ANGULAR MOMENTA, DYNAMICAL MASSES, AND MERGERS OF BRIGHTEST CLUSTER GALAXIES^*

Brightest cluster galaxies (BCGs) exist in an extreme environment. They are typically located in the center of their cluster, near the bottom of the cluster's potential well. As a result of their location, BCGs are expected to undergo more mergers than a typical galaxy. The increased frequency of merging events in BCGs should make them sensitive to the relation between galaxy mergers and angular momentum.

The large number of merging events is also likely the cause of the high mass and early-type classification observed in a typical BCG. Observations by Lidman et al. (2012) have shown that BCGs have grown in mass since z = 1.6 and on average have grown by a factor of 1.8 ± 0.3 from z = 0.9 to z = 0.2. Similarly, simulations of galaxy growth via mergers by De Lucia & Blaizot (2007) have shown that only half the mass of a BCG is in place by z = 0.5. The same simulations found that the majority of stars within a BCG formed very early, with 50% of the stars being formed before z = 5 (De Lucia & Blaizot 2007). In this scenario, BCGs would have accumulated half of their mass in the time frame from z = 0.5 to z = 0 via gas-poor mergers.

The fact that multiple minor dry mergers have been shown to remove angular momentum from a galaxy when the mergers are isotropically distributed, combined with the fact that BCGs in simulations undergo a higher than average number of dry merger events, leads us to hypothesize that BCGs are likely to have a lower average angular momentum when compared to standard elliptical galaxies. In other studies BCGs have been shown to be remarkably uniform and exhibit special characteristics when compared to elliptical galaxies not at the center of their cluster. von der Linden et al. (2007) showed that a typical BCG is more likely to host a radio-loud active galactic nucleus (AGN), is larger in size, and has higher velocity dispersion on average than a typical elliptical galaxy in the same mass range. These currently observed differences are likely the result of the combination of two effects: BCGs' large stellar masses and BCGs' location at the bottom of their host clusters' potential well, although it is uncertain which effect dominates.

Specific features of BCGs, such as their large radii and low surface brightnesses compared to normal elliptical galaxies are consistent with the products of major dissipationless mergers (e.g., Oegerle & Hoessel 1991; Brough et al. 2005; von der Linden et al. 2007; Lauer et al. 2007; Tran et al. 2008). Their sizes and velocity dispersions may have also evolved faster than less-massive early-type galaxies since z ∼ 0.3 (Bernardi 2009; although, see also Stott et al. 2011). These two factors cause the Faber–Jackson relation between luminosity and velocity dispersion to be flatter for BCGs. Similar studies of the fundamental plane (FP) have found differing slopes between populations of BCGs and elliptical galaxies (Desroches et al. 2007). The differing slope of the Faber–Jackson relationship and the FP for BCGs compared to ellipticals supports the scenario where BCGs form mainly via dissipationless mergers (Boylan-Kolchin et al. 2006).

By combining spectroscopic information with photometric information, we examine the recent merger history of a sample of 10 BCGs. Integral field unit (IFU) spectroscopy allows us to observe the kinematic properties of galaxies in two dimensions, showing whether they are largely rotation-supported or dispersion-supported. The SAURON team have developed the λ_R parameter, which utilizes the increased spatial information from IFU spectroscopy to quantify the observed stellar angular momentum in galaxies (Emsellem et al. 2007). They used this parameter to quantify the angular momentum of 48 early-type galaxies. This work was followed by the ATLAS^3D team, which performed a similar analysis on 260 galaxies (Emsellem et al. 2011). In both studies they classified galaxies as either fast rotators (FRs) or slow rotators (SRs) based on their λ_R value within 1R_e.

The λ_R parameter provides a quantitative method to compare the stellar kinematics of BCGs to those of other early-type galaxies, however the SAURON sample has only four galaxies with M_dyn > 10^11.5 M_☉ and only includes one BCG (M87). The ATLAS^3D survey studies an additional 10 galaxies with M_dyn > 10^11.5 M_☉. Above 10^11.5 M_☉ only 23% of galaxies from the ATLAS^3D survey are classified as FRs (Emsellem et al. 2011). The mass limit of 10^11.5 may be important as Peng et al. (2010) have shown that above M_* = 10^11.5 M_☉ the majority of galaxies have undergone a major merger after their star formation was quenched. If dry mergers preferentially remove angular momentum, then we would expect to see fewer BCGs above 10^11.5 M_☉ with a high λ_R value.

