Instrumentation Concepts Ground-based Optical Telescopes

1
Instrumentation ConceptsGround-based Optical
Telescopes

• Keith Taylor
• (IAG/USP)
• Aug-Nov, 2008

Aug-Sep, 2008
IAG-USP (Keith Taylor)
2
Imaging Fourier Transform Spectrographs (IFTS)
FTS Michelson Interferometer IFTS Imaging
IFTS over solid angle, ?.

• Beam-splitter produces 2 arms
• Light recombined to form interference fringes on
  detector
• One arm is adjustable to give path length
  variations
• Detected intensity is determined by the path
  difference, ?x, between the 2 arms.

3
IFTS theory (simple version)
Given that frequency, ? 1/? (unit units of
c) Phase difference between two mirrors
2???x So recorded intensity, I, is given by
Now, if we vary x in the range ?? ? x/2 ? ?,
continuously then
These represent Fourier Transform pairs.
Spectrum B(?) is obtained from the cosine
transformation of the Interferogram I(x)
4
IFTS reality (simple version)

• At x 0 the IFTS operates simply as an imager
• White light fringes all wavelengths behave the
  same
• At all other x-values, a subset of wavelengths
  constructively/destructively interfere
• For a particular ?, the intensity varies
  sinusoidally according to the simple
  relationship

In reality, of course, x goes from 0 ? ?xmax
which limits the spectral resolving power to
eg if ?xmax 100mm and ? 500nm then R0 ?
1.105
5
IFTS in practice
Since we are talking here about an imaging FTS
then what is its imaging FoV? Circular symmetry
of the IFTS is identical to the FP and
hence 2?l.cos? m? And also R?
gtgt 2? limited only by the wavelength variation,
??, across a pixel However, in anaolgy to the
FP ? Phase-correction is required in order to
accommodate path difference variations over the
image surface.
6
Pros Cons of an IFTS

• Advantages
• Arbitary wavelength resolution to the R limit set
  by ?xmax
• A large 2D field of view
• A very clean sinc function, instrumental profile
• cf the FPs Airy Function
• A finesse N 2??/?? which can have values higher
  than 103
• Disadvantages
• Sequential scanning like the FP. However, the
  effective integration time of each interferogram
  image can be monitored through a separate
  complementary channel, if required
• Very accurate control of scanned phase delay is
  required
• Especially problematic in the optical
• At all times, the detector sees the full spectrum
  and hence each interferogram receives integrated
  noise from the source and the sky
• This compensates for the fact that all
  wavelengths are observed simultaneously which is
  why there is no SNR advantage over an FP
• Also sky lines produce even more noise, all the
  time.

7
Michelson Interfermeter(N 2 interference n
gtgt1)
8
Hybrid and Exotic Systems

• FP IFTS are classical 3D imaging spectrographs
• ie Sequential detection of images to create 3D
  datat cubes
• FP Wavelength scanning
• IFTS Phase delay scanning

There are, however, techniques which use a 2D
area detector to sample 2D spatial information
with spectral information, symultaneously. These
we refer to as Hybrid Systems

• Examples of this are Integral Field Units
  (IFUs). These can use either
• Lenslets
• Fibres
• Lenslets Fibres
• Mirror Slicers

9
Integral Field Spectroscopy

• Extended (diffuse) object with lots of spectra
• Use contiguous 2D array of fibres or mirror
  slicer to obtain a spectrum at each point in an
  image

10
Lenslet array (example)
LIMO (glass) Pitch 1mm Some manufacturers use
plastic lenses. Pitches down to 50?m
Used in SPIRAL (AAT) VIMOS (VLT) Eucalyptus (OPD)
11
Tiger (Courtes, Marseille)

• Technique reimages telescope focal plane onto a
  micro-lens array
• Feeds a classical, focal reducer, grism
  spectrograph
• Micro-lens array
• Dissects image into a 2D array of small regions
  in the focal surface
• Forms multiple images of the telescope pupil
  which are imaged through the grism spectrograph.
• This gives a spectrum for each small region of
  the image (or lenslet)
• Without the grism, each telescope pupil image
  would be recorded as a grid of points on the
  detector in the image plane
• The grism acts to disperse the light from each
  section of the image independently

So, why dont the spectra all overlap?
12
Tiger (in practice)
Enlarger
Detector
Camera
Lenslet array
Collimator
Grism
13
Avoiding overlap
?-direction

• The grism is angled (slightly) so that the
  spectra can be extended in the ?-direction
• Each pupil image is small enough so theres no
  overlap orthogonal to the dispersion direction

Represents a neat/clever optical trick
14
Tiger constraints

• The number and length of the Tiger spectra is
  constrained by a combination of
• detector format
• micro-lens format
• spectral resolution
• spectral range
• Nevertheless a very effective and practical
  solution can be obtained

