Mirror Technologies for Giant Telescopes · Jim Burge Professor College of Optical Sciences and ......

Mirror Technologies

for

Giant Telescopes

Jim BurgeProfessor

College of Optical Sciences and

Steward Observatory

[email protected]

J. H. Burge University of Arizona 2

Today’s talk

• Introduce giant astronomical telescopes

planning to be built in the next decade

• Summarize challenges of making the mirrors for

these telescopes

• Present technologies and capabilities at

University of Arizona that enable manufacture of

the mirrors


The Giant Magellan Telescope

Site in Chile

2018 first light

Consortium with UA


Consortium

Telescopes

Large Binocular 2 x 8.4m Telescope

Mt. Graham AZHobby-Eberly 10m Telescope

McDonald Observatory, TX

MMT 6.5m

Mt. Hopkins AZ

Magellan twin 6.5m Telescopes

Las Campanas, Chile


GMT Design36 meters high, 25.3 meters across

25-m Primary mirror (f/0.7)

3.2-m segmented secondary mirrorcorrects for PM position errors

deformable mirror for adaptive optics

Alt-Az structure

~1000 tons moving mass

21.3 m azimuth disk

21 m elevation C-ring

Steel + CFRP secondary support

Instruments mount below primary at

the Gregorian focus


GMT Optical

DesignAplanatic Gregorian design

Primary mirrorf/0.7 near-paraboloid

Secondary mirrorellipsoid

segmented like primary

conjugate to 160 above M1

Internal f/0.7 focus

f/8 system focal ratio

Field of view: ~24 arc-min

BFD = 5.5 meters


Aperture – diffraction limit~350 m2 Collecting area

25 meters across

Point spread function

14 mas FWHM at 1.65 µm

(same FWHM as from

24.5-m filled aperture)(same collecting area as 21.9-m filled aperture)


GMT is Optimized for adaptive optics3.2-meter secondary mirror

- Segmentation matches primary, 1.1-m segments correct tip/tilt/piston

- Lightweight Zerodur segments with fast steering capability

- Deformable segments with 672 actuators/segment, 1 kHz controlLaser projector


Key optical challenges for GMT

• Primary mirror segments

– Off-axis asphere with ~15 mm aspheric departure

– Segments must match in radius to work together correctly

• Secondary mirror segments

– 1-m off-axis aspheres

– Two sets of secondaries, rigid + adaptive

– Thin shell, supported by actuators


MMT, the early pathfinderMultiple Mirror Telescope

(1979-1998)

6.5 meter MMT Conversion

(2000- )

6 telescopes, each 1.8 meters

System pointed as Alt/Az

Collecting area of 4.5-m

telescope

Resolution of 6-m telescope

6.5-m conventional telescope

First of the big AZ mirrors

Double collecting area

Increase FOV

Adaptive Optics development


LBT – Today’s challengeLarge Binocular Telescope

Pair of 8.4-m telescopes

Coherently combined

On Mt. Graham, AZ


13 m 21 m

LBT to GMT, a natural step

LBT GMT

Baseline 22.8 m 25.4 m

Mass 600 tons 1000 tons

Enclosure 30 x 50m 56 x 65 m

LBT and GMT at same scale


Primary mirror manufacturing

Cast lightweight mirror blanks

High stiffness, low weight

Thermal responsive

Low expansion borosilicate glass

Spin cast for near net shape

GMT mirrors follow procedures

already used for 9 large mirrors,

now in optical systems


Casting: 1. Mold assembly

Machine and install 1681

ceramic fiber boxes in

silicon carbide tub.

Tops of boxes follow

shape of aspheric

surface; no two are

identical.


Casting: 2. Glass loading

Close furnace, prepare

to melt and spin.

Inspect, weigh, and

load 18 t of E6

borosilicate glass.


Casting: 3. Glass melting

Glass heated to 1160C melts and flows into mold

UV cameras mounted in furnace lid monitor the casting.


Casting: 4. SpinningHeat to 1160˚C, spin at 4.9 rpm for 4 hours

Cool rapidly to 900˚C then slowly for 3

months,


Finished blank on the oven hearth


Lift and washout

Tilted into vertical planeRear surface with floor tiles attached

Segment with lifting fixture

Removal of floor tiles


Preparation of rear surface

Rear surface machined and polished flat.

