serendipitous discovery of a tidal tail in ngc632 · found in early type galaxies (e/s0), while the...

25

* European Southern Observatory, Garching, Germany.# Department of Astronomy, University of Padova, Italy.

Nulla a che vedere con il vuoto angoscioso della pagina bianca, per la snervante attesa che una parola si accenda nell’oscurità deltuo cielo, come lo scoppio di una supernova.

Paolo Maurensig, Venere Lesa

Abstract. While searching for a light echo of the Ia SN1998es type in NGC632, we dis-covered a very faint tidal tail, which extends for at least 50 kpc from the galaxy nucle-us. This finding, together with the presence of a strong nuclear starburst, the boxy shapeof the isophotes and their twisting, strongly suggests that what we see in NGC632 is theresult of a galaxy merger. NGC632 was previously classified as an S0, while the deepVLT imaging discussed in this paper shows very clearly the presence of spiral arms, ontop of which rather blue knots appear to be projected. They are possible sites of star for-mation, probably triggered by the galaxy merger. Further observations needed to clari-fy the open questions are outlined.

Keywords. NGC632, starburst, tidal tails.

Friulian Journal of Science 2. 2002, 25-38

Serendipitous discovery of a tidal tail in NGC6321

F E R D I N A N D O P A T A T * , G I O V A N N I C A R R A R O #

& M A R T I N O R O M A N I E L L O *

1. Introduction. In the last few yearsthe study of light echoes in Super-novae (SNe) has become rather fash-ionable, since they provide a poten-tial tool to perform a detailed tomog-raphy of the SN environment (see, forexample, Sugerman & Crotts 2002)

and, in turn, can provide importantinsights into the progenitor’s nature, amatter which is still under debate.

Due to the typical number densityof dust particles which are responsi-ble for the light scattering, the echoesare expected to have an integrated

brightness about ten magnitudesfainter than the SN at maximum. Forthis reason, a SN Ia in the Virgo clus-ter is supposed to produce, if any, anecho at a magnitude V~21.0. As asimple consequence, it is much easierto observe such a phenomenon in anIa than in any other SN type, due toits high intrinsic luminosity. As a mat-ter of fact, only four cases are known:the SNe Ia 1991T (Schmidt et al. 1992,Sparks et al. 1999) and 1998bu (Cap-pellaro et al. 2000), and the IISN1987A (Xu et al. 1994) and 1993Jtypes (Sugerman & Crotts 2002). Asexpected, the light echo detections forthe two core-collapse events occurredin nearby galaxies: LMC (d=50 Kpc)and M81 (d=3.6 Mpc), respectively.

While a dusty environment is ex-pected to result from explosions aris-ing in massive and short-lived prog-enitors, this is in principle not thecase for the kind of stars which arecommonly supposed to generate Iaevents, i.e. old population and lowmass stars. This is why the detectionof substantial dust in the immediatesurroundings of thermonuclear SNe,or at least in a fraction of them, wouldindicate a different scenario.

Actually, several authors havepointed out that the observed charac-teristics of Ia’s, such as intrinsic lumi-nosity, color, decline rate, expansionvelocity and so on, appear to be relat-ed to the morphological type of thehost galaxy (van den Bergh & Pazder1992; Hamuy et al. 1996, 2000; How-ell 2001). Since these objects repre-sent a fundamental tool in Cosmology(see for instance Leibundgut 2000), itis clear that a full understanding of

the underlying physics is mandatoryin order to exclude possible biases inthe determination of cosmologicalparameters like Ω and Λ. In thisframework, disentangling betweenpossible Ia sub-classes is a fundamen-tal step.

In this respect, an important yearin the SN history is 1991, when twoextreme objects were discovered, i.e.SN1991T (Filippenko et al. 1992a)and SN1991bg (Filippenko et al1992b). The former was an intrinsi-cally blue, slow declining and spec-troscopically peculiar event, while thelatter was intrinsically red, fast declin-ing and also showing some spectralpeculiarities.

