NUMBER 82 JULY 1997                                 NEWSLETTER


On June 8-9 the spectrum shown above was obtained using the Anglo-Australian Telescope of a brown dwarf candidate identified by the DEep Near Infrared Southern sky survey. The clear lithium detection indicates a mass of less than 0.06 solar masses — making it unquestionably a brown dwarf!

Director's message

AAT confirms DENIS brown dwarf study (Chris Tinney et al.)

The enigma of the red quasars (Paul Francis et al.)

Raman scattering in symbiotic stars (Tim Harries & Ian Howarth.)

Optical spectroscopy of supersoft X-ray sources (Phil Charles et al.)

AAT news

Schmidt news



Brian Boyle

As we move into semester 97B, there is a great deal of encouraging progress to report at the AAO. The success of 2dF commissioning over the past few months has brought us to the point where we are able to offer the full 400 fibre, 2 spectrograph system to the UK and Australian users from September. As reported by Karl Glazebrook and Ian Lewis in this newsletter, the growing reliability of the 2dF system has also allowed the AAO to offer 2dF to specific observing programs on a `shared risk` basis. The move away from `commissioning' mode for scientific observations is a significant step forward for 2dF, and should result in 2dF beginning to realise its full scientific potential in advance of a fully commissioned system in 1998.

It is also heartening to see the scientific productivity of the AAO in other areas. In this newsletter, there are four reports of successful programs carried out on the `old workhorse' — the RGO spectrograph. The discovery of a brown dwarf from the DENIS survey by Tinney et al. is particularly exciting, with the promise of more to come. Programs as varied as the search for red quasars and Raman scattering in symbiotic stars typify the AAO's ongoing commitment to supporting as wide a range of competitive scientific programs as possible and demonstrate the central role that Cassegrain spectroscopy still plays at the AAO.

IRIS-2 is progressing with its preliminary design review phase, with the review currently scheduled for the last week in September. The optical design has been received from Damien Jones and the mechanical design is currently in progress. Full details of the proposed scientific capabilities of IRIS-2 are provided by Stuart Lumsden in this newsletter.

There was a very encouraging response from both the UK and Australian communities to the announcement of the availability of the MPE 3D instrument on the AAT. As a result, 3D has been scheduled for 27 nights during semester 97B, over two lunations in November and I am sure that the collaboration between the MPE team and the AAO will prove extremely productive for all concerned.

As reported by Quentin Parker in this newsletter, the UK Schmidt H alpha survey has also begun. At their last meeting, the Schmidt Telescope Panel and the H alpha consortium agreed to upgrade the program to full survey status, thus assuring high priority and rigorous quality control for the survey. The data will be made available to the community and there will no proprietary period for any survey data.

We have also continued to upgrade our support service to observers. Articles in this newsletter by Chris Tinney and Ray Stathakis highlight the provision of new data-assessment routines and software tools (the QIKLOOK package for CCD imaging, AATRGO for spectroscopy) and the move towards the provision of a Web-accessible archive and database. However, these initiatives do use significant staff resources and it is inevitable that, in devoting resources to upgrading, some existing services can no longer be maintained at current levels. For example, we will no longer offer a tape-copying service to observers, and archive requests which require complex searches of the existing tape-based archive will be given a low priority. We trust that observers will be sympathetic towards these moves, since they are designed to enable the AAO to provide an archive/database system suitable for a world-class optical observatory, particularly one specialising in wide-field astronomy, well into the 21st century.


Stuart Lumsden

The AAO is pleased to announce the continuing availability of the MPE infrared spectrometer 3D at the AAT for semester 98A. 3D will only be available for two lunations, in the first half of the semester. For the earlier run, one or more of the high resolution grisms will be installed (it is unlikely however that the long K grism will be used at this time of year). For the second lunation, the choice of grisms has not yet been made. This choice depends, in part, on the requests made in the accepted programs. Proposers should state which alternative grisms would be useful for their program, in the event that their first choice grism(s) is (are) not available.

We will accept proposals for half nights, to allow programs involving the Galactic centre the chance of observing time. These will be balanced by suitable projects seeking access to earlier RAs by the scheduler in consultation with MPE after the time allocation committees have allocated time.

Full details of 3D can be found at Details of the original agreement between the AAO and MPE regarding the usage of 3D at the AAT can be found at


Chris Tinney (AAO), Xavier Delfosse and 
Thierry Forveille (Grenoble)

Brown dwarfs — "stellar" objects too small to burn hydrogen — have been predicted to exist now for over thirty years. They were first proposed by Shiv Kumar in 1963 (ApJ,137,1121). (Though they were called in that paper "black" dwarfs; Jill Tarter is generally credited with coining the term "brown dwarf" in her 1975 PhD thesis.) However, since then actually claiming a brown dwarf detection has turned out to be a dangerous activity. The journals are littered with brown dwarf detection claims and brown dwarf candidates which have subsequently turned out not to hold water.

There turned out to be a few major problems in finding brown dwarfs. Firstly they are very faint — it is really only in the last few years that telescopes and instruments have become sensitive enough to sample sufficient volumes to really expect to find brown dwarfs. The second problem is that it is hard to tell an object right at the star-brown dwarf boundary (about 0.08 solar masses) from a more massive star. Because brown dwarfs cool, their luminosity decreases with time. So while young brown dwarfs are easier to find, they have similar luminosities and spectra to stars, making them hard to distinguish.

To claim a really solid brown dwarf detection you need to either find something so faint it can't possibly be a star, or satisfy the even more stringent "lithium test". This test relies on the fact that lithium is destroyed at central temperatures above 2x106K. Since stars burn hydrogen only above 3x106K, and low mass stars are fully mixed by convection, finding a low mass star with lithium guarantees a mass of less than 0.06 solar masses.

