I n 1999, a new and uniqueintegral-field spectrograph,SAURON, made its debut at the

WHT. SAURON has a large field ofview and high throughput, and wasdeveloped as a private instrument ina collaborative effort between groupsat Observatoire de Lyon, LeidenObservatory, and the University ofDurham for the systematic study ofthe stellar and gaseous kinematicsand the line-strength distributions ofnearby early-type galaxies.

Galaxy dynamics andintegral-field spectroscopy

Many elliptical galaxies and spiralbulges are triaxial and/or displaymultiple kinematic components.Triaxiality leads to velocity structuresthat are difficult to map withtraditional long-slit spectroscopy.Furthermore, the central fewarcseconds are often �kinematicallydecoupled�: the inner and outerregions appear to rotate arounddifferent axes. This makes two-dimensional (integral-field)spectroscopy of stars and gasessential for determining the internalstructure of these systems, and forunderstanding their formation andevolution.

Substantial instrumental effort hasgone into high-spatial resolutionspectroscopy with slits (e.g., STIS onHST), or in small areas (e.g., theintegral-field units (IFUs) TIGER

and OASIS on the CFHT), to studygalactic nuclei. By contrast, studiesof the large-scale kinematics ofgalaxies still make do with long-slitspectroscopy along at most a fewposition angles. Therefore, the galaxydynamics groups at the Observatoirede Lyon, Leiden Observatory, andthe University of Durham joinedforces to build SAURON (SpectroscopicAreal Unit for Research on OpticalNebulae), an IFU optimized forstudies of the kinematics of gas andstars in galaxies, with highthroughput, and most importantly,with a large field of view.

SAURON

The optical layout of SAURON isillustrated in Figure 1. A filterselects a fixed wavelength range,after which an enlarger images thesky on the heart of the instrument,the lenslet array (Figure 2). Eachlenslet produces a micropupil. Afterthe collimator, the light of eachmicropupil is dispersed by a grism,

and a camera then images theresulting spectra onto a CCD. TheCCD is rotated slightly to avoidoverlap of adjacent spectra. Thisdesign is similar to that of theprototype IFU TIGER, whichoperated on the CFHT between 1987and 1996 (Bacon et al., 1995), and ofOASIS, a common-user IFU for usewith natural guide star adaptiveoptics at the CFHT. Table 1summarizes the basic specificationsof SAURON.

The SAURON optics are optimizedfor the wavelength range 4500�7000Å.Currently only one filter and onegrism are available, for the range4810� 5400Å. This allowssimultaneous observation of the OIIIand Hβ emission lines to probe thegas kinematics, and a number ofabsorption features (the Mg b band,various Fe lines and again Hβ) formeasurements of the stellarkinematics (mean velocity, velocitydispersion, and the full line-of-sightvelocity distribution), and the line

THE ING NEWSLETTER No. 2, March 2000

1 1

TELESCOPES AND INSTRUMENTATION

The One Eye that Sees All: Integral Field Spectroscopywith SAURON on the WHT

P. T. de Zeeuw1, J. R. Allington-Smith2, R. Bacon3, M. Bureau1, C. M. Carollo4, Y. Copin3,R. L. Davies2, E. Emsellem3, H. Kuntschner2, B. W. Miller5, G. Monnet6, R. F. Peletier7,E. K. Verolme1

1: Leiden Observatory; 2: Durham University; 3: Observatoire de Lyon; 4: Columbia University; 5: Gemini International Project; 6: University ofNottingham; 7: European Southern Observatory

Spatial sampling 0.26″ 0.94″Field of view 9″ × 11″ 33″ × 41″Spectral resolution (FWHM) 2.8 Å 3.6 ÅInstrumental dispersion (σ) 70 km/s 90 km/sGrism 514 lines/mmSpectral sampling 0.9 Å/pix 1.1 Å/pixWavelength coverage 4810– 5400 ÅCalibration lamps Ne, Ar, WDetector EEV12 2148 × 4200Pixel size 13.5 µmEfficiency (optics / total) ~35% / 14%

Table 1. SAURON specifications.

rebalance the WHT when SAURONis mounted, four counterweightstotalling 580 kg are bolted on withthe instrument: these are visible as asteel ring surrounding the top ofSAURON. The dewar, made availableby the ING, can be seen at the verybottom of the photograph.

