UKIRT NEWSLETTER 1

UNITED KINGDOM INFRAREDTELESCOPE

NEWSLETTERIssue 7, Autumn 2000

Gravitational arcs associated with the core of Abell 383.

2 UKIRT NEWSLETTER

TOP ENDAndy AdamsonHEAD OF UKIRT OPERATIONS, JAC, HILO, HAWAII

Top End...............................................2Research Articles...............................3

T Tauri Stars seen through high resolutionspectroscopy.................................................3

Surveying the Distant Universe throughMassive Cluster Lenses.............................5

UKIRT and CGS4 define the new Methane-Dwarf sequence(s).....................................8

UKIRT News.....................................12 Multi-Wavelength observing with

TRISPEC........................................................12What the Fast-Guider Strip-Chart can do for

you..............................................................14View from the top............................................15People..............................................................15

Special Report.................................16ORAC - a new data acquisition and reduction

system for UKIRT.....................................16And Finally........................................19

The past few months - indeed the entire period since the previousNewsletter - have been dominated by the installation andcommissioning of the ORAC control system and the new telescopecontrol interface. ORAC is fully described in a “Special Report”later in this issue. At the time of writing, the first few PATT runsusing the new systems have been completed. In the new paradigm,observers are able (and expected) to prepare their programmes muchmore in advance than previously, and it is clear that this contributesto the considerable improvement in observational efficiency weare already seeing with ORAC. UKIRT and its observers owe adebt of gratitude to the ORAC team, at the ATC in Edinburgh andthe JAC.

While ORAC has dominated our efforts in both software andtraining, it would be remiss not to note that we have also installeda powerful Linux PC at the summit. This system carries out all datareduction, and copes very well with the data rate from UFTI, whichhad been an occasional issue with the Sun workstation whichpreceded it (and which is still used to sequence ORAC programmes).“kauwa”, as it is known, is equipped with twin heads, giving itsdisplay a sweeping, landscape quality which is certainly necessarywhen using the ORAC tools to prepare programmes and reducedata (see the centre-spread in this Newsletter).

We have welcomed Paul Hirst as a support astronomer, takingresponsibility for CGS4. Paul is settling into the role very well,and has established excellent relations with his colleagues and thosehe has supported.

New instrument development continues at the ATC. Michelle wasthe subject of a series of reviews in the early summer, and is nowworking to a formal cost limit set by PPARC. The main problem

area - the grating-exchange drum bearings - were redesigned andconstruction has just been completed. Both UIST and WFCAMhave suffered slippage due to the ATC’s concentration of effort onMichelle; at the time of writing, the WFCAM team has just beenre-assembled and work is commencing in earnest. Work on definingthe WFCAM surveys has been gathering momentum since earlythis year; the second, and highly productive, meeting of theWFCAM consortium took place at the Preston wide field meetingheld in late August.

UKIRT’s web pages have been under continual development sincethe commissioning of IRCAM/TUFTI and the arrival of ORAC.Many of the changes were required to take account of ORAC, butwe have also been working on single-page printable versions ofinstrument manuals in response to requests from observers, whoclearly still wish to print out a single document in advance of theirruns. We appreciate this desire to read the documentation, but issueone caution: these documents evolve as the instruments evolve,and a printout taken now should not be taken as definitive a yearfrom now!

TRISPEC, which had its first run on the telescope in February, wasawarded eight nights in Semester 00B, and is back on the telescopeat present. The instrument is more fully described in an article inthis issue.

Flexible observing, which proved its worth two semesters ago, isonce again being done for selected programmes. Three pairs of00B programmes have been identified which have complementaryrequirements in terms of either seeing or thermal-IR transparency.We have been working to establish a relationship between submmoptical depth and transmission in the three-micron window, andhave a working definition for “good” three micron weather whichwill be tested and refined over the coming semester. We alsoanticipate working with the JCMT on a collaborative softwareproject to handle the data and feedback loop between absent P.I.sand the observatory, which will be needed if we are to flex properlybetween a wider range of programmes than is possible at present.For now, we thank the P.I.s of the various programmes identifiedand hope that their experience with flexing is a positive andproductive one.

UKIRT NEWSLETTER 3

T TAURI STARS

SEEN THROUGH

HIGH RESOLUTION

SPECTROSCOPY

DANIEL FOLHACENTRO DE ASTROFISICA DA

UNIVERSIDADE DO PORTO,PORTO, PORTUGAL.

SETTING THE SCENE

UKIRT’s CGS4 is one of the few in-struments available to the astro-nomical community capable of ex-cellent performance at high spectralresolution in the near infrared. Afield of study that can take great ad-vantage of this almost unique capa-bility is, of course, star formation.

T Tauri stars (TTS) represent thepopulation of low mass young starsin the final phases of formation.Their spectra may potentially includeradiation from several components:a stellar photosphere, shocks result-ing from material accreting onto thestar, infalling and outflowing mate-rial, an accretion disk, starspots,strong chromospheric activity and,possibly, a residual circumstellar en-velope. All these are affected by ex-tinction. Understanding these sys-tems requires a multi-wavelengthapproach, of which high resolutionNIR spectroscopy plays a very im-portant role. Here I give an accountof the results from a project devel-oped in collaboration with Jim

Fig.1. Distribution of the veiling measurements for the observedsample of T Tauri stars. Top panel - J-band veiling; Bottom panel- K-band veiling.

Emerson (Queen Mary & WestfieldCollege, U.K.) aimed at studying i)the origin of Hydrogen emissionlines in TTS, ii) the kinematic infor-mation provided by these lines, andiii) any excess emission relative tothe stellar photospheric contribution.For this purpose CGS4 was used inits echelle mode to observe a sam-ple of TTS from the Taurus-Aurigacomplex. 49 stars were observedat J and 36 were observed at K, re-spectively, around Paβ and Brγ, withR~25000, during three nights inDecember 1995. A full descriptionof this work can be found in Folha &Emerson (1999, 2000).