Both these requirements were dramatically filled for the first time in 1995, when Basri, Marcy & Graham (ApJ, 458, 600) announced the detection of Li in a faint Pleiad, and Nakajima et al. (Nature, 378, 463) announced the discovery of a companion to the star Gl 299, which was far too low in luminosity to be a star itself. The latter object was subsequently shown to display methane in its infrared spectrum — clearly indicating an effective temperature near 1000K. However, these objects left several of the leading brown dwarf questions unanswered. In particular, they didn't tell us what an old (i.e. ~1 Gyr) brown dwarf near the star/brown dwarf boundary looks like. Gl 299B is clearly either very old or very low in mass, and the Pleiades brown dwarfs are all ~100Myr old. This has been reflected in the still ambivalent status of one of the leading brown dwarf candidates for many years — GD 165B. First announced as a brown dwarf candidate in 1988 (Nature, 336, 656), GD 165B has remained mysterious. Its infrared spectrum looks similar to that of a low-mass star — however, its optical spectrum looks like no low-mass star known. And its bright companion (the white dwarf GD 165A) has prevented a lithium detection. So we still can't say what the objects we are most likely to find in a brown dwarf survey (i.e. field brown dwarfs, just below the hydrogen burning limit) will look like.

These surveys are now being carried out. Both the DEep Near Infrared Southern sky survey (DENIS) and the Two Micron All-Sky Survey (2MASS) will survey the entire southern hemisphere in the infrared to a sufficient depth to find brown dwarfs. The first set of brown dwarf candidates has already been produced by DENIS. Observations of these objects were carried out in March and October 1996 using IRIS on the AAT, and were able to confirm that at least three of the objects are as faint as, or fainter than, the lowest mass stars known
(c.f. ). In June 1997, the clinching observational test (searching for lithium) was carried out for two of these objects using the RGO Spectrograph. The result was the spectacular detection of a 0.23nm equivalent-width Li I 670.8nm line in one object (the largest Li detection in any low mass dwarf), and a non-detection in the other. The Li detection means that DENISP J1228-1547 has a mass less than 0.065 solar masses, and is clearly a brown dwarf, though its effective temperature is still significantly higher than that of GD 165B (c.f. ).

Moreover, it shows an optical spectrum which is remarkably similar to GD 165B. The same lack of TiO and VO bands is seen, and the same strong lines of Cs and Rb I are seen. This implies that GD 165B is not a "bizarre" object due to any effects of mass loss in its white dwarf companion — this is simply what objects with Teff ~ 1500K look like. Even more excitingly, the candidate list for this study was based on less than 1 percent of the DENIS survey data! It is clear that many more results, and a fundamentally improved understanding of the density of brown dwarfs can be expected as this important survey proceeds.



Paul Francis (ANU), Rachel Webster (Melbourne), Michael Drinkwater (UNSW),
 Frank Masci (Melbourne) and Bruce Peterson (Mt Stromlo)


Quasars are blue; everyone knows that. This is, after all, how most quasars are found. In this article, however, we describe our discovery that at least one type of quasar can have extraordinarily red optical/near-IR colours, and our ongoing attempts to understand why.

This project started out as an attempt to compile a complete sample of radio-selected quasars, in order to do a gravitational lensing survey. We used as our basis the venerable Parkes half-Jansky (2.7 GHz) survey of the southern sky, and defined a sub-sample of 323 sources with flat radio spectra. By comparing accurate radio positions with COSMOS scans of AAO Schmidt plates, and building upon the decades-long work of Ann Savage and collaborators, we were able to identify around 90% of the sample, mostly as quasars with 0 < z < 3.8 (Drinkwater et al. 1996).

The remaining 10% of the sample, however, had no counterpart on the Schmidt plates down to B=22.5. In an attempt to identify these remaining "empty field" sources, we imaged the radio coordinates with IRIS. Somewhat to our surprise, remarkably bright (K~16) point sources were seen at virtually all the radio positions. We were unaware at the time of the pioneering work of Lebofsky, Rieke & Walsh (1983), and of Wright, Ables & Allen (1983) both of whom had found the same result a decade earlier.

Had we discovered a new class of extremely red (B-K>6) radio sources? Further IR observations showed that the whole sample had an extremely large dispersion in its optical/near-IR colours (Fig. 1), and that the empty field sources were simply the tail of this distribution (Webster et al. 1995).


The enormous range of B - K colours shown by our radio selected sample contrasts dramatically with the much narrower colour distribution of optically selected QSO surveys, which all have colours near the blue envelope of our colour distribution (Fig. 1). Why are our sources so different? A number of theories have been advanced:

• Host galaxy light: an elliptical host galaxy would have very red B-K colours. Any host galaxy would, however, have to be far more luminous than known radio galaxies to dominate the light from our sources.

• Synchrotron emission: our sample is selected on the basis of a flat radio spectrum, and many such sources are thought to be dominated in the optical by relativistically beamed synchrotron emission. If this synchrotron emission is red, and is much stronger than other emission sources, it could produce the red colours (Serjeant & Rawlings 1996).

• Dust: these could be normal quasars obscured by dust somewhere along the line of sight. The absence of red optically selected QSOs could be a selection effect: if dusty radio-quiet QSOs exist, they will be enormously under-represented in any survey with an optical magnitude limit, due to the steepness of the QSO luminosity function.

Generally speaking, quasar pundits preferred the synchrotron model (these are flat spectrum sources, after all), while most other astronomers preferred the dust model (is it really surprising that things that live in the middle of galaxies should be dusty?).

Figure 1: The distribution of B-K colours for our sample (top panel) and for a sample of optically selected QSOs (bottom panel, from Francis 1997).


Figure 2: Balmer decrements as a function of B-K colour; note that the reddest sources have large Balmer decrements, indicating substantial reddening.

The Spectra of Red Quasars

An obvious test of these models was to obtain spectra of the red sources. Most, alas, are very faint (B>22), but the RGO spectrograph/FORS combination, ably assisted by the AAT's superbly precise blind offsetting, proved well up to the task, and slowly, painfully, we obtained spectra of many of the red sources.

