sSFR profiles and slow quenching in the green valley

SDSS IV MaNGA - sSFR profiles and the slow quenching of discs in green valley galaxies

Abstract

We study radial profiles in H equivalent width and specific star formation rate (sSFR) derived from spatially-resolved SDSS-IV MaNGA spectroscopy to gain insight on the physical mechanisms that suppress star formation and determine a galaxy’s location in the SFR- diagram. Even within the star-forming blue cloud, the typical sSFR profile depends on stellar mass. Flat radial profiles are observed for , while star-forming galaxies of higher mass show a significant decrease in sSFR in the central regions, a likely consequence of both larger bulges and an inside-out growth history. Our primary focus is the green valley where, at all masses, we find sSFR profiles that are suppressed with respect to normal star-forming galaxies at galactocentric distances out to 2 effective radii. The responsible quenching mechanism therefore appears to affect the entire galaxy, not simply an expanding central region. However, those galaxies in which central star formation has shut down (classified spectroscopically as central low-ionisation emission-line regions, or cLIERs) show an even stronger suppression in sSFR across the whole disc. These systems are also more suppressed than green valley galaxies with residual, central star-formation. In fact, compared to structural parameters like (the mass surface density within 1 kpc), central quiescence is a stronger predictor of ongoing sSFR suppression at fixed . The green valley hosts both quiescent bulges and strongly suppressed star forming discs, supporting a scenario in which a slow quenching process affects the entire galaxy, and not just the central regions.

keywords:
galaxies: ISM – galaxies: evolution – galaxies: fundamental parameters – galaxies: survey
123

1 Introduction

Galaxies are observed to be strongly bimodal in many of their fundamental properties, including star formation rate (SFR), colour, morphology and the mean age of their stellar population. Galaxy bimodality was already evident in the seminal work of Hubble (1936), and later confirmed with exquisite statistics by large-scale galaxy surveys, like the Sloan digital sky survey (SDSS, Strateva et al. 2001; Kauffmann et al. 2003b; Blanton & Moustakas 2009). These studies motivated the division of galaxies in the colour-magnitude diagram into the star forming ‘blue cloud’ and the passive ‘red sequence’.

Thanks to its superior sensitivity to young hot stars, UV photometry from the GALEX satellite has convincingly demonstrated the existence of a third population of galaxies lying at intermediate colours, in the so-called ‘green valley’ (GV, Wyder et al. 2007; Martin et al. 2007). Since its discovery, the GV has been widely interpreted as a cross-road in galaxy evolution, being populated by galaxies in transition. The transitional nature of GV objects is supported by the fact that the mass flux through the GV is roughly comparable to that needed to assemble the red sequence (Martin et al., 2007), and that GV objects are intermediate in terms of other physical properties, such as Sérsic index/concentration and stellar population ages (Schiminovich et al., 2007; Mendez et al., 2011; Pan et al., 2013). As expected from their intermediate colours, GV galaxies lie mostly below the star formation main sequence (SFMS), the tight relation in the SFR- plane inhabited by star forming galaxies (Noeske et al., 2007; Salim et al., 2007; Renzini & Peng, 2015).

While GV galaxies have larger bulge fractions and are more concentrated than their blue cloud progenitors (Schiminovich et al., 2007; Fang et al., 2013), the role of galactic subcomponents (the disc and the bulge) in the cessation of star formation (‘quenching’) in this population remains unclear. Dressler & Abramson (2014) recently suggested that the intermediate colours and depressed sSFR typical of the GV are the natural consequence of the coexistence of red, passive bulges and blue, star forming discs. According to this scenario the green valley may be explained by a change in bulge to total mass ratio, since the contribution of the (red) bulge increases relative to the (blue) disc moving away from the blue cloud towards the GV. This interpretation of the GV as a bulge-disc ‘purple’ composite is echoed in the work of Abramson et al. (2014), who suggest that galaxy discs form stars at constant sSFR, when the mass of the passive bulge is correctly accounted for. More detailed analysis of the optical colours and bulge and discs in SDSS in presented by Morselli et al. (2017). They find that mean disc colours get redder when moving from the SFMS towards the quiescent region, pointing towards a systematic difference in the star formation rate properties of GV discs. Unfortunately, these works are based on photometry and spatially-limited spectroscopy, and have thus been unable to map the SFR distribution in discs and determine the role of galactic subcomponents.

Intriguingly, in Belfiore et al. (2017b) we have demonstrated, using integral field spectroscopy (IFS) from the MaNGA survey (Bundy et al., 2015), that a large fraction of galaxies in the GV () have quiescent central regions, while hosting star formation in extended outer discs. The central regions in these sources are not devoid of line emission, but are characterised by low H equivalent widths and line ratios typical of low ionisation emission-line regions (LIERs). The line emission is generally not due to an active galactic nucleus (AGN), but is caused by UV radiation emitted by old hot stars (Trinchieri & di Serego Alighieri, 1991; Binette et al., 1994; Cid Fernandes et al., 2011; Belfiore et al., 2016). Galaxies displaying LIER-like central regions are referred to as central LIERs (cLIERs, Belfiore et al. 2016), reflecting their selection criterion. They represent a population of centrally quiescent galaxies, as the regions associated with LIER emission in these objects are largely devoid of young stellar populations, as inferred from stellar continuum features (like the 4000 Å break, as reported in Belfiore et al. 2016, 2017b).

