The SLUGGS survey00footnotemark: 0††thanks: http://sluggs.swin.edu.au/: Exploring the globular cluster systems of the Leo II group and their global relationships
We present an investigation of the globular cluster (GC) systems of NGC 3607 and NGC 3608 as part of the ongoing SLUGGS survey. We use wide-field imaging data from the Subaru telescope in the g, r, and i filters to analyse the radial density, colour and azimuthal distributions of both GC systems. With the complementary kinematic data obtained from the Keck II telescope, we measure the radial velocities of a total of 81 GCs.
Our results show that the GC systems of NGC 3607 and NGC 3608 have a detectable spatial extent of 15, and 13 galaxy effective radii, respectively. Both GC systems show a clear bimodal colour distribution. We detect a significant radial colour gradient for the GC subpopulations in both galaxies. NGC 3607 exhibits an overabundance of red GCs on the galaxy minor axis and NGC 3608 shows a misalignment in the GC subpopulation position angles with respect to the galaxy stellar component.
With the aid of literature data, we discuss several relationships between the properties of GC systems and their host galaxies. A one-to-one relation between the ellipticities of red GCs and the galaxy stellar light emphasises the evolutionary similarities between them. In our sample of four slowly rotating galaxies with kinematically decoupled cores, we observe a higher ellipticity for the blue GC subpopulation than their red counterparts. Also, we notice the flattening of negative colour gradients for the blue GC subpopulations with increasing galaxy stellar mass. Finally, we discuss the formation scenarios associated with the blue GC subpopulation.
keywords:galaxies: elliptical and lenticular, cD - galaxies: star clusters: individual - galaxies: individual: NGC 3605, NGC 3607, NGC 3608
To better understand the formation of early-type galaxies (ETGs), it is useful to study their oldest stellar components, such as globular clusters (GCs). GCs are mostly old (10 Gyr, Strader et al. 2005), luminous ( 10 L, Brodie et al. 2011) and compact (R 3pc) star clusters. Establishing connections between the properties of GC systems and their host galaxies can help in understanding the formation of GC systems and hence, their parent galaxies.
In almost all massive galaxies, GC systems are found to be bimodal in colour (Larsen et al., 2001; Peng et al., 2006; Kim et al., 2013; Hargis & Rhode, 2014). Transforming colours to metallicity connects this bimodality with two stages of GC formation. The colour/metallicity distribution peaks are represented by blue/metal-poor and red/metal-rich GC subpopulations (Brodie et al., 2012). In a recent work, Forbes et al. (2015) determined the mean ages of blue and red GC subpopulations as 12.2 12.8 and 11.5 Gyr respectively, suggesting that both subpopulations are very old. However, the two peaks differ in colour (e.g. 0.8 and 1.1 in (gi)) and metallicity (i.e. ([Z/H] peaks at 1.5 and 0.4 dex) values. Other properties of the two subpopulations differ such as azimuthal distribution, spatial distribution, radial colour distribution (Strader et al., 2011; Park & Lee, 2013) and also kinematics (Pota et al., 2013). Three scenarios have been suggested to explain the formation of these two distinct GC subpopulations: major-merger (Ashman & Zepf, 1992), multi-phase collapse (Forbes et al., 1997) and accretion (Côté et al., 1998, 2000). See Brodie & Strader (2006) and Harris (2010) for reviews. Many cosmological simulations of hierarchical galaxy formation have been used to investigate the characteristics of GC systems (physical, dynamical, chemical etc.) in ETGs (Beasley et al., 2002; Bekki et al., 2005, 2008; Muratov & Gnedin, 2010; Griffen et al., 2010; Tonini, 2013; Katz & Ricotti, 2013; Gnedin et al., 2014; Trenti et al., 2015).
In order to associate the formation of different GC subpopulations with galaxy formation events, we explore the radial and azimuthal distributions of both subpopulations and compare with galaxy stellar light. Generally, the radial distribution of blue GCs is found to be more extended than that of both the red GC subpopulation and the galaxy stellar light. The red GC subpopulation is similar in radial distribution to the galaxy light (Bassino et al., 2006b; Strader et al., 2011), whereas the profiles of the blue GC subpopulation show similarities with the X-ray surface brightness profiles (Forbes et al., 2012). The radially extended blue GC subpopulation residing in galaxy haloes suggests that they are very old stellar components formed early in an in-situ dissipative collapse galaxy formation scenario, or accreted later, into the galaxy outskirts, in galaxy build up by minor merger formation scenario. The similarity of red GCs with the galaxy stellar light supports coeval formation, but their origin (from the enriched gas of a parent galaxy or from accreted gas) is not clear.
Park & Lee (2013) studied the azimuthal distribution of GC systems in 23 ETGs using the data from the Advanced Camera for Surveys (ACS) mounted on the Hubble Space Telescope (HST). They found that the ellipticities of the red GC subpopulation match the galaxy stellar light ellipticities with a one-to-one correspondence. They also found that blue GC subpopulations show a similar but less tight relation. Wang et al. (2013) using the ETGs, from the ACS Virgo Cluster Survey (VCS), concluded that both red and blue GC subpopulations align in position angle with the galaxy stellar light, although in a weaker way for blue GC subpopulations. Several single galaxy studies concluded that the galaxy stellar light is mimicked by red GC subpopulation in position angle and ellipticity, but the blue GC subpopulation is differently distributed (e.g. NGC 720, NGC 1023: Kartha et al. 2014, NGC 5813: Hargis & Rhode 2014, NGC 4365: Blom et al. 2012).
The decreasing mean colour of GC subpopulations with increasing galactocentric radius is a key diagnostic observation (Bassino et al., 2006a; Harris, 2009; Arnold et al., 2011; Faifer et al., 2011; Forbes et al., 2011; Blom et al., 2012; Usher et al., 2013; Hargis & Rhode, 2014). More specifically, the degree of steepness allows us to distinguish between two different formation processes, dissipation and accretion/merger (Tortora et al., 2010). Recently, Forbes et al. (2011) studied the colour gradient for NGC 1407 GC subpopulations and found that both GC subpopulations have a steep negative gradient within 8.5 effective radii (R) and a constant colour to larger radii. They explained this colour trend as being compatible with two-phase galaxy formation (Oser et al., 2010). This implies that the inner GCs are formed during a dissipative collapse phase, whereas the outer GCs are acquired during late accretion/mergers. Thus, exploring the radial colour distribution can reveal clues about formation events that happened in the host galaxy’s history.
The layout of this paper is as follows. A brief literature review of the target galaxies is presented in Section 1.1. Section 2 describes the observations, data reduction techniques and initial analysis of imaging and spectroscopic data. Section 3 and 4 present the GC selection techniques and methods used to select the GC systems of individual galaxies. A detailed analysis of various GC system distributions (radial density, colour and azimuthal) for the selected GC systems is presented in Section 5. In Section 6, we discuss connections between the characteristics of galaxy stellar light and GC systems followed, in Section 7, by the conclusion.
1.1 NGC 3607 and NGC 3608 in the Leo II group
Here, we focus on the GC systems of the Leo II group. NGC 3607 and NGC 3608 are the brightest ETGs in the Leo II group. NGC 3607 is a near face-on lenticular galaxy while NGC 3608 is an E12 elliptical galaxy. In the same system there is a third galaxy, NGC 3605, which is a low mass galaxy of E45 morphology. Table 1 presents the main characteristics of the three galaxies with NGC 3607 as the central galaxy in the group. NGC 3608 and NGC 3605 are situated at a distance of 6 arcmin north-east and 2 arcmin south-west from NGC 3607. Kundu & Whitmore (2001a, b) investigated the GC systems of 57 ETGs including NGC 3607 and NGC 3608 using HST/Wide-Field Planetary Camera 2 (WFPC2) data in V and I filters. For these galaxies they detected 130 and 370 GCs, respectively, from single pointing imaging. They did not find a sign of a bimodal colour distribution in either galaxy.
With the same HST/WFPC2 data, Lauer et al. (2005) investigated the surface brightness profiles of NGC 3607 and NGC 3608. They mentioned that NGC 3607 contains a symmetric, old and tightly wrapped outer dusty disk to which a second disk is settling in a perpendicular direction. They explained this observation as an infall of gas directly to the centre of galaxy with no interaction with the outer disk. They also detected the remnants of a pre-existing dusty disk in NGC 3608. Terlevich & Forbes (2002) derived the ages of 150 galaxies using the spectral line indices and found 5.8, 3.6 and 10 Gyr ages for NGC 3605, NGC 3607 and NGC 3608 respectively. Later, Rickes et al. (2009) investigated the metallicity distribution, stellar population and ionised gas in NGC 3607 using long-slit spectroscopy. They found stellar components ranging in age from 1 to 13 Gyr between the centre and a 30 arcsec radius of the galaxy’s centre. As part of the ATLAS survey, McDermid et al. (2015) determined the mass-weighted ages for NGC 3605, NGC 3607 and NGC 3608 as 8.1 0.8, 13.5 0.7 and 13.0 0.7 Gyr respectively. Also, from the ATLAS survey Duc et al. (2015) studied these galaxies using the deep multi-band images from the Canada France Hawaii Telescope. They mentioned that NGC 3607 and NGC 3608 are interacting galaxies with the presence of weak dust lanes and ripples.
Based on the ROSAT data, two peaks were detected in diffuse hot X-rays on NGC 3607 and NGC 3608 (Mulchaey et al., 2003). They proposed that the two galaxies are undergoing a merger. Later, Forbes et al. (2006a) detected an extended diffuse X-ray emission around the Leo II group.
The GC systems of the ETGs at the centre of the Leo II group have not yet been studied using wide-field imaging data. As part of the SLUGGS (SAGES Legacy Unifying Globulars and GalaxieS) survey (Brodie et al., 2014), we obtained wide-field data in three optical filters covering the central region of the Leo II group using the Suprime-Cam instrument on the Subaru telescope. With the aid of the imaging and spectroscopic data, we aim to understand the properties of the GC systems associated with each galaxy.
2.1 Observations and reduction techniques
Photometric data for the Leo II group was obtained using the Subaru Prime Focus Camera (Suprime-Cam; Miyazaki et al. 2002) mounted on the 8-meter Subaru telescope. The Suprime-Cam imager consists of ten CCDs with individual sizes of 2048 x 4096 pixels and a pixel scale of 0.202 arcsec, covering a maximum sky area of 34 x 27 square arcmin. Wide-field images were obtained during 2011 January 3 and 4. The sky conditions were good with an average seeing of 0.81, 0.80 and 0.80 for the g, r, i filters, respectively. Multiple exposures in a dithered pattern were taken to fill the gaps between individual CCDs. In g, r and i filters the total exposure times were 3743, 1560 and 1200 sec, respectively.
The individual exposures were reduced and combined using the Suprime-Cam Deep Field Reduction package 2 (SDFRED2; Ouchi et al. 2004) for each of the three filters. The pre-processing of images included flat fielding, distortion corrections and atmospheric dispersion corrections. The pipeline also features custom-made modifications to improve the sky subtraction and alignment between multiple exposures. We employed SExtractor (Bertin & Arnouts, 1996) 111http://www.astromatic.net/software/ and Montage 222http://montage.ipac.caltech.edu/index.html for the alignment process. All point sources three sigma above the background level are identified on each CCD image using SExtractor. The positions of these point sources are matched with a reference catalogue (here we use the Sloan Digital Sky Survey) to create an astrometric solution. The astrometric solution is used by the Montage program to align and combine the individual images, generating mosaic images in the three filters. A combination of g, r and i filter mosaic images is shown in Figure 1.
