INTEGRATION OF SAXO, THE VLT-SPHERE EXTREME AO: FINAL PERFORMANCE

Transcription

1 Florence, Italy. Adaptive May 2013 Optics for Extremely Large Telescopes III ISBN: DOI: /AO4ELT INTEGRATION OF SAXO, THE VLT-SPHERE EXTREME AO: FINAL PERFORMANCE Thierry Fusco 1,2,a, Jean-Francois Sauvage 1, Cyril Petit 1, Anne Costille 2, Kjetil Dohlen 2, Jean-Luc Beuzit 3, David Mouillet 3, Pascal Puget 3, Laurence Gluck 3, Sylvain Rochat 3, Andréa Barrufolo 4, Bernardo Salasnich 4, Markus Kasper 5, Marcos Suarez 5, Christian Soenke 5, Enrico Fedrigo 5, Pierre Baudoz 6, Arnaud Sevin 6, Denis Perret 6, François Wildi 7 1 ONERA (Office National d Etudes et de Recherches Aérospatiales),B.P.72, F Chatillon, France 2 LAM, UMR6110, CNRS/Université de Provence, B.P. 8, F Marseille Cedex 12, France 3 IPAG, UMR5274, CNRS/Université J. Fourier, B.P. 53, F Grenoble Cedex 9, France; 4 Osservatorio Astronomico di Padova, INAF, Vicolo dell Osservatorio 5, I Padova, Italy; 5 European Southern Observatory, Karl-Schwarzschild-Strasse 2, D Garching, Germany; 6 LESIA, UMR8109, CNRS/Observatoire de Paris, 5 place J. Janssen, F Meudon, France; 7 Observatoire de Genève, 51 chemin des Maillettes, CH-1290 Sauverny, Switzerland Abstract. Direct imaging of exoplanet is one of the most exciting fields of astronomy today. The light coming from exoplanet orbiting their host star witnesses for the chemical composition of the atmosphere, and the potential biomarkers for life. However, the faint flux to be imaged, very close to the huge flux of the parent star, makes this kind of observation extremely difficult to perform from the ground. The direct imaging instruments (SPHERE [1], GPI [2]) are nowadays reaching lab maturity. Such instruments imply the coordination of XAO for atmospherical turbulence real-time correction, coronagraphy for star light extinction, IR Dual band camera, IFS, and visible polarimetry. The imaging modes include single and double difference (spectral and angular). The SPHERE project is now at the end of AIT phase. This paper presents the very last results obtained in laboratory, with realistic working conditions. These AIT results allow one to predict on-sky performance, that should come within the next weeks after re-installation at Very Large Telescope at Paranal. Keywords: High contrast imaging, extreme adaptive optics, exoplanet direct imaging 1. Introduction The primary goal of Extrasolar Planet science of the next decade is a better understanding of the mechanism of formation and evolution of planetary systems. SPHERE will give a primary contribution in this area. Determination of the frequency of giant planets in wide orbits (> 5-10 AU) will allow testing basic aspects of the planet formation models. High-resolution, high contrast imaging like that provided by SPHERE is expected to be the most efficient technique to discover planets in the outer regions of planetary systems. With its enhanced capabilities (a gain of two orders of magnitudes in contrast with respect to existing instruments) and a list of potential targets including several hundred stars, SPHERE will provide a clear view of the frequency of giant planets in wide orbits. Both evolved and young planetary systems will be detected, respectively through their reflected light (mostly by visible differential polarimetry [8]) and through the intrinsic planet emission (using IR differential a thierry.fusco@onera.fr