In this work we exploit IFS data from the MaNGA survey to study the equivalent width of H [EW(H)] and sSFR profiles of nearby galaxies in the blue cloud and the GV and assess the relative importance of quiescent central regions and star forming discs. In particular, we study whether the intermediate sSFR of GV galaxies is due their being composite ‘purple’ systems, with quiescent bulges and star forming discs, or whether discs form stars with different sSFRs depending on their galaxy host. This work is part of a series of papers from the MaNGA team dedicated to mapping sSFR on resolved scales. Spindler et al. (submitted) focus on the study of sSFR profiles for the general galaxy population and on the effects of environment, while Sánchez et al. (in prep.) study sSFR profiles in AGN hosts. Finally in Lin et al. (submitted) we present the sSFR properties for a sample of three green valley galaxies observed by MaNGA and ALMA with matched resolutions.

This work is structured as follows. In Sec. 2 we describe the MaNGA IFS data used in this work and the sample selection. In Sec. 3 we describe the EW(H) and sSFR profiles obtained for the blue cloud and the GV and the role of quiescent central regions, while in Sec. 4 we discuss these results and their influence on our understanding of the quenching of star formation. In Sec. 5 we present our conclusions.

2 Data and sample

2.1 The MaNGA data

The MaNGA survey (Bundy et al., 2015; Yan et al., 2016b), part of SDSS-IV (Blanton et al., 2017), aims to obtain spatially resolved spectroscopy for a representative sample of 10 000 galaxies in the redshift range 0.01 z 0.15 by 2020. The MaNGA instrument operates on the SDSS 2.5m telescope at Apache Point Observatory (Gunn et al., 2006) and consists of a set of 17 hexagonal fibre bundles of different sizes, plus a set of mini-bundles and sky fibres used for flux calibration and sky subtraction respectively (Drory et al., 2015; Law et al., 2015; Yan et al., 2016a). All fibres are fed into the dual beam BOSS spectrographs covering the wavelength range from 3600 Å to 10300 Å with a spectral resolution R 2000 (Smee et al., 2013).

MaNGA galaxies are selected from an extended version of the NSA catalogue and are observed out to 1.5 (primary sample, comprising 2/3 of the total sample) or 2.5 (secondary sample, comprising 1/3 of the total sample). Targets are selected to be representative of the overall galaxy population at each stellar mass in the range (in practice is used for sample selection to avoid the systematic uncertainty intrinsic in deriving stellar masses, Wake et al. 2017).

The MaNGA data used in this work was reduced using version v1_5_1 of the MaNGA reduction pipeline (Law et al., 2016). Our starting sample consists of all MaNGA galaxies observed within the first 2 years of operation, corresponding to the publicly available SDSS data release 13 (DR13, Albareti et al. 2016), which includes 1352 unique galaxies.

Figure 1: The relation between sSFR and EW(H) in emission for all the spaxels classified as star forming using the [SII]/H BPT diagram in the sample of MaNGA galaxies considered in this work. Red points with black error bars show medians and scatter in bins of log[EW(H)] and the red dotted line is a linear fit to the median relation. The derived parameters of the linear fit are also shown. The obtained relation is superlinear, and matches the data extremely well, demonstrating that EW(H) can be used successfully to trace the sSFR.

2.2 Spectral fitting

Physical parameters of the continuum and the emission lines are obtained via a customised spectral fitting procedure described in Belfiore et al. 2016, with some minor differences described below. For each galaxy, the stellar continuum is binned to a minimum signal to noise () and fitted using a set of simple stellar population models (Vazdekis et al., 2012). The fit is performed with pPXF (Cappellari & Emsellem, 2004; Cappellari, 2016) using regularisation (McDermid et al., 2015; Cappellari, 2017). We fit the stellar continuum with a Calzetti (2001) extinction law and derive the stellar mass for each resolved region within the galaxy assuming a Chabrier (2003) initial mass function. When integrating over the whole field of view of the MaNGA bundle, the derived stellar masses are in excellent agreement with those reported in the MPA-JHU catalogue (Kauffmann et al. 2003a, with a median offset of 0.02 dex and scatter of 0.25 dex).

After subtracting the stellar continuum the emission lines are fitted on a different binning scheme. To increase the ability to fit weaker lines the velocities of all lines are tied together, thus effectively using the stronger lines to constrain the kinematics of the weaker ones. The reddening of the emission lines is calculated from the Balmer decrement, using the ratio and a Calzetti (2001) attenuation curve with . The theoretical value for the Balmer line ratio is taken from Osterbrock & Ferland (2006), assuming case B recombination (). In order to obtain a reliable extinction correction we select only spaxels with S/N 3 on both H and H.

2.3 Deriving EW(H) and sSFR

The EW of the nebular lines is measured by dividing the measured line flux (with no extinction correction) by a measure of the local continuum. For simplicity of notation, we define equivalent widths to be positive in emission. The limiting detectable EW depends on the relative ratio of the S/N on the line and on the continuum, and on the velocity dispersion of the line (Sarzi et al., 2006). For a barely detected H line (S/N 2) with typical dispersion (, close to the instrumental resolution), the limiting EW is for a continuum S/N 6. Lower EWs are routinely detected in MaNGA in spaxels where the continuum S/N is higher.

