A Spitzer-IRS view of early-type galaxies with
cuspy/core nuclei and with fast/slow rotation

A relatively large fraction of ETGs at high-redshift show clear evidence of interaction and/or merger morphologies and active star formation (e.g. Treu et al., 2005) supporting the model view these galaxies are produced by a halo merger process (see e.g. Mihos, 2004; Cox et al., 2008; Khochfar et al., 2011; De Lucia et al., 2011, and reference therein). A further element for high-redshift formation scenarios comes from their measured [$\alpha$/Fe] ratios, encoding information about the time-scale of star formation. In massive ETGs this ratio has super-solar values, suggesting that they formed on relatively short time-scales (see e.g Chiosi et al., 1998; Granato et al., 2004; Thomas et al., 2005; Annibali et al., 2007; Clemens et al., 2009). Annibali et al. (2007) estimated that a fair upper limit to the recent rejuvenation episodes is $\sim$25% of the total galaxy mass but that they are typically much less intense than that (see e.g. the Spitzer-IRS study of NGC 4435 by Panuzzo et al., 2007). However, rejuvenation signatures in ETGs are often detected, not only in the galaxy nucleus, but also in the disk, rings and even in galaxy outskirts, as clearly shown by GALEX (e.g Rampazzo et al., 2007; Marino et al., 2009; Salim & Rich, 2010; Thilker et al., 2010; Marino et al., 2011), so that the different phases of galaxy assembly/evolution, and their link to morphological and kinematical signatures, are vivaciously debated.

The merger process may involve either relatively few (major) or multiple (minor) events during the galaxy assembly. Furthermore, it may or may not include dissipation (and star formation), two possibilities often called “wet” and “dry” mergers, respectively (see e.g. van Dokkum, 2005). Other mechanisms, however, like conversion of late-type galaxies into ETGs by environmental effects, like strangulation, ram-pressure etc. (e.g. Boselli & Gavazzi, 2006), and by energy feedback from supernovae may also be important (e.g. Kormendy et al., 2009).

Two observable quantities are thought to distinguish ETGs produced by “wet” and “dry” mergers. The first, mainly fruit of high resolution observations with the Hubble Space Telescope and of high precision photometric analyses, is the presence of either a cusp or a core in the inner galaxy luminosity profile (Lauer et al., 1991, 1992; Côté et al., 2006; Turner et al., 2012). In contrast to cuspy profiles, the surface brightness in core profiles becomes shallower as $r\rightarrow 0$. The same concept is considered by Kormendy et al. (2009) who divide ETGs into cuspy–core and core–less, depending on whether the luminosity profile misses light or has an extra-light component with respect to the extrapolation of the Sersic’s law at small radii. Kormendy et al. (2009) suggest that cuspy–core nuclei have been scoured by binary black holes (BHs) during (the last) dissipationless, “dry”, major merger. In contrast, core–less nuclei originate from “wet” mergers. Analogously, Côté et al. (2006) and Turner et al. (2012) found an extra stellar nucleus in the profile decomposition of ETGs in their Virgo (ACSVCS) and Fornax (ACSFCS) surveys in addition to simple Sersic profiles, in particular in low-luminosity(/mass) ETGs. They proposed that the most important mechanism for the assembly of a stellar nucleus is the infall of star clusters through dynamical friction, while for more luminous(/massive) galaxies a “wet scenario” (gas accretion by mergers/accretions and tidal torques) dominates.

The second observable quantity is the kinematical class. The class is defined by a parameter describing the specific baryonic angular momentum defined as follows, $\lambda_{r}$=$\langle r|V|\rangle/\langle r\sqrt{V^{2}+\sigma^{2}}\rangle$, where $r$ is the galacto-centric distance, $V$ and $\sigma$ are luminosity weighted averages of the rotation velocity and velocity dispersion over a two-dimensional kinematical field. The measure refers to the inner part of the galaxy, typically of the order or less than 1 effective radius, $r_{e}$, i.e. significantly larger than the regions where cusps and cores are detected. $\lambda_{r}$ divides ETGs into the two classes of fast (FR) and slow (SR) rotators (Emsellem et al., 2011, and reference therein). FR are by far the majority of ETGs (86$\pm$2% in the ATLAS^3D survey). SR represent massive ETGs that might have suffered from significant merging without being able to rebuild a fast rotating component. Khochfar et al. (2011) find that the underlying physical reason for the different growth histories is the slowing down, and ultimately complete shut-down, of gas cooling in massive, SR galaxies. On average, the last gas-rich major merger interaction in SR happens at $z>1.5$, followed by a series of minor mergers which build-up the outer layers of the remnant, i.e. do not feed the inner part of the galaxy.

