UCDs in the Perseus Cluster core: evidence for mulple formaon mechanisms Samantha J. Penny

ARC Super Science Fellow Monash Centre for Astrophysics School of Physics, Monash University

Overview- What are UCDs?

• An extension of globular clusters to higher magnitudes and larger sizes

• Massive star clusters formed during gas rich mergers between galaxies (Fellhauer & Kroupa 2002, Brüns et al. 2011)

• The nuclear star clusters of stripped dwarf ellipcals- “Galaxy Threshing” (e.g. Bekki, Couch & Drinkwater 2001) The Perseus Cluster

D = 70 Mpc Not as well studied as Virgo, Fornax and Coma due to its low galacc latude (b = -13°) NGC 1275 • Surrounded by a complex system of Hα filaments

• Star clusters in the nucleus of NGC 1275 with masses 107 to 8 10 M¤ (e.g. Carlson et al. 1998)

Could galaxies like NGC 1275 be the formaon sites of the most massive globular clusters? NGC 1275- a possible formaon site

for UCDs? 118 R. E. A. Canning et al.

Table 1. Aperture corrections for the three ACS filters used.

Filter Aperture correction

ABC F435W 1.07 1.06 1.05 F550M 1.07 1.06 1.04 F625W 1.08 1.07 1.04 Note. ‘A’ corresponds to the Blue Loop, ‘B’ the Southern filament and ‘C’ the south-west portion of the nucleus. This correction was determined from the 10 brightest sources in each region.

1 than 0.3 counts s− . The variation of background to peak for this innermost cluster is less than 2 per cent. In the inner region, the bright galaxy nucleus was masked out, as was a bright saturated

star near the top of the eastern arm of the Blue Loop. Downloaded from Average aperture corrections are determined using the 10 bright- est sources in each of the three regions. These are corrections from a Image credit: NASA, ESA, and the Hubble Heritage (STScI/AURA)- Canning et al. 2010 2pixelradiusaperturecorrespondingtoa0.1arcsecradiusaperture ESA/Hubble Collaboraon flux to a flux determined with a 0.6 arcsec radius. These corrections are listed in Table 1. The corrections are similar but slightly larger http://mnras.oxfordjournals.org/ Figure 2. Image showing the star clusters (blue) and the Hα emission (red) than those found by Carlson et al. (1998) in the core of the galaxy. that make up the Blue Loop. Reproduced from Fabian et al. (2008). The aperture correction for the three stellar regions was determined èA number of star forming regions in the filamentary structure separately and is found to be marginally larger in the outer regions. the clusters are spatially offset from the filament by approximately The point spread function changes with position on the chip (errors surrounding the BCG 10 arcsec corresponding to a distance of 3.5 kpc at a redshift of z of a few per cent), due to both optical aberrations and geometric dis- 1 1 = 0.0176 (we adopt H 0 71 km s− Mpc− ). Clusters on the eastern tortions which may be responsible for the small differences detected arm coincide directly (in= the line of sight) with the Hα filament. in aperture correction across the field. We detected star clusters in the south-west region of NGC 1275 Our colour–magnitude results for the three regions are shown in using IRAF DAOFIND and the method of Carlson et al. (1998). We Figs 3 and 4. Fig. 5 shows a colour–magnitude plot of the outer find that the clusters in the Blue Loop and southern filament re- star formation regions accompanied by photometry of neighbour- at UQ Library on April 30, 2013 gion are more compact than the clusters in the centre. The point- ing control regions. We do not find any objects as blue as our blue spread-function fitting technique for crowded fields employed by population of clusters in any of these fields; however, the red pop- IRAF DAOFIND was not used as many of the clusters are partially re- ulation appears ubiquitous throughout the field. Contamination of solved, and we required that a single technique should be used for the young star cluster population from foreground and background the whole field. As in Carlson et al. (1998), we required the DAOFIND sources is therefore highly improbable due to both the lack of young parameters of roundness between 1and1andsharpnessbetween blue sources in other regions and the spatial distribution of the blue 0.2 and 1. In the central image, only− the south-west region was con- population in the Blue Loop and southern filament regions. This is sidered; it is both less obviously ‘dusty’ and allows a more direct discussed further in Section 3.1 and Fig. 7. comparison of our colour–magnitude results with previous results. The ACS filters differ significantly from other filter systems, and IRAF PHOT was employed to perform aperture photometry using this needs to be taken into account when determining reddening an aperture with a 2 pixel radius in all regions and all filters. The corrections from the galactic foreground. Due to the differences background in the Blue Loop and southern filament regions was in filter transmission curves between ground-based and the ACS estimated using the modal value in an annulus 10 pixels wide with filter systems, the extinction coefficients are calculated in the native an inner radius of 15 pixels. In all regions, growth curves of the photometric system and all corrections applied before converting to brightest, isolated star clusters in the field were constructed to test another system. the background estimation. The galaxy background near the outer For comparison with the stellar evolutionary synthesis models, regions of star formation, far from the nucleus, does not vary much we transform the ACS filter system to the Johnson–Cousins UBVRI with position; however, the field here is crowded. The large inner system using the method below, described in Holtzman et al. (1995) radius of the background annulus ensures we are sampling much and Sirianni et al. (2005), more background than surrounding stars. In the central region, an TMAG SMAG c0 c1 TCOL c2 TCOL2. (1) inner radius of 12 pixels and annulus 2 pixels wide, as in Carlson = + + × + × et al. (1998), was found sufficient to estimate the background. The The F435W filter can be transformed to a B-band filter and F625W small sky radius here is used to minimize the variation of the back- to a R-band filter; however, uncertainties of a few per cent are in- ground due to the galaxy. No subtraction of the smooth galaxy light troduced during this transformation. An additional systematic error in the inner region was done as the error introduced in the bright is introduced when using the synthetic transformation coefficients sources of interest is small. The worst case scenario of a non-linear for stars with B V<0.5 due to uncertainties in the shape of the background variation over the 28 pixels for the innermost and there- total response curve− in the F435W filter (see Sirianni et al. 2005, fore most steeply varying background introduces an error of less their fig. 21). The overall uncertainty in these transformations is of