Few, if any, are galaxies; most show broad emission lines, and few show any sign of a 4000Å break or of any other stellar spectral features. We are able to show that host galaxy light is not a significant contributor to the wide observed range of B-K colours.

The spectra of the Parkes sources seem to fall into three loose (and overlapping) classes:

(1) Normal blue (B-K ~ 2.5) quasars (~40% of the sample); (2) Moderately red (B-K ~ 4) quasars with low equivalent width broad emission lines (~30% of the sample); a few BL Lac objects form the extreme of this population; (3) Very red (B-K > 4) quasars with reddened broad emission-line ratios (Figure 2; ~30% of the sample)

If you take a class 1 spectrum and add a featureless red synchrotron emission component, you obtain a class 2 spectrum, while if you redden a class 1 spectrum with some standard extinction law, you obtain a class 3 spectrum. Thus it tentatively seems that both the synchrotron and dust models are true for different sub-sets of the sample. (Francis et al. in preparation).


Is this two population model valid? In principle, measurements of both the optical/IR spectral energy distribution, and the wavelength dependence of polarisation in the near IR, have the power to discriminate between the different reddening mechanisms, and projects are underway to determine both.

If some of our quasars are indeed dusty, where is the dust? Our statistics are currently too poor to tell for certain, but there are intriguing signs that the dustiness of our population may evolve with time, higher redshift quasars being less dusty. This would suggest that the dust is in the quasar host and that we may be seeing chemical evolution in progress. Signs of a comparably dusty population of radio-quiet QSOs are now being found in a variety of X-ray surveys (e.g. Puchnarewicz et al. 1996); the population of such hitherto undetected QSOs is potentially very large.

Finally, can we trace a continuous progression of properties between the BL Lac objects in our sample and the normal blue quasars? Can we determine the spectrum of the synchrotron component, and what physical mechanism makes it so red? Watch this space…


Drinkwater, M. J., Webster, R. L., Francis, P. J., Condon, J. J., Ellison, S. L., Jauncey, D. L., Lowell, J., Peterson, B. A. & Savage, A. 1997, MNRAS 284, 85

Francis, P. J. 1997, Publ. A.S.A., 13, 212

Lebofsky, M. J., Rieke, G. H. & Walsh, D. 1983, MNRAS 203, 727

Puchnarewicz, E. M., Mason, K. O., Romero-Colmenero, E., Carrera, F. J., Hasinger, G., McMahon, R., Mittaz, J. P. D., Page, M. J. & Carballo, R. 1996, MNRAS, 281, 1243

Serjeant, S. & Rawlings, S. 1996, Nature, 379, 304

Webster, R. L., Francis, P. J., Peterson, B. A., Drinkwater, M. J. & Masci, F. J. 1995, Nature, 375, 469

Wright, A. E., Ables, J. G. & Allen, D. A. 1983, MNRAS 205, 793


Tim Harries (St Andrews) and Ian Howarth (UCL)


Symbiotic stars are binary systems consisting of a hot star which ionizes material shed by a cool, giant companion, resulting in a characteristic "combination spectrum" of sharp, nebular emission lines and stellar molecular absorption bands. The binary periods are typically of the order of years and longer, and hence accurate orbital parameters have been established for only a handful of systems. Furthermore, the mechanisms that drive the mass outflows from red giant stars are poorly understood. It is therefore highly desirable to obtain observational constraints on both the binary geometry and wind structure of these systems.

In a seminal paper Schmid (1989) attributed two previously unidentified nebular lines at ~wavelengths 6825,7080Å to Raman scattering of resonance-line O vi photons (wavelengths 1032,1038 Å), originating in the ionized nebula, off neutral hydrogen associated with the cool giant. Raman scattering is the inelastic analogue of Rayleigh scattering; the frequency of the scattered photon is different to that of the incident photon and the scatterer is left in an altered quantum-mechanical state. Raman scattering occurs according to the Rayleigh scattering phase matrix, and since the scattering geometry is markedly asymmetric (as photons from the nebula are scattered in the locale of the red giant) the emission lines show significant linear polarization.


We have used the RGO spectrograph (25cm camera and 1200R grating) in combination with the polarimetry module (which comprises a calcite analyzer and super-achromatic half-wave plate modulator) in order to obtain high-resolution, high-S/N spectropolarimetric observations of 28 Galactic symbiotic systems. (This sample represents rather more than half those known to display the Raman lines). The observational setup fully resolves the line profiles, since as a bonus of the Raman scattering process, the effective velocity resolution is increased by a factor of 7 (~6825/1032) over `normal' lines. In fact, the velocity resolutions we obtain (~6 kms-1) are comparable to UCLES! We have obtained further observations in follow-up campaigns in semesters 95A+B and 97A+B.


Figure 1: Observations of the l6825Å line in V455 Sco and H2-38. The data shown here have been corrected for interstellar polarizations, and are binned to a constant error of 0.5% in P.


Our 1994 survey (Harries & Howarth 1996b) revealed that:

• The line polarizations are in general quite large (up to ~10%), with the strongest and most highly polarized lines found in D-type systems (i.e., those showing dust signatures in the IR, with Mira-type cool components).

• The intensity and polarization morphologies are often complex in nature, with multi-peaked polarization profiles observed across a broad range of spectral types. The commonest structure is a triple-peaked line.

• The Raman lines show a systematic redshift, typically by 200-300 kms-1.

• A position angle discontinuity (often of 90o) is seen in about half the stars observed. Many of the objects display a rotation of the position angle through the line.

The diversity of polarized line profile morphologies is illustrated in Figure 1. In particular note the 90o PA flip in V455 Sco, the monotonic PA rotation in H2-38, and the change in the relative strengths of the polarization peaks.

The position angle (PA) of the polarization of the blue wing of the Raman lines (which is formed in the region between binary components, where the red giant wind is approaching the O vi photon source) is perpendicular to the plane containing the observer's line-of-sight and the binary line-of-centres. Binary parameters may be determined by measuring the orbital variation in this PA, in a method analogous to that used for visual binaries. Figure 2 shows the spectropolarimetric orbit of SY Mus, a relatively short period (~600d) system. From observations obtained in 1992, 1994, and 1995 (two measurements) we are able to derive an ephemeris and inclination that are in excellent agreement with independent photometric and spectroscopic measurements (Harries & Howarth 1996a).