FRs in the models of Khochfar et al. (2011) have different formation paths. The majority (78%) have bulge-to-total stellar mass ratios (B/T) larger than 0.5 and manage to grow stellar discs due to continued gas cooling as a results of frequent minor mergers. The remaining 22% live in high–density environments and consist of low B/T galaxies with gas fractions below 15%, that have exhausted their cold gas reservoir and have no hot halo from which gas can cool. Summarizing, a dissipative merging and/or a gas accretion episode from interacting companions, could be the way for the galaxy to rebuild a fast-rotating disk-like component. SR and FR basically correspond to the paradigms of “dry” vs. “wet” accretions/mergers respectively.

Recently, Lauer (2012) attempted to unite the structural and kinematical views, claiming that they are the two aspects of the same process. Using the specific angular momentum $\lambda_{r_{e}/2}$, computed from the 2D kinematics within half the effective radius by Emsellem et al. (2011), Lauer (2012) showed that core galaxies have rotation amplitudes $\lambda_{r_{e}/2}\leq 0.25$ while all galaxies with $\lambda_{r_{e}/2}>0.25$ and ellipticity $\epsilon_{r_{e}/2}>0.2$ lack cores. Some FR have a core profile but they argue that both figure of rotation and the central structure of ETGs should be used together to separate systems that appear to have formed from “wet” and “dry” mergers. Krajnović et al. (2013b) show, however, that there is a genuine population of FR with cores. They suggest that the cores of both FR and SR are made of old stars and are found in galaxies typically lacking molecular and atomic gas, with few exceptions.

For the sake of simplicity throughout the paper, we will call core ETGs those galaxies for which the luminosity profile shallows out as $r\rightarrow 0$ (i.e. cuspy-core in Kormendy et al. (2009), core in Lauer (2012), non-nucleated in Côté et al. (2006); Turner et al. (2012)). We will refer to cuspy ETGs as those which present an extra central luminosity component (i.e. core-less in Kormendy et al. (2009), power-law + intermediate in Lauer (2012), nucleated in Côté et al. (2006); Turner et al. (2012)) with respect to a fit of a Sersic model. Depending on the accurate surface brightness profile decomposition performed by the above authors, rarely the cuspy versus core classification given by different authors for the same ETG is discrepant.

This note aims to contribute to the debate on the origin of core/cuspy and FR/SR ETGs, and the connection to the “wet” vs. “dry” merger hypotheses, using mid-infrared (MIR) spectra of well-studied ETGs. The paper is organized as follows. In § 2 we briefly describe how Spitzer-IRS spectra trace the recent few Gyr evolution in ETGs. We present the MIR vs. the cuspy/core nuclear properties of ETGs (Kormendy et al., 2009; Côté et al., 2006; Turner et al., 2012; Lauer, 2012) in § 3, and vs. the FR/SR kinematical classes (Emsellem et al., 2011) in § 4. Conclusions are presented in § 5.

Using Spitzer-IRS, Rampazzo et al. (2013) have recently produced an atlas of low resolution MIR spectra, of 91 nearby (D$\leq$72 Mpc) ETGs in the Revised Shapley-Ames Catalog (RSA). Spectra are extracted within a common aperture (3.6″$\times$18″) providing an integral view of the galaxy’s inner region. On average, the area covered by the rectangular IRS aperture is about 2-3 times the area of a circle of radius $r_{e}/8$. MIR spectra then cover a region where the cuspy/core profiles and/or the departures from the Sersic’s profile are measured, as well as a significant part of the region where FR vs. SR are separated, using e.g. $\lambda_{r_{e}/2}$. In addition, “dry” vs. “wet” mergers maximize their possible signatures in the galaxy nucleus, as shown by the study of line-strength indices (e.g. Annibali et al., 2007, and references therein), by counter-rotating components (e.g. Emsellem et al., 2011) and morphological peculiarities such as irregular/chaotic dust-lanes (e.g. Côté et al., 2006; Turner et al., 2012).

For each galaxy, the atlas provides the fully reduced and calibrated spectrum, the intensity of nebular and molecular emission lines and Polycyclic Aromatic Hydrocarbons (PAHs), after a template spectrum of a passively evolving ETG has been subtracted. Spectra are classified into five mid-infrared classes, ranging from AGN (class-4) and star forming nuclei (class-3), transition class-2 (with PAHs) and class-1 (no-PAHs) to passively evolving nuclei (class-0) (Bressan et al., 2006). Panuzzo et al. (2011) suggest that each of the five MIR classes is a snapshot of the evolution of ETG nuclei during an accretion episode, starting from, and ending with, a class-0 spectrum (see their Figure 11). Vega et al. (2010) suggest that anomalous PAHs in class-2 spectra are produced by carbon stars which are present in stellar populations with ages in the range of a few Gyr (Nanni et al., 2013, see their Table 1), depending on the metallicity, corresponding, on average, to a redshift coverage $z\lesssim 0.2$.