THE NIR EXCESS FLUXTHROUGH VEILING ANALYSIS

The analysis of photospheric linesin a TTS spectrum is the tool usedto study the excess emission inthese stars. Excess emission veilsthe photospheric lines, decreasingtheir equivalent width relative towhat they would be in an undis-turbed photospheric spectrum. The“veiling” at a given wavelength isdefined as the ratio of the excessflux to the photospheric flux at thatwavelength. It can be determinedby comparing the photosphericspectrum of a TTS with that of atemplate star of the appropriatespectral type.

4 UKIRT NEWSLETTER

Fig.2 Examples of T Tauri stars’ Paβ and Brγ line profiles from ourextensive survey performed with CGS4. a) Four sample spectrashowing generally symmetric profiles with weak signs of absorp-tion features; b) four examples of IPC profiles.

Since the late nineteen eightiesveiling has been studied in the opti-cal region of the spectrum of TTS(eg. Hartigan et al. 1995) and ex-plained as a result of shocks wherematter accretes onto the stellar sur-face. The contribution of this ex-cess flux to the NIR is less than 0.1for a typical accreting TTS (eg.Calvet & Gullbring 1998). Emissionfrom disks in TTS is also not ex-pected to produce significant veil-ing in the NIR, especially at the J-band, since the inner disk holes thatare thought to exist do not containenough material emitting at thesewavelengths (Meyer et al. 1997);accretion rates through these disksare not likely to be high enough toproduce the necessary radial tem-perature profile. On the basis ofthese results, NIR veiling in TTSshould be very small. What do theobservations reveal? Figure 1shows the J- and K-band veiling dis-tributions for the stars for which itwas possible to determine the veil-ing. Clearly, the excess flux thatmanifests itself via these veiling

measurements is relatively high, afact not easily explained by currentmodels for TTS and their environ-ment.

Paβ β β β β and Br γγγγγ LINE PROFILES

Traditionally, the strong hydrogenemission lines found in many TTSwere interpreted in terms of massloss (eg. Hartmann et al. 1982).The last decade saw a shift in inter-pretation towards mass accretion(eg. Calvet & Hartmann 1992), fol-lowing the magnetospheric accre-tion (MA) model (eg. Camenzind1990). However, the debate is farfrom settled. These results havebeen based mostly on observationsat optical wavelengths. NIR hydro-gen lines impose strong constraintson models and contribute to the un-derstanding of the origin of hydro-gen emission in TTS.

Using CGS4 we have carried out thefirst extensive survey of NIR hydro-gen line profiles. It shows that Paβ

and Brγ l ines are broad(FWHM~200 km/s) centrallypeaked and slightly blueshifted.Their wings extend up to an aver-age velocity of about 250 km/s butoften reach velocities in excess of350 km/s. More than 85% of theobserved Paβ and Brγ profiles fallwithin two main groups: generallysymmetric profiles with weak signsof absorption features (eg. Figure2a) and inverse P-Cygni (IPC) pro-files (eg. Figure 2b) which exhibit atypical redshifted absorption featuredipping below the continuum. Fromthe profiles that fall in either of thesetwo groups, about 60% of the Paβand 80% of the Brγ profiles belongto the former. These numbers arein complete contrast to what is ob-served in Hα, the most studied hy-drogen line in TTS. Most Hα pro-files in TTS (> 50%) displayblueshifted absorption features,about 25% have a generally sym-metric shape and only about 5% areIPC. IPC profiles reveal the pres-ence of infalling material, the veloc-ity of which can be estimated by thevelocity at which the redshifted ab-sorption feature occurs. Typicalvelocities for the redshifted absorp-tion in Pα and Brγ lines are of theorder of 200 to 300 km/s, consist-ent with free fall from a few stellarradii, as expected from the MAmodel. Comparing the observedNIR line profiles with results fromradiative transfer models found inthe literature (eg. Muzerolle et al.1998) shows that MA models pro-vide a qualitative insight into howthese lines might be produced in theTTS environment; however they failwhen quantitative comparisons aremade.

The data set discussed here dem-onstrate that current knowledgeabout the formation of hydrogenlines in TTS is far from providing adetailed explanation for their char-acteristics and origin. Also, it con-stitutes a solid database with whichmodel results should be comparedif they are to succeed in their goals.

UKIRT NEWSLETTER 5

CONCLUSION

High resolution NIR spectroscopyrevealed unexpected results both interms of the excess emission andthe hydrogen line profile shapes.Current models for TTS have to berefined/reviewed in order to accountfor these observations. From anobservational point of view, ex-tremely important clues regardingthe origin of the excess emissionand hydrogen line emission will re-sult from variability studies of both

veiling and line profiles. Good sam-pling over, at least, one rotation pe-riod of a typical TTS, i.e. up to aweek is necessary. UKIRT togetherwith CGS4 is where it can be done!

References

Calvet & Hartmann (1992), ApJ,386, 239Calvet & Gullbring (1998), ApJ, 509,802Camenzind (1990), Rev. Mod.Astron., 3, 24

Folha & Emerson (1999), A&A, 352,517Folha & Emerson (2000), to appearin A&AHartigan, Edwards & Ghandour,ApJ, 452, 736Hartmann, Edwards & Avrett(1982), ApJ, 261, 279Meyer, Calvet & Hillenbrand (1997),AJ, 114, 288Muzerolle, Calvet & Hartmann(1998), ApJ, 492, 743.

SURVEYING THE

DISTANT UNIVERSE

THROUGH MASSIVE

CLUSTER LENSES

IAN SMAIL1,GRAHAM SMITH1,JEAN-PAUL KNEIB2,OLIVER CZOSKE2 &HARALD EBELING3

1DURHAM UNIVERSITY, U.K.2OBS. MIDI-PYRENEES, FRANCE.3UNIVERSITY OF HAWAII, USA.

Fig.1 A true colour image of the core of the massive cluster Abell 383, createdfrom ground-based B- and K-band frames and the HST R-band exposure. ThecD galaxy in the cluster centre clearly dominates this region and our lensing analy-sis shows that it makes a large contribution to the mass profile in the central 20kpc of the cluster. The z=1.01 giant arc (B0) is visible to the south of the cD andhas an apparently blue nucleus at the right-hand end. Several other multiply-imaged sources are visible by virtue of their blue colours in the halos of brightcluster ellipticals near the bottom edge of the frame (see also Figure 2). However,arguably the galaxy with the most unusual colours is the Extremely Red Object(ERO, called B14) in the bottom-left of the field. This galaxy has (R-K)=6.0 and isbarely detected on our HST frame with R~26. This field is 50"x50" in size.

