Mem. S.A.It. Suppl. Vol. 9, 409c© SAIt 2006 Memorie della

Supplementi

The Telescopio Nazionale Galileo

E. Oliva

INAF, Telescopio Nazionale Galileo, Apartado de Correos 565, E-38700 Santa Cruz de laPalma, Spain e-mail: oliva@tng.iac.es

Abstract. This report briefly summarizes the status of the � 3.58m Telescopio NazionaleGalileo (TNG) and provides statistical information on the downtimes, observing requests,costs and publications rate. It also presents a few scientific results which particularly high-light the virtually unique capabilities of some of its instruments. The overall picture is thatof a competitive, efficient and cost-effective facility which is very much appreciated andused by the Italian and international astrophysical community.

Key words. Telescopes – Instrumentation

1. The TNG instruments at a glance

Five years ago, in the summer of 2000, theTelescopio Nazionale Galileo (TNG) was com-pleted with all its baseline instruments whichconsist of two efficient and versatile multi-mode imager/spectrometers working at optical(“DOLORES”) and infrared (“NICS”) wave-lengths, a high resolution optical spectrograph(“SARG”), a direct optical imager (“OIG”) andan adaptive optics module (“AdOpt”).

DOLORES acts as the workhorse for allthose programs requiring fairly wide fieldand/or low-medium resolution spectroscopy.The performances of this instrument are in linewith those of the numerous, high efficiencymulti-mode facilities available on other 4mclass telescopes.

SARG, the high resolution spectrometer, isa remarkably high-efficiency and high-stabilityinstrument designed with the specific purposeof obtaining very accurate radial velocity mea-surements. At wavelengths longer than 4000 Åit is one of the best high resolution instrumentspresently available on 4m class telescopes.

OIG is optimized for high quality imagingand is the only optical imager which can takefull advantage of the relatively frequent peri-ods of good seeing when the PSF cannot besampled by the much coarser scale and poorerimage quality offered by DOLORES

NICS is one of the most efficient and ver-satile multi-mode near IR instruments operat-ing worldwide. In low resolution spectroscopyit can directly compete with 10m-class facili-ties.

AdOpt is an optical bench permanentlymounted on the Nasmyth-A derotator and canfeed NICS or OIG. It is the only systemworldwide employing a pyramid wavefrontsensor and operating on night-time telescopededicated to scientific observations. AdOpthas been extensively used to demonstrate,with great success, the technical viability ofthis system, but had so far limited scien-tific applications due to the difficulty to oper-ate it. Nonetheless, a thorough upgrading ofits control system is in progress and shouldsoon allow routine observations also in ser-vice/queuing.

410 E. Oliva: The Telescopio Nazionale Galileo

Fig. 1. Weather downtimes per month averagedover the last five years (dots+curve, based on TNGdata) and pver the 1989-1999 period (based on datafrom ING).

Fig. 2. Weather and technical down-times averagedover the last 9 observing periods. Each observingperiod covers a period of six months, the data forAOT11/05A are for 5 months (Feb-Jun 2005).

All the instruments, except AdOpt, werecommissioned in a very short time and, sincethe beginning of 2001, all the system has beenworking on a routine basis offering to theItalian and international community one of themost complete and flexible combination of in-strumentation worldwide.

2. Operations and budgeting

TNG is one of the few, if not the only, 4m-class telescope worldwide heavily employing

service observations with flexible and queuingscheduling. The fraction of programs sched-uled in service is close to 50% of the totalnight-time available. All the operations, fromthe assistance to the preparation of the observ-ing blocks to the actual observations and ex-traction of the data from the archive, are han-dled by only 7 astronomers. The same staff as-tronomers are also responsible of the qualitycontrol of the data and of the development ofthe calibration plans. Such an heavy duty workresults in a significant cost saving but, however,limits the development of scientific activitiesled by local astronomers. In other words, theprime aim and activity of TNG is to offer a ser-vice to the community.

