Kernel-phase in space

This week, the results of the proposal selection for the Cycle 1 General Observers (GO) program with JWST were announced. In total, some 6000 hours of observing time were awarded to a large number of programs. The details of the GO program can be found on the dedicated STScI webpage.

All of that is great news on its own: the community has been waiting for JWST for a while now and now everybody is really getting ready to do the first observations with this amazing observing facility… but the reason I bring this up here is because, of all of these programs, it turns out that three are directly relevant to what we do in the context of the KERNEL project… and have the word “kernel-phase” in their title!

Two of these programs have been awarded their own time.

The first one is called “Multiplicity Survey of 20 Y Dwarfs with NIRCam Kernel Phase Interferometry“. It is a program led by Loïc Albert (Université de Montreal) that greatly benefited from the paper we published in 2019. As a result, no less than 38.8 hours of observing time were awarded to this program, that will use the kernel-phase technique, combined with the detection algorithm outlined in the paper to look for companions around a sample of 20 very cool sub-stellar objects known as Y-dwarfs. Our team will naturally contribute to the analysis of the data, using the kernel-phase pipeline developed over the course of the KERNEL project.

The second project is called “High resolution, high contrast kernel phase imaging with NIRCam“. It is a smaller program of 4.3 hours of observing time, led by Jens Kammerer, a recent PhD graduate (co-supervised by me) who recently joined STScI. Here, the plan is to target one specific object (HIP 75056 A), known to host a 20-30 Jupiter mass brown dwarf companion to fine tune the kernel-phase analysis procedures.

And an archival proposal… already?!

I was also surprised to discover that, although no data obviously already exists, there is already an archival research proposal called “Kernel-Phase Detection Limits for Planet Discovery with JWST” that was awarded to Samuel Factor, from the University of Texas.

Kernel-phase: a new standard?

To see three different projects directly bear the name “kernel-phase” in their title for the first observing programs by one of the most important observing facilities of the decade to come is a very nice thing to witness! To think that out of the 6000 hours of GO time, more than 40 (not even accounting for the commissioning) aim to take full advantage of kernel-phase is humbling.

I guess after having been a very unusual and marginal observing technique for over a decade, the idea is finally making its way through the brains of observers… who see it as a valid alternative to sparse aperture masking interferometry, particularly onboard a space borne telescope. This means that we have approximately one year to make sure that our pipeline and our detection procedures are razor sharp and ready to be used the moment the data becomes available!

KERNEL-NULLER prototype!

En avril 2020, en plein milieu de la première période de confinement à laquelle nous avons été confrontés, j’avais réalisé une vidéo décrivant le principe de fonctionnement du kernel-nuller. Notre équipe avait lancé le partenariat industriel qui devait nous mener à la livraison imminente d’un prototype…

La crise sanitaire de 2020 a affecté le calendrier de développement de notre prototype: l’étude industrielle pourtant terminée, les usines de fabrication se sont temporairement arrêtées. En pratique les retards se sont accumulés et le prototype, qu’on espérait pouvoir intégrer à notre laboratoire courant 2020 s’est fait attendre.

Mais en ce début d’année 2021, ça y est: notre composant nous a enfin été livré! C’est avec beaucoup de soulagement que je peux enfin poster ici, une première image de ce à quoi ce fameux prototype ressemble, en réalité et comment il se compare à ce qui était spécifié:

Premier prototype de kernel-nuller fabriqué en optique intégrée: l’image de gauche montre le design mis au point par l’équipe KERNEL en collaboration avec l’entreprise Bright Photonics qui a assuré le suivi de la fabrication. L’image de droite est une photographie mosaïque de la surface du composant réel. Le prototype intègre en réalité une demi-douzaine de versions différentes de kernel-nullers dont certaines offrent des versions actives, pouvant être contrôlées par des électrodes (apparaissant en jaune sur le design de gauche). Le composant réel ne mesure que 16 millimètres de côté. Réunir les mêmes fonctionnalités avec des composants optiques classiques (lentilles, miroirs, séparateurs de faisceaux, …) aurait demandé de remplir un volume de plusieurs mètres cube!

Nous avons en réalité fait fabriquer quatre copies de ce composant: deux de ces composants nous ont été livrés nus à la toute fin de l’année 2020 et c’est la photographie de l’un d’eux que vous venez de voir. Les deux autres ont pris un peu plus de temps pour nous être livrés: ils ont été soudés à des borniers électroniques, qui vont nous permettre de contrôler les parties actives du composant en appliquant des tensions. Ce sont ces deux composants “soudés” qui vont faire l’objet du travail d’intégration dans le cadre du projet KERNEL.