The properties of massive clustersof galaxies are expected to predomi-nantly reflect gravitational processesand can thus provide unique insightsinto the nature and distribution ofdark matter. One particularly strik-ing demonstration of the masses ofthe richest clusters is their ability toact as gravitational lenses and fo-cus light from background sources,creating highly-magnified (and inrare cases multiple) images of dis-tant galaxies. Moreover, any distantgalaxy seen through the lens is mag-nified (in a tangential direction)shearing the images of backgroundgalaxies into ‘arclets’ - the distortionof these arclets can be used to trace

6 UKIRT NEWSLETTER

Fig.2 A false colour version of the HST R-band frame, overlaid with an isodensity representation of ourlens model (Smith et al. 2000).Contours correspond to projected surface mass densities of 3, 4, 6, 8,11,15 x 109 Mo kpc-2. Note the well-defined circular symmetry exhibited by the cluster mass, as well as thelocal perturbations produced by individual cluster ellipticals. The numerical labels indicate the clustermembers used in the lens model. The alpha-numeric labels identify the multiply-imaged systems usedto constrain the model, and a number of other singly-imaged arclets. The very faint object labelled B14is the ERO which is so clearly visible in Figure 1.

the shape of the mass distributionwithin the lens. To understand moreabout the distribution of mass withinthese massive clusters we aretherefore undertaking a lensing sur-

vey with Hubble Space Telescopeof 12 X-ray luminous clusters in anarrow redshift slice at z~0.2. Thegravitationally-lensed features iden-tified in our HST exposures allow us

to construct detailed mass maps forthe central ~1Mpc regions of theseclusters. By incorporating wide-fieldimaging from CFHT and sensitiveX-ray observations from Newton we

UKIRT NEWSLETTER 7

will trace the dark matter andbaryonic profiles of the clusters outto the point where they merge intothe surrounding field.

Our HST observations provide infor-mation not only about the mass dis-tribution within the lens, but also canbe analysed to determine the redshifts for any background galaxywhich is significantly magnified (andthus distorted) by the lens. The im-ages of some of these galaxies canbe magnified by >10-times and thusoffer an enormous increase in sen-sitivity for the detailed study of faint,high-redshift galaxies. But more typi-cally the galaxies in the HST fieldsare boosted by factors of 2 to 3-times. This amplification is depend-ent upon both the mass distributionin the lens and the redshift of thebackground galaxy. Thus, if we canconstruct an accurate mass modelfor the cluster we can ‘invert’ it topredict redshifts for the faint galax-ies seen in the background. Theoriginal predictions of the redshiftsof galaxies seen through the clusterlens Abell 2218 using this ‘inversion’technique (Kneib et al. 1996) weretested and confirmed with very-deepspectroscopic observations byEbbels et al. (1998).

We are now using UFTI on UKIRTto obtain high quality K-bandimaging of gravitationally-lensedgalaxies seen through the rich clus-ters in our HST survey. This articledeals with the observations of theX-ray luminous cluster Abell 383 (atz=0.19), the analysis of which wehave recently completed (Smith etal. 2000). The UFTI image of thecluster core, combined with ground-based and HST optical images isshown in Figure 1. The medianseeing for the stacked 2.4-ks K-band observation was 0.42", com-pared to 0.17" in the 7.5-ks HSTimage and 0.88" in the 7.2-ks CFHTB-band exposure.

The most obvious lensed feature inFigures 1 & 2 is the giant arc (la-

belled as B0a+B1a/b/c). This com-prises four images of two back-ground galaxies; we have obtaineda redshift for B0a which places it atz=1.01 and provides an absolutenormalisation for the mass modelof the cluster. We find that B1 mustlie at a similar redshift to B0, z~1.Several other unusual lensed gal-axies are visible in the HST image .Of these the brightest are the five-images of a z~3.5 galaxy seen asB2a/b/c/d/e, and the four visibleimages from a five-image systemof a galaxy at a similar redshift seenas B3/a/b/c/d.

However, our UFTI imaging turnedup an equally unusual galaxy seenthrough the core of the cluster - anExtremely Red Object (ERO, Fig-ure 2); this galaxy has K=19.7 andR-K=6.0, meaning it is only just vis-ible in our deep HST R-band expo-sure. EROs comprise a rare classof galaxies which is believed to con-sist of a mixture of evolved early-type galaxies at z=1-2 and dustystarbursts at similar (and higher)redshifts. The latter class includesmany members of the faint submmpopulation identified by theSubmillimeter Common-User Bo-lometer Array at the JCMT. How-ever, by definition EROs are faint atoptical wavelengths and hence onlya couple have reliable redshiftmeasurements. This lack of infor-mation about their redshifts andhence luminosities and restframecolours has hampered efforts to dis-entangle the properties of this unu-sual population.

Using the positions and shapes ofthe various images of the lensedgalaxies seen in Figure 2 we haveconstructed a detail model of themass distribution in the central re-gions of the cluster (Smith et al.2000). This is overlayed on the HSTframe in Figure 2 and shows thatthe cluster’s gravitational potentialwell appears relaxed and circular.We can then employ this model tointerpret the shapes of other

gravitationally lensed arclets seenin this region and predict redshiftsfor these faint, background galax-ies. Using this approach we canconstrain the redshift of the ERO,B14. As only a single image of thisgalaxy is visible it must lie at z<3.9.Moreover, the very faint arclet vis-ible at the position of B14 in our HSTimage appears to be highly elon-gated (a/b~4); this would suggest aredshift of around z~3 and wouldgive the galaxy unmagnified appar-ent magnitudes of K=21 and R=27,well-beyond the reach of even 8-mclass telescopes. When magnifiedto K=19.7, however, this galaxy isjust within reach of efficient near-infrared spectrographs on 8-m tel-escopes and so our estimate of theredshift of this ERO could be con-firmed in the near future.

