Why we need it now

1
Spectrometer simulation

• Why we need it now
• What should be simulated
• How to do it
• Work plan
• Conclusion

A.Bonissent A.Ealet C.Macaire E.Prieto A.Tilquin
Note in http//www.astrsp-mrs/snap/spectro/spectro
sim.ps
2
Past

• Previous stage, only laboratory tests and
  simulation of slicer alone have been performed.
• This is sufficient to ensure that an instrument
  can be built with adequate performance.
• Now to study the real performances on the full
  instrument, we need a complete simulation

3
Why

• Needed in the present phase for
• Optimizing the design balance cost and
  simplicity (reliability) for best possible
  physics
• compute realistic efficiency
• evaluate tolerance
• evaluate calibration procedure
• produce realistic data to develop and test data
  processing algorithms
• At term, it will be used in detailed MC studies
  for physic analysis

4
Optimisation process
Specifications
Optical design
No
OK ?
yes
5
Développement plan

• At the end of phase A, we need a Final design of
  the instrument with estimated (and justified)
  performances
• Simulation and data reduction software for
  evaluation should be ready well before
• simulation spring 2003
• data reduction prototype spring 2004

6

• Full SNAP simulation

instrument
Cosmo models
analysis
Lightcurve spectra
Physic parameters
Detector pixel data
propagation
physic
Data cube
Data reduction
Lightcurve spectra
SN simulation
7

• Spectrometer simulation

x,y,l
optical sim
i,j,Qij
telescope
readout
Parameterization constants
x,y,l psf1
Pixel parameterization
fit
slicer
Data cube i,j,adc
x,y,l psf2
i,j,Qij
x,y,l psf3
pixellisation
spectrograph
8
Method
Compute Psf and transmission at each x,y,l
Telescope xt,yt,
Slicer xs,ys
Slit xf,yf
prism xl,yl
Detector xd,yd
pupil xp,yp
TF
TF
TF
TF
TF
peif
p
peif
peif
p
psf
TF of amplitude from object plane to pupil plane
then to image Apply geometry and phase (zernike)
on pupil Apply geometry on image Compute
intensity to evaluate efficiency Interpolate
position x,y at each step (need
parametrisation) Output is position on the
detector for each point and wavelength with an
associated PSF Very long and CPU intensive
9

• Psf shape and size depends on x,y,l
• (small amount of) energy is lost by diffraction
• Geometry affects
• performance

10
0.9 mm
Slice 0
Slice 2
1.7 mm
11
Zernike polynomial of slicer for l 1.7 mm
Zernike Polynomia from Zeemax are used to
introduce aberrations Depend of l,x,y They need
to be extrapolate on each point of the image plan
Use Neural Network technique to do extrapolation
12
Efficiency study
Gobal efficiency Telescopeslicerspectrograph
13
Simulation checking spectral resolution
R l/dlpixel
14
Used for

• DESIGN OPTIMISATION
• Test optic
• Play with optic to study tolerance
• Efficiency/nb of pixel
• Visible/IR efficiency vs spectral
  resolution/detector
• optimise spatial resolution gt detector noise
  optimisation
• Reduce Nb of mirrors better transmission but
  may need more space, more complex optics
• TEST DATA
• Slit effect Position of SN in slice gt
  translation of spectrum
• SN may cover several slices need to add
  translated spectra
• Optical distorsions
• Pixellization
• Dithering
• Detector and electronics efficiency, noise,
  cosmics ...

15
Distorsions on the detector
20 pixel/slice
U spatial dimension
Detector pixels do not coincide with l Cte or x
Cte
V spectral dimension
16
Current Status

• Full simulation of slicer unit
  OK
• Full simulation of telescope and spectrometer
  OK
• Interpolate for intermediate points
• using Neural Network technique.
  OK
• 
  
• library of PSF for a grid of x,y,l
  under work
• From library of PSF geometry (x,y,l -gt detector
  indices) to be done
• Pixellisation integrate over pixels
  
• Add dark current, readout noise etc...
  
• Include galaxy
  
• Dithering (spatial, spectral)
  

17
Conclusion

• Detailed simulation of the spectrometer is needed
  in this phase to quantify performances
• CPU intensive not appropriate for physics
  simulation
• 
  
• Parametric simulation under development, based on
  the library of PSF should be appropriate for a
  full SNAP simulation (not for SNAPfast).

18
Spectrograph Performances
telescope Relay optics Slicer Optic straylight diffra. Slicer Optic straylight diffra. Slicer Optic straylight diffra. Spectro Mirrors prism dichroic Spectro Mirrors prism dichroic Spectro Mirrors prism dichroic Detector Vis / NIR Detector Vis / NIR
elements 4 1 3 1 1 2 1 1 1 1
Efficiency/ 0.98 0.98 0.98 0.99 0.95 0.98 0.81 0.95 0.9 0.6 (0.8)
cumulative 0.92 0.90 0.85 0.84 0.80 0.77 0.62 0.59 0.53 0.35 (0.47)
Gain on mirror transmission, loose on
diffraction/prism (complete simulation) Globally
equivalent
19
Design issues

• Spectral resolution optimization visible /IR
• ( R(IR 1-1.4 mm) lt 100 but dont need to join
  the 2 detectors )
• Polarization specification needed impact on
  spectrograph
• Design
• Spatial resolution 0.15. Issue vs the
  radiation rate
• Wavelength range 1.7 mm short for the Si line ,
  1.8 mm better
• but detector l cut issue , issue on temperature