En plus des guides d’onde gravés dans le substrat et qui sont représentés par les tracés en rouge sur le schéma de principe, et dans lesquels on va venir coupler la lumière en provenance de plusieurs télescopes, certaines régions du composant peuvent être chauffées par des petites résistances déposées sur la surface du substrat. La tension électrique nécessaire au passage du courant dans les résistances est appliquée via les électrodes représentées par les pistes en jaune sur le même schéma de principe. Le composant étant tellement petit et délicat, nous avons du sous-traiter la soudure et la conception de cartes d’interface.

Prototype de kernel-nuller, connecté à deux borniers électroniques. En appliquant des tensions électriques deux bornes, il est possible de venir chauffer localement le substrat constituant le composant. La variation de l’indice de réfraction qui en résulte peut être utilisée pour moduler la phase du champ électrique de la lumière présente dans les guides d’onde!

En appliquant des tensions aux bornes de ces connections électriques, on peut chauffer localement le substrat constituant le composant (verre SiN, optimisé pour une utilisation dans la bande H) et moduler la phase du champ électrique de la lumière se propageant dans les guides d’onde (globalement de la gauche vers la droite). Cette propriété du kernel-nuller offre un nombre important de possibilités: un élément de chauffage à l’entrée peut se substituer à une ligne à retard optique avec un temps de réponse de l’ordre de la milli-seconde . Un élément de chauffage situé entre les fonctions réalisant les interférences peut permettre de compenser des imperfections du design ou des dérives optiques induites par des fluctuations de l’environnement thermique.

Cette technologie thermo-active, fréquemment mise en œuvre dans le cadre des télécommunications n’a jamais été exploitée dans un contexte astro: si on arrive à le mener sur le ciel, le kernel-nuller pourrait être le premier exemple d’une famille de calculateurs électro-optique à lumière stellaire 😉

Maintenant, nous avons du pain sur la planche: notre banc de test KERNEL est en train d’être upgradé pour permettre l’intégration du composant avec le reste des éléments (contrôle actif du front d’onde et moyens de caractérisation). Si les avantages théoriques de l’architecture kernel-nuller se confirment, il nous faudra envisager une démonstration technique sur le ciel! Ce travail est principalement mené par Nick Cvetojevic (employé du projet ERC KERNEL) et Peter Chingaipe (doctorant UCA bénéficiant d’une bourse 80-PRIME du CNRS).

KERNEL publication: Kernel-nullers for all interferometers!

In this new study, led by PhD candidate Romain Laugier, our team has looked into what one needs to do in order to build a kernel-nuller, that is a high-contrast interferometric recombiner robust to residual piston excursions, for an array featuring an arbitrary number of telescopes.

The paper takes advantage of nice graphical representations of the inner workings of a nuller called “complex matrix plots” (but that we’ve nicknamed “Laugier-grams”) to find the properties a recombiner must exhibit in order to be robust to piston perturbations.

The same graphs were also featured a few months ago in a video recorded in French (first announced in this blog post) and later translated into English that explains why the 4-beam kernel-nuller of Martinache & Ireland (2018) works. Both videos were posted on Youtube:

Congratulations to Romain for another published paper to append to his upcoming PhD dissertation! The preprint version of the paper is available for download on the arXiv.org website: https://arxiv.org/abs/2008.07920. Romain’s PhD defense will take place in a little over a month on September 22, 2020.

Postdoctoral position in High Angular Resolution Astronomy

The KERNEL project, hosted by Observatoire de la Cote d’Azur (OCA) invites applications for a postdoctoral research position in the field of high-angular resolution astronomy starting no later than July 1, 2020. This position is funded by the European Research Council (ERC – CoG – grand agreement #683029) under the European Union’s Horizon 2020 research and innovation program.

The KERNEL project

KERNEL aims at enabling every optical and infrared astronomical facility to reach its ultimate angular resolution potential, often pushing beyond the formal diffraction limit, while preserving the full sensitivity. By looking at astronomical data as the result of an interferometric process, the KERNEL framework brings much needed robustness to high-performance observing techniques, required for instance for the direct detection of extrasolar planets.

The mission

The KERNEL framework offers a wide range applications that go from the post-processing of available archival data to high-performance focal plane metrology, partly coupled with high-contrast imaging. KERNEL now benefits from advanced data reduction tools developed over the course of the project whose effectiveness has been demonstrated by results featured in multiple peer reviewed publications.

