Steven Vogt é astrónomo e professor de astronomia e astrofísica na Universidade da Califórnia, Santa Cruz. É conhecido mundialmente pelo seu trabalho pioneiro no desenvolvimento de espectrógrafos de alta resolução dos quais se destaca o HIRES (High Resolution Echelle Spectrograph) instalado no telescópio Keck I no Hawaii. Foi orientador de doutoramento de Geoffrey Marcy e é uma figura de referência na área dos exoplanetas, fazendo parte de equipas creditadas com mais de 100 descobertas. É o investigador principal do projecto APF (Automated Planet Finder), um telescópio optimizado para o estudo de exoplanetas instalado no Observatório Lick, e desenhou um novo espectrógrafo para o TMT (Thirty Meter Telescope). O professor Vogt concedeu recentemente uma entrevista ao AstroPT que aqui reproduzimos.
[AstroPT.org] – You are principal investigator of the APF (Automated Planet Finder) telescope project. Can you tell us about the history of the project, from its conception to its current status ?
The APF project got its start back in 2001 at UC Berkeley thanks to efforts from Geoff Marcy, Debra Fischer, and Bernie Walp. They managed to get major funding from the Department of Defense for an Automated California Planet Finding telescope. Initially, we were going to just purchase one of the 1.8-meter “outrigger” telescopes built by EOS Corp. for the Keck Interferometer project (which were in storage in Tucson and never going to be used at Keck). Our intent was to install the 1.8-m in a dome adjacent to the Shane 3-m telescope at Lick Observatory and simply fiber-couple the telescope into the Shane’s Hamilton spectrometer. But we really wanted a larger telescope with its own modern, highly optimized spectrometer, so I was able to garner enough additional funds from NASA to allow us to purchase a 2.4-meter telescope, fitted with its own highly-optimized RV spectrometer.
This path was a much more ambitious undertaking, and required the procurement of a new 2.4-m mirror, and building an entirely new telescope from scratch. Thus, what was originally expected to be a 1-2 year project unfortunately ended up stretching now to almost a decade. At present, we have a fully-working dome up at Mt. Hamilton, with a telescope installed. Both the dome and telescope have passed their Final Acceptance Tests with flying colors. The spectrometer housing and support systems have also been installed, and the spectrometer optical train is expected to be installed over the coming 1-2 months.
(APF em construção)
[AstroPT.org] – The APF (Levy) spectrograph is perhaps its most innovative feature. What were the requisites for the instrument? Can you describe its design and the features that allow it to achieve the planned precision?
I wanted to build a spectrometer that was specifically optimized for high precision radial velocity work, both from the standpoint of limiting precision as well as optical efficiency. Following on the MIKE concept pioneered by Steve Shectman and Rebecca Bernstein of The Carnegie Institution of Washington, I decided to make the instrument a totally unobstructed all-dioptric style (all-refracting collimator and camera). This optical configuration allowed the echelle grating to be used in quasi-littrow mode for maximum peak efficiency. I also incorporated the use of high-tech coatings to minimize photon loss at all air-glass surfaces. Most of the surfaces are coated with a sol-gel coating process that we developed in the Lick Labs. The sol-gel is spun on in the case of the large lenses, and dip-coated for the large prisms. The two flat fold-mirrors in the spectrometer are coated with proprietary multi-layer reflective coatings for very high efficiency across a huge bandpass. One of the 6″ diameter fold flat mirrors has a coating we developed in collaboration with MLD Coatings in Sunnyvale. This coating has about 150 layers and produces a reflectivity above 98% from the UV all the way up into the IR. It is, as far as I am aware, the highest performance optical reflecting mirror in use in astronomy.
Another unique feature of this spectrometer is extensive passive athermalization of the optical train. Stability is of paramount importance in a spectrometer that is expected to be able to track movements of the spectrum of less than 1/1000 of a pixel from season to season and year to year. Lenses are made of glass and that glass responds to temperature variations by 1) changing its thickness, 2) changing the radii of curvature of its two surfaces, and 3) changing its index of refraction (the amount by which it bends light). By incorporating all these variables into the optical design directly, I was able to work out how much each lens needed to move axially to stabilize the overall focus and image scale of the optical train, thereby ensuring high stability at the final focal plane. This required supporting the various optical elements using a determinate structure of struts. The thermal behavior of this “spaceframe” can be accurately modeled in a computer, and specifically tuned to provide the requisite de-spacings of the optical elements to provide stability of both focus and scale in the face of temperature changes. The strut structure is also exceedingly stiff and light weight.
