My work has also been my hobby
As told by Jørgen Christensen-Dalsgaard

2022 KAVLI PRIZE IN ASTROPHYSICS

MORE ABOUT JØRGEN CHRISTENSEN-DALSGAARD

Over the past more than four decades, my efforts have focused on understanding the structure and evolution of stars, and the way that this understanding affects other areas of astrophysics. I have played a substantial role in the development of helio- and asteroseismology, the studies of solar and stellar properties based on their oscillations. These procedures are now central to investigations of stellar physics.

This has included work on the development of the basic theory of stellar oscillations and their diagnostic potential, techniques for modelling accurately stars and their oscillations, and development and application of techniques for the analysis of seismic data. In addition, I have contributed to the establishment of facilities for obtaining such data, and the organization of a community for the planning and use of space-based seismic observations.

It has been my amazingly good fortune to be part of helioseismology from the inception of the field at the start of my career, and near its end to have contributed to the revolution in asteroseismology resulting from space observations. I have had great satisfaction in following these fields over time, with stellar astrophysics now being a thriving and vigorous area of research, with major impact on other fields.

Early life
I was born in 1950 and grew up in Kolding, a provincial Danish town 100 km north of the border to Germany. My parents were teachers at, and my father became headmaster of, a family-owned private school funded by my grandfather’s oldest brother in 1890. Despite being an only child, it was quite early made clear that there was no expectation that I would follow in the family footsteps. My interest in the physical sciences started early, also inspired by two close friends from my father’s university days, one of them professor of the history of science, the second professor of physics. My activities included circuit making with the then relatively new transistors and basement chemistry, making use of the rather liberal access to buying chemicals at the time.

My interest in astronomy made me build a 15 cm Newtonian telescope, grinding and polishing the mirror (with several attempts to suppress an astigmatism), silver-coating the mirror, and constructing a simple but efficient equatorial mount. I also constructed a simple astrograph, taking images of the sky as a basis for stellar mapping, and carried out visual observations of variable stars. The focus on astronomy and astrophysics for my further studies grew during high school, with the added realization that this offered a combination of physics and mathematical analysis with observations. This combination of observations and theory has been a major factor and pleasure in my whole career.

“My interest in astronomy made me build a 15 cm Newtonian telescope.”

My childhood and youth coincided with the early parts of space exploration, culminating with the moon landings from 1969. I remember, at the age of 7, seeing the first Sputnik, and I followed the evolution of the US and Soviet space programme in the news, including live transmissions of launches and an early space walk.

I entered Aarhus University in 1969 for a first degree in mathematics and physics. An important event was the introduction to computing, using Algol on the Danish-built GIER computer, where programmes were coded on punched tapes. My coding included a programme to analyse and visualize some of my variable-star observations and a simple code aiming at, and failing, to launch a rocket to the moon. This was followed by an introduction to Fortran, with calculations of planetary motion and stellar atmospheres, developing subroutines a few of which I still use.

When I started MSc studies in astronomy in Aarhus I found stellar astrophysics particularly appealing as a possibility of applying physics to astrophysical objects that were relatively well observed. This included the realization of the importance of hydrodynamics in astrophysics, brought home by courses in stellar structure and evolution and stellar atmospheres, which demonstrated the pitiful state of the modelling of stellar convection. I felt that one should be able to do better than this, which motivated plans to take up convection modelling as a research topic.

Cambridge and helioseismology
Astrophysics at Aarhus University was predominantly observational at the time, so I began considering options for a Ph.D. abroad. A friend of my parents was studying English in Cambridge; during the autumn break in 1972 we visited him, and I used the opportunity to look into the possibilities for Ph.D. studies. I managed to get an interview with the Head of the Department of Applied Mathematics and Theoretical Physics (DAMTP) to enquire about this, but, understandably, he was not impressed by this Danish student walking in from the street and could only advice the official route of applications. At a dinner at Darwin College, with which our friend was associated, the evening before our return I happened to sit next to a Ph.D. student of astrophysics and told him about my failed attempt. He adviced me to contact a young assistant professor, Douglas Gough, at the Institute of Astronomy, and we went there the following day on the way to the ferry in Harwich. When I knocked on Douglas’s door he invited me in, and we had an hour-long chat about my ideas and his work, at the end of which he agreed to help me with the application procedures, being listed as a supervisor if the application was successful. In my application I selected Darwin College as an international graduate college, on Douglas’s recommendation. With my interest in hydrodynamics he suggested that I apply to DAMTP. The application was successful, and I started as a Cambridge student in the autumn of 1973, supported by grants from Aarhus University, and with a room in college in Silver Street, a few hundred meters from DAMTP.