The SFR is calculated from extinction-corrected H using the conversion formula from Kennicutt (1998) and a Chabrier (2003) initial mass function. We note that the use of the Salpeter (1955) initial mass function would imply SFR higher by a factor of 0.2 dex, but also an increase in stellar mass by a similar factor (0.22 dex, Madau & Dickinson 2014), eventually resulting in very similar sSFR values.

In calculating the SFR we exclude spaxels with LIER-like excitation, as determined by the Kewley et al. (2001) demarcation line in the [SII]6717,31/H versus [OIII]5007/H Baldwin-Phillips-Terlevich (BPT) diagnostic diagram (Baldwin et al., 1981; Veilleux & Osterbrock, 1987; Kauffmann et al., 2003b; Kewley et al., 2006; Belfiore et al., 2016).

Using repeat observations Yan et al. (2016b) demonstrate that for (converting to a Chabrier IMF) and the uncertainty in the SFR is less than 0.2 dex. This uncertainty is dominated by the uncertainty in the line fluxes and not by the absolute and relative spectrophotometric calibration (Yan et al., 2006). Using the current MaNGA data we obtain that a S/N 3 for H corresponds to a median . We take this to represent the median SFR sensitivity limit in MaNGA, but in the case of low extinction and for the lowest redshift objects lower SFR can be measured.

In Fig. 1 we show the relationship between EW(H) and sSFR for all the star forming spaxels in the current MaNGA sample. The relationship between sSFR and EW(H) is well-fitted by a power law over almost 2 dex in sSFR, given by

(1)

Our best fit coefficients agree well with those reported by Sánchez et al. (2013) in a study based on CALIFA data. The relation between sSFR and EW(H) is superlinear, possibly due to the M/L dependence of the EW. In fact, EW/sSFR where is the light emitted in the continuum passband around H. Higher EW(H) corresponds to younger stellar populations and hence lower M/L, going in the direction of the observed trend. Differences in relative extinction between gas and the stellar component will also affect the relation between sSFR and EW(H), but the magnitude of this effect is difficult to quantify and remains disputed (Calzetti et al., 1994, 2000).

In light of these limitations, in this work we consider both the EW(H) profiles, which are less model-dependent and easier to compare with other (higher-redshift) datasets, and the sSFR profiles, which provide a more physical picture, but are subject to modelling assumptions (e.g. the derivations of the M/L for the stellar population and the dust extinction correction). Moreover, we only calculate the sSFR for spaxels classified as star forming using the [SII] BPT diagram. Therefore, quiescent regions will have a different effect in the EW(H) and sSFR profiles shown in the next sections. However, as demonstrated by the tight relation between sSFR and EW(H), the trends derived in this work are robust to the choice of tracer.

Figure 2: Top: The MaNGA DR13 sample in th versus diagram, subdivided according to spectroscopic class in star forming (SF, blue), central LIER galaxies (cLIERs, green), extended LIER galaxies (eLIERs, red) and passive (black). Contours are drawn to emphasise the respective position of SF (blue) and cLIER (green) galaxies in this diagram, demonstrating that cLIERs preferentially live at the high-mass end of the green valley (GV, delimited by the dashed lines). Bottom: The position of SF (blue) and cLIER (green) galaxies in the SFR versus diagram. SF galaxies lie on a linear relation (the SFMS, fitted with the red line). The dashed black lines represent the 1 width of the SFMS. Most cLIER galaxies lie below the SFMS (as demonstrated by the green density contours).
Figure 3: Top: The stacked EW(H) radial profiles for different mass bins in the blue cloud (left) and the GV (right). Profiles are computed by deprojecting the gradient using the photometric inclinations from the NASA-Sloan catalogue and normalised to the elliptical Petrosian effective radius (). In these plots we consider both star forming and LIER regions with detected line emission. Bottom: Same for sSFR. Only regions classified as star forming using the [SII] BPT diagram and the Kewley et al. (2001) demarcation line are considered.
Type N. galaxies
Sample used in this work 579
Blue Cloud 443 (76%)
Green Valley (GV) 109 (19%)
Red Sequence 27 (5%)
star forming (SF) 455 (79%)
central LIER (cLIER) 124 (21%)
Table 1: Sample of MaNGA galaxies used in this work.

2.4 The galaxy sample

In this work we define the blue cloud and the GV using UV-optical colours ( and respectively, where is the magnitude in GALEX near-ultraviolet band, see Fig. 2, top panel). The use of optical colours alone is inadvisable to select galaxies with intermediate sSFR, as the high-mass end of the blue cloud, the GV and the red sequence all have equivalent optical (e.g. ) colours (see Appendix A, also e.g. Kauffmann et al. 2007). We note that the UV-upturn phenomenon, observed in sub-samples of early-type galaxies, and thought to be caused by radiation from old hot stars (O’Connell, 1999), is not sufficient to move red sequence galaxies into the UV-optical GV. Hence the UV-optical green valley is truly inhabited by intermediate sSFR systems.

