Tribute to François Roddier

Published on October 19, 2023

A founding father of high angular resolution astronomy

At the end of summer 2023, we were saddened to learn of the passing of François Roddier (23/09/1936 - 19/08/2023), one of the founding fathers of what has become an emblematic field of expertise for France, and one which still justifies the existence of our ASHRA organization (action spécifique pour la haute résolution angulaire) today.

Whatever our favorite flavour of high angular resolution astronomy may be - whether atmospheric optics, adaptive optics, optical interferometry, high contrast imaging or signal and data processing techniques - we find, at the origin of many of these islands of technological and methodological wealth, references to founding publications by François.

To help us appreciate the richness, diversity, depth and scope of François Roddier's work, the ASHRA Directorate invited four of his collaborators and/or close acquaintances to give us a personal account of a key moment in his career. We are very grateful for the contributions offered by Claude Aime, Gérard Rousset, Olivier Guyon and Olivier Lai, who have agreed to share with us, episodes and anecdotes covering about François Roddier's early days in the astrophysics department at the University of Nice; his pioneering work on adaptive optics on the large optical telescopes at MaunaKea; the great generalizations and bridges he built between interferometry and high contrast imaging; and his fascinating post-retirement deep reflections, on the thermodynamics of evolution and its consequences for the emergence of life and its impact on human economy. It's enough to make you dizzy to think of all these contributions, many of which are the at the origin of the "raison d'être" of our ASHRA organization.

With such a track record, François Roddier has contributed to the training of many young researchers, which naturally resulted in a large number of collaborations. A search of his name in ADS reveals an impressive bibliography of 275 articles to which he is associated. In collaboration with our four contributors, we also offer, at the bottom of this page, a selection of writings considered seminal, of which François is the first or sole author.

Frantz Martinache & Elodie Choquet, on behalf of the ASHRA Scientific Council

A few important dates for the life and career of François Roddier

  • 1936: Born in Paris on September 23
  • 1960: ENS graduate
  • 1964: Doctoral thesis
  • 1965: Professor at the University of Nice
  • 1984: Secondment to NOAO in Tucson (Arizona, USA)
  • 1989: Secondment to the Institute of Astrophysics at the University of Hawaii (USA)
  • 2001: Retirement and return to France
  • 2023: Death on August 19 in Hyères

François Roddier, professeur at the University of Nice and pioneer of high angular resolution astronomy

Contribution offered by Claude Aime

François Roddier, a graduate of the Ecole Normale Supérieure and an associate professor of physics, wrote his doctoral thesis in Jacques Blamont's laboratory on the development of an atomic jet spectrograph and its application to solar observation. He defended his thesis in 1964 and became a professor at the University of Nice the following year. This marked the creation of the Faculty of Sciences, with mathematician Jean Dieudonné as Dean, and the rebirth of the Nice Observatory under the direction of Jean-Claude Pecker.

François Roddier created the Department of Astrophysics at Valrose and surrounded himself with young students and researchers, soon joined by assistant masters from another laboratory. They all became his students, recognizing him as an undisputed research director, both demanding and benevolent. He set out the main lines of research, supervised everyone very closely, found solutions to stumbling blocks and made very few demands in terms of participation in publications, which was not at all the practice of laboratory directors at the time.

His initial research theme, developed by his first PhD student Éric Fossat, is in continuity with his early solar observations at very high spectral resolution, by replacing atomic jet systems with a sodium resonance cell. Using the Zeeman effect, the two wings of the sodium D1 line are analyzed in time to deduce velocities using the Doppler effect. The system was highly stable, enabling very precise measurements of solar oscillations. To obtain observations over long periods, François Roddier led the transformation of the Nice Observatory's bent equatorial telescope into a solar telescope. Eric Fossat and his team then went on to make very long-duration observations in Dome C at the South Pole, marking the start of the development of helioseismology.

In parallel with this research, François Roddier became interested in the work of Antoine Labeyrie, who in 1970 had invented the speckle interferometry technique. By calculating the spatial spectral density of images, it became possible to obtain the resolution limit of a telescope, despite atmospheric turbulence. Like many opticians around the world, François Roddier set out to study interferometric techniques and the propagation of light through the Earth's atmosphere, with the mathematical formalism necessary for a rigorous understanding of the phenomena. It was he who demonstrated the similarities between speckle masking (triple correlation) and closure phase.