Tiger (on CFHT) SAURON (on WHT) OSIRIS (on
Keck)
True 3D spectroscopy does NOT use time-domain
as the 3rd axis (like FP IFTS) very limited
FoV, as a result
15
Tiger Results (SAURON WHT)
16
Fibres in Astronomy
Optical fibre technology offers the astronomical
spectrograph designer vast opportunities. Astrono
mical Spectroscopy is the art of recording
spatial and spectral information simultaneously
onto a 2D area detector. In other words it
requires the re-formatting of information to suit
the detector and the astronomical goals. If we
could arbitrarily define the geometry of our
detectors (even to make them 3D!) then none of
the sophisticated optical design would be
necessary. This is where fibres come into their
own They are the perfect image
re-formatters, taking any shape of object and
re-forming it into a spectrograph slit.
17
Types of Fibre
Fiber operates as an optical wave-guide
Operates by total internal reflections
18
Focal Ratio Degradation(FRD)
Input f-ratio Output f-ratio (A? is preserved)
But not, unfortunately, in a fibre

• Note
• Input f-ratio is not preserved Fin (slower) gt
  Fout (faster)
• Central obstruction is filled in
• A?in lt A?out to compensate, R decreases or d
  increases

19
Numerical Aperture (NA)
For the fibre to operate as an optical waveguide,
total internal reflection (TIR) has to be
maintained throughout the passage of light along
the fibre. TIR then requires
Note tan?max 1/2Fin NA, the numerical
aperture NA 0.22 (Fin is slower than gt 2.3)
for normal fibres
20
Using Fibres to link Telescopesto Spectrographs.

• Advantages
• Spectrograph independent from telescope. Bench
  Spectrographs, no weight or volume restrictions.
  
• High spectral stability.
• Fibres are easy to use and install (once
  prepared!)
• Possibility to perform two-dimension
  spectroscopy with fibre bundles.
• Drawbacks
• Transmission losses.
• Focal Ratio Degradation.
• Circular aperture losses.
• Poor sky subtraction.
• Fixed slit aperture.
• Difficult to prepare if not proper tools are
  available.
• Fragile!

21
Fibre slicer(the simplest approach)
A simple fibre re-formatter from sky to
spectrograph slit
Re-formatted onto a long-slit of the spectrograph
2D array of fibres at the telescope focal plane
22
Fibre slicer attributes and examples

• Captures light over a full seeing disk (and more)
  without degrading the intrinsic resolving power
  of the spectrograph
• Facilitates spatially resolved spectroscopy
• No requirement to centre a point object on a
  slit
• No requirement to match slit width to the seeing
• Effectively detaches spectral and spatial
  information
• Facilitates spatially integrated spectroscopy
• Integral field spectroscopy (IFS)
• Supplies robust spectrophotometry
• Objects aligned along the slit

• Examples
• F.I.S. (on AAT - 1981) 100 fibres
• SILFID (on CFHT - 1988) 400 fibres
• HEXAFLEX (on WHT 1991) 61 fibres
• 2D-FIS (on WHT 1994) 125 fibres

23
Fibre Spectral Image
24
and now some numbers!
Clearly for a fibre diameter, ?fibre , each
individual fibre aperture (?fibre) on the sky is
given by
where Ffibre is the input focal ratio of the
fibre.

• Example
• Take ?fibre 0.5 D 8m and Ffibre 5
• ? ?fibre 80?m
• This integral field unit (IFU) fibre can be
  retro-fitted to existing long-slit spectrographs,
  however there are 3 problems
• Focal ratio degradation (FRD) which requires fast
  f-ratios
• Collimator speeds which are matched to normal
  Cassegrain f-ratios, which require slow f-ratios
• Spatial information is lost in the inter-fibre
  gaps

25
Coupling fibres with micro-lenses
If ? spatial sampling on sky (subtended by
micro-lens), then ?fibre ?.DT.Ffibre ?lens
?.DT.FTel
26
The SIFS IFcourtesyCL de Oliveira, (LNA)?
27
Down-side oflenslet/fibre coupling
Ffibre
The fibre
Fin (slower) gt Fspec (faster) because of FRD
But fibre receives light from micro-lens
significantly faster than Fin (where Ffibre
(faster) lt Fin (slower) - see red rays)
Take ?fibre dia. of fibre ?lens dia. of
micro-lens
Conclusion dont make ?lens too small Use
macro-lenses (!)
28
Mirror Image Slicers
Pioneered by MPI (3D) (Gensel)
Compact Efficient Slicer of choice but Cannot
be retrofitted to existing spectrographs
29
Slicer Promo (The End)