The combination of an all-dioptric path, together with highly efficient optical coatings and use of the echelle grating near littrow should allow us to reach an overall efficiency of about 30-35%, a good factor of 3 higher than existing conventional high resolution spectrometers used for this kind of work. So even though the telescope is comparatively small, it will pack a pretty serious punch for exoplanet hunting. And since we are not sharing it with other groups and have total access to it, the increased number of nights also helps to offset its size disadvantage.
[AstroPT.org] – The APF was designed to follow up “interesting” systems discovered with other programs. How do you define “interesting” ? In other words, what are the criteria used to select systems for the observing program ?
Interesting in this context means a star that shows itself to be intrinsically quiet (in a radial velocity sense), but that is slightly noisier than expected from its chromospheric activity index. That slight extra “noise” is likely the signature of undiscovered planets. Other interesting candidates are stars that have already revealed one or more planets, but none in or near the star’s habitable zone. What is needed to reveal these additional planets is generally much higher cadence, and that will be made possible with APF.
[AstroPT.org] – Since APF is a relatively small telescope, it will be limited to observing moderately bright stars. Can you describe its limitations and trade-offs ? When is it expected to start taking data?
Yes, APF is a rather small telescope by comparison with telescopes such as the Keck 10-m telescope, or the HARPS 3.5-m telescope. But some of what we lose in aperture, we hope to make up in spectrometer efficiency. We should be able to work routinely down to at least V = 9 and that will include a very large number of interesting stars. It can probably be pushed even fainter, but will require correspondingly longer exposures, and thus fewer stars per evening. We hope to be taking our first data by June of this year.
[AstroPT.org] – Together with Paul Butler you are leading the Lick-Carnegie Exoplanet Team that has recently announced Gliese 581g. Can you describe the project and its goals?
We have been observing GJ 581 for over 11 years now. The Swiss group has already found 4 interesting low-mass planets around this star, and we have been shadowing their work, monitoring and checking each of their emerging planets as they appeared. In September of 2010, we decided to finally join the conversation on GJ 581 by publishing our 11-year data set that, in combination with the 4-year Swiss HARPS data set, revealed two more planet candidates in this remarkable system. One was GJ 581f, a 433d planet. The other was GJ 581g, a 37-d planet smack in the middle of the habitable zone, and with a mass only 3.1 times that of the Earth. GJ 581g thus is a highly promising candidate as the first known potentially habitable Earth-like planet. And at a relatively close distance of only 20 light years, its early discovery portends that there must be many Earth-like planets orbiting the nearby stars. Indeed, our Milky Way galaxy must be teeming with such potentially habitable planets, with obvious implications for the possibilities of life elsewhere in the universe.
[AstroPT.org] – You have designed and built HIRES (High Resolution Echelle Spectrograph), the powerful spectrograph installed in the Keck I telescope. This has been a workhorse for many ongoing exoplanet programs. Can you describe the instrument, its performance, and how it is modified for exoplanetary work? How long does it take to fully calibrate and optimize such an instrument?
I built HIRES as a general-purpose workhorse instrument for Keck. And since its commissioning in 1993, it has been used for quite a wide range of scientific projects. Early on, it was primarily used for extragalactic research, mostly involving faint quasar absorption line spectroscopy. It has also been quite productive in studies of dark matter, and in the temperature of the Cosmic Microwave Background at early times. Much stellar abundance work has also been carried out with HIRES. The faintest I have ever used it is at about V=22 on stars in and around the Andromeda galaxy. But it is mostly used on objects in the V= 15-20 range.
HIRES was not specifically optimized for high precision radial velocity work, but I installed an iodine cell precision wavelength reference system for this type of work from the outset, and it is this device that has turned HIRES into a quite powerful and productive instrument for exoplanet hunting. I also installed an exposure meter that calculates the intensity-weighted time centroid of each observation, necessary for avoiding errors due to barycentric corrections. Both the Iodine cell and exposure meter are available to all planet hunting users and make HIRES a quite powerful facility for this type of research. HIRES is now used by many different groups for exoplanet research, most notably in recent months for Kepler follow-up through large blocks of time purchased by NASA.
(Célula de Iodo)
Calibration of HIRES is done nightly and takes about an hour before sunset for all the necessary calibration lamps. Optimization of HIRES was done only once, when the instrument was initially installed and took several months of careful alignment and focussing. Since HIRES is also passively athermalized, the focus is essentially constant and need not be re-done by observers. We do however have our own proprietary fine-tuning optimization that we do each evening as part of our setup, that helps us reach precisions of better than 1 m/s.
[AstroPT.org] – You are also working on the MTHR (Moderate-To-High-Resolution Spectrometer) spectrograph to be installed in the future TMT (Thirty Meter Telescope). What are the innovative aspects of this spectrograph and its capabilities ? Is it designed from scratch for exoplanetary work ?