At the time the discrepancy between the recently detected neutrino flux from the sun and the flux predicted by solar models was a severe problem in astrophysics. Douglas had made the suggestion, named the «solar spoon», that oscillations in the solar core could become unstable and grow to a level where they caused mixing of the core and hence temporarily changed solar structure such as to reduce the neutrino flux. As an initial project Douglas suggested that I continue work on the stability of solar oscillations, started by his previous Ph.D. student, Fisher Dilke. He had written a code to calculate the instability of the relevant oscillation modes, verifying simpler estimates that he and Douglas had developed, but had not completed the calculation before the end of his Ph.D. studies. With further calculations and tests of the code I managed to locate a sign error in the calculation, changing the sign of the stability coefficient! With this correction the results made sense, confirming the instability and relating it to the compositional structure of the solar core. This was the start of my work on stellar oscillations, which I have never really left.

An early significant result was my rigorous demonstration that dipolar modes of oscillation do not move the centre of mass of the star. This is, in a sense, a trivial result, but suspicion that the centre of mass might move had previously led to these modes being somewhat neglected. A consequence is that dipolar modes must have at least one nodal surface in the radial direction.

As a continuation of the project on oscillation instability I started to develop an improved, more flexible code to model solar structure, and worked on more realistic modelling of solar oscillations, taking into account physical processes in the near-surface layers of the star. Both these issues have been, and remain, ongoing themes in the research of my collaborators and me. At the end of my first year I presented the results at a major conference in Liège, Belgium, organized by the group of Paul Ledoux, the founding father of the study of non-radial stellar oscillations. Douglas did not attend but instructed me to make contact at the conference with Wojciech Dziembowski from Warsaw, perhaps the leading authority on the physics of stellar oscillations. Locating him amongst the large number of participants was a challenge, but I had a very interesting conversation with him and he invited me to visit him in Warsaw the following year.

A major change
A major change in my research occurred in the summer of 1975, with the realization of the possibility for seismic probing of the sun. Oscillatory motions in the solar atmosphere with periods around five minutes had been detected in the 1960s and it was proposed by Roger Ulrich in 1970, with a roughly simultaneous suggestion by Leibacher and Stein, that they reflected standing acoustic waves in the outer parts of the sun, with a clear diagnostic potential. This identification was confirmed by Franz Deubner in 1975 through more extensive observations. In parallel, Henry Hill had detected what appeared to be oscillations in the solar diameter. He presented the results at a conference in Cambridge in June 1975 organized by Douglas, showing clear peaks in a power spectrum reflecting well-defined frequencies. It was immediately clear to Douglas and me that these could be global oscillations, involving the whole sun, and hence having the potential to explore also the solar core where solar neutrinos were generated. Having already a solar model and a code to compute the oscillations I produced a spectrum of model frequencies and compared them, apparently favourably, with the observations the following day. It has since been found that Hill’s observations reflected fluctuations in the Earth’s atmosphere. Even so, his early claim was a major impetus for the development of helioseismology, taken up also by several other groups.

These early indications of the possibility of helioseismology changed the direction of my work. The computation of accurate solar models remained a central aspect, with increased relevance: stellar oscillation frequencies can be determined with very high accuracy and reflect the inner properties of the star; but to draw relevant inferences about the physics of stellar interiors from a comparison of the observations with model results, the numerical accuracy of the model frequencies, at given assumptions about the physics, should match the observational accuracy. This remains a key goal of much of my work. A second important aspect was to investigate through modelling and analysis how changes to stellar internal properties are reflected in the oscillation frequencies.

