Star Formation and the Interstellar Medium in Nearby Tidal Streams (SAINTS)
S. J. U. Higdon, J. L. Higdon, B. J. Smith, M. Hancock, C. Struck
aa r X i v : . [ a s t r o - ph . C O ] D ec **FULL TITLE**ASP Conference Series, Vol. **VOLUME**, c (cid:13) **YEAR OF PUBLICATION****NAMES OF EDITORS** Star Formation and the Interstellar Medium in NearbyTidal Streams (SAINTS)
S. J. U. Higdon, J. L. Higdon, B. J. Smith, M. Hancock, and C.Struck Abstract.
We compare Spitzer Infrared Spectrograph observations of SQ-A & SQ-Bin Stephan’s Quintet, Ambartzumian’s knot in Arp 105, Arp 242-N3, Arp 87-N1,a bridge star forming region, NGC 5291 N and NGC 5291 S. The PAHs tendto be mainly neutral grains with a typical size of 50 - 100 carbon atoms. Theinterstellar radiation field is harder than typical starburst galaxies, being similarto that found in dwarf galaxies. The neon line ratios are consistent with a recentepisode of star formation ∼ < ∼ M ⊙ of warm H in SQ-A and Arp 87N1 and ∼ M ⊙ in SQ-B. These resultsare similar to those derived for the tidal dwarf galaxies (TDGs) NGC 5291N and NGC 5291 S and are consistent with emission from photodissociationregions. Using our 8 µ m images of 14 interacting systems we identify 62 tidalstar forming knots (TSFKs). The estimated stellar masses range from superstar cluster (10 − M ⊙ ) to TDG ( ∼ M ⊙ ) sizes. The median stellar massis 10 M ⊙ . The stellar mass, with some scatter, scales with the 8 µ m luminosityand tends to be an order of magnitude smaller than the KISS sample of starforming dwarfs. An exception to this are the TSFKs in Arp 242 which havestellar masses similar to the KISS dwarfs. The TSFKs have “notched” 3.6 - 8 µ m spectral energy distributions (SEDs) characteristic of star forming regions.The TSFKs, form two distinct clumps in a mid-infrared color diagram. There are38 red-TSFKs with [4 .
5] - [8 . > .
6] - [4 . < < [4 .
5] - [8 . < .
6] - [4 . < .
5] - [8 .
0] population tends to have the sources with a rising 8-24 µ m SEDwhile the blue population tends to contain the sources with a descending SED.The rising SED is typical of spiral and starburst galaxies with a dominant 40 −
60 K dust component and the declining SED probably indicates a dominant hotdust component. Georgia Southern University, Department of Physics, Statesboro, GA 30458 East Tennessee State University, Department of Physics and Astronomy, Johnson City, TN37614 University of California, Riverside, Riverside, CA 92521 Iowa State University, Department of Physics and Astronomy, Ames, IA 50011 Higdon et al.
In addition to triggering starbursts and active galactic nuclei, mergers of dusty,gas rich disk galaxies frequently lead to the formation of tidal tails that canstretch many disk diameters from the site of the collision (Toomre 1972; Schweizer1972; Sanders & Mirabel 1996). These structures tend to be HI rich withblue optical colors, reflecting both their origin in the outer spiral disk and on-going star formation (van der Hulst 1979; Schombert, Wallin, & Struck 1990;Mirabel, Lutz, & Maza 1991; Hibbard & van Gorkom 1996). Zwicky (1956) pro-posed that dwarf galaxies might form out of self-gravitating clumps within tidaltails, and indeed, concentrations of gas and star forming regions are commonlyfound there, ranging in size from super star clusters (SSCs, 10 − M ⊙ ) totidal dwarf galaxies (TDGs, ∼ M ⊙ ).Tidal star forming knots (TSFKs) and