MIR spectral classes show a dichotomy between the Es and S0s. i.e. the classical morphological classification provided by RSA. Rampazzo et al. (2013) found that 46${}^{+11}_{-10}$% of Es and 20${}^{+11}_{-7}$% of S0s have a spectrum of class-0, i.e. Es are significantly more passive than S0s. A small, similar ($\sim$9${}^{+4}_{-3}$%) fraction of Es and S0s have PAH ratios typical of star forming galaxies. PAHs are detected in 47${}^{+8}_{-7}$% of ETGs and with a statistically similar fraction in Es and S0s. Basically half of ETGs (41${}^{+8}_{-7}$%) in the sample show kinematical and/or morphological scars of recent accretion episodes.

In this note we investigate ETGs subdivided into the new, well defined photometric (cuspy vs. core) and kinematical (FR/SR) classes. To this end, we cross-correlate the MIR RSA sample described above with the Kormendy et al. (2009), Lauer (2012), Côté et al. (2006), Turner et al. (2012) and ATLAS^3D samples (see e.g. Emsellem et al., 2011, and references therein) to investigate the recent nuclear star formation history. 14 ETGs in Virgo, in particular the brighter ones, are in common with Kormendy et al. (2009) who classified them as either cuspy-core or core-less (i.e. core and cuspy in the present paper). 26 are in common with the ACSVCS+ACSFCS surveys in which galaxies are separated into core+Sersic and Sersic type nuclei (Côté et al., 2006; Turner et al., 2012). 41 out of 91 ETGs in the RSA MIR atlas are in common with the ATLAS^3D survey sample. For these galaxies Emsellem et al. (2011) provide a classification into FR or SR. The inner luminosity profiles of these ETGs have been classified into cuspy (power-law and intermediate) or core by Lauer (2012) (23 out of 41) and by Krajnović et al. (2013b) for a total of 44 objects.

With the aim of investigating the MIR properties of FRs vs. SRs in the $r_{e}/2$ region, we cross-correlate the samples adding a measure of the specific angular momentum to our sample with a MIR nuclear classification. 41 out of 91 ETGs in the Rampazzo et al. (2013, Table 5) sample of MIR spectra have a measure of the specific angular momentum. Table 2 provides the list of ETGs with a FR/SR classification within $r_{e}/2$ (column 8 from Emsellem et al. (2011, ATLAS^3D: Table B1)) and their MIR class (column 10). The list includes 29 FRs and 12 SRs. The Fornax ETGs lack the kinematical classification in the $r_{e}/2$ region.

In Figure 3 the specific angular momentum at $r_{e}/2$, $\lambda_{r_{e}/2}$, is plotted against the corresponding average ellipticity, $\epsilon_{r_{e}/2}$, (Table 2, columns 5 and 6). According to Lauer (2012) the dashed line in the figure separates the core and cuspy ETGs into different rotation classes, leaving the core set contaminated only by face-on cuspy galaxies. The figure shows the 23 galaxies of Lauer (2012), among the 41 ETGs in Emsellem et al. (2011) (large squares for SR, large circles for FR). Figure 3 plainly shows that different MIR classes are found associated with both FR and SR, as well as in core and cuspy ETGs.

25${}^{+24}_{-14}$% of SRs and 52${}^{+17}_{-13}$% of FRs have a class-0 spectrum i.e. are passively evolving. On the other hand, of those MIR classes showing emission lines (1, 2, 3 and 4), i.e. the complement of passive ETGs, 75${}^{+25}_{-24}$% of SRs and 48${}^{+17}_{-12}$% of FRs are gas rich in their nucleus. In the whole sample of Rampazzo et al. (2013), 64${}^{+12}_{-6}$% of the nuclei show emission lines, although with different intensity. This results are in agreement with optical studies, in which, depending on the sample the ionized gas is detected in $\approx$50-90% of ETGs (Phillips et al., 1986; Macchetto et al., 1996; Sarzi et al., 2006, 2010; Yan et al., 2006; Serra et al., 2008; Annibali et al., 2010). This suggests that wet-mergers/accretions play a role in the recent few Gyrs history of both SR and FR.

50${}^{+30}_{-20}$% of SRs and 38${}^{+15}_{-11}$% of FRs show PAH features (MIR classes 2, 3, 4), indicating that in about half of both SR and FR a star formation episode has recently occurred (Kaneda et al., 2008; Panuzzo et al., 2011; Rampazzo et al., 2013). Galaxies with normal PAH ratios, class-3 spectra, are a minority: only 10${}^{+8}_{-5}$% of FR show star forming spectra and 8${}^{+19}_{-2}$% SR. All MIR class-4 spectra, 17${}^{+22}_{-11}$% of SR and 3${}^{+8}_{-1}$% of FRs, show PAHs. The optical study of 65 ETGs of Annibali et al. (2010) provided indications that the AGN phenomenon is associated with star formation.