From that time on, several otherobjects sharing the characteristics ofone or the other event were discov-ered, indicating that these deviationsfrom the standard Ia were not so rare.Of course, one of the most importantissues generated by the discovery ofsuch theme variations concerned theexplosion mechanism. The growingevidence produced by the observa-tions in the last ten years has clearlydemonstrated that subluminousevents (1991bg-like) are preferentiallyfound in early type galaxies (E/S0),while the superluminous ones (1991T-like) tend to occur in spirals (Sbc or lat-er). This has an immediate conse-quence on the progenitors, in thesense that subluminous events appearto arise from an old population, whilesuper-luminous ones would ratheroccur in star-forming environmentsand would thus be associated with ayounger population. This importanttopic has been recently discussed in a

N . P a t a t , G . C a r r a n o & M . R o m a n i e l l o

26

work by Howell (2001), to which thereader is referred for a more detailedreview.

What is important to emphasizehere is that 1991T-like events tend tobe associated with young environ-ments and are, therefore, the mostpromising candidates for the study oflight echoes. Or, in turn, if lightechoes are detected around such kindof SNe, this would strengthen theirassociation with sites of relatively re-cent star formation.

The first case of light echo in an Ia(1991T) seems to confirm this hy-pothesis, in the sense that the SN wasoverluminous and the host galaxy(NGC4527) is an Sbc and a liner too.Slightly less convincing is the otherknown case (1998bu), since thegalaxy (NGC3368) is both an Sbcand a liner, but the SN is neitheroverluminous nor spectroscopicallypeculiar. The only characteristic incommon with SN1991T is its declinerate ∆m15, which is lower than aver-age, even though not so extreme as inthe case of 1991T. Nevertheless, theHST observations by Garnavich et al.(2001) show that some significantamount of dust must be present with-in 10pc from the SN.

Of course, no statistically signifi-cant conclusion can be drawn fromsuch a small sample, which definitelyneeds to be enlarged. For this reason,during the past years, we have beenlooking for new cases, the mostpromising of which is represented bySN1998es in NGC632. This SN, infact, was classified as a 1991T-like byJha et al. (1998) who also noticed thatthe parent galaxy was classified as an

S0, hosting a nuclear star-burst (Pogge& Eskridge 1993). Moreover, the SNwas reported to be projected very closeto a star-forming region and an Echellespectrum obtained with FEROS at LaSilla (Patat et al. in preparation)showed very strong NaID lines at theSN rest frame velocity. The latter indi-cates the presence of significant ab-sorbing material in front of the SN.

All these facts were pointing to-wards the possible appearance of alight echo. If this were of the same na-ture as those seen in 1991T and1998bu, due to the higher distance ofNGC632 (d=42.2 Mpc, H0=75 km s-1

Mpc-1), it would be expected to havea magnitude V~24. Due to the rela-tive faintness of the target, the obser-vations were carried out with one ofthe 8.2m ESO VLTs equipped withFORS2.

In the following we report aboutthe results from these observationswhich, if on the one hand did not leadto light echo detection, on the otherthey brought to our attention the ex-istence of a very faint filament, ex-tending for at least 160,000 light yearsfrom the nucleus of NGC632 in theSW direction. While the implicationson the SN environment will be dis-cussed in a forthcoming paper, wewill focus on a preliminary analysis ofthe tidal tail and the inner morpholo-gy of NGC632 in connection with itsstar-burst nature.

2. Observations and Data Reduc-tion. NGC632 was observed in B andV Bessel filters on September 25,2001 with FORS2 mounted at theCassegrain focus of the ESO 8.2m


27

UT3-Yepun2 telescope, located atCerro Paranal, Chile. FORS2 isequipped with a 2048x2048 pixel(px) TK2048EB4-1 backside thinnedCCD. When, as in our case, the stan-dard resolution collimator is used, aprojected scale of 0″.2 per pixel (24µm x 24 µm) is obtained, which cor-responds to a field of view of 6’.8 x6’.8. To allow for a proper removal ofcosmic-ray hits, three images of 300sec and 600 sec each were obtained ineach passband; an image quality ofabout 0″.62 and 0″.75 in B and V rawframes was achieved respectively.

The images were bias and flat-field corrected by the FORS pipeline,while frame registration and stackingwere performed using IRAF3. The fi-nal image quality achieved in both Band V stacked images is about 0″.64and 0″.78, while the limiting magni-tude for unresolved objects (5σ peaklevel) is about 27.0 and 25.0 in B andV, respectively.

Photometric calibration (zeropoints and color terms) was achievedthrough the observation of standardfields (Landolt 1992).