In addition to coronagraphic imaging, observing campaigns carried out over the last five years by exoplanet hunting high-contrast imaging instruments have also made extensive use of non-coronagraphic observing modes. The KERNEL project is therefore looking for a postdoc candidate with experience in the processing of data produced by such high-contrast imaging instruments, for further processing exploit the KERNEL software tools. Anticipated applications include the KERNEL processing of sparse aperture masking (SAM) or non-redundant masking (NRM) interferometry data as well as non-saturated well-sampled full-aperture imaging data. In both cases, careful processing using the KERNEL tools will make it possible to probe for the presence of companions and asymmetric structures at very small angular separations.

How to apply

A Ph.D. in astronomy, physics, or a closely related field is mandatory. We are interested in individuals with several years of post-PhD research experience in the exploitation of high angular resolution astronomy instrumentation. The candidate would be joining an international team and be willing to collaborate with graduate students and other postdocs. The candidate will also be encouraged to find ways to apply the tools of the KERNEL project to pursue his/her personal research interests.

The candidate must possess a strong background in the modeling, reduction and interpretation of diffraction dominated data. Experience with the Python and/or the C programming language is highly desirable.

The initial appointment will be for one year. The successful candidate will be hosted by the Lagrange Laboratory, with a lab located on the campus of Valrose, downtown the beautiful city of Nice, France.

To apply, please send a copy of your curriculum vitae, list of publications and a summary of your research interests. Also arrange for two to three reference letters to be sent to Frantz Martinache (frantz.martinache@oca.eu). For full consideration, applications should be received before May 4, 2020, although applications will be reviewed up until the position is filled.

This job offer has been posted on the Euraxess website.

Two Kernel-phase A&A papers out this month!

The April 2020 issue of Astronomy & Astrophysics will feature two papers from the Nice KERNEL team!

Paper #1: Angular Differential Kernel-phase

The first paper features the results of a study led by graduate student Romain Laugier who’s successfully adapted an angular differential observing technique commonly used in high-contrast imaging to the kernel-phase scenario. This approach, coined angular differential kernel-phase (ADK) takes advantage of the sky rotation experienced by the SCExAO instrument at the Nasmyth focus of the Subaru Telescope when the field rotator is turned off. The technique makes it possible to better calibrate the biasing effect introduced by AO-residuals in the presence of quasi-static aberrations. Whereas interferometric observations typically require to alternate between a target of interest and a calibration star, this new approach spends 100% of the observing time on the target of interest, making it a more efficient alternative.

Figure extracted from the Laugier et al (2020) publication introducing the angular differential kernel-phase observing mode.

The publication is available in open access on the Astronomy and Astrophysics website!

Paper #2: Kernel-phase… version 2.0?

The second paper features the result of a study led by KERNEL project PI Frantz Martinache. This paper goes back to the roots of kernel-phase. After several years of development of the XARA pipeline carried out in the context of the KERNEL project, it was time to revisit previous analysis results in the light of its latest developments. The paper shows that while overall successful, early uses of kernel-phase were not particularly careful. The paper shows that refined descriptions of the diffractive apertures by instruments leads to a major improvement of the kernel-phase analysis and reduces the importance of systematic errors.

Illustraction extracted from the Martinache et al (2020) publication, showing from top to bottom, how a better model of the diffractive aperture can reduce the amount of systematic error. By either increasing the density of the aperture model (middle row) or by introducing a transmission model (bottom row), the magnitude calibration signal (the red or the orange curves on the right hand side plots) can be considerably reduced in comparison with the astrophysical signal (the blue curve).

Using these new aperture modeling prescriptions, the authors then reprocess previously published observations from ground-based and space-borne observatories and shows major improvements in both cases!

In the same vein as the ADK idea at the core of the Laugier et al (2020) publication, the paper quickly explores the possibility offered by consecutive observations at multiple wavelengths. For a target whose aspect would change depending on the wavelength, spectral differential kernel-phase (SDK?) would be a powerful observing mode that would spend

The publication is of course also available in open-access on the Astronomy & Astrophysics website!

KERNEL-Nuller: vidéo explicative

Il y maintenant bientôt deux ans, j’annonçais sur ce site l’acceptation d’un article publié avec mon collègue Mike Ireland présentant un mode d’observation interférométrique haut contraste robuste aux petites erreurs de correction par un suiveur de franges: le kernel-nuller.

Notre équipe à Nice est, en collaborant avec l’entreprise Bright Photonics en train de faire fabriquer un premier prototype de kernel-nuller, sous forme de composant d’optique intégrée. La situation sanitaire du COVID-19 complique un peu le calendrier de cette activité et de la thèse de doctorat qui y est liée, mais nous devrions pouvoir annoncer cette année, des résultats partiels d’une première intégration d’un tel composant sur un banc optique.

En attendant de voir ce composant en action, voici une vidéo mise en ligne il y a quelques jours, expliquant ce qui distingue le kernel-nuller du nuller interférométrique initialement imaginé par Ronald Bracewell à la fin des années 1970… et illustre comment le concept fonctionne!