Honeymoon in Poland
My wife Birte and I got married in 1975, and after our wedding we spent a month in Poland, partly in Warsaw and partly at an observatory in the countryside outside the city. This offered excellent opportunities for discussions with Wojciech on the properties of nonradial oscillations and the prospects for helioseismology. We returned together to Cambridge, where Birte carried on her studies for an MSc in atomic physics. One consequence of this was that I was introduced to the mathematical properties of avoided crossings of eigenvalues of physical systems, well-known in atomic physics but just appearing in astrophysics, including in a problem that I was working on.

When I had to set the time for the submission of my Ph.D. dissertation, Birte and I were expecting our first child, our daughter Karen, and I somewhat foolhardily set the deadline for my dissertation to the date where birth was expected. Fortunately, Karen was late by a few more days than the dissertation.

Postdoc in Liège and Boulder
After I finished my Ph.D. in Cambridge, we moved back to Aarhus, where I had a one-year association with the Department of Astronomy, continuing work started in Cambridge and started developing a new code for computing adiabatic oscillations of stars, which has seen widespread use. I succeeded in obtaining Danish support for a two-year postdoctoral position at Institut d’Astrophysique, Université de Liège, in Belgium, in the group of Paul Ledoux. This was an excellent opportunity to work with leading scientists in the field of stellar oscillations, further developing the prospects for helioseismology, and my understanding of the properties of stellar oscillations. The computational resources were based on a computing centre, the institute being serviced by a daily van to bring punched cards to the centre and return the output, so my work shifted more in the direction of analytical studies of the properties of stellar oscillations.

First trip to the US
An important event during this period was a three-week visit to Henry Hill’s group in Tucson, my first trip to the US, which included a workshop on stellar pulsations. Here George Isaac from Birmingham presented the first true observations of global solar oscillations. Like Deubner’s observations they focused on oscillations with periods near five minutes. The Birmingham observations, however, analysed light integrated over the solar disk, as in stellar observations, and hence were sensitive to modes of the largest scale on the solar surface, which penetrate the core of the sun. Douglas, who also attended the workshop, and I immediately recognized the observed frequency pattern as reflecting the expected asymptotic behaviour of high-order, low-degree acoustic modes. Further, more detailed, observations of these oscillations were obtained by Gerard Grec, Eric Fossat and Martin Pomerantz from five days of nearly continuous observations from the geographical South Pole. Douglas and I developed the interpretation of these modes over the following year, comparing with solar models, and concluding that models proposed to explain the low measured flux of solar neutrinos were unlikely to be correct.

High Altitude Observatory
The trip was also an occasion to visit Boulder, Colorado, as part of planning a postdoctoral position there. With offers from both the Joint Institute for Laboratory Astrophysics (JILA) and the National Center for Atmospheric Research (NCAR), I accepted the offer from the Advanced Study Program at NCAR to work at the High Altitude Observatory (HAO). This was the beginning of a close association with NCAR over more than two decades, including three years as postdoctoral associate and later an appointment as affiliate scientist, involving month-long visits almost every summer over an extended period. The High Altitude Observatory has a strong programme in solar physics, with a focus on solar magnetic activity, and involves observations from the ground and space as well as numerical modelling. At the time of my appointment there was also a substantial activity in related areas of stellar research. My main collaborator at HAO was Tim Brown, originally from Henry Hill’s group, who has played a major role in the development of observational helioseismology. I also established contact with the group of Juri Toomre at JILA, which worked on detailed modelling of solar and stellar convection and some aspects of helioseismology. My connection with this group remained an important feature of my summer visits to Boulder.

The High Altitude provided an excellent opportunity for me to get familiar with other aspects of solar physics and, through Tim Brown, with details of helioseismic observations. On the other hand, I brought insight into the analysis and prospects of the helioseismic data, which was not developed there at the time. This complementarity strongly improved my understanding of the observational aspects of helioseismology, including the challenges in carrying out the observations at the required very high level of sensitivity.

Some of my most important work at HAO involved also collaboration with Søren Frandsen, on sabbatical leave from Aarhus University. In one project, we developed tools for realistic studies of the effect of stellar atmospheres on the oscillations, including a detailed treatment of the radiative transfer. We published initial results of this work, and it was subsequently taken up by two South African Ph.D. students that I supervised, but it still remains to be completed, including proper comparisons with the observed properties of the oscillations. A critical issue is the effect of convection, to which I return below. A second project with Søren was inspired by the early understanding of the possible origin of the solar oscillations, developed by Goldreich and Keeley, in terms of acoustic noise generated by convection. This strongly suggested that the oscillations should be a common feature of solar-like stars, opening the possibility for similar seismic studies of other stars. Søren and I made estimates of the expected amplitudes of oscillations as a function of stellar properties, based on the Goldreich-Keeley analysis. Our hope was that our estimates could be helpful in the attempts to detect solar-like oscillations in other stars. An important result was that the amplitude increased with stellar luminosity, as confirmed by later observations.