the formation of gravitationallybound TDGs, formed either via tidal interactions between the parent galaxiesor from ram-sweeping of debris material, may play an important role in galaxyformation and evolution. Higdon et al. (2006a), hereafter HHM06, found tidalbridges and tails associated with TDGs and TSFKs in NGC 5291, indicating fur-ther tidal interaction amongst the star forming knots. TDGs may be useful aslocal analogs of the multiple mergers of small dwarf-like galaxies at high redshift.More importantly, dwarf galaxies are the most common galaxy type in the cur-rent epoch, and TDGs may contribute significantly to this population in some en-vironments, for example, in compact groups (Hunsberger, Charlton, & Zaritsky1996).Here we present some new results from our Spitzer study of Star Forma-tion and the Interstellar Medium in Nearby Tidal Streams (SAINTS). We haveselected 12 pre-merger binary pairs with prominent optical tails and/or bridges(Smith et al. 2007; Higdon & Higdon 2008). This is complemented by our studyof two more complex systems: NGC 5291 where ram-sweeping is occurring inaddition to a tidal interaction, and TSFKs in Stephan’s Quintet. Using theSpitzer 8 µ m images we have identified 62 TSFKs. We discuss the InfraredSpectrometer (IRS, Houck et al. 2004) observations of seven bright TDG can-didates along with the Infrared Array Camera (IRAC, Fazio et al. 2004) andMultiband Imaging Photometer for Spitzer (MIPS, Rieke et al. 2004) images ofthe full sample and address some key questions concerning the nature of TDGsand star formation in tidal streams. Figure 1 shows the IRS Short-Low (IRS-SL) observations of NGC 5291 N andNGC 5291 S (from HHM06), SQ-A & SQ-B in Stephan’s Quintet and Am-bartzumian’s knot in Arp 105 (Higdon & Higdon 2008), and Arp 242-N3, andArp 87-N1, the bridge star forming region in Arp 87. The TDGs/TSFKs are richin atomic and molecular emission features from the ISM, including fine structurelines, e.g., [Ne ii ] 12.81 µ m, [Ne iii ] 15.56 µ m, polycyclic aromatic hydrocarbons(PAHs) and warm H . The mid-infrared spectra of dwarf galaxies are knownto differ substantially from those of spirals, with weaker PAH emission featuresand higher [Ne iii ]/[Ne ii ] ratios (Madden et al. 2006). However, it is unknown AINTS
Wavelength ( � m) F l u x D e n s i t y ( m J y ) + c o n s t . Figure 1. IRS-SL Spectra in the rest-frame wavelength. From top to bot-tom, NGC 5291 N, NGC 5291 S, Stephan’s Quintet-A, Stephan’s Quintet-B,Arp 87-N1, Ambartzumian’s Knot in Arp 105 & Arp 242-N3.
Draine & Li (2001) model the ratio of the 6.2 to 7.7 µ m features, which are bothfrom C-C stretching modes, to derive the PAH ion / grain size. The number ofcarbon atoms decreases as the ratio increases. The 11.3 µ m feature is from a C-H out of plane bending mode and the ratio of the 11.3 to 7.7 µ m band strengthgives a measure of the PAH ion fraction. The ratio of neutral to ionized PAHsincreases as the ratio increases. The PAHs in the TSFKs tend to have a strongC-H out-of-plane bending mode resulting in a mainly neutral population of PAHgrains. PAH N ∼
80 - 90 % in Ambartzumian’s knot, Arp 242-N3, SQ-B and Arp87-N1 and around PAH N ∼
60 % in the remainder of the sample. Arp 242-N3and Ambartzumian’s knot in Arp 105 have small PAHs, N C ∼
50. The restof the sample have larger PAHs, N C ∼
100 (the model fit to SQ-A is not wellconstrained).