Following the classification scheme proposed in Belfiore et al. (2016) we subdivide galaxies according to their line emission in star forming (SF, dominated by star formation at all radii), central LIER (cLIER, LIER emission in the central regions but star forming at larger galactocentric distances), extended LIER (eLIER, LIER emission at all radii) and line-less (no line emission detected). We exclude galaxies hosting Seyfert nuclei, as they are the subject of a parallel study (Sanchez et al., in prep.). Operationally, galaxies are classified as cLIERs if their line emission from a central aperture of 3 in diameter is classified as LI(N)ER in the [SII]-BPT diagram using the demarcation lines of Kewley et al. (2006), but star forming regions are also present at larger galactocentric distance. As already noted in Belfiore et al. (2016), all galaxies in MaNGA hosting star formation have some line emission in the bulge/central regions, showing either star-forming, Seyfert or LIER line ratios. This means that there exist no line-less bulges in the local Universe. cLIER galaxies represent, therefore, the only class of objects with no spectroscopically-detected star formation in the central regions.

cLIER galaxies are found at the high-mass end of the blue cloud and across the GV, and tend to lie below the SFMS (Belfiore et al., 2017b), as seen in Fig. 2. We note that cLIER are virtually absent for , except for a small population (13 galaxies), which tend to lie at the low-mass end of the blue cloud. We comment further on these galaxies in Appendix B. Considering the low number statistics and the possible contaminant nature of these galaxies, only cLIERs with are considered in this work.

The SFMS obtained here using the integrated SFR from the MaNGA data is sub-linear (slope of , Fig. 2, bottom panel). Moreover, it does not show a flattening at the high mass end if one only selects star forming galaxies. A flattening would be obtained, on the other hand, if galaxies with quiescent central regions (cLIERs) were included in the fit.

eLIER and line-less galaxies are defined to host no spectroscopically-detected star formation, and are thus excluded from the current study. We emphasize that in this work a galaxy is defined as SF based on its MaNGA spectroscopy and not by its position with respect to the SFMS, although, as seen in Fig. 2, galaxies classified as SF lies mostly on the SFMS.

For a reliable study of radial gradients we impose further cuts on our galaxy sample, as in Belfiore et al. (2017a). In detail we select galaxies to have a major to minor axis ratio (b/a) greater than 0.4 (to exclude high-inclination systems), and exclude visually-classified mergers or closely interacting systems. These cuts lead to a final sample of 579 galaxies. The relative mix of blue cloud and GV, as well as SF and cLIER sources in this sample is detailed in Table 1.

Elliptical Petrosian effective radii () and inclinations from the NASA-Sloan catalogue (Blanton et al., 2011) are used throughout this work to construct de-projected radial gradients, following the procedure utilised in Belfiore et al. (2017a). We have checked that the blue cloud and GV samples have comparable inclinations and our results are robust to a change in the b/a cut described above. For each galaxy, we compute deprojected radial profiles of the EW(H), (SFR surface density) and (stellar mass surface density). The (sSFR surface density) profile is then obtained by dividing the by the in each annulus. If no star forming regions are found within a given annulus, then the (and hence ) are not computed. This procedure leads to an inner truncation of the sSFR profiles for cLIER galaxies, since no star formation is detected in the centres of these systems. We refrain from assigning an sSFR upper limit to these regions. In Spindler et al. (submitted) an upper limit is instead derived based on the strength of 4000 Å break.

3 sSFR and EW(H) profiles

3.1 The blue cloud and the green valley

The EW(H) and sSFR profiles for blue cloud and GV galaxies in five stellar mass bins are presented in Fig. 3. For each stellar mass bin, the radial profile is computed as the Tukey biweigth of the galaxies contributing to the bin at that radius. Error bars are obtained by calculating a robust estimator for the sample standard deviation and dividing by , where N is the number of profiles at each radius.

Not all galaxies contribute to all radii, since some galaxies are not covered out to 2.0 and the SFR is not computed in regions classified as LIERs. Both these effects may lead to biases in the stacked line profiles. The bias resulting from the truncation of the sSFR profiles in the central regions due to central LIERs is discussed in Sec. 3.2 below. The bias arising from the truncation of sSFR profiles at larger radii is due to the fact that primary sample galaxies are only covered to 1.5 in MaNGA. At larger radii, only some primary sample galaxies will be covered, and these may represent a biased subset of the overall galaxy population. To address this bias we computed the profiles obtained using only the secondary sample, which consists of a representative sample of galaxies covered to 2.5 . We find that the qualitative features of the profiles at large radii are preserved.

For blue cloud galaxies we observe a remarkably regular change of the shape of the EW(H) and sSFR radial profiles with total stellar mass. Both EW(H) and sSFR profiles show two key features: a sharp, mass-dependent decrease in the central regions, and a slow decrease in the sSFR profile at large radii, where the profiles have little to no mass dependence.

In the central regions of blue cloud galaxies the sSFR difference between the profiles of the lowest and highest stellar mass bin reaches 0.8 dex. This mass-dependent decrease in sSFR in the central regions may reflect the presence of a bulge component, but also the decrease in sSFR associated with the inside-out growth of discs (see Sec. 4). To put the absolute scale of the sSFR into context, a useful number to keep in mind is that characteristic sSFR for a main sequence galaxy of in SDSS is (Peng et al., 2010; Renzini & Peng, 2015). The upward-bending profile in sSFR observed in the innermost radial bins for the highest masses is partially an artefact caused by the stacking of galaxies with different profiles, as discussed in Sec 3.2.