We also obtained a single pointing covering the central region of NGC 3607 from the Hubble Legacy Archive (HLA). This was taken in the F814W (I) filter using the ACS instrument. The Wide-Field Channel on the ACS consists of two 2048 x 4096 CCDs with a 0.049 arcsec pixel scale, and 3.37 x 3.37 square arcmin field of view. A custom-made pipeline (for detailed explanation see Spitler et al. 2006) is employed to reduce the ACS data. The pipeline provides source positions and half light radii for all the detected sources, which are utilised for a preliminary selection of GCs in the Subaru/Suprime-Cam imaging (see Section 3).
Complementary spectroscopic data were obtained using the DEep Imaging Multi-Object Spectrograph (DEIMOS, Faber et al. 2003) on the Keck II telescope. The field of NGC 3607 was targeted on five nights during 2013 January 10 – 12 and 2014 January 26 and 27 as part of the SLUGGS survey. We used five slit-masks for good azimuthal coverage and the seeing per night was 0.87 FWHM 1.15 arcsec with a total exposure time of 10 hours. DEIMOS was used with 1200 l/mm grating centered on 7800 Å, with slit widths of 1 arcsec. In this way, we have a wavelength coverage from 6500 – 8700 Å and spectral resolution of Å. We reduced the raw spectra using the IDL SPEC2D reduction pipeline together with dome flats and arc lamp spectra. The pipeline produces sky-subtracted GC spectra that covers the CaT absorption lines in the near-infrared (8498, 8542, 8662 Å) and H line at 6563 Å(where possible).
We obtain the radial velocities from our science spectra using the FXCOR task in IRAF by measuring the doppler shift of the CaT lines, cross-correlating each Fourier transformed science spectrum with the Fourier transformed spectra of 13 Galactic template stars. In practice, we require that the strongest CaT lines (8542, 8662 Å) be present and where possible the H line as well. Where the lines are not properly defined, but the velocity is consistent with either galaxy, the GC is classified as marginal. Objects with velocities less than 350 km/s are classified as Galactic stars and those with velocities greater than 1800 km/s as background galaxies. Our final catalogue has 75 GCs and 7 ambiguous objects (see Table 8 in Appendix A). Here, ’ambiguous’ denotes that either the velocity or position has a mismatch with the target galaxies, but it has confirmed characteristics of a GC.
Before carrying out any photometric analysis, the galaxy light was subtracted in each of the three mosaic images. The two large galaxies are individually modelled using IRAF task ELLIPSE with the center of the galaxy, the major axis position angle (PA) and the ellipticity () as free fitting parameters. During the fitting process the bright stars were masked before modelling the galaxy light. The best fit galaxy model produces radial profiles of surface brightness, position angle and ellipticity measurements for both the galaxies. We made use of galaxy light subtracted images to improve the source detection in the central regions of target galaxies.
We utilised SExtractor for source identification and photometry. We instructed SExtractor to identify a probable source only if it has a minimum of 5 adjacent pixels with a flux higher than three sigma above the local background. SExtractor estimates the total instrumental magnitudes for the detected sources using Kron radii (Kron, 1980) in the automatic aperture magnitude mode. It provides an output list of point sources with position and magnitude. As standard stars were not observed for zeropoint calibration, we exploited the bright stars (i 22) present in the galaxy field. A match between these bright stars and the Sloan Digital Sky Survey catalogue (data release 7 version) was used for the flux calibration in all three mosaic images. Photometric zeropoint magnitudes in three filters are derived from the best-fit linear relationship between the instrumental magnitudes of bright stars and calibrated magnitudes from the SDSS catalogue. Estimated zeropoints in g, r, i bands are 28.68 0.08, 28.92 0.12, 28.78 0.15 magnitudes, respectively. All magnitudes have had the zeropoint correction applied. The galaxy photometry is corrected for Galactic extinction using the dust extinction maps from Schlegel et al. (1998). Hereafter, all the magnitudes and colours cited are extinction corrected.
3 Globular cluster selection
The large galaxies, NGC 3607 and NGC 3608, are at an assumed distance of 22.2 0.1 Mpc (Brodie et al., 2014) and NGC 3605 taken to be 20.1 Mpc (Cappellari et al., 2011). For GC identification, a match of object positions between the three bands is carried out at first, in order to eliminate the false detections. Afterwards a separation between extended objects (galaxies) and point source objects (both GCs and stars) is incorporated. This separation is based on the surplus light detected beyond the extraction aperture. Objects showing large difference between the extraction aperture and an outer aperture are considered as extended sources and are removed (see Kartha et al. 2014 for details).
We employ a colour-colour selection as the next step to identify the GC candidates. To aid this selection, we used the position and half light radius of the sources from the HST/ACS data. An upper limit of 9 pc at the distance of NGC 3607, for GC candidature is applied, and the selected objects are visually verified. A positional match between the Subaru objects and the GCs selected on the HST/ACS image is carried out and then the half light radius is attached to the Subaru list for the common objects. Hence we create a list of probable GCs with their positions, three magnitudes from the Subaru/Suprime-Cam data, and half light radii from the HST/ACS data. From earlier studies, e.g., figure 6 in Blom et al. (2012) and figure 3 in Pota et al. (2013), it is evident that the GCs populate a particular region in the colour-colour diagram. With the above list we identify the locus of GCs in (ri) verses (gi) colour space, implementing similar procedures as explained in Spitler et al. (2008) and Blom et al. (2012). The GC candidates, along with neighbouring objects showing a 2 deviation from the selected region, are chosen as final GC candidates. The selected GCs range over 0.6 (gi) 1.4, which corresponds to a metallicity range of 1.94 [Z/H] 0.86 using the empirical relation given in Usher et al. (2012). The upper and lower cut off in i band magnitude are 20.4 and 24.4 magnitudes, respectively. At the distance of NGC 3607 objects brighter than 20.4 magnitude include ultra compact dwarfs (Brodie et al., 2011) while the lower limit is one magnitude fainter than the turnover magnitude for the GC system. This final list of GC candidates include 1000 objects from NGC 3605, NGC 3607 and NGC 3608.
4 Defining the GC systems of each galaxy
We derive the stellar mass of NGC 3605 as log(M) = 10.76 M from the galaxy V-band magnitude (see Table 1) and the mass to light ratio from Zepf & Ashman (1993). The extent of the GC system of NGC 3605 is calculated from the stellar mass in the empirical relation between GC system extent and the galaxy stellar mass (equation 7 in Kartha et al. 2014). A GC system extent of 40 arcsec is derived from the calculation and we assume a maximum of 1 arcmin extent for NGC 3605. We detect 10 objects in the 1 arcmin region around NGC 3605 and eliminate them from the following calculations. The surface density distribution of GCs around NGC 3605 has been investigated and we find a constant GC density, implying no contamination from NGC 3605 to the NGC 3607 or NGC 3608 GC systems.
The remaining GC candidates are a combination of objects from NGC 3607 and NGC 3608. In order to classify their individual GC systems, we invoke two methods, based on surface brightness and position angle of the host galaxies.
4.1 Surface brightness method
The galaxy light for both galaxies is modelled and extracted using the IRAF task ELLIPSE. The individual surface brightness profiles are fit with Sérsic profiles (Graham & Driver, 2005). We extrapolate these profiles to larger galactocentric radius ( 15 arcmin) and use these extrapolated profiles to represent the stellar light profiles of individual galaxies to large radius. Figure 2 shows the surface brightness profiles of NGC 3607 and NGC 3608. Based on the position of each GC, its membership probability is computed from the ratio of surface brightness of NGC 3607 to NGC 3608. Hereafter we refer to this as the surface brightness (SB) method. GCs with a probability greater than 55 percent are counted as members of NGC 3607, while less than 50 percent are classified as members of NGC 3608. The 6 R ellipses overlap around 55 percent SB probability (see Figure 3). We classify the GCs with probability between 55 and 50 percent as ambiguous objects.
4.2 Major axis method
We employed a second method called the major axis (hereafter MA) method, to separate the GC systems of the two galaxies. In this method, we divided the GCs along the photometric major axis (125 and 82 degrees for NGC 3607 and NGC 3608, respectively) and selected the hemisphere pointing away from the other galaxy. Thus, the selection of GCs for NGC 3607 includes GCs in the position angles 125 to 305 degrees and for NGC 3608 GCs from 0 to 82 and 262 to 360 degrees. This method excludes the region of maximum tidal interaction between the two galaxies. Coccato et al. (2009) adopted a similar method for disentangling the planetary nebulae (PNe) of NGC 3608. To eliminate the contaminants from NGC 3607, they excluded the PNe on the southern side of NGC 3608, which is equivalent to the major axis method used here.
4.3 Analysis of kinematic data
We obtained the radial velocity measurements for 82 (confirmed plus marginal) GCs in the field of the Leo II group. The galaxy systemic velocities for NGC 3607 and NGC 3608 are 942 and 1226 km/s (Brodie et al., 2014), respectively. To assign the membership of GCs to individual galaxies, we performed a biweight estimator distribution (following Walker et al. 2006) based on the right ascension, declination and line of sight velocity of each GC. The GCs within 2 ( is the standard deviation calculated from the velocity distribution) from the central galaxy velocity are assigned membership to the corresponding galaxy, while keeping marginal members as velocities between 2 to 3. Figure 3 displays positions of spectroscopically confirmed GCs on a SB probability map. The background map shows the SB probability used in the separation of GCs (see Section 2). The positions of individual galaxy GCs (as determined using velocities) fall on the same region derived from the SB method, confirming the robustness of the SB probability method for classifying the GCs. The distribution gives 43 and 32 GCs, respectively, as NGC 3607 and NGC 3608 members.
In addition, we classified the 7 ambiguous objects as 6 GCs and one extreme member. The extreme member S41879 has a velocity of 1822 22 km/s, but positionally it is projected near the centre of NGC 3607 (see Figure 3) in the 2D map. Assuming it lies at the distance of NGC 3607 (D = 22.2 Mpc), then it has M = 9.97 mag. From the line of sight velocity and H = 70 (km/s)/Mpc, we calculate the distance as 26 Mpc and hence the magnitude M = 10.31 mag. This suggests that it is a possible UCD (see Brodie et al. 2011). To confirm this, we checked the HST image for an estimation of its size. Unfortunately, this object is placed in the central gap region of the HST pointing. We examined the Subaru image and found that the object is very circular in shape. Another possibility is an intra-group GC, as it is blue (gi) = 0.623, circular in shape and lies in the projected region between NGC 3607 and NGC 3605. With the above information, we suggest that this extreme object might be a background UCD or an intra-group GC. Eliminating this extreme object, we have 81 spectroscopically confirmed GCs for NGC 3607 and NGC 3608.