The TSFKs in SQ-A and SQ-B have [S iv ] / [S iii ] and [Ne iii ] / [Ne ii ] line ra-tios indicating moderate excitation consistent with our earlier results for NGC5291 N and NGC 5291 S (HHM06) and with the overlap region in the Anten-nae (Verma et al. 2003). The ISRF in TSFKs is harder than typical starburst Higdon et al. galaxies, being similar to that found in dwarfs, but softer than the ISRFs foundin the extreme low metallicity Blue Compact Dwarfs (Verma et al. 2003). Theneon line ratios in the TSFKs are consistent with the Starburst99 models inThornley et al. (2000) for a recent episode of star formation ∼ < do the TSFKs Contain? We detect emission from ∼ M ⊙ of warm H in SQ-A and Arp 87-N1 and ∼ M ⊙ in SQ-B. These results are similar to those derived by HHM06 forNGC 5291 N and NGC 5291 S (HHM06) The warm H masses of 10 − M ⊙ are about 100 - 1000 times smaller than the average warm molecular mass of 2 × M ⊙ measured in a sample of 59 ULIRGs (Higdon et al. 2006b), but theresults for both galaxy types are consistent with an origin in photodissociationregions (PDRs). Arp 105 Arp 242 Stephan's Quintet
Figure 2. IRAC images of (left-right) Arp 105, Arp 242 & Stephan’s Quin-tet. Top row 3.6 µ m and bottom row 8 µ m. TSFKs are identified with awhite circle. North is top and East is left. A square root transform is used. Figure 2 shows three examples of the 3.6 µ m and 8 µ m images for Arp 105,Arp 242 (The Mice) and Stephan’s Quintet. The locations of the TSFKs areindicated with white circles on the figures. 95% of the sources in the 14 systemshave a “notched” IRAC SED characteristic of star forming regions. 79% have 24 µ m detections or meaningful upper limits. These are classified into four groupsdepending on the 8 - 24 µ m slope. There are 18 rising (f − f > . × f ), 12declining SED (f − f > . × f ), 9 flat/slowly rising (f > f and f − f < . × f ), and 8 flat/slowly declining (f < f and f − f < . × f ). Therising SED is typical of spiral and starburst galaxies with a dominant 40 −
60 Kdust component and the declining SED probably indicates a dominant hot dustcomponent. Far-infrared observations are needed to confirm this interpretation.Figure 3 shows that TSFKs form two distinct clumps on an IRAC 2-colordiagram. The first group has 38 red-TSFKs with [4 .
5] - [8 . > . AINTS Figure 3. The horizontal and vertical lines mark the color zero-points (i.e.where the two flux densities are equal). The lower right quadrant containssources with active star formation. The TSFKs are coded according to 8- 24 µ m SED shape: Rising - filled diamond, Flat/Rising - diamond; De-scending - solid square; Flat/Descending - square; unknown - X. For com-parison the Pahre et al. (2004) galaxy sample are shown as crosses and theRosenberg et al. (2006) star forming dwarfs are shown as solid circles. - [4 . < < [4 . . < .
6] - [4 . < .
5] -[8 .
0] population tends to have the sources with a rising 8-24 µ m SED while theblue population tends to contain the sources with a descending SED. µ m Luminosity? Preliminary stellar masses are estimated using the 4.5 µ m data (e.g., Oh et al.2008) and assuming minimal non-stellar emission in the band. We find a medianM ∗ = 10 M ⊙ , with a wide range: 4.0 × to 1.3 × M ⊙ . The 18 risingsources have a median stellar mass of 1.0 × M ⊙ . The 12 declining sourceshave a higher median stellar mass of 4.0 × M ⊙ . Half of the declining sourcesare in Arp 242, which probably biases the median stellar mass to be greater thanthe average for the whole sample. The sources with flat or slowly rising/decliningSEDs have stellar masses similar to that of the rising sources.The stellar mass, with some scatter, tends to scale with the 8 µ m luminos-ity. For a given 8 µ m luminosity the KISS dwarfs have a more massive stellarcomponent. The median stellar mass for the KISS dwarfs is 1.0 × M ⊙ , whichis a factor of ten larger than the median value for the TSFKs. 6/8 TSFKs in Higdon et al.
Arp 242 have declining SEDs. Four of these TSFKs overlap with 3 of the dwarfsfrom the KISS sample.
Studies of TDGs and smaller knots of star formation in tidal features help usbetter understand the dwarf galaxy population as a whole. The mid-infraredemission from the ISM in TSFKs are rich in atomic, molecular and PAH emissionfeatures. The study of these systems is still in its infancy as few have beenstudied in detail. The next generation of telescopes, for example Herschel, JWSTand ALMA offer the promise of studying both fainter objects and larger samples.In particular, such studies will provide information about what fraction are trulyprimordial building blocks of massive galaxies left over from an early epoch ofgalaxy formation, and what fraction may have been born in tidal interactions.In time we will perform a more comprehensive census of star formation andthe interstellar medium in tidal streams and be able to refine models of bothintergalactic enrichment and star formation triggering and regulation.
Acknowledgments. research was supported by Spitzer/NASA grants RSANo.s 1346930 (Higdon & Higdon), 1353814 (Smith) & 1347980 (Struck).