“Our hope was that our estimates could be helpful in the attempts to detect solar-like oscillations in other stars.”

This period saw a dramatic evolution in the observations of solar oscillations. Particularly important were observations by Jack Harvey and Tom Duvall of modes of intermediate degree, essentially linking the low-degree modes observed by the Birmingham group and from the South Pole to the high degree-modes observed by Deubner. In this way the full characteristics of the observed modes could be identified, allowing a reliable comparison with modes computed for a solar model. From a preliminary analysis of such data, Douglas and I identified a dominant source of the differences between observed and model frequencies to arise from the near-surface layers of the sun, the modelling of which suffered from uncertainties in the treatment of convection and the energetics of the oscillations. This surface error remains a source of uncertainty in helio- and asteroseismic analyses and an active area of research.

It was evident that full realization of the potential of helioseismology would require observations over very long periods, ideally without interruption. To achieve this, the Birmingham group started establishing the Birmingham Solar Oscillations Network (BiSON) network of observing sites, using observations in integrated light, and planning started for the Global Oscillation Network Group (GONG) to be organized by National Solar Observatory in Tucson, for spatially resolved observations. I had some involvement in the planning for the GONG network; it started observations in 1995 and, like the BiSON network, is still operating. Observations from space would have obvious advantages in terms of continuity and lack of atmospheric interference, and plans started to develop instrumentation for such observations. Together with Roger Ulrich, I was involved in a working group, led by Phil Scherrer, at Stanford, to define such an instrument for observations with high spatial resolution. This led to the Michelson Doppler Imager, essentially using a technique developed by Tim Brown, which was flown on the joint ESA/NASA mission Solar and Heliospheric Observatory (SoHO) launched at the end of 1995.

With the increased range of observed modes, it became realistic to carry out inverse analyses to infer properties of the solar interior. Douglas had pointed out the power of techniques for such analysis developed in geophysics and I implemented a simple technique, which Douglas and I tested on artificial data reflecting the frequency splitting induced by solar rotation. The results showed great promise for detailed inferences of solar internal rotation, with the data expected from the observational projects under development. Douglas also developed a technique that directly allowed determination of the solar internal sound speed from observed frequencies. Applying that to the available data, he and I, with other collaborators, in 1985 published a first estimate of the variation of sound speed with position in the sun. Comparing that with solar models showed reasonable agreement, but strongly indicating that the model opacity was too low in the region below the solar convection zone. This has since been confirmed by newer opacity calculations.

Returning to Denmark
I returned to Denmark in 1983, for a position as a fellow of the Nordic Institute for Theoretical Physics, best known as NORDITA. Here I developed a diagram showing the evolution of stars in terms of differences between frequencies of low-degree solar-like modes. Given observed differences, a first estimate of the mass and age of a star can be obtained from its position in the diagram. I presented an early version of the diagram, by some known as a ‘C-D diagram’, at a meeting in Paris in connection with a French space project for observation of solar-like oscillations in stars (unfortunately the mission failed to reach orbit).

Associate professor
In 1984, I was appointed associate professor at the Department of Physics and Astronomy at Aarhus University, returning to my alma mater. To establish helio- and asteroseismology firmly in Aarhus, I organized the first major international conference in the field, Symposium 123 of the International Astronomical Union, in 1986 with about 130 participants. This demonstrated the substantial recent advances in helioseismology, and the prospects for seismology of solar-like stars as well as investigations of other types of oscillating stars.