2 imaging IRDIS [6] and integral field spectroscopy IFS [7]). Beside frequency it would also be interesting to derive the distributions of planets parameters such as mass, semi-major axis and eccentricities. The main scientific goal of SPHERE will then be the description of the properties of young planets in the expected peak region of gas giant formation and in the outer regions of the systems. Furthermore, a direct imager like SPHERE provides the only way of obtaining spectral characteristics for outer planets. Finally, a few planets shining by reflecting stellar light might be detected by the SPHERE polarimetric channel (ZIMPOL). SPHERE will be then highly complementary to current and contemporaneous studies of extrasolar planets. The present paper presents the status of the SPHERE instrument in section 2, then demonstrates the system performance in section 3, and ends up with a study of the relevance of a laboratory demonstration for an instrument dedicated to complex sky observations in section Sphere overview Common Path IRDIS CPI ZIMPOL IFS Fore optics SAXO Vis Coronagraph ZIMPOL NIR Coronagraph IFS IRDIS Figure 1 SPHERE four sub-systems Including in the common path (CPI) the high order adaptive optics system (SAXO), coronagraphs, and the three observing subsystems IRDIS, IFS and ZIMPOL (see text). Left: block diagram. Right: schematic view on the Nasmyth platform. The SPHERE instrument lies on the VLT Nasmyth platform and is divided into four subsystems (Figure 1): The Common Path and Infrastructure (CPI) receives the light from the telescope, supports the 3 other sub-systems and feed them with highly stabilized, AO-corrected and coronagraphic beams. ZIMPOL produces either images or differential polarimetric images in the visible range. The integral field spectrograph (IFS) produces spectra on each point of the internal FoV in NIR. The infrared dual-band imager and spectrograph (IRDIS) also works in NIR with a larger FoV and various modes: classical imaging (CI), dual-band imaging (DBI), dual-polarization imaging (DPI), or long slit spectroscopy (LSS). The whole SPHERE design and its more severe specifications have been driven by the primary science case of exoplanet imaging (wide and efficient detection surveys and characterization capabilities). It provides a very high image quality of a narrow field around bright targets (corrected from turbulence and highly stabilized) that can be observed in visible or near infrared. Observations can be made without or with a variety of coronagraphs. In visible with ZIMPOL, imaging (centered on the bright star or offset up to a radius of 4 ) is possible in a variety of narrow band to broad band filters, down to 15 mas diffraction-limited angular resolution obtained in good conditions of AO turbulence correction. Interestingly, very accurate differential

3 polarimetric imaging, obtained quasi-simultaneously by fast modulation, can reveal very faint circumstellar linearly polarized light (such as reflected light). ZIMPOL cannot be used simultaneously with the other sub-systems IFS or IRDIS in the NIR. This imager IRDIS can also be used alone in the following modes: Dual-band imaging (DBI) with other pairs of filters, covering the expected cool companion spectral features over the Y to K band domain Dual-polarization imaging (DPI) where the two quadrants of the detector image simultaneously two orthogonal linear polarization states. Classical imaging (CI) in a variety of narrow band to broad band filters from Y to K band Long slit spectroscopy (LSS) at low resolution (LRS) over Y-K band or at medium resolution (MRS) over Y-H band. 3. SPHERE SAXO status and performance The SPHERE instrument is now at the very end of AIT phase. During the last weeks, all the performance have been demonstrated and validated at IPAG laboratory, for both infrared and visible path. The operational scheme of the instrument is now being validated. The acceptance phase is ongoing on September and October SAXO high level requirements [3] have been extracted from the SPHERE Technical Specification document and the SPHERE sub-system functional requirements. They are listed hereafter: residual TT, 3 mas rms (goal 1.5 mas) turbulent residual wave front variance on corrected modes: 60 nm rms SR(1.6 mm) > 15% in poor or faint conditions Ability to stabilize pupil in translation: < 0.2% (goal = 0.1%) of pupil diameter Ability to reproduce an image position and to stabilize the image in translation, (hence compensate image movements due to thermo-mechanical effects and differential atmospheric dispersion between Vis and NIR bands) with accuracy better than 0.5 mas (goal 0.2 mas) The residual non common path aberrations after phase diversity measurement and AO precompensation shall be lower than 0.8 (goal 0.4) nm per mode. AO system shall pre-compensate for 50 nm rms of non-common path defocus and 40 nm rms of the 55 first Zernike modes. These values have been driven all the SAXO design. The SAXO system is composed by 3 loops + one off line calibration Main AO loop (1.2 khz): correct for atmospheric, telescope and common path defects. The main impact is the increase of detection signal to noise ratio through the reduction of the smooth PSF halo due to turbulence effects HODM and TTM rejection transfer function shows that specification in terms of sampling frequency (1200 Hz) and pure delay (100 µs) have been achieved and even more exceeded since (80 µs of pure delay has been measured on the SPARTA RTC). Figure 2 Rejection Transfer Function for HODM (left) and TTM (right). For HODM, the average of RTF per actuator is plotted for various AO loop gain. For TTM, the average of Tip and Tilt is plotted for a 0.3 gain. Both

