1.1

Zerodur main properties and advantages

Integrated CTE at cryo-temperature (integrated values from 293K):

α= 0.11 E-6 °K^-1 at 120K

α= 0.16 E-6 °K^-1 at 50K

Moreover, Zerodur offers an excellent homogeneity of thermal expansion.

This very stable and homogeneous integrated CTE property leads to a high stability of the optical shape from the manufacturing/polishing at ambient to the operational at cryo-temperature.

Mechanical characteristics (E/ ρ) are high enough to allow the design of rather stiff and lightweight mirror structures.

Density: ρ= 2.53 g/cm3

Young’s modulus:

E= 90 GPa at 293K

E= 85 GPa at 50K

Strength: fragile material but very well known and predictable with Weibull distribution. Attention must be paid on the application of the appropriate surface treatment (etching, polishing).

Conservative bending strength value for long term application is 10 MPa, value used for the global mirror.

For the allowable stress at the fixation interface we have considered a D35 surface treatment (especially for the prediction of the bonded fixation strength). This grade qualifies a median treatment in terms of strength performances. Associated weibull parameters are:

The weibull distribution law gives the following allowables for the interface area:

1.2

Zerodur manufacturing improvement

IBF (Ion Beam Figuring) allows to decrease drastically the quilting effect of a classical polishing.

This quilting effect is one of the main limitation for the definition of the cells diameter, then for the mirror lightweighting.

The IBF correction has been simulated on different cell shape, tria or hexa, and for several size (inner diameter) and front face thickness.

Globally IBF correction reduces from more than a ratio of 10 the RMS value link to the quilting effect of the classical polishing.

This allows to increase the cells diameter for more than 30% (for a same front face thickness).

Fig. 1.

Quilting on tria cell – 15 g/cm²

Fig. 2.

Residual quilting on tria cell – after IBF

We also push at their limits the ‘’conventional” glass manufacturing capabilities by making:

Fig. 3.

Mock up definition

Fig. 4.

Manufactured mock up for the lightening

1.3

Mock up for lightening feasibility check

A mock up dedicated to the validation of the glass lightweighting has been developped by Alcatel/Sagem- REOSC and manufactured by Sagem-REOSC.

This activity has validated the feasibility of manufacturing the following geometry:

1.4

Test sample and results for the Zerodur behaviour at operational cryo-temperature

A cylindrical flat mirror made of Zerodur have been optically measured at ambient and at cryo-temperature in order to check the good stability of the material: image quality and hysteresis.

Mirror characteristics:

Zerodur class 0, thickness 15 mm, dia 250 mm

Fig. 5.

Mirror mounted into the cooled cavity

Mirror supporting is reduced to the minimum, based on ISO mirror cryo-testing heritage, to avoid constraint induced during cool down and provide a good thermal decoupling (3 permaglass cylinders).

The optical measurement is performed with a Fizeau interferometer. The optical reference surface of the interferometer is accommodated inside the vacuum chamber to have a measurement free from air turbulence and vacuum shroud porthole deformation. This optical reference is mounted on the optical bench and is protected by MLI. This configuration is inherited from ISO experience. The measurement accuracy is close to 5 nm.

The Wave-front error measurements are performed on sub-areas of 100mm diameter. The seven positions selected are sketched on to cover a maximal part of the surface. The five measurements A, B, C, D and E give the mirror material behaviour while F and G monitor the mirror supporting.

Fig. 6.

Measurement areas location on the front face

From wave front error map computations between ambient and cryogenic temperature (105K), it appears that the values are close to the measurement accuracy: between 5 and 8 nm. The residual effects can be assigned more realistically at mechanical configuration (slight residual constraint at I/F) or optical variation on phase reference plate under vacuum.

Moreover, delta WFE computation between room and cryogenic temperature evidences no hysteresis effect.

Zerodur class zero confirms its good behaviour whatever the temperature is (here down to 105K).

1.5

Design and optimization for 1.5m² mirror

Based on the manufacturing performances here above, we designed two rectangular mirrors (primary mirror type for large telescope) which have the necessary level of mechanical and mecha-optical performences to cope with visible wavelengths mission.

1^st concept: Evolutive thickness with tria cells:

Fig. 7.

Evolutive thickness – rear view.

Fig. 8.

Evolutive thickness – cross section.

2^nd concept: Circular Frame

Fig. 9.

Circular frame – rear view.

Fig. 10.

Circular frame – cross section.

Both concepts have the same mass but:

The mass and strength for both are:

Here after, an exemple of the performances for the ‘’Evolutive thickness’’ concept, are:

2.1

Design and justification

The fixation definition at the interface with Zerodur is very critical since it must be compatible in strength with:

A trade-off on different concept of fixation has shown that the best compromise in term of performences is the following one:

An INVAR ring is bonded on a Zerodur boss. The bonding thickness is calibrated by a groove in the metallic ring.

The adhesive was selected for its particular good mechanical characteristics in the working temperature range (100 K to 300 K): it is the best compromise for strength and stiffness vs. low temperature.

Fig. 11.

Fixation – cross section (sketch)

2.2

Main characteristics and capability

Mechanical strength of the bonded interface:

Glue:

Zerodur:

WFE impact at Operational temperature:

Thermoelastic effect of the 3 bonded fixations on the mirror optical face at Operational after cool down 293K -> 100K (by analysis with the mirror FEM):

2.2

Mock up and test results

Fig. 12.

Fixation mock up

Thermal cycling:

More than 10 thermal cycles from ambient temperature down to 100 K have been performed successfully. The temperature rate was 2° / min which leads to a cycling duration of more than 3.5 days, when including stabilisation time at 100 K and at 313 K.

Static loading, performed after the thermal cycling:

Fig. 13.

Fixation mock up during traction test

Out-of-plane direction:

the maximal effort of 6600 N was reached and maintained a few seconds.

That demonstrates a margin of + 56% wrt the highest axial design load.

In-plane direction:

the test was performed in several steps up to 17000 N. At this level (twice the tangential design load) no damage has been detected.