Equation of state
An important aspect of stellar modelling is the so-called equation of state (EoS), i.e., the relations between pressure, density and temperature and other thermodynamic properties. Together with Werner Däppen, I implemented a more advanced version of the EoS in my model and frequency calculations. Comparing that with the observed frequencies showed substantial improvements over the simpler formulation used previously. Also, it highlighted the dominant role of the near-surface layers in the remaining discrepancies between the sun and the model. Such studies of the EoS have remained an active area of my research; there are still significant discrepancies between the most recent formulations and the observations. By characterizing the properties of matter under the extreme conditions found in the sun, the results can also find use in high-temperature applications on Earth.

Further developments in the observations of solar oscillations by Ken Libbrecht opened the possibilities for more detailed investigations of solar rotation. My MSc student Jesper Schou and I carried out a first inverse analysis of rotational splittings to determine both the radial and latitudinal variation of solar rotation. The results were consistent with the known latitude variation of the surface rotation frequency and showed that this variation persisted through the convection zone, with a transition to latitude-independent rotation in the deeper interior. This gradient, later named the tachocline by Spiegel and Zahn, appears to play an important role in the generation of the solar magnetic activity.

A key collaboration was established when Michael Thompson, a former Ph.D. student of Douglas, arrived for a one-year postdoc position in Aarhus. Michael was for many years a close collaborator and friend, until his untimely death in 2018. We typically shared summer visits to Boulder, with work that also involved Juri Toomre’s group. Central aspects were the further development and testing of inversion techniques in a collaboration that also involved Jesper Schou, who continued as my Ph.D. student before taking up a position at Phil Scherrer’s group at Stanford. This provided a much improved understanding of the relation between different inversion techniques and the statistical and systematic uncertainties in the results, aspects that are still important in my work. Together with Douglas, Michael and I also made a first accurate determination of the depth of the solar convection zone. Additional activities concerned a better understanding of the effect of uncertainties in stellar modelling on the oscillation frequencies and the inversion results. A particularly active period of the collaboration was the programme on helioseismology organized by the then Institute for Theoretical Physics, Santa Barbara, in the spring of 1990 (now the Kavli Institute for Theoretical Physics as of 2001).

An important result that Michael and I obtained, in collaboration with Charles Proffitt, was that inclusion of helium settling in the modelling substantially improved the agreement between the model and the helioseismically inferred sound speed. Settling was a known but often ignored process in the solar interior, in the calculation of what was known as standard solar models; our results clearly demonstrated that neglecting it was not justified, and diffusion and settling are now as a standard included in solar and much stellar modelling.

The early observations of the Global Oscillation Network Group, or GONG network, were marked by a special issue of the journal Science in 1996. Here I participated in inverse analyses for solar structure and rotation and led a paper on the modelling of the sun. This included the computation of a new model of the sun, now known as Model S, based on physics that was fully up to date at the time. I made the model easily available on the internet, and as a result it has seen remarkably broad use in various areas of solar physics. The model is clearly no longer fully up to date; however, amongst models of the sun computed without adjustments to fit the observations, it is still one of those that most closely matches the seismic inferences of solar structure.

Developing the Aarhus group
Funding for my research got a major boost in the period 1994 to 2004, with the establishment of the Theoretical Astrophysics Center (TAC), a centre of excellence funded by the Danish National Research Foundation. The activity was shared between Copenhagen University, with Igor Novikov as director of TAC, and an Aarhus branch led by me. This allowed employing Frank Pijpers on a long-term contract and a number of postdocs. With Sarbani Basu, and later Maria Pia di Mauro, I further developed inverse analysis for solar structure, using the increasingly extensive observational data. One project was a detailed inverse analysis specifically addressing aspects of the equation of state, documenting that neither of the two detailed formulations then available agreed with solar properties. Despite further updates to the EoS, this remains an open issue. My Ph.D. student Rasmus Munk Larsen, shared with the Department of Computer Science, and I also developed sophisticated codes for the inversion of solar rotational data, which still see substantial use.

Hans Kjeldsen was hired as a TAC postdoc at the start of the center and has been a central person to my activities since then. He did his Ph.D. in Aarhus supervised by Søren Frandsen and focused on observational aspects of asteroseismology. The strength of our collaboration is in fact the interplay between his observational expertise and my activities more aimed at theoretical and modelling aspects. In addition, he has played a key role in organizational aspects of the work of my group. In 1995, Hans led the first detection of individual solar-like modes of oscillation in a star other than the sun, using the Nordic Optical Telescope to observe the star η Bootis, a subgiant somewhat more evolved than the sun. I carried out modelling of the observed frequencies and found the first observational indication of avoided crossings in the so-called mixed modes. These are modes that have the character of acoustic waves in the outer part, and internal gravity waves in the core, of the star. This makes the frequencies much more sensitive to the core properties than the purely acoustic modes seen in the sun. Mixed modes have since played a major role in asteroseismic investigations using the extensive observational data that are now available.