References:

Ebbels, T.M.D., Ellis, R.S., Kneib, J.-P., Le Borgne, J.-F., Pello, R.,Smail,I., Sanahuja, B., 1998, MNRAS,295, 79Kneib, J.-P., Ellis, R.S., Smail, I.,Couch, W.J., Sharples, R.M.,1996,ApJ, 471, 643Smith, G.P., Kneib, J.-P., Ebeling,H., Czoske, O., Smail, I., 2000,ApJ,submitted.

Image Credits:

The UFTI, HST and CFHT imagingof this cluster was taken for a col-laborative study by our group.These observations were obtainedwith UKIRT, the CFHT (operated bythe National Research Council ofCanada, the Centre National de laRecherche Scientifique de Franceand the University of Hawaii) and theNASA/ESA Hubble Space Tel-escope (operated by the Space Tel-escope Science Institute for theAssociation of Universities for Re-search in Astronomy Inc., underNASA contract NAS 5-26555).

8 UKIRT NEWSLETTER

UKIRT continues to producedefinitive data for the rapidlyadvancing field of brown dwarfstudies. In the last Newslet-ter (March 2000) Phil Lucas &Pat Roche described their dis-covery of a population of veryyoung brown dwarfs in Orion,possibly with masses as smallas 8 Jupiter-masses (see alsoMNRAS 314, 858). This workused UFTI with I, J and H fil-ters. In the September 1999Newsletter I describedUKIRT’s contribution to thediscovery of a population ofbrown dwarfs cool enough atT

eff ~1000 K to show methane

absorption in their near-infra-red spectra. Such objects areprovisionally being called Tdwarfs. Prior to 1999 only onemethane dwarf was known,Gliese 229B, which was dis-covered in 1995 (Nakajima etal. Nature, v.378). In 1999 theSloan Sky Survey reportedtheir discovery, confirmed withCGS4 spectra, of two 229B-like brown dwarfs (Strauss etal. ApJ 522; Tsvetanov et al.ApJ 531), followed by an-nouncements from the 2—Mi-cron All Sky Survey (2MASS)of 4 others (Burgasser et al.ApJ 522), and of another de-tected by the NTT/VLT (Cubyet al. A&A 349). 2MASS havefound further examples of229B-like methane dwarfs thisyear. All of these T dwarfshave spectral energy distribu-tions that are extremely simi-lar to each other. They all haveinfrared colours of J-H ~ H-K ~ 0, and their spectra show

Fig.1 CGS4 spectra showing the L- to T-dwarf spectral sequenceswith weakening CO and strengthening CH4 (based on Leggett etal. 2000, ApJ, 536).

extremely strong absorption bands dueto water and methane, with atomic fea-tures due to absorption by the alkalis

cesium and potassium. This is incontrast to the warmer L dwarfswhose near-infrared spectra showno indication of absorption by meth-ane and which have extremely redcolours of J-K ~1.5. For these ob-jects carbon is predominantly in theform of carbon monoxide and CObands are seen at 2.3µm - 2.4µm.

It was suggested by Kirkpatrick etal. (2000 AJ 120) that the lack of ob-jects with near-infrared spectra

UKIRT AND CGS4 DEFINE THE NEW

METHANE-DWARF SEQUENCE(S)SANDY LEGGETTUKIRT/JOINT ASTRONOMY CENTRE, HAWAII, U.S.A.

UKIRT NEWSLETTER 9

Fig.2 CGS4 spectra showing the onset of methane absorptionat 3.30µm in L-dwarfs (based on Noll et al. 2000; astro-ph/0007449.

3.10 3.20 3.30 3.40 3.50 3.60 3.70 3.80wavelength (µm)

0.0

0.5

1.0

1.5

2.0

2.5

No

rmal

ized

Fλ

2M1506 L3V

2M1507 L5V

2M0825 L7V

showing both COand CH

4 implies that

they are rare andthat the spectraltransition from L toT must occur over avery small tempera-ture range. How-ever in March TomGeballe (Gemini)and myself, usingCGS4 and workingwith Princeton rep-resentatives of theSloan consortiumJill Knapp and AlexMcDaniel, disprovedthis by finding threeexamples of the so-called transition ob-jects. These objectslink the L dwarfs withT

eff~2000 K to the

229B—like objectswith T

eff ~ 1000 K

(Leggett et al. 2000ApJ 536). TheirJHK colours arevery similar to themuch hotter, andvery numerous, Kand M dwarfs, andhence they are ex-tremely difficult fornear-infrared sur-veys to identify; however in the Sloansurvey they are distinguished by ex-tremely red I-Z colours. The three,with the previously known L and Ttypes, form a clear sequence,shown in the figure, of strengthen-ing CH

4 absorption at 1.6µm and

2.2µm (Figure 1) while the CObandhead at 2.3µm is still seen butweakens. H

2O absorption at around

1.15µm, 1.35µm, 1.85µm and2.40µm strengthens. These objectswill form the hot end of the T dwarfspectral sequence, where T dwarfsare defined by the presence of CH

4

absorption in the H and K bands.

There was still the question: at whichspectral type can methane first bedetected? The ν

3 methane band at

3.3µm is two orders of magnitudestronger than the bands seen in thenear-infrared. Keith Noll (STScI) ini-tiated a program with CGS4 tosearch for the onset of methane andin May we detected methane in anL5 and an L7 dwarf (and not in anL3). The spectra are shown in Fig-ure 2 and are reported in Noll et al.2000 (ApJ accepted, astro-ph/0007449). There is telluric CH4 ab-sorption at 3.312µm—3.323µm butthe (sub)stellar band is broader;preliminary models can reproducethe shape of the band quite well.Clearly the term “methane browndwarf” must be used with caution.As these CGS4 spectra show,methane is detectable in thephotospheres of the later L dwarfs.Use of the label “T dwarfs” to mean

detected at wavelengths shortwardsof 2.5µm is a useful definition forclassification , but it should be re-membered that the warmer L dwarfsare not methane-free.

The CGS4 spectra shown in the fig-ures reveal vital diagnostic featuresin the atmospheres of the very coolL and T dwarfs. Certainly sky sur-veys will continue to add to theknown population of free-floatingbrown dwarfs, and we can look for-ward to detecting objects with tem-peratures between the T dwarfs andthat of Jupiter, while theorists workhard to develop models that incor-porate clouds so that we can betterunderstand the physics of theseplanet-like atmospheres.

