Pupil-tracking Mode Research Articles

MODEL&CO: exoplanet detection in angular differential imaging by learning across multiple observations

ABSTRACT Direct imaging of exoplanets is particularly challenging due to the high contrast between the planet and the star luminosities, and their small angular separation. In addition to tailored instrumental facilities implementing adaptive optics and coronagraphy, post-processing methods combining several images recorded in pupil tracking mode are needed to attenuate the nuisances corrupting the signals of interest. Most of these post-processing methods build a model of the nuisances from the target observations themselves, resulting in strongly limited detection sensitivity at short angular separations due to the lack of angular diversity. To address this issue, we propose to build the nuisance model from an archive of multiple observations by leveraging supervised deep learning techniques. The proposed approach casts the detection problem as a reconstruction task and captures the structure of the nuisance from two complementary representations of the data. Unlike methods inspired by reference differential imaging, the proposed model is highly non-linear and does not resort to explicit image-to-image similarity measurements and subtractions. The proposed approach also encompasses statistical modelling of learnable spatial features. The latter is beneficial to improve both the detection sensitivity and the robustness against heterogeneous data. We apply the proposed algorithm to several data sets from the VLT/SPHERE instrument, and demonstrate a superior precision-recall trade-off compared to the PACO algorithm. Interestingly, the gain is especially important when the diversity induced by ADI is the most limited, thus supporting the ability of the proposed approach to learn information across multiple observations.

PACOME: Optimal multi-epoch combination of direct imaging observations for joint exoplanet detection and orbit estimation

Context.Exoplanet detections and characterizations via direct imaging require high contrast and high angular resolution. These requirements are typically pursued by combining (i) cutting-edge instrumental facilities equipped with extreme adaptive optics and coronagraphic systems, (ii) optimized differential imaging to introduce a diversity between the signals of the sought-for objects and that of the star, and (iii) dedicated (post-)processing algorithms to further eliminate the residual stellar leakages.Aims.With respect to the third technique, substantial efforts have been undertaken over this last decade on the design of more efficient post-processing algorithms. The whole data collection and retrieval processes currently allow to detect massive exoplanets at angular separations greater than a few tenths of au. The performance remains upper-bounded at shorter angular separations due to the lack of diversity induced by the processing of each epoch of observations individually. We aim to propose a new algorithm that is able to combine several observations of the same star by accounting for the Keplerian orbital motion across epochs of the sought-for exoplanets in order to constructively co-add their weak signals.Methods.The proposed algorithm,PACOME, integrates an exploration of the plausible orbits of the sought-for objects within an end-to-end statistical detection and estimation formalism. The latter is extended to a multi-epoch combination of the maximum likelihood framework ofPACO, which is a post-processing algorithm of single-epoch observations. From this, we derived a reliable multi-epoch detection criterion, interpretable both in terms of probability of detection and of false alarm. In addition,PACOMEis able to produce a few plausible estimates of the orbital elements of the detected sources and provide their local error bars.Results.We tested the proposed algorithm on several datasets obtained from the VLT/SPHERE instrument with IRDIS and IFS using the pupil tracking mode of the telescope. By resorting to injections of synthetic exoplanets, we show thatPACOMEis able to detect sources remaining undetectable by the most advanced post-processing of each individual epoch. The gain in detection sensitivity scales as high as the square root of the number of epochs. We also appliedPACOMEon a set of observations from the HR 8799 star hosting four known exoplanets, which can be detected by our algorithm with very high signal-to-noise ratios.Conclusions.PACOMEis an algorithm for combining multi-epoch high-contrast observations of a given star. Its sensitivity and the reliability of its astrophysical outputs permits the detection of new candidate companions at a statistically grounded confidence level. In addition, its implementation is efficient, fast, and fully automatized.