The first work on atmospheric turbulence was carried out with Jean Vernin, on the multidimensional analysis of stellar scintillation. Numerous studies followed, opening up new avenues for the research and qualification of astronomical sites. To meet the needs of intensive statistical calculations, François Roddier had the laboratory develop an electronic correlator capable of very fast calculations, competing with what was then the state of the art, by Princeton Applied Research, now light years away from today's fast GPU calculations.

François Roddier then directed my thesis work on the high-angular-resolution study of solar granulation, and together we developed various statistical studies, from the spatial distribution of granules to the study of differential speckle interferometry. These techniques were eventually superseded by the progress on adaptive optics systems, which are producing remarkable results, including for solar observations, and of course by space observations.

In the same field of high angular resolution, Claude and François Roddier developed a very delicate rotation shearing interferometer, for which they described the theory and carried out astrophysical applications, notably for the high-angular-resolution observation of Betelgeuse.

François Roddier, who was totally committed to research, was equally committed to teaching. When the university was founded, he was in charge of the course on thermodynamics, a subject he would eventually return to later in his career. He created what was then known as a DEA (a post-graduate diploma) in astrophysics, which was to change its name as the various degrees were awarded. François was particularly involved in mathematics courses for the physics master's degree, which led to the publication of his remarkable book "Distributions and Fourier transform". In the introduction of this book published in 1971, he humbly presents himself as an experimental physicist, supplementing this overly mathematical title with a modest "for the use of physicists and engineers". In it, he shows how, thanks to distribution theory, convolution and the Fourier transform are becoming indispensable tools in experimental disciplines such as electronics and optics. His little "book with a green-cover" was to become a reference for several generations of students and researchers. The presentation of the distributions is uncompromisingly mathematical, and is complemented by numerous exercises, some of which are directly applicable to research experiments. The solutions he gives are very (too?) concise, just like François Roddier's explanations during laboratory meetings, during which he would often develop mathematical ideas, to later be surprised by the fact that people often found it hard to follow him!

François also gave very rigorous courses on Fourier optics. He was the driving force behind practical lab courses in optics that have helped students of several generations to gain practical experience with ideas like Fresnel diffraction and optical filtering, and to experimentally realize what a two-dimensional Fourier transform is. At DEA level, he lectured on the propagation of light through the earth's atmosphere, the extension of which led to his now often cited 1981 publication in Progress in Optics.

François Roddier and the beginnings of Adaptive Optics

Contribution offered by Gérard Rousset

It was during his stay at NOAO in Tucson, Arizona, between 1984 and 1989, that François Roddier became involved in an American research program for the deployment of adaptive optics (AO) for astronomy. Building on his work on the properties of atmospheric turbulence in astronomy (Roddier 1981), he carried out an initial theoretical study analyzing intrinsic errors in AO linked to atmospheric structure and chromatic effects. He also sought to design an AO system that would be as efficient as possible for a given number of degrees of freedom in the system.

He then turned his attention to the measurement of wavefront curvature, which exhibits strong spatial decorrelation in contrast to the slopes measured by a Shack-Hartmann wavefront sensor. He realized that the curvature could also be directly the command to be applied to a deformable mirror (bimorph, for example) whose mechanical deformation is governed by a Poisson equation. François thus came up with a system that had a near perfect correspondence between measurements and commands, leading to a quasi-diagonal "command matrix".

As for the curvature sensor itself, he simply proposed to measure in a defocused plane (on either side of the focus) the over- or under-intensities induced by wavefront curvatures present in the pupil. This is the concept François presented at a LEST Foundation workshop in 1987, and published in 1988. Like the rest of his work, these ideas were particularly elegant, drawing on his perfect mastery of physics.

From then on, he set about developing the technologies needed to make this new AO system operational on a telescope. He did this at the University of Hawaii in the 1990s, where he surrounded himself with a team of engineers and young researchers.

To make the best possible use of the photons from the observed source, he was inventive. He proposed the use of photon-counting avalanche photodiodes (APDs) with good quantum efficiency as detectors. These had just become commercially available at the time. Their performance is far superior to the intensified CCDs of the time commonly used for the Shack-Hartmann wavefront sensor. A single photodiode is used per sub-aperture. To couple the photons to the photodiodes, the pupil is sampled with a circular microlens array that injects the light into optical fibers, transporting it to the photodiodes.