Interview de Romain Laugier, doctorant du projet KERNEL par le magazine CIEL & Espace

Photo de Romain Laugier

Le magazine d’astronomie populaire Ciel & Espace propose sous forme de podcast, des interviews d’étudiants en cours de thèse de doctorat. Un des épisodes de ce podcast intituté “La Science en Chemin” propose une interview de Romain Laugier, doctorant du projet KERNEL qui parle de son travail et de son parcours. Cet épisode est disponible à l’adresse suivante: https://www.cieletespace.fr/actualites/podcast-la-science-en-chemin-avec-romain-laugier

KERNEL Publication: Statistical Tests for high-contrast detection using KERNEL-PHASE!

In this new publication, KERNEL team PhD student Alban Ceau and collaborators formulate new formal tools and techniques to determine the contrast detection limits for a planned JWST program designed to look for companions around ultracool Y-dwarfs.

The paper notably introduces an operational test that turns out to be only marginally less powerful than the theoretical optimal (aka Neyman-Pearson) test. The operational test, applied to simulated JWST images of the NIRISS instrument expected to be representative of planned observations of faint Y-dwarfs suggests that companions of less than 1 Jupiter mass are well within the grasp of this observing mode.

Parameter estimation by the kernel-phase method for 10:1 contrast companions located inside or outside lambda/D (~ 150 mas at the wavelength of the simulation). For super-resolved detections, the contrast ratio and the angular separation are strongly correlated: a behavior that is similar to what has been observed for closure-phase when using a non-redundant aperture mask. Beyond one resolution element, the interpretation of the data is unequivocal.

Congratulations to Alban whose paper was accepted for publication by Astronomy & Astrophysics! The preprint version of the paper is available for download on the arXiv.org website: https://arxiv.org/abs/1908.03130

Kernel-phase imaging VLT/NACO paper!

In this new publication using the tools developed in the context of the KERNEL project, ANU-based PhD student Jens Kammerer announces the detection of eight low-mass companions to stars observed using VLT/NACO in the L-band, five of which were previously unknown,

Among these new companions, two appear at angular separations ranging between 0.8 and 1.2 λ/D (i.e. 80 and 110mas⁠), demonstrating that from the ground, kernel-phase makes it possible to achieve a resolution better than the traditionally accepted diffraction limit of a telescope.

Corner-plot of the best-fitting parameters for the two close companion candidates HIP50156 (top) and HIP39718 (bottom) recovered by the kernel-analysis of archival VLT/NACO L-band images.

Congratulations to Jens whose paper was accepted for publication by the Monthly Notices of the Royal Astronomical Society (MNRAS) and should appear in the June 2019 edition of the journal! The preprint version of the paper is already available for download on the arXiv.org website: https://arxiv.org/abs/1903.11252

My Innovation Is… on Youtube?

Back in November last year, the concept of kernel-nuller was featured as one of the ten projects selected by the SATT Sud-Est in the context of the innovation challenge called “My Innovation Is…”. Yesterday, the following video was posted on youtube… enjoy!

The following text was extracted from the description of the video (available here):

SATT Sud-Est and its partners present the Superheroes of Innovation, winners of the My innovation is… 2018 competition. During the third edition of the “My Innovation Is…” competition, held on November 27, 2018 at the Palais des Papes in Avignon, France, 10 researchers and future startuppers from the public laboratories of the South & Corsica Regions presented their innovative projects before a jury of experts.

Ignacio CASUSO, Inserm, is able to visualize the nervous system at the molecular level. Sébastien GIRAUD (Aix-Marseille Université) is able to evaluate proprioception to better communicate with our body. Olivier CHUZEL of Aix-Marseille Université has the power to selectively kill cancer cells. Soraya MEZOUAR (Aix-Marseille University) is developing a new therapy using placental stem cells. Eric LECHEVALLIER (AP-HM) can describe and quantify the intensity of a bleed using a connected tool. Manuel ESPINOSA of the University of Corsica controls the storage of solar energy. Christine CONTINO-PEPIN from Avignon Université is able to extract and stabilize compounds of plant origin in an eco-compatible way. Michel Alain BARTOLI (AP-HM) is developing a stent graft dedicated to the endovascular treatment of aortic diseases. Mikaël CHELLI of the Centre Hospitalier Universitaire de Nice masters medical statistics thanks to an autonomous application. Frantz MARTINACHE (Observatoire de la Côte d’Azur) reveals the presence of extrasolar planets in search of habitable worlds. Faced with a merciless jury, they had to fight to convince and ensure the next generation of innovation.

It’s not every day you are called a “super-hero” of innovation 😉