The plan to develop SAURON washatched in the summer of 1995, andthe first work started in late spring1996. The instrument was completedin January 1999. This fast schedulewas possible because SAURON is aspecial-purpose instrument with fewmodes, and because much use couldbe made of the expertise developed atObservatoire de Lyon through thebuilding of TIGER and OASIS. WhileSAURON was constructed as a privateinstrument, plans are being developedto make it accessible to a widercommunity.

Commissioning the Eye

SAURON was commissioned on theWHT in early February of 1999,during time shared with the ESASTJ team (see ING Newsletter, 1, 13).In the preceding days, the instrumentwas checked, together with thesoftware to run it, and the interfaceto the WHT environment. We paidparticular attention to the opticalalignment of SAURON and thedetector, to be sure that the spectrawere aligned accurately with thecolumns of the CCD.

The instrument performed very wellfrom the moment of first light, onFebruary 1. We modified an internalbaffle slightly, to avoid some residuallight leak between the main field andthe sky part. A small tilt of the filterremoved an unwanted ghost image.The slow read-out of the CCD led tosome inefficiencies, but this is beingfixed by the ING.

Figure 4 presents some calibrationdata taken during commissioning.Panel a) shows a small part of anexposure with the grism taken out,which produces an image of themicropupils. Panels b) and c) showspectra obtained with the internal

No. 2, March 2000 THE ING NEWSLETTER

1 2

Figure 1 (top). Optical layout ofSAURON.

Figure 2 (bottom). The SAURON lensletarray is made of fused silica, andconsists of over 1600 square lenslets,each 1.35 mm on the side.

Figure 3 (right). SAURON mounted onthe Cassegrain port of the WHT.

strengths (age, metallicity, andabundance ratios).

A new feature of SAURON comparedto existing IFUs is the ability to dosimultaneous sky subtraction. This isachieved by means of an extraenlarger which observes a patch ofsky 1.7 arcmin away from the mainfield, using about 100 lenslets.

Each of the lenslets normally images0.94 × 0.94 arcsec on the sky. Thisundersamples the typical seeing atLa Palma, but provides essentiallyall one can extract from the integratedlight of a galaxy over a remarkablylarge field of view, 33 × 41 arcsec,observed with a filling factor of 100%(1500 lenslets). SAURON thereforedoes justice to its name, as itsmythical predecessor was also knownas �The one eye that sees all�.

SAURON contains a mechanism forswitching to another enlarger,resulting in a 9 × 11 arcsec field ofview, sampled at 0.26 × 0.26 arcsec.This mode can be used in excellentseeing to study galactic nuclei withthe highest spatial resolutionachievable at La Palma withoutadaptive optics.

The total throughput of SAURON,including atmosphere, WHT anddetector, is 14%.

The mechanical design of SAURONis similar to that of OASIS; it is lightand strong, and minimizes flexure.Figure 3 shows the instrumentmounted on the Cassegrain port ofthe WHT. SAURON weighs less than300 kg, as compared to about 1500 kgfor ISIS, which normally occupiesthis port. In order not to have to

tungsten and neon lamps, respectively.The former shows the nicely alignedcontinuum spectra, while the lattershows the emission lines used forwavelength calibration.

Data Analysis

Each SAURON exposure produces aCCD frame containing about 1600individual spectra. These are extractedby means of an optimal algorithmbased on a full optical model of theinstrument. The resulting data cubeis reduced in the standard manner,archived, and analysed by means ofa set of programs modelled on theXOASIS package.

We have developed a special pipeline,fittingly called Palantir, to analysethe SAURON data cubes in a verysystematic manner. The first moduleof Palantir removes the instrumentalsignature, the cosmic rays, and alsodoes the wavelength and fluxcalibration and the sky subtraction.The second module allows merging ofindividual exposures and mosaicingof different pointings for the sameobject. Measurement of the kinematicquantities and the line strengths isthen done with dedicated algorithms.The reduced data is archived atLeiden Observatory, and will bemade public in due course.