In parallel to its routine astronomical work,TNG is also undergoing a thorough upgradeof a significant fraction of its hardware andsoftware systems. The aim is to increase thereliability and efficiency of the system, fol-lowing the prescriptions of the 2001 report ofthe gruppo istruttorio per il TNG led by R.Pallavicini. The work is performed by 13 tech-nical staff members who also take care of theregular maintenance of the telescope and in-struments. While much progress has alreadybeen achieved in terms reliability (the techni-cal downtimes in the last two years are below2%, see Fig. 2), a lot of work is still neededto achieve a higher efficiency. In particular,one needs to increase the speed of CCD con-trol systems, parallelize the setup operationsof the instruments and eliminate (automate) alarge number of manual operations which arenormally performed during an observing night.Thanks to the use of flexible scheduling, all thetechnical works are handled with virtually noimpact on the scientific observations.

From the budgeting point of view, TNG ispossibly the most cost-effective 4m-class tele-scope among European astronomical facilities.Its cost per night (about 7300 Euros, see Fig. 3)is much lower than competing telescopes atMauna Kea or La Palma and virtually identi-cal to that of ESO-NTT whose running costs,however, are mitigated by the fact that the samestaff takes also care of the 3.6m and 2.2m ofLaSilla. This result is particularly importantif one considers that service observing with

E. Oliva: The Telescopio Nazionale Galileo 411

Fig. 3. Costs per night of 4m class telescopes par-ticipating into the OPTICON access program.

Fig. 4. Evolution of time requests in proposals tothe TNG-TAC compared with the proposals submit-ted by Italian PI’s to ESO.

queuing scheduling, normally considered as avery expensive activity, is performed at impres-sively high levels by TNG.

3. Downtimes

The monthly weather downtimes at TNG aver-aged over the last 5 years are similar to thoserecorded by the Isaac Newton Group (ING)group in the last decade (see Fig. 1). Theresults confirm that the site of La Palma isstrongly seasonal, with extremely good condi-tions in summer but with a quite high fractionof useless nights in winter.

Fig. 5. Distribution of time requests to the TNG-TAC by instruments. The AOT calls are semiannualand span the 2001 (AOT3,4) to 2005 (AOT11,12)years.

Fig. 6. Pubblications on refereed journals based ondata collected at the Telescopio Nazionale Galileo.

This has created quite some frustrationto astronomers interested in extragalactic re-search - indeed a large fraction among thoseapplying to TNG time - who usually point to-ward objects which are best visible betweenlate autumn and early spring. Nonetheless, theintensive use of queuing observations has al-leviated the problem and made it possible tocomplete most of the proposals to which theTNG time allocation committee has awarderthe highest scientific merit.

The down-times by semester are displayedin Fig. 2. The time lost for bad weather is onaverage 30% but with quite large variations

412 E. Oliva: The Telescopio Nazionale Galileo

Fig. 7. Left: average, continuum subtractedNICS/TNG spectra of high-z quasars showing theFeII and other metal lines. Right: deduced Fe/Mgabundance ratio as a function of redshift, filled dotsare solely based on NICS/TNG data.

among different observing periods caused bythe succeeding between remarkably mild andstormy winters, such as those of 2000-2001and 2001-2002.

The technical downtimes, defined as thescientific time loss due to the failure of the tele-scope and/or instrument system, show a def-inite trend from about 11% in early 2001 tobelow 2% in 2004 (see Fig. 2). This reflectsthe good results of the detailed maintenancework aimed at stabilizing the telescope and fo-cal plane systems. The drop in 2003 is also re-lated to the activation of the service mode withflexible scheduling which allows a much moreefficient handling of the technical and scientifictimes.

4. Proposals

The statistics of time requests to the TNG-TACare summarized in Figures 4. There a definitetrend in the total time requested which has de-creased by almost a factor of 2 between 2000and 2005, and the over-subscription rate hasconsequently dropped from a factor of about4 to 2. A similar, though less pronounced trendis also visible in the time requested by Italianastronomers to the ESO 3.6m and NTT tele-scopes at La Silla (red squares in Fig. 4), whilethe Italian requests for the ESO-VLT has in-creased and reached values close to those ofTNG.