Deep PACO: combining statistical models with deep learning for exoplanet detection and characterization in direct imaging at high contrast

ABSTRACT Direct imaging is an active research topic in astronomy for the detection and the characterization of young substellar objects. The very high contrast between the host star and its companions makes the observations particularly challenging. In this context, post-processing methods combining several images recorded with the pupil tracking mode of telescope are needed. In previous works, we have presented a data-driven algorithm, PACO, capturing locally the spatial correlations of the data with a multivariate Gaussian model. PACO delivers better detection sensitivity and confidence than the standard post-processing methods of the field. However, there is room for improvement due to the approximate fidelity of the PACO statistical model to the time evolving observations. In this paper, we propose to combine the statistical model of PACO with supervised deep learning. The data are first pre-processed with the PACO framework to improve the stationarity and the contrast. A convolutional neural network (CNN) is then trained in a supervised fashion to detect the residual signature of synthetic sources. Finally, the trained network delivers a detection map. The photometry of detected sources is estimated by a second CNN. We apply the proposed approach to several data sets from the VLT/SPHERE instrument. Our results show that its detection stage performs significantly better than baseline methods (cADI and PCA), and leads to a contrast improvement up to half a magnitude compared to PACO. The characterization stage of the proposed method performs on average on par with or better than the comparative algorithms (PCA and PACO) for angular separation above 0.5 arcsec.

Half-sibling regression meets exoplanet imaging: PSF modeling and subtraction using a flexible, domain knowledge-driven, causal framework

Context. High-contrast imaging of exoplanets hinges on powerful post-processing methods to denoise the data and separate the signal of a companion from its host star, which is typically orders of magnitude brighter. Aims. Existing post-processing algorithms do not use all prior domain knowledge that is available about the problem. We propose a new method that builds on our understanding of the systematic noise and the causal structure of the data-generating process. Methods. Our algorithm is based on a modified version of half-sibling regression (HSR), a flexible denoising framework that combines ideas from the fields of machine learning and causality. We adapted the method to address the specific requirements of high-contrast exoplanet imaging data obtained in pupil tracking mode. The key idea is to estimate the systematic noise in a pixel by regressing the time series of this pixel onto a set of causally independent, signal-free predictor pixels. We use regularized linear models in this work; however, other (nonlinear) models are also possible. In a second step, we demonstrate how the HSR framework allows us to incorporate observing conditions such as wind speed or air temperature as additional predictors. Results. When we applied our method to four data sets from the VLT/NACO instrument, our algorithm provided a better false-positive fraction than a popular baseline method in the field. Additionally, we found that the HSR-based method provides direct and accurate estimates for the contrast of the exoplanets without the need to insert artificial companions for calibration in the data sets. Finally, we present a first piece of evidence that using the observing conditions as additional predictors can improve the results. Conclusions. Our HSR-based method provides an alternative, flexible, and promising approach to the challenge of modeling and subtracting the stellar PSF and systematic noise in exoplanet imaging data.

Detection of Polarization due to Cloud Bands in the Nearby Luhman 16 Brown Dwarf Binary

Abstract Brown dwarfs exhibit patchy or spatially varying banded cloud structures that are inferred through photometric and spectroscopic variability modeling techniques. However, these methods are insensitive to rotationally invariant structures, such as the bands seen in Jupiter. Here, we present H-band Very Large Telescope/NaCo linear polarization measurements of the nearby Luhman 16 L/T transition binary, which suggest that Luhman 16A exhibits constant longitudinal cloud bands. The instrument was operated in pupil tracking mode, allowing us to unambiguously distinguish between a small astrophysical polarization and the ∼2% instrumental linear polarization. We measure the degree and angle of linear polarization of Luhman 16A and B to be p A = 0.031% ± 0.004% and ψ A = −32° ± 4°, and p B = 0.010% ± 0.004% and , respectively. Using known physical parameters of the system, we demonstrate that an oblate homogeneous atmosphere cannot account for the polarization measured in Luhman 16A, but could be responsible for that of the B component. Through a nonexhaustive search of banded cloud morphologies, we demonstrate a two-banded scenario that can achieve a degree of linear polarization of p = 0.03% and conclude that the measured polarization of the A component must be predominantly due to cloud banding. For Luhman 16B, either oblateness or cloud banding could be the dominant source of the measured polarization. The misaligned polarization angles of the two binary components tentatively suggest spin–orbit misalignment. These measurements provide new evidence for the prevalence of cloud banding in brown dwarfs while at the same time demonstrating a new method—complementary to photometric and spectroscopic variability methods—for characterizing the cloud morphologies of substellar objects without signs of variability.