In choosing the bimorph deformable mirror, he emphasized its advantage of having a large stroke for low spatial frequencies, perfectly matching the observed properties of the turbulence to be compensated. He is also optimized the layout of electrodes to maximize the efficiency of the correction. He initiated a collaboration with Laserdot (now CILAS) in France to manufacture relatively low-cost mirrors, while a member of his team also developed in-house solutions.

A series of curvature AO systems have been built by François' team during this period, starting with a system featuring 13 degrees of freedom (13 APDs and 13 electrodes on the bimorph), then 36 and eventually 85 on the well known Hokupa'a system. Each of these systems is a step on François's quest for optimum performance at the lowest possible cost. The first telescope tests took place in 1991, but the first operational system was tested at the CFHT in 1994. Regular observations followed, either at the CFHT or at the UH 88'' telescope. This was followed by a wide range of observations: solar system objects (Io, Saturn's rings, Neptune…), young stars such as the GG Tau binary and its disk, or star clusters (Trapezium…), whose images revealed highly interesting details.

Curvature AO systems have enjoyed remarkable success in the astronomical community, with the following major developments:

  • the 19-electrode PUEO system installed at CFHT in 1996
  • the SUBARU telescope with a 188-electrode mirror (2006)
  • at the VLT (ESO), the 6 MACAO 60-electrode systems (2003) equipping the 4 UTs of the interferometer and two other instruments.

In addition, François and his wife Claude proposed to use the principles of the curvature wavefront sensor to test the optical quality of telescopes using defocused images, which gave them the opportunity to contribute to the identification of HST spherical aberration.

François Roddier's contribution to adaptive optics was original and highly relevant. To minimize the cost of his systems, he came up with a design strategy to maximize their efficiency, ideally matched to the known properties of turbulence. The curvature sensor he invented was the answer. He was a strong advocate of the widespread use of natural star AO systems for IR astronomy, given that at the time, laser star systems were still not very efficient. Before retiring, he collected all his ideas and published a book on AO with several contributors: Adaptive Optics in Astronomy (1999).

The enf of his career in the USA

Contribution offered by Olivier Guyon

I was François' last doctoral student, from early 1998 until François' retirement from the University of Hawaii's Institute of Astronomy in 2001.

François was one of the first to realize that adaptive optics was essential to enable efficient interferometric recombination between large telescopes. He thus encouraged the deployment of AO on the CHARA interferometer and, together with the pioneers of optical interferometry, advocated that the large AO-equipped telescopes atop MaunaKea be interferometrically linked by single-mode fibers, which became the OHANA project.

François had a deep understanding of the fundamentals of astronomical optics, which enabled him to identify elegant and simple solutions in interferometry, adaptive optics and coronagraphy. He simplified problems that most of us considered too complex to tackle, and conceptualized practical solutions that combined simplicity and optimality. This made François a great mentor, who taught me to approach new problems by starting with the essential physical principles.

The curvature-based AO systems he developed were unrivaled in terms of performance and sensitivity, making optimal use of a small number of photon-counting detectors to monitor almost as many wavefront modes. By combining curvature-based wavefront detection with curvature-deformable mirrors, the systems he built could operate at high speed with the hardware available at the time. François showed that astronomical adaptive optics could provide sharp images of faint targets, considerably extending its reach before laser-guided stellar systems matured.

François came up with the idea of unifying coronagraphy and nulling interferometry by replacing the opaque focal plane mask of a Lyot coronagraph with a partially phase-shifting mask. This circular mask, smaller than the characteristic size of diffraction, induces destructive interference between starlight passing through the dephasing area and that avoiding it, resulting in an image with an enhanced dynamic image. This new approach, now universally refered to as "the Roddier coronagraph", was the first of what is now a very rich family of coronagraphs. Such masks are now used on large telescopes on the ground and in space, enabling highly efficient direct imaging of exoplanets in angular proximity to their host star.

Works and reflections on the thermodynamics of evolution and economics

Contribution offered by Olivier Lai

I was lucky enough to work with François and Claude during my post-doc at Keck when we tried to implement their old pupil rotation interferometer (which they used in the 80s at CFHT visually) on the wavefront sensor arm of the KeckAO system (but there was too much vibration on the sky). We became friends and kept in touch after he left Hawaii and returned to France.

