The Coming of Age of 2dF

Science Highlights Of 2dF Commissioning
During the December and January runs a significant milestone was achieved during 2dF runs: we finally changed over from routine instrument testing to routine science observing. 2dF has been performing much more reliably over the last few months and is beginning to show its potential productivity.

In these runs useful data was obtained for 9 different PATT and ATAC projects, in addition to the early data of Ray Sharples mentioned in the last Newsletter. There are somewhere around 4000 science spectra to date. No doubt we will soon lose track of this total! The choice of science projects was driven by RA, grating convenience, seeing, telescope panel ratings and suitability for early 2dF observing. We tried to cover as wide a range of science as possible. As this was done in service mode we took advantage of the extra flexibility and tried to select projects in anticipation of good or mediocre seeing.

Figure 1: A random sample of galaxy and QSO spectra from the SGP 463 Survey field observed in January. These are drawn from a 1.5 hour exposure in 1.5'' seeing.

Our first major success is that we have now obtained test fields for the Bright Galaxy/QSO redshift survey -- i.e. more than 700 galaxies and 400 QSOs. Sample galaxy/QSO spectra are shown on the cover and in Figure 1. Two of these fields have now been reduced and have had redshifts measured, thanks to Robert Smith who has come to AAO for the final year of his PhD Thesis, supervised by Brian Boyle. We show the galaxy/QSO n(z)s in Figure 2 -- they compare very well with model predictions. As expected the galaxy survey will sample local structure out to 600Mpc. The goal in the early stage of this survey is to get 5 long-exposure fields for templates and 5 short-exposure fields for test purposes. It is expected this will be completed in the forthcoming March run. We have also obtained 2 fields for the galaxy redshift survey's ancillary deep (R<21) survey. Prime redshift hunting season opens next spring when the SGP returns and the 400-fibre system is expected to be available.

Figure 2: The redshift distribution of galaxies (top panel) and QSOs (bottom panel) in the SGP 246 and SGP 463 fields from November and January 2dF runs. Model n(z) predictions are shown (note: these do not allow for galaxy clustering nor the UVX QSO redshift cutoff at z=2.2). At the left and right are shown bars representing the number of non-target and unidentified objects respectively. The completeness of identifications is about 75% for the galaxies and 80% for the QSOs -- we expect these figures to increase substantially when the spectrographs and CCD are fully commissioned.

Of course the 2dF is not just for redshift surveys. Moving on to other projects: Michael Drinkwater from UNSW has obtained several fields in the nearby Fornax galaxy cluster (see separate article). Background QSOs seem quite popular -- Chris Tinney from AAO has used 2dF to find several new QSOs behind the Carina and Fornax satellite galaxies and hence lay the groundwork for measuring the proper motions of these galaxies over the next 5 years. He commented that ``although the added overhead in preparing for these observations is high (ie. preparing accurate co-ordinates for CCD UV-excess candidates), the speed of 2dF (compared to previous `one-at-a-time' experience with the RGO spectrograph) more than compensates''. In another AAO project (3 out of the top 4 ATAC projects involved AAO staff!) Helen Johnston has obtained 2dF data to look for QSOs behind the Globular Cluster 47 Tucanae -- again for proper motion work.

There are also other, more specialised, redshift surveys being undertaken with 2dF. Karl Glazebrook has observed 3 fields from the Hawaii K-band imaging survey which covers 10 square degrees to K=16. The goal of this work is to obtain redshifts and measure the K-band galaxy luminosity function and examine galaxy evolution out to z=0.3. A few brown dwarfs may not go amiss either. Andrew Hopkins and Lawrence Cram from the University of Sydney have obtained one field for their Phoenix Deep Survey -- a 2 degree CCD field with additional Molonglo and Parkes radio data. Their project aims to study the relation between faint radio galaxies and IRAS starburst galaxies.

Moving back to PATT, Mike Irwin (Cambridge) has obtained some high-resolution 2dF spectra of the popular Carina dwarf galaxy in order to measure radial velocities and hence the galaxy's dynamic structure and history. Finally, Steve Warren from Imperial College is looking for some optical Einstein Rings -- of which one is now known. He has obtained spectra of a sample of very massive and red early-type field galaxies and is looking for lensed emission lines in their spectra. His survey is sensitive to lenses from z=2.7 out to z=6. We wish him every success and hope he, and other early 2dF beneficiaries, will show us their early results in future issues of the Newsletter.

Instrument Development
The current status as of January 1997 is that we have 200 fibres on each of the two fieldplates feeding one spectrograph. We can routinely reconfigure the fibres on one fieldplate while observing on the second without problems. The time to reconfigure the fibres is between 90 and 120 minutes which with careful planning allows us to observe almost continuously throughout the night as long as the target science exposure times are of this order. The whole observing program from reconfiguring the fibres to pointing the telescope, acquiring the field and starting the CCD exposure is now fully controlled from a user interface running on a Sparc-10 in the telescope control room.

