Transcript No Slide Title - The University of Arizona College of Optical Sciences

Optical Design of
Giant Telescopes for Space
Jim Burge, Erin Sabatke
Optical Sciences Center
Roger Angel, Neville Woolf
Steward Observatory
University of Arizona
The need for large telescopes
• Push back the frontier for astrophysics
– We want to study what we can barely detect
– We know that increased technology will detect new things
• Imaging planets around other stars
– Requires blocking or nulling the star light
• Laser projectors for interstellar vehicles
– Use light momentum to push the sail for interstellar travel
• Earth observations from geosynchronous
Telescopes in space
Hubble Space Telescope
1990
Hubble’s telescope
Mt. Wilson 100-in
1917
Natural evolution to large telescopes
• Make the primary larger
– keep it in the shade
– make the f/number faster to
limit length
– effective optical surface using
smaller segments that can be
launched and deployed
– maintain weight while
increasing area
– Requires primary mirror with
density ~15 kg/m2
Next Generation Space
Telescope 2009
8-m aperture
Optical design issues for NGST
• Three mirror anastigmat, 10 arc min FOV
• Fine steering mirror at a pupil
– image stabilization limited by field rotation, distortion
• Fast primary is highly aspheric and difficult to
fabricate and test
8-m primary
mirror
Tertiary
mirror
Secondary
mirror
Science
Instruments
Fine Steering
Mirror
Multiple Aperture Systems
Collecting
telescope
Combining
mirrors
Combining
telescope
Coherent image
Coherent image
Interferometer that combines two telescopes
Collecting
mirror
Interferometer that directly combines light from two
reflectors.
• Increase baseline and collecting area by
combining multiple apertures
• 100-m array
• Use nulling (destructive
interference) to cancel star
light
• Detect planets and obtain
low resolution spectra,
looking for familiar
atmospheric constituents
• Special Purpose
Very small field of view
optimized for exoplanets
• Solar orbit, benign thermal
and gravity environment
Intensity
Terrestrial Planet Finder
7
8
9
10 11 12 13 14 15 16
Wavelength (µm)
TPF as free flying array of 3.5-m telescopes
What about giant telescopes
• Size is limited by mass from mirror technology
• NGST mirror technology could get to 5 kg/m2
50 cm diameter mirror under construction
1 mm thick glass
7 gram actuators
1 kg/m2 composite support
• For economical launch with existing technology,
need mass << 1 kg/m2
Ultralight mirrors for space optics
• Lower mass mirrors require thinner substrates
(<< 1 mm)
• The difficulty is support and control
• Curved optics intrinsically require shape control
• Flat optics can be made by simply stretching a
thin membrane
Error in reflective surface = half of thickness variation
Control for flat membrane mirrors
• Start with thin, reflective membrane of uniform thickness
• Hold it in tension from a plane at the perimeter
Membrane with
reflective coating
Tension control
Rigid frame
Shape control with
actuators at nodes
• Define the perimeter with multiple points, each one under
active control.
• Reduces shape control to 1 dimension - perimeter
Membrane mirror technology
• Numerous developments underway at University
of Arizona
(Stamper et al. In Imaging Technology and Telescopes,
presented Sunday).
What good are flats?
• Collect light using diffraction
– from Rod Hyde, Livermore
– limited bandwidth, contrast
• Or use an array of flats to approximate a
paraboloidal reflector
– like solar collectors
– downstream optics compensate for non-curvature
Primary made from flat segments
Optical design issues for
primary made from flats
• On axis - easy
– make different segments come to focus at the same place
with the same path length
• For field of view - tricky. For each subaperture
system, must also
– meet sine condition (constant mapping of entrance pupil to
exit pupil)
– match image scale and distortion
– match field curvatures
• The general solution is to make the effective focal
ratio of the primary as long as possible
Telescope with free flying elements
1 km
S u n s h ie ld
S u n s h ie ld
1 0 m c o m b in in g
te le s c o p e
5 0 m p rim a ry
Faster telescopes for “conventional” rigid systems
Slower designs for telescope with free flying elements
Transition to Membranes
The length comes at the price of system agility
Telescope diameter
surface density
F/1 systems
2.6m
8.0m
150Kg/m2 16Kg/m2
Mass
800Kg
Moment of Inertia
1 unit
Rotation period for same
thruster expended
1
Rotation period for same
reaction wheel use
1
800Kg
10
14m
5Kg/m2
F/20 systems
25m
100m
1.6Kg/m2
800Kg
30
800Kg
36,000
3
5
10
30
0.1Kg/m2
800Kg
600,000
190
800
36,000
600,000
Membrane telescopes are for long observations of ultra-faint objects
only. General Purpose telescopes should be restricted to rigid
mirrors.