10 UKIRT NEWSLETTER

The ORAC-OT

UKIRT NEWSLETTER 11

See the Special Report on Page 16 of thisNewsletter!

12 UKIRT NEWSLETTER

MULTI-WAVELENGTH OB-SERVING WITH TRISPECSHUJI SATONAGOYA UNIVERSITY, JAPAN

Fig.1.TRISPEC mounted on the north port at UKIRT.

TRISPEC (the Triple Range Imager andSPECtrograph), shown inset in Figure1 mounted on the UKIRT mirror cell,contains three optical channels, eachwith its own detector, to cover the wave-

length range 0.46to 2.5 microns si-multaneously. OneCCD and two near-IR InSb arrays areused. The incomingbeam from the tel-escope is focused,collimated and thensplit into threebeams by two dich-roic mirrors (Figure2). Each of thethree channels hasa filter wheel-boxcontaining three orfive wheelsequippped withbroad-band filters,grisms and polari-zation analysers(Wollaston prisms).TRISPEC is thuscapable of imaging,

spectroscopy and spectro/imaging po-larimetry at three bands simultaneously.Spectral resolutions are around 100,depending on the slit widths used (de-tails are available from the UKIRT “visi-tor instruments” web page). Linear andcircular polarimetry can be carried outusing an upstream (warm) waveplate(s)followed by (cold) Wollaston prisms in-

stalled inside TRISPEC.This polarization modeof operation is providedby Prof. Jim Hough(University of Hertford-shire, U.K.).

The first trial withTRISPEC at UKIRT wascarried out on February4-7, 2000. Imaging and

spectro-polarimetry data were obtained,although some aberration and light losswere experienced in the infrared chan-nels due to the mis-alignment of theoptical axes, specifically the pupil stops

and the focal planes, relative to the sec-ondary mirror. Nevertheless, this wasa positive first visit to UKIRT.

To illustrate the unique capabilities ofTRISPEC we show some of thespectropolarimetry data - covering op-tical and near-IR wavebands - obtainedearlier this year at UKIRT (Figure 3).

We will carry out a second trial at UKIRTon August 23-30, 2000, when adjust-able trusses, similar to those used withCGS4, will be used to support the 450kgweight of TRISPEC and thereby allevi-ate the misalignment problems experi-enced in February.

Technical Specifications

The detectors used by TRISPEC arean SITe 512x512 pixel CCD and twoSBRC 256x256 pixel InSb arrays (en-

gineering grade).The simultaneousw a v e l e n g t hcoverages are;optical 0.46-0.90microns; J-Hband 0.90-1.90microns; K-Lband 1.90-2.50microns. TheF i e l d - o f - V i e wmeasures 56arcsec on UKIRT(F/36: f=135m).Pixel scales are0.1 arcsec/pixel inthe optical and 0.2arcsec/pixel in theinfrared. Notethat 1) the F/36beam and 135-mfocal length atUKIRT currentlymakes the pixelscales slightly

oversampled for the typical slit-width of0.9arcsec in the spectrophotometrymode, and that 2) atmospheric disper-sion degrades photometric accuraciesat zenith angles larger than 30 degreesdue to the wide coverage of wave-lengths.

UKIRT NEWSLETTER 13

DM2M2

M1

F F GR

P F GR

FP

GR

P

DM1

SM

WIN

CM1CL1

CL2

CM4

OPTICAL: V-I 0.46-0.90 microns

IR1: J-H 0.90-1.80 microns

CM3

CM2

512x512CCD

256x256InSb

256x256InSb

512x512CCD

SLIT VIEWER (OPTICAL)

WP

WIN: DEWAR WINDOW

P: WOLLASTON PRISM/FILTER WHEELGR: GRISM WHEEL

F: FILTER WHEEL

WP: WAVE PLATE

SM: SLIT/MASK WHEEL

CM1/CM2/CM3/CM4: CAMERACL1/CL2: COLLIMATOR

M1/M2: MIRROR

DM2: IR1/IR2 DICHROICDM1: OPT/IR DICHROIC

IR2: K 1.90-2.50 microns

Fig.2. The Optical and Infra-red trains within TRISPEC .Two dichroic mirrors mountedsplit the incoming beam intothree channels.

0.4 0.5 0.6 0.7 0.8 0.9 1.0 2.010

20

40

100

200

400

1000[×10-15]0.5 0.6 0.7 0.8 0.9 1.0 2.0

F λ [W

/m2 /µ

m]

λ (Rest frame) [µm]

Quasar 3C273Redshift : 0.158Slit size : 0.9 x 2.0 arcsecExp. time : 121 s × 12 (0.46 - 0.90 µm) 120 s × 12 (0.90 - 2.50 µm)

2000.02.04λ (Observed frame) [µm]

Hα

HβHγ

PaβHe I

Paα

-6

-3

0

3

6

Qλ

[%]

0.4 0.5 0.6 0.7 0.8 0.9 1.0 2.0-6

-3

0

3

6

Uλ

[%]

λ (Rest frame) [µm]

Fig.3. Optical/Infraredspectropolarimetric observa-tions of 3C273, obtained atUKIRT in February 2000.Note the simultaneous detec-tion of Balmer and Paschenlines in the top pannel.

Acknowledgements: Since 1996TRISPEC has been developed by threegraduate students, H. Nakaya, M.Watanabe and T. Yamamuro, fromNagoya University.

14 UKIRT NEWSLETTER

WHAT THE FAST

GUIDER STRIP-CHART CAN DO FOR

YOU

TIM HAWARDENUKIRT/JAC, HILO, HAWAII

The UKIRT Fast Guider has an optionto display, in a strip-chart format, sig-nals from the fast guider CCD. The“stripchart” display can be very informa-tive and its general use by observers isencouraged. The TSS can set up thedisplay for you: whenever the FastGuider is operating, a strip-chart dis-play is available.

