Spontaneous and Stimulated Star Formation in Galaxies
Barry F. Madore, Samuel Boissier, Armando Gil de Paz, Erica Nelson, Kristen Petrillo
aa r X i v : . [ a s t r o - ph ] J a n **FULL TITLE**ASP Conference Series, Vol. **VOLUME**, **YEAR OF PUBLICATION****NAMES OF EDITORS** Spontaneous and Stimulated Star Formation in Galaxies:Ultraviolet Limits on Star Formation Thresholds andOptical Constraints on Lambda-CDM CosmologicalSimulations of Galaxy Formation.
Barry F. Madore
The ObservatoriesCarnegie Institution of WashingtonandNASA/IPAC Extragalactic Database (NED)California Institute of TechnologyPasadena CA
Samuel Boissier
Labratoire d’Astrophysique de MarseilleMarseille, France
Armando Gil de Paz
Dept. de AstrofisicaUniversidad Complutense de MadridMadrid, E-28040 Spain
Erica Nelson & Kristen Petrillo
The ObservatoriesCarnegie Institution of WashingtonandPomona CollegeClaremont, CA
Abstract.
We present recent results from several on-going studies: The first addressesthe question of gas-density thresholds for star formation, as probed by the outerdisks of normal nearby galaxies. The second concerns the observational evidencefor the existence of gravitating non-luminous (GNL) galaxies, as predicted bymost recent simulations of galaxy formation in Lambda-CDM cosmologies. Wefind that (1) If star formation is traced by far-ultraviolet light, then there isno evidence for a threshold to star formation at any gas density so far probed,and (2) there is no evidence for GNL galaxies gravitationally interacting withknown optical systems based on the observations (a) that there are no ringgalaxies without plausible optically visible intruders, (b) all peculiar galaxiesin the Arp Atlas that are bodily distorted have nearby plausibly interactingcompanions, and (c) there are no convincingly distorted/peculiar galaxies withinKarachentsev’s sample of more than 1,000 apparently/optically isolated galaxies. Madore, Boissier, Gil de Paz, Nelson & Petrillo
1. Introduction
Galaxies continue to surprise us, both in their observed structure and by theirinferred evolution. Some of these surprises come from moving to new wave-lengths and re-observing old friends; while other surprises come staying at fa-miliar wavelengths but looking at the galaxies from a somewhat new perspectiveor in a slightly revised context.With the launch of the GALEX satellite into low-Earth orbit on April 28,2003 it has become possible to image a large number and a wide range of dif-ferent types of galaxies at largely unexplored wavelengths: in the near and farultraviolet. Given the high sensitivity of the detectors and the very low sky back-ground (especially in the far ultraviolet channel) it is possible to see features inthe ultraviolet out to surface brightness levels unequalled by ground-based opti-cal observations. Moreover, the wide (one-degree) field of view of GALEX alsomaximizes the possibility of serendipitous discovers, of which there have beenmany.
2. Radially Extended Star Formation
All of the above aspects of GALEX contributed to the “discovery” of extendedultraviolet (XUV) disks around a number of nearby spirals. Prime examples,found early in the mission are M83 (Thilker et al. et al.
Figure 1. NGC 4625. Right panel shows the optical image of the galaxy,while the left panel shows the NUV (GALEX) of this same galaxy where theextended UV disk of star formation is clearly highlighted tar-Formation Thresholds and Constraints LambdaCDM Cosmology While these features are striking, they should not have come as a surprise (al-thogh they did) as they are not unheralded. For instance, based UV imagingdata obtained from a balloon-borne experiment (SCAP 2000) Donas et al. (1981)asked the question “How Far Does M101 Extend?” Clearly they were seeing ex-tended UV features in the outer disk of a well known nearby galaxy. And laterFerguson et al. (1998), announced “The Discovery of Recent Star Formation inthe Extreme Outer Regions of Disk Galaxies” based on deep H α surveys of threenearby late-type galaxies: NGC 0628, NGC 1058 and NGC 6946. And all of thismight have been seen to be a natural consequence of the discovery of extendedHI disks around many galaxies (e.g., NGC 2915, Muere et al. 1996) except per-haps for the interesting interpretive paper by Martin & Kennicutt (2001) whichseemed to put an end to star formation in the outer disks of galaxies by titlingtheir paper “Star Formation Thresholds in Galactic Disks.” With hindsight,which we all now possess, it might have been more robust to have added “asTraced by HII Regions” to the title. Recently, using GALEX ultraviolet imaging data, Boissier et al. (2007) havere-examined the correlation of gas density with star formation ( as traced byFUV/NUV light ) and have come to very different conclusions regarding theexistence of any threshold to star formation at low gas densities. As can be seenin Figure 2 the left panel shows the run of star formation rates as a function oftotal gas density for a combined sample of 43 galaxies. The star formation rateis based on extinction-corrected UV surface brightness obtained by the GALEXsatellite and the total gas density combines neutral and molecular hydrogencontributions. The right panel shows a selection of individual galaxies wherethe run of UV and the H α star formation rates with surface density of gasare intercompared on an individual basis. Clearly there is a difference. Notruncation and no threshold is apparent in the UV data. The surface density ofstar formation as traced by hot, blue stars (which may or may not) include Ostars (which power the brightest HII regions) is continuous with the gas surfacedensity, to the limits of both surveys.The hard cut-off in H α has been challenged, of course, by Ferguson et al.