Truss for large primary mirror
(Tom Connors, Steward Observatory)
Optical design and analysis
• Simulations using Optima (from Lockheed Martin)
• Test case with several flat mirrors
Analysis of flat mirror telescope
On axis
OPD
PSF
1 arc minute
Pupil mapping and phase errors
Mapping distortion of entrance pupil (h) to exit pupil (h’)
couples with wavefront tilt to cause phase errors
Distorted
Ideal mapping
Entrance
Entrance
mapping
pupil
pupil
h
Exit pupil with
non-linear
mapping
Exit pupil
No distortion
h’
From Object
a‘
a
To image
’
Wavefront is
preserved
Wavefront
Phase from
WF ‘tilt’
Phase error
in wavefront
Definition of “sine condition”
Entrance
pupil
Exit pupil
To Image
From Object
Sin(U)
Sin(U’)
Spherical entrance pupil, coordinates of sin(U)
Spherical exit pupil, coordinates of sin(U’)
Sine condition requires linear mapping of
sin(U) -> sin(U’)
Sine condition violation
• Geometric pupil distortion causes violation of sine
condition, which varies with field
• This causes images to dephase for < 1 arcmin FOV
Quantifying sine condition violation
1.6
Sine condition violation:
% deviation from on-axis value of sine condition
1.4
1.2
1
0.8
0.6
0.4
0.2
0
0
0.5
1
1.5
2
Field angle (arcminutes)
2.5
3
3.5
Optical design summary
The design with flat segments works!
However, the field of view is limited by the apertures dephasing
with field, from the sine condition violation
Preliminary results indicate that a 100-m telescope of this type
could be made with
6 meter segments
2 km length
7 m secondary (spherical relay)
10 meter corrector (60 cm elements)
0.3 µm rms wavefront errors for 20 arcsec FOV
(40 nm rms for 6 arc seconds)
Curving the primary
The system works much better if the primary can be curved
Electrostatics can be used to do this
A two-mirror 100-m telescope can achieve 5 arc minute FOV
at f/20 with
2 km length
10 m concave secondary
0.4 µm rms wavefront error
The 6 m primary segments have 4 mm sag
This has 40 nm rms wavefront error at 1.6 arc minute FOV
Stretched Membrane with Electrostatic Curvature
• Primary mirrors with membrane reflectors can be made from
slightly curved segments (using electrostatics)
The moon imaged at Steward Observatory with the first
telescope to use a primary mirror of Stretched Membrane
with Electrostatic Curvature (SMEC). The silicon nitride
membrane was 0.7 um thick and curved to a 3 m focal length
by a field of 2 MV/m
What about a strip mirror
Simulated performance 100m x 2m
HST
Slot,
1 exposure
NGST
Slot
18 exposures
(Keith Hege, Steward Observatory)
Truss modeled for strip mirror
(Tom Connors, Steward Observatory)
Summary
There is no evolutionary path from today’s systems to
giant telescopes in space.
Launch constraints require low mass, leading to optics
made from membranes.
Orbital mechanics allows the use of free flying
elements and sunshields.
Primary mirrors, made from arrays of flat mirrors can
provide corrected images.
With added weight and complexity, the membranes can
be moderately curved, gaining an order of magnitude in
field of view.