Figure 4 shows the velocity distribution of GCs with galactocentric radius measured from the centre of NGC 3608. The six marginal GCs are labelled in Figures 3 and 4. Based on both these figures, we assign a membership for the marginal GCs. Note here that this manual membership assignment is unimportant for any broad conclusions of this study. S51178 is positionally close towards NGC 3607 with velocity > 1300 km/s. But according to the SB probability, this GC has > 80 percent probability to be associated with NGC 3607. Hence, considering these facts we assign it to NGC 3607 as GC44 (name given in Table 8). Based on the SB probability and velocity measurement, S53407 is assigned to NGC 3607 (GC45). The position of S64467 is close to NGC 3608 with 50 percent probability, but having a velocity of 807 km/s supports a membership with NGC 3607 (GC46). S60023 has a 70 percent probability with NGC 3608 and with a velocity of 1160 km/s. Hence, S60023 is a probable member of NGC 3608 (GC33). S55434 (GC34) and S57144 (GC35) are GCs with velocities 1281 and 1229 km/s, respectively. Both fall on the probability region of 60 percent for NGC 3607. However, a membership to NGC 3608 is allocated for these GCs based on the positional closeness and velocities. Hence, S60023, S55434, S57144 are NGC 3608 members and S51178, S64467, S53407 are NGC 3607 members. Finally, NGC 3608 and NGC 3607 have 35 and 46 spectroscopically confirmed GCs, respectively.
The mean velocities estimated from the GC systems of NGC 3607 and NGC 3608 are 963 and 1220 km/s, respectively, in good agreement with galaxy central velocities. Estimates of the GC system velocity dispersions for NGC 3607 and NGC 3608 are 167 and 147 km/s, respectively. Cappellari et al. (2013) found central velocity dispersions of 206.5 10 and 169.0 9 km/s from the galaxy stars, respectively, for NGC 3607 and NGC 3608.
4.3.1 GC subpopulations
Currently, we have 46 and 35 spectroscopically confirmed GCs, respectively, for NGC 3607 and NGC 3608. We have classified the GCs into blue and red subpopulations based on a constant colour division with galactocentric radius due to small number statistics. The GMM algorithm (explained in Section 5.1.2) gives a (gi) dividing colour of 0.87 for NGC 3607 and 0.93 mag for NGC 3608 (from photometric measurements). We used these colours to separate the blue and the red subpopulations of the two galaxies as shown in Figure 5. From the photometric analysis of the GC subpopulations, we obtained 62 and 38 percent blue and red subpopulations (see Section 5.2.2), respectively.
5 Analysis of photometric data
Below we describe the radial density, colour and azimuthal distributions of the NGC 3607 and NCG 3608 GC systems. Note here that the GC systems are selected from the colour-colour space discussed in Section 3.
5.1 GC system of NGC 3607
5.1.1 Radial density distribution
To derive the radial distribution of the GC system, we define radial bins up to a galactocentric radius of 16.9 arcmin. Then the effective area coverage is obtained for each radial annulus. The area is corrected for the presence of saturated stars and for any regions outside the detection area. The GC number in each annulus is then divided by the effective spatial area to determine the spatial density in that particular annulus. The errors are calculated using Poisson statistics.
We obtained the GC system surface density using two methods. In the SB method (refer Section 2), a correction is applied for the missing area due to NGC 3608 and NGC 3605. In the MA method (refer Section 4.2), the number density is doubled in each radial bin. The radial density distribution is fitted with a combination of Sérsic profile plus a background parameter to estimate the effective radius and the background value for the GC system. The fitted surface density profile is:
where N is the surface density of the GCs at the effective radius R, n is Sérsic index or the shape parameter for the profile, b is given by the term 1.9992n 0.3271 and bg represents the background parameter. Note that the radius R is the centre of each radial bin.
Figure 6 shows the density profile of the GC system for NGC 3607 only. The GCs brighter than the turnover magnitude, i = 23.5, only are considered. The plot displays the density values derived from the two different methods, i.e. SB and MA methods. Both are fitted with the profile given in Equation 1. In the density distribution plot for NGC 3607, the SB and MA methods used 1170 and 907 objects, respectively. It is evident from the figure that both methods yield consistent results and the profile reaches the background at a galactocentric radius of 9.5 0.6 arcmin (61 3 kpc).
Kartha et al. (2014) found an empirical relation between the galaxy stellar mass and the extent of its GC system. The relation is as follows:
|SB||2.45 0.54||2.74 1.76||1.70 0.15||9.4 0.6|
|MA||1.99 0.29||1.97 1.19||1.68 0.08||9.6 0.5|
NGC 3607, an S0 galaxy, with absolute V-band magnitude M = 21.87 and assumed mass to light ratio of 7.6 (given in Zepf & Ashman 1993) has a host galaxy mass, log(M/M) = 11.56. The GC system extent for NGC 3607 determined using the above equation is 57 3 kpc, in good agreement with the direct estimation using the wide-field Subaru/Suprime-Cam image (61 3 kpc).
5.1.2 GC bimodality
Figure 7 shows the colour magnitude diagram of NGC 3607 GCs. The GCs, brighter than M = 7.75 mag, within the GC system extent of NGC 3607 are shown in the diagram. The bottom panel contains the histogram of GCs which are brighter than the turnover magnitude (M = 8.23 mag) along with the background contamination. To estimate the background contamination, we made use of the detected objects beyond the GC system extent of the galaxy. For NGC 3607, the objects beyond 11 arcmin (as GC system extent is 9.5 0.6 arcmin) are considered as background contamination. We applied an areal correction, if needed. The background corrected colour histogram is also shown in Figure 7 and it includes 611 GCs.
To quantify the colour distribution of the GC system, we used the gaussian mixture modeling (GMM, Li & Gnedin 2014; Muratov & Gnedin 2010) algorithm on the GC system (gi) colour, after background correction. The algorithm tests for a multimodal colour distribution over unimodal. To be a significant multimodal GC system distribution, the following three statistics should be, 1. low values for the confidence level from the parametric bootstrap method, 2. the separation (D) between the means and the respective widths greater than 2, and 3. negative kurtosis for the input distribution.
For NGC 3607, the GMM algorithm confirmed a bimodal colour distribution from the SB and MA method selected GCs, based on the following statistics : with less than 0.001 percent confidence level, D > 2.6 0.3 and negative kurtosis. The blue and red GC subpopulations peak in (gi) colour at 0.74 0.04 and 1.03 0.03, respectively. The (gi) colour of separation between the blue and the red subpopulations is at 0.87 0.02. The total GC system is classified into 45 9 and 55 8 percent, respectively, blue and red subpopulations.
|3607||Blue||1.59 0.94||4.14 2.32||0.36 0.12|
|Red||0.67 0.52||3.38 1.35||0.48 0.05|
|3608||Blue||1.42 0.31||1.03 0.89||0.50 0.05|
|Red||0.91 0.72||1.98 0.82||0.35 0.05|
The radial density distribution for both GC subpopulations (from the MA method) are estimated and plotted in Figure 8. Both subpopulation distributions are fitted with Sérsic profile given in Equation 1. The parameters derived from the Sérsic fit are tabulated in Table 3. The red subpopulation is centrally concentrated while the blue subpopulation is more extended. The red subpopulation appears to have higher number density for most galactocentric radii. The galaxy stellar light profile is in better agreement with the density distribution of red subpopulation than blue subpopulation. Also the effective radius of the galaxy stellar light (39 arcsec) matches more with the red subpopulation (40 29 arcsec) than the blue ones (95 50 arcsec).
5.1.3 Azimuthal distribution
To quantify the azimuthal distribution of GCs, they are initially folded along the North to South direction, then binned in equal angular intervals. The azimuthal distribution, , is then fitted with a profile (McLaughlin et al., 1994) of the form:
where is the power law index fitted to the surface density of GCs, bg is the background estimated from the Sérsic fits (see Section 5.1.1) and k is the normalization constant. The profile is iterated with the position angle of the GC system (PA) and the ellipticity () as free parameters. For the analysis, only the GCs within the extent of GC system (i.e., 9.5 arcmin) are included. The number of GCs in each angular bin is corrected for the missing area due to NGC 3608 in the SB method, and is doubled in the MA method. Here we used 980 and 564 GCs, respectively, in the SB and MA methods.
Figure (a)a shows the azimuthal distribution of GCs selected based on the SB method. The GCs are aligned to a position angle of 110 7 degrees, which is in reasonable agreement with the stellar light (125 degrees) of the galaxy. The alignment of GC system is more elliptical (0.39 0.08) than the stars (0.13). The GCs also show an enhancement along the minor axis (35 degrees), which is either a genuine feature or possibly a contamination from the GCs of NGC 3608 and NGC 3605 (both positioned around the minor axis of NGC 3607). We already found a constant surface density around NGC 3605 and hence, we assume that NGC 3605 is not contributing to the overabundance.
The only other possible contributor for this minor axis overabundance is NGC 3608, situated in the NE direction. We have eliminated the maximum contamination from NGC 3608 in the MA method, as it counts only the hemisphere away from the other galaxy. Hence, if the enhancement of GCs is not genuine, then we should not observe the same in the MA method. Figure (b)b displays the azimuthal distribution of GCs selected in the MA method, including only the GCs from 125 to 305 degrees counted from North in counter-clockwise direction. It is evident from this plot that the enhancement along the minor axis is a genuine feature, with decreased strength which is consistent within error bars. The position angle of GCs from the MA method also aligns with the galaxy stellar light. Similarly, from the SB method, the GCs are found to be more elongated than the arrangement of stellar light. Table 4 summarises the best fit sinusoidal profile parameters.
|Method||NGC 3607||NGC 3608|
|SB||Total||110 7||0.39 0.09||104 15||0.20 0.09|
|Blue||112 14||0.37 0.11||106 11||0.31 0.10|
|Red||108 11||0.47 0.09||97 18||0.14 0.16|
|MA||Total||109 8||0.42 0.07||66 7||0.39 0.10|
|Blue||108 10||0.45 0.11||67 8||0.45 0.09|
|Red||109 8||0.48 0.11||64 10||0.44 0.13|
Figure 9 also shows the azimuthal distribution of blue and red GC subpopulations from the two methods. The subpopulations are separated at a (gi) colour of 0.87, obtained from the GMM algorithm. Both subpopulations have similar position angles for the total GC system and are more elliptical than the galaxy stars.
Summarising, the total GC system and both subpopulations follow the galaxy stellar light in position angle. But the distribution of GCs is not as circular as the galaxy stellar component. The red GC subpopulation shows a more flattened distribution than the blue subpopulation for NGC 3607.
5.1.4 Radial colour distribution
Figure 10 shows the radial distribution of GC colours from the centre of NGC 3607. The GCs brighter than the turnover magnitude in the MA method only are included. The GC subpopulations are divided with a moving colour with radius technique. In each radial bin, the average colour for both subpopulations are determined (keeping a constant number of GCs per radial bin). For NGC 3607, we used 350 GCs to plot the colour distribution with 45 GCs in each bin.
As seen from the plot, for the total extent of the GC system, the average colour for the blue subpopulation decreases with radius from the centre, while a flat colour gradient is seen for the red subpopulation. The colour distribution for the blue subpopulation is fitted with a logarithmic relation (following Forbes et al. 2011) as:
where R is the effective radius for NGC 3607 equal to 39 arcsec (Brodie et al., 2014), a and b are, respectively, intercept and slope of the fit. We obtained a best fit line using the bootstrap technique and derived the parameters for the blue subpopulation as a = 0.82 0.018 and b = 0.036 0.009 mag per dex. Maraston (2005) derived a relation between (gi) and [Z/H] over the metallicity range [Z/H] 0.2, using single stellar population models, of (gi)/[Z/H] = 0.21 0.05 mag per dex. Using this we obtained for the blue subpopulation a metallicity gradient of 0.17 0.04 dex per dex to the total extent of the GC system. But, we did not detect a significant colour gradient for the red subpopulation and the total population in the total extent of GC system (0.01 0.01 and 0.013 0.011 mag per dex for red and total GCs).