C C # 2010 The Authors. Journal compilation # 2010 RAS, MNRAS 405, 115–128 Survey APOD 12/09.2011 • Perseus Cluster D = 70 Mpc • 5 fields V & I- 1 orbit per field per band (PI Conselice) • 2 fields in B & R of NGC 1275. 2 orbits per field per band Idenficaon of UCDs

• No selecon by luminosity or ellipcity • No selecon by colour – We do not restrict our sample to old, passive systems • Objects with effecve

radii 10 pc < re < 150 pc are classified as UCDs è We idenfy a sample of 84 UCD candidates with half-light radii 10 pc < re < 93 pc The massive star cluster populaon of NGC 1275

• We idenfy both normal globular clusters and massive young star clusters • “Normal” UCDs follow the blue and red globular cluster subpopulaons

But what about the very blue (B-R < 0.6) candidates? UCD formaon in NGC 1275 • Blue proto-UCDs: 34 sizes R = 10-30 pc e 33 29

31 N

E 2 kpc

è (B − R)0 ∼0.5 consistent with ages ~ 100 Myr Massive blue star clusters in NGC 1275

Normal UCD Very blue UCD • Filaments contain Star Cluster regions of acve star formaon • Young clusters with 7 masses 10 M¤ (Canning et al. 2010) • Remarkably round- no evidence for N dal elongaon

5 kpc E èThrough mergers, young, blue star clusters could reach sufficient mass such to survive the dal forces of NGC 1275 to appear as normal UCDs. The size-magnitude relaon

4.0

3.5 • Keck-DEIMOS

3.0 spectroscopy

2.5 confirms cluster membership for (pc)

) 2.0 e R

( 14 UCDs using

log 1.5 Giant Elliptical Hα and the CaT Dwarf Elliptical 1.0 Globular Cluster feautre UCD 0.5 Compact Elliptical dE Nuclei (Cotˆ e06)´ Perseus UCDs 0.0 UCD13 (This work)

6 8 10 12 14 16 18 20 22 − − − − − − − − − MB Penny et al, in prep è One UCD stands out as being parcularly extended UCD 13 • Most extended UCD in the Penny et al. (2012) sample, with Re = 93 pc from a Sérsic fit • Keck-DEIMOS spectroscopy 2.5’’ provides an internal velocity dispersion σ=35 ± 8 km s−1 èWhat is the origin of this object? A UCD or a Compact Ellipcal?

2.5 • Intermediate

2.0 in velocity )

1 dispersion − between a cE 1.5 )(kms

σ and a UCD

log( • Likely a 1.0 Giant Elliptical dally Globular Cluster Dwarf Elliptical stripped dE Compact Elliptical 0.5 UCD UCD 13 6 8 10 12 14 16 18 20 22 − − − − − − − − − MB Penny et al, in prep Is dal stripping likely?

v = 5292 ± 13.7 km s−1, vs v = 5276 km s−1 for NGC 1275 Angular separaon 1.5ʹ èminimum physical separaon of 30 kpc. The nuclei of stripped dwarf ellipcals? N • The colours of Perseus UCDs are E similar to those of dE nuclei, similar to the findings of Brodie et al. (2011).

• Perseus is an ideal environment to strip nucleated dEs (Penny et al. 2009) 4 kpc

èWe can not easily discriminate a dal origin from a massive star cluster one via colours alone Conclusions • Mulple formaon channels for UCDs exist in the core of Perseus: – Massive globular clusters – Stripped dEs – Formaon in gaseous filaments

• The complex filaments around NGC 1275 provide a possible formaon site for UCDs and globular clusters • However, the most massive Perseus UCD is likely a stripped galaxy • The UCD populaon in Perseus is likely a composite populaon of bright, extended globular clusters and stripped dEs