Future work

Additional observations of the target systems were obtained in 97A and 97B, and combining these with the excellent results of our 1994 and 1995 runs we will be able to compute ephemerides and inclinations of the majority of the shorter period systems in our sample, and provide estimates of the periods of the wider (P> 50 yr) binaries. Combining our ephemerides with spectroscopic mass functions we will be able to gauge the relative importance of Roche lobe overflow and wind-fed mass-transfer in producing symbiotic phenomena.

We are also making progress in modelling the polarized line profiles, which we have found are sensitive to the mass-loss rate and velocity law of the cool giant wind (Harries & Howarth 1997). Observations at different orbital phases will allow us to isolate the geometrical and physical variables and will enable us to construct models `tailored' to particular systems, and hence to estimate mass-loss rates and the extent and shape of the ionized region.


Schmid, H.M. 1989, A&A, 221, L31

Harries, T.J., Howarth, I.D. 1996a, A&A, 310, 235

Harries, T.J., Howarth, I.D. 1996b, A&AS, 119, 61

Harries, T.J., Howarth, I.D. 1997, A&AS, 121, 15


Figure 2: The polarimetric orbit of SY Mus. The open circle shows the location of the O vi source (hot component), the large shaded circle the location (and approximate size) of the red giant at phase zero, and filled dots every 0.01 P. Short thick lines show the calculated position angles, and long thin lines the measured PAs. The arrow indicates the direction of motion for the nearer half of the orbit.


Phil Charles, Karen Southwell (Oxford)  and Mario Livio (STScI)


The "supersoft X-ray sources'' (SSS) were discovered in the LMC by the Einstein Observatory about 20 years ago. Their very high (near Eddington-limited) luminosities, but extremely low blackbody temperatures (a few tens of eV), marked them as a highly unusual member of the group of low-mass X-ray binaries (LMXBs).

Initially considered to be potential black-hole or neutron star candidates, the model which has come to prevail establishes them as quite different objects. Unlike regular LMXBs, the SSS are believed to contain white dwarf accretors which support (quasi) steady hydrogen burning at their surfaces. This is possible for a limited regime of mass transfer rates (~10-7 solar masses yr-1), which are achievable if the donor star mass exceeds that of the white dwarf, so that the mass transfer proceeds on a thermal timescale. However, this interpretation has never been confirmed and so, in order to derive constraints on the nature of the SSS compact object, we have been using the RGO spectrograph on the AAT to undertake intermediate resolution (FWHM ~ 1.3Å) optical studies of the SSS. Our aim was to test this model by using spectroscopy to explore the binary parameters.

Observations and Interpretation

In spite of very poor weather conditions during two out of three runs, we obtained blue spectra of the two brightest LMC SSSs, CAL 83 and RX J0513-69, in Dec 94 and Jan 97 (see Fig. 1). Clearly, there are many similarities, both spectra exhibiting strong emission lines of He ii and the Balmer series, indicative of high luminosity LMXBs.

RX J0513-69

The spectrum of RX J0513-69 displays additional features marked S+ and S-. These were first noted by Cowley and Pakull, and interpreted as highly collimated outflows of He II 4686 and H beta. Clearly, the strength of the features is variable, since they are almost absent in the 1/97 spectrum. More importantly, by measuring the velocities of the outflow (or jet) components, and associating this with an escape velocity, we have been able to show from these data that the compact object is almost certainly a white dwarf. Moreover, a red spectrum of RX J0513-69 obtained at this time indicated that the S- component of H alpha appeared in absorption. This complex behaviour is not yet fully understood, but may perhaps result from a precession of the disk, and hence also the jets.

CAL 83 (bottom curve)

The spectrum of CAL 83, whilst lacking such sharply collimated outflows, does show highly asymmetric profiles with excess red emission. This may be a broadly collimated outflow, since the emission also sometimes appears as a blue wing (e.g. Crampton et al. 1987, ApJ 321, 745). We have suggested a radiation-induced disc warping as a possible precession mechanism (Southwell, Livio & Pringle 1997, ApJL 478, L29).


Whilst we have insufficient spectra for compiling good radial velocity curves for the SSS, we have made good progress in advancing our understanding of these systems. The outflow components in the spectrum of RX J0513-69 lend strong observational support for the white dwarf model for the SSSs. Moreover, comparison of the spectra of RX J0513-69 and CAL 83 have played a role in enabling us to develop a general framework for the SSSs, in which individual source characteristics are explicable through subtle differences in the mass accretion rate.


Figure 1: Observations of RX J0513-69 (top 2 curves) and CAL 83 (bottom curve).


Karl Glazebrook and Ian Lewis


The first three nights of the July 2dF commissioning run were totally lost to weather, part of the fourth night was lost to foggy conditions, and the last four nights were clear (less the equivalent of half a night for fog) but had highly variable seeing (<1.0" to >3.0").


No fundamental hardware problems were experienced with either the spectrograph or positioner, just minor unreliabilities. However these minor problems seemed to aggravate problems in the drama software which caused us a major loss of time during the 4-6th nights.

Extensive testing of the drama software during the first 6 nights enabled Tony Farrell and Keith Shortridge to release a final version of the software which ran without a single failure for the last two nights. No positioner hardware failures occurred during the last two nights either.

The servo performance was thoroughly investigated during the first three afternoons/nights, yielding detailed information regarding the speeding up of the positioner required to operate with 400 fibres; this investigation showed some encouraging results.

In general the positioner operated at the 18-20s/fibre regime for the entire run without problems. On the last four nights the engineering and software staff felt sufficiently confident with the growing reliability of the positioner to depart around midnight and leave the system in the care of mere astronomers. Only on the first of these did they have to come back!