We also quantified the colour/metallicity gradient in the central ( 6.5 arcmin) region, only including the common galactocentric radii between the two galaxies. The colour gradient for the blue, red and the total population are 0.070 0.013, 0.033 0.015 and 0.039 0.018 mag per dex. In the inner 6.5 arcmin region, the blue subpopulation has a higher metallicity gradient (0.33 0.06 dex per dex) compared to the red subpopulation (0.16 0.07 dex per dex). Hence, we conclude that a significant colour/metallicity gradient is obtained for the blue and the red subpopulations of NGC 3607.
5.2 GC system of NGC 3608
5.2.1 Radial density distribution
Figure 11 displays the radial density of GCs selected with the SB and the MA methods for NGC 3608 fitted with the profile given in Equation 1 (fitted parameters are given in Table 5). In the SB method, the selection of GCs for NGC 3608 gives a maximum galactocentric radius of 5.5 arcmin (as seen from Section 2). But the MA method identifies objects to a distance of 12.8 arcmin from the galaxy centre (thus extends up to the edge of the detection area). In both methods, the GCs with i 23.5 mag (turnover magnitude) are counted for studying this distribution. In the density distribution plot, the SB and MA methods used 304 and 402 objects, respectively. The density distribution of GCs in radial annuli, after applying respective corrections for both methods, are shown in Figure 11. The GC system reaches a background level of 1.65 0.1 GCs per square arcmin to a galactocentric radius of 6.6 0.8 arcmin (43 5 kpc), from the MA method. But the density value for the final data point from the SB method is 1.82 0.36 GCs per square arcmin implying that the distribution has not reached the background level. The elimination of marginal GCs (SB probability between 50 and 55 percent) in the SB method might be the reason for this discrepancy in the extent of GC system. Another point from the figure is that the surface density values estimated from both methods are consistent within error bars, up to 5.5 arcmin.
|SB||1.29 0.15||0.66 0.36||1.82 0.36||-|
|MA||1.50 0.15||0.93 0.56||1.65 0.10||6.6 0.8|
5.2.2 GC bimodality
The colour magnitude diagram for the selected GCs of NGC 3608, within the GC system extent (43 kpc) and brighter than M = 7.75 mag, is shown in Figure 12. The figure displays 250 GCs. The background contamination in the GC system selection is quantified, as explained in Section 5.1.2, and are corrected for this contamination. The bottom right panel displays the colour histograms of GCs which are brighter than the turnover magnitude with and without background correction. The estimated background correction is also illustrated in the same figure.
The GMM algorithm fit to NGC 3608 GCs selected from the MA method gives a bimodal colour distribution with peaks at (gi) = 0.80 0.02 and 1.12 0.04. The total GC system contains 65 6 and 35 6 percent, respectively, blue and red subpopulations. The blue and red subpopulations are divided at (gi) = 0.93.
The radial surface densities (GCs from the MA method) are fitted with Sérsic profiles and are displayed in Figure 13. The parameters estimated from the Sérsic fit are tabulated in Table 3. For NGC 3608, the blue subpopulation shows a higher density than the red subpopulation throughout the extent of the GC system. The red subpopulation is found to be more centrally concentrated, and their density profile is in good agreement with the galaxy stellar light.
5.2.3 Azimuthal distribution
The range of galactocentric radii for the selected GCs in the SB method is from 0.5 to 5.5 arcmin. The selection of GCs in all position angles is complete up to 2.2 arcmin and hence, an areal correction is applied for the missing area outside that radius. Here we used 378 and 275 GCs, respectively, in the SB and MA methods. Figure (a)a shows the azimuthal density distribution of GCs from the SB method. The histograms are fitted with the sinusoidal profile given in Equation 3. Table 4 gives the position angles and ellipticities obtained from the sinusoidal fit. The galaxy stellar light has a major axis of 82 degrees and ellipticity of 0.20. As seen from Table 4, the total GC system and both subpopulations are arranged along a different position angle of 100 degrees for the SB method. When the distribution is examined over 0 to 360 degrees rather than 0 to 180 degrees (i.e., without folding along the North to South direction), an overabundance is evident in the position angles between 90 and 230 degrees. This is in the direction towards NGC 3607 and also the direction in which the area correction is largest. Hence, this overabundance is either due to contamination from NGC 3607 (or due to overestimation of missing area). Also a scarcity of GCs is observed in both major axis position angles (82 and 262 degrees). The ellipticity for the total GC system is 0.20 0.09, matching with the galaxy stellar light.
Figure (b)b shows the azimuthal density distribution of GCs selected in the MA method, for which GCs in the position angles 80 to 260 degrees are under abundant. The GCs within the extent of GC system (6.6 arcmin) are included in the azimuthal distribution. As seen in Table 4, the best fit sinusoidal profile gives a position angle of 66 7 degrees for the total GC system and an ellipticity of 0.39 0.10. The GCs selected in the MA method includes GCs of NGC 3608 placed at a position angle pointing away from NGC 3607, implying minimum contamination. The arrangement of GCs in the MA method is along the position angle matching the galaxy stars, but the distribution is more elliptical. Since we observed an overabundance in GCs for both galaxies, in the region towards each other, an interaction may be occurring between the two.
The total GC system is separated into subpopulations at (gi) = 0.93 (obtained from the GMM algorithm). Regarding the azimuthal distribution of GC subpopulations, both subpopulations are aligned along the position angle of the total GC system in the two methods. Also the ellipticity of both subpopulations matches with the total GC system. In the MA method, the total and both subpopulations are more elliptically aligned than the galaxy stellar light.
5.2.4 Radial colour distribution
Figure 15 shows the radial (gi) colour distribution of GC system of NGC 3608 selected on the MA method. To study this distribution, GCs (from the MA method) brighter than the turnover magnitude are selected and is 215 GCs. The total GC system and the red subpopulation show a null gradient, while the blue subpopulation shows a strong gradient along the total radial extent of the GC system. The colour distribution of the blue subpopulation is fitted with the logarithmic relation given in Equation 4, where R = 30 arcsec (Brodie et al., 2014). The parameters, a = 0.823 0.019 mag and b = 0.052 0.011 mag per dex, are derived from the best fit profile using the bootstrap technique (shown in Figure 15). The colour gradient, when converted to a metallicity gradient, gives [Z/H] = 0.25 0.05 dex per dex.
6.1 GC system distribution and galaxy effective radius
In this study of two group galaxies (NGC 3607 and NGC 3608 situated within a projected distance of 39 kpc), we introduce two methods, the Surface Brightness and the Major Axis methods, to separate the individual GC systems. For NGC 3607, the radial GC system extent determined from both methods are consistent with each other and in good agreement with the empirical relation for GC system extent (Equation 2), initially presented in Kartha et al. (2014). From the radial surface density distribution, the red subpopulation is more centrally concentrated than the blue subpopulation. The galaxy surface brightness distribution is in agreement with the density distribution profile of the red subpopulation than the blue subpopulation (Figure 8). Also, the effective radius of the galaxy stars (39 arcsec) is consistent with that of the red GC subpopulation (40 29 arcsec), while for the blue GC subpopulation it is 95 50 arcsec. Both the spatial distribution and the effective radius measurements support the idea that the red GC subpopulation has evolutionary similarities with the galaxy stellar component (Forbes & Forte, 2001; Larsen et al., 2001; Brodie & Strader, 2006; Spitler, 2010; Forbes et al., 2012).
For NGC 3608, the blue GC subpopulation is more extended than the red GC subpopulation. It is evident from Figure 13 that the density distribution of the red GC subpopulation follows the galaxy stellar light distribution. However, the effective radius of galaxy light (30 arcsec) is half of the red subpopulation (59 40 arcsec) and one third of the blue subpopulation (85 18 arcsec). The effective radius of the red GC subpopulation is therefore not consistent with the stellar light component. Even so the resemblance of the density distribution profile with the galaxy stellar light might imply a significant association.
|NGC||GC system (kpc)||Stellar light (kpc)|
Figure 16 displays the total GC system effective radius versus the galaxy effective radius and is an updated version (with the addition of six galaxies) of figure 20 in Kartha et al. (2014). In this figure, the GC system effective radii are determined from Sérsic profile fits to the density distribution, which is currently available for twelve galaxies. The positions of the newly added galaxies, tabulated in Table 6, are compatible with the existing linear relation (R = [(5.2 3.7) R] (8.5 6.5)). The updated relation for the twelve galaxies is as follows:
where both Rs are measured in kpc. When compared with the relation in Kartha et al. (2014), the above relation has a similar slope within error bars. The effective radii for both GC subpopulations are determined only for six galaxies. With the available data, we could not detect any significant relation between the effective radius of GC subpopulations and the host galaxy stellar light.
From Equation 5, we can infer that the GC system effective radius is 6 times the galaxy effective radius, which confirms that the GC system of a galaxy extends further out than the bulk of its stellar component (Harris et al., 2000; Forbes et al., 2006b; Brodie & Strader, 2006; Alamo-Martínez et al., 2012; Cantiello et al., 2015). A byproduct from the above relation is that we can estimate the GC system effective radius by knowing the galaxy effective radius.
6.2 GC system ellipticity and galaxy ellipticity
To further address the association of galaxy stellar light with GC subpopulations, we need to study the two dimensional spatial distribution of these systems. Different studies of two dimensional distributions (position angle and ellipticity) have confirmed an association of both subpopulations with galaxy stellar light (e.g. NGC 2768 by Kartha et al. 2014, NGC 4636 by Dirsch et al. 2005). Park & Lee (2013) analysed the two dimensional shape parameters of 23 early-type galaxies using the HST/ACSVCS. They found that the arrangement of both subpopulations is aligned with the photometric major axis of galaxies. Also, the red GC subpopulations show a tight relation in ellipticity with galaxy stellar light, while the blue GC subpopulations show a less tight relation. Concurrently, Kartha et al. (2014) obtained a similar relation for the red subpopulations from a sample of six galaxies using wide-field imaging.
Figure 17 demonstrates the relation between GC subpopulation ellipticity and galaxy stellar light ellipticity for ten early-type galaxies. The plot is an updated version of figure 22 in Kartha et al. (2014), with the addition of four galaxies (NGC 3607, NGC 3608, NGC 4406 and NGC 5813). Data for NGC 4406 and NGC 5813 are taken from Hargis & Rhode (2014). For NGC 4406, the stellar, blue and red subpopulation ellipticities are 0.4 0.03, 0.39 0.06 and 0.36 0.07, respectively. Similarly, the ellipticities for stellar, blue and red GC subpopulations of NGC 5813 are 0.3 0.03, 0.52 0.15 and 0.36 0.11, respectively. With the addition of four galaxies, we observe a tight one-to-one relation between red GC ellipticity and galaxy stellar light. The relation is
The intrinsic scatter in the above relation is estimated as 0.10. The one-to-one relation signifies that the red subpopulations are affiliated with the stellar light of the parent galaxies (Park & Lee, 2013). In other words, both the red subpopulation and galaxy stellar light might have a similar origin. In contrast to Park & Lee (2013), we notice a poor association of blue GC subpopulation ellipticity with galaxy stellar light ellipticity. We explain this as a consequence of our wide-field imaging, as the ACSVCS data used by Park & Lee (2013) does not reach far out enough to detect the whole blue GC subpopulation for the most extended galaxies (Peng et al., 2006).
