the telescopes
that are developed by ASTRIUM. SiC has been identified as the most suitable material for manufacturing very
large cryogenic telescopes. In this paper we present the development of the Φ 3.5 m telescope for Herschel
Mission. Herschel main goal is to study how the first stars and galaxies were formed and evolved. The Herschel
Space telescope, using silicon carbide technology will be the largest space imagery telescope ever launched. The
Herschel telescope will weight 300 kg rather than the 1.5 tons required with standard technology. The Herschel
telescope is to be delivered in 2005 for a launch planned for 2007
Keywords: Large Telescope, Space Application, Lightweighted reflector, SiC
1 INTRODUCTION
Over the past ten years ASTRIUM have developed the SiC technology for Space Applications1, 2 in collaboration
with a ceramic company BOOSTEC Industries (Tarbes, France) The joint venture SiCSPACE gathers the
expertise of both companies for supplying optics and telescopes for both space and ground applications. The
technology is now fully mastered and ESA selected ASTRIUM in 2001 for manufacturing HERSCHEL large
telescope. HERSCHEL satellite is part of HERSCHEL/PLANCK program of the European Space Agency (ESA)
devoted to far infrared astronomy. Both satellites, HERSCHEL and PLANCK will be launched in 2007 by
Ariane 5 in a dual launch and will operate at an orbit about Lagrange L2 point.
HERSCHEL telescope is of Cassegrain type with a parabolic primary reflector of diameter 3.5 meters. This
unprecedented telescope will be mounted on the top of HERSCHEL helium cryostat vessel that contains three
science instruments. The telescope operates at cryogenic temperature (~ 80 K) in the far infrared wavelength
range (80 µm to 600 µm) and it will be the largest telescope in orbit in 2007. The overall telescope mass is
below 300 kg and this puts in evidence the exceptional lightweighting capability of SiC technology: for
comparison, Hubble Space Telescope primary mirror mass is about 900 kg for a diameter 2.4 m
During previous study phases (1995–2000), ESA conducted extensive trade-offs on the telescope material,
supported by the development of breadboards of CFRP composite and lightweight aluminium mirrors. It was
clearly established that ASTRIUM/BOOSTEC Silicon Carbide technology was the only viable solution for
meeting mass and performance requirements.
After a brief review of the material properties, we present the development status of this telescope and the results
of the HERSCHEL mirror demonstrator that was manufactured and tested two years ago.
2 PROPERTIES OF THE SINTERED SIC
Silicon Carbide is now a mature technology for space applications that presents a high growth potential. The
material can be obtained by several processes, which can significantly affect either, its physical properties or its
cost .We shall only consider here the sintered Silicon Carbide manufactured by BOOSTEC industries (located at
Tarbes) according to a well-defined and cost efficient process.
2.1 Material Basic Properties
Some key figures measured by ASTRIUM are summarized in Table N°1.
Material Characteristics Room 110 K
Temp.
Density ρ (Kg/m²) 3160 3160
Young E (Gpa) 420 120
Modulus
Bending σr (Mpa) 374 405
strength
Toughness K1C 2.75 2.83
(Mpa.m1/2)
Weibull m 14 14
Modulus
CTE α (ppm/K) 2 0.65
Thermal λ 190 180
conductivity (W/m/K)
Specific Cp (J/kg/K) 700 135
Heat
Table N°1: Some basic properties of SiC (ASTRIUM measurements)
Two classical figures of merit have been added :
Thermal distortion ratio λ/α: The higher is the better. It physically reflects that thermal distortions are not only
proportional to the CTE, but also to the thermal gradients, which, under given thermal environment, drop down
when the thermal conductivity (λ) increases.
Specific stiffness E/ρ : The higher is this value, the better is the material lightweighting capability, for equal
mechanical behaviour (e.g. equal first resonance frequency). However, a comparison only based on specific
stiffness is rather theoretical, since it does not include manufacturing limitations such as the aspect ratio (rib
thickness/ height). Actually, such limitations are not very constraining for SiC: an aspect ratio about 20-25 can
easily be reached, and values as high as 80 with a rib thickness of 1 mm have been achieved.