The key result of this work, evident from Fig. 3, is the existence of a systematic difference in the sSFR [and EW(H)] profiles between the blue cloud and the GV at all radii. For all stellar mass bins, EW(H) and sSFR are suppressed in the GV with respect to the blue cloud at all radii, even though the radial profiles maintain a qualitatively similar shape. The decrease in EW(H) and sSFR in the GV with respect to the blue cloud has only a weak radial dependence and, importantly, is prominent out to the largest radii probed in this work (out to , where the disc is the dominant mass component). For example, at 1.0 , the suppression in sSFR between the blue cloud and the GV median profiles is between 0.35-0.8 dex, depending on the mass bin considered. At 2.0 the GV data is noisier and the median suppression is between 0.55-0.65 dex, with only a weak dependence on stellar mass. We note that the very central regions of blue cloud and GV galaxies for the highest mass bin () have comparable sSFR. The reason for this behaviour in the highest-mass bin is discussed in the next section, and is due to the presence of centrally suppressed cLIER galaxies in both the blue cloud and the GV.

We note that the main differences in the EW(H) and sSFR profiles are preserved if instead of defining the blue cloud and the GV using colours, galaxies are split into a population lying on the SFMS and another below the SFMS (at intermediate integrated sSFR).

3.2 The role of centrally quiescent regions

In GV galaxies the EW(H) and sSFR profiles have a different shapes at small galactocentric radii. This discrepancy originates from the inclusion of LIER regions in the EW(H) profiles. These regions are, on the other hand, intentionally excluded when computing the sSFR. The EW(H) profiles in the GV are therefore capturing not only the decrease in sSFR, but also the increasing contribution of central LIER regions, which have low EW(H) Å (Belfiore et al. 2017b and Appendix B).

To quantify the impact of LIER-like quiescent regions in the GV, in Fig. 4 we show the fraction of spaxels classified as LIERs as a function of radius for galaxies of different stellar mass in the GV. The figure demonstrates the increase in LIER fraction at small radii. The increase is mass-dependent, being more pronounced at higher masses. The highest mass bin [] is dominated by LIER ionisation (i.e. more than 50% LIER fraction) for , and reaches % LIER fraction in the innermost radial bins. This feature is also evident by looking at the EW(H) profile of the GV galaxies in the highest mass bin (in Fig. 3, top), which reaches the characteristic value of EW(H) Å at small radii.

Figure 4: The fraction of spaxels classified as LIERs (according to the [SII]-BPT diagram) in GV galaxies of different stellar masses as a function of galactocentric radius, normalised to the effective radius.
Figure 5: The sSFR radial profiles for three stellar mass bins () for GV galaxies subdivided into (centrally) star forming (GV SF, in green) and central LIER (GV cLIER, in red). Profiles for the blue cloud galaxies are also shown (in black).
Figure 6: Same as Fig. 3 bottom, but subdividing galaxies into star forming (SF) and cLIER, following the classification scheme introduced in Belfiore et al. (2016), which classifies galaxies according to their resolved excitation morphology making use of the BPT diagnostic diagram.

The increasing predominance of quiescent LIER-like central regions in the GV leads to a natural bimodality in the SFR profiles of galaxies: centrally star forming and cLIERs, which are centrally quiescent. We show the sSFR profiles for both classes of galaxies in the GV in Fig. 6. We only consider mass bins of , for which we have sufficient statistics in the GV. In this mass range we have a sample of 96 GV galaxies, of which 50 (52%) are cLIER and the remaining 46 (48%) are SF.

A sharp decline in sSFR is observed at small radii in cLIER galaxies, since we do not measure sSFR in LIER regions. The increase of the sSFR at small radii observed in the previous section in the GV stacks can, therefore, be interpreted as the combination of strongly declining sSFR profiles for cLIERs and flatter profiles for centrally star forming GV galaxies. Since cLIER do not have measurable sSFR profiles in the central regions, the central regions in a mixed stack only receive contributions from the flatter profiles of the star forming galaxies. We note that, while in this work we simply discard annuli with no star formation, in Spindler et al. (submitted) we use the observed of these regions to define an upper limit for their sSFR []. Taking this difference into account, the results presented here are equivalent to those presented in Spindler et al. for centrally quiescent galaxies.

We note that cLIERs are also present in the blue cloud, although they are rare and have a negligible effect on the sSFR profiles expect for our highest stellar mass bin []. The central increase in sSFR profile associated with this bin can, therefore, be also interpreted as an artefact of stacking the bimodal sSFR profiles of cLIER and SF galaxies.

In Fig. 6 we show the sSFR profiles for galaxies classified as SF and cLIER, irrespective of their position in the blue cloud or GV. The profiles of star forming galaxies are similar to those of blue cloud galaxies, since the contamination of cLIERs to the blue cloud is small, except in our highest-mass bin. Overall, cLIER galaxies have suppressed sSFR profiles with respect to star forming galaxies of the same mass, both in their bulges, but also in their star forming discs, at large . Their sSFR profiles are remarkably similar as a function of stellar mass. Notably, the degree of sSFR suppression of cLIER with respect to star forming galaxies is larger than that of general GV population (as also evident for the GV sub-sample in Fig. 6). This points to central LIER emission as a tell-tale sign of the quenching process, as further discussed in Sec. 4.3.

in Belfiore et al. (2017b) we showed that cLIER galaxies have higher concentration and Sérsic indices than star forming galaxies of the same mass. It is, therefore, interesting to ask whether the sSFR difference between star forming and cLIER galaxies persists when considering galaxies of the same mass and also the same central concentration. To make such a comparison we considered different observational proxies for bulge prominence, including the Sérsic index (measured from -band SDSS photometry), the median velocity dispersion within 1.0 and the mean stellar mass surface density within 1 kpc from the galactic centre (, as defined in Fang et al. 2013, and measured here directly from the MaNGA stellar mass maps). For each cLIER galaxy, we associate a blue cloud galaxy within the same stellar mass bin and with similar bulge proxy (within 0.15 dex). We then compare the sSFR radial profiles of cLIER and the selected comparison sample of star forming galaxies in the different mass bins. The results for the different bulge proxies considered are similar, so we show in Fig. 7 only the results obtained applying this procedure with as a proxy for the bulge. We conclude from this analysis that the sSFR profiles of cLIER galaxies are suppressed with respect to those of star forming galaxies of comparable stellar mass and at all radii.