In the following years, solar-like oscillations were detected and studied in some details by my group in a few other stars, using observations of radial velocity of stellar surfaces through the Doppler effect; this was possible largely thanks to advances in spectroscopy aiming at the detection of planets outside the solar system, the so-called exoplanets. The work involved a number of collaborators, with very important contributions from Tim Bedding, in Sydney. A major advance just after the turn of the millennium was the first definite detection, led by Søren Frandsen and his MSc student Dennis Stello, of solar-like oscillations in a red-giant star. With space-based observations, asteroseismology of red giants has become a very active field, of major importance to the study of the evolution of the Milky Way Galaxy.

The issues of the modelling of the outer layers of stars have remained an important area of research for my collaborators and me. Günter Houdek, then Ph.D. student in Vienna partly supervised by me, adapted a code developed by Douglas to model the interaction between convection and stellar pulsations and taking into account the energetics of the oscillations. He carried out extensive calculations of the excitation and damping of stellar oscillations, accounting for the so-called instability region where modes are self-excited, including the large-amplitude δ Scuti and δ Cephei stars. His calculations also largely reproduced the damping rates of solar oscillations as inferred from the width of their peaks in a power spectrum.

A second major source of uncertainty in the modelling of near-surface layers of stars is the structure of convection, typically treated using a highly simplified procedure. My postdoc Colin Rosenthal, another former Ph.D. student of Douglas, and I collaborated with Åke Nordlund, in Copenhagen, and others to improve this by using detailed hydrodynamical simulations of convection, restricted to the region near the stellar surface. Colin devised a way to attach the average structure of the simulations to models of the deeper parts of the star. This demonstrated that much, but not all, the frequency differences between the observations and the model could be accounted for by such a correction to the model. I return to this issue below. The technique of matching convection simulations to stellar interiors has been extensively used in recent years, with the potential routinely to calculate stellar models with an appropriate treatment of convection.

New observations from space and from the ground
It was clear early on that asteroseismology of solar-like stars would in principle be possible with photometric observations, but only from space owing to the effects of Earth’s atmosphere in ground-based observations. Thus efforts to define and launch a space mission to carry out such observations started quite early, with several proposals to the European Space Agency (ESA). I was involved in some of these proposals, none of which made it to launch. However, in the early 2000s the Eddington mission, with the joint goal of asteroseismology and search for extra-solar planetary systems, made it far enough to be included in ESA’s programme as a reserve mission to be adopted given adequate funding in ESA’s Science Program, which for a while seemed possible. This possibility ended due to severe cost overruns in another mission. Being a member of ESA’s Space Science Advisory Committee at the time, I had the painful duty to support the cancellation of Eddington, in favour of another mission of higher priority in the programme. Roughly parallel with this, Hans and I were developing the Rømer mission, a Danish proposal for a micro-satellite dedicated to asteroseismology. The design was highly developed, demonstrating the feasibility of the mission, but at this point the proposal was also cancelled due to lack of funds.

These setbacks were obviously frustrating, but the experience with Eddington and Rømer gave my group a strong background for work in space asteroseismology. This caught the attention of the leaders of the NASA Kepler mission, William Borucki and Dave Koch. Kepler was under development to search for exoplanets using very precise photometry of a huge number of stars, to catch the rare transits of planets in front of the stellar disk. The photometric precision would be sufficient to study oscillations at the level of those seen in the sun, and the requirements for asteroseismology were taken into account in the design of the mission, but the project lacked the funding fully to utilize the asteroseismic data. With the support from US participants in the mission, including Tim Brown, Borucki and Koch approached us to suggest that we participate, with responsibility for the asteroseismic use of the data. In return for data access we committed to carry out asteroseismic analysis of stars found to host planetary systems, to determine stellar radii, masses and, if possible, ages. This was clearly a huge opportunity for my group, which we accepted. It was clear that the expected amount of data was far beyond what we could analyse in a satisfactory manner. With Borucki’s approval, we therefore established the Kepler Asteroseismic Science Consortium (KASC), to organize the work with the data, setting up separate working groups to deal with the different types of pulsating stars, and with a total membership of around 500 scientists from all over the world. In support of the work, we also set up the Kepler Asteroseismic Science Operation Centre (KASOC) to distribute the data and organize the results and the publications resulting from them. This was largely established by Rasmus Handberg, who at the time was a Ph.D. student supervised by Hans.