I designed MTHR to also be a workhorse instrument for our upcoming TMT (30-m telescope). The spectrometer size has to scale with the size of the telescope, so moving from a 10-m to a 30-m means a spectrometer at least 3 times bigger than HIRES (which is already the size of a typical faculty office). Figuring out how to make an instrument that large, using available optical elements was the big challenge. I ended up devising a scheme that was a mix of the dual-white-pupil concept of the VLT’s UVES spectrometer, combined with the large catadioptric cameras pioneered by HIRES. This combination allows then a collimated beam diameter of about 1-meter. It uses camera that have 60″ diameter fused silica lenses, and mirrors that are 2-3 meters in diameter. The echelle gratings are large mosaics, each about the size of a good-sized office bookcase, some 3′ by 8′ in area. The entire instrument sits in a thermally insulated housing with a footprint about the size of a tennis court, and sits up on the telescope’s Nasmyth platform.
MTHR has a number of aspects that will help optimize it for exoplanet hunting, such as the usual Iodine cell, coupled with a high degree of passive athermalization of the optical train. I also designed a fiber-optic image scrambler/slicer to stabilize the pupil and produce even further radial velocity precision. The main drawback is available observing time for any one group. The 30-meter will be in such high demand, among such a large pool of partner institutions, that no one person or group will get many nights each year. That will make it very hard to carry out an exoplanet survey. So MTHR’s main use will be to target a rather small number of the most interesting stars to study in more detail. It will be used for example to do follow-up on Kepler transit candidates, which, at V=14, are just too faint for today’s largest telescopes. Another exoplanet niche for MTHR is absorption spectroscopy of transiting planets, obtaining spectra of the light that streams through the atmosphere of a planet while in transit. This is very challenging observationally, and will require the reach of such a large telescope mated with a powerful spectrometer. Or another possible use might be in a queue-scheduled mode where the instrument is switched in for 5-10 minutes a night, or every few nights, to monitor M dwarfs for earth-sized planets. There, with a very small amount of total time, one could imagine carrying out a reasonable survey on a respectable number of the nearest M dwarfs for habitable Earth-sized planets.
Unfortunately, MTHR was not selected for first-light of the 30-m telescope project. And by the time it might be selected, I will probably be retired. So it is unlikely I will have a hand in actually building MTHR, though it would be an amazing and powerful instrument for the 30-m, and I hope that someday it gets built.
Resumo da entrevista em português com pequenos apontamentos, assinalados por (astropt: …), que pretendem esclarecer alguma referência menos explícita no texto. O texto é da exclusiva responsabilidade do autor da entrevista.
O professor Steve Vogt começou por falar-nos do projecto APF (Automated Planet Finder) que teve início em 2001 na Universidade de Berkeley pela iniciativa de Geoff Marcy, Debra Fischer e Bernie Walp. Inicialmente tratava-se de construir um telescópio de 1.8 metros ligado ao espectrógrafo Hamilton, instalado noutro telescópio do Observatório Lick, através de fibra óptica. Entretanto os planos mudaram e o projecto inicial foi reconfigurado em algo mais ambicioso e necessariamente mais demorado: um telescópio de 2.4 metros com um espectrógrafo próprio optimizado especificamente para o estudo de exoplanetas, para o qual contribuiram fundos da NASA. Actualmente a cúpula e o telescópio estão instalados e passaram todos os testes de aceitação final. O espectrógrafo está a ser instalado. Em princípio as primeiras observações poderão ser realizadas em Junho deste ano.
Steve Vogt entra depois em detalhes mais técnicos descrevendo o espectrógrafo Levy (em honra de um mecenas) do APF, um dos mais sofisticados do mundo. O instrumento foi optimizado para medições de velocidade radial de estrelas com grande precisão e ao longo de anos. Vogt desenhou o sistema óptico e incluiu “coatings” especiais nas componentes ópticas para maximizar a transmissão de luz e concentrar o máximo de luz na rede de difracção que decompõe a luz do espectro das estrelas. Os espelhos usados no instrumento têm também “coatings” de muito elevada reflectividade, cerca de 98% desde o infravermelho ao ultravioleta. Para além as optimizações no sistema óptico, a estrutura de suporte do espectrógrafo, que garante o alinhamento preciso das componentes ópticas e um foco preciso, foi também alvo de atenção especial. A dita foi desenhada com base num modelo simulado em computador que permitiu optimizá-la no sentido de obter consistentemente medições de alta precisão ao longo de anos. O espectrógrafo terá uma eficiência global de 30%-35%, cerca de 3 vezes superior aos espectrógrafos convencionais usados no estudo de exoplanetas. Esta eficiência, diz Vogt, ajuda a contrabalançar o facto do telescópio do APF não ser tão grande como os utilizados por outras equipas (e.g. o HARPS está instalado num telescópio de 3.6 metros no Observatório de La Silla, no Chile). Por outro lado, o APF vai fazer exclusivamente medições de velocidade radial, numa cadência muito superior à que a maioria das outras equipas é capaz. Vogt está assim muito confiante que o APF se vai tornar num instrumento de referência para o estudo dos exoplanetas.