4 Discussion

4.1 The role of the bulge in the blue cloud and the green valley

The decrease in the EW(H) and sSFR profiles observed in the central regions of blue cloud galaxies can be attributed to the combined effects of radial dependent disc formation timescales and bulge assembly.

According to the expectations from growth of structure in the cosmological context, the central regions of galactic discs are formed at earlier times (‘inside-out’ growth), and are therefore more evolved, less gas rich and will show lower sSFR than the outer disc, which is still un-evolved and gas rich (Muñoz-Mateos et al., 2011; Pezzulli et al., 2015). The inside-out growth paradigm is supported by extensive observational evidence, including disc colour gradients (de Jong, 1996; Bell & de Jong, 2000; Muñoz-Mateos et al., 2007), studies of their stellar populations fossil record (Sánchez-Blázquez et al., 2014; González Delgado et al., 2015; Goddard et al., 2017; González Delgado et al., 2016) and gas phase metallicity gradients (Boissier & Prantzos, 1999; Chiappini et al., 2001; Mollá & Díaz, 2005).

The assembly of a bulge component, which may contribute stellar mass, but no star formation, would also be responsible for the observed decrease in the sSFR profiles of star forming galaxies at small galactocentric distances. The importance of the bulge in setting the integrated sSFR of galaxies has recently been discussed in Abramson et al. (2014). They argue, based on single-fibre SDSS data and a set of bulge-disc decompositions, that the integrated sSFR of galaxies is seen to decrease as a function of stellar mass, because of the increasing contribution of quiescent bulges in high-mass systems. Indeed, in this work we find that the sSFR of galaxies of different masses shows a very small dependence on integrated stellar mass at large radii (). At , for example, we find sSFR values going from for to for . Therefore our data is in agreement with the idea, introduced in Abramson et al. (2014), that the sSFR of discs does not strongly depend on , at least in the outer disc.

Based on studies of their Sérsic indices and the stellar mass profiles, GV galaxies are known to host larger bulges than blue cloud galaxies of the same stellar mass (Schiminovich et al., 2007; Fang et al., 2013). Hence their sSFR profiles are expected to be more centrally suppressed than those of blue cloud galaxies of same total mass. In this work we have additionally shown that the sSFR profiles of GV galaxies differ from those of the blue cloud also in their discs (i.e. at large galactocentric radii). The implication is that the suppressed integrated sSFR of GV galaxies cannot be reproduced simply by adding more mass in the quiescent bulge, but a change in the properties of the disc is also necessary.

Comparing to higher redshift work is challenging, but our results here echo those of Nelson et al. (2016), who measure EW(H) and sSFR profiles for galaxies on and below the SFMS at , and conclude that at this redshift the sSFR of galaxies lying below the SFMS is already suppressed with respect to galaxies on the SFMS, at all radii. The similarities in the qualitative shape of the sSFR profiles as a function of mass between this work and Nelson et al. (2016) is in fact striking.

A ‘slow quenching’ mechanisms, which preserves the disc morphology while building up the bulge (possible via interactions and minor mergers, Bundy et al. 2010, or bar-driven secular evolution) seems favoured by our observations. In this scenario, GV galaxies, once fully quenched, may give rise to the population of passive discs seen both locally (Masters et al., 2010; Morselli et al., 2017) and at high redshift (Bundy et al., 2010).

Figure 7: Same as Fig. 6, but comparing the cLIER sample with a sample of star forming galaxies matched in mass and (the stellar mass surface density within 1 kpc, as defined in Fang et al. 2013). The comparison demonstrates that GV galaxies have lower sSFR in their discs even when compared to a sample of blue cloud galaxies with comparable total mass and central mass surface density ().

4.2 The green valley as a quasi-static population

Models are able to reproduce the tight nature of the SFMS by assuming galaxies live in an equilibrium between cosmological accretion rate, star formation and outflow rate (White & Frenk, 1991; Bouché et al., 2010; Lilly et al., 2013; Peng & Maiolino, 2014). Galaxies below the SFMS, which live in the GV, are generally assumed to have been perturbed off this equilibrium, and in the process of transitioning towards the red sequence. Galaxies transiting in the opposite direction (from the red sequence to the blue cloud) as a consequence of a ‘rejuvenation’ event are found by models to be a sub-dominant population in the GV (Trayford et al., 2016; Pandya et al., 2017).

Following a somewhat different approach, quenching can also be defined as the smooth crossing of galaxies to arbitrarily low sSFR (Gladders et al., 2013; Abramson et al., 2016). While this definition explicitly neglects very fast quenching events, it has the advantage of regarding quenching as a more ‘natural’ endpoint of galaxy evolution, which may not necessarily require a triggering event.