There are two traces: the first trace isupdated about once a second. It showsthe GUIDE SIGNAL, the number usedin the fast guider signal processing. Itrepresents the value, from one readout,of the total signal from the sixteensuperpixels on the Fast Guider CCDless the BACKGROUND SIGNAL. Thislast is the second signal, which ismeasured from the four corner pixelsof the sixteen and smoothed over a 10second period.

CLOUD DETECTION

The most important use of the guidesignal strip chart is as a transparencymonitor. The presence of cloud is allbut unmistakable as the signal levelchanges over timescales of a few sec-onds. Quite thin cirrus cloud (see theFigure inset) is readily detected, whichis a boon in a dark sky when thin cloudis undetectable by eye, even with ex-tensive dark adaption (especially sincehypoxia on Mauna Kea impairs the low-light performance of the retina). Thestrip chart should probably not be re-lied upon absolutely if one is attempt-ing to do 1% photometry, but it is a veryadequate indication of how things aregoing as one pursues one’s favouritegalaxy down into the murk in search ofa spectrum with improved S/N. (Gen-erally a mug’s game, as the strip-chartdisplay will rapidly convince you, if yourspectra have not done so already.)

This is punctuated by rather rapid down-ward excurions of the guide signal.

SEEING MONITORING

On moderately bright objects the Guidesignal can be regarded as the signalfrom the star through a 3.77 arcsecaperture. Photon noise reaches ~5% ata bit below V=15, i.e. anything from theGSC should give a moderately goodtrace.

The effective aperture is small enoughthat even with excellent Mauna Keaseeing and the stabilised images de-livered by UKIRT, there is usually someloss of light from the 16-pixel box asthe image size varies. The fluctuationsof the level of the blue line reflect this,and the size of its wiggles gives an im-mediate (subjective) indication of theseeing: small fluctuations = good see-ing, large fluctuations (but on a gener-

chart display must be set manually onthe screen. If one is observing numer-ous bright sources with short exposuresthis can be quite onerous for the TSSand should probably be avoided. It maybe possible to automate the scale-set-ting, but we have not as yet plucked upthe courage to ask the software sec-tion to address this!

PHOTOMETRY CORRECTIONS

There have been suggestions that theguide star signal could be logged andused to correct for transparency varia-tions when doing photometry. Thiscould work, though only when condi-tions were good enough so that all theguide signal fell in the guide box on theCCD. One would also have to assumethat the clouds were perfectly neutralattenuators. As an old photometrist, Iwould, however, view this stratagemwith grave suspicion.

Excess light (i.e. an upwards excursionof the strip chart) in the “background”signal accompanies the downwardswing of the guider signal when thincloud intervenes, even in a dark sky.This is because of scattering by thecloud of guide star light into the cornerpixels. Bigger excursions happen inbright time, as the cloud is also illumi-nated by the moon, and an increase inthickness is signalled by a rise in back-ground brightness (noted in the figure).

Another effect which can mimmic thesignature of thin clouds is windshake.

ally flat trace) = poor seeing (Modulothe photon noise).

This is a useful facility as the autofocusZrms seeing masurement is availableonly when checking focus, while the“guider trace” method is availablewhenever observations are takingplace. It is not linear (or calibratable!),but it is certainly a useful indicator.

SETTING SCALES

Currently the vertical scale of the strip-

UKIRT NEWSLETTER 15

VIEW FROM THE TOPThor WoldTELESCOPE SYSTEM SPECIALIST,UKIRT,/JAC, HILO, HAWAII

PEOPLE

Another semester is underway, and weare now committed to ORAC and thePTCS. As of this writing (late August),things are progressing nicely and itseems the visitors are catching onquickly. It seems the TSSs are alsocatching on... The hands-on assistancefrom our software group has been mostwelcome!

It appears La Nina has gone (if it wereever here). Since around June, wehave had moisture occasionally at theelevation of Hale Pohaku. This has al-leviated the severe years-long droughtthat has plagued the upcountry, and hasgreened up the pastures on Humu’ulaRanch. This has meant the re-introduc-tion of my favorite black cows. How-ever, it appears the ranchers are finallyweary of having their animals driveninto. They have erected a three-strandelectric fence along the road. I amsomewhat suspicious though, as theyhave set this somewhat far from theroad, leaving a nice stretch of long,green grass, which I bet they turn thecows loose on should it turn dry again.So, no excuse not to drive warily.

I had been asked to make a listing ofsome ‘rules of observing etiquitte’ as apossibly interesting subject, so in col-laboration with some of my compatri-ots (not only at JAC), I came up with arather lenghty list. When I first wrotethis column, I went ahead and wrote upthe entire listing. I then had third andfourth thoughts on this. I certainly didnot want to seem rude and crude anddid not want to imply that visiting ob-servers did not have any sense of ba-sic etiquitte. It is true, though, that somevisitors seem to never have beentrained as children in social graces.These are definitely in the small minor-ity, though they had an overwhelminginfluence in the items on this list, whichI thought was unfair.

And so, rather than get into the nitty-gritty of some small number of people’s

behavior, I thought it best to re-writethis in more general terms. The laun-dry list of “don’ts” shall remain unpub-lished.

A. Safety: Please remember that yourTSS is absolutely in control. This ap-plies to everyone, regardless of theirpolitical stature. For example, in theevent that the TSS has decided the

conditions are dangerous enoughtowarrent evacuation, no dissention willbe tolerated. Discussion over the mer-its of this decision can happen later. Ifyou do not want to leave, you can cer-tainly stay - by yourself! We are not inthis business to put our lives in danger.If you choose to not wear your seat beltin the car, you will find yourself suddenlywalking.

B. Personal space: If you are ill andinfectious, you should remove yourselfwithout being told to do so. We haveall been made sick by visitors bringingexotic bugs here and being most ill-mannered in expectorating their bugsall over the place. Our standing de-mand is a bottle of single-malt scotchfor being infected...and we have col-lected. There are ways to get aroundthis situation without your infecting us,which causes great hardship onUKIRT’s operations. Basically, pleaseremember that we are spending 11 to13hours in a small room, so have consid-eration for everyone’s personal spacein all ways and at all times.

