The 1998 C.S. Beals lecture:
CHALLENGES FOR THE NEW MILLENNIUM

I had the privilege of knowing Carl Beals in the early 1960's when he was Dominion Astronomer and I was a Post Doctoral Fellow at the Dominion Astrophysical Observatory. He had an abiding interest in the nature of the Diffuse Interstellar Bands (DIB), having established in the late 1930's that the strongest of them at 4430Ź was interstellar. As the DIB had been the preoccupation of my PhD thesis, we often talked about them.

As many as 200 such diffuse features have now been detected in spectra of stars seen through clouds of interstellar dust. They vary from a few, to tens, of Ångstrom in width2. Figure 1, taken from 4, shows several strong DIB (including a broad one at 5778Å) in the yellow spectral region of several stars lightly obscured by single interstellar dust clouds. The source of the DIBs must be intimately connected with this dust because their intensities correlate so closely with it. They are too broad to be resonance lines of simple molecules, atoms or ions, and they are unlikely to arise in the grains themselves because, as van de Hulst3 pointed out in the 1940's, the anomalous dispersion introduced by the absorption within the grains would result in an apparent blue emission wing to each line, something never convincingly seen. This impasse opened the door for a wide range of speculation, largely about complex molecules, perhaps precursors of the graphite in interstellar grains. My modest contribution to understanding this puzzle was a collaboration with Jacek Krelowski in 19874. We demonstrated the existence of three families among the stronger DIB, listed in Table 1, within each of which the members vary in unison. The distinction between the families seemed particularly marked between single clouds where: can be weak or even absent (especially in regions of high UV-flux), can be weak (sometimes 1/3 normal), is unchanging. Some of these variations are quite obvious in Figure 1.

Figure 1 Spectra of five B-stars, each lightly obscured by dust in single interstellar clouds. The narrower DIB are identified by wavelength. There is also a broad DIB centred at 5778Å. The other lines arise in the stars atmosphere.

Figure 2

A composite spectrum of the `Red Rectangle' nebula (shown in Figure 3). A linear background, probably caused by hot dust grains, has been removed. The four arrows indicate the emission features which coincide with DIB family 3 of Table 1.

Figure 3 An infrared (1.65 micron image of the `Red Rectangle' taken with the University of Hawaii adaptive optics system on the CFH telescope6. The tick marks corresponds to 0.1 arcsec and the lines on the right outline the likely hollow bi-cone, thick disc, dust distribution.

An image of the Red Rectangle taken at 1.65 micron by the Roddiers6 with their adaptive optics system on CFHT is shown in Figure 3 together with a model of its likely cloud structure. A ninth magnitude A0 sub-giant, HD 44179, is immersed in the small X-shaped nebulosity which appeared rectangular in photographic images - hence the name. The nebulosity consists largely of carbonaceous material in a hollow bi-cone and an obscuring disc which almost certainly emanates from a close, unseen companion carbon star smoking like a giant candle.

The association of DIBs, apparently in emission, with clouds of complex organic molecules seemed to confirm the suggestion made by Douglas⁷ in 1977 that long Carbon chains might be responsible for at least some of the DIB. The Radio Astronomers at the Herzberg Institute using the 46-m Algonquin Radio telescope had discovered a series of polyacetylenes^8,9 viz. HC₅N, HC₇N, and HC₉N, with other groups finding HC₃N, and HC₁₁N. Douglas pointed out that these molecules are large enough to be stable against photodissociation from ultra-violet photons by internal conversion.

The most generally accepted source of the bands now seems to be an increasing complexity of graphite precursors from chains to polycyclic aromomatic hydrocarbons (PAH)¹³, to fullerenes¹⁴, which could well form a series of increasing complexity all the way to the fine, graphite particles responsible for interstellar extinction^10,11. The greatly improving skill of laboratory groups in isolating various species by mass and measuring their spectra, has resulted in coincidences with the DIB and infrared-emission bands being presented with increasing confidence. Allamandola¹³ and his colleagues have shown a series of coincidences with PAH cation spectra of such substances as Napthalene and Coronene which can also explain certain interstellar infrared emission bands.

The laboratory spectroscopic group at Basle under John Maier has demonstrated intriguing coincidences between laboratory spectra of the long Carbon-chain anion, C₇^-, and several DIB where the DIB have similar widths and relative intensities to the lab values¹².