The camera operated at a very low level of halation for the entire run with a heater voltage of over 12V. The camera focus was tweaked after the spectrograph reached ambient dome temperature with very good results. It seems best to put the centre of the curved CCD further out of focus than the corners as the out-of-focus aberrations are smaller on-axis than in the corners of the chip.

One notable milestone on the last night was that for the first time a long exposure field was observed on BOTH plates — i.e. reconfigured for a later hour angle as per the 2dF concept. This was due to the astronomers' growing confidence in the stability and reliability of the positioner.


Half a night was devoted to standard 2dF setup (including pointing model), which was done with extra care because of plans for astrometric investigations. The offsets between plates and focal plane imager (FPI) are now confirmed as stable, so it should be possible to short-cut the standard setup to <1 hour when 2dF settles down.

Two nights were spent doing investigations of reported problems with 2dF astrometry in the SW corner of the plate to further understand the results obtained in June. The seeing was typically 2-3", so the conditions were not optimal for science; centroids to 0.2" were still possible so it was fine for astrometry.

Initially we did a new tracking test of the telescope with the 2dF top end. Tracking appears stable with a maximum excursion of 0.5", always recovering, and a typical excursion of 0.2-0.3" even at considerable hour angle. This was as expected.

Experiments with fibre peakups on PPM stars confirmed earlier observations of an intermittent problem in the south-west region of the field. We then did a astrometric FPI calibration using SDSS stars, supplied for this run by Steve Warren. This gave a similar internal fit RMS (0.5") but, with 3-4 times as many stars, fills in the focal plane much better. The RMS versus the earlier PPM-derived plate model was large (0.8") with evidence of a 1-2" skew in the SW. This is attributed to the poor plate-filling factor of the PPM survey. We then did various fibre setups with the SDSS calibrations on the SDSS fields. The RMS from peakup tests came out at 0.2" (slightly better than PPM) but still gave evidence of a problem in the south-west. Tests at various telescope positions show the effect varies between 0 and 1" residual in the SW, with a vague correlation with hour angle. The fact the effect can disappear probably explains why it was not picked up last year. (It may even be a new effect).

The cause of the effect is still not understood but SDSS stars provide dense enough coverage to map it. Our philosophy was to take an average map so that at least no target in the south-west will miss by more than 0.5". KGB wrote a quick and dirty program to fudge all the input coordinates to allow for the observed distortion; this was tested on the next night sandwiched between science exposures, and the results were gratifyingly correct. We got all the signs right! Subsequent science fields were observed with fudged coordinates, so the "sine-wave" missing light problem ought to be somewhat better, though perhaps not completely cured. Information from observers is awaited as there were some minor problems running the 2dF data reduction software during the run.

In August the entire system is being taken apart for the 400-fibre upgrade, so the entire issue will have to be revisited. With careful monitoring, it is possible to correct for this effect, even though its origin is not currently understood. Hypotheses include (a) non-flatness of the field plates (the effect is similar for both); (b) tilt of the field plate associated with tumbles; (c) irregularity in the grid defined by fiducial holes.

The last two nights were spent mostly doing service science observations.


Chris Tinney 

Observers may like to note that a new "quick look" data examination tool has been provided at the AAT. Called (imaginatively enough) QIKLOOK, it provides a simple graphical user interface to the "imexam" program within the Figaro data reduction package. The facility can be used on any of the AAT's Sparcstations by logging in to the obsred account and typing "qiklook". Alternatively, you can start it from the "Reduction" pull-down menu on the "AAT" information GUI.

What you will get is a small terminal window with logging information, a figdisp display window, and an interface as shown above.

You can use the file chooser and the buttons to switch between the files on the NFS-mounted VAX disks or the files transferred automatically to the Sparc, and between the "data" and "inst" and "ccd_1" and "ccd_2" files. Using the "Show" buttons you can display the selected file.

You can also use the buttons on the right to obtain information about your data: seeing, photometry, X and Y profiles. If you haven't "Show"n the file you've chosen before using these, the display will be updated. In each case you can use the "…" button to set the default values for parameters like the image scale. You should be able to do this once at the start of your run, and then forget about these buttons. The "Focus" button allows you to estimate the best focus from a focus sequence and the "Offset" button allows you to calculate the telescope offsets required to put a given object in your image at a given position — this can be used with IRIS to acquire your objects directly onto the slit in the infrared, or with the Auxiliary CCD to acquire RGO objects. All you have to do is work out once at the start of the night what pixel position on the chip corresponds to your preferred slit location.

QIKLOOK was mainly designed with imaging in mind, but may be useful to spectroscopic observers as well. Please give it a try, and let Chris Tinney ( know if you have any problems.



Chris Tinney 

If you've tried observing a faint target with the RGO Spectrograph lately, you'll know that when it comes to trying to put a 21st magnitude star down the slit, the current TV system just doesn't cut the mustard.

So as an interim solution to the AAT's long-standing acquisition difficulties, we have begun permanently mounting the Blue Thomson CCD at Auxiliary focus during all RGO Spectrograph runs. When used in conjunction with the QIKLOOK (see above) offsetting function, it allows you to acquire very faint targets in under about a minute, which is considerably faster than you can do it with the TV. You also get a nice CCD image of your field, so that later you can be SURE just which object you observed.

The only difficulty with using the Aux CCD is that you must be very careful not to over-saturate it. The Thomson chips suffer from significant residual images for up to a day, so the CCD is not useful for bright objects — luckily that's just what the TV is perfect for. The other difficulty is that currently you can not acquire with the instrument rotated to any angle other than 0, 90, 180 or 270 degrees. If there is a need for this, QIKLOOK will be updated. If you find the Aux CCD useful — or if you don't — please let us know on the feedback form at the end of your run.


Chris Tinney (

Observers are warned that both ATAC and PATT have awarded time to over-ride programs in Semester 1997B. These programs can over-ride ANY other program for a period of up to 3 hours. Pay-back time is not provided to programs over-ridden for less than 3 hours. No program will be over-ridden for more than 3 hours without the provision of pay-back time, and without the approval of the Director.