Load spreaders attached to provide interface to 165 support actuators


Generate optical surface• Using high speed diamond wheels, grind blank to final thickness

• Use NC control to created the nominal aspheric surface

• Tricky for GMT because it was cast as on-axis parabola

Large Optics Generator at Steward Mirror Lab 14.5 mm shape variation generated

into GMT off axis segments


Stressed lap grinding and polishing

Loose-abrasive grinding and polishing uses a deformable

(stressed) lap which actively changes shape so that it fits

the aspheric mirror surface.

8.4-m LBT2 primary

finished to 17 nm rms


Optical testing for LBT

Axisymmetric

Test optics at ~20 meters

Light from optical test is only 200

mm diameter near the test optics –

allows direct measurement of test

system

Test

optics

~1.4 mm aspheric departure

Test wavefront

defined to match

aspheric shape

of mirror

20 m

Null lens assembly, used

with interferometer

23 cm CGH to calibrate

LBT null corrector


Ship to telescope - LBT


Integration into telescope cell• 8.4-meter LBT mirror is supported with 165 actuators

• Maintains figure while changing elevation angle

• Controls mirror position to ~10µm

• Blows temperature controlled air into each cell in mirror


Reflective coating


Active opticsMirror is supported with 6 “hard point actuators and 165 force actuators.

Shape is controlled based on measurements from wavefront sensors

• Calculate actuator influence functions: deflection of segment’s optical

surface caused by 1 N force at one actuator.

• Calculate natural bending modes: combinations of influence functions that

have most impact with least force (normalized to 1 N rms force).

nm

surf

ace

nm

surf

ace

Actuator influence functions

Lowest bending modes

(Buddy Martin)


“These mirrors work!” Wendy Freeman

NGC2359 ; Thor's Helmet

64 Mpixel imager on

Magellan

(Steeghs, Jan'04 )


“These mirrors work!” Wendy Freeman

LBT first light image

October 2005


Optical testing of GMT segmentsHeritage (LBT) GMT

Axisymmetric

Test optics at ~20 meters

Light from optical test is only 200

mm diameter near the test optics –

allows direct measurement of test

system

No Axisymmetry

Light path defined by GMT is much

larger(~3.5 meters across at the top of our tower)

Test

optics

~1.4 mm aspheric departure ~14.5 mm aspheric departure

Test wavefront defined

to match aspheric

shape

of mirror

20 m

roof


Surface measurement using interferometer and

reflective null corrector with CGH

GMT segment

0.75 m sphere

interferometer

computer-generated

hologram

GMT segment

0.75 m sphere

interferometer

computer-generated

hologram

GMT segment

Interferometer

CGH

3.75-m M1 fold sphere

tilted 14.2°

25

meters

75-cm mirror

Axis

of

par

ent

elli

psi

od

M1

center of curvature

Sam

Interferometer

at M1 center

of curvature


Test tower at Steward

Observatory Mirror LabOriginal tower

New tower

New tower

28 meters tall, 80 tons of steel

floated on 400 ton concrete pad

accommodates other UA projects (LBT, LSST)

lowest resonance of 4.8 Hz with 9 ton 3.75-m fold sphere + cell


CGH test of small optics system

• CGH inserted into light coming from Sam for alignment test

• Reflection back through system is used to verify wavefront

• CGH mounted on invar plate with other references for M1 alignment


Active alignment relies on laser tracker

• Reference hologram is aligned to

Sam. Then it is used to represent

Sam.

• Laser tracker used to co-align Sam,

fold sphere, GMT

Laser trackerSphere

Mounted

Retroreflector

Measures to determine

position to ~25 µm


3.75 m fold sphere• UA produced mirror, mounted at the top of the tower

• Shape is actively controlled based on surface measurements from the center of curvature

Polished, measured at the

Mirror Lab

Cast in the Mirror Lab

spinning oven Coated at Kitt Peak


GMT interferometry• Test is operational!

• Mirror is polished out, but surface figure is not yet within

dynamic range of interferometer

(M. Tuell)


Measurement of center segment

50 mm CGH

compensates

20µm aspheric

departure

Cone defined by

light from outer

edge of mirror

Vibration

insensitive

interferometer

Cone defined by

light from edge of

central hole

Tilt the fold sphere to nadir


Scanning pentaprism test

Scanning pentaprism measures slope errors

by producing collimated beams parallel to

parent axis. Displacement of focused spot is

measured with camera in focal plane.