The intensity of a DIB family may well depend on the degree of hydrogenation or the formation of ions and species containing Nitrogen, Oxygen etc., because the strengths of the different families seem to be related to the relative abundance of different interstellar molecules, particularly H₂. Jacek Krelowski and I hope that the high sensitivity of FUSE in the measurement of H₂ (and many other species) in single clouds will provide some important clues.

Maybe we shall see closure on the mystery of the diffuse interstellar bands in the next few years.

The First Glimpse of an Extra-Solar Planet:

direct imaging

Let me turn to another long-standing challenge, the direct detection of extra-solar planets. To the best of my knowledge, the only evidence of William Shakespeare is the existence of a few of his signatures. There is no contemporary portrait. Somewhat reminiscent of the situation for the short-period extra-solar system planets - where there are very convincing dynamical signatures for nine short-period planets in the apparent reflex radial velocities of nine Solar-type stars and two milli-second pulsars. But the only convincing portrait of a sub-stellar companion is of the cool brown dwarf close to the M flare-star, Gliese 229 at 5.7 parsecs¹⁵.

I have recently tried with my Montreal colleagues, René Doyon, René Racine, Daniel Nadeau and Philippe Valleé, and Scott Chapman at UBC to make direct detections of any other cool brown dwarf companions to nearby stars using a sensitive dual imaging, double filter technique with the University of Montreal IR cameras and the adaptive optics system on the Canada France Hawaii 3.6-m telescope, so far without success. Cool brown dwarfs seem rare indeed, much rarer than white dwarfs.

Our technique of simultaneously imaging through closely spaced filters exploits the strong distortion by methane absorption in brown dwarf and giant planet spectra while defeating the serious speckle shot-noise introduced by the Earth's atmosphere. The latter is critically important for ground-based observations. Figure 4 shows a low resolution spectrum near 1.6 micron of Gliese 229B and Saturn's satellite, Titan. They are closely similar, with a large absorption decrement caused by CH₄, Jupiter (not shown) is similar. The filter band passes which we selected are shown on, and off, CH₄.

Figure 4

Low resolution spectra showing the methane absorption band at 1.6 micron for the cool brown dwarf, Gliese 229B, and Titan, a satellite of Saturn with a methane atmosphere. The band-passes of the two filters used in our imaging are shown.

Figure 5 shows in spectacular style, how effective our technique is. The left hand image is taken through the CH4 filter while that on the right is in the continuum. The right hand image shows the brown dwarf in high contrast. In fact, Figure 5 is a composite of a field with two faint stars in Orion and another of Gliese 229B. If we had made a real discovery, I would have devoted most of my lecture to it! An even greater challenge is to detect planets directly. My Montreal colleagues and I feel that our technique, particularly in its ability to defeat speckle noise, has great potential and plans are afoot to equip KIR with a dual imaging coronagraphic module.

Figure 5 A composite dual image of a field in Orion and the known brown dwarf Gliese 229B. The left hand image is taken through the CH4 filter - the right hand in the continuum. Judicious selection of contrast in SAO Image really makes the brown dwarf stand out on the right.

spectra

Figure 6 The near-infrared spectrum of a hot giant planet such as tau Boötis B (Te=1600 K) compared to that of the Sun. The solid lines are for an equivalent black-body.

A conservative estimate of sensitivity suggests that with 100 nm spectral coverage near 1.6 micron on the CFHT and an hour's exposure per night, we could easily detect the spectrum of the secondary phase-sensitively over a single cycle of the 3.3 night orbit.

Someone is bound to succeed with this soon.

What about detection of a habitable satellite?

Just over a year ago, Williams, Kasting and Wade19 made the provocative suggestion that extra-solar giant planets might have habitable moons if they lie in, or pass through, the habitable zone of the parent star. So far, radial velocity studies appear to rule out any giant planets within the habitable zones of any nearby stars other than 16 Cygnus B20 or 47 UMa B21. Since their spectra will be dominated by reflected starlight, they would be some 12 mag less luminous than companions such as tau Boötis B, making detection of their spectra impossible, let alone looking for second order perturbations.