2.2 Optical Properties
BOOSTEC Silicon Carbide is polishable (polishing convergence as good as for glass) and the surface roughness
can be made as low as 1 nm. The material can therefore be used for optics in the visible range. It can be metal-
coated by using the same process than for silica glass and with similar performances. Several types of reflecting
coating have been space qualified on the sintered SiC (Al, Ag, Au) without any problem.
2.3 Joining Techniques For Manufacturing Large Sic Telescope
Available facilities allow manufacturing monolithic silicon carbide pieces of dimensions up to 1 m x 1.6 m. This
covers practically most of the needs. Manufacturing larger monolithic pieces would require a significant
industrial investment. Therefore, the cost effective and safe approach for the realisation of large pieces, such as
HERSCHEL primary reflector, is to assemble together smaller pieces, which are well within manufacturing
capabilities.
In addition such joining techniques are absolutely necessary for telescope manufacturing where we have to
integrate SiC mirror onto SiC mechanical structure. Several techniques can be envisaged and have been
investigated:
i) Mechanical assembly of two SiC pieces, by bolting with shear pins ( this type of assembly has been space
qualified)
ii) Epoxy bonding
iii) High temperature brazing
Brazing technique consists of adding a material between two SiC pieces. ASTRIUM technique is a high
temperature brazing that provides several remarkable properties:
• Its CTE can be matched to that of Silicon Carbide, say within 0.1 ppm/K or so,
• The brazing joint can be very thin : few µm to few tens of µm (Fig. 6). But thick joints of thickness as high as
300 µm have been achieved. For thin joints, the brazing strength is comparable or better than for SiC.
Numerous tests performed at liquid nitrogen temperature showed that brazing strength is practically not
affected at low temperatures.
20µm
Microscope inspection of a
brazing joint.
Figure N°2: Φ630 mm brazed mirror
( This mirror is composed of two identical semi-spherical lightweighted mirrors brazed on a diameter)
• The brazing is non-reactive, i.e. SiC is not attacked. Therefore, de-brazing is possible without any
damage of the SiC parts.
3 THE Φ 3.5 M SIC HERSCHEL TELESCOPE
The main characteristics of this telescope are given in table N°16 .The telescope is a Cassegrainian telescope
which elements ( reflector and structure) will be manufactured in SiC.
Parameters
Diameter 3.5 m
Mass of the telescope <300 kg
Operating temperature 70 – 90 K
Operating wavelengths 80 µm to 600 µm
Telescope total WFE < 6 µm rms
Transmission > 0.95
Longitudinal frequency > 60 Hz
Lateral frequency > 45 Hz
Table N° 3. HERSCHEL Telescope Characteristics
Figure N ° 4 : Herschel Telescope
3.1 Optical Design
The HERSCHEL telescope is a pure Cassegrain design composed of a parabolic primary and a hyperbolic
secondary. The primary makes a perfect image of an on axis object point at infinity and the secondary re-images
this point in a perfect manner in the focal plane. Standard Cassegrains can be improved by higher order
aspherisation of the mirrors as in the Ritchey-Chretien design. In our case however, the difference between the
two designs is negligible and purely theoretical so that the simple Cassegrain design has been selected. The main
optical characteristic is given in the following figure (Figure N° 5)
Herschel Telescope Cassegrain
optical parameters
Primary reflector
Radius of curvature 3500 mm
Conic constant -1
Distance to M2 1587.998 mm
Secondary reflector
Radius of curvature 345.2 mm
Conic constant -1.279
Diameter 308.12 mm
Image surface
Radius of curvature -165 mm
Conic constant -1
Diameter 246 mm
Distance to M1 -1050 mm
Figure N° 5: HERSCHEL telescope optical design.
The coating is made of a first adhesive layer on NiCr (less than 10 nm thickness) and a 150 to 300 nm layer of
aluminium. The coating will be performed at the CALAR ALTO Observatory facilities in Spain where a 4 m
deposition vacuum chamber is available.