Pentaprism rail lies in plane perpendicular to parent axis.

Scanning pentaprism test as implemented

for GMT off-axis segments. Pentaprism

rail is suspended from tower.A

xis

of p

are

nt p

ara

bolo

id

parent paraboloid

Off-axis mirror

Collimated laser

CCD camera at

focus of paraboloid

Fixed reference

pentaprism

with beamsplitter

Scanning

pentaprism

Image at CCD

Axis

of p

are

nt p

ara

bolo

id

parent paraboloid

Off-axis mirror

Collimated laser

CCD camera at

focus of paraboloid

Fixed reference

pentaprism

with beamsplitter

Scanning

pentaprism

Image at CCD

Uses the “magic” of the pentaprism


Pentaprism test of 1.7 m off-axis NST mirror

• 1/5 scale GMT pentaprism test (f/0.7 off axis paraboloid)

• The pentaprism test only samples lowest order aberrations

• The PP results corroborate the results from interferometry!

pentaprism measurementinterferometric test

nm

surface

102 nm rms 113 nm rms(Peng Su)


sphere-mounted retro-

reflector for laser tracker

Retroreflector for interferometer

and position sensing detector (PSD)

assemblies in 4 places at edge of

mirror

laser tracker & distance-measuring

interferometers (DMI)

PSD 10% BS

DMI retroreflector

DMI laser and remote

receivers

laser tracker

DMIs

Laser Tracker Plus

T. Zobrist dissertation


GMT secondary mirrors

• 1.1 m diameter segments

• Deformable mirrors use technology from MMT, LBT– Make aspheric surface on a thick glass substrate

– Thin the glass from behind to ~ 2 mm

• Rigid mirror blanks are conventional machined class

• Polishing of optical surfaces requires efficiency (14 mirrors plus spares)

• Testing with interferometer and CGH (easy!)


Adaptive Secondary Mirror


Swingarm computer controlled

polishing machines at Optical Sciences

UA Swingarm computer

controlled polisher

Mounting OAP onto CCP

CCP in operation

Two of these machines are

currently operation #3 is being built


Swingarm computer controlled polishers

80 cm mirror for GMT test43 terms removed,Fizeau map rms=0.0083898um

50 100 150 200

20

40

60

80

100

120

140

160

180

200

-0.05

-0.04

-0.03

-0.02

-0.01

0

0.01

0.02

0.03

0.04

1.4-m off axis parabola8 nm rms after removing low order terms

Use small, engineered polishing laps.

Polishing runs based on computer optimization

Run on the high spots!

Final figuring in 6 weeks with 150 machine hours


On-board metrology with

Swingarm Optical CMM• Scans surface with optical displacement probe

• Continuous arc scans create profiles

• Profiles stitched together to give surface maps

• Works for convex or concave surface

probe and alignment stages

convex asphere

center of curvature

optical axis

axis of rotation

probe trajectory

rotary stage

arm


Surface maps from SOC data for 1.4-m off axis asphere

Pattern of 64 scans

Interpolated data

75 nm rmsGrid map rms=0.07471um

100 200 300 400

50

100

150

200

250

300

350

400

-0.1

0

0.1

0.2

0.3

Grid map power removed rms=0.07245um

100 200 300 400

50

100

150

200

250

300

350

400

-0.1

0

0.1

0.2

0.3

0.4

Grid map power astigmatism and coma removed rms=0.031665um

100 200 300 400

50

100

150

200

250

300

350

400

-0.1

-0.05

0

0.05

0.1

0.15

4 terms removed, rms=0.07857um

50 100 150 200

20

40

60

80

100

120

140

160

180

200 -0.2

-0.1

0

0.1

0.2

0.3

0.4

896 term reconstruction

78 nm rms43 terms removed, rms=0.0065192um

50 100 150 200

20

40

60

80

100

120

140

160

180

200

-0.04

-0.02

0

0.02

0.04

0.06

Low order terms removed

6 nm rms

(P. Su)

-50 -40 -30 -20 -10 0 10 20-0.02

-0.015

-0.01

-0.005

0

0.005

0.01

0.015

0.02rms=6nm

Encoder angle in degrees

Repeatability ~ 6 nm rms/scan

Surface

measurement

in µm

Repeatable errors of 34 nm rms are calibrated out


The Giant Magellan Telescope

• Uses proven mirror technology

• Challenge of optical testing is being solved

• Challenge of rapid fabrication for

secondary mirrors is solved

• The casting of the second primary mirror

segment is currently being planned


Large Synoptic Survey Telescope

8.4-m aperture

Uses three-mirror design to attain

3.5° field of view

15-sec exposures to survey the sky

30TB of data per night!