This note investigates whether the photometric segregation between cuspy and core nuclei and/or the dynamical segregation between fast and slow rotators can be attributed to formation via “wet” or “dry” mergers.

We explore the question by comparing the MIR spectral characteristics with their cuspy/core morphology (Kormendy et al., 2009; Lauer, 2012; Côté et al., 2006; Turner et al., 2012; Krajnović et al., 2013a) and FR/SR characterization (ATLAS^3D). We use Spitzer-IRS spectra and MIR classes discussed in Panuzzo et al. (2011) and Rampazzo et al. (2013). These spectra are sensitive to the recent few Gyr ($z\lesssim 0.2$) evolution of the ETGs. We find the following:

• •

  With the exception of NGC 4365, which is passively evolving, MIR spectra of all the bright core ETGs in the Kormendy et al. (2009) sample show nebular emission lines, and PAH features are detected in 5 out of 9 objects. These types of nuclei should have recently accreted gas-rich material. If such objects formed via “dry” mergers, the process was completed before $z\sim 0.2$ and “wet” accretions have happened since. AGN feedback does not prevent a late star formation episode in the bright Es NGC 4374, NGC 4636 and NGC 4552. The few (6) faint, cuspy ETGs in the Kormendy et al. (2009) sample all show passively evolving spectra, irrespective of their magnitude.

• •

  MIR spectra of the total cuspy/core sample (44 ETGs), confirm that ETGs fainter than M${}_{K_{s}}$=$-24$ mag, mostly cuspy, are predominantly passively evolving. This fact is particularly significant in Virgo and Fornax where 82${}^{+18}_{-10}$% are cuspy, the majority also FRs. Kormendy et al. (2009) noticed that cuspy ETGs have disky (positive $a_{4}/a$ Fourier coefficient) isophotes in the nuclear region, a structure that suggests some form of dissipation during the formation. The passive MIR spectra suggest that either this infall was sterile, i.e. without star formation, or happened at $z\gtrsim 0.2$, so as to leave no trace in the present MIR spectra. Khochfar et al. (2011) models suggest that a fraction of FR lying in high–density environments have a residual gas fraction below 15%, i.e. they have exhausted their cold gas reservoir and have no hot halo from which gas can cool. Counterparts in low density environments (Figure 2 right panel) show the tendency to be more gas rich and hence more active.

• •

  A significant fraction of both FR (38${}^{+18}_{-11}$%) and SR (50${}^{+34}_{-21}$%) shows PAH features in their MIR spectra. Ionized and molecular gas are also commonly detected. Recent star formation episodes are not a rare phenomenon in either FR or SR, even in those dominated by AGN activity (see also. Annibali et al., 2010). Recently, observing HI rich ETGs, Serra et al. (2014) found that SRs are detected as often, host as much H I and have a similar rate of H I discs/rings as FRs.

Despite the expectation that the signature of “wet” or “dry” merger is strongest in the galaxy nucleus, the nuclear MIR spectra do not clearly link the core versus cuspy morphology and the FR versus SR kinematical class to these alternative formation scenarios. Within the last few Gyrs, only at the two extremes of the ETG luminosity, does the dichotomy emerge: the brightest core, mostly SR and faint cuspy, mostly FR (in the Virgo and Fornax clusters) separate into mostly active and passive ETGs, respectively. The result, however, is in contrast to what is expected for core-SR versus cuspy-FR, i.e. a “dry” versus “wet” accretion scenarios.

The obvious possibility is that these photometric and kinematical classes are signatures generated by the two different evolutionary scenarios at $z\gtrsim 0.2$ (see e.g. Khochfar et al., 2011), so that they do not affect the MIR spectra. On the other hand, adopting the traditional E/S0s morphological subdivision Rampazzo et al. (2013) found that Es are significantly more passive than S0s in the same epoch. FR/SR classes may smooth away differences between Es and S0s, since a large fraction of Es transit into the FR class. At the same time, recent observations tend to emphasize the complexity of ETGs when their study is extended to large radii (see also Serra et al., 2014). Arnold et al. (2014) recently obtained the extended, up to 2-4 $r_{e}$, kinematics of 22 ETGs in RSA. They find that only SRs remain slowly rotating in their halos, while the specific angular momentum of ETGs classified as FR within $r_{e}=1$, may dramatically change at larger radii. Arnold et al. (2014) suggest that the traditional E/S0 classification better accounts for the observed kinematics up to large radii and likely of their complex evolutionary scenario.

RR acknowledges partial financial support by contracts ASI-INAF I/016/07/0 and ASI-INAF I/009/10/0. OV acknowledges support from the Conacyt grant CB-2012-183013. This research has made use of the NASA/ IPAC Infrared Science Archive, which is operated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration.