Nicolas B LIND

Texte intégral

(1)Observation de binaires en interaction à très haute résolution angulaire Suivi de franges pour l’interférométrie infrarouge Soutenance de thèse. Nicolas BLIND. Directeurs de thèse Jean-Philippe Alain. 1. BERGER CHELLI.

(2) The need for very high angular resolution. 2.

(3) The need for very high angular resolution. 8 meters 50 mas. 3.

(4) The need for very high angular resolution T Tauri star ~10 mas. 8 meters 50 mas. Stellar surface. Interacting binary ~5 mas. ~20 mas. 4.

(5) The need for very high angular resolution Synthetizing a giant telescope by combining several telescopes → angular resolution x25. Baselin. eB. 200 m. 5.

(6) Interferometric observables φ phase. = TF(source)(B/λ). 1.0. Normalized amplitude. V. eiφ. 0.0. Single telescope. Baselin. Spatia l. V visibility. 0.5. −5. 0. 5. 10. opd [µm]. eB. freque. ncy B/. λ. Interferometer. Fringe pattern. 6.

(7) Interferometric observables 3T. 5.. 10+7. 0.. Parametric modeling. Visibility. Spatial frequency East [m]. 1.0. 2mas. Parametric modeling 0.5. −5. −5.. 0.. 10+7. 0.0 0.. 5.. 2.. 4.. 6.. Spatial frequencies. Spatial frequency North [m]. 3T. 7. 10+7. 8..

(8) Interferometric observables 1.0. 4T. 10+7 2.. 0.. Parametric modeling. −2.. Visibility. Spatial frequency East [m]. 4.. 2mas. Parametric modeling 0.5. −4. −4.. −2.. 0.. 2.. 10+7. 0.0 0.. 4.. 2.. 4.. 6.. Spatial frequencies. Spatial frequency North [m]. 4T. 8. 10+7. 8..

(9) Interferometric observables 1.0. 1.0. 0.5. 0.0. Parametric modeling. −0.5. −1.0 −1.0. −0.5. 0.0. 0.5. 10+8. Visibility. Spatial frequency East [m]. 10+8. 2mas 0.5. 0.0 0.. 1.0. 2.. 4.. 6.. Spatial frequencies. Spatial frequency North [m]. 6T. 9. 10+7. 8..

(10) Interferometric observables 1.0. 1.0. V eiφ = TF(objet)(B/λ). 0.5. Visibility. Spatial frequency East [m]. 10+8. 0.0. 2mas 0.5. −0.5. −1.0 −1.0. −0.5. 0.0. 0.5. 10+8. 0.0 0.. 1.0. Spatial frequency North [m]. 2.. 4.. 6.. Spatial frequencies. 10 7. 10+7. 8..

(11) Interferometric observables 1.0. 1.0. V eiφ = TF(source)(B/λ). 0.5. Visibility. Spatial frequency East [m]. 10+8. 0.0. 2mas 0.5. −0.5. −1.0 −1.0. −0.5. 0.0. 0.5. 10+8. 0.0 0.. 1.0. Spatial frequency North [m]. 2.. 4.. 6.. 10+7. 8.. Spatial frequencies. Image reconstruction. Parametric modeling. T Leporis: Mira star. Altair: Fast rotator. Interference fringes. as. as. 2m. 10 m Lebouquin et al A&A 2009 11 7. Monnier et al. Science 2007.

(12) Presentation outline An introduction to interferometry PART I. Study of interacting binaries • Interest of interacting binaries • The promises of interferometry • The case of SS Leporis PART II. Fringe trackers for imaging instruments Conclusions and perspectives. 12.

(13) Why to study interacting binaries? At least 50% of stars are binaries Determining mass Constraints on stellar evolution models. 13.

(14) Why to study interacting binaries? Interacting binary: 2 stars exchanging matter + complex structures (accretion disk, jets, nebula, etc.). Excellent laboratories to study numerous physical processes. Properties relevant to many astrophysical objects. Evolution dominated by mass transfer processes. 14.

(15) Mass transfer processes Evolution dominated by mass transfer processes Roche lobe. Stellar wind accretion. Roche lobe overflow (RLOF) 15.



(18) Issues with indirect observables Spectroscopy or Photometry → Indirect observables → Assumptions required Can lead to conflicting observations, e.g.: Difference by a factor of 2 between stellar radii derived from Light Curve & Rotationnal Velocities. Radii from RV. Mikolajewska, Baltic Astro, 2007. Radii fromvelocities LC rotationnal 18.

(19) The breakthrough of interferometry Spectroscopy or Photometry → Indirect observables → Assumptions required. ~ 1 mas. < 1 AU Optical interferometry → Direct observables. β Lyr (Zhao et al. 2008) Image. Constraints on physical sizes, morphology High benefits from new imaging capability. Model. 1 mas Accretion disk 19. Distorted giant.

(20) The case of SS Leporis M giant + oversized A dwarf + dusty disk. SED. Algol paradox Mass ratio MA/MM~ 2 to 4 → hints for mass transfer. Roche lobe overflow Distance ~ 270 to 370 pc. A star. Orbit: - P = 260d - Quasi circular - Inclination estimated to 30°±10° - Separation ???. Envelope. evolved M star A star. ??? ~12 mas. Verhoelst et al. 2007, Welty et al. 1995, Jura et al. 2001 20. circumbinary material.

(21) VLTI observations 8 observations over 3 revolutions: • 4 AMBER (3T), H&K, R~40 • 4 PIONIER (4T), H, R~40 Method: 1. Images with PIONIER 2. Parametric modeling: morphology, orbit & energy balance. Fundamental parameters (M, T). Constraints on the mass transfer 21.

(22) VLTI observations PIONIER images: SS Lep as a visual binary @ 3 epochs Commissionning data: resolution ~1mas Thanks to 4T, image reconstructions in <4h hours with MIRA (Thiebaut SPIE 2008) → with AMBER (3T) several nights required... 1st images of an interacting binary & orbital motion @ VLTI 28−10−2010. 07−12−2010. 22−12−2010. 5. N (mas) −−−>. V [106 rad−1]. 50. ~5mas. 0. A star. −50. −50. 0. U. [106. M giant. 50. rad−1]. 0. ~2mas −5 −5. 0 E (mas) −−−>. 5. −5. 0 E (mas) −−−>. 22. 5. −5. 0 E (mas) −−−>. 5.

(23) VLTI observations Parametric modeling Resolved M star + unresolved A star + circumbinary material Free parameters: • Binary separation & orientation • M giant diameter (Uniform disk) • Dusty envelope size (gaussian shape) + Wavelength dependency of parameters: color Starting point: PIONIER images 1.0. PIONIER UV plan. Binary (~5mas) M giant (~2mas). Envelope (~12mas) 50. Visibility. 0. Phase closure. V2. V [106 rad−1]. 50. 0.5. 0. + Data - Model. −50. −50. 0. 50. 0.0. Closure Phase 0. 10. 20. U [106 rad−1]. B/λ. 23. 30. 40. −50. 1.60. 1.65. 1.70. λ. 1.75. 1.80.

(24) The energy balance Individual spectroscopy of the 3 components at a 1-mas resolution 5. M star MARCS model 3200±200K Metallicity?. 2MASS photometry. λ Fλ [W.m−2]. 2. M star. 10−10. A star Rayleigh-Jeans, 9000K (SED) 10x oversized (∅~18R⊙) ? OR accretion disk ?. 5. Envelope 2. Envelope BB@1700K, gaussian FWHM ~8mas. 10−11. A star 1.6. 1.8. 2.0. λ [microns]. Sharper analysis: need for a high resolution spectrum 24. 2.2. 2.4.

(25) The orbit and masses Radial velocity + astrometry PIONIER: unambiguous positions. 4. a~. N (mas) −−−>. 2. 1.3. 0. AU. AMBER: good orbit sampling. −2. −4. i ~ 37° −4. −2. 0. 2. 4. E (mas) −−−>. d [pc] MA [M⊙] MM [M⊙] MA/MM. Before 330±70 2~3 0.35~1 4±1. Now 280±25 (Hipparcos) 2.7 ± 0.3 1.3 ± 0.3 2.2 ± 0.3. Errors dominated by the distance uncertainty 25.

(26) Mass transfer: stellar wind accretion! ∅M [mas] d [pc] ∅M [R⊙] Roche lobe filling. Before 3.1 ± 0.3 330±70 220±60 140±20 %. Now 2.2 ± 0.01 280±25 130±7 85±3 %. Errors dominated by the distance uncertainty. Α. No Roche lobe overflow. Μ. Stellar wind accretion. a ~ 1.3 AU. 26.

(27) A new vision of SS Lep Coll. H. BOFFIN Scenario for the accretion process Ideal candidate to test theories of accretion & mass loss. Enhanced mass loss ~10-6 M⊙/year (Tout & Eggleton 1988). Wind accretion efficiency >> 10% (Nagae et al. 2004). <1mas Accretion disk 27.

(28) Perspectives for SS Lep New vision of the system, important constraints on the mass transfer Related publications and communications:. • Conf Evolution of compact binaries (03/2011) • Conf 10 years VLTI (10/2011) • Publi. Blind et al. A&A accepted • Press release ESO (30/11/2011), A&A Future work on SS Lep: • Circumbinary envelope morphology • Tidal distortion of the M giant ? • Accretion disk or oversized star? + simultaneous spectro/photometry. → NaCo/SAM + PIONIER → PIONIER → VEGA/CHARA. 10/2011 - SS Lep observed with VEGA-3T «P Cygni» Hα line spatially & spectrally resolved 28.

(29) How to go further ? To do more physics, we need:. • More objects: • Spectro-imaging:. Increasing limiting magnitude More telescopes + spectral resolution. Increasing sensitivity of the observations. Need for a multi-telescope fringe tracker. 29.

(30) Presentation outline An introduction to interferometry PART I. Study of interacting binaries PART II. Fringe trackers for imaging instruments • Context • Definition of a fringe tracker concept • The POPS concept Conclusions and perspectives. 30.

(31) Dealing with the atmosphere.... Turbulence. Integration time < 10 ms → low sensitivity. 31.

(32) ... by using a fringe tracker Fringe tracking: measuring and compensating in real time the randomly variing fringe position → sensitivity x1000. Turbulence. Fringe tracker OFF Integration time ~10ms. Fringe tracker ON Integration time ~5s !. Faint emission line SPECTROSCOPY !. 32.

(33) Fringe tracking at VLTI today FINITO. PRIMA-FSU. 3T - 2 baselines Temporal fringe sampling. 2T Static fringe sampling Off-axis tracking. Lebouquin et al SPIE 2008. Performances. Sahlmann et al A&A 2009. FINITO + ATs. Fringe tracking 5.5 Mag limit: H=5.5 @ ATs Fringe detection -. 33. PRIMA + ATs. PRIMA + UTs. ~8. 9. ~ 10. 11.7.

(34) A fringe tracker at VLTI tomorrow... More telescopes → increasing complexity. 2T 2T → 1 baseline 3T → 3 baselines 4T → 6 baselines 6T → 15 baselines ... 34.




(38) A fringe tracker at VLTI tomorrow... Context: phase A study for ESO 2GFT My implication: system analysis Detector Interferometric combination scheme. Telescopes Wavefront Reference source. 2nd generation fringe tracker To be defined. Fringe position estimation. Closed-loop control architecture Fringes tracked 38.

(39) A fringe tracker at VLTI tomorrow... Context: phase A study for ESO 2GFT My implication: system analysis Detector Interferometric combination scheme. Telescopes Wavefront Reference source. 2nd generation fringe tracker To be defined. Fringe position estimation. Closed-loop control architecture Fringes tracked 39.

(40) 1. To filter or not to filter? Coll. E. TATULLI. Flux. Without filtering Bulk optics flat WaveFront. Low visibility High flux. disturbed WF. (AO) partially corrected WF. With filtering Fibers Integrated optics Telescope 40.

(41) 1. To filter or not to filter? Coll. E. TATULLI. Flux. Without filtering Bulk optics flat WaveFront. disturbed WF. (AO). High visibility. partially corrected WF. With filtering Fibers Integrated optics Telescope 41. Low flux.

(42) 1. To filter or not to filter? Coll. E. TATULLI. Flux. Without filtering Bulk optics flat WaveFront. disturbed WF. (AO) partially corrected WF. ?. With filtering Fibers Integrated optics. Telescope 42.

(43) 1. To filter or not to filter... Coll. E. TATULLI. Without filtering Bulk optics. magnitude 15. flat WaveFront. 10. 5. 0. phase error ratio. 4. disturbed WF. 3. no AO, S = 0.2 S = 0.5 S = 0.9 S = 0.99. 2. 1. (AO). 10+0. partially corrected WF. 10+2. 10+4. 10+6. # photo−events. With filtering Fibers Integrated optics Telescope 43. Equivalent performances Tatulli, Blind et al. A&A, 2010. 10+8.

(44) 1. More important: use an AO! Coll. E. TATULLI. Without filtering Bulk optics. magnitude 15. flat WaveFront. 5. 0. phase error ratio. 4. disturbed WF. 3. 10+0. partially corrected WF. no AO, S = 0.2 S = 0.5 S = 0.9 S = 0.99. 2. 1. (AO). AO: gain up to 3 mag on ATs !!! (sensitivity x15). 10. Tatulli, Blind et al. A&A, 2010 10+2. 10+4. 10+6. # photo−events. With filtering Fibers Integrated optics. 44. Equivalent performances Tatulli, Blind et al. A&A, 2010. 10+8.

(45) 2. Combination concept: the issue 4T case All-in-one multi-axial. Pairwise co-axial. 24 pixels max. ~100 pixels Tarmoul et al. SPIE 2010. 45.

(46) 2. Co-axial combination concept 4T case All-in-one multi-axial. Pairwise co-axial. 24 pixels max. ~100 pixels Tarmoul et al. SPIE 2010. Limitation: slow and noisy detectors (HAWAI II RG) → All-in-one multi-axial handicaped by the high number of pixels. 46.

(47) 2. Co-axial pairwise combination schemes Cophasing N telescopes requires N-1 to N(N-1)/2 baselines. 47.

(48) 2. Co-axial pairwise combination schemes Cophasing N telescopes requires N-1 to N(N-1)/2 baselines. Precision. 1. Ideal conditions: every baseline equivalent 2. Resolved source: 1 baseline with low visibilities 3. Low flux: several baselines with low flux and visibility. 48.

(49) 2. Co-axial pairwise combination schemes Cophasing N telescopes requires N-1 to N(N-1)/2 baselines. Precision Photometry. Extracting the photometry from the fringe signal is only possible for telescopes part of a sub-array constituted of an odd number of telescopes. 49.

(50) 2. Co-axial pairwise combination schemes Cophasing N telescopes requires N-1 to N(N-1)/2 baselines. Precision Photometry Robustness. More baselines = more robustness 50.

(51) 2. Co-axial pairwise combination schemes Cophasing N telescopes requires N-1 to N(N-1)/2 baselines. Precision Photometry Robustness. More baselines = more robustness. 51.

(52) 2. Co-axial pairwise combination schemes Cophasing N telescopes requires N-1 to N(N-1)/2 baselines. Precision Photometry Robustness. Blind et al, A&A, 2011. 52.

(53) 3. Measuring the fringe position In practice, 2 different fringe positions: Polychromatic fringes. Group Delay Envelope position. fringes → phase for precision envelope → group delay for SNR. Phase Fringe position. Position measurement 53.

(54) 3a. Measuring the phase: issue Temporal. Static A. D. B. C. C. A. D. B. Time Delayed measurements → disturbances between samples. Simultaneous measurements → each sample is equivalent. Ex: FINITO, CHAMP. Ex: PRIMA, GRAVITY. Fundamental noises (photon, RON, etc.): strictly equivalent concepts !. 54.

(55) 3a. Measuring the phase: be static! Temporal. Static A. D. B. C. C. A. D. B. Time Delayed measurements → disturbances between samples. Simultaneous measurements → each sample is equivalent. Ex: FINITO, CHAMP. Ex: PRIMA, GRAVITY. Static estimator far less sensitive to external disturbances 55.

(56) 3b. Measuring the Group Delay: issue Temporal: broadband fringes. Static: dispersed fringes. λ. Time Delayed measurements → disturbances between samples. Simultaneous measurements → each sample is equivalent. Ex: FINITO, CHAMP. Ex: PRIMA, GRAVITY. 56.

(57) t. 3b. Measuring the Group Delay: be static! Temporal: broadband fringes. Static: dispersed fringes. λ. Time Delayed measurements → disturbances between samples. Simultaneous measurements → each sample is equivalent. Ex: FINITO, CHAMP. Ex: PRIMA, GRAVITY. Static estimator far less sensitive to external disturbances 57.