In this small sample of ten galaxies, NGC 3607 shows the lowest ellipticity (nearly circular at = 0.13) for the galaxy stellar component. For NGC 3607, the ellipticities of both GC subpopulations show a deviation from the galaxy stellar light, although both are arranged along the photometric major axis of the galaxy. So, NGC 3607 supports the idea that galaxies with low ellipticities might have randomly arranged GC subpopulations (Wang et al., 2013). The difference in spatial distribution of GC subpopulations from the galaxy stellar component suggests that a major fraction of both GC subpopulations might have formed separately from the galactic stars and later settled in the host galaxies. In the case of NGC 3608, both GC subpopulations show deviations from the galaxy stellar light in position angle. In addition, the blue GC subpopulation shows a more elongated distribution than the red GC subpopulation.
In addition to NGC 3608, three other galaxies - NGC 4365, NGC 4406 and NGC 5813 - also have blue GC subpopulations more elongated in shape than the red GC subpopulations. The elongated shape of the blue GC subpopulation suggests that it shows spatial distribution similarities with the red subpopulation that mostly follows the distribution of galaxy stellar component. If the distribution of blue GCs is not spherical, Wang et al. (2013) suggest that it may not have been built from accretions that were equally distributed in all directions. Instead they might formed through local filamentary structures in particular directions. This points out that directional dependent accretion or minor mergers might have occurred in these galaxies, altering the shape of blue GC subpopulations.
In addition, these four elliptical galaxies are all slow rotators with kinematically distinct cores, KDCs, (Emsellem et al., 2011; Krajnović et al., 2011). Naab et al. (2014) carried out hydrodynamical simulations to kinematically study the centres of early-type galaxies. They suggested that KDCs were generally formed in slow rotators that had experienced multiple gas-poor minor mergers. They proposed that their recent mass assembly histories are devoid of any major mergers and are expected to have older stellar populations. Few, if any GCs, are expected to have formed from such mergers. It is unclear whether blue GCs from the accreted galaxies would form a more elongated distribution than the host galaxy starlight as we observe.
In summary, the ellipticities of red GC subpopulations have a one-to-one relation with the galaxy stellar light ellipticities, whereas only a weak relation is seen for the blue GC subpopulation. Additionally, slowly rotating galaxies with a KDC have larger values for blue GC subpopulation ellipticities than their red GC counterparts. The elongated shape of the blue GC subpopulations may be due to recent minor mergers that were asymmetric in direction (Tempel et al., 2015).
6.3 GC metallicity gradients and galaxy stellar mass
Colour gradients are important observational features for exploring the formation history of GC subpopulations and are clues to galaxy mass assembly. A negative colour gradient (GCs are redder at the centre of the galaxy than the outskirts) represents either the presence of younger (or more metal-rich) GCs at the galaxy centre or older (more metal-poor) GCs at the outskirts. As GCs are observed to be mostly old ( 10 Gyr, Strader et al. 2005; Dotter et al. 2010; Forbes et al. 2015), the colour gradients are basically caused by metallicity gradients rather than age gradients.
The observed gradients in GC subpopulations help discriminate between the different galaxy formation processes e.g., a negative gradient is predicted when the GCs are formed from a dissipative collapse (Pipino et al., 2010), while a gas-poor major merger will wash away any existing gradient (Di Matteo et al., 2009), a gas-rich major merger may remake a new gradient different from the original one (Hopkins et al., 2009), etc. Also, minor mergers (accretions) can deposit GCs in the outskirts of galaxies (Hirschmann et al., 2015; Pastorello et al., 2015) which will alter the existing gradient, perhaps resulting in an inner negative gradient and a flat outer gradient (Oser et al., 2010; Forbes et al., 2011).
The first detection of a radial colour gradient in a GC system was by Geisler et al. (1996) in NGC 4472. With ground based data, GC colour gradients have been detected in other massive galaxies (NGC 4486: Harris 2009, NGC 1407: Forbes et al. 2011, NGC 4365: Blom et al. 2012), while only seen in a handful of intermediate mass galaxies (NGC 3115: Arnold et al. 2011, NGC 4564: Hargis & Rhode 2014) to date.
In NGC 3607, another intermediate mass galaxy, the mean colours of both the blue and the red GC subpopulations reveal a significant colour gradient in the inner 6.5 arcmin (10 R). The colour gradient for the blue subpopulation is steeper than the red subpopulation. Within the total extent of the GC system (beyond 10 R), only the blue subpopulation has a significant colour gradient. We detect a significant colour gradient only for the blue GC subpopulation of NGC 3608.
|NGC||Blue GCs||Red GCs|
|(M)||(dex dex)||(dex dex)|
References: 1 - Liu et al. (2011); 2 - Bassino et al. (2006a); 3 - Forbes et al. (2011); 4 - Arnold et al. (2011); 5 - Faifer et al. (2011); 6 - This paper; 7 - Usher et al. (2013); 8 - Blom et al. (2012); 9 - Geisler et al. (1996); 10 - Harris (2009); 11 - Forte et al. (2012); 12 - Hargis & Rhode (2014); 13 - Strader et al. (2012); 14 - Usher et al. (2013)
Table 7 compiles the list of galaxies in which colour gradients are detected for the GC subpopulations. The colour gradients are detected in different colour filters. For a uniform comparison, the different colour gradients are converted to metallicity gradients ([Z/H]) using colour-metallicity transformation equations (references are given in Table 7). The converted metallicity gradients are given in Table 7. The list includes the gradients obtained only from wide-field imaging data and hence we exclude galaxies with single pointing HST/ACS imaging. This criterion excludes most galaxies from Liu et al. (2011), for which only the central regions of target galaxies were covered. The sample of galaxies in the Table includes two galaxies from this paper plus another ten galaxies. The galaxy stellar masses are also tabulated. In order to calculate the galaxy stellar masses, we used their distance and visual magnitude from NED and the mass to light ratios of Zepf & Ashman (1993). In the sample of twelve galaxies, half have multiple measurements of their metallicity gradients. We include all measured GC gradients and their quoted uncertainties for the 12 galaxies. This comprises 18 measurements of blue GC gradients and 15 measurements of red GC gradients for 12 galaxies.
In Table 7, multiple measurements are given for five galaxies. For the same galaxy, the observed gradients are not always consistent among different works. For example, in the case of NGC 4649 the gradients for the blue subpopulation are 0.00 0.04 (Faifer et al., 2011) and 0.21 0.05 (Strader et al., 2011). Both studies extended to 7 R. In another example, the gradient for the red subpopulation of NGC 3115 is quoted in Arnold et al. (2011) as 0.240.06, while Faifer et al. (2011) quoted 0.110.1. But in the case of NGC 4472, Geisler et al. (1996) and Liu et al. (2011) find consistent gradients for the red subpopulation.
Figure 18 shows the metallicity gradients for blue/red GC subpopulations versus the galaxy stellar mass. We plot multiple measurements for individual galaxies. Linear fits are carried out separately for the blue and the red GCs with uncertainties estimated from the bootstrap technique. The technique uses the errors associated with individual gradients. Best fit relations are:
The galaxies in our sample have a mass range 11.0 < log(M) < 12.0 M. From the above relations, we find that the metallicity gradient for the blue subpopulation has a significant correlation with stellar mass; the negative gradients flattens with increasing stellar mass. As more massive galaxies are expected to accrete more satellites (Oser et al., 2010), we expect more GC accretion to have taken place in these galaxies. This addition of mostly blue GCs at different galactocentric radii may make the initial gradient of the blue GCs shallower. In addition, Hirschmann et al. (2015) found that gradients resulting from major mergers are shallower in more massive galaxies due to radial mixing of GCs. From the spectroscopic metallicities of GC subpopulations in twelve ETGs, Pastorello et al. (2015) observe a similar trend of decreasing metallicity gradient with increasing galaxy stellar mass.
For the red GC subpopulation, we are unable to find a significant relation between the metallicity gradient and galaxy stellar mass. In comparison with the blue GC subpopulations, the metallicity gradients for the red GC subpopulations have higher errors and also a lower number of data points. In Table 7, the least significant gradient measurement is for the red subpopulation of NGC 3115 (Faifer et al., 2011). Hence, we carried out another fitting for the red GCs without that measurement and the best fitted relation is
From the above relation, we infer that the gradients for the red GC subpopulation show a very weak dependence on galaxy stellar mass.
The galaxy stellar mass is derived from the M/L ratios that are given in Zepf & Ashman (1993). We appreciate that the Zepf & Ashman (1993) values are an approximation but have chosen to use them as this is the approach we took in Kartha et al. (2014), which follows from the same approach as used by Rhode et al. (2010) and Spitler et al. (2008). So in order to match the results with the above mentioned publications, we use the same method. Bell et al. (2003) derived the relationships to estimate the stellar mass from galaxy colours (see Appendix B for details). We used their relationship to derive the galaxy stellar mass from (B - V) colour. We find that the equations 7 – 9 are statistically unchanged when using Bell et al. (2003) to derive galaxy stellar masses.
In summary, we suggest that the subsequent addition of GCs from minor mergers may weaken any pre-existing gradients in metallicity (from an early dissipative formation event) both for the blue and the red GC subpopulations.
6.4 Ratio of blue to red GCs with environment density and galaxy mass
Figure 19 shows the ratio of blue to red GCs versus the environment density for a sample of 42 galaxies (two from this paper and forty from Kartha et al. 2014). We note that the blue to red GC ratio is largely insensitive to any GC magnitude incompleteness. With this sample of galaxies, it is evident that neither spiral nor elliptical galaxies show any particular trend in the ratio of blue to red GCs with environment density. NGC 3608, with N/N = 1.85 and density = 0.56 Mpc (Tully, 1988), is consistent with the other elliptical galaxies.
Kartha et al. (2014) found that the fraction of blue to red GCs in lenticular galaxies decreases with local density of environment. This suggests that lenticular galaxies residing in high density environments accommodate a higher fraction of red GCs. NGC 3607 is a lenticular galaxy with a relatively high fraction of red GCs (N/N = 0.79 and density = 0.34 Mpc from Tully 1988). The position of NGC 3607 is consistent with the trend of decreasing fraction of blue to red GCs with increasing environment density.
In Figure 20, the ratio of blue to red GCs is plotted against the host galaxy stellar mass for the above sample of 42 galaxies. There is no obvious correlation between the ratio of blue to red GCs with the galaxy stellar mass. We divide the galaxies into three mass bins of size 0.5 M and derive the mean value for the ratio of blue to red GCs. The mean ratio of blue to red GCs in the low (log(M)< 11 M), intermediate (11 < log(M) < 11.5 M) and high mass (log(M) > 11.5 M) bins are respectively, 1.7 0.8, 2.0 1.2 and 1.6 0.4. The mean ratio of blue to red GCs for the total sample of forty two galaxies is 1.76.
Using cosmological simulations, Bekki et al. (2008) investigated the structural, kinematical and chemical properties of GC systems in different Hubble type galaxies. They estimated the ratio of blue to red GCs, in the host galaxy luminosity range 14 < M < 22 to vary from 50 – 0.25, with an average of 1.5. Using the ACSVCS, Peng et al. (2006) also investigated the ratio of blue to red GCs in a similar luminosity range and determined that the fraction varies from 5.6 to 0.67 percent from low to high luminosity galaxies, suggesting an average ratio of 1.5 blue to red GCs over the total luminosity range.