The sky background was comput-ed using the mode of the pixel inten-sity distribution, using areas of theimages free from extended objects.Due to the FORS2 field of view,which is much larger than the actualextension of NGC632, it was possibleto select large regions where the in-tensity distribution was dominated bybackground emission.

In this kind of studies, the so-called unsharp masking is widely used(see, for example, Peng et al. 2002).The application of this technique has

the effect of revealing fine structuresoverimposed on smooth back-grounds, which would otherwise behardly detectable.

Since NGC632 has structureswith extremely different scales (seenext section) and intensities, weadopted another procedure, which isbasically a differential compression ofthe image dynamics in order to bringhighly differently exposed regions toa similar contrast range.

We achieved this by using the ra-dial profile of the diffuse componentof the galaxy by means of the medianvalue calculation within ellipticalcoronae centered on the galaxy nucle-us. For the position angle and ellip-ticity we adopted the values deducedby Chitre & Joshi (1999) from theisophotes’ fitting. Using this profilewe then constructed a bi-dimensionalfunction having the same averageisophotal properties of the galaxy. Fi-nally, we divided the original imageby this function scaled by a suitableconstant and thus applied a differen-tial compression factor, which in theend brings all regions to a similar dy-namical range.

This process, of course, does notconserve the flux and is not suitablefor photometry, but allows one to dis-play on a single plot features whichoriginally had a very large intensityrange (see Figure 5).

In fact, as we will see in the nextsection, NGC632 has a stronglypeaked diffuse component, on top ofwhich stellar-like knots, spiral armsand other diffuse structures withcomplex morphology are overim-posed.


28


29

Figure 1. VLT+FORS2 B image of NGC632 obtained by stacking 3 frames of 600 secondseach. The field of view is 7’x7’; North is up and East to the left. The original images were ob-tained on September 25, 2001 with a seeing of 0″.6. To enhance the tidal tail structure, the com-bined image was binned 2x2. At the distance of NGC632, 1’ corresponds to about 12.3 kpc.

3. Results. The images were reducedand analyzed immediately after theobservations to detect the light echoand follow it up with spectroscopy. Acareful inspection of the SN location,though, did not show any trace of apoint source down to B=25.5, i.e. 1.5magnitudes fainter than a 1991T-likeecho. Nevertheless, a re-analysis ofthe data that was carried out aboutone year later, revealed the presenceof two very faint tidal tails, one ofwhich clearly extending beyond the

FORS2 image. This is clearly shownin Figure 1, where we set the imagecuts to rather extreme values in orderto show the faint structures. Thelonger tail, coming out in the SE di-rection, has a typical intensity ofabout 20 counts above the sky back-ground in B, which turns into a sur-face brightness of about 27 mag arc-sec-2. At the NGC632distance thisimplies a projected length of at least50 kpc, which is enormous, beingcomparable to the distance from the

Large Magellanic Cloud. The othertail is brighter but much shorter andshows a widening at its end.

Another fact which is clearly dis-played in Figure 1 is the slight defor-mation of the galaxy in the SW direc-tion, which becomes more evident inFigure 2, where we present a contourplot for the B band. There are severalinteresting facts that emerge from thisFigure. The first is the clear detectionof a well-defined spiral structure,

within 20″ from the nucleus. In thisrespect, we noticed that the existenceof a rudimentary spiral structure waspointed out by Chitre & Joshi (1999).The second aspect concerns theshape of the isophotes, which tend totwist counterclockwise and to be-come more and more box-shaped.Chitre & Joshi (1999) found a similarresult, even though their data had amuch smaller signal-to-noise ratioand a lower spatial resolution which


30

Figure 2. B contour plot of NGC632. The saturated part of the nucleus is indicated by theinner contour, while the outer contour is at a 3σ level on the sky background.

made the conclusion rather question-able (see their Fig. 1h). They inter-preted these two findings as a possi-ble indication for a merger and ourdetection of a tidal tail seems indeedto agree with their suggestion.

Finally, the isophotes, especially inthe outer parts, are distorted in theSW direction and this breaks thesymmetry around the nucleus, whichroughly holds up to a semi-minor ax-is of about 17″. At larger distances a

clear deformation is seen and thejoining of the smaller tidal tail withthe galaxy appears at about 30″ NWof the nucleus.