The telescope design has to be optimised in order to avoid narcissus effect and to restrict the impact of spurious
source and self-emission of the structure below the background induced by the reflecting mirrors. Therefore a
thermal foil or multi-layer blankets of low emissivity envelope all non-optical parts (un-coated). This naturally
ensures a thermal decoupling of the telescope from the rest of the satellite. Heaters are also attached to the
telescope parts for a contamination release of optical surfaces in-orbit. The contamination release phase is
achieved at the beginning of life prior to the cryostat opening by heating up the telescope parts to ~ 45 °C. After
contamination release phase, heaters are switched off and the thermal control of the telescope is purely passive.
3.2 Primary reflector design
The most complex element is obviously the Φ 3.5 m primary reflector that is composed of 12 SiC segments
brazed together. The segment is open-back lightweighted with triangular cells of inner diameter ~ 120mm. The
rib heights and thickness are optimised for minimising mass while meeting frequency requirement. As an
illustrative example, the rib height varies from 110 mm in the centre of the segment to 10 mm at the edge. The
thickness of the optical surface is 2.5 mm. For minimising manufacturing cost and schedule, the design has been
constrained so as to include only two types of segments (figure 11) with minor differences between the two
types: Three segments include interfaces for mounting the Secondary Mirror hexapod and the telescope isostatic
mounts, while the nine other are “without interfaces”.
Figure N°6– HERSCHEL Primary Reflector design – segments with or without I/F
FigureN°7– HERSCHEL Primary Reflector
The main parameters of the primary reflector are detailed in the following table:
Diameter 3.5 m
Mass 210kg
Areal density 21.8 kg/m²
First Frequency
Lateral 50 Hz
Axial 100Hz
Table N°8 – HERSCHEL Primary Reflector Main Characteristic
The primary reflector blank will be ground by BOOSTEC then polished by OPTEON (Turku, Finland). The
lapping and polishing are efficiently performed with a technique specific to OPTEON where the polishing tool
consists of several sub-tools disposed in a linear array across the mirror radius. The sub-tools are moved in a
short radial stroke while the mirror is rotating during lapping/polishing. The sub-tool sucking produces the
surface pressure by generating a partial vacuum to the reflector face-sheet. There are no unbalanced forces
acting perpendicularly to the surface and the quilting problem is minimised. The partial vacuum is varied under
computer control to regulate the surface pressure and the local removal efficiency. It is necessary for the
vacuum induced surfacing pressure that the three interface holes of the face-sheet are blocked with silicon
carbide pieces. During lapping phase, the measurements of the mirror shape are made with a swing arm
profilometer. Fine measurements of the polished mirror are achieved with a pentaprism test device.
OPTEON polishing technique using partial vacuum between the tool and the optical surface was demonstrated
on the 1.35 m reflector in 1999 (see section 4)
Figure N°9. Polishing machine showing the polishing tools and their carriage.
3.3 Secondary reflector design
The secondary reflector is a 308 mm diameter hyperbolic convex mirror.
Its dimensions are compatible with a simple central fixation concept integrated to the mirror. This kind of
interface has been developed and validated by ASTRIUM for various other instruments and features several
advantages: simple mounting by bolting on the hexapod SiC barrel, no isostatic mounts, accurate and
reproducible shimming during telescope alignment, cost effective design.
The optical face thickness 2.4 mm, leading to a weight below 2.1kg.
The mechanical sizing of this piece is not critical with regard to frequency as well as to strength.
For avoiding narcissus effect by reflection at the central area of the reflector, a special shaped SiC piece is added
at the vertex. Herschel wavelength range is such that this cap piece is not required to be polished and may be
only ground (roughness < 0.5 µm). Therefore the piece is simply fixed to the mirror after polishing.
Figure N°10 – M2 design
3.4 Mechanical structure design
-The Hexapod
- The hexapod ensures the relative position of the primary and secondary mirrors, which has to be stable within
less than 10 µm at operational temperature without any refocusing mechanism. The selection of hexapod
configuration was driven by mass, stiffness and obscuration ratio constraints. The reflector material choice
drives that of the hexapod and the use of a unique material for the whole telescope practically suppresses thermal
stresses and distortions. The defocus error is then almost insensitive to average temperature variations of the
telescope. Furthermore, the defocus error induced by residual thermal gradients at operational temperature is
negligible, because silicon carbide material combines a low thermal expansion (CTE) with a high thermal
conductivity.
The hexapod concept is based on:
• 6 SiC Legs .The legs are “U” shaped and slightly rotated by for straylight minimisation
• 1 SiC barrel glued to the 6 legs
• 3 flexible Invar fittings glued onto 2 legs each and bolted onto M1 upper inserts via an adjustment shim to
minimize integration distortions
The legs and the barrel are covered by kapton foil for thermal and optical purposes. SiC/invar gluing was
validated down to 50 K through a dedicated qualification plan.
Figure N°11– SiC HERSCHEL Hexapod with its interface fittings
-Isostatic Mount and Hexapod Fittings
The telescope isostatic mount function is to enable the telescope integration on the satellite cryostat structure
with negligible optical distortions. The isostatic mounts consist of titanium bipods with a 49° optimal angle.
They will be fixed through specific INVAR inserts bolted on the M1 reflector interfaces. The bipod rod section
is designed so as to achieve a proper isostatic mounting while meeting frequency requirements.
The purpose of the Hexapod fittings is to isostatically link the hexapod onto the M1 reflector interfaces for
minimising the primary reflector optical distortion during integration. The Hexapod fittings are made of invar for
matching as close as possible the SiC CTE and minimizing the telescope defocus in cold related to SiC/invar
CTE mismatch.
Figure N°12: Telescope bipods and Hexapod Fittings
3.5 WFE test set-up
The most challenging verification process will be the optical WFE measurement near operational temperature
(110 K). The Herschel telescope will be tested in auto-collimation on a flat mirror in vacuum chamber with a
Shack-Hartmann wavefront sensor. The test will be performed by ASTRIUM in the Centre Spatial de Liege
(Belgium) test facility. The optical test is performed in autocollimation configuration and in cold environment
ideally fully representative of operational conditions. Since the whole telescope is practically made of a single
homogeneous material, residual elastic cool-down distortions are expected to be small with respect to the overall
WFE budget, with a smooth dependence with temperature. The telescope behaviour at 80 K can safely be
deduced from thermal cycling tests and sample tests. The auto collimation set-up is the major difficulty of the
optical test. ASTRIUM excluded for obvious economical reasons the manufacturing of a 3.5 m flat mirror,
which quality must be better than that of the telescope. It is proposed to replace the flat mirror by a set of
distributed liquid mirrors that cover the entire telescope pupil. This set-up is compatible with a Shack-Hartmann
measurement technique, which relies on ray slope measurements: actually the liquid mirror height can be
arbitrary, while the gravitation field ensures its surface quality. The liquid mirror sub-assembly is made up of 48
identical mirrors of 100 mm diameter distributed on a square grid. The grid period is 440 mm. Each mirror is
made up of a cell containing mercury sealed by a BK7 window and the test will be made in the visible range for
simplicity and improved accuracy.
.
Figure N° 13: HERCHEL telescope Vacuum test configuration
4 HERSCHEL PRIMARY REFLECTOR DEMONSTRATOR
During the preliminary design phase several breadboard and mock-up were developed in order to
demonstrate that a large cryogenic telescope could be manufactured in SiC.
4.1 Breadboard Φ1.35m
For validating the manufacturing process3 of the primary reflector we realised a Φ1.35m diameter
demonstrator. This demonstrator will be fully representative of the HERSCHEL manufacturing
technology. The breadboard experience was funded in the frame of ESTEC contract.
The design was made of 9 SiC segments brazed together in order to realize a half scale demonstrator of
the primary reflector. The reflector breadboard has sustained all the qualification tests planned for the
Herschel Telescope showing not only the compliance of the proposed technology with the environmental
requirements but also the mastery of the analysis and design process of Silicon Carbide pieces.