4 results are consistent with an overall AO loop delay of 2.14 frame at 1200 Hz which is beyond the initial specifications (2.2 frames at 1200 Hz). In addition to classical Optimal modal gain feature a dedicated LQG controller has been implemented for TT in order to be able to deal with any (up to twelve vibration peaks between 0 and 300 Hz) exogenous vibrations as well as to optimally correct for turbulence itself. Tilt residual PSD Cumulative PSD Figure 3 PSD [Left] and cumulative PDS [Right] of residual Tilt. Black : classical integrator with the optimal gain, Red : LQG. Vibration peak is clearly filtered and the low temporal frequency better attenuated with LQG With a typical seeing condition of 0.85 arcsec and 10m/s of wind speed, it has been demonstrated that the residual jiter is going from 1.5 mas rms in the case of a pure integrator down to less than 1 mas rms with the LQG. Again this value is by far better than the original specification of 2.1 mas rms per axis. The DTT loop for fine centering on coronagraph mask (correction of differential tip-tilt between VIS and IR channel). It will ensure an optimal performance of the coronagraph device The PTT loop for pupil shift correction (telescope and instrument). It will ensure that the uncorrected instrumental aberrations effects (in the focal plane) will always be located at the same position and thus will be canceled out by a clever post-processing procedure NCPA pre-compensation which will lead to the reduction of persistent speckle [4]. The measurement is done thank to a optimized iterative phase diversity process. Defocus is applied to the DM and a pseudo closed loop (measurements by phase diversity and correction by a modification of the reference slopes) is performed till the residual NCPA measured directly on the images are as close as possible to zero. The residual rms obtained with this process is lower than 5 mas rms. The measured SR on the image itself is larger than 99 % in H! The SAXO nominal performance is define for the following conditions: Vmag~9, i.e. 100 photo-e- per sub-aperture and per frame on the WFS (@1.2 khz) Seeing = 0.85 and wind speed = 12.5 m/s In these conditions, with a fully operational and optimized SAXO system (all loops closed, all optimization process on including Kalman Filtering on Tip Tilt) we have recorded classical PSFs and coronagraphic images on IRDIS (H2H3 mode). Direct imaging will give us Strehl ratio and will allow us quantify AO loop residuals in terms of nm rms. The coronagraphic image will provide the final performance in terms of detectivity High Flux (mag<9) / Ultimate performance Figure 4 and

5 Figure 5 show the experimental results for typical seeing conditions. Figure 5 also compares experimental coronagraphic profile with a simulated one. On classical imaging, it is shown that AO performance exceeds 90% of SR in H band whereas the coronagraphic profile demonstrate that the SAXO system behaves as expected since the simulation tools allows to accurately reproduce the experimental curves. Figure 4 Classical and coronagraphic PSF (turbulent and AO corrected) for typical seeing conditions (0.85, 10m/s wind speed, 100 ph/sub-ap/frame (magv= 9)) The pure coronagraphic contrast is below10-4 in the corrected area (it goes down around mas). On top of that, the full system will add the post processing aspects (Spectral Dual-band Imaging [SDI], Angular Dual-band Imaging [ADI], spectral deconvolution etc ) which give us confident in the ability of the full SPHERE system to reach or even exceed its top level requirements (10-6 extinction). The small difference (plateau around mas in the experimental data) is due to the 15 (over 1377 active) dead actuators in the Cilas high order DM. These dead actuators (stuck at given values) create speckles in the focal plane. Note that these speckles should be removed using SDI or ADI and at the end of the day will only lead to a small increase of photon noise. F Focal plane obstructed area Figure 5 PSF profile (classical and coronagraphic). In red: measured coronagraphic profile, Blue: End2end simulation Flux dependency Low Flux performance The previous Section has demonstrated that, despite the DM issues) SAXO is achieving its required performance for nominal conditions (exceeding the goals in some aspects). We are now going to focus on performance in faint conditions.

6 Figure 6 shows the evolution of SAXO performance (in terms of SR and PSF shape) for various SNR conditions. R magnitudes have been computed from flux measured on VIS-WFS sub-apertures and system measured transmissions. It shows that ultimate performance is achieved (in H band) for mag < 9 and that the performance smoothly drops down to 15 % for mag > 15.5 which is a remarkable result. This is especially due to the very good system transmission and the exceptional performance of the WFS in terms of RON (<< 1e-) and QE (deep depletion device). Figure 6[UP] SR evolution as a function of GS magnitude for various wavelengths. [DOWN] example of PSF for various GS magnitude (in H band) 3.3. Performance in poor atmospheric conditions In this section, the performance (as well as the robustness) of the system is studied in the poor condition regime, i.e. a seeing of 1.12 arcsec and two wind speed values of 12.5 and 30 m/s. First of all, despite some actuators in saturation (well handled by the anti-wind up and Garbage Collector processes), the loop was stable and robust during all the acquisition process (a few tens of minutes).