In either case, the timescale associated with crossing the GV remains a point of contention. This timescale was originally thought to be very short, in order to explain the lack of a substantial population of galaxies at intermediate optical colours (Baldry et al., 2004). As shown in Wyder et al. (2007) and Salim et al. (2007), the use of UV-optical colours highlights the presence of a substantial population of galaxies at intermediate sSFR, which was previously not evident in the optical colour-magnitude diagram. Exploiting the selection criterion, Martin et al. (2007) derived timescales of the order of Gyr for galaxies to cross the GV. Intermediate redshift data suggests that this timescale should decrease at higher redshift (Gonçalves et al., 2012).

Schawinski et al. (2009) compared UV-optical colours of GV galaxies with stellar population models to argue that, while early-type galaxies may have transitioned quickly trough the GV (see also Schawinski et al. 2007), late type galaxies are likely to be on a slow pathway to quiescence. They suggest ‘strangulation’ (i.e. the reduction of the cold gas supply from the cosmological accretion) as a viable mechanism for the slow quenching mode. More recently Smethurst et al. (2015) used UV-optical colours in a Bayesian framework to fit two-parameter star formation histories and argue in favour of the existence of both a fast and a slow quenching channel through the GV. Abramson et al. (2016) use an ensemble of log-normal star formation histories, matching the overall star formation history of the Universe, to argue that galaxies which appear quenched today simply had more compressed star formation histories, and have therefore already exhausted their cosmic supply of gas (although they also acknowledge the need to maintain these galaxies quiescent after they gas supply has been exhausted, possibly via AGN heating). This simple model naturally predicts that galaxies crossing the GV at higher redshift had more compressed star formation histories and were thus transitioning on faster timescales (‘fast quenching’). Galaxies crossing the GV today, on the other hand, correspond to more extended star formation histories and are therefore exhausting their gas on slower timescales.

The existence of widespread star formation in the discs of GV galaxies with quiescent central regions is now well established, thanks to UV observations first (Kauffmann et al., 2007; Thilker et al., 2007; Salim & Rich, 2010; Salim et al., 2012; Fang et al., 2012), and large IFS surveys later (Belfiore et al., 2017b). Moreover, in Belfiore et al. (2017b) we have studied the kinematic misalignment between gas and stars in cLIER galaxies, and find no strong misalignments and regular disc kinematics (this conclusion is to be contrasted with the kinematics of extended LIER galaxies, which are found to be strongly misaligned). Further insight can be obtained from a more detailed study of the stellar population ages in the GV. Although this is beyond the scope of this work, we point out that the central regions of cLIER galaxies show uniformly old stellar populations, implying that star formation has been absent in these regions for Gyr (Belfiore et al., 2017b). This evidence suggests that neither fast gas expulsion nor a recent major merger can easily explain the star formation in the green valley, and instead supports the idea of the GV as a ‘quasi-static’ population subject to a slow quenching process (borrowing the terminology from Salim 2014).

These observations do not preclude the existence of a fast quenching channel, exemplified, for example, by the population of post-starburst galaxies (Wild et al., 2017). Since these galaxies would transition through the GV on fast timescales, the majority of galaxies currently observed in the GV would still belong the quasi-static population.

4.3 The physics of the slow quenching mode

In light of the need for a slow quenching mechanism two main classes of processes can be responsible for the observed reduction in sSFR in GV discs: ‘preventive’ processes, leading to a reduction in the gas content of GV galaxies, or ‘sterilising’ processes, predicating a reduction in the star formation efficiency of the cold gas. Purely ‘ejective’ processes, which would require the expulsion of the gas, are unlikely to lead to slow quenching, may still act in combination with other (e.g. preventive) processes to lead to the observed quenching signature.

AGN feedback, for example, is not only ejective but can also be preventive in nature, and lead to reduced rates of gas accretion onto GV galaxies, causing a slow fading of their discs and decrease in their overall sSFR. More generally, any process that prevents cold gas accretion from the IGM from reaching the galaxy will lead to slow quenching over the entire galaxy disc (this process is sometimes referred to as ‘strangulation’, Peng et al. 2015).

Alternatively a mechanism is required to make star formation more inefficient in bulge-dominated galaxies (Martig et al., 2009) or galaxies which host bars (Cheung et al., 2013; Emsellem et al., 2015; Gavazzi et al., 2015). Based on studies of the integrated sSFR and gas content of galaxies below the SFMS, it is likely that both processes are at play, since galaxies below the SFMS are found to have both lower gas fractions and lower star formation efficiencies (Saintonge et al., 2012; Genzel et al., 2015; Saintonge et al., 2016). A pilot study of resolution-matched ALMA and MaNGA data (Lin et al., submitted) in three GV galaxies confirms this scenario, although larger samples of resolved molecular gas observations would prove useful in understanding the relative changes in star formation efficiency and gas fraction between the bulge and disc of GV galaxies.

Finally, cLIERs are prototypical GV bulge+disc systems, where the inner regions have stopped forming stars and residual line emission is powered by hot evolved stars. We note that in MaNGA quiescent bulges are detected as LIERs since, after visual inspection, all galaxies that host a star forming disc also have line detections in their central regions. As demonstrated in this work, cLIERs lie off the SFMS not only because of their quiescent bulges, but also because of the strongly suppressed sSFR in the discs. Importantly the discs of cLIERs have sSFR which are even lower than those of an average GV galaxy of the same mass, making this class of objects the most deviant from the SFMS.