This is most probably a direct consequenceof the stunning growth of new instruments of-fered by VLT during the period under consid-

Fig. 8. NICS/TNG spectrum (black solid line) ofone of the most distant object known in the uni-verse (z=6.2). The dotted and dashed lines are thespectra of similar objects much closer to us. The reddashed line corresponds to dusty quasars, while theblue dotted line to dust free objects.

Fig. 9. False colour image (red=K, green=J,blue=I) and NICS-Amici spectrum of the extremelyred object (ERO) at the center of the field. Thedata (black histogram) are remarkably similar to thespectrum of a giant elliptical galaxy at the recordredshift of z=1.7 (green line).

eration as well of the increasing difficulty ofastronomers to “digest” the always increasingamount of data produced at the telescope.

E. Oliva: The Telescopio Nazionale Galileo 413

In this respect ESO has the main advantageof providing already reduced data to the as-tronomer who, therefore, can save a lot of timeby-passing the, sometimes painful, task of datareduction. A last effect could be also relatedto the weather downtimes at La Palma whichhave repeatedly frustrated the attempts of sev-eral groups to complete their investigations onobjects which are best visible in winter.

The distribution of time requests by instru-ments is displayed in Fig. 5. The “classical”optical imager and low resolution spectrome-ter (DOLORES) is the most requested instru-ment, immediately followed by the multi-modeinfrared instrument (NICS) and the high reso-lution optical spectrometer (SARG).

The direct CCD imager (OIG) attracts verylittle interest, possibly because its field of view(5’ x 5’) is much smaller than that availablewith DOLORES (9’ x 9’).

The very few requests for the adaptive op-tics module may reflect the difficulty of operat-ing the system and achieving a reasonably uni-form standard of data quality with just a fewnights a year scheduled in visitor mode. Sincethe AdOpt instrument can only work in goodseeing conditions, the only way to achieve sci-entifically useful results is to perform the ob-servations in service/queuing mode. To achievethis aim the instrument is undergoing intensiverefurbishment works aimed at creating a reli-able and user-friendly system.

5. Publications statistics

The number of articles based on TNG datapublished on refereed journals has steadily in-creased in the last 5 years, and reached in 2005the remarkable rate of 5 papers per month (seeFig. 6). The first, important step forward oc-curred in 2001, just one year after the first lightof the 3 spectrometers (summer 2000), whilethe major increase started in 2003. This 2-3years time-lag is similar to what occurred inother telescopes and reflects the typical timenormally necessary by astronomer to properlyanalyze data coming from new instruments.

Remarkable is the anti-correlation betweenthe publication rate, which has increased by afactor of ∼3 from 2001, and the requests for

observations which have decreased by a factorof ∼2 over the same period (see Fig. 4).

6. Science highlights

This section does not pretend to select the“best” scientific results achieved using TNGdata, a task which is indeed far beyond the aimof this report. Rather, it summarizes a few, re-cent publications which emphasize the qualityof the data obtained at TNG relative to otherworks based on observations at other tele-scopes, including Keck, VLT and other 10m-class telescopes.

6.1. High redshift quasars

The TNG/NICS combination has produced themost complete data-base of low and mediumresolution spectra of the farthest quasars (z>5)discovered by the SLOAN survey. These spec-tra, whose quality is comparable or even bet-ter than those obtained at 10m telescopes bycompeting groups, yielded important informa-tion on the physical and chemical conditions inthe early (5 (see Fig. 7). The produc-tion of such a large amount of iron requiresa major episode of star formation occur-ring at redshifts z>9 in the host galaxies ofthese quasars. This starburst possibly cor-responds to the violent birth of the very firstgeneration of massive stars which rapidlysynthesized, via thermo-nuclear reactionsin their interiors, all the elements heavierthan Helium – including the magnesiumand iron observed in the quasars – and ex-pelled them into the interstellar medium byexploding as supernovae at the end of theirshort (few million years) lives.