Figure N°14: Φ1.35 m HERSCHEL demonstrator (before brazing and after, brazing, machining and polishing)
The vibration test was performed at ESTEC including low level frequency research survey and static tests at
fixed frequency sustained up to 29g laterally and 27g axially. The resulting stress sustained by the Reflector is
estimated to 100Mpa in SiC and 6Mpa in Braze joints. This test has also enabled the measurement of the
amplification factor (Q=110 to 175 on mirror modes). Furthermore the measured natural frequencies were well
correlated with the model predictions (<5%). This gives good confidence for the future Herschel vibration test
predictions and proposed notching.
Figure N°15– HERSCHEL Breadboard during Vibration Test
The Breadboard has also sustained the acoustic Ariane5 proto-qualification spectrum in free-free
boundary conditions, in ESTEC chamber. Once again the measurements were in good accordance with
the predictions. This test has shown that, despite the high acceleration measured on the external contour (56g at 3
sigma), this environment is less critical for the Reflector (stress level: < 11 MPa).
The proposed technology (SiC+brazing) has been also tested at 110°K, close to the Herschel operating
temperature, without any damage on the SiC parts or brazed joints. In addition no distortion resulting from
brazing or from CTE variation between segments has been measured. However a trefoil distortion of WFE =
2.6µm due to local effects on the 3 Invar fittings has been detected and is also recovered with the mechanical
model. This has been avoided on the Herschel design by adapting the design of the fittings.
Figure N°16– WFE distortion Map between ambient and operational temperature
If the pure trefoil distortion is removed, the residual cool-down distortion is within the measurement noise of the
interferometer (0.3 µm rms). Neither effect due to the reflector segmentation nor to the brazing was detectable,
in accordance with the predictions and measurements on samples.
4.2 3.5m segment mock-up
During this phase we also design and manufacture a full size petal .The achieved design takes into account not
only the HERSCHEL telescope specifications but also all the manufacturing requirements and the 1.35 m
breadboard experience.
Figure N°17: HERSCHEL telescope full-scale petal. Size 1.5 m x 0.91m.
5 FUTURE TRENDS & CONCLUSION
The SiC technology developed by ASTRIUM and BOOSTEC is now fully mature. Space qualified telescope and
mirrors are currently proposed for a wide range of applications. The Φ 3.5 m HERSCHEL telescope is currently
under development .The critical design review has been successful .The primary reflector is expected to be
completed on June 2003. The full telescope will be delivered to our customer on march 2005 after acceptance
test campaign. During this program, all the preliminary validation activities have demonstrated that SiC material
is ideally suited the development for very large cryogenic telescope.
We are presently pushing up the technology for manufacturing ultra-lightweighted mirrors with areal density
down to 9 kg/m² and very large instrument as GAIA (where the complete payload will be build in Sic),
DARWIN telescopes. Further more, the sintered Sic manufacturing is really cost effective and it will a very
attractive solution for future ground based telescope as the 100 m diameter telescope OWL planned by ESO
AKNOWLEDGEMENTS
The development of the HERSCHEL telescope is part of the HERSCHEL-PLANK program that is funded by the
European Space Agency (ESA).
The authors thanks T. Korhonen (OPTEON Oy) and H. Vayssade (SAGEIS CSO) for their fruitful discussions
on SiC polishing process and telescope performance validation.
REFERENCES
1- F. Safa, F. Levallois, M. Bougoin, D. Castel: "Silicon Carbide Technology for Submillimetre Space Based
Telescopes", 48th International Astronautical Congress, October 97,Turin. (Italy)
2- E. Sein, F. Safa, D. Castel, P. Deny; “Silicon Carbide, A Sound Solution For Space Optics”, 52nd
International Astronautical Congress, 1-5 Oct 2001/Toulouse, (France)
3- F. Safa, M. Bougoin “Development And Test Of A 1.35 M Silicon Carbide Reflector”, ICSO 2000,3-6
December 2000,Toulouse (France)
Didier CASTEL,
[email protected], Pierre DENY,
[email protected]**Daniel De Chambure,
[email protected], Thomas Passvogel,
[email protected], Göran
Pilbratt,
[email protected]