C. General: With the consideration ofthe above; safety and our being con-fined in a small room for long periodsof time (and, of course at altitude andthrough the night), there are some sub-tle ways to ensure that the whole op-eration proceeds smoothly. For exam-ple, remember that while we allegedlydo speak the same language, a lot isoften lost in the translation. It certainlydoes not help that sometimes peopledo not enunciate clearly and think thatthey can face the wall and speak to theother side of the room and be under-stood. Personally, I find it embarassingto keep having to ask people to repeatthings to me; after this happens a fewtimes - get the hint.

Finally, you are reponsible for your clut-ter. We are not janitors here, so pleaseclean up your coffee cups, uneatensandwiches, etc. And a universal

pet peeve: please do not start doingsomething at the summit at the end ofthe night, after we have closed down,that you could do from Hale Pohaku.

So if you follow the safety rulesand pay attention to this situation ofbeing in a confined space, etc., I seeno reason why the whole operationcould not run smoothly. Then again,the reason I (and others) have broughtthis subject up in the first place is thatthere are individuals who apparently donot know how to conduct themselvessensibly. This said, I hope your nextobserving run is a productive and en-joyable! Aloha!

The most recent addition to the UKIRTSupport Scientist team is Paul Hirst,who joined us in March of this year.Paul will largely be responsible forCGS4 as it faces its twilight years(though in the UIST era to come it willbe maintained for use with the echelle).Paul obtained his PhD from LeicesterUniversity and also has an MSc fromJodrell Bank. His research interests in-clude active galaxies, starbursts in AGNand multi-wavelength astronomy. Paulalso enjoys mountaineering - rock andice climbing - so he’ll be a handy guy tohave around should your car breakdown between HP and the Summit.

Kynan Delorey started at the JAC inJune. His official title is ‘ComputerHelper’ and he’s responsible for mak-ing sure that observers get their UKIRTdata and that non goes astray (he’llsoon be responsible for JCMT data aswell). He’s also been writing some niftyutilities in Perl; new on the scene atUKIRT are ‘whoson’ and ‘nightlog’.Kynan is a Sega Dreamcast and DVDfanatic, and he thinks the petting zooat Panaewa Zoo is top notch (hey, sodo my kids, though when do we get topat the white tiger, that’s what I want toknow! - Ed.).

16 UKIRT NEWSLETTER

ORAC - A NEW DATA ACQUISITION

AND REDUCTION SYSTEM FOR UKIRT1ANDY ADAMSON & 2GILLIAN WRIGHT1UKIRT/JAC, HILO, HAWAII2ASTRONOMY TECHNOLOGY CENTRE, EDINBURGH, U.K.

Introduction

ORAC is an acronym which standsfor Observatory Reduction and Ac-quisition Control. The ORAC soft-ware was integrated with all UKIRTinstruments in May 2000, and is nowin use for all UKIRT observing. Itreplaces all the software at UKIRTthat interacts with users, providinga modern observing interface.

The goal of the ORAC project wassimply to improve the publicationrate of UKIRT. It will do this by mak-ing it simpler to prepare and carryout observations, and by giving ob-servers excellent feedback on theirobserving as it goes along, thus re-ducing time wasted, and by produc-ing near-publication quality reduceddata at the telescope. There aremany other ways in which ORACwas designed to provide a goodbase for future operations, however.The above elements are all requiredif we are to consider scheduling thetelescope flexibly: (i) excellent re-mote preparation tools such that en-tire observing programmes can beprepared in the UK or elsewhere, toa level that someone else knows ex-actly what to do, and (ii) sufficientlyaccurate,revealing and non-interac-tive data reduction pipelining so thata non-expert can monitor data qual-

ity on behalf of an absent P.I. Inaddition to the benefits for the user,the ORAC project was designed toprovide benefits to UKIRT opera-tions “under the hood”, by makingsoftware support and future devel-opment easier - it does this by us-ing modular designs to achieve flex-ibility and extensibility. For exam-ple, future new instruments will onlyneed to provide the software that isspecific to that instrument meetingthe interfaces defined by ORAC andeverything will work together. Addi-tional software for flexible schedul-ing could be added in a straight for-ward way.

For many years, UKIRT observershave prepared their observationsand executed them using “EXECs”and “CONFIGs”, generated usingthe UKIRT_PREP system. This sys-tem, based on ASCII files generatedusing tools running on VT terminals,has stood UKIRT in good stead formany years, but it is decidedly longin the tooth and is not appropriateto operations in the 21st Century.ORAC replaces both of these as-pects of data acquisition at UKIRTand also provides a data reductionsystem which is data-driven and intowhich all future UKIRT facility instru-ments will slot. The new prepara-tion system provides far more help

than was possible withUKIRT_PREP - for example it hasa powerful position editor with whichmosaics can be set up and stored,and displayed on digital sky surveyimages. ORAC does not supersedethe telescope control system; how-ever, it does interface to the TCSand is capable of executing slews,offsets, etc. Since UKIRT plans toupgrade the telescope control sys-tem to use the “PTCS” in the nearfuture, ORAC was implementedwith an interface to this new system.Although the original UKIRT tel-escope control system is currentlystill in use, it has been modified to“look like” the PTCS in some re-spects. Of course ORAC had to becompatible with existing UKIRT in-struments, and some aspects of theold software were still good ideas,so in some ways ORAC will have afamiliar “feel”.

Naming conventions

The various components of ORACwith which observers now interactare:

ORAC-OT: Observing Tool (prepa-ration of “Observations” and sub-mission to database).

ORAC-OM: Observation Manager(selection of individual “Observa-tions” at the telescope) and OMSequence Console (display andcontrol of detailed breakdown of ob-serving sequences).

ORAC-DR: Data Reduction sys-tem.

An “Observation” in the ORAC

UKIRT NEWSLETTER 17

Fig.2. An example ORAC-OM session showing the OM tool to the left andthe OM sequencer (the larger window) to the right. Note that the observa-tions being undertaken are the same as those in the example OT sessiondisplayed as a centre spread in this Newsletter (Figure 1).

sense can embody a fairly lengthysequence of events: for example, acycle of 40 quads with CGS4 or anextended 5x5 UFTI jitter patternwould both be described as “Obser-vations”.