There is a decreasing trend in the ratio of blue to red GCs with host galaxy luminosity, both observationally (Peng et al., 2006) and theoretically (Bekki et al., 2008). We observe a nearly constant ratio of blue to red GCs in our sample of 42 galaxies because our luminosity range is much more restricted. As seen in the inset of Figure 20, our sample of forty two galaxies lie in the galaxy luminosity range 18.5 M 22 mag, whereas the faint end extends to M = 14 mag for both Peng et al. (2006) and Bekki et al. (2008).
6.5 Formation of GC systems
6.5.1 NGC 3607 and NGC 3608
In the Leo II group, NGC 3607 is the massive central galaxy and has a red GC subpopulation fraction higher than the blue, while the neighbouring galaxy NGC 3608 is less massive and has a higher fraction of blue GCs. An overabundance of red GCs is observed along the minor axis of NGC 3607 (even after removing the GCs in the direction towards NGC 3608). From the azimuthal distribution of GCs of NGC 3608, it is found that both GC subpopulations are aligned in position angle and that angle is different from the position angle of the galaxy stellar light. These results (overabundance and misalignment) suggest a possible interaction between the galaxies in the group.
Using HST data, Lauer et al. (2005) carried out an imaging study of 77 early-type galaxies, including NGC 3607. They detected an additional gas disk settling in NGC 3607 perpendicular to the existing dusty disk. They commented that the dusty disk is in a transition phase merging with the gas disk. They explained this process as gas infalling directly onto the centre of NGC 3607 without disturbing the dusty disk and without any obvious features of interaction.
Later, Annibali et al. (2007) studied the stellar population properties of 66 early-type galaxies. They estimated the age, metallicity and alpha enhancement using the Lick indices with updated simple stellar population models (including the non-solar element abundance patterns). They estimated a very young age, 3.1 0.5 Gyr, for NGC 3607 and suggested it had experienced a recent episode of star formation. Rickes et al. (2009) carried out long slit spectroscopy, out to galactocentric radii of 30.5 arcsec, and claimed that NGC 3607 has undergone a minimum of three star formation episodes with ages ranging from 1 to 13 Gyr. The young age for the stellar population of NGC 3607 and the detection of a central gas disk indicate that NGC 3607 has experienced a recent star formation episode and the overabundance of red GCs may be due to GCs formed in that episode.
From the ATLAS survey, McDermid et al. (2015) estimated the mass-weighted ages for NGC 3605, NGC 3607 and NGC 3608 as 8.1 0.8, 13.5 0.7 and 13.0 0.7 Gyr respectively. They utilised the spectra within 1R to fit the single stellar population models and hence derive the mass-weighted ages, metallicity and star formation histories of 260 ETGs. Using the Lick indices, they estimated the age for NGC 3607 as 7.3 1.3 Gyr that contradicts the young age determined by Annibali et al. (2007).
Forbes et al. (2006a) carried out a multi-wavelength (X-ray, optical and Hi imaging) study of 60 galaxy groups, including the Leo II group. They investigated the evolutionary connections between different groups and the influence of group environment. In their study, they detected extended X-ray emission associated with the Leo II group but did not resolve individual galaxies. Recently, using Chandra X-ray data, Jang et al. (2014) observed X-ray emission from the central AGN in NGC 3607 and diffuse emission around NGC 3608. The detection of extended X-ray emissions confirms the presence of hot intergalactic gas.
The misalignment in the position angles of the GCs relative to the galaxy in NGC 3608 might be another sign of interaction with NGC 3607. Additionally, each galaxy shows an overabundance of GCs in the direct of the other, again suggesting a possible interaction between them. Jedrzejewski & Schechter (1988) proposed a close encounter between these two galaxies. They studied the absorption line kinematics for the stellar component of NGC 3608 and found a change in direction of the rotation curve between the core and outside region. They proposed that the reversal might be due to an interaction with the nearby NGC 3607.
We conclude that our results also support a possible interaction between the two galaxies. To confirm this proposition, deep surface photometric and detailed kinematic studies are needed.
6.5.2 Formation scenarios
As described in the introduction, three ’classic’ formation scenarios were proposed to explain bimodality in globular cluster systems. In the major merger model (Ashman & Zepf, 1992), the blue GCs already exist in the merging galaxies, while the red GCs form during the merging process. In the multi-phase collapse scenario (Forbes et al., 1997), the blue GCs are formed early, followed by a quiescent phase. After a few Gyrs, star formation is restarted with the formation of red GCs, which can be followed by accretion of additional blue GCs. According to Côté et al. (1998, 2000), the red GCs are inherent to the parent galaxies and the blue GCs are purely accreted from dwarf galaxies.
The three classic scenarios were explored in cosmological simulations which addressed a variety of GC system properties: structural and kinematical (Bekki et al., 2005), dynamical and chemical (Bekki et al., 2008), colour and metallicity bimodality (Muratov & Gnedin, 2010; Tonini, 2013), as well as physical relationships with the host galaxies (Beasley et al., 2002). Recently, Trenti et al. (2015) proposed another scenario for GC formation from the merging of multiple gas rich mini halos.
In all the classic formation scenarios, there is a strong association between red GC subpopulations and the parent galaxy. This relationship is established from different observations such as the strong relation between red GC peak colour and galaxy luminosity (Peng et al., 2006; Strader et al., 2006; Faifer et al., 2011), position angle arrangement of red GCs and the galaxy stellar component (Wang et al., 2013), connection between rotation velocity for red GCs and field stars (Pota et al., 2013) etc. On the other hand, the association between blue GC subpopulations and parent galaxy stars is weak. Peacock et al. (2015) found that the blue GC subpopulations of NGC 3115 are consistent with the stellar halo in metallicity and spatial distributions. However, the origin of the blue GC subpopulation is quite controversial. Côté et al. (1998, 2000) and Tonini (2013) proposed a dissipationless accretion origin whereas dissipational in-situ formation (Forbes et al., 1997; Beasley et al., 2003) is suggested for the formation of blue GCs in the inner regions. This distinction in region (inner or outer) is mentioned since accretion of blue GCs to the galaxy outskirts in the later phase is also included in the multi-phase scenario (Forbes et al., 1997).
Strader et al. (2004, 2005) investigated the feasibility of the above formation scenarios using observational data for massive elliptical galaxies. From the GC colour-galaxy luminosity relation and the age-metallicity relation, they proposed an in-situ plus accretion model for the formation of inner blue GCs which were then truncated by reionization, whereas the red GCs formed along with the bulk of field stars. They suggested that dwarf galaxies residing in overdense regions collapse before dwarfs in less dense regions, and then accrete more enriched gas from nearby star forming regions. These dwarf galaxies, along with their blue GCs, are later accreted into the halo of a massive galaxy forming part of the main system. This implies an in-situ+accretion origin for blue GCs. Hence, the origin of blue GCs in the inner regions could be due to one of three proposed processes, i.e. completely in-situ, fully accreted or in-situ+accretion.
In the following paragraphs, we try to differentiate between these three formation processes for blue GCs based on their global properties. In particular, we measure radial density, radial colour and azimuthal distributions in relation to their parent galaxies.
From the azimuthal distribution of GC subpopulations, both blue and red GCs have a positional arrangement in common with the galaxy stellar light component (Wang et al., 2013). This suggests that the blue GC subpopulation and galaxy stellar component have similar evolutionary histories. For galaxies in which the blue GCs and stars accreted from satellite dwarfs, this similarity is expected (Côté et al., 2001). From the derived ellipticities it is seen that red GC subpopulations have a one-to-one relation with the galaxy stellar component whereas the relation is not tight for blue GC subpopulations (see Figure 17). If the galaxy has accreted its blue GCs recently, then a complete one-to-one correlation with host galaxy properties is not expected. Park & Lee (2013) also investigated this relationship for 23 early-type galaxies using ACSVCS data and found an approximate one-to-one relation between blue GC ellipticity and the galaxy stellar component. As is well-known, the ACS field of view does not provide anywhere near complete coverage for massive nearby galaxies (Peng et al., 2006). That means a nearly one-to-one relation between inner blue GCs and galaxy stellar component suggests a common origin for both and hence, supports the in-situ formation scenario.
Another diagnostic trend is the GC subpopulation peak colour versus galaxy luminosity. The peak colour of the red GC subpopulation gets redder with increasing galaxy luminosity. Perhaps a weaker correlation exists for the blue GC subpopulation. Liu et al. (2011) found that projection effects tend to flatten GC radial trends, particularly for the blue subpopulation because if its extended nature. Hence, the slope of the relation between the blue GC subpopulation peak colour and galaxy luminosity is reduced to half of the earlier value (0.0126 0.0025: Peng et al. 2006), making the relation between peak colour of the blue GC subpopulation and galaxy luminosity insignificant. This result weakens the idea that the formation of blue GCs is via in-situ processes.
Radial colour gradients may also reveal the origin of blue GCs. The colour gradients for blue GCs formed in-situ are expected to be steeper than for a subpopulation formed from in-situ+accretion or completely accreted processes. We expect this because the addition of GCs through accretion can dilute (in the case of in-situ+accretion) the existing colour gradient for the blue GC subpopulation. In the case of complete accretion, we assume zero colour gradient for the blue subpopulation. Hence, to disentangle the origin of blue GCs, the steepness of the gradient needs to be quantified with large samples of galaxies where the colour gradients are measured with maximum accuracy. Our present work is limited by a small sample of 10 galaxies collected from the literature (Geisler et al., 1996; Bassino et al., 2006a; Harris, 2009; Forbes et al., 2011; Faifer et al., 2011; Arnold et al., 2011; Blom et al., 2012; Usher et al., 2013; Hargis & Rhode, 2014) and two from this work. Liu et al. (2011) carried out an analysis of the colour gradients for 76 early-type galaxies using ACSVCS and ACS Fornax Cluster Survey (FCS). Even though the sample size is impressive, only three galaxies have more than one pointing and we have included them in the above sample. Hence, significant color gradients are detected in a total of 12 galaxies, five of which have multiple measurement. Gradient values are provided in Table 7.
Figure 18 show this sample of GC metallicity gradients plotted against host galaxy stellar mass. The blue GC subpopulation shows a trend of decreasing gradient with increasing galaxy stellar mass. This implies that high mass galaxies have shallower gradients, whereas low mass (log(M) 11.0 M) galaxies have steeper gradients. As the metallicity gradients show a dependency on galaxy stellar mass, both the GC subpopulations are expected to have some formational similarities with the galaxy stellar component. This means that a completely accreted origin (Côté et al., 1998, 2000; Tonini, 2013) may not be the best scenario to explain the formation of blue GCs. Also, we notice that both GC subpopulation gradients show a dependence on galaxy stellar mass. Thus, a common, or in-situ origin (Forbes et al., 1997; Beasley et al., 2003), is probably involved in the formation of blue and red GC subpopulations (Pastorello et al., 2015). However, we note that large red (early-type) galaxies tend to preferentially accrete red satellite galaxies (Hearin et al., 2014; Hudson et al., 2015). Thus GC system metallicity gradients may also reflect the gradients of the accreted satellites, if they are preserved in the accretion process. In the in-situ+accretion formation scenario (Strader et al., 2004, 2005) for the blue GCs, we expect the gradient to be shallower than for the blue GCs formed completely in-situ, but a reference scale is not yet established by models.