Regime-switching model detection map for direct exoplanet detection in ADI sequences

Context. Beyond the choice of wavefront control systems or coronographs, advanced data processing methods play a crucial role in disentangling potential planetary signals from bright quasi-static speckles. Among these methods, angular differential imaging (ADI) for data sets obtained in pupil tracking mode (ADI sequences) is one of the foremost research avenues, considering the many observing programs performed with ADI-based techniques and the associated discoveries. Aims. Inspired by the field of econometrics, here we propose a new detection algorithm for ADI sequences, deriving from the regime-switching model first proposed in the 1980s. Methods. The proposed model is very versatile as it allows the use of PSF-subtracted data sets (residual cubes) provided by various ADI-based techniques, separately or together, to provide a single detection map. The temporal structure of the residual cubes is used for the detection as the model is fed with a concatenated series of pixel-wise time sequences. The algorithm provides a detection probability map by considering two possible regimes for concentric annuli, the first one accounting for the residual noise and the second one for the planetary signal in addition to the residual noise. Results. The algorithm performance is tested on data sets from two instruments, VLT/NACO and VLT/SPHERE. The results show an overall better performance in the receiver operating characteristic space when compared with standard signal-to-noise-ratio maps for several state-of-the-art ADI-based post-processing algorithms.

Characterization of the Low-Mass Companion HD 142527 B

Context. The circumstellar disk of the Herbig Fe star HD 142527 is host to several remarkable features including a warped inner disk, a 120 au-wide annular gap, a prominent dust trap and several spiral arms. A low-mass companion, HD 142527 B, was also found orbiting the primary star at ~14 au. Aims. This study aims to better characterize this companion, which could help explain its impact on the peculiar geometry of the disk. Method. We observed the source with VLT/SINFONI in H + K band in pupil-tracking mode. Data were post-processed with several algorithms based on angular differential imaging (ADI). Results. HD 142527 B is conspicuously re-detected in most spectral channels, which enables us to extract the first medium-resolution spectrum of a low-mass companion within 0.″1 from its central star. Fitting our spectrum with both template and synthetic spectra suggests that the companion is a young M2.5 ± 1.0 star with an effective temperature of 3500 ± 100 K, possibly surrounded with a hot (1700 K) circum-secondary environment. Pre-main sequence evolutionary tracks provide a mass estimate of 0.34 ± 0.06 M⊙, independent of the presence of a hot environment. However, the estimated stellar radius and age do depend on that assumption; we find a radius of 1.37 ± 0.05 R⊙ (resp. 1.96 ± 0.10 R⊙) and an age of 1.8_(-0.5)^(+1.2) Myr (resp. 0.75 ± 0.25 Myr) in the case of the presence (resp. absence) of a hot environment contributing in H + K. Our new values for the mass and radius of the companion yield a mass accretion rate of 4.1–5.8×10^(−9) M⊙ yr^(−1) (2–3% that of the primary). Conclusions. We have constrained the physical properties of HD 142527 B, thereby illustrating the potential for SINFONI+ADI to characterize faint close-in companions. The new spectral type makes HD 142527 B a twin of the well-known TW Hya T Tauri star, and the revision of its mass to higher values further supports its role in shaping the disk.

Characterization of low-mass companion HD 142527 B

Context. The circumstellar disk of the Herbig Fe starHD 142527is host to several remarkable features including a warped inner disk, a 120 au-wide annular gap, a prominent dust trap and several spiral arms. A low-mass companion,HD 142527B, was also found orbiting the primary star at ~14 au.Aims. This study aims to better characterize this companion, which could help explain its impact on the peculiar geometry of the disk.Method. We observed the source with VLT/SINFONI inH + Kband in pupil-tracking mode. Data were post-processed with several algorithms based on angular differential imaging (ADI).Results.HD 142527B is conspicuously re-detected in most spectral channels, which enables us to extract the first medium-resolution spectrum of a low-mass companion within 0.″1 from its central star. Fitting our spectrum with both template and synthetic spectra suggests that the companion is a young M2.5 ± 1.0 star with an effective temperature of 3500 ± 100 K, possibly surrounded with a hot (1700 K) circum-secondary environment. Pre-main sequence evolutionary tracks provide a mass estimate of 0.34 ± 0.06M⊙, independent of the presence of a hot environment. However, the estimated stellar radius and age do depend on that assumption; we find a radius of 1.37 ± 0.05R⊙(resp. 1.96 ± 0.10R⊙) and an age of 1.8-0.5+1.2Myr (resp. 0.75 ± 0.25 Myr) in the case of the presence (resp. absence) of a hot environment contributing inH+K. Our new values for the mass and radius of the companion yield a mass accretion rate of 4.1–5.8 × 10−9M⊙yr−1(2–3% that of the primary).Conclusions. We have constrained the physical properties ofHD 142527B, thereby illustrating the potential for SINFONI+ADI to characterize faint close-in companions. The new spectral type makesHD 142527B a twin of the well-known TW Hya T Tauri star, and the revision of its mass to higher values further supports its role in shaping the disk.

