Point spread function reconstruction at W.M. Keck Observatory : progress and beyond

Point spread function reconstruction at W.M. Keck Observatory : progress and beyond Olivier Beltramo-Martin Aix-Marseille Univ., LAM, A*MIDEX, Extra November 9th, 217 - LAM Olivier Beltramo-Martin (LAM) PSF-R at Keck 1/19/217 1 / 2

Outline 1 PSF-R: Scientific motivations 2 Methodology : from telemetry to PSF 3 Towards best-fit PSF-R 4 Summary Olivier Beltramo-Martin (LAM) PSF-R at Keck 1/19/217 2 / 2

What do we care about getting the system PSF? Photometry Galaxies assembly Interstellar med. Morphology PSF Stellar Population Black hole masses Asteroids Companions Astrometry The object brightness distribution is convolved with the PSF : Im = Obj PSF Getting the real object properties requires a deconvolution process of focal-plane images Olivier Beltramo-Martin (LAM) PSF-R at Keck 1/19/217 3 / 2

What are exactly the needs? Galactic center Quasi stellar objects Aims : measuring astrometry at.1 mas accuracy for tracking orbits around the central black hole Issue : sources confusion caused by overlapping Method : PSF model-fitting Aims : measuring masses of the host galaxy bulge and central black hole to evaluate potential correlation Issue : 3-4 order of magnitude on QSO/host galaxy contrast Method : PSF subtraction Olivier Beltramo-Martin (LAM) PSF-R at Keck 1/19/217 4 / 2

What are exactly the needs? Galactic center Quasi stellar objects Aims : measuring astrometry at.1 mas accuracy for tracking orbits around the central black hole Issue : sources confusion caused by overlapping Method : PSF model-fitting Aims : measuring masses of the host galaxy bulge and central black hole to evaluate potential correlation Issue : 3-4 order of magnitude on QSO/host galaxy contrast Method : PSF subtraction Olivier Beltramo-Martin (LAM) PSF-R at Keck 1/19/217 4 / 2

What are exactly the needs? Galactic center Quasi stellar objects Aims : measuring astrometry at.1 mas accuracy for tracking orbits around the central black hole Issue : sources confusion caused by overlapping Method : PSF model-fitting Aims : measuring masses of the host galaxy bulge and central black hole to evaluate potential correlation Issue : 3-4 order of magnitude on QSO/host galaxy contrast Method : PSF subtraction Olivier Beltramo-Martin (LAM) PSF-R at Keck 1/19/217 4 / 2

What does the PSF look like Diffraction-limited Seeing-limited λ/dλ/d λ/r λ/r PSF characterized by the pupil shape, FWHM = λ/d PSF characterized by the atmosphere seeing, FWHM = λ/r Olivier Beltramo-Martin (LAM) PSF-R at Keck 1/19/217 5 / 2

What about the PSF after AO correction? GS Astronomical target Flat wavefront Atmospheric turbulence Distorded wavefront Telescope AO System Corrected wavefront Beam splitter Science Residual camera wavefront Deformable mirror Atmospheric turbulence : Introduce OPD of 1-1 µm Deformable mirror : Restores the wavefront flatness by pushing/pulling on actuators Wavefront sensor : Provides the phase gradient over the pupil Real time Real-time computer : computer Converts recursively WFS Wavefront measurements to DM commandsensor

1 PSF-R: Scientific motivations 2 Methodology : from telemetry to PSF 3 Towards best-fit PSF-R 4 Summary Olivier Beltramo-Martin (LAM) PSF-R at Keck 1/19/217 7 / 2

AO residual Anisoplanatism Static aberrations Fitting error From telemetry to PSF GS Astronomical target Flat wavefront Atmospheric turbulence Distorded wavefront Telescope AO System Corrected wavefront Beam splitter Science Residual camera wavefront Deformable mirror Generally we have : PSFε = F [P(r) exp(iφε(r))] 2 Residual phase : φε = φε }{{} + φncpa + φfield }{{} Real time computer + φ }{{} + φ }{{} WFS measurements : s = Gφε + η = φε = R s Static aberrations : Must be calibrated Anisoplanatism : Statistics known from the Cn 2 (h) profile and the separation Wavefront sensor Fitting error : Statistics known from the seeing

The OTF calculation PSF obtained from the OTF : PSF ε = F 1 [OTF ε ] Residual OTF is expressed from the covariance matrix of the residual phase : OTF ε (u/λ) = drp(ρ)p(r + u) exp (C ε (r, r + u) C ε (, )) P Getting the OTF/PSF is a matter of estimating the covariance of the residual phase C ε (θ) = C ε }{{} AO residual + C (θ, Cn 2 (h)) + C NCPA + C Field (θ) }{{}}{{} Aniso model static aberrations + C (r ) }{{} Fitting model C ε = R ss t R t C η + C alias Olivier Beltramo-Martin (LAM) PSF-R at Keck 1/19/217 9 / 2