(1988), but it may also be that other factors are in play. In the outer partsof galaxies we may be seeing small number statistics force the mass functionof the molecular clouds down to a mass level that while they can support starformation they are not individually large enough to produce even a single Ostar capable of ionizing the surrounding medium. The plane thickness mayhave grown so much, or the intercloud separation may be so large that thestar formation regions are density-bounded and that large fraction of the UVradiation is leaking out before it can produce a detectable HII region or that theradiation that is intercepted is only reradiated at a very low emission measureand perhaps and widely distributed. Very compact HII regions have been foundin the XUV disk of M83 and they are consistent with single-star ionization (Gilde Paz et al. 2005). Studies are underway to probe deeper and to lower surfacebrightness levels in search of any stray radiation. Madore, Boissier, Gil de Paz, Nelson & Petrillo
Figure 2. The left panel shows the radial drop-off of star formation ratesas a function of gas density for ** galaxies as reproduced from Boissier et al.(2007). The vertical grey zone shows the gas-density where the star-formationthreshold of Martin & Kennicutt (2001)is expected. For the UV data nothreshold is observed. The right panels shows an intercomparison of UV starformation rates (dotted lines) and H α star formation rates (solid lines) as afunction of radius (scaled to the Martin-Kennicutt threshold.) Again, the UVstar formation shows no sign of any thresholding at any gas density plotted.
3. Gravitating Non-Luminous (GNL) Galaxies
Most contemporary models of structure formation within the framework of a
Lambda -CDM cosmology predict a steep power-law increase of lower mass galax-ies fainter than L ∗ (e.g., Kravtsov, Gnedin & Klypin 2004). Most of the satel-lites are deemed “missing” because they not seen in optical surveys. This hasled to the speculation that they are currently somehow devoid of baryons, andthus invisible by construction. Invisible but not undetectable. The presence of Gravitating Non-luminous (GNL) galaxies can in principle (and in practice) beinferred (and seen) by their gravitational interactions with other nearby (visible)galaxies. Below we explore a few such tests for their presence ... or absence.
Twenty years ago Arp & Madore (1987) published a catalog and photographicatlas of peculiar galaxies based on a visual inspection of over 94,000 optical im-ages of southern-hemisphere galaxies. More recently Madore, Nelson & Petrillo(2007 in prep) extracted from that a pure sample of about 100 ring galaxies (seeFigure 3 for a sampling). According to models by Theys & Spiegel (1976) andby Lynds & Toomre (1976) they can be explained by galaxy-galaxy interactions,fine-tuned to be head-on collisions between a disk galaxy and a lower-mass in-truder. The ring galaxies culled from the Arp-Madore Catalogue were listedby those original discoverers because of their ring morphology not because they tar-Formation Thresholds and Constraints LambdaCDM Cosmology
Figure 3. Examples of ring galaxies and their adjacent companions fromthe soon to be published atlas of Madore, Nelson & Petrillo (2007)
While this is all good news for the theory of ring formation in specific, itis not so good news for the theory of galaxy formation in general. Withouta single convincing example of an isolated ring, one obvious conclusion is thatGNL galaxies do not exist in the numbers predicted by theory. While it may becounter-argued that ring galaxies require such peculiar circumstances for theirformation (mass ratios, orbital parameters and galaxy types, etc.) and that they
Madore, Boissier, Gil de Paz, Nelson & Petrillo can only arise from a collision involving concentrated baryonic satellite intrudersfor their formation, the same cannot be said for collision-induced peculiaritiesin general. The CDM simulations predict the observed number of optical galaxyonly at one mass. Above and below this there is a divergence at all massesbewteen theory and observation, reach more than a factor of 10 descrepancyat the fainest end. Wehether is is high-mass or low-mass companions that arebeing looked for in GNL-galaxy interactions they are predicted to be there atthe level of factors more than their optical counterparts rather than occasionallyoccurring at very low levels of incidence, as appears to be the case.