Kepler
Kepler was launched from Cape Canaveral in March 2009. Hans and I witnessed the launch, my first such experience. After my early fascination of the space programme where Cape Canaveral played a major role, being present at this launch with its strong implications for my own work was a major experience. The organization of KASC and KASOC led to a strong asteroseismic research activity already in the early phases of the mission, covering a broad range of stars and fully confirming the potential of space asteroseismology.

Although photometry from space provides simultaneous data from large numbers of stars, radial-velocity measurements from the ground still have substantial advantages in terms of sensitivity and a lower noise from other phenomena in the stellar atmospheres, as demonstrated by Hans and his collaborators with observations using the largest telescopes in the world. Based on this experience we found that a similar precision for bright stars could be obtained with modest telescopes, if they were optimized for this type of investigation. This realization led us to start the Stellar Observations Network Group (SONG) project, with the goal of setting up a global network of one-meter telescopes with optimized instrumentation for radial-velocity observations. We obtained funding to build a prototype site at the Teide Observatory on Tenerife, in collaboration with Instituto de Astrofísica de Canarias and University of Copenhagen. The telescope is fully robotic, carrying out observations based on a list of possible targets and the immediate observing conditions, and then after a night’s observations the data are transferred to Aarhus for further analysis.

The Tenerife telescope, named the Hertzsprung SONG telescope, was officially inaugurated in 2014 and has operated successfully since then. The results have largely lived up to expectations and include the most detailed seismic data for any star apart from the sun, for the subgiant μ Hercules. Following the prototype, a node in the network was set up by our Chinese collaborators; it is currently being moved to a better site. Also, a node in collaboration with Australian collaborators using two smaller telescopes is being commissioned in Queensland, and we have started collaboration with US scientists to set up nodes in New Mexico and Hawai’i.

Two grants
In 2011, I received a Senior Advanced Grant from the European Research Council. Also, in 2012 the Danish National Research Council awarded me a grant to set up the Stellar Astrophysics Centre (SAC), a 10-year centre of excellence. These two grants led to a major increase of the range and staff of my group. For my own research, key people were Günter Houdek who returned as postdoc and later associate professor, and Victor Silva Aguirre (now Victor Aguirre Børsen-Koch) who came as a postdoc just after his Ph.D. and later obtained a permanent position as associate professor at Aarhus University.

Günter and I, together with Regner Trampedach, in Boulder, and my Ph.D. student Magnus Aarslev, have continued work on the near-surface physics of solar and stellar models. By including an averaged hydrodynamic simulation in the model, and in addition using the procedures originally developed by Douglas for the interaction between convection and oscillations, we were able very substantially to reduce the difference between the modelled and observed solar oscillation frequencies, at the same time providing a good match of the observed spectral line widths. We also demonstrated a good match between the computed and observed line widths for several well-observed Kepler stars. These results indicate that we are on the right track towards understanding the properties of these critical layers and provide a hope that we may eventually eliminate this source of uncertainty in asteroseismic investigations.

Based on my early asteroseismic analysis of a Kepler star, Victor led a project, including me as well as collaborator from KASC, to determine asteroseismic properties of stars found by Kepler to host planetary systems. This determined relatively precise masses, radii and ages of around 30 stars, finding stars with ages of more than 10 billion years (twice as old as the sun) to have planets. Victor and I also led an analysis of the so-called Kepler Legacy stars –around 60 stars with exquisite asteroseismic data. This is now a reference set of results and a basis for further extensive investigations of these stars.