First Results

Figure 5 shows results from the firstSAURON observing run (February14�20, 1999). Reconstructed totalintensity, stellar Mg b line strength,mean velocity and velocity dispersion,and emission-line gas intensity andvelocity maps are presented for theE6 galaxy NGC 3377. The stars showa striking rotating disk pattern witha spin axis misaligned ∼10º from thephotometric minor axis. The Mg bisophotes seem to follow the continuumlight. A comparison with the Hβisophotes will reveal whether thisgalaxy contains a younger stellardisk. The gas also reveals non-axisymmetric structures and motions.These results are based on a total ofonly four 30 min exposures (no spatialbinning was applied). The large

angular extent of the maps should benoted.

During our run in October 1999, weobserved the E3 galaxy M32 to testthe high-resolution mode. Figure 6shows the integrated intensity andthe mean velocity field of the centralregion of M32, derived from a single45 min exposure with good seeing.The velocity field is very regular, isconsistent with axisymmetry, andhas a peak amplitude of about 50km/s.Individual velocity measurements areaccurate to a few km/s. The velocitydispersion in M32 is significantlysmaller than the SAURONinstrumental dispersion, except inthe inner few arcseconds (e.g., vander Marel et al., 1998), and we do notshow this here.

Figures 5 and 6 also illustrate a keyadvantage of integral-field spectroscopyover traditional aperture and long-slit spectroscopy: by integrating theflux in each spectrum, the surfacebrightness distribution of the galaxyis recovered. As this is derived fromthe same spectra that are used toobtain the kinematics and linestrengths, there is never any doubtabout the relative location of thesemeasurements and the galaxymorphology.

The SAURON project:structure of E/S0/SaGalaxies

We built SAURON to measure theintrinsic shapes and internal velocityand metallicity distributions of early-

type galaxies, and to gain insightinto the relation between the stellarand gaseous kinematics and thestellar populations in spheroids. Ourstrategy is to study a representativesample of nearby ellipticals, lenticulars,and early-type spiral bulges. TheSAURON data are combined withhigh spatial resolution spectra of thenuclei, and interpreted throughstate-of-the-art dynamical and stellarpopulation modelling.

Our sample was constructed as follows.We first compiled a complete list ofaccessible E/S0 galaxies and Sa bulgesfor which SAURON can measure thestellar kinematics: �6º ≤ δ ≤ 64º (zenithdistance), cz < 3000 kms-1 (spectralcoverage), σ > 90 kms-1 (spectralresolution). The objects span a factor50 in luminosity (MB < �18) andcover the full range of environment,nuclear cusp slope, rotational support,and apparent flattening. The galaxieswere then divided into six categories(E/S0/Sa; field /cluster) and arepresentative sample of 72 objectswas selected to populate the ellipticityversus absolute magnitude planeshomogeneously (Figure 7). Based onour two observing runs in 1999, wehope to complete the observations ofthe entire sample by early 2002.

The detailed measurements will becompared with fully general galaxymodels constructed by means ofSchwarzschild�s numerical orbitsuperposition technique (e.g., Crettonet al., 1999). The dynamicalmodelling uses all appropriateimaging and spectral data that areavailable, including HST and OASIS

THE ING NEWSLETTER No. 2, March 2000

1 3

Figure 4. Examples of various SAURON calibration measurements, taken duringcommissioning. Each panel shows only a small part of the entire CCD frame sothat details can be seen. a) Micropupil image taken with the grism out. Each dotis the diffraction pattern from one square lenslet. b) Continuum image using thetungsten lamp. c) Neon arc.