7 Figure 7 Non-coronoagraphic PSF, for poor conditions (1.12 arcsec seeing) and two wind speed values: 12.5 m/s [left] and 30m/s [right]. Estimates SR are respectively 85.5± 2 and 73.3 ± 2 % Measured SR are quite impressive with 85 % for 12.5 m/s (mag = 8) and 73 % for 30 m/s. Both results are fully compatible with theoritical values and show that the system behaves perfectly well at large seeing or wind speed. In addition to the SR value, a coronapraphic profile (see Figure 8) is plotted showing a smooth and expected behavior. Figure 8 Coronagraphic extinction in poor conditions (30m/s, mag 8, seeing 1.12 arcsec). The apodizer transmission has been artificially accounted 4. Conclusion The intensive tests made on the SAXO system during years have shown that, despite the DM defects (shape at rest, dead actuators) the SAXO performance is well within the specifications and for some of them even exceed the goals. It has been shown that, in addition to its ultimate performance, the AO system was stable (loop closed for hours without problem) and robust to the evolution of operation conditions (wind speed change, poor turbulence conditions etc ). The SAXO system is now considered as fully operational (simple calibration and operation processes through templates without the need of any AO specialist), meeting all its initial requirements and ready to go on sky. The SPHERE instrument is now at the end of AIT phase and has demonstrated full performance in laboratory scheme. The demonstration of operational scheme is now ongoing. Acceptance visits are planned until the beginning of November 2013, preliminary acceptance is foreseen in November The first light of the instrument is foreseen in spring 2014.

8 5. References [1] Beuzit, J.-L., Mouillet, D., Moutou, C., Dohlen, K., Puget, P., Fusco, T. and Boccaletti, A., "A planet finder instrument for the VLT," Proc. of IAU Colloquium 200, Direct Imaging of Exoplanets: Science & Techniques, Cambridge University Press, pp , (2005) [2] Bruce A. Macintosh ; Andre Anthony ; Jennifer Atwood ; Nicolas Barriga ; Brian Bauman et al " The Gemini Planet Imager: integration and status ", Proc. SPIE 8446, Ground-based and Airborne Instrumentation for Astronomy IV, 84461U (September 24, 2012); doi: / ; [3] Fusco, T., Rousset, G.,Sauvage, J.-F.,Petit, C., Beuzit, J.-L., Dohlen, K., Mouillet, D., Charton, J., Nicolle, M., Kasper, M., Baudoz, P. and Puget, P.,"High-order adaptive optics requirements for direct detection of extrasolar planets: Application to the SPHERE instrument, " Opt. Express 14, (2006) [5] Sauvage, J.-F., Fusco, T., LeMignant, D., Petit, C., Sevin, A., Dohlen, K., Robert, C., Mugnier, L.,"SPHERE non-common path aberrations measurement and pre-compensation with optimized phase diversity processes: experimental results," AO for ELT Proc., second edition, (2011) [6] Kjetil Dohlen, Maud P. Langlois, Michel Saisse, The infra red dual imaging and spectrograph for SPHERE: design and performance, in Ground-based and Airborne Instrumentation for Astronomy II, SPIE (2008) [7] Riccardo U. Claudi, Massimo Turatto, Raffaele G. Gratton, Jacopo Antichi,; Enrico Cascone,; Vincenzo De Caprio,; Silvano Desidera, Dino Mesa, Salvatore Scuderi, SPHERE IFS: the spectro differential imager of the VLT for exoplanets search, in Ground-based and Airborne Instrumentation for Astronomy II, SPIE (2008) [8] Christian Thalmann, Hans M. Schmid, Anthony Boccaletti, David Mouillet, Kjetil Dohlen, Ronald Roelfsema, Marcel Carbillet, Daniel Gisler, Jean-Luc Beuzit, Markus Feldt, Raffaele Gratton, Franco Joos, Christoph U. Keller, Jan Kragt, Johan H. Pragt, Pascal Puget, Florence Rigal, Frans Snik, Rens Waters, Francois Wildi, SPHERE ZIMPOL: Overview and performance simulation, in Ground-based and Airborne Instrumentation for Astronomy II, SPIE (2008)