In order to determine the recent merger history of our galaxies, we use a combination of Source Extractor (SExtractor; Bertin & Arnouts 1996), GALFIT (Peng et al. 2002), and PyMorph (Vikram et al. 2010) to produce photometric models of these galaxies. We then subtract our models from the original image to identify tidal tails, multiple cores, and excess intra-cluster light, which would all be signs of merging that can be identified visually. A more quantitative detection of merging comes from the G − M₂₀ (Gini coefficient minus the second order moment of the brightest 20% of pixels) value for each galaxy. According to simulations, these parameters would tell us if there was a dry merger within the last 0.2 Gyr (Lotz et al. 2011). After 0.2 Gyr merger information will have been erased by dynamical friction, meaning this selection criterion is only able to inform us of very recent or currently ongoing mergers.

It is our goal to measure λ_R, dynamical mass, and recent merger history for our sample of BCGs and compare our kinematic results to the SAURON and ATLAS^3D results for early-type galaxies. We present targeted observations of BCG stellar kinematics and photometrics for seven BCGs with close companions and three BCGs with no close companion. In our initial selection, we specifically chose BCGs with close companions to determine if they were bound to each other. We also determine if morphological signatures of merging are correlated with λ_Re measurements.

We assume a Hubble constant of H₀ = 70 km s⁻¹ Mpc⁻¹ and Ω_M = 0.3, Ω_Λ = 0.7.

In von der Linden et al. (2007), a sample of 625 BCGs (z < 0.1) were selected from the C4 cluster catalog (Miller et al. 2005) of the Third Data Release of the Sloan Digital Sky Survey (SDSS; York et al. 2000). In Brough et al. (2011) BCGs with companions within ∼10'' (18 kpc at z ∼ 0.1) were chosen from the von der Linden sample. We are using the same four galaxies from Brough et al. (2011) as well as seven new galaxies chosen by the same criteria as Brough et al. (2011). The redshifts of these objects are in the range 0.04 < z < 0.1. For simplicity, we will drop the SDSS-C4-DR3 prefix from all target BCG names, and simply use the last four digits that are unique to each cluster. In total, our initial sample size is seven BCGs with companions and four BCGs without companions within ∼10''. One target, BCG 1067, will be omitted from analysis because it had excessively noisy observations and we were unable to obtain results on that BCG above our signal-to-noise cut of 10.

The first set of BCGs were observed with VIMOS (Le Fèvre et al. 2003) on the Very Large Telescope (VLT) from April to August of 2008 (Prog. ID 381.B-0728) and the second set were observed, also with VIMOS, from April to July of 2011 (Prog. ID 087.B-0366). VIMOS was used in IFU mode with the high-resolution blue grism and a spatial sampling of 067 pixel⁻¹ (BCGs 1153 and 1067 were observed with a spatial sampling of 033 pixel⁻¹). This gives a field-of-view (FOV) of 27'' × 27'' (13'' × 13'' for 1153 and 1067). The VIMOS HR blue grism (pre 2012 March 15 version) has a spectral range of 4150–6200 Å and a spectral resolution of 0.51 Å pixel⁻¹. Observations were made during dark time, with an average seeing of 09. Average seeing for individual galaxies can be seen in Table 1.

2.1.1. IFU Data Reduction

2.1.2. Stellar Kinematics

The stellar kinematics (velocity, V, and line-of-sight velocity dispersion, σ) of the galaxy are computed from the binned spectra using a penalized pixel fitting scheme (pPXF; Cappellari & Emsellem 2004) and the Medium-resolution Isaac Newton Telescope Library of Empirical Spectra (MILES; Sánchez-Blázquez et al. 2006) evolutionary stellar population templates. This method determines the stellar kinematics by fitting templates to the absorption line features in the spectra (Figure 1). We choose 10 templates of stars in the range of G6 to M1 (luminosity classes III, IV, and V) from the MILES library and convolve them using the quadratic difference between the resolution of the library and the observations. The pPXF fits are performed in the region around the observed G-band and calcium H and K absorption features (4200–4900 Å). The MILES templates cover a similar wavelength range to VIMOS and have a similar spectral sampling (FWHM = 1.8 Å).

