Entrevista com Gordon Walker

[Crédito: foto cedida por Gordon Walker.]

Em 1987, no meio da excitação provocada pelo aparecimento da supernova 1987A na Grande Nuvem de Magalhães, a notável descoberta de um planeta em torno da estrela Gama Cephei foi anunciada por um grupo de três astrónomos canadianos: Gordon Walker, Bruce Campbell e Stephenson Yang. Lembro-me como se fosse hoje de ver a notícia nas revistas New Scientist e Astronomy e de ter ficado fascinado com esta primeira detecção de planetas fora do sistema solar, apesar de estranhar o aparente pouco entusiasmo com que foi recebida pela comunidade científica.

[Artigo na revista NewScientist (1987) anunciando a descoberta dos primeiros exoplanetas. Crédito: New Scientist.]

Fiquei ainda mais intrigado com a engenhosa técnica inventada por Walker e Campbell para conseguir fazer a detecção. Na verdade muitas das equipas que actualmente fazem detecção de exoplanetas pela medição da velocidade radial usam uma técnica essencialmente idêntica à dos canadianos. Deles Geoff Marcy disse “(They) invented the technique that we stole” (entrevista ao The Globe & Mail, 2009) . Alan Boss, outro especialista de renome na área diz “(Gordon) Walker and Bruce Campbell […] were true pioneers in the field of searching for planets around other stars” (The Crowded Universe, 2009).

Gordon Walker, o líder da equipa canadiana, é actualmente professor “Emeritus” (aposentado mas com uma ligação activa à universidade), na Universidade de British Columbia, no Canadá. Iniciou o seu percurso académico com um bacharelato em Física em 1958, pela Universidade de Edinburgh, seguido de um doutoramento em Astrofísica pela Universidade de Cambridge em 1962. Durante a sua longa e produtiva carreira foi autor de mais de 160 artigos sujeitos a revisão independente. Participou ainda em vários projectos importantes: membro do Scientific Advisory Committee e do Board of Directors do Canada-France-Hawaii Telescope (CFHT); Project Scientist no desenvolvimento dos telescópios Gemini (dois gigantes de 8 metros no Chile e no Hawaii); membro do Gemini Scientific Advisory Committee e do Gemini Board of Directors; membro da Science Team do telescópio MOST, o primeiro telescópio espacial canadiano, e; consultor no desenvolvimento e selecção de instrumentos para novos telescópios, como o TMT (Thirty Meter Telescope).

Contactado pelo AstroPT, Gordon Walker aceitou amavelmente responder a algumas perguntas sobre a sua contribuição seminal para a detecção de exoplanetas e sobre os seus projectos mais recentes.

[AstroPT] – When did you first became interested in science, namely astronomy ? What is your scientific background ? What research areas interest you the most ?

[Gordon Walker] – My introduction to astronomy is still a vivid memory. It was 1943 walking home one night with my father during the war time blackout. I was seven years old, it was clear and I asked him about the stars – he told me that each one was a sun but so far away that they just seemed points of light. Soon after I learnt about the light year and developed an abiding interest in astronomy. I grew up in Scotland where I went to school and University graduating with a degree in Natural Philosophy (Physics) and completed a PhD in Astrophysics at Cambridge in 1962. I have always been interested in astronomical instrumentation and actually wrote a text book on it published in 1980. My areas of research apart from extra-solar planets have included interstellar dust and molecules, and stellar oscillations and rotation.

[O Canada-France-Hawaii Telescope. Crédito: Canada-France-Hawaii Telescope.]

[AstroPT] – What was the state of the art of radial velocity work by 1980, when you started your observing program at the CFHT (Canada-France-Hawaii Telescope) ? At the time, why did you feel you could make an important contribution to the field ? It seems that you correctly assessed that, by the late 70s, several technologies had matured sufficiently to allow a major breakthrough in radial velocity measurements. Can you comment on this ?