Noise-weighted Angular Differential Imaging

Angular differential imaging (ADI) (Marois et al. 2006) is an observational technique in high contrast imaging where the telescope is used in pupil tracking mode so that the image of the sky rotates with respect to the optical surfaces. Bright speckle light caused by optical errors remains fixed in the image, while planets and disks rotate with the sky. The resulting dataset is then post-processed to remove the speckles, de-rotated to undo the sky motion, and median-collapsed to create a final data product. The postprocessing algorithms to remove the speckles are an active area of research and beyond the scope of this note. We consider the derotation and median-combination, where we show gains in signal-to-noise ratio are possible with a small change to the algorithm.

SPHERE/SHINE reveals concentric rings in the debris disk of HIP 73145

The debris disk of HIP73145 has been detected in scattered light in the near-IR, and at far-IR wavelengths before, but no substructure has been seen so far. Detection of such substructures in combination with detailed modeling can hint at the presence of perturbing planetary bodies, or reveal other mechanisms acting to replenish gas and dust reservoirs and forming structures such as spirals or rings. We obtained multiwavelength images with SPHERE in the near-IR in the H2 and H3 bands with the IRDIS camera and a 0.95-1.35 micron spectral cube with the IFS. Data were acquired in pupil-tracking mode, thus allowing for angular differential imaging. The SPHERE standard suite of angular differential imaging algorithms was applied. ALMA Band 6 observations complement the SPHERE data. We detect a bright ring of scattered light plus more structures inside, at least one of them forming a secondary, concentric ring with the first. This is the first detection of this disk in total-intensity scattered light. A second object is detected in the field at high contrast but concluded to be a background star. Forward modeling yields information on the primary parameters of the disk and confirms that the detected substructures are not due to the data analysis approach, which sometimes leads to spurious structures. We detect a series of concentric rings in the disk around HIP73145. This is one of the rare cases where multiple components are necessary to fit the SED and are also detected in scattered light. The presence of such ring structures somewhat questions the nature of the object as a pure debris disk, but the gas and dust content would presumably offer sufficient explanations for such structures to form.

On-sky performance of the QACITS pointing control technique with the Keck/NIRC2 vortex coronagraph

A vortex coronagraph is now available for high contrast observations with the Keck/NIRC2 instrument at L band. Reaching the optimal performance of the coronagraph requires fine control of the wavefront incident on the phase mask. In particular, centering errors can lead to significant stellar light leakage that degrades the contrast performance and prevents the observation of faint planetary companions around the observed stars. It is thus critical to correct for the possible slow drift of the star image from the phase mask center, generally due to mechanical flexures induced by temperature and/or gravity field variation, or to misalignment between the optics that rotate in pupil tracking mode. A control loop based on the QACITS algorithm for the vortex coronagraph has thus been developed and deployed for the Keck/NIRC2 instrument. This algorithm executes the entire observing sequence, including the calibration steps, initial centering of the star on the vortex center and stabilisation during the acquisition of science frames. On-sky data show that the QACITS control loop stabilizes the position of the star image down to 2.4 mas rms at a frequency of about 0.02 Hz. However, the accuracy of the estimator is probably limited by a systematic error due to a misalignment of the Lyot stop with respect to the entrance pupil, estimated to be on the order of 4.5 mas. A method to reduce the amplitude of this bias down to 1 mas is proposed. The QACITS control loop has been successfully implemented and provides a robust method to center and stabilize the star image on the vortex mask. In addition, QACITS ensures a repeatable pointing quality and significantly improves the observing efficiency compared to manual operations. It is now routinely used for vortex coronagraph observations at Keck/NIRC2, providing contrast and angular resolution capabilities suited for exoplanet and disk imaging.