Over the past few months considerable effort has been put into improving the reliability of some weak mechanisms (notably the gripper jaw) and testing and improving the software control. A usable spectrograph collimation and focus has been achieved (by no means a trivial effort) and the light-proofing of the back-illumination has been successful (also non trivial as we have the equivalent of a 15W lamp in the light path of the spectrograph!).

Over the next six months 2dF will be spending a considerable amount of time on the telescope. However between these spells we are not going to be idle, the `down time' is even more tightly scheduled than the telescope time. This is because we have to ready the complete system for the upgrade from 200 to 400 fibres per fieldplate. This process may be summarised as follows:

1. The current spectrograph is under limited manual remote control (grating angle, slit traverse and calibration lamps), before the upgrade to 400 fibres we must commission the second spectrograph with full software control and then retrofit the control system to the current spectrograph.

2. Improve the speed of the fibre positioning process to enable the full complement of fibres to be configured in a useful time. We would like initially to be able to configure all the fibres within the previous target exposure. For most of the sky and for most science projects this puts a requirement of configuring all the fibres within 90-120 minutes (a factor of two faster than we are currently positioning fibres). We are planning this upgrade (and hopefully will exceed this requirement) well before the upgrade to 400 fibres.

3. Install and commission field-plate rotation. In order to remove refraction effects we will rotate each fieldplate while tracking. Although partially fitted and tested this has not yet been commissioned on sky. This will in fact make observing much easier as we have been acquiring fields without an instrument rotator and (successfully) predicting the field rotation using a combination of telescope pointing model, image scale and distortion map as part of the 2dF Observing System software.

4. Install and align the optics in the second spectrograph and fit science grade CCD detectors.

If all goes to plan we will be providing a 400 fibre capability in time for the prime SGP dark time in September/October. Following this there will be a further shakedown period with little visible change in the 2dF status but be assured that there will be considerable behind the scenes work going on in order to make 2dF usable by visitors rather than local experts.

2dF System Efficiency
During the January run we finally managed to obtain a decent set of well-acquired spectra of Landolt standards in good seeing (1.0'') and photometric conditions. Averaging over 11 stars and scaling we obtain the following count rates: 0.6, 0.6 and 0.4 electrons/Angstrom/sec for B=17, V=17 and R=17 stars with the 300B grating. Note these counts are for an engineering CCD and the spectrograph is not yet optimised. We expect these to improve, especially in the B-band where we expect an increase of 30% in the final system.

Figure 3: Signal/Noise of some random galaxies and QSOs from SGP 463 (1.5 hours). The S/N has been measured very carefully from the spectra, avoiding strong features. The prediction from the above new sensitivity figures is also shown (solid line), assuming the fibre aperture picks up 70% of the light (appropriate for point sources in 1.5 arcsec seeing) and adopting a mean B-V = 0.5 colour.

From this we can calculate expected signal to noise ratios for various exposure times, object types, gratings (using the known relative throughputs from the RGO spectrograph) and sky brightnesses. These match very well the S/N measured from the fainter Galaxy/QSO data (Figure 3). Rather than publish here several pages full of tables for all possible combinations of the 6 variables we have created an interactive Web-based signal to noise calculator which can be reached here.

Observers are requested to use these new, empirical, figures in planning future 2dF applications rather than the old 1994 estimates.

The sensitivity figures and the Web calculator will be revised when new measurements are made.

Pipeline Data Reduction Software
The pipeline reduction software is currently working but not yet producing as good results as standard methods -- such as the IRAF `dofibres' package. Now we have achieved a reasonably stable spectrograph focus we have obtained the necessary maps of tramline curvature and fibre profile required to optimise the software. This is now being worked on and we will be testing this on the December and January data and validating it against independent reductions done by the various observers.

The other major software component required in-place for fully automatic reduction is communication of fibre configuration information from the 2dF control task, running on a UNIX workstation, to the CCD headers which are written on the site VAX. Work on this has started and we hope this will be available by the March run.

Future Observing
During semester 97A and the first part of 97B 2dF observations will continue to be done in service mode in parallel with the commissioning of the continuing upgrades to the system. Prospective observers can expect to be contacted by email by AAO 2dF astronomers for target lists several weeks in advance of a run. Ancillary information about how to choose guide stars, what to do about sky positions, etc., will also be supplied. We re-emphasise the importance of good RA, DEC coordinate lists. 2dF now positions to <0.3 arcsec over the 2 degree field with full allowance for the telescope model, field distortion and atmospheric refraction. The actual fibre to target allocation will be done by AAO staff as the list of valid fibres is currently variable. We are now using the `Iterative Swapping Algorithm' of Gavin Dalton from Oxford which generates very efficient allocation tables with minimal fibre crossings. We warmly thank him for all the time and effort he has placed in developing this algorithm and for coming out to Australia to help us integrate it into our software. We also thank PATT for supporting his visit.

Our goal, assuming the successful commissioning of the full 400 fibre and two spectrograph system by September, is to move towards normal scheduled, shared-risks observing in the latter half of semester 97B.

Karl Glazebrook, Ian Lewis

Back to contents page