[Gordon Walker] – There was a long tradition of measuring radial velocities from photographic spectra by setting the cross hair of a travelling microscope on each absorption line and on comparison emission lines imposed on either side of the stellar spectrum. This was labour intensive and yielded radial velocities with a precision of 1 km/s at best, but that was adequate to help solve the rotational properties of our Milky Way Galaxy since the distances to individual stars were not very certain. Not only were photographic emulsions extremely inefficient – detecting at most 1% of the incoming light – the emulsion would shift during the development process. Electronic detectors had a very much higher detection efficiency and, with the advent of digitising electronics (analog to digital converters), could be directly transformed into the digital images which are so familiar today. In a digital format it was then possible to directly compare the spectra of a star taken on different nights to see how the radial velocity of the star had changed. The challenge was to eliminate systematic shifts caused by the erratic motion of the star image on the spectrograph slit and, in the earliest TV systems, due to jitter in the electron reading beam. The key was to impose fiducial lines right in the stellar spectrum before the light passed into the spectrograph. One could rely on them not to have changed in wavelength and lines caused by (telluric) gases such as water vapour and oxygen in the Earth’s atmosphere fulfilled this condition. Precise radial velocity programs became possible with the introduction of low light level multichannel detectors (first LLTV, then solid state print readers, finally charge coupled devices CCDs), high speed analog to digital converters, and digital computers for the data manipulation. We also benefited from the superb, large, static spectrographs developed at the coudé foci of most large telescopes.

[O foco coudé permite desviar o feixe de luz do telescópio, independentemente da posição deste último, ao longo de um eixo imóvel até um compartimento no observatório. Aí, o feixe de luz é analisado por instrumentos que, pelo seu tamanho, peso ou outros requisitos, não podem ser montados directamente no telescópio. Na imagem vê-se desenhado o percurso da luz até ao foco coudé do telescópio Cassegrain de 3.6 metros do ESO. No foco coudé existe uma sala com o famoso espectrógrafo HARPS (High Accuracy Radial velocity Planet Searcher). Crédito: Vik Dhillon.]

[A célula de fluoreto de hidrogénio no foco coudé do CFHT. Crédito: fotografia cedida por Gordon Walker.]

[AstroPT] – Can you describe the setup you used during the 80s at the CFHT ? When did your team came up with the idea of using a gas cell to imprint a stationary reference spectrum for measuring radial velocities ? Can you describe how the technique works and how the cell is positioned in the light path from telescope to detector ? What precision were you able to achieve using this technique ?

[Gordon Walker] – My contribution was to suggest that we could look for extra-solar planets using telluric lines but Bruce Campbell made the important step of introducing a captive gas such that one could take spectra with and without it – an essential step for the best precision – and impossible with Earth’s atmosphere. He chose hydrogen fluoride in consultation with Gerhard Herzberg and Alex Douglas. The gas was contained in a 90 cm long, heated cell with sapphire windows and it could be raised and lowered into the horizontal coude telescope beam delivered to the spectrograph. On any given night at CFHT we achieved a precision of about 10 m/s and a bit larger from year to year.

[AstroPT] – Most gas cell installations today use (molecular) iodine for the reference spectrum. Why did you choose to go with hydrogen fluoride, especially given its toxicity. What is it about its spectrum that makes it so appealing and worth the risk when compared with iodine ?

[Gordon Walker] – Our observations were made in the pre-CCD era when the best solid state detectors were linear arrays of self-scanned silicon diodes. They were excellent detectors but only available in arrays of up to 1872 diodes – so this restricted us to some 120 Angstrom at the spectral resolution available to us – the HF lines gave a well spaced comb of lines with similar widths to the absorption lines in the stellar spectra. This all made the data analysis comparatively straightforward. With the advent of large, square CCDs and echelle gratings one could stack a whole series of spectra on the detector and cover an order of magnitude more spectrum. For such a setup a quite different gas is required which was why Marcy and Butler chose iodine. These latter elaborations introduced many more complications into the data processing, admirably overcome by Paul Butler, but it did slow their development of the iodine technique.

[A velocidade radial de Gamma Cephei com dados da equipa canadiana e da equipa de Artie Hatzes. Crédito: observatório de McDonald.]

[AstroPT] – In 1987, together with Bruce Campbell and Stephenson Yang, you announced the discovery of planetary companions to Gamma Cephei, Beta Geminorum (Pollux) and Epsilon Eridani. In retrospect, these detections seemed robust and indeed the planets were confirmed later (about 10 years later) by Artie Hatzes’ team with parameters similar to those derived by your team. Yet, by 1992, you essentially retracted the discoveries, claiming that the data was not good enough to make a clear case for any of the planets. What happened ?