414 E. Oliva: The Telescopio Nazionale Galileo

– Broad-absorption line quasars (BAL-QSO)are relatively rare in the local universe butbecome the most common form of quasarsat z>5. The huge amount of outflowing gas,responsible for the absorption lines, indi-cates that a large-scale star formation eventhas formed the massive galaxy hosting theactive nucleus at epochs much earlier thanpredicted by the popular CDM models.

– The continuum spectral shape of the highredshift BAL-QSO’s is remarkably differ-ent than that observed in lower redshift(i.e. older) quasars (see Fig. 8). At longwavelengths (>1800 Å rest frame) theyare remarkably similar to local dust-freefree active nuclei, but show clear evidenceof dust absorption at shorter wavelengths.This demonstrates that the very first gen-eration of star-burst supernovae also pro-duced a large amount of dust grains withoptical and chemical properties very dif-ferent than at present times. The forma-tion of dust grains at such early epochshas fundamental implications for the sub-sequent evolution of the galaxies, sincethese grains strongly favored the formationof molecules, of low-mass stars and of theirplanetary systems.

6.2. Old and massive galaxies at z>1.5

Faint, spatially extended and very red obejcts(called EROS) are commonly found in opti-cal/IR imaging surveys. Their nature is con-troversial. Their colors are compatible withrelatively small and dusty star-burst galax-ies, but also with giant galaxies with a domi-nant population of old stars at redshifts z=1-2. The latter interpretation implies that largescale star-bursts and massive galaxies have al-ready formed at very early epochs (z>7), instrict contrast to what predicted by the popu-lar CDM model which postulates that star for-mation starts on small scales and that largegalaxies are formed at later time by the merg-ing of smaller systems. Hence the importance

of obtaining quantitative spectroscopic infor-mation of EROS. However, this is a very dif-ficult task because of their low fluxes and redcolors. Infrared spectroscopic study carried outat 10m-class telescopes failed to produce un-ambiguous results. The only quantitative mea-surements are based on optical (CCD) spectra,but these are limited to the least red objects.

A major step forward in this field was re-cently performed at TNG, thanks to the uniqueperformances and capabilities of the NICS-Amici disperser. Using these spectra (see Fig. 9for an example) Saracco et al. (2005) identi-fied 10 among the most massive (>3 1011 M�)and distant (z=1.2-1.7) field elliptical galax-ies in the Universe. The presence of such gi-ant evolved galaxies at these redshifts suggeststhat the assembly of massive spheroids hastaken place at red-shifts z>7 thus supportinga high efficiency in the accretion of the stellarmass in massive halos in the early Universe.Moreover, one object is the most massive anddistant early-type galaxy (z=1.7) discovered sofar. In this galaxy new X-rays data from theXMM-Newton satellite unveiled the presenceof a strongly obscured and high luminosity ac-tive galactic nucleus (AGN), thus corroborat-ing the link between the Quasar activity andthe formation of massive galaxies at very highredshifts.

References

Maiolino, R., Juarez, J., Mujica, R., Nagar, N.,Oliva, E. 2003, ApJ, 596, L155

Maiolino, R., Oliva, E., Ghinassi, F., Pedani,M., Mannucci, F., Mujica, R., Juarez, Y.2004, A&A, 420, 889

Maiolino, R., Schneider, R., Oliva, E., Bianchi,S., Ferrara, A., Mannucci, F., Pedani, M.,Roca Sogorb, M. 2004, Nature, 431, 533

Saracco, P., Longhetti, M., Severgnini, P.,Della Ceca, R., Braito, V., Mannucci, F.,Bender, R., Drory, N., Feulner, G., Hopp, U.,Maraston, C. 2005, MNRAS, 357, L40