A escolha de sistemas para observação no APF é feita da seguinte forma. Sistemas já descobertos por outros programas e intrinsecamente interessantes, e.g. sistemas múltiplos e sistemas com planetas na zona habitável. Outros candidatos são estrelas que não apresentam variações óbvias da velocidade radial, mas que apresentam um “ruído” espectral ligeiramente superior ao que seria de esperar da sua actividade cromosférica. Parte deste “ruído” adicional pode ser na realidade um sinal muito fraco de planetas ainda não identificados e de massa tipicamente baixa. O sinal destes planetas só pode ser extraído do “ruído cromosférico” da estrela através de observações muito precisas e com grande cadência temporal, exactamente o tipo de tarefa para o qual o APF foi concebido.
Um destes sistemas particularmente interessantes é GJ581. Vogt fala-nos da sua colaboração com Paul Butler na “Lick-Carnegie Exoplanet Team” e a recente descoberta de dois novos planetas no sistema: GJ581f e GJ581g. Este último, ainda envolto em polémica, poderá ser o primeiro planeta do tipo terrestre na zona habitável da estrela hospedeira a ser detectado. A descoberta resultou da análise de 11 anos de observações realizadas por Vogt e Butler combinadas com 4 anos de observações realizadas pela equipa Suiça com o espectrógrafo HARPS. Vogt diz-nos que a descoberta precoce de GJ581g, na vizinhança imediata do Sol, parece implicar que a Via Láctea tem um enorme número destes planetas potencialmente habitáveis.
Vogt fala também no desenvolvimento do espectrógrafo HIRES (High Resolution Echelle Spectrograph), instalado no telescópio Keck I no Hawaii. O instrumento foi desenhado para ser utilizado em todo o tipo de trabalho espectroscópio e é o espectrógrafo de serviço no Keck I. Foi utilizado para estudos variados, desde a astronomia extragaláctica até ao estudo de populações estelares. A sua adaptação para o estudo de exoplanetas, i.e. medição precisa da velocidade radial das estrelas hospedeiras, foi feita com a instalação de uma célula de iodo (astropt: uma célula de iodo é um recipiente em vidro contendo iodo molecular (I2) no estado gasoso. A célula é colocada no percurso da luz da estrela até ao espectrógrafo pelo que, sobreposto ao espectro da estrela, fica o espectro de absorção do iodo molecular que serve de referência para a medição dos desvios de Doppler nas linhas espectrais da estrela. As moléculas de iodo produzem um espectro muito complexo, com muitas linhas e muito finas de comprimentos de onda bem conhecidos.). O HIRES assim modificado tem sido utilizado por vários grupos que usam o Keck I para estudar exoplanetas e é o instrumento utilizado na confirmação dos candidatos descobertos pela missão Kepler. A calibração do HIRES é feita diariamente, uma hora antes do Sol se pôr. A sua optimização foi feita uma vez, aquando da sua instalação no Keck I, e demorou vários meses para atingir uma performance óptima.
Finalmente, Vogt fala-nos do seu trabalho no desenho de um novo espectrógrafo designado por MTHR (Moderate-To-High-Resolution Spectrometer) para o TMT (Thirty Meter Telescope), um telescópio gigante que será construído no cume do Mauna Kea, no Hawaii. O MTHR é um instrumento gigante, três vezes maior do que o HIRES. O primeiro desafio no seu desenho foi precisamente o seu tamanho enorme. No que diz respeito à sua optimização para o estudo de exoplanetas, o MTHR deverá ter uma célula de iodo semelhante à instalada no HIRES e algumas optimizações no sentido de manter as componentes ópticas termicamente estáveis. O principal problema com o TMT será no entanto a grande competição por tempo de observação num telescópio com estas características de topo. Vogt diz-nos que o MHTR não foi selecionado para a primeira fase do TMT e caso o seja mais tarde, ele estará provavelmente reformado e não terá um papel activo na sua construção. No entanto, não resiste a dizer que gostaria de ver o MTHR construído e em serviço no TMT.
1 comentário
Excelente entrevista! 🙂