People

The ORAC was a collaboration be-tween the UK-ATC and JAC. All ofthe following have contributed sig-nificantly to the new system: AlanBridger, Gillian Wright, FrossieEconomou, Andy Adamson, Min

Tan, Malcolm Currie, Alan Pickup,Maren Purves, Russell Kackley andNick Rees.

Data acquisition: what the usersees

ORAC-OT: The ORAC ObservingTool is based on the Gemini OT, withextensions to allow preparation ofprogrammes for the UKIRT instru-ments. The centre-spread in thisNewsletter shows an OT running onKauwa, the Linux PC at the sum-mit. Your prepared observations arestored to a database using this fa-

cility.

ORAC-OM: This runs at the tel-escope (see the screen shot in Fig-ure 2). Here you load your observ-ing programme from the database,and select individual observationsfor execution - this is shown on theleft of the figure.

ORAC-OM Sequence Console:Submitting an observation for ex-ecution from the OM pops up a ver-sion of the Sequence Console (ifone is not present already for theinstrument requested). This console

18 UKIRT NEWSLETTER

(the much larger window on theright in Figure 2) shows you the ex-ecutable form of the currently-se-lected observation, unwrapped intoits component parts, and gives youcontrol of the observing sequence.For example, in this screen you canrun a sequence from a given high-light position, abort a sequence, re-quest that a sequence stops at alogical breakpoint (no more stoppingquads by waiting for the last expo-sure!) and restart the sequence ifyou need to increase signal-to-noise. The sequence console is alsoresponsible for showing you the sta-tus of the instrument - in the exam-ple shown in Figure 2 CGS4 is be-ing used; the status information isat the bottom right of the consolewindow.

Data reduction

ORAC-DR: This part of ORAC hasbeen in use (and under continualdevelopment) since the arrival ofUFTI in 1998. The pipeline is de-signed to require minimal interactionwith the user - indeed once a copyof the pipeline is set into execution,the only interaction possible, apartfrom adjusting display parametersand doing slices, etc. in GAIA, is toterminate the pipeline. This is de-liberate; the ORAC philosophy is toimplement well-defined observingsequences and to produce stand-ard reduction recipes which accom-pany them. The data reduction proc-ess then requires very little in theway of human interaction. This is notto say that the results of the pipe-line cannot be easily inspected in aninteractive way; the STARLINKGAIA display tool is used for mostimage display purposes, and thisprovides an increasing array of in-spection tools (slices, spectral ex-traction, image analysis, seeingmeasurements etc.).

Because of the above philosophy,ORAC-DR takes its cue from thedata themselves as to how it shouldreduce a given sequence of frames.

Each OT observation contains aDRRECIPE instruction, which doesnothing other than to make sure thateach data frame has in its FITSheader the name of a pre-definedDR recipe to be used on that frame.For example, a set of jittered UFTIframes might be flagged with a DRRecipe name ofJITTER_SELF_FLAT. When thoseframes arrive on the raw data disk,ORAC-DR knows what to do withthem without being told. Indeed theonly argument (as opposed to com-mand-line switch) which the ORAC-DR command takes is a recipename, so that you can over-ride therecipe specified in the FITS header.This gives the flexibility required, forexample, to use a non-flat-fieldingreduction recipe if one exists.

Initial Results

ORAC has been used by all PATTprogrammes since 1 Aug 2000, andhad been used by JAC staff mem-bers for engineering and their ownprogrammes since the initial instal-lation in May. Initial perceptions arethat ORAC offers a considerable ef-ficiency gain (of the order of 20%;we are working on calibrating this).A further gain is to be had when car-rying out combined imaging/spectroscopy programmes: switch-ing from CGS4 to UFTI is now ex-tremely quick, limited only by thespeed with which one can rotate thedichroic and set up on the target.Obviously the better prepared yourprogram is, the more likely you areto see efficiency improvements.

Notes for Observers

Remember that the ORAC philoso-phy is to match observing se-quences to suitable data reductionrecipes. If you require a non-stand-ard observing sequence (or believethat UKIRT should implement yoursequence as standard!), contact theJAC ahead of time to enable us toimplement data reduction recipesappropriate to the data your se-

quence will produce. A good exam-ple might be multi-position noddingalong the CGS4 slit, which whiletrivial to set up in the observing tool,is not as yet catered for by the cur-rent set of DR recipes.

Significant Web Sites

You can get more detailed informa-tion on the ORAC project, its inter-nal workings and details of the soft-ware, from the following site:

http://www.roe.ac.uk/atc/projects/orac/

and its JAC Mirror:

h t tp : / /www. jach .hawa i i .edu /JACpublic/UKIRT/software/orac/

Also, information pertaining to theuse of ORAC with specific instru-ments is given in the individual in-strument web pages.

UKIRT NEWSLETTER 19

UNITED KINGDOM INFRARED TELESCOPEJoint Astronomy Centre660 N.A’ohoku Place

University ParkHilo, Hawaii 96720

U.SA.

http://www.jach.hawaii.edu/UKIRT/

Newsletter Edited by Chris Davis

And Finally...Could it be UKIRT’s answer to theBackstreet Boys!? Alas - no. Inactual fact its UKIRT’s finest: theTSS team (from left to right): WatsonVaricattu, Tim Carroll, (John Davies- Support Scientist - in his lucky ob-serving trousers), Olga Kuhn andThor Wold.

The demand for tables at HalePohaku from astronomers and tour-ists alike forces the queue out intothe car park. Seen here (from leftto right) are Paul Hirst, WatsonVaricattu, Frossie Economou, TomKerr, Sandy Leggett, RussellKackley and (in green) AndyAdamson.

20 UKIRT NEWSLETTER

UFTI snaps another High-Mass Star Forming Region

Perhaps not the most appetis-ing colour-scheme! The red,blue and green channels of this“true-colour”, narrow-bandimage were assigned to H2,[FeII] and NB K-continuum fil-ters respectively, the aim beingto search for signs of dynamicstar formation activity viathese hot-gas tracers. Thefetching green shade of thenebulosity results from themore efficient scattering oflight at the shorter (H-band)wavelengths.

Data courtesy M.S. NandaKumar & Chris Davis