Zoom In Zoom Out Reset image size

Velocity dispersion for a galaxy as a whole is found by co-adding all fibers within the effective radius of a galaxy into one bin, and performing a pPXF fit on the binned spectrum as in Cappellari et al. (2006). As an alternative method for calculating the velocity dispersion of the whole galaxy, we also tried calculating the luminosity weighted mean of each individual bin's velocity dispersion. These results were often not in agreement with the effective spectrum method. However, by integrating (preliminary) inclination information as described in Equation (29) of Cappellari et al. (2013), the two velocity dispersion measuring methods appear to be in agreement. We choose to use the effective spectrum method instead of the weighted mean method because of the higher S/N afforded by binning many spaxels together. The effective spectrum method also simulates an observation with a single aperture of radius R_e.

We perform 100 Monte Carlo realizations on each observation to estimate the uncertainty in our results. This process involves adding random noise on the order of the background noise to each binned spectrum, and performing the pPXF fit again on the noisy data. We repeat this process 100 times, then take a weighted mean of all the fits with the random noise added. The standard deviation of all the fits is our uncertainty for our velocity and velocity dispersion results. Typical uncertainties of both values are on the order of 20 km s⁻¹. Errors on M_dyn and λ_R are propagated from these errors using standard Taylor series error propagation techniques.

The IDL code described above to reduce our data is publicly available.⁹

Photometric observations are taken from SDSS Data Release 3. For our analysis we use only the r-band images of each galaxy. Each galaxy is modeled twice, once using a combination of SExtractor (Bertin & Arnouts 1996) and GALFIT (Peng et al. 2002) with a single component de Vaucouleurs profile, and a second time using PyMorph (Vikram et al. 2010) to obtain the Gini and M₂₀ coefficients. The de Vaucouleurs profile is used to find the effective radius of each galaxy.

Effective radius measurements are found by first feeding the SDSS r-band image into SExtractor to obtain a catalog of all sources identified in the image. We use the objects catalog to determine the centers of the galaxies we wish to model in GALFIT. We also invert the segmentation map produced by SExtractor (non-zero pixels are set to zero, and pixels that were originally zero are set to one) to generate the bad pixel mask that will be used with GALFIT. The default values and keyword searches were used as inputs to SExtractor except the following: MAG_ZEROPOINT = 25.61, BACK_SIZE = 256, and PIXEL_SCALE = 0.396. The original SDSS r-band images are then fed into GALFIT using the position, magnitude, effective radius, axis ratio, and position angle found by SExtractor as initial assumptions. We model each galaxy using a single component de Vaucouleurs profile in GALFIT. We generate our point-spread function by using a 2–4 Gaussian fit in GALFIT to model a point source near each galaxy. Residual maps are inspected by eye to determine if the fit appears to be reasonable, and if not, initial parameters are adjusted until a reasonable fit is found.

PyMorph combines SExtractor, GALFIT, and other added analysis routines in a python wrapper in order to automate the photometric analysis process. It is used to determine the merger properties of our sample of galaxies. We use it to make two component models for each galaxy. We then perform a visual inspection of the residual image to check for the goodness of the fit.

PyMorph also reports the concentration, asymmetry, clumpiness, Gini coefficient, and M₂₀ for each galaxy that is modeled. We use the relation between the Gini coefficient and M₂₀ to search for recent mergers. The Gini coefficient measures uniformity of the distribution of a galaxys light (Abraham et al. 2003) and M₂₀ measures the distribution of the brightest 20% of pixels. The relation between these two properties tells us if there are irregularities in the galaxy's light distribution, signaling some type of violent interaction. We can rule out the possibility that anomalies in these parameters are due to projection effects because we can measure the redshift of each galaxy to confirm that they are indeed located near each other in physical space.

We begin our spectroscopic analysis with the velocity maps seen in Figures 2 and 3. Our first goal is to search for visual signs of galaxy rotation. Figure 2 shows the velocity maps of the galaxies without companion measurements. This sample contains three of our control galaxies, BCGs 1042, 1050, and 1153, which were selected to not have any companions within 10''. The S/N on the companion galaxies of BCGs 2001, 2039, and 2086 was below 10 after binning, meaning we were unable to achieve accurate measurements of the companion's kinematic properties.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

By visual inspection of Figure 2, we see that five of the BCGs (1042, 1050, 2001, 2039, and 2086) appear to have little to no rotation. They have a uniform velocity across the entire BCG and the velocities measured in each spaxel of the BCG are very near 0 km s⁻¹. These observations are consistent with a dispersion-supported early-type galaxy. Two of the BCGs (1153 and 1261) in Figure 2 appear to have velocity gradients across their surface, with BCG 1261 being the most obvious example.