To summarise, from the present study it is difficult to ascribe either a completely in-situ or an in-situ+accretion origin for the blue GC subpopulations. A homogeneous large sample with accurate GC properties is needed to address this issue in depth.
We present wide-field imaging data from the Subaru telescope with which we carry out an investigation of the GC systems in the Leo II group to large galactocentric radii ( 120 kpc). Using the multi-band wide-field images in g, r and i filters, we analysed the radial density, radial colour and azimuthal distributions of GC systems in the two brightest galaxies of the group, NGC 3607 and NGC 3608. Our study is complemented with spectroscopic data obtained from DEIMOS on the Keck II telescope. We present the main conclusions here.
The GC systems of NGC 3607 and NGC 3808 are found to have radial extents of 9.5 0.6 arcmin (equivalent to 61 5 kpc or 4.4 R) and 6.6 0.8 arcmin (equivalent to 43 5 kpc or 4.7 R), respectively. The derived values are in agreement with estimates obtained from the empirical relation between the effective radius of the GC system and galaxy stellar light given in Kartha et al. (2014).
The GC system colours of both galaxies are fitted with the GMM algorithm and we detect a bimodal distribution with confidence level greater than 99.99 percent. NGC 3607 is observed to have 45 9 and 55 8 percent of blue and red GC subpopulations, while for NGC 3608 the blue and red GC subpopulations contribute 65 6 and 35 6 percent to the total GC system.
From the radial velocity measurements, we detect 81 GCs in the field of the Leo II group. We assign 46 and 35 GCs, respectively, to NGC 3607 and NGC 3608. We estimate a mean velocity of 963 and 1220 km/s for NGC 3607 and NGC 3608, respectively. Also, the mean GC velocity dispersions for the respective galaxies are 167 and 147 km/s.
From the radial density distributions of the GC subpopulations of NGC 3607, the red subpopulation is more centrally located while the blue subpopulation is more extended. Also, the effective radius of the red GC subpopulation (40 29 arcsec) and the galaxy stellar light (39 arcsec) are in good agreement, compared to the blue subpopulation (95 50 arcsec).
For NGC 3608, the blue subpopulation is more extended in radius than the centrally concentrated red subpopulation. The red subpopulation distribution shows similarities with the galaxy surface brightness distribution. However, the effective radius of the red subpopulation (59 40 arcsec) is larger than the galaxy stellar light (30 arcsec).
The azimuthal distribution of the NGC 3607 GC system reveals that both subpopulations are aligned along a position angle ( 110 degrees), which is in reasonable agreement with the galaxy stellar light (125 degrees). However, the distribution of the GC system is more elliptical in comparison with the circular distribution of galaxy stellar light. The red subpopulation shows a more elliptical distribution when compared with the blue subpopulation.
For NGC 3608, the GCs are arranged along position angles that are different from the galaxy stellar population. Using two different methods of GC selection, the position angles for the total GC system are found to be along 104 15 and 67 7 degrees, while the galaxy major axis is at 82 degrees. One method of GC selection suggests that the GCs have an ellipticity = 0.20 0.09, while the other shows an ellipticity of 0.39 0.10. By comparison, the stellar light ellipticity is 0.20. In NGC 3608, the blue subpopulation has a more elliptical arrangement than the red subpopulation.
The total GC system, and both subpopulations of NGC 3607, become bluer in colour with increasing galactocentric radius; a significant metallicity gradient is observed for both subpopulations. We find that the blue subpopulation has a steeper gradient than the red subpopulation. We also detect a strong colour gradient only for the blue subpopulation of NGC 3608. The colour gradient for the blue subpopulation in NGC 3608 is steeper than that in NGC 3607.
We compare different global properties of the GC systems and their parent galaxies. We reconfirm that the extent of the GC system is a function of galaxy size and the effective radius of a GC system is nearly 6 times the effective radius of parent galaxy. We obtain a one-to-one relation between the parent red GC ellipticities and galaxy stellar light ellipticities. Also, the blue GC ellipticities of slow rotators with kinematically decoupled cores are more elongated than their red GC subpopulation ellipticities. We propose that they might have experienced recent minor mergers from anisotropic directions (Tempel et al., 2015).
From a sample of twelve galaxies, we investigate the relationship between the metallicity gradients and host galaxy stellar mass. We found that the gradients of both GC subpopulations become shallower with increasing stellar mass. The average ratio of blue to red GCs in galaxies in the mass range 11.0 log(M) 12.0 M is nearly 1.7. These findings agree with the predictions from the simulations of Bekki et al. (2008) and also with the findings from other observations (Peng et al., 2006). We also carried out a study to disentangle the formation of blue GC subpopulations (i.e. completely in-situ versus in-situ+accretion versus completely accreted), which have not given conclusive results and need to be followed up with a homogeneous, large sample.
We thank the anonymous referee for his/her careful read- ing of the manuscript and the valuable feedbacks. The authors extend their gratitude to Jacob A. Arnold and Kristin A. Woodley regarding their help in observations. We thank Blesson Mathew and Nicola Pastorello for the careful reading of the manuscript. We also acknowledge the members of SAGES group, especially Christopher Usher, Vincenzo Pota and Joachim Janz, for the support and enlightening discussions. SSK thanks the Swinburne University for the SUPRA fellowship. DAF thanks the ARC for support via DP-130100388. This work was supported by NSF grant AST-1211995. This paper was based in part on data collected at Subaru Telescope, which is operated by the National Astronomical Observatory of Japan. This paper uses data products produced by the OIR Telescope Data Center, supported by the Smithsonian Astrophysical Observatory. Based on observations made with the NASA/ESA Hubble Space Telescope, and obtained from the Hubble Legacy Archive, which is a collaboration between the Space Telescope Science Institute (STScI/NASA), the Space Telescope European Coordinating Facility (ST-ECF/ESA) and the Canadian Astronomy Data Centre (CADC/NRC/CSA). Some of the data presented herein were obtained at the W. M. Keck Observatory, operated as a scientific partnership among the California Institute of Technology, the University of California and the National Aeronautics and Space Administration, and made possible by the generous financial support of the W. M. Keck Foundation. The authors wish to recognize and acknowledge the very significant cultural role and reverence that the summit of Mauna Kea has always had within the indigenous Hawaiian community. We are most fortunate to have the opportunity to conduct observations from this mountain. This research has made use of the NASA/IPAC Extragalactic Data base (NED) operated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration. The analysis pipeline used to reduce the DEIMOS data was developed at UC Berkeley with support from NSF grant AST-0071048. We acknowledge the usage of the HyperLeda database (http://leda.univ-lyon1.fr).
Appendix A List of spectroscopically confirmed objects around the Leo II group
Table 8 presents the photometric magnitudes g, r and i and the radial velocities for GCs, Galactic stars and background galaxies detected around NGC 3607 and NGC 3608 in the Leo II group.
Appendix B M/l ratio calculation using Bell et al. (2003)
Relationships between stellar M/L values and various colours in SDSS and 2MASS passbands are given in Bell et al. (2003). They derived these relationships by fitting galaxy evolution models to a large sample of 22679 galaxies from the SDSS Early Data Release (Stoughton et al. 2002) and 2MASS extended source catalog (Jarrett et al. 2000). To estimate the stellar mass for our sample of 42 galaxies, we utilize the relationship between M/L ratio and (B V) colour which is given below.
We find that the Bell et al. (2003) M/L ratios are about a factor of 2 times lower for ellipticals and a factor of 1.5 times lower for lenticulars than Zepf & Ashman (1993) values. This affects the X-axes of Figures 18 and 20. Hence, we fit the trends in Figure 18 after incorporating the stellar mass from Bell et al. (2003). The fits are given below which can be compared to Equations 7 – 9.
We find that even if the stellar mass varies between Zepf & Ashman (1993) and Bell et al. (2003), the relationships shown by blue and red GCs with metallicity remains statistically the same. This also implies that our results remain unchanged between different M/L ratio estimations.
- Alamo-Martínez et al. (2012) Alamo-Martínez K. A., West M. J., Blakeslee J. P., González-Lópezlira R. A., Jordán A., Gregg M., Côté P., Drinkwater M. J., van den Bergh S., 2012, A&A, 546, A15
- Annibali et al. (2007) Annibali F., Bressan A., Rampazzo R., Zeilinger W. W., Danese L., 2007, A&A, 463, 455
- Arnold et al. (2011) Arnold J. A., Romanowsky A. J., Brodie J. P., Chomiuk L., Spitler L. R., Strader J., Benson A. J., Forbes D. A., 2011, ApJL, 736, L26
- Ashman & Zepf (1992) Ashman K. M., Zepf S. E., 1992, ApJ, 384, 50
- Bassino et al. (2006a) Bassino L. P., Faifer F. R., Forte J. C., Dirsch B., Richtler T., Geisler D., Schuberth Y., 2006a, A&A, 451, 789
- Bassino et al. (2006b) Bassino L. P., Richtler T., Dirsch B., 2006b, MNRAS, 367, 156
- Beasley et al. (2002) Beasley M. A., Baugh C. M., Forbes D. A., Sharples R. M., Frenk C. S., 2002, MNRAS, 333, 383
- Beasley et al. (2003) Beasley M. A., Harris W. E., Harris G. L. H., Forbes D. A., 2003, MNRAS, 340, 341
- Bekki et al. (2005) Bekki K., Beasley M. A., Brodie J. P., Forbes D. A., 2005, MNRAS, 363, 1211
- Bekki et al. (2008) Bekki K., Yahagi H., Nagashima M., Forbes D. A., 2008, MNRAS, 387, 1131
- Bell et al. (2003) Bell E. F., McIntosh D. H., Katz N., Weinberg M. D., 2003, ApJS, 149, 289
- Bertin & Arnouts (1996) Bertin E., Arnouts S., 1996, A&AS, 117, 393
- Blom et al. (2012) Blom C., Spitler L. R., Forbes D. A., 2012, MNRAS, 420, 37
- Brodie et al. (2011) Brodie J. P., Romanowsky A. J., Strader J., Forbes D. A., 2011, AJ, 142, 199
- Brodie et al. (2014) Brodie J. P. et al., 2014, ApJ, 796, 52
- Brodie & Strader (2006) Brodie J. P., Strader J., 2006, ARAA, 44, 193
- Brodie et al. (2012) Brodie J. P., Usher C., Conroy C., Strader J., Arnold J. A., Forbes D. A., Romanowsky A. J., 2012, ApJL, 759, L33
- Cantiello et al. (2015) Cantiello M. et al., 2015, A&A, 576, A14
- Cappellari et al. (2011) Cappellari M. et al., 2011, MNRAS, 413, 813
- Cappellari et al. (2013) Cappellari M. et al., 2013, MNRAS, 432, 1862
- Coccato et al. (2009) Coccato L. et al., 2009, MNRAS, 394, 1249
- Côté et al. (1998) Côté P., Marzke R. O., West M. J., 1998, ApJ, 501, 554
- Côté et al. (2000) Côté P., Marzke R. O., West M. J., Minniti D., 2000, ApJ, 533, 869
- Côté et al. (2001) Côté P., McLaughlin D. E., Hanes D. A., Bridges T. J., Geisler D., Merritt D., Hesser J. E., Harris G. L. H., Lee M. G., 2001, ApJ, 559, 828
- de Vaucouleurs et al. (1991) de Vaucouleurs G., de Vaucouleurs A., Corwin, Jr. H. G., Buta R. J., Paturel G., Fouqué P., 1991, Third Reference Catalogue of Bright Galaxies. Volume I: Explanations and references. Volume II: Data for galaxies between 0 and 12. Volume III: Data for galaxies between 12 and 24.