[Gordon Walker] – While the Gamma Cephei result looked intriguing in 1988, the data were hardly convincing – there were unknowns in terms of the binary period and the possible stability of such a triple system. In the mean time, we had found that all of the giant stars in our target list had long period (months to years) radial velocity variations – ironically, these had all been designated velocity standards by the International Astronomical Union for the days of photographic radial velocities. Gamma Cephei was designated a giant and so there was a real possibility that the variations were intrinsic to the star. Further, as a by-product of our spectra we were able to measure chromospheric activity. One, the nearby star Kappa-1 Ceti, showed a very marked correlation of its radial velocity with its level of chromospheric activity, very similar to the the solar 11 year sunspot cycle but with a shorter period. Gamma Cephei also showed a weak level of chromospheric activity with a similar period to the radial velocity variations. The paper we wrote in 1992 says `planet or rotation’ in the title – when writing the paper I felt convinced that we had a planet but the presence of the weak chromospheric signal tipped the balance for one colleague whose opinion I respected in favour of rotation. Subsequently, the chromospheric signature proved spurious, the star was reclassified as a sub-giant, and the binary period proved to be significantly longer than we had thought – all of this came from the much longer time base and careful work by Artie Hatzes. In science one goes with the model that appears to work best at the time. It has to be remembered that the planet search program was only one of many astronomical programs we were pursuing at that time. Bruce had left a couple of years earlier and we were faced with the difficult task of re-analysing all of the data.

[A anã vermelha companheira de Gama Cephei em 2007. Crédito: NAOJ.]

[AstroPT] – How did you feel when you received word of the discovery of 51 Pegasi b ? Had you ever considered the possibility that Jupiter-sized planets might exist in such short orbits ? Could you have detected Hot Jupiters in your data or was your observing schedule fine-tuned for finding long period planets ?

[Gordon Walker] – I was one of the referees. I have to admit that at first I was very skeptical given our experience with Gamma Cephei. The Geneva group didn’t use an imposed absorption fiducial spectrum as we, Marcy and Butler, and the Texas group had done but relied on the stability of their spectrograph. But they were right – their spectrograph was fed by an optical fibre which had the effect of scrambling the incoming light which largely eliminated erratic slit illumination by the star. Their data raised a number of questions in my mind but I quickly withdrew my reservations when I saw the velocity curve measured independently by Marcy and Butler. Like the Geneva group our observing runs were limited to three or four pairs of nights per year. We could have detected such a short period Jupiter – but such a system was completely unexpected.

[MOST, o telescópio espacial canadiano especializado em fotometria. Crédito: Projecto MOST.]

[AstroPT] – In recent years you were involved in the design of MOST (Microvariability and Oscillations of STars), Canada’s first space telescope. Can you describe the instrument and its capabilities ? What exactly was your contribution ?

[Gordon Walker] – MOST is a 15 cm aperture white light photometric satellite telescope in a 104 minute orbit. It can stare for weeks at a time at individual stars within a broad band of the sky. The science detector is a CCD and to increase the precision for bright stars, the star is isolated by a pinhole and an image of the telescope entrance pupil in the light of the star is projected onto the CCD. The original goal was to detect the natural frequencies of various stars, particularly those like the Sun. Stars are like musical instruments, they have natural, resonant frequencies and overtones related to the star’s internal structure. From the frequency spectrum one can determine a great deal about the structure and chemistry of the star. While the mission was only planned for two years it still continues the original science objective with an extensive range of other programs. I designed the science experiment and the electronic detectors were built in my lab.

[AstroPT] – You have used MOST to study the interaction between stellar atmospheres and close-in Hot Jupiters. What stars have you studied and what results did you get ?

[Gordon Walker] – I have been involved in several attempts to detect photometric signals from parent stars associated with the orbital periods of their close-in planets. The effects are subtle and the only one of note may be Tau Bootis in which an active spot appears to have tracked its giant jupiter mass planet. The spot is not at the sub-planet point but precedes it by some 60 degrees suggesting that the interaction is magnetohydrodynamic. Although I stepped down from MOST two years ago others are continuing to look for such effects.

Segue-se um resumo da entrevista em português da responsabilidade exclusiva do autor da entrevista.

Gordon Walker despertou para a astronomia em plena 2ª Guerra Mundial, numa noite em que vigorava o “blackout”. Quando voltava para casa com o seu pai perguntou-lhe o que eram as estrelas. O pai respondeu-lhe que cada uma delas era um sol, mas tão distante que era visível apenas como um pequeno ponto de luz. Esse primeiro encontro com as estrelas levou-o a um interesse vitalício pela astronomia e a uma carreira na astrofísica.