Figure 3 shows velocity maps of the BCGs that have bright enough companions to measure the companion's kinematic properties. In each collapsed IFU FOV, the BCG is labeled with the letter A, and then brightest companion is labeled B, and so on. BCG 1048 appears to have one side very red-shifted and the opposite side very blue-shifted, suggesting rotation. Note also that all the companions in Figure 3 appear to exhibit rotation.

The velocity dispersion map (Figure 4) shows very high dispersions in BCG 1048, suggesting a possible ongoing interaction. We also see in Figure 4 a peak in the dispersion near the center of BCG 1261. This peak could be the result of seeing both the positive and negative velocities on either side of the axis of rotation in the same fiber, enhancing the observed dispersion.

Zoom In Zoom Out Reset image size

In order to quantify the rotation seen in some of the BCGs mentioned above, we use the λ_R parameter developed by the SAURON team (Emsellem et al. 2007). This parameter acts as a proxy for angular momentum, and is defined as

$$\begin{equation} \lambda _{R} \equiv \frac{\langle R | V | \rangle }{\langle R \sqrt{V^2+\sigma ^2} \rangle }, \end{equation} \tag{ 1 }$$

where R is the distance of the spaxel to the galaxy center, V is the velocity of the spaxel, and σ is the velocity dispersion. The numerator acts as a surrogate for the angular momentum L, and the denominator acts as a mass normalization. The brackets in the numerator and the denominator denote a luminosity weighted average.

The λ_R profile for each BCG and companion galaxy, plotted along with the SAURON results, can be seen in Figure 5. A higher λ_R value indicates higher angular momentum. As expected, angular momentum tends to increase with radius, especially in galaxies classified as FRs. Galaxies that fit into the FR category appear to have a convex profile as λ_R increases with radius, whereas most SR galaxies have a concave profile. Most of our BCGs appear to have profiles consistent with the SR category.

Zoom In Zoom Out Reset image size

λ_Re is the measured angular momentum at the effective radius. In cases where S/N of our measurements drops below 10 before we reach 1R_e, we assume that value of λ_Re is the furthest measured λ_R in that galaxy. This likely gives a minimum value for λ_Re because λ_Re tends to increase with radius within 1R_e as observed in the SAURON results in Figure 5.

The SAURON survey was followed by the ATLAS^3D survey, which refined the definition of a fast rotating galaxy to take into account the ellipticity of the galaxy, which the original SAURON definition does not. According to the ATLAS^3D definition, the threshold for a FR is

$$\begin{equation} \lambda _{Re} > (0.31 \pm 0.01) \times \sqrt{\epsilon _e} \end{equation} \tag{ 2 }$$

(Emsellem et al. 2011), where _e is the ellipticity at the effective radius (R_e). is measured by the IDL routine find_galaxy.pro written by Michele Cappellari and available as part of the mge_fit_sectors package.¹⁰

The λ_Re versus _e plot in Figure 6 shows our quantitative determination of galaxy rotation. Using the ATLAS^3D definition, all galaxies above the blue line are classified as FRs. We find that three BCGs and all four companions are FRs. Seven of our BCGs are within one standard deviation of the dividing line, causing us to doubt their classification. For them we consider their λ_R profiles to see if they are concave or convex in order to make our determinations. By the curve criteria, we consider six of these ambiguous galaxies to be SRs. Giving us a final total of three fast rotating BCGs (30%) and four fast rotating companion galaxies (100%).

Zoom In Zoom Out Reset image size

Next we examine the relation between a galaxy's dynamical mass and its λ_Re measurement. We determine the dynamical mass of the galaxy by using the equation provided by Cappellari et al. (2006):

$$\begin{equation} M_{{\rm dyn}} = \frac{5R_e\sigma _e^2}{G}. \end{equation} \tag{ 3 }$$

R_e is the effective radius, σ_e is the aperture corrected velocity dispersion at the effective radius, and G is the gravitational constant. The factor of five is a parameter that scales between the virial and Schwarzschild M/L estimates. According to theory it should be 5.953, however, those calculations assume perfect one-component isotropic spherical systems, whereas 5 is found to be a good fit for observed galaxies (Cappellari et al. 2006). In order to compensate for the fact that most of our velocity dispersions reported are measured at R < R_e, we have applied the aperture correction from Equation (1) of Cappellari et al. (2006) to our velocity dispersion results.