Ray Stathakis

There is growing interest in archival data as a source of observations, in particular as survey-style research gains popularity. The AAT is unusual in having archived its data since its inception, over twenty years ago. The result is a huge and ever-growing database. Unfortunately access is a problem with much of the data still on nine-track tape, accompanied by hand-written log books.

An extensive upgrade is planned. A web-accessible database will be established, initially with new data, but eventually with details of the complete archive available on line. In parallel, data is being copied to CD-ROM, a more accessible medium. A related project is to provide an on-line observing log, melding instrumental parameters from the data headers and interactively entered additional information from the night assistant.

In the meantime, send archive requests to Please expect some delay with large requests while the archive is being upgraded.


Ray Stathakis

A large number of southern fields are now available through the local APM web page ( Note that the galactic plane and Magellanic clouds are not included.

CD-ROMs of the 2nd Epoch Digitized Sky Survey by STScI are available at the Epping Laboratory. About half the sky is covered at present. (As these CD-ROMs are masters for distribution to other Australian groups, access is limited). Please contact Ray Stathakis ( or Jessica Chapman ( for access. The DSS II contains scans of the recent red survey carried out by the UK Schmidt and Palomar telescopes. The scans are made at higher resolution than the DSS I, with one arcsec pixels. Copies of the set are being made (with permission from STScI) for the Schmidt, MSSSO, U. Sydney, U. Melbourne and ATNF.


Jessica Chapman

The recently relaunched service observing program for the AAT, announced in the last newsletter, has got off to a good start, with a substantial number of service proposals received for the July deadline. During semester 97B (August 1st 1997 to January 31st 1998) seven service nights have been scheduled for the following instruments: RGO spectrograph (not including time series or polarimetry modes), FORS, Prime Focus CCD imaging and the Taurus Tunable Filter. The maximum length of time that will be allocated to a service project is three hours.

Full information on the AAT service observing program and application forms are available on the web at URL From now on service applications will only be accepted using the web-based forms. Service proposal deadlines will be set at regular 2-month intervals.

The next deadline for service proposals is

September 15th.


Ray Stathakis (

No… not a workers' revolt. The Consolidation Project recently started at the AAO tries to fill in the gaps between the large instrumentation projects. Areas of interest at present are upgrades to the AAT observing area, providing a suite of software tools for observing, and calibrating the growing suite of diffraction gratings. Main contributors are Ray Stathakis as project manager, Steve Lee, Frank Freeman and Bob Dean.

AAT Observing Area

Recent observers at the AAT will find that the usable control room area has grown to almost twice the previous area, and with much more benchspace. Work is continuing, with reduction of ambient noise and upgrade of furniture priorities for the next year.


Observing Software Tools

A suite of software tools are available on aatssf, the Sun workstation in the control room. A simple menu system is provided: type AAT (note case!) at the prompt. Routines are available for setting up your instrument and quick-look data reduction, and useful information and links are provided. In particular, it is no longer necessary to use the VAX to set up the RGO spectrograph. Pipeline data reduction packages provided are QIKLOOK and AATRGO. More utilities are planned, such as a quick-look package for UCLES, and an on-line arc atlas. Though the AAT menu is only available from aatssf at present, it will be available on all common-user Sun workstations by the end of the year.


Ray Stathakis (

A pipeline software package is now available for the RGO spectrograph. AATRGO is a package written by Jim Lewis, and supported at the AAT by Ray Stathakis. At present it is recommended for quick-look applications during observing, but it has the potential to be used as a full-look reduction system.

To start the package, start up IRAF on an xgterm (either yourself or through the utilities menu AAT) and type `aatrgo'. After a short delay, a `gui' interface appears (see figure) followed by an empty monitor window. As the AAT data headers are incomplete, the grating details need to be entered, as well as the file name of the observation and the calibration arc (if any). I recommend you activate the radio buttons `dispersion correct data', `plot resulting spectra' and `redo previous reductions' for a quick-look result. Other options need more tweaking of parameters. Then just press `reduce' and you'll get a wavelength-calibrated, extracted spectrum drawn in `splot'. Type `q' in the splot window to return to the gui and reduce the next spectrum. You cannot reduce a spectrum more than once (with different parameters, say) unless the redo option is selected. Once the spectrum is plotted, the usual, and many, options of splot are now available — type `?' in the splot window to get a list of options in the monitor window. Type `q' to exit splot help in the monitor window, and then `q' in the splot window to exit splot.

Progress of the processing is given in the monitor window, so it is useful to keep the window open. More details of parameters used can be found in the log which is available as a button in the gui. Other buttons provide choice of parameters used in the different stages of processing, similarly to the usual IRAF epar command. The buttons toggle the parameter windows open and closed. For full-look reduction, you'll need to fine tune these parameters to suit your data set. Input lists of files can be entered instead of individual files using the `@' symbol.

I encourage observers to try out this software, and would appreciate feedback on problems or suggestions for enhancements. At present it is available at the AAT on aatssz (Solaris version) or on aatssf (Sunos, in the control room). The package will be more widely distributed with the next release.



Stuart Lumsden

As regular readers of this Newsletter are no doubt aware, the next major instrumentation project at the AAO will be a new infrared camera/spectrometer, IRIS-2. This instrument has been through several stages of approval, and we are currently in the preliminary design phase. We felt it was worth telling the community briefly about the various capabilities we plan for IRIS-2 at this time: we welcome comments on any aspect of IRIS-2. Actual construction of the instrument will start sometime early next year. We have at all times tried to take account of possible changes that may occur in the future to ensure that IRIS-2 can be upgraded to improve its performance. We would expect the lifetime of the instrument to be at least ten years and it is clear that scope for upgrades is vital in such a long timespan. One of the major failings of the original IRIS was the relative lack of provision for modifying it to take advantage of, e.g., larger arrays.