Probably, the complex structureof this galaxy is better shown in Fig-ure 3, where we applied to our Bstack the dynamics compression tech-nique described in the Introduction.The spiral structure appears veryclearly and it is not confined to the in-ner regions only. Close to the nucleus


31

Figure 3. NGC632 after the application of dynamic compression (see the Introduction).

several bright point sources are visi-ble on top of the spiral arms; in par-ticular, there is a short filament,slightly off-centered with respect tothe nucleus, which points towards thestarting point of the upper tidal tail.

Even though the noise producedby the high level of the diffuse galac-tic component tends to mask it, onecan follow this tail until it almostreaches the spiral structure. Besides

the spiral arms visible in the inner re-gion, we notice the presence of an al-most symmetric elliptical ring with asemi-minor axis of about 19″, whichcauses a kink in the radial profile (seealso Figure 5). The NW symmetrybreaking is evident in Figure 3: itstarts with an enhancement of the in-tensity in this elliptical ring and is fol-lowed by the appearance of a diffuseand extended component, which


32

Figure 4. (B-V) color image of NGC632 computed from the two B and V stacks. Thegreyscale is linear and displays the color range 0.1-1.1 (black-white). The white region in thecenter is the saturated part of the nucleus.

seems to join clockwise the SE tidaltail at its origin. Another relevant as-pect in Figure 3 is the presence of se-mi-stellar objects on top of the diffusecomponent in the outer regions of thegalaxy. These are probably star clus-ters, which certainly deserve a multi-color study.

Even though a complete set ofdeeper UBVRI images would be nec-essary to analyze the stellar popula-tion present in the galaxy, some inter-esting results can be obtained alreadyfrom the inspection of the (B–V) col-or image (see Figure 4). Several unre-solved, very blue objects with 0.1 ≤(B–V) ≤ 0.4 are detected on top of theinner spiral structure, while the dif-fuse component appears to have atypical (B–V)~0.7. For comparison,the color of the nuclear region is(B–V)=0.38 (Chitre & Joshi 1999),consistent with the estimate we getfrom a few unsaturated pixels in ourimages. NGC632 is known to showHα emission both in the nucleus andin the clumpy circumnuclear region(Balzano 1983) and this has been in-terpreted as a sign of starburst activi-ty (Pogge & Eskridge 1993). For thisreason, we tend to believe that theblue spots we see in the inner regionare sites of recent star formation. Be-sides these knots and the diffuse com-ponent, Figure 4 shows some redderfilaments with a typical (B–V)=0.95,which are probably due to the pres-ence of dust.

From what we have been dis-cussing so far, it is clear that we aredealing with two kinds of compo-nents, i.e. a very faint spiral structureand an underlying diffuse and

smooth component. To get the ap-proximate properties of the latter, wecomputed its radial profile within el-liptical coronae, assuming the posi-tion angle and ellipticity given byChitre & Joshi (1999), i.e. 163° and0.26 respectively. The result is shownin Figure 5, where we plotted the ra-dial surface brightness profiles of thediffused component as a function ofthe isophotes semi-minor axis for Band V passbands (upper panels).

Due to the large number of pixelsincluded in the elliptical coronae atlarge radii, it is possible to follow theprofile with a reasonable accuracy outto at least 50″ which, at the distanceof NGC632, correspond to about10.4 kpc. In these outer regions, thesurface brightness is about 27.7 and27.2 mag arcsec-2 in B and V, respec-tively.

With the exception of the inner 5″(~1 kpc), which are dominated by thepeaked nucleus, both profiles showan exponential decrease out to 35″(~7 kpc), where they tend to bend.The very low signal to noise, however,does not make it possible to establishwhether this is due to the onset of anadditional component following adifferent law. Here we only note thatthe exponential disk we see in the in-ner part of NGC632 is typical of spi-ral galaxies.

One remarkable feature dis-played by both B and V profiles isthe appearance of a kink at about17″ (3.5 kpc), which corresponds tothe elliptical ring we have mentionedbefore.

Finally, in the lower panel of Fi-gure 5, we have shown the (B–V) color


33

profile for the diffused component,which appears to be roughly con-stant, at a value of (B–V)=0.7.

This integrated color is consistentwith a population of stars whoseturn-off mass is of the order of 0.8 so-lar masses and an age of some Gyr(Bressan, Chiosi & Fagotto 1994).