(58) Result: the POPS concept Blind et al, A&A, 2011. Combination Scheme. Spatial filtering Tatulli, Blind et al A&A, 2010. ag M. lim. it:. Inte grat opti ed cs. m 3 +. Adaptive Optics Tatulli, Blind et al A&A, 2010. Pairwise co-axial multiaxial: Tarmoul et al SPIE, 2010. Detector. 4T3. 6T3B Fringe position K-band dispersed estimators on 5 pixels. Static ABCD. ag. Blind et al A&A, 2011. Design: Kern, Jocou. ‣. ‣. POPS/2GFT ESO documentation VLT-TRE-POP-15461-001. Blind et al, SPIE, 2010. 2GFT status: discussions on-going 58. Control architecture Vincent et al, SPIE, 2010.

(59) Result: expected POPS performance End-to-end simulations of O. Absil. (details in Blind et al, A&A, 2011). 4T concept Limiting magnitude. AT. UT. Fringe tracking. ~7★. ~9. Fringe detection. ~9. ~ 11. Fringe tracking: SNR>4, 95% (dit=10ms). ★. Fringe detection: SNR~4 (dit=25ms). ★ FINITO & ATs: fringe tracking limiting magnitude = 5.5 3T - 2 baselines (open scheme); temporal estimators.. ★ PRIMA & UTs: fringe detected at K=11.7 2T; static estimators.. Limiting magnitude -1mag from excellent to bad conditions 59.

(60) Presentation outline An introduction to interferometry PART I. Study of interacting binaries PART II. Fringe trackers for imaging instruments Conclusions and perspectives. 60.

(61) Summary System analysis. Signal processing Data reduction (not presented). Data analysis & modeling SS Lep. POPS. • Signal processing for static ABCD • Bench validation. •. • Interest of spatial filtering • High benefit of AO • Benefit of static estimators • Optimal combinations. •. Potential of interferometry for interacting binaries Interest of imaging 18/03/2011 4UTs & PIONIER. Blind et al, SPIE, 2010 Tatulli, Blind et al A&A, 2010 Blind et al, A&A, 2011. Blind et al, A&A, accepted 61.

(62) Perspectives Interacting binaries Important potential of interferometry for interacting binaries with model-independent imaging:. • Wind, molecular layers around giant → spectro-interfero • Precise morphology (tidal distortions, etc) → imaging • Detection of accretion disks → visible, longer baselines • Very compact systems → astrometry Future work:. ‣ Deeper studies of SS Lep with H. Boffin ‣ Survey @ VLTI proposed with T. Verhoelst, M. Hillen et al.. Wish list for the VLTI: limiting magnitude → AO !!!, fringe trackers longer baselines 62.

(63) Perspectives Fringe tracking ‣ Rapid & low noise IR detectors:. • Sensitivity gain ~1mag • Multi-axial solution to reconsider: simpler, 6T ~optimal for 4T. ‣ Off-axis fringe tracking (PRIMA, GRAVITY) → blind integration ‣ Predictive algorithms, real-time characterization of atmosphere (?). VLTI in 2025 will really need a good fringe tracker... 63.

(64) 64.

(65) 65.

(66) Laboratory fringes Realistic data reduction for IOC: compensating for the defaults Validation on an interferometric testbench K-band 2T-ABCD IOC. 100. BD phase [deg]. 3. 2. 50. AC. 1. 0. 0 1.9. 2.0. 2.1. 2.2. 1.95. 2.05. 2.10. 2.15. AC vs BD. λ [µm]. wavelength [µm]. 2. 10+10 1.. Calibration. 2.00. A vs C. 0. −1. −2.. 66. B vs D.

(67) Laboratory fringes Realistic data reduction for IOC: compensating for the defaults Validation on an interferometric testbench 1.0. K-band 2T-ABCD IOC. Phase. 4 2. V. 0. 0.5. −2 −4. Visibility. −6 0.0. −4. −2. 0. 2. 4. 6. OPD [µm]. 0. 50. 100. 150. 200. scan index. Group delay. 6. Observables. 10. 4. measure opd [µm]. −6. measured OPD [µm]. measured OPD [µm]. 6. 2 0 −2. 5. 0. −5. −4. Data. −6 −6. −4. −2. 0. OPD [µm]. 67. 2. 4. Simulation. −10 6. −10. −5. 0. input opd [µm]. 5. 10.

(68) SS Lep: Results from Verhoelst et al 2007 VINCI observations: • Single baseline, no phase information • Limited number of spatial frequencies + position angle 55° to 90° • 1-year obs VS binary period ~260d !. M star A star Envelope. Not possible to disentangle the M giant signature from the binary one in visibilities. M giant + A dwarf rotating from one obs to the other. 68.

(69)