Discovery of concentric broken rings at sub-arcsec separations in the HD 141569A gas-rich, debris disk with VLT/SPHERE

Transition disks correspond to a short stage between the young protoplanetary phase and older debris phase. Along this evolutionary sequence, the gas component disappears leaving room for a dust-dominated environment where already-formed planets signpost their gravitational perturbations. We endeavor to study the very inner region of the well-known and complex debris, but still gas-rich disk, around HD 141569A using the exquisite high-contrast capability of SPHERE at the VLT. Recent near-infrared (IR) images suggest a relatively depleted cavity within ~200 au, while former mid-IR data indicate the presence of dust at separations shorter than ~100 au. We obtained multi-wavelength images in the near-IR in J, H2, H3 and Ks bands with the IRDIS camera and a 0.95-1.35 micrometers spectral data cube with the IFS. Data were acquired in pupil-tracking mode, thus allowing for angular differential imaging. We discovered several new structures inside 1", of which the most prominent is a bright ring with sharp edges (semi-major axis: 0.4") featuring a strong north-south brightness asymmetry. Other faint structures are also detected from 0.4" to 1" in the form of concentric ringlets and at least one spiral arm. Finally, the VISIR data at 8.6 micrometers suggests the presence of an additional dust population closer in. Besides, we do not detect companions more massive than 1-3 mass of Jupiter. The performance of SPHERE allows us to resolve the extended dust component, which was previously detected at thermal and visible wavelengths, into very complex patterns with strong asymmetries ; the nature of these asymmetries remains to be understood. Scenarios involving shepherding by planets or dust-gas interactions will have to be tested against these observations.

Discovery of a low-mass companion to the F7V star HD 984

We report the discovery of a low-mass companion to the nearby (d = 47 pc) F7V star HD 984. The companion is detected 0.19" away from its host star in the L' band with the Apodizing Phase Plate on NaCo/VLT and was recovered by L'-band non-coronagraphic imaging data taken a few days later. We confirm the companion is co-moving with the star with SINFONI integral field spectrograph H+K data. We present the first published data obtained with SINFONI in pupil-tracking mode. HD 984 has been argued to be a kinematic member of the 30 Myr-old Columba group, and its HR diagram position is not altogether inconsistent with being a ZAMS star of this age. By consolidating different age indicators, including isochronal age, coronal X-ray emission, and stellar rotation, we independently estimate a main sequence age of 115$\pm$85 Myr (95% CL) which does not rely on this kinematic association. The mass of directly imaged companions are usually inferred from theoretical evolutionary tracks, which are highly dependent on the age of the star. Based on the age extrema, we demonstrate that with our photometric data alone, the companion's mass is highly uncertain: between 33 and 96 M$_{\rm Jup}$ (0.03-0.09 M$_{\odot}$) using the COND evolutionary models. We compare the companion's SINFONI spectrum with field dwarf spectra to break this degeneracy. Based on the slope and shape of the spectrum in the H-band, we conclude that the companion is an M$6.0\pm0.5$ dwarf. The age of the system is not further constrained by the companion, as M dwarfs are poorly fit on low-mass evolutionary tracks. This discovery emphasizes the importance of obtaining a spectrum to spectral type companions around F-stars.

Searching for companions down to 2 AU fromβPictoris using theL′-band AGPM coronagraph on VLT/NACO

Context. The orbit of the giant planet discovered around beta Pic is slightly inclined with respect to the outer parts of the debris disc, which creates a warp in the inner debris disc. This inclination might be explained by gravitational interactions with other planets. Aims. We aim to search for additional giant planets located at smaller angular separations from the star. Methods. We used the new L'-band AGPM coronagraph on VLT/NACO, which provides an exquisite inner working angle. A long observing sequence was obtained on beta Pic in pupil-tracking mode. To derive sensitivity limits, the collected images were processed using a principal-component analysis technique specifically tailored to angular differential imaging. Results. No additional planet is detected down to an angular separation of 0.2" with a sensitivity better than 5 MJup. Meaningful upper limits (< 10 MJup) are derived down to an angular separation of 0.1", which corresponds to 2 AU at the distance of beta Pic.