IRIS-2 will use a 10242 HgCdTe Rockwell Hawaii array, which limits our wavelength coverage in the red to 2.5mm, the same as for IRIS. The array however should have better uniformity, higher quantum efficiency, and lower read noise and dark current. This should mean considerably improved sensitivity compared to IRIS.

As part of the initial approval process, ACIAAT made recommendations as to the relative ranking of the various suggested science aspects that IRIS-2 might cover. The highest of these was for wide field imaging. We have therefore used this as the basis for the optical design. When used with the f/8 secondary, IRIS-2 will have a pixel scale of 0.5 arcsec/pixel and a field of view of 8.5arcmin x 8.5arcmin. Unlike IRIS, there will be no facility for changing the pixel scale within the basic camera itself. However, the camera will work with three different secondaries (f/8, f/15, f/36) enabling a range of image scales to be achieved. This will require observers to plan ahead, since it will not be possible to change secondaries during the night.

In addition to straight imaging modes, IRIS-2 will also offer facilities for spectroscopy. Initially, these will only be low resolution (R~1000-1500), but this may be upgraded in future. Further potential upgrades include provision for an integral field unit or multi-slit spectroscopy. To some extent the provision of these facilities depends on the potential demand from users, so if you feel you would want to use these please let us know.

Polarimetry (both imaging and spectroscopic) will also be possible with IRIS-2. The exact implementation of the polarimetric modes is still under consideration, but we should be able to allow imaging polarimetry over half the available field of view without difficulty. However, the requirements for polarimetry may clash with the new guiding system being built for IRIS-2, and it may be that guiding will not be available for some polarimetric observations.

As users of IRIS will be aware, the AAT is not ideally set up for guiding when infrared observations are being carried out. Since this is clearly a limitation that we don't want to keep for IRIS-2, we have designed a new guiding system for IRIS-2 itself (which will also provide a fairly crude but usable optical imaging capability at the same time). This system allows for a possible upgrade to full tip-tilt correction. Since the median seeing in the near infrared at the AAT is 1 arcsecond or better, tip-tilt is clearly an attractive option for the future, and should provide considerable gains in sensitivity, even if we never achieve the image quality that an excellent site such as Hawaii can.

The current plan would see IRIS-2 on the telescope two years from now, with it being made fully available as a commissioned instrument sometime during the first semester of 2000. We will keep you posted on the progress of IRIS-2 through regular updates in the Newsletter. In the meantime, if you have any queries or suggestions for the IRIS-2 team, please contact us.


Q.A. Parker (AAO), S. Phillipps (Bristol) and D.F. Malin (AAO)

The new AAO/UKST high-specification, large, single-element H alpha interference filter has recently been delivered, tested, accepted and commissioned on the UKST. The filter is coated on a 356 x 356mm piece of RG 610 filter glass and is probably the largest optical interference filter for astronomy in the world.

Preliminary test exposures taken in March/April reveal excellent imaging and uniformity, the high quality of the filter and the clear (circular) aperture of close to 300mm, or a little over five degrees on the sky. The CSIRO National Measurement Laboratory in Sydney have quantitatively confirmed that the manufacturer, Barr Associates, closely met the stringent optical specifications we set, especially the 7nm bandwidth.

Used in conjunction with Kodak's fine-grain Tech Pan film (which has a sensitivity peak around H alpha), the filter will enable us to make H alpha images with an unprecedented combination of area coverage, depth and resolution. Amongst other projects, we plan a major survey to include the Southern Galactic plane and Magellanic Clouds. This should result in a survey superior to any previous attempt to image ionized gas in the Galaxy or outside. Many new discoveries and avenues of research are expected to be opened up.

Given the 5o diameter circular field of the filter, standard 5o-centred survey fields were not possible for full contiguous sky coverage. We have a thus adopted a conservative 4o field centre separation to ensure proper field overlap. As a consequence, we will need about 200 such fields to cover the Southern Galactic Plane. We estimate that the survey, which can make excellent use of good-seeing grey time, will take about 3 years to complete.

Announcement of Opportunity

To maximise the availability and usefulness of the exciting new H alpha survey on the UKST, the AAO, Schmidt Telescope Panel and H alpha survey consortium have agreed to upgrade the Galactic plane/Magellanic Clouds proposal to a fully fledged AAO survey. This will ensure rigorous quality control of the survey and its immediate accessibility by the community. In addition, the consortium still intend to provide the community with a fully calibrated digitised database of 10mm resolution pixel data which will be released as a CD-ROM atlas from SuperCOSMOS scans. The CD-ROM atlas will be released in instalments for faster community access. Film copies of the survey will also be available.

The AAO invites the community to apply for non-survey photographic time to use the filter for specific fields/areas of interest.

WWW H alpha pages

Up-to-date information concerning the survey including current progress, examples of preliminary images, the filter specifications and details of the H alpha survey consortium membership can be found at:

International Workshop

A very successful 3 day International Workshop was held in Sydney, Australia from April 16-18, 1997 to highlight the science that can be expected from the new UKST H alpha survey. Around 50 participants, half from overseas, gathered at CSIRO Radiophysics laboratory in Epping. The aims of the workshop were primarily to showcase the new survey, to discuss the science likely to be produced and hopefully to generate new ideas and collaborations, particularly in the radio and millimetre regimes (and hence to tie in with these associated surveys). Some preliminary results from the survey were presented as well as an assessment of the H alpha imaging work performed on other telescopes. It became clear that our new survey would set the benchmark for wide-field high resolution imaging of our Galaxy in ionized gas.

The Scientific Organising Committee comprised Q.A.Parker (AAO, co-chair), S.Phillipps (Bristol, co-chair), J. Bland-Hawthorn (AAO), D.F. Malin (AAO), D. Morgan (ROE), L. Staveley-Smith (ATNF) and W.Zealey (Wollongong), while the Local Organising Committee was M.Hartley (chair), Q.A.Parker, H. Woods, F. Watson, V. Thompson, K. Larson (AAO) plus A. Green (Univ. Sydney)

The workshop proceedings will be published in a special edition of the Publications of the Astronomical Society of Australia (PASA), which is a refereed journal, early in 1998.