Depois de duas décadas de trabalho de investigação, no final dos anos 70, Walker e dois colaboradores, Bruce Campbell e Stephenson Yang, resolveram aplicar as tecnologias digitais emergentes à medição das velocidades radiais das estrelas. Até então este processo era realizado manualmente sobre espectros obtidos em chapas fotográficas e com a ajuda de um microscópio. A precisão das velocidades assim medidas era da ordem de 1 km/s.

Durante a década de 80, a equipa desenvolveu uma técnica de medição que permitiu atingir uma precisão cerca de 100 vezes superior, na ordem dos 10 m/s, suficiente para detectar planetas gigantes em torno de outras estrelas. O equipamento de medição envolvia a introdução de uma célula de vidro com um gás cuidadosamente escolhido no foco coudé de um telescópio, imediatamente antes do feixe entrar num espectrógrafo. Desta forma, sobreposto ao espectro da estrela alvo, aparecia um conjunto de linhas espectrais estacionárias devidas ao gás contido na célula. Estas linhas, com comprimentos de onda bem conhecidos, funcionavam como referências para a medição dos movimentos das linhas no espectro da estrela. O gás escolhido pela equipa, em função de restrições no equipamento, foi o fluoreto de hidrogénio, uma substância altamente tóxica e que corrói inclusivé o vidro comum, daí utilizarem uma célula de vidro de safira. Esta técnica é no essencial idêntica à utilizada por Marcy e Butler anos mais tarde e é actualmente utilizada por muitas outras equipas de investigadores.

No final dos anos 80 a equipa anunciou a descoberta de três planetas em torno de Gama Cephei, Pollux (Beta Geminorum) e Epsilon Eridani. Depois de inúmeras tentativas falhadas de detecção de exoplanetas ao longo das décadas anteriores, a comunidade científica encarou com suspeição este anúncio. Havia também algumas incógnitas relativamente às propriedades das estrelas hospedeiras que tornaram a análise dos dados complicada. Em 1992, já sem Campbell, e numa demonstração de grande honestidade científica, a equipa escreveu um artigo em que dizia não poder inequivocamente demonstrar que as variações detectadas na velocidade radial das estrelas resultavam da presença de companheiros planetários.

Em 1995, Gordon Walker foi um dos revisores do artigo de Michel Mayor e Didier Queloz, submetido à revista Nature, dando conta da descoberta do 51 Pegasi b. Dada a natureza do resultado, Walker teve algumas reservas iniciais que rapidamente foram dissipadas pela análise do método experimental utilizado pela equipa suiça, pela qualidade dos dados e, em especial, pela imediata confirmação das observações por Marcy e Butler. Walker refere que, tal como a equipa suiça, a sua equipa tinha apenas acesso ao CFHT alguns pares de noites por ano, e poderia ter detectado Júpiteres Quentes, caso existissem em torno das estrelas no seu programa de observação. No entanto, acrescenta, ninguém esperava encontrar planetas gigantes numa tal configuração orbital.

Recentemente, Walker fez parte da equipa que desenvolveu o telescópio MOST, o primeiro telescópio espacial canadiano, do tamanho de uma mala de viagem. O pequeno satélite tem um telescópio de apenas 15 cm de diâmetro e pode observar o mesmo objecto, ininterruptamente, durante semanas, com uma precisão fotométrica excelente. O objectivo principal da missão era o de estudar as oscilações das estrelas usando técnicas da asterosismologia. As frequências e sobretons observados estão intimamente relacionados com a estrutura interna e densidade média das estrelas. Walker foi responsável pelo conceito da missão e os detectores utilizados pelo MOST foram construídos no seu laboratório. Enquanto cientista da missão, utilizou ainda o telescópio para realizar observações de estrelas hospedeiras de Júpiteres Quentes, na tentativa de identificar casos em que a actividade cromosférica da estrela pudesse ser relacionada com a órbita planetária. Tal indicaria uma possível interacção entre os campos magnéticos da estrela e do planeta. Walker refere que são observações difíceis de realizar e que actualmente o caso mais convincente é o da interacção entre Tau Bootis e o Júpiter Quente que a orbita.

Deixe uma resposta

O seu endereço de email não será publicado.

Este site utiliza o Akismet para reduzir spam. Fica a saber como são processados os dados dos comentários.