Anisoplanatism effect Anisoplanatism is produced by altitude turbulence during AO closed-loop operations and caused by the spatial correlation of the residual phase across the field : Angular anisoplanatism : phase correlation between two separated cylinders Focal anisoplanatism : phase correlation between a cone (LGS) and a cylinder Tip-tilt anisoplanatism : angular anisoplanatism on tip-tilt modes only Olivier Beltramo-Martin (LAM) PSF-R at Keck 1/19/217 1 / 2

Separation [mas] PSF elongation -5 Angular anisoplanatism - GS at (,) -4-3 -2-1 1 2 3 4 5-5 5 Separation [mas] Olivier Beltramo-Martin (LAM) PSF-R at Keck 1/19/217 11 / 2

Separation [mas] PSF elongation -5 Angular-Focal anisoplanatism - GS at (2,) -4-3 -2-1 1 2 3 4 5-5 5 Separation [mas] Olivier Beltramo-Martin (LAM) PSF-R at Keck 1/19/217 11 / 2

Anisoplanatism filter Focal + TT anisoplantism 1 High-order/tip-tilt modes measured using a LGS/NGS at a different location C (C 2 n (h)) = C lgs (C 2 n (h)) + C TT (C 2 n (h)).9.8.7 E2E simulation - FocalxTT Zonal - FocalxTT E2E simulation - Focal Zonal - Focal E2E simulation - TT Zonal - TT.6 Extrapolating the PSF anywhere in the field requires the knowledge of the C 2 n (h) profile : could be provided by external profiler (MASS/DIMM@Mauna Kea).5.4.3.2.2.4.6.8 1 Spatial frequency [D/ 6 units] Olivier Beltramo-Martin (LAM) PSF-R at Keck 1/19/217 12 / 2

Separation [arcsec] Which metrics for evaluating PSF-R? How can we evaluate the efficiency of PSF-R? which metrics should we consider? PSF characteristics Strehl ratio - FWHM - Encircled energy - PSF profile - Reconstruction residual Science estimates Photo-astrometry accuracy B (α, PSF, p) = p(1) (PSF (α x, α y ) + PSF (α x + p(2), α y + p(3))) ε 2 ) (p) = B (α, PSF ε, p ref ) B (α, PŜFε, p 2 2 -.5 Clean Binary on-sky #1 5 4 -.5 Noisy Binary on-sky #1 5 5 -.5 Binary model best-fitted #1 5 5 -.5 Residual #1 4 5 4 4 3 2 3 2 1 3 2 1 1.5 -.5.5.5 -.5.5-1.5 -.5.5.5 -.5.5-5 Olivier Beltramo-Martin (LAM) PSF-R at Keck 1/19/217 13 / 2

PSF profile [max = Strehl] Y axis [mas] Y axis [mas] Y axis [mas] NGSs results KECK NGS PSF SR [H] = 52.9% Reconstructed PSF SR = 56.1% Residual -5-1 -5-1 -5-1 -2-2 -2-3 -3-3 5-4 5-4 5-4 -5-5 -5-5 5-5 5-5 5 X axis [mas] X axis [mas] X axis [mas] 1 Reconstructed PSF Simulated PSF Residual 1-5 1 2 3 4 5 6 7 8 9 1 Angular separation [mas] Olivier Beltramo-Martin (LAM) PSF-R at Keck 1/19/217 14 / 2

Reconstructed FWHM [H-band] Counts Counts Reconstructed Strehl ratio [H-band] NGSs results Residual Units Bias (R. vs G.) 1-σ std SR pts 1.4 4 FWHM mas 2.4 2.8 Photo mag.1/.75.8/.74 Astro mas 1.5/2.5 3.9/5.5 6 5 4 3 2 1 1 2 3 4 5 6 Sky image Strehl ratio [H-band] Single star magnitude accuracy 12 18 Reconstructed PSF Gaussian-fitted PSF 18 Reconstructed PSF Gaussian-fitted PSF 11 16 16 1 14 14 9 12 12 8 1 1 7 8 8 6 6 6 5 4 4 4 2 2 4 5 6 7 8 9 1 11 Sky image FWHM [H-band] -.4 -.2.2.4.6.8 1 Magnitude accuracy [mag] 5 1 15 2 25 Astrometric accuracy [mas] Olivier Beltramo-Martin (LAM) PSF-R at Keck 1/19/217 14 / 2

PSF profile [max = Strehl] Y axis [mas] Y axis [mas] Y axis [mas] LGSs results KECK LGS PSF SR [H] = 25.6% Reconstructed PSF SR = 27.7% Residual -5-1 -5-1 -5-1 -2-2 -2-3 -3-3 5-4 5-4 5-4 -5-5 -5-5 5-5 5-5 5 X axis [mas] X axis [mas] X axis [mas] 1 Reconstructed PSF Simulated PSF Residual 1-2 1-4 1 2 3 4 5 6 7 8 9 1 Angular separation [mas] Olivier Beltramo-Martin (LAM) PSF-R at Keck 1/19/217 15 / 2