Dynamical mass results for each galaxy are listed in Table 2. In Figure 7 we have plotted λ_Re as a function of dynamical mass. We also plot dynamical masses from the SAURON survey (Emsellem et al. 2007) and derived dynamical masses from the reported ATLAS^3D velocity dispersions (Cappellari et al. 2013). Figure 7 shows that the majority of the galaxies reported in this study fall on the high mass end of the ATLAS^3D survey sample. Eight galaxies in our sample have dynamical masses above 10^11.5 M_☉. In the SAURON and ATLAS^3D results there appears to be a ceiling at approximately M_dyn = 10^11.5 M_☉ at which point galaxies were much less likely to be classified as FRs (3/13 (23%) in Emsellem et al. 2011). In agreement with the findings of the ATLAS^3D survey, we find that only 25% of galaxies in our sample above 10^11.5 M_☉ are classified as fast rotating.

Zoom In Zoom Out Reset image size

Table 2. Kinematic Properties of BCGs and Companions

Galaxy	Re ('')	Redshift	σe	log(M) (log(M☉))		λRe	FR/SR
	± 0.01	± 0.0001	(km s−1)	± 0.01	± 0.01
1027 BCG*	6.98	0.0900	227 ± 7	11.79	0.08	0.12 ± 0.08	SR
1027 Comp*	4.39	0.0909	190 ± 6	11.17	0.08	0.24 ± 0.06	FR
1042 BCG	7.22	0.0947	227 ± 7	11.83	0.05	0.08 ± 0.06	SR
1048 (A4)a BCG*	5.17 (3.40)a	0.0774	319 ± 7	11.59	0.33	0.46 ± 0.06	FR
1048 Comp	1.08	0.0801	167 ± 3	10.51	0.20	0.34 ± 0.05	FR
1048 Comp	1.24	0.0746	124 ± 3	10.54	0.12	0.21 ± 0.06	FR
1050 BCG	8.43	0.0722	291 ± 11	11.78	0.14	0.07 ± 0.07	SR
1066 BCG*	5.07	0.0838	186 ± 9	11.62	0.35	0.12 ± 0.05	SR
1066 Comp*	11.25	0.0836	176 ± 13	11.55	0.37	0.60 ± 0.04	FR
1153 BCG	2.39	0.0591	226 ± 7	11.14	0.28	0.21 ± 0.08	FR
1261 BCG	5.76	0.0371	271 ± 1	11.32	0.29	0.35 ± 0.05	FR
2001 BCG	5.84	0.0415	200 ± 4	11.38	0.09	0.13 ± 0.07	SR
2039 BCG	8.82	0.0829	248 ± 6	11.86	0.06	0.10 ± 0.07	SR
2086 BCG*	4.83	0.0839	203 ± 7	11.60	0.08	0.09 ± 0.06	SR

Notes. Kinematic results from a combination of IDL routines. R_e is derived from photometric results, but listed here as it is relevant to many values calculated in this table. (*) Galaxies with an asterisk next to their name are gravitationally bound to their neighboring galaxy. ^aThe effective radius reported in parenthesis for BCG 1048 is for component A4 only, and not the entire BCG.

Download table as:  ASCIITypeset image

To perform our analysis of the spectroscopic data, we have to also examine the photometric data in order to determine the effective radius of each galaxy. The single component de Vaucouleurs models produced by GALFITS and residual images can be seen in Figure 8. In Figure 8 one can also see the tidal tails (BCG 1027) and multiple cores (BCG 1048) which motivated us to search for a quantifiable way to measure recent and ongoing mergers. We find that a single de Vaucouleurs profile is unable to properly model BCG 1048, and the residual image shows multiple hidden cores. To properly model BCG 1048, we instead use four spatially separated de Vaucouleurs profiles. We report both the single de Vaucouleurs results as well as the results for component A4, brightest of the four profiles, in parenthesis in Tables 2 and 3.