Turning back to the time at which the original
Atlas of Peculiar Galaxies waspublished (Arp 1966) it is fair to say that the topic of peculiar galaxies was stillin its formative stages and that the sample illustrated was not premised upontheir being or not being nearby galaxies to qualify them to be include in theAtlas. Indeed by moder standards many of the galaxies in the
Atlas are not nowconsidered to be particularly peculiar at all ( e.g., low-surface-brightness galaxies[ARP 001-004], dwarf irregular galaxies [ARP 005-006], and certain alignmentsin small groups and clusters [ARP 311-332]). Rare does not necessarily mean pe-culiar, but the
Atlas of Peculiar Galaxies did include many rare types of objects.Our point here is however that of the 338 objects that are included in the
Atlas because the galaxy in question is bodily distorted, is considerably asymmetricor has extended tails, the vast majority of those have very nearby companionsthat can easily be implicated as the source of the interaction and distortion. Itis the absence of isolated peculiar galaxies that is in itself noteworthy. Clearlythe vast majority of peculiarities seen as bodily deformations of the galaxy inquestion can be explained by interactions with nearby optical companions. Andthis simple fact leaves no room (and no evidence for) bodily deformed galaxiesresulting from their interaction with gravitating non-luminous (GNL) galaxies(i.e., pure dark-matter halos).
Figure 4. Arp 172 (left), Arp 107 (center) and Arp 173 (right) typify thetypes of bodily distorted galaxies in the Arp Atlas of Peculiar Galaxies thatalmost without exception have obvious interactions on-going between twooptically visible galaxies tar-Formation Thresholds and Constraints LambdaCDM Cosmology There is an alternative path to follow here. Karachentsev (1988) published alist of 1,000 galaxies that are optically isolated from other comparable-sized (op-tically visible) galaxies. This, of course, is not to say that apparently isolatedcannot and do not have GNL galaxies of comparable (or even larger) sizes orbit-ing and interacting with them. However this catalog would suggest that this isnot the case. With complete certainty we can say that out of the entire sampleof optically isolated galaxies there are no examples of Arp-like bodily-distortedsystems.There are a handful of isolated galaxies that are peculiar to some lesserdegree. But even this is to be expected without having to invoke GNL galaxies;mergers will deplete the apparent population of visible interactors while stillleaving evidence of the collision in the form of tidal debris or lingering asym-metries. For example, even a cursory examination of the 2MASS near-infraredimage of KIG 0022 (an object that has large ‘tidal’ arms in the optical) showsthat it has a double nucleus; presumably the result of a recent merger of two(previously) visible galaxies. Details of these samples and their analysis will begiven in a forthcoming paper (Madore. Petrillo & Nelson 2007).
4. Conclusions
Using UV light as a tracer for star formation, FUV and NUV imaging observa-tions of nearby galaxies using the GALEX satellite show a smooth and monotonicdecline of star formation with total gas surface density. No thresholding of starformation is visible in this sample, at these projected surface densities.The summary observations of peculiar galaxies viewed in the context of
Lamda -CDM simulations are as follows: (1) All cataloged ring galaxies haveplausible colliders that are optically visible. (2) All peculiar galaxies (in theArp Atlas) that are bodily deformed have visible nearby companions that areplausibly responsible for the interaction-induced deformities. (3) Virtually allisolated galaxies are not peculiar, distorted or interacting to any noticeabledegree.The peculiar galaxy study leads to the following general conclusions: (1)No ring galaxy is being produced from a head-on collision between a spiral andpure dark-matter GNL galaxy (2) No bodily deformed galaxies are the resultof collisions and/or near encounters between optical and pure dark-matter GNLgalaxies. (3) Optically isolated galaxies show no signs of bodily interactions withpure dark matter GNL galaxies.
Madore, Boissier, Gil de Paz, Nelson & Petrillo
Acknowledgments.
We sincerely thank all of our colleagues on the GALEXMission for their support in enabling the science being conducted by NASA’sGalaxy Evolution Explorer. Major portions of this research would not have beenpossible without the NASA/IPAC Extragalactic database (NED) which is oper-ated by the Jet Propulsion Laboratory, California Institute of Technology, undercontract with the National Aeronautics and Space Administration. Significantsupport for this research was also provided by the Observatories of the CarnegieInstitution of Washington.
References
Arp, H.C. 1966, Ap.J.Suppl., 14, 1Arp, H.C. & Madore, B.F. 1987,