On average, the color within 20″from the nucleus is 0.1 mag redderthan in the outer regions, even

though the large error bars do not al-low a firm conclusion.

4. Discussion and conclusions.NGC632 is classified as E1 in theMarkarian catalogue (Mrk 1002) andas S0 by Balzano (1983), while wehave shown that there are indeedtraces of spiral arms, even thoughthey are masked by the dominant dif-fuse component.


34

Figure 5. Upper panel: surface brightness radial profile for the diffuse component in B (left)and V (right). The dashed lines trace a fitting with an exponential law. Lower panel: (B–V)profile of the diffuse component. The dotted line is placed at the average value.

Traditionally, S0 galaxies havebeen interpreted as fossile disk sys-tems, with a small amount of gas andare therefore considered as dead fromthe point of view of star formation.This scenario has started to change,since more and more evidence of starforming activity is being gathered.The case of NGC632 is emblematic:the nucleus shows a strong HII re-gion spectrum (Balzano 1983) andprobably hosts a nuclear starburstwith considerable circumnuclear starformation (Pogge & Eskridge 1993).Of course, one of the main issues hereis how such a dead galaxy can gener-ate massive stars, given the fact that avery small amount of gas is supposedto be left.

Possible explanations have beenprovided. For example, Larson &Tinsley (1978) have first proposedthat tidal forces in interacting galax-ies could trigger such events, whileBournaud & Combes (2002, and ref-erences therein) have shown that inthe presence of bar-like structures,gravitational torques can drive the gasleftovers into the central regions andgive similar results.

In particular, the most active star-bursts are found almost exclusively inon-going mergers (Hibbard 1997 andreferences therein), a fact that clearlyestablishes the link between the twophenomena.

In the case of NGC632, Chitre &Joshi (1999) have already pointed outthat the cause of the observed star-forming activity was probably the re-mote merging of a gaseous disk intothe main galaxy. Our tidal tail detec-tion gives a definite support to this

hypothesis and strengthens the ideathat mergers are responsible for star-bursts even in the cases where directevidence of close galaxy encounters(v.g. strong asymmetries, tidal tailsand double nuclei) has been totallyspoiled. A deep survey of S0 starburstgalaxies carried out with an 8m-classtelescope could probably demon-strate whether this is the case or not.

Since our data were obtained for acompletely different purpose, we arecertainly not in the conditions of ful-ly solving the case of NGC632. Thereis in fact a number of fundamentalpoints which need to be addressed.

First of all, a full set of UBVRIvery deep imaging is required toproperly study the stellar and globu-lar cluster populations both in thegalaxy and in the tails, solving theage-metallicity degeneracy. This re-quires several hours of integrationtime per filter even with an 8m-classtelescope in order to get a sufficientsignal-to-noise ratio in the outer partsof the galaxy and in the tidal tail. Thelatter needs also to be imaged at larg-er distances, to get its full extension.

Another important piece of infor-mation would be given by deep Hαimaging, which will allow one to de-tect all emitting regions to be studiedlater on with spectroscopy. We sus-pect that the blue knots we detectedin the inner part of the galaxy are cir-cum-nuclear regions of active starformation. High-resolution spec-troscopy of those clumps would un-veil their nature, provide fundamen-tal hints on their chemical composi-tion and physical conditions. Finally,a kinematical map of the inner re-


35

gions would provide us with a funda-mental tool for the reconstruction ofthe merging-event dynamics.

Nowadays hierarchical clusteringof galaxies is reckoned as one of the

fundamental large-scale processesin Astrophysics. Therefore, the studyof cases where this mechanism is stillat work is of great relevance for un-derstanding Universe evolution.


36

Notes

1 Based on observations collected at ESO-Paranal.2 Yepun in the Mapuche word for Sirius.

3 IRAF is distributed by National Optical As-tronomy Observatories, which is operated by theAssociation of Universities Inc. (AURA) undercooperative agreement with the National ScienceFoundation, U.S.A.


37

References/ Bibliografie

Bournaud F., Combes F. (2002). Gas accretion on spiral galaxies: bar formation and renewal.Astronomy & Astrophysics, in press (astro-ph/0206273).