But surely, Gliese 229B, at 7.2 arcsec from 229A, is largely uncontaminated by light from A, which makes it a much easier candidate to test right away. Currently the accepted physical elements for B are: radius = 7 x 10⁴ km (very similar to Jupiter), effective temperature 1000 K, and mass 7 x 10²⁸ kg. The radius of its habitable zone, 7 x 10⁵ km, is only ten times the radius of the brown dwarf. At this distance the satellite would have a period of only about 14 hours with a velocity amplitude of 75 km s^-1. The low mass of the brown dwarf results in a reflex motion induced by a Jupiter-mass satellite of some ± 2 km s^-1, which should not be difficult to detect for a m_H = 14 star! Indeed, as one can see from Table 2 one could detect satellites all the way down to an Earth mass without too much trouble.

Notice that these observations do not demand the high precision and long-term stability of precise radial velocity programs because of the very short period involved. The relevant accelerations are given in m s^-1 hr^-1 under Delta RV_max in the third column.

The likelihood of such short period satellites of the right mass seems high with appropriate scaling by primary mass from Jupiter's Gallilean, and the Saturnian satellites and the seven short period planetary companions already implied for other stars.

If there is an `Earth' revolving around Gleise 229B, its rotation would probably be locked at the same period which means that ocean tides would be tens of kilometres high, but fairly static. While Gleise 229B would subtend some 11° in the sky and provide the warmth, `daylight' about ten times brighter than our full moon would come from Gleise 229A, 44 astronomical units away.

So, if someone does discover a habitable satellite around Gleise 229B, remember you heard it here first!

acknowledgement: very many thanks to Gerry Grieve for all his help in preparing this article for publication.

¹ Beals CS, & Blanchet GH, 1938, MNRAS, 98, 398-407

² Herbig GH, 1995, Ann. Rev. Astrophys., 33, 19-73

³ van de Hulst HC, 1949, Recherches Ast. d'Utrecht, 11, pt. 2

⁴ Krelowski J, & Walker GAH, 1987 Ap. J., 312, 860-67

⁵ Scarrott SM, Watkin S, Miles JR, & Sarre PJ, 1992, MNRAS, 255, 11P-16P

⁶ Roddier F, Roddier C, Graves JE, & Northcott MJ, 1995, Ap. J., 443, 249-60

⁷ Douglas AE, 1977, Nature, 269, 130-2

⁸ Avery LW, Broten NW, Macleod JM, Oka T, & Kroto HW, 1976, Ap. J., 205, L173-L175

⁹ Kroto HW, 1981, Int. Rev. Phys. Chem., 1, 309

¹⁰ Webster A, 1993, MNRAS, 262, 831-38

¹¹ Maier JP, 1997, Chem. Soc. Rev., 26, 21-26

¹² Kirkwood DA, Linnartz H, Grutter M, Dopfer O, Motylewski TD, Pachkov M, Tulej M, Wyss M, & Maier JP, 1998, Faraday Transactions, 1998 in press

¹³ Allamandola L, 1998, http:www-space.arc.nasa.gov/ astrochem/PAHs.html

¹⁴ Foing BH, & Ehrenfreund P, 1997, Astron. Astrophys., 317, L59-L62

¹⁵ Nakajima T, Oppenheimer BR, Kulkarni SR, Golmowski DA, Matthews K, & Durrance ST, 1995, Nature, 378, 463

¹⁶ Mayor M, & Queloz D, 1995, Nature, 378, 355-9

¹⁷ Butler P, Marcy G, Williams E, Hauser H, & Shirts P, 1997, Ap.J Lett, 474, L115

¹⁸ Hauschildt PH, Allard F, 1998, http://dilbert.physast.uga.edu/ yeti/

¹⁹ Williams DM, Kasting JF, & Wade RA, 1997, Nature, 385, 234-235

²⁰ Cochran W, Hatzes A, Butler P, & Marcy G, 1997, Ap.J., 483, 457

²¹ Butler P, & Marcy G, 1996, Ap.J. Lett., 464, L153-6

Gordon Walker      <walker@astro.ubc.ca>   When I was seven years old, my father aroused a lifetime interest in Astronomy by telling me about the stars. After a BSc in Natural Philosophy from Edinburgh University (1954) and a PhD from Cambridge (1962), I spent seven years at DAO before coming to UBC (1969). I retired in 1997. Observational techniques have always fascinated me and my lab developed many instruments, particularly detection systems. I was closely involved in the realisation of the CFH and Gemini large telescopes, and Starlab, a space telescope which never happened. My current research: interstellar dust, nonradial pulsations, and the search for extra-solar planets.