After leaving the field of optics and astronomy in 2001, François retired with Claude to their family home in Carqueiranne, Var. Being above all a passionate scientist, he continued to ask questions about how the world worked, and perhaps even more so with the time freed up by his retirement. In particular, he was interested in societal issues, but being a physicist at heart, he approached them from a thermodynamic point of view.

Between 2005 and 2019, he maintained a blog entitled "Point de vue d'un astronome", which many of us followed assiduously; there were elements of open systems thermodynamics, evolutionary biology, enriching comments on economics, on the study of self-organized structures or on societal collapses such as Jared Diamond's book "How societies choose to fail or succeed" (his first post, dated October 9, 2005), and we witnessed the blossoming of a profound idea: any structure in the universe appears only to maximize the overall entropy of the system, and dissipate the energy available to do work (Gibbs energy).

Indeed, the second law of thermodynamics (S>0, entropy can only increase) only applies to closed systems. But we are witnessing self-organizing structures (with "negative entropy") in our universe (galaxies, stars, life), seemingly at odds with the second law. The solution to this apparent paradox is to consider that in open systems, entropy can decrease locally, by importing information from the environment to increase the overall dissipation of Gibbs free energy and maximize the rate of entropy increase. The theoretical foundations of this "third law" of thermodynamics are only recent, and are due to a Scottish biologist working in Bordeaux, Roderick Dewar: "Information theory explanation of the fluctuation theorem, maximum entropy production and self-organized criticality in non-equilibrium stationary states", Journal of Physics A, 36, 631 (2003). While reading François' blog, I kept thinking of the famous physicist John Archibald Wheeler's phrase: "To my mind there must be, at the bottom of it all, not an equation, but an utterly simple idea. And to me that idea, when we finally discover it, will be so compelling, so inevitable, that we will say to one another, 'Oh, how beautiful. How could it have been otherwise?'"

In his usual didactic way, François showed how this principle of importing information from the environment in order to better dissipate available energy can be applied to physical, chemical, biological and even societal systems. In this paradigm, we find many concepts familiar from complex systems, such as bifurcations, micro- and macro-evolution, the theory of punctuated equilibria, natural selection processes with their feedback (the red queen effect) and Zipf's Law. With a level of abstraction befitting a visionary scientist, he described how life could have arisen under the influence of these thermodynamic principles, explaining how autocatalytic structures (enzymes) could have developed during convective cycles around the critical point of water, leading to the formation of RNA and finally DNA, under conditions found in underwater geothermal vents. With this same level of abstraction, he took up and extended Richard Dawkins' ideas on memetic evolution, which can transmit far more information than genes, and adapt far more rapidly to better dissipate available energy and resources.From this point on, it was natural to perceive that these same laws also apply to cultural evolution. Seeing the history of the world through the prism of thermodynamics is striking, because we find the same cycles, but above all we understand why these patterns repeat, and that there is an underlying principle that organizes structures, from the scale of molecules to organisms to societies.

François first presented his ideas to the scientific community at a conference at IAP on October 5, 2010, and went on to write three books, "Bread, Leaven, Genes" (2007), "The Thermodynamics of Evolution" (2012) and "From Thermodynamics to Economics" (2018), laying the foundations for these new ideas. Many scientists with whom I have spoken have understood the significance of this work as applied to physical systems (laminarity to turbulence transition, hurricane development, etc.) and biological systems (evolution, natural selection and adaptation to the environment). The thermodynamic principle behind these ideas is so convincing in such a wide range of subjects that it seemed natural to François to extend them to culture and economics, which some have criticized for going beyond the scope of validity of the assumptions of such a simple principle. However, in his last book on the application of thermodynamics to economics, he uses the Van der Waals gas equation to convincingly and accurately describe the depression -' inflation -' stagflation -' collapse cycle that is evident in every culture and civilization but is not captured by classical economics. François also clearly understood the implications of his work, and appreciated the difficult situation facing humanity today.

These works show just that François had a scientific mind, in the noblest sense of the word: he sought out the most elegant solutions, even if they went against the grain, and never hesitated to be truly creative in finding solutions to problems that aroused his curiosity. It was a real honor for me to share his homemade sourdough bread in his company, knowing that it was both the culmination of 10,000 years of human evolution and the inspiration for his ideas on the driving principles of that same evolution.

Selected publications