Analysis of the Kepler data is still ongoing. Furthermore, the activities of my group have been continued for the NASA Transiting Exoplanet Survey Satellite (TESS) mission, with similar goals as Kepler, in the corresponding TASC collaboration and with data made available in TASOC. We are also strongly involved in preparations for the ESA PLAnetary Transits and Oscillations of stars (PLATO) mission with expected launch in 2027; here I serve as Danish representative on the PLATO Mission Consortium Board.

The future of my research group
The establishment of the Stellar Astrophysics Centre (SAC) aimed at a major extension of the activities of my group, since SAC, in addition to addressing stellar physics, also included exoplanets and astrobiology, with the goal to establish or strengthen these fields at Aarhus University. The astrobiology part of the centre was in collaboration with Kai Finster at the Department of Bioscience, and in addition SAC included formal collaboration with six external nodes in Europe, the US and Australia supplementing the local expertise.

This broadening of activity has been successful. Victor has taken up the study of the evolution of the Milky Way Galaxy, using his expertise in stellar physics for asteroseismic characterization of red-giant stars. Simon Albrecht was hired to establish exoplanet research and now has a permanent position at our department. Christoffer Karoff, who has worked in SAC on solar and stellar activity and its influence on Earth and exoplanets, is now also in a permanent position as associate professor at the Department of Geoscience. At the Department of Bioscience Tina Šantl-Temkiv, working on the effects of bacteria on cloud formation, is in a tenure-track position. Victor, Simon and Tina have been able to attract personal grants, further adding postdocs and Ph.D. students to the group. Thus, the research activities started in the centre are set to continue after its conclusion in September 2022, with broad impact on the university.

Emeritus status
My personal employment as professor ends in June this year, after which I transition to an emeritus status. For my personal research, this will not make much difference. Having been gradually relieved of formal duties I am focusing on further development of my codes for calculating stellar evolution and oscillations, probing aspects of the models and their oscillation modes in the hope of understanding them better, and in this way also improve the potential for asteroseismology. In addition, I find myself getting involved in various projects by international colleagues; it is difficult to say no to an interesting project! And, of course, I enjoy discussing research with the local community of younger scientists.

Beyond research
I have never had the ambition to be involved in broader administrative leadership roles, and I have had the very good fortune that Hans Kjeldsen has dealt with the administrative duties of my various funding projects, lately as Centre Manager for SAC. However, I have had some administrative duties on a national and international level. I was member of the Danish Natural Science Research Council in the 1990s and member of various space advisory bodies, including a period as the head of the committee. The latter position gave me the opportunity to attend the launch of the first Danish astronaut from Baikonur in 2015, an overwhelming experience owing to the relatively small Russian safety distance for launches. I was member of the Scientific and Technical Committee of ESO for four years from 1988, and of council of the Nordic Optical Telescope for an extended period up to 2013. In the International Astronomical Union I was president of Commission 27 for three years from 2000 and subsequently president of Division V. I was member of the ESA Space Science Advisory Committee 2001-2003, and Danish Delegate to the ESA Science Programme Committee 2005-2017, for the last year as its president. All these activities have in their different ways provided interesting insights into other areas of research and the way that international science is managed.

“Looking at the starry sky on clear moon-less nights is a never-ending joy.”

As dedicated to (some might say obsessed with) my research, this has also been my hobby. However, I do enjoy reading, mainly fiction and biographies, and listening to classical music. As an attempt at distraction from work, only partly successful, my wife found a summer house in a wonderful location on the west coast of Denmark. This is amongst the regions least affected by light pollution in Denmark, and looking at the starry sky on clear moon-less nights is a never-ending joy. And of course I treasure spending time with our daughters Karen and Signe and their daughters, in, respectively, Edmonton, Canada and Trondheim, Norway.

Concluding remarks
This is perhaps not the appropriate location for giving thanks, but let me at least note that all my research activities would not have been possible without the wonderful mentors and collaborators, some of which are mentioned above, starting with Douglas Gough who introduced me to the field of stellar physics, and including Michael Thompson who played a key role in the development of helio- and asteroseismology and Hans Kjeldsen who has been invaluable in the development of the activities at Aarhus University. The funding from the Danish National Research Foundation has played a seminal role in the development of my activities, as has other more specific sources of funding, in particular for the development of the SONG network.

And last, but far from least, all of this has relied on the never-failing support and tolerance of my wife Birte.