- Di Matteo et al. (2009) Di Matteo P., Pipino A., Lehnert M. D., Combes F., Semelin B., 2009, A&A, 499, 427
- Dirsch et al. (2005) Dirsch B., Schuberth Y., Richtler T., 2005, A&A, 433, 43
- Dotter et al. (2010) Dotter A. et al., 2010, ApJ, 708, 698
- Duc et al. (2015) Duc P.-A. et al., 2015, MNRAS, 446, 120
- Emsellem et al. (2011) Emsellem E. et al., 2011, MNRAS, 414, 888
- Faber et al. (2003) Faber S. M. et al., 2003, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, Vol. 4841, Instrument Design and Performance for Optical/Infrared Ground-based Telescopes, Iye M., Moorwood A. F. M., eds., pp. 1657–1669
- Faifer et al. (2011) Faifer F. R., Forte J. C., Norris M. A., Bridges T., Forbes D. A., Zepf S. E., Beasley M., Gebhardt K., Hanes D. A., Sharples R. M., 2011, MNRAS, 416, 155
- Forbes et al. (1997) Forbes D. A., Brodie J. P., Grillmair C. J., 1997, AJ, 113, 1652
- Forbes & Forte (2001) Forbes D. A., Forte J. C., 2001, MNRAS, 322, 257
- Forbes et al. (2015) Forbes D. A., Pastorello N., Romanowsky A. J., Usher C., Brodie J. P., Strader J., 2015, MNRAS, 452, 1045
- Forbes et al. (2012) Forbes D. A., Ponman T., O’Sullivan E., 2012, MNRAS, 425, 66
- Forbes et al. (2006a) Forbes D. A., Ponman T., Pearce F., Osmond J., Kilborn V., Brough S., Raychaudhury S., Mundell C., Miles T., Kern K., 2006a, PASA, 23, 38
- Forbes et al. (2006b) Forbes D. A., Sánchez-Blázquez P., Phan A. T. T., Brodie J. P., Strader J., Spitler L., 2006b, MNRAS, 366, 1230
- Forbes et al. (2011) Forbes D. A., Spitler L. R., Strader J., Romanowsky A. J., Brodie J. P., Foster C., 2011, MNRAS, 413, 2943
- Forte et al. (2012) Forte J. C., Vega E. I., Faifer F., 2012, MNRAS, 421, 635
- Geisler et al. (1996) Geisler D., Lee M. G., Kim E., 1996, AJ, 111, 1529
- Gnedin et al. (2014) Gnedin O. Y., Ostriker J. P., Tremaine S., 2014, ApJ, 785, 71
- Graham & Driver (2005) Graham A. W., Driver S. P., 2005, PASA, 22, 118
- Griffen et al. (2010) Griffen B. F., Drinkwater M. J., Thomas P. A., Helly J. C., Pimbblet K. A., 2010, MNRAS, 405, 375
- Hargis & Rhode (2014) Hargis J. R., Rhode K. L., 2014, ApJ, 796, 62
- Harris (2009) Harris W. E., 2009, ApJ, 703, 939
- Harris (2010) Harris W. E., 2010, Royal Society of London Philosophical Transactions Series A, 368, 889
- Harris et al. (2000) Harris W. E., Kavelaars J. J., Hanes D. A., Hesser J. E., Pritchet C. J., 2000, ApJ, 533, 137
- Hearin et al. (2014) Hearin A. P., Watson D. F., Becker M. R., Reyes R., Berlind A. A., Zentner A. R., 2014, MNRAS, 444, 729
- Hirschmann et al. (2015) Hirschmann M., Naab T., Ostriker J. P., Forbes D. A., Duc P.-A., Davé R., Oser L., Karabal E., 2015, MNRAS, 449, 528
- Hopkins et al. (2009) Hopkins P. F., Cox T. J., Dutta S. N., Hernquist L., Kormendy J., Lauer T. R., 2009, ApJS, 181, 135
- Hudson et al. (2015) Hudson M. J. et al., 2015, MNRAS, 447, 298
- Jang et al. (2014) Jang I., Gliozzi M., Hughes C., Titarchuk L., 2014, MNRAS, 443, 72
- Jarrett et al. (2000) Jarrett T. H., Chester T., Cutri R., Schneider S., Skrutskie M., Huchra J. P., 2000, AJ, 119, 2498
- Jedrzejewski & Schechter (1988) Jedrzejewski R., Schechter P. L., 1988, ApJL, 330, L87
- Kartha et al. (2014) Kartha S. S., Forbes D. A., Spitler L. R., Romanowsky A. J., Arnold J. A., Brodie J. P., 2014, MNRAS, 437, 273
- Katz & Ricotti (2013) Katz H., Ricotti M., 2013, MNRAS, 432, 3250
- Kim et al. (2013) Kim H.-S., Yoon S.-J., Sohn S. T., Kim S. C., Kim E., Chung C., Lee S.-Y., Lee Y.-W., 2013, ApJ, 763, 40
- Krajnović et al. (2011) Krajnović D. et al., 2011, MNRAS, 414, 2923
- Kron (1980) Kron R. G., 1980, ApJS, 43, 305
- Kundu & Whitmore (2001a) Kundu A., Whitmore B. C., 2001a, AJ, 121, 2950
- Kundu & Whitmore (2001b) Kundu A., Whitmore B. C., 2001b, AJ, 122, 1251
- Larsen et al. (2001) Larsen S. S., Brodie J. P., Huchra J. P., Forbes D. A., Grillmair C. J., 2001, AJ, 121, 2974
- Lauer et al. (2005) Lauer T. R. et al., 2005, AJ, 129, 2138
- Li & Gnedin (2014) Li H., Gnedin O. Y., 2014, ApJ, 796, 10
- Liu et al. (2011) Liu C., Peng E. W., Jordán A., Ferrarese L., Blakeslee J. P., Côté P., Mei S., 2011, ApJ, 728, 116
- Maraston (2005) Maraston C., 2005, MNRAS, 362, 799
- McDermid et al. (2015) McDermid R. M. et al., 2015, MNRAS, 448, 3484
- McLaughlin et al. (1994) McLaughlin D. E., Harris W. E., Hanes D. A., 1994, ApJ, 422, 486
- Miyazaki et al. (2002) Miyazaki S. et al., 2002, PASJ, 54, 833
- Mulchaey et al. (2003) Mulchaey J. S., Davis D. S., Mushotzky R. F., Burstein D., 2003, ApJS, 145, 39
- Muratov & Gnedin (2010) Muratov A. L., Gnedin O. Y., 2010, ApJ, 718, 1266
- Naab et al. (2014) Naab T. et al., 2014, MNRAS, 444, 3357
- Oser et al. (2010) Oser L., Ostriker J. P., Naab T., Johansson P. H., Burkert A., 2010, ApJ, 725, 2312
- Ouchi et al. (2004) Ouchi M. et al., 2004, ApJ, 611, 660
- Park & Lee (2013) Park H. S., Lee M. G., 2013, ApJL, 773, L27
- Pastorello et al. (2015) Pastorello N. et al., 2015, MNRAS, 451, 2625
- Paturel et al. (2003) Paturel G., Petit C., Prugniel P., Theureau G., Rousseau J., Brouty M., Dubois P., Cambrésy L., 2003, A&A, 412, 45
- Peacock et al. (2015) Peacock M. B., Strader J., Romanowsky A. J., Brodie J. P., 2015, ApJ, 800, 13
- Peng et al. (2006) Peng E. W., Jordán A., Côté P., Blakeslee J. P., Ferrarese L., Mei S., West M. J., Merritt D., Milosavljević M., Tonry J. L., 2006, ApJ, 639, 95
- Pipino et al. (2010) Pipino A., D’Ercole A., Chiappini C., Matteucci F., 2010, MNRAS, 407, 1347
- Pota et al. (2013) Pota V. et al., 2013, MNRAS, 428, 389
- Rhode et al. (2010) Rhode K. L., Windschitl J. L., Young M. D., 2010, AJ, 140, 430
- Rickes et al. (2009) Rickes M. G., Pastoriza M. G., Bonatto C., 2009, A&A, 505, 73
- Schlegel et al. (1998) Schlegel D. J., Finkbeiner D. P., Davis M., 1998, ApJ, 500, 525
- Spitler (2010) Spitler L. R., 2010, MNRAS, 406, 1125
- Spitler et al. (2008) Spitler L. R., Forbes D. A., Strader J., Brodie J. P., Gallagher J. S., 2008, MNRAS, 385, 361
- Spitler et al. (2006) Spitler L. R., Larsen S. S., Strader J., Brodie J. P., Forbes D. A., Beasley M. A., 2006, AJ, 132, 1593
- Stoughton et al. (2002) Stoughton C. et al., 2002, AJ, 123, 485
- Strader et al. (2005) Strader J., Brodie J. P., Cenarro A. J., Beasley M. A., Forbes D. A., 2005, AJ, 130, 1315
- Strader et al. (2004) Strader J., Brodie J. P., Forbes D. A., 2004, AJ, 127, 3431
- Strader et al. (2006) Strader J., Brodie J. P., Spitler L., Beasley M. A., 2006, AJ, 132, 2333
- Strader et al. (2012) Strader J., Fabbiano G., Luo B., Kim D.-W., Brodie J. P., Fragos T., Gallagher J. S., Kalogera V., King A., Zezas A., 2012, ApJ, 760, 87
- Strader et al. (2011) Strader J., Romanowsky A. J., Brodie J. P., Spitler L. R., Beasley M. A., Arnold J. A., Tamura N., Sharples R. M., Arimoto N., 2011, ApJS, 197, 33
- Tempel et al. (2015) Tempel E., Guo Q., Kipper R., Libeskind N. I., 2015, MNRAS, 450, 2727
- Terlevich & Forbes (2002) Terlevich A. I., Forbes D. A., 2002, MNRAS, 330, 547
- Tonini (2013) Tonini C., 2013, ApJ, 762, 39
- Tortora et al. (2010) Tortora C., Napolitano N. R., Cardone V. F., Capaccioli M., Jetzer P., Molinaro R., 2010, MNRAS, 407, 144
- Trenti et al. (2015) Trenti M., Padoan P., Jimenez R., 2015, ApJL, 808, L35
- Tully (1988) Tully R. B., 1988, Nearby galaxies catalog, Cambridge and New York, Cambridge University Press, 221 p
- Usher et al. (2012) Usher C., Forbes D. A., Brodie J. P., Foster C., Spitler L. R., Arnold J. A., Romanowsky A. J., Strader J., Pota V., 2012, MNRAS, 426, 1475
- Usher et al. (2013) Usher C., Forbes D. A., Spitler L. R., Brodie J. P., Romanowsky A. J., Strader J., Woodley K. A., 2013, MNRAS, 436, 1172
- Walker et al. (2006) Walker M. G., Mateo M., Olszewski E. W., Pal J. K., Sen B., Woodroofe M., 2006, ApJL, 642, L41
- Wang et al. (2013) Wang Q., Peng E. W., Blakeslee J. P., Côté P., Ferrarese L., Jordán A., Mei S., West M. J., 2013, ApJ, 769, 145
- Zepf & Ashman (1993) Zepf S. E., Ashman K. M., 1993, MNRAS, 264, 611