No. 2, March 2000 THE ING NEWSLETTER

1 4

Figure 5. SAURON measurements of the E6 galaxy NGC 3377. Each pixel is 0.94″× 0.94″ and the FOV shown is 30″× 39″. a)Reconstructed total intensity. b) Stellar Mg b index. c) Stellar mean velocity. d) Stellar velocity dispersion. e) Gas total intensity([OIII] λ5007). f) Gas mean velocity.

a b

f

d

e

c

Figure 6. SAURON measurements of M32. a) Reconstructed total intensity.b) Stellar mean velocity. The FOV is 9″×11″, and the spatial sampling is0.26″× 0.26″.

spectra to constrain the mass of acentral black hole. When combinedwith the constraints on the stellarpopulations derived from the line-strength distributions (Kuntschner &Davies, 1998), this will shed newlight on the fundamental connectionsbetween the large and small scaledynamics, the formation (andexistence) of supermassive blackholes and galactic nuclei, and thehistory of metal enrichment in early-type galaxies.

It is a pleasure to thank RenéRutten, the ING staff, in particularTom Gregory, and Didier Boudon andRene Godon for enthusiastic andcompetent support on La Palma. TheSAURON project is made possiblethrough financial contributions fromNWO, the Institut National desSciences de l�Univers, the UniversitéClaude Bernard Lyon I, the universitiesof Durham and Leiden, and PPARC.

a b

THE ING NEWSLETTER No. 2, March 2000

1 5

Figure 7. Distribution of E, S0, and Sa galaxies in the SAURON sample in theplane of ellipticity ε versus absolute blue magnitude MB. Open circles: fieldgalaxies. Solid circles: cluster galaxies.

References:

Bacon, R., et al., 1995, A&AS, 113, 347

Cretton, N., de Zeeuw, P. T., van derMarel, R. P., Rix, H-W., 1999, ApJS,124, 383

Kuntschner, H., Davies, R. L., 1998,MNRAS, 295, 29

van der Marel, R. P., Cretton, N., deZeeuw, P. T., Rix, H-W., 1998, ApJ, 493,613 ¤

Tim de Zeeuw (tim@strw.leidenuniv.nl)

First Sodium Laser Beacon at La Palma

L. Michaille, J. C. Quartel, J. C. Dainty, N. J. Wooder (Imperial College London),R. W. Wilson (ATC), T. Gregory (ING)

I n the ING Newsletter No. 1, anongoing project to implement anadaptive optics system on the

WHT was described. Called NAOMI(Nasmyth Adaptive Optics for MultipleInstrumentation), this system will becapable of eliminating, to a largedegree, the image degrading effectsof atmospheric turbulence.

As with any adaptive optics system,NAOMI will rely on measuring lightfrom a relatively bright referenceobject, or �guide star�, close to thedesired target in order to make thenecessary correction. At visiblewavelengths, however, it is often notpossible to find a sufficiently brightstar in the neighbourhood of theastronomical region of interest. One

would like to be able to simply dial-up a guide star exactly where it isneeded, and this is possible with theuse of a �sodium laser beacon�.

At an altitude of 90km, well abovethe troublesome turbulence (whichrarely reaches beyond 25km altitude),is a 15km thick layer of sodiumatoms which, just as in a sodiumlamp, can be excited to spontaneouslyemit light. To achieve this from theground we can use a laser tuned toan absorption wavelength (589.0nm)of the sodium atoms. The excitedregion of atoms creates a bright spotin the sky, the sodium laser beacon,which can be used as the guide starfor an adaptive optics system(Figure 1).

A Laser Guide Star (LGS) system isbeing planned for the WHT tocompliment the NAOMI system. Inpreparation, the Applied Optics groupat Imperial College London has beeninvestigating some practical aspectsof creating and using sodium beaconsfor adaptive optics. Amongst these isa study of the sodium layer dynamicsat La Palma using observations ofan experimental sodium laser beacon.We have built a laser system whichlaunches a 25cm diameter, diffractionlimited beam from the liquid nitrogenbuilding located roughly midwaybetween the JKT and the WHT(Figure 2). Part of the laser beam isdirected through a small sodium ovenand by measuring the absorption wecan lock the (tuneable) laser outputto precisely the correct wavelength

Figures 1 and 2. Left: A laser, tuned exactly to the D2 transition of sodium, can be used to excite the sodium layer at 90kmaltitude. The beacon thus created can be used as the guide star for an adaptive optics system. Right: The Liquid Nitrogen (LN)building which houses the laser. The beam is launched through a hole in the roof.