Reconstructed FWHM [H-band] Counts Counts Reconstructed Strehl ratio [K-band] LGSs results Residual Units Bias (R. vs G.) 1-σ std SR pts -.13 2.7 FWHM mas 3.5 14.4 Photo mag -.13/.57.15/.61 Astro mas 2.5/7.6 2.8/8. 3 28 26 24 22 2 18 16 14 12 12 14 16 18 2 22 24 26 28 3 Sky image Strehl ratio [K-band] Single star magnitude accuracy 12 25 Reconstructed PSF Gaussian-fitted PSF 4 Reconstructed PSF Gaussian-fitted PSF 35 11 2 3 1 15 25 2 9 1 15 8 5 1 5 7 7 8 9 1 11 12 Sky image FWHM [H-band] -.4 -.2.2.4.6.8 Magnitude accuracy [mag] 5 1 15 2 25 3 35 4 Astrometric accuracy [mas] Olivier Beltramo-Martin (LAM) PSF-R at Keck 1/19/217 15 / 2

1 PSF-R: Scientific motivations 2 Methodology : from telemetry to PSF 3 Towards best-fit PSF-R 4 Summary Olivier Beltramo-Martin (LAM) PSF-R at Keck 1/19/217 16 / 2

Tracking the anisoplanatism from focal-plane images Seeing limited AO-correction λ/dλ/d λ/r λ/r Seeing is measurable from focal-plane PSF FWHM C 2 n (h) profile is measurable from focal-plane PSF morphology Olivier Beltramo-Martin (LAM) PSF-R at Keck 1/19/217 17 / 2

Focal Plane Profiler: concept Classical PSF-R Reconstruct the on-axis PSF from the telemetry Grab the C 2 n (h) profile from an external profiler Extrapolate the PSF off-axis from an anisoplanatism model Errors on anisoplanatism model and C 2 n (h) estimation generate reconstruction deviations Focal Plane Profiler Reconstruct the on-axis PSF from the telemetry Pre-compute normalized (C 2 n (h) = 1) C and Fitting/Aliasing PSDs Model-fit the C 2 n (h) by minimizing difference PSF model/observations Errors on anisoplanatism model are compensated and the C 2 n (h) profile is estimated from focal-plane images Olivier Beltramo-Martin (LAM) PSF-R at Keck 1/19/217 18 / 2

Application to the HeNOS bench off-axis Bench PSFs 1 Reconstructed PSFs 1.9.9 Asterism radius 5 Sources wavelength 67 nm r.8297 67 nm fractional r 74.3%,17.4%,8.2% altitude layer (.6, 5.2, 16.3) km source height 98.5 km Telescope diameter 8.13 m DM actuator pitch.813 m 5 1 15 2.8.7.6.5.4.3.2 5 1 15 2.8.7.6.5.4.3.2.1.1 5 1 15 2 5 1 15 2 LGS On-axis LGS 2 LGS 3 LGS 4 Ref PSFR FPP Ref PSFR FPP Ref PSFR FPP Ref PSFR FPP FWHM 21.1 22.6 22.6 116.6 97 112 9 8 93 1 77 91 r fr (1) fr (2) fr (3) Ref 82.97 cm 74.3% 17.4% 8.2% Fitted 82.88 cm ± 2.7 68.1%±4.5 23.35%±2.7 8.5%±1.2 Olivier Beltramo-Martin (LAM) PSF-R at Keck 1/19/217 19 / 2

Summary PSF-R 4 Science GC : estimate photo-astrometry in crowded field to get stars orbits around the BH QSO : subtract the PSF to reveal the host galaxy and obtain galaxy/smbh masses PSF-R : provide the AO-compensated PSF from telemetry + models + calibrations Results on Keck in NGS/LGS mode Preliminary results : reconstruction in on-axis NGS/LGS with 3 pts/1 mas of accuracy on Strehl/FWHM.1mag/2 mas -ish of accuracy on photometry/astrometry Next step : getting a better evaluation of the photo-astrometry accuracy through StarFinder Focal plane profiler Cn 2 (h) retrieval from the AO-compensated focal plane images Successfully applied to HeNOS : error FWHM 2% > 5% Next step : test FPP on NIRC2/OSIRIS images to compare with MASS/DIMM Olivier Beltramo-Martin (LAM) PSF-R at Keck 1/19/217 2 / 2

Summary PSF-R 4 Science GC : estimate photo-astrometry in crowded field to get stars orbits around the BH QSO : subtract the PSF to reveal the host galaxy and obtain galaxy/smbh masses PSF-R : provide the AO-compensated PSF from telemetry + models + calibrations Results on Keck in NGS/LGS mode Preliminary results : reconstruction in on-axis NGS/LGS with 3 pts/1 mas of accuracy on Strehl/FWHM.1mag/2 mas -ish of accuracy on photometry/astrometry Next step : getting a better evaluation of the photo-astrometry accuracy through StarFinder Focal plane profiler Cn 2 (h) retrieval from the AO-compensated focal plane images Successfully applied to HeNOS : error FWHM 2% > 5% Next step : test FPP on NIRC2/OSIRIS images to compare with MASS/DIMM Thank you! Olivier Beltramo-Martin (LAM) PSF-R at Keck 1/19/217 2 / 2