Survey Coverage (sq.deg) Resolution Sensitivity FOV Timescale Telescope/Detector
AAO/UKST ~4000 ~1" 5-10R? 6.5° x 6.5 º 1999 1.2m UKST + TP film
70Å FWHM at H alpha
WHAM ~30000 ~1º 0.25R 1º Sep 1998 0.6m + Fabry-Perot + CCD 0.25Å FWHM at H alpha (12km/s)
MSSSO ~4000 ~12"? ? 7 º x
7 º
1998 400mm F/4.5 lens +2K CCD 55Å FWHM at H alpha
Marseille ~280 ~9" 0.2R 38' x 38' 1999 36cm + photon counter + Fabry-Perot 5 km/s
Virginia ~4000 ~1.6' 1 pc/cm -6 10º x 10º ? f/1.2 lens + CCD
17Å FWHM at H alpha
Michigan ~100 ~3" * 1.1º x 1.1º 1999 Curtis Schmidt + 2K CCD

30Å FWHM at H alpha

Current major H alpha surveys: This is not an exhaustive list, though most major current H alpha-type surveys are included. This clearly illustrates the unique combination of area coverage, resolution and sensitivity offered by the new AAO/UKST survey. Further details of these surveys will be found in the upcoming H alpha workshop PASA proceedings.

N.B. Sensitivity for the Michigan survey: 3x10-17 ergs/cm2/s/arcsec.

Note the lack of consistent units for measuring H alpha sensitivity levels! e.g. Rayleighs, pc/cm-6, and the Michigan survey's ergs/cm2/s/arcsec.



Rhonda Martin

Here we were, looking forward to the usual rain and mayhem that 2dF provides whenever it is on the telescope, and the weather was superb. Granted, we did get an inch of precipitation — for which we thank this marvellous instrument — but after that it was glorious weather.

Previous to this last 2dF run, we actually had observers complaining that the weather was so good they didn't get any chance to rest on these long winter nights — now, that is truly something!

It's a sad thing that no-one gets all hot and bothered any more about 2dF, a sure sign that it is going well — all the entertainment value has gone out of it. Stepping over prostrate bodies of exhausted technicians in the corridor or pouring coffee into hollow-eyed electronics types, soothing the highly-strung ones who threatened nervous breakdowns was the norm and now it is all just too quiet.

In preparation for the 800 fibres in September, a new engineering grade chip is going into one spectrograph camera and a science grade chip into the other. The spectrograph controller racks are now installed and are being commissioned. I suppose it is on to the next marvel of optical engineering now, with all the attendant excitement and brains in overdrive.

We have a new electronics officer in Allan Lane, from Perth. One of the first things that happened to Allan was that he was struck down with the latest rotten flu and has subsequently felt miserable ever since — welcome Allan!

While Ian Bates is on sick leave, Trevor Lindsay has taken over the carpentering jobs in Coona. While he was doing a lovely job on the new Director's Office he livened up a quiet afternoon no end when he put a power drill through a power cable — causing Wayne Clarke some trepidation when the resulting flames and sparks licked at his moustache ends. So a hundred thousand welcomes (not volts) Trevor!

Our intrepid Night Assistant Gordon Schafer is back from sick leave, sporting a re-designed ankle and a pain-free smile. We have missed Gordon's wry sense of humour and are glad to see him back.

Coming back from Long Service Leave (yes, there are still people actually at the telescope) is John Sullivan, while Frank Freeman is off to England to catch up on his family and all his old Rugby friends.


Helen Woods

The AAO is pleased to welcome Roger Bell, our new Public Relations Officer and AATB Secretary. Roger joined the AAO in July, coming from a public relations position at the NSW Department of Fisheries from where he has brought considerable public relations and media expertise. Interestingly, Roger has also worked at the Museum of Applied Arts and Sciences where he enjoyed a year at the Sydney Observatory in a public relations capacity, concentrating, among other things, on the development of visitor programs.

The AAO very much misses Wendy Carmichael who left at the end of May after 15 months as a permanent and very active member of the AAO's administration team. Before she took up a permanent position with us, Wendy previously worked temporarily as the AAO's Receptionist and Travel Clerk and it is from this position that many readers may also have encountered her. We wish Wendy all the best for her future.


David Malin (

Information about nightly activities on the AAT is now available on the WWW. The initial motivation for this was to provide some text for the AAT Visitor's Gallery so that casual onlookers would know what the amazing device before them did during the night. However, it was clear that there was a wider audience, hence the WWW page. See .

It is intended to ask all future recipients of AAT time for a brief, non-technical, general interest piece about their project so we can continue this. Comments and suggestions on improving the first iteration are welcome.

Sample H alpha Image from a recent Vela Supernova test exposure: The figure is a contrast enhanced derivative from a 1.8o x 2o part of the Vela supernova remnant showing the astonishing detail visible on Tech Pan images made with the new H alpha filter. The original was a two hour test exposure taken with the focus out by 70mm from optimum! Routine survey exposures will be of three hours, so exciting images and discoveries are expected. See page 13 for details.


1997 Oct 1 - 5 6 - 8 9 - 20 21 - 24 25 - 31
Nov 1 - 3 4 - 7 8 - 19 20 - 23 24 - 30
Dec 1 - 3 4 - 6 7 - 20 21 - 24 25 - 31
1998 Jan 1 - 2 3 - 6 7 - 20 21 - 23 23 - 31
Feb 1 2 - 5 6 - 19 20- 22 23 - 28
Mar 1 - 2 3 - 6 7 - 20 21 - 23 24 - 31

editor HELEN JOHNSTON editorial assistant SANDRA RICKETTS

ISSN 0728-5833


PO Box 296 Epping, NSW 2121 Australia

Epping Lab

Telephone +61 2 9372 4800 Fax +61 2 9372 4880 email < >


Telephone +61 2 6842 6291 Fax +61 2 6884 2298 email < >

URL < >