Zoom In Zoom Out Reset image size

3.4.1. G − M₂₀ Analysis

The G − M₂₀ analysis searches for irregularities in the distribution of light which correlate to morphological signatures of merging. Our G − M₂₀ analysis is performed using PyMorph. For a galaxy to be classified as a merger candidate it must have

$$\begin{equation} G > -0.14M_{20}+0.33 \end{equation} \tag{ 4 }$$

(Lotz et al. 2008). As can be seen in Table 3, BCG 1048 (FR) is confirmed to be a merger case by the G − M₂₀ selection criteria, as is BCG 2086 (SR). The other merging candidates are companions to BCGs 1027 (SR), 1066 (SR), and 2086 (SR).

BCG 1261, our fastest rotating galaxy, shows no signs of merging in the G − M₂₀ classification. According to Lotz et al. (2011), for our gas-poor galaxies, this suggests no recent mergers within the last 0.2 Gyr by the G − M₂₀ classification. That would suggest that any mergers with this z = 0.04 galaxy would have had to happen before z = 0.06.

Therefore, 40% (4/10) of the systems in our sample show morphological signs of current or very recent mergers. Based on the original unbiased sample of 625 BCGs from von der Linden et al. (2007), we find a lower limit of 0.64% ± 0.32% of BCGs are currently undergoing or have undergone a merger within the last 0.2 Gyr. Although this subsample is biased toward merging galaxies and not representative of the entire BCG sample, we consider it significant that it is consistent with simulations of BCGs from De Lucia & Blaizot (2007) that showed a 1.47% merger rate within the same 0.2 Gyr time period (assuming a constant merger rate from z = 0.5 to z = 0) as well as observational results from Liu et al. (2009) that showed a 0.63% BCG major merger rate within 0.2 Gyr (assuming a constant merger rate from z = 0.12 to z = 0.03).

By utilizing the information from both the photometric and the kinematic results, we are able to determine whether the companion galaxies are likely to be bound to the BCG. The criteria for being gravitationally bound is

$$\begin{equation} V_rR_p \le 2{GM} \sin ^2\alpha \ \cos \alpha \end{equation} \tag{ 5 }$$

(Beers et al. 1982). V_r is the radial velocity offset between the two galaxies, R_p is the projected radial separation between the two galaxies, M is the dynamical mass, G is the gravitational constant, and α is the angular separation between the two galaxies. By integrating over all possible values of α, we can find the probability that a companion galaxy will be gravitationally bound to the BCG.

The same analysis was performed by Brough et al. (2011) for BCGs 1027, 1066, and 2086, in which they found that systems 1027 and 1066 were likely to be bound, but 2086 was not. In performing our analysis, we find that indeed the companion to BCG 1027 has a 65% chance of being bound, the companion to BCG 1066 has a 94% chance of being bound, and the companion to BCG 2086 has a 56% chance of being bound. We also find in addition that the two outer companions to BCG 1048 (1048 B and 1048 C) never cross the threshold outlined in Equation (5) and are therefore unlikely to be bound to BCG 1048 A. As seen in Figure 8, BCG 1048 A is comprised of four components, the most massive being 1048 A2, A3, and A4 which together make up the galaxy which we call BCG 1048. We find that component 1048 A3 has an 80% chance of being bound to 1048 A4.

From the combination of kinematic and photometric results, it appears that BCGs have diverse recent merging histories. This is unexpected considering their similarities in luminosities and stellar population. BCGs 1261, 1048, and 1153 are our highest λ_Re BCGs and it appears that their high λ_Re values come from different sources. Simulations show that very gas-rich major mergers (Cox et al. 2006; Robertson et al. 2006) and certain gas-poor major mergers (Bois et al. 2011) produce fast rotating galaxies. Since we do not see any evidence for gas emission lines in any of the galaxies in our sample, we assume that they are still gas-poor. Below we will discuss specific examples of the various combinations of merging and angular momentum, demonstrating that there is little correlation between the two observables (Figure 9).

Zoom In Zoom Out Reset image size

4.1.1. Fast Rotators with No Mergers

4.1.2. Fast Rotators with Mergers

4.1.3. Slow Rotators without Mergers

4.1.4. Slow Rotators with Mergers

Companions of BCGs 1066, 1027, and 1048 show clear signs of rotation, both visually and in their λ_R results. Every BCG companion that we have sufficient data to measure λ_R is a fast rotating galaxy. Although we have a limited number of spaxels to measure the rotation in BCG 1066, the analysis presented in Brough et al. (2011) provides more spaxels, showing that both companions of 1066 and 1027 are FRs. Our data indicates that all elliptical galaxies near BCGs are FRs, however we have an admittedly small sample size of only four companion galaxies to make that determination.