Primary-tertiary combination

made at UA


LSST PM-TM

• Spin cast blank at UA

• PM curvature from spinning,

TM shape must be generated

• To be polished at UA with

stress lap

• Separate optical tests being

developed for TM and PM


The Thirty Meter Telescope Led by Universities of California

(formerly known as the

California Extremely Large Telescope)

Plan for Mauna Kea site

2018 first light


TMT optical challenges

3.5-m flat

tertiary mirror

3.1-m convex

secondary mirror

30 meter f/1 primary

mirror, made of 492

hexagonal segments


Fabrication of large flat mirrors

Flat

surface

under test

Reference

mirror

(spherical)

TMT tertiary: 3.5 x 2.5 meter flat mirror

Conventional test of large flats uses

Ritchey-Common test, not practical

Measurement difficulties are solved by technologies developed and

proven at UA


1-m vibration insensitive

interferometer + subaperture stitching

• Commercial instantaneous Fizeau

interferometer (uses 2 circularly

polarized beams)

• Modified to use external 1-m

reference

• Demonstrated on 1.6-m flat mirror

• World’s best large flat!

Large flat mirror

1 m reference flat

H1000 Fizeau

interferometer

Fold flat

1 m illumination OAP

Rotary table

53

Flat Surface by Stitching Method - power/astigmatism removed

100 200 300 400 500 600 700 800

100

200

300

400

500

600

700

800 0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.1

7 nm rms

[R. Spowl]

(J. Yellowhair, P. Su)


UA Fizeau test for TMT tertiary

Layout of subapertures 3 nm rms typical

measurement noise

Mode 1 2.9 nm rms

Mode 2 2.4 nm rms

Mode 3 1.4 nm rms

Mode 4 1.1 nm rms

Mode 5 1.4 nm rms

Mode 6 0.8 nm rms

Mode 7 0.8 nm rms

Mode 8 0.8 nm rms

Mode 9 0.5 nm rms

Mode 10 0.8 nm rms

RSS for all modes: 4.6 nm rms

Residual from fitting all modes 3 nm rms

4.6 nm rms for all modes

3 nm rms residual

Monte Carlo analysis to evaluate

coupling of alignment, noise to

surface reconstruction

5.5 nm rms measurement accuracy

with proven UA hardware!

(C. Zhao)


Scanning pentaprism test for flat mirrors

• Demonstrated performance is 2

nrad rms

• Power measurement for 1.6-m

flat was 11 nm rms

Shutters

Autocollimator system

Fixed prism

(reference )

Scanning

prism

Feedback

mirror

Mechanical

supports

Coupling

wedge

ELCOMAT

(Measuring AC)

UDT

(Alignment AC)

(J. Yellowhair)


Fabrication of large convex aspheres

TMT secondary: 3.1-m convex

hyperboloid

LSST secondary: 3.4-m diameter

convex hyperboloid

Conventional test of such mirrors uses

the Hindle test, not practical at these

sizes

LSST SM blank at Corning

TMT secondary mirror in telescope


Measurement with Swingarm Optical CMM

• Original system built at Optical Sciences was made with 3.8-m

capacity.

• This was designed to be integrated with 4-m polishing machine

• Performance is expected to be < 20 nm rms at this size


Fizeau interferometry for large secondary mirrors

Reference surface on test plate has concave matching asphere for off axis portion of

the secondary mirror

We have also developed concept for using a spherical reference, corrected with CGH


4-m polishing capability at Optical Sciences Primary mirror for Discovery Channel Telescope

16 nm rms

0 50 100 150 200

20

40

60

80

100

120

140

160

180

200

-60

-40

-20

0

20

40

60

80

100

We can make the TMT and LSST flat and convex mirrors!


