-SPID- - cfht.hawaii.edu fileCPNG:Comptage de Photons NouvelleG¶en¶eration 10 0 20 30 40 50 wavelength (nm) 400 500 600 700 800 900 quantum efficiency (%) ITT FS9910C (ultra blue

- SPID -

A high spectral resolution

monochromator

for diffraction limitedimaging

R. Foy, M. Tallon, I. Tallon-Bosc, E. Thiebaut, F.-C. Foy SPID / PUEO NUI Grenoble, 23/05/2003

Goal of speckle interferometryTo get diffraction limited images through

turbulent atmosphere

=⇒ Specs :

– sampling of speckles : magnification

– spectral coherence : filters or

spectrograph

– temporal coherence : fast detectorspecklegraph @

= monochromator for diffraction limited imaging


SPID specs :– Spatial sampling : 0.012 marcsec at

CFHT

– Field : 3”×2.25”

– Wavelength range : 400 - 750 nm (850

with filters)

– At once, 2 spectral channels within 15nm

– Spectral resolution : 60 . λ/δλ . 3000


SPID scheme

Telescope

Interface

Imaging / spectrograph

devices

Wavefront sensor Calibration


SPID at CFHT


CPNG : Comptage dePhotonsNouvelle Generation

• High quantum efficiency ρ ≈35%

500.550nnm, ρ &30% 440.720nnm

• High frame rate 262 Hz

• Fully digital

• Easy to use Peltier cooling, small and

light, remote controled, ...





10

0

20

30

40

50

wavelength (nm)

500 600 700 800 900400

qu

antu

m e

ffic

ien

cy (

%)

ITT FS9910C (ultra blue gen III)

Hamamatsu V8070U−64 (gen IIIwith GaAsP − CS photocathode)



CCDs or photon counting ? ?


SPID limiting magnitudes

• Telescope diameter : D =3.60m

• Spectral resolution : λ/δλ =3000

• Spatial coherence : r0 =0.15m• Coherence time : τ0 = 20ms• Signal/noise : S/N = 4

Wavelength 550nm 656.3nm

S/N per spatial frequency mv . 11.0 mHα. 11.3

S/N for one parameter model mv . 15.7 mHα. 16.0

At λ/δλ = 60, m =⇒ m + 4.2At τ0 = 40ms, m =⇒ m + 0.75At S/N = 10, m =⇒ m - 1


About the imagecoherence time at CFHT

100 200 300 400 5000.0

0.2

0.4

0.6

0.8

1.0

=⇒ τ0 = 72ms


SPID : first results

-1.0 -0.5 0.0 0.5 1.0

-1.0

-0.5

0.0

0.5

ADS 8804 - separation = 0.63’’. Levels : 200, 400, 700, 900. SAO 6m telescope.

ADS 8804 - separation = 0.63A: 200B: 400C: 700D: 900



-5 0 5

-5

0

5

arcsec -1

arcs

ec -

1

R Cas -Solid contours : TiO band λ=618nm,∆λ=3.5nnmDotted contours : continuum λ=613nm,∆λ=1nnm


Our goals :Atmosphere vertical structure, shockwavepropagation, geometry and homogeneity

CFHT/SPID, 3.60m ESO/SPID,Keck/AO/NIRSPEC, VLTI/AMBER, ...



R Aqr -Left : [OIII] λ=500.65nm, ∆λ=0.5nmRight : TiO λ=504.9nm, ∆λ=3nm


SPID : WFSlost photons=⇒in board SH WFS5 microlens arrays 9× 9 to 40× 40 subpupilsSH data =⇒(residual) wavefront surface=⇒PSF=⇒Deconvolution

Strehl from the phase : 37%, SPID WFS atLyon 1m telescope, r0 ≈4cm (Lane et al.)


SPID : regimesTwo observing/processing regimes

• Diffraction limited raw images : long

exposure, + WFS, multiframe blind

deconvolution

• Seeing limited raw images : short exposure

+ speckle-like, + WFS, + blind

deconvolution


Blind deconvolutionYou don’t need to know your PSF !


distributionobject brightness

CFHT, Roddier et al. 1994)(UH adaptive optics at

PSF

blu

rred

imag

e

restored image restored PSF

convolution + noise

blind deconvolution


SPID : Gain at PUEO NUI

Measurements with BOA (80 act.) at OHP 1.52m.r0 ≈ 5cm, τ0(pupil)&10ms



=⇒Gain by a factor of 2 !

... and with PUEO NUI,how much more ? ? ?


SPID : going to PUEO NUI

• Optics f/20 → f/8

• Cooling water, or chiller

• GUI in case of untrained observers

• Regime tradeoff mostly at the processing

time

• Blind deconvolution Multiframe


SPID and PUEO NUI

There is a dream ! ...


SPID and PUEO NUI

– A laser guide star

– Speckle data processing insensitive to

the tilt

– Now robust and reasonably expansive

thanks to the modeless laser (Pique,

2003)=⇒Much fainter limiting magnitudes


Documents

-SPID- - cfht.hawaii.edu fileCPNG:Comptage de Photons NouvelleG¶en¶eration 10 0 20 30 40 50 wavelength (nm) 400 500 600 700 800 900 quantum efficiency (%) ITT FS9910C (ultra blue