The presence of fast rotating BCGs in our sample was quite unexpected considering our initial assumption that BCGs should be mostly dispersion supported. Although, when comparing our measurements of λ_Re and M_dyn in BCGs to the SAURON and ATLAS^3D samples of early-type galaxies (Figure 7), we find that our sample is generally consistent with their results. There appears to be a limit around M_dyn = 10^11.5 M_☉ above which there are virtually no galaxies with a high angular momentum, independent of whether or not they are BCGs. The two galaxies that we do find slightly above this limit are both within 1σ, and both currently in the process of merging, causing some doubt that they are actually as massive as measured.

In observing these very massive galaxies, in extreme environments, at distances further than what had been measured before, our results suggest that λ_Re would be influenced more by a galaxy's dynamical mass than its location within a cluster, and that this is consistent across samples of galaxies that were selected using differing criteria. The SAURON and ATLAS^3D surveys focused on early-type galaxies in general and measured only one BCG, whereas we examined a sample of 10 BCGs at higher redshift. Larger galaxies likely formed as a result of more mergers than smaller galaxies, therefore they have a higher likelihood of losing their initial angular momentum. Peng et al. (2010) have also shown that above M_* = 10¹¹ M_☉ there is a change in the ratio of galaxy mass assembled via post-quenched galaxy mergers, also suggesting that these dry merging events could be responsible for removing the angular momentum.

Conversely, Martizzi et al. (2012) have used simulations to show that AGN are a possible mechanism for removing angular momentum from large galaxies. Having shown in our sample examples of dry mergers that exhibit a high λ_Re measurement, it is possible that a different process, such as AGN, are responsible for turning fast rotating galaxies into SRs. It would be interesting and relatively straightforward to test for the presence of the M_dyn = 10^11.5 M_☉ limit for FRs observed within these samples using such simulations, and to see if the presence of AGN in simulations has any effect on the reproducibility of this limit.

In a further study using higher resolution photometry, we would like to measure which galaxies are cuspy and which are cored as in Lauer (2012). It is possible that AGN are responsible for both building the core, as well as removing angular momentum at the same time. Initial results suggest that our sample of galaxies follow the same relationship seen in the ATLAS^3D sample, in that all cored galaxies are SRs, but not all SRs are cored. However higher resolution photometry than the SDSS imaging is needed in order to properly make the cuspy/cored determination.

Also in Lauer (2012) they discuss the possibility that the division between FRs and SRs should be set at a constant λ_Re/2 = 0.25. Using this alternative selection criteria, we would find two fast rotating BCGs (20%), and two fast rotating companion galaxies (50%). Under this higher threshold, the main message of our results would change little. The majority of the BCGs that we studied would be slow rotating, however we would still find fast rotating BCGs. Similarly, 25% of the galaxies we studied above M_dyn = 10^11.5 M_☉ would continue to be FRs.

Another possible issue to consider is that in some cases, such as BCG 1048 and the companion to BCG 1066, the galaxies contain multiple cores from ongoing mergers, which could cause the λ_Re measurements to be artificially inflated. This would also artificially increase our measurements of M_dyn. It is also possible that our measurements for λ_Re for these galaxies will shift as they become dynamically relaxed post-merger. An examination of λ_Re values pre-merger, peri-merger, and post-merger in simulations may shed light on how much of an effect dry mergers would have on λ_Re results. In our analysis, we consider these merging systems, such as BCG 1048 and companion 1066 to be one galaxy because they are currently in the process of merging.

Another factor to consider would be the angular momentum of the cluster as a whole. If the angular momentum of a BCG at the kinematic center of a cluster coincides with the average angular momentum of the entire cluster, then infalling galaxies merging with the BCG would likely contribute to the angular momentum of the BCG instead of subtracting from it. It would be interesting to study the angular momentum of the whole cluster of our best example of a fast rotating BCGs, such as 1261, to see if the angular momentum of the cluster aligns with the angular momentum of the BCG. Such measurements would require new observations with a much wider FOV. An instrument such as SAMI (Croom et al. 2012) with deployable fiber bundles would allow us to collect IFU observations of spatially separated objects across a much wider FOV, e.g., to observations beyond 1R_e for many of the galaxies in our sample.

We have observed that BCGs are surprisingly diverse in their photometric and kinematic properties. We intend to continue our study of these BCGs by conducting an analysis of their stellar populations in order to further understand their merger history.