Bressan A., Chiosi C., Fagotto F. (1994). Spectrophotometric evolution of elliptical galaxies.I. Ultraviolet excess and color-magnitude-redshift relations. Astrophysical Journal Sup-plements, 94: 63-115.

Cappellaro E., Patat F., Mazzali P.A., Benetti S., Danziger I.J., Pastorello A., Rizzi L., SalvoM. & Turatto M. (2001). Detection of a light echo from SN 1998bu. Astrophysical Jour-nal, 549: L215-218.

Balzano V.A. (1983). Star-burst galactic nuclei. Astrophysical Journal, 268: 602-627.Chevalier R.A. (1986). The scattered light echo of a supernova. Astrophysical Journal, 308:

225-231.Chitre A. & Joshi U.C. (1999). Photometric studies of some starburst galaxies. Astronomy

and Astrophysics Suppl. Ser., 139: 105-133.Filippenko A.V. et al. (1992a). The peculiar type Ia SN 1991T- Detonation of a white dwarf?

Astrophysical Journal, 384: L15.Filippenko A.V. et al. (1992b). The sublumious, spectroscopically peculiar type Ia Superno-

va 1991bg in the elliptical galaxy NGC4374. Astronomical Journal, 104: 1453.Garnavich P., et al. (2001). Echos is SN 1998bu. American Astronomical Society Meeting,

199, #47.01.Hamuy M., Phillips M.M., Suntzeff N.B., Schommer R.A., Maza J. & Aviles R. (1996).The absolute luminosities of the Calan/Tololo type Ia supernovae. Astronomical Journal, 112:

2391-2397.Hamuy M., Trager S.C., Pinto P.A., Phillips M.M., Schommer M.M., Ivanov V. & Suntzeff

N.B. (2000). A search for environmental effects on type Ia supernovae. AstronomicalJournal, 120: 1479-1486.

Hibbard J.E. (1997). Mergers, interactions, and the fueling of starbursts. In Star formationnear and far: seventh astrophysics conference. In Host S.S., Mundy L.G. (Ed.). WoodburyN.Y.: AIP Conference Series, v. 393, pp. 259.

Howell D.A. (2001). The progenitors of subluminous type Ia supernovae. Astrophysical Jour-nal, 554: L193-196.

Jha S., Garnavich P., Challis P. & Kirshner R.P. (1998). Supernova 1998es in NGC632.IAU Circ. N. 7054.Landolt A.U. (1992). UBVRI Photometric Standard Stars. Astronomical Journal, 104: 340-491.Larson R.B. & Tinsley B.M. (1973). Star formation rates in normal and peculiar galaxies. Astrophysical Journal, 219: L46-59.Leibundgut B. (2000). Type Ia Supernovae. Astronomy and Astrophysics Review, 10: 179-209.Peng E.W., Ford H.C., Freeman K.C. & White R.H. (2002). A young blue tidal stream in

NGC 5128. Astronomical Journal, in press. Astro-ph/0208422.Pogge R.W. & Eskridge P.B. (1993). Star formation in the disks of HI-rich S0 galaxies. Astro-

nomical Journal, 106: 1405-1419.Schmidt B.P., Kirshner R.P., Leibundgut B., Wells L.A., Porter A.C., Ruiz-Lapuente P., Chal-

lis P. & Filippenko A. (1994). SN 1991T: reflections of past glory. Astrophysical Journal,434: L2-23.

Sparks W.B. (1994). A direct way to measure the distances of galaxies. Astrophysical Journal,433: 19-28.

Sparks W.B., Macchetto F., Panagia N., Boffi F., Branch D., Hazen M.L. & Della Valle M.(1999). Evolution of the light echo of SN 1991T. Astrophysical Journal, 523: 585-592.

Sugerman B.E.K. & Krotts A.P.S. (2002). Multiple light echoes from Supernova 1993J.Astrophysical Journal, in press. Astro-ph/0207497.

Van den Bergh S. & Pazder J. (1992). The luminosity of supernovae of type Ia – Apparentmagnitudes at maximum light. Astrophysical Journal, 390: 34-38.

Xu J., Crotts A.P.S. & Kunkel W.E. (1994). Two new light echo structures from SN1987A atlarge distances. Astrophysical Journal, 435: 274-278.


38

serendipitous discovery of a tidal tail in ngc632 · found in early type galaxies (e/s0), while the...

Documents