TMT primary mirror

30 –m f/1 near paraboloid

492 segments

1.44-m segments, 45 mm thick

Each is supported by 27-point whiffle tree, warped with 21 actuators

Segment tip/tilt/piston adjusted with 3 position actuators, based on edge sensors


European large telescopes

Last decade’s new telescopes

VLT Very Large Telescope

4 telescopes with 8-m apertures

Giant Magellan

Telescope

California Extremely

Large Telescope

What’s next?


The Overwhelmingly Large Telescope

• 100 m aperture, made from >3000 spherical segments

• 25-m flat secondary mirror

• 4-mirror corrector

• Operates outside, uses roll away enclosure


The European Extremely Large Telescope

42-m aperture

European Southern Observatories

2018 First light


E-ELT telescope

6 meter convex aspheric

secondary mirror

42-meter f/1

segmented

primary mirror

(plus other mirrors)


E-ELT primary mirror• 42-m f/1 near paraboloid

• 984 segments

1.43-m segments, 50 mm thick

Each is supported by 18 or 27-point whiffle tree

Segment tip/tilt/piston adjusted with 3 position actuators, based on edge sensors


The challenges of mirror segments

• Measurement of the surface

– Off-axis aspheres, with different prescription (curvature changes from center to edge)

– To work together, the radius of curvature must match

• Efficient fabrication of hundreds of parts

• Support and operational control of thousands of degrees of freedom


CGH Fizeau test for

primary mirror segments

Collimator

Measurement

CGH

Projection

Lens

CCD camera

Aperture

Reference wavefront

Zero order from CGH

Reflects from reference sphere

Test wavefront

First order from CGH

Reflects from segment

Return : common path

Both wavefronts coincide

The difference between these gives

the shape error in the segment

Common

CGH

Convex

reference

sphere

m = 1

m = 0

m = 0 from segment

m = 1 from sphere

Blocked by aperture:

Reference and test wavefronts come to

focus and pass through aperture

All other orders and reflections are blocked

Mirror

segment

Aspherical

surface

Focusing lens

objective

diffuser

Test Plate


CGH Fizeau test

• Common path – low noise

• Radius matching is easy, all segments

compared with the same reference

• Detailed engineering analysis for TMT,

E-ELT predicts 14 nm rms overall

accuracy, 5 nm rms after some low order

correction


OAP

Collimator

Measurement

CGH

Lens

CCD camera

Common

CGH

Aspheric

surface

m = 1

m = 0

m = 0 from OAP

m = 1 from sphere

Blocked by aperture:

Reference

sphere

Spherical

surface

objective

diffuser

Full scale demonstration of CGH Fizeau test

Test convex, off axis aspheres by

measuring though the substrate,


Logistics for Fizeau test at COS

1.8-m Reference

sphere, face

down

Beam launch &

imaging system

0.9-m fold flat

OAP and

reference

sphere

Roll OAP

into tower

Hoise frame

with OAP

Fasten OAP

into position


Up, up, and away..


UA achieved very low noise measurements with

CGH Fizeau system

Excellent fringe visibility

Excellent spatial resolution

Overall accuracy of < 4 nm rms

< 1 nm rms noise per measurement (average of 50 maps)

Verified effects of straie in glass to cause < 1 nm rms

Largest sources of error: ghost reflection in CGH, coating irregularity on fold flat


Comparison of Fizeau, SOC for off-axis aspheres

• The Fizeau test was budgeted as < 3.3 nm rms uncertainty, after

correction for low order terms.

• SOC measurements of the OAPs are consistent with this.

Fizeau SOC

Raw data

After

removing low

order terms

Difference109 nm rms 117 nm rms

14 nm rms 16 nm rms 7 nm rms

Largest errors in Fizeau came from

coating defect on large fold flat 1 nm rms

ghost fringes 1 nm rms

Astigmatism and coma from

alignment were not needed to be

controlled accurately


What next?Combine several elements:

• Fizeau interferometry with spherical reference, corrected by CGH

• Vibration insensitive interferometry using polarization

Design for such a test for LSST was presented last summer. Prototype work is underway.

(M. Dubin)


Conclusion

• University of Arizona technology is

enabling the Giant Magellan Telescope

and the Large Synoptic Survey Telescope

• We are prepared to support TMT, E-ELT if

those projects move forward.

Mirror Technologies for Giant Telescopes · Jim Burge Professor College of Optical Sciences and ......

Documents

Transcript of Mirror Technologies for Giant Telescopes · Jim Burge Professor College of Optical Sciences and ......