In summary, despite the cLIER label being attributed according to the properties of the central regions, cLIER discs are also the most deviant from the quasi-universal sSFR observed to characterise discs of star forming galaxies at large radii (sec. 3.1). Galaxies, therefore, do not become cLIERs just by quenching their bulge, but there is a close relation between the quenching of the central regions and the suppression of star formation in the disc. The required inside-out quenching process does not act as a step function in radius, only affecting the innermost regions and gradually moving outwards, but has an effect on the entire galaxy.

5 Conclusions

We derive sSFR and EW(H) radial profiles in a representative sample of nearby () galaxies with resolved spectroscopy from SDSS-IV MaNGA in order to compare the stacked profiles of blue cloud () and green valley galaxies (). We make use of the MaNGA IFS data to elucidate the role of galactic subcomponents (the bulge and the disc) in transitioning galaxies. Central LIER (cLIER) galaxies, which are defined to have quiescent, LIER-like central regions and star forming outer regions, are a particularly interesting class of objects in this context, as they live preferentially in the GV and the upper-mass end of the blue cloud. Below we summarise the main findings of this work.

  1. The sSFR and EW(H) profiles in blue cloud (and in BPT-classified star forming) galaxies evolve smoothly as a function of total stellar mass. At small galactocentric radii, these profiles show a decrease for galaxies of higher masses, which is particularly evident at . This trend may reflect the increasing importance of the bulge component, but also the natural evolution of the sSFR in the context of inside-out growth, where central regions are more evolved and therefore less gas rich. At large radii () the sSFR of galaxies of different masses shows a gradual decline and converges to a roughly constant value, independent of stellar mass ( at ).

  2. The sSFR and EW(H) profiles of GV galaxies show a suppression with respect to mass-matched blue cloud galaxies at all radii. The fact that this suppression persists out to the largest radii probed in this work (i.e. ) indicates that the star formation properties of the discs of GV galaxies are fundamentally different than those of the discs of blue cloud galaxies of the same mass.

  3. cLIER galaxies not only show absence of star formation in the central regions (by definition), but also show a suppression in sSFR across their entire disc. Such suppression is on average stronger than that observed for the full sample of GV galaxies. Even when controlling for both and , cLIER galaxies exhibit significantly suppressed star forming discs. A cLIER classification is, therefore, a better proxy for suppressed star formation than mass or alone.

These observations demonstrate that the GV is a complex space. While GV galaxies have larger bulges than blue cloud galaxies at fixed mass, this work demonstrates they also have fundamentally different discs, with suppressed sSFR. This observation supports a view of the GV as a quasi-static population, requiring a slow quenching process, uniformly affecting the entire galaxy.

Acknowledgements

We wish acknowledge Alvio Renzini, for his suggestion to further quantify the role of central LIER galaxies on the star formation main sequence. F.B. and R.M. acknowledge funding from the Science and Technology Facilities Council (STFC). R.M. acknowledges funding from the European Research Council (ERC), Advanced Grant 695671 ‘QUENCH’. M.B. acknowledges funding from NSF/AST-1517006. This work makes use of data from SDSS-IV. Funding for SDSS has been provided by the Alfred P. Sloan Foundation and Participating Institutions. Additional funding towards SDSS-IV has been provided by the U.S. Department of Energy Office of Science. SDSS-IV acknowledges support and resources from the Centre for High-Performance Computing at the University of Utah. The SDSS web site is www.sdss.org. This research made use of Marvin, a core Python package and web framework for MaNGA data, developed by Brian Cherinka, José Sánchez-Gallego, and Brett Andrews (Cherinka et al., 2017). SDSS-IV is managed by the Astrophysical Research Consortium for the Participating Institutions of the SDSS Collaboration including the Brazilian Participation Group, the Carnegie Institution for Science, Carnegie Mellon University, the Chilean Participation Group, the French Participation Group, Harvard-Smithsonian Center for Astrophysics, Instituto de Astrofísica de Canarias, The Johns Hopkins University, Kavli Institute for the Physics and Mathematics of the Universe (IPMU) / University of Tokyo, Lawrence Berkeley National Laboratory, Leibniz Institut für Astrophysik Potsdam (AIP), Max-Planck-Institut für Astronomie (MPIA Heidelberg), Max-Planck-Institut für Astrophysik (MPA Garching), Max-Planck-Institut für Extraterrestrische Physik (MPE), National Astronomical Observatory of China, New Mexico State University, New York University, University of Notre Dame, Observatário Nacional / MCTI, The Ohio State University, Pennsylvania State University, Shanghai Astronomical Observatory, United Kingdom Participation Group, Universidad Nacional Autónoma de México, University of Arizona, University of Colorado Boulder, University of Oxford, University of Portsmouth, University of Utah, University of Virginia, University of Washington, University of Wisconsin, Vanderbilt University, and Yale University.

The MaNGA data used in this work is publicly available at http://www.sdss.org/dr13/manga/manga-data/.

Appendix A Defining the green valley

In this work we have used integrated colours to define the GV. Previous work has sometimes defined the GV using optical colours (e.g. ) or by considering galaxies lying below the SFMS at fixed mass. In this appendix we show that the GV corresponds closely to the GV as defined by galaxies below the SFMS, while the GV defined in optical colours is heavily contaminated by blue cloud star forming galaxies.

In order to test different definitions of the GV, we consider the MaNGA targeting catalogue, consisting of all SDSS galaxies meeting the MaNGA target selection criteria. For these galaxies, we recover elliptical Petrosian photometry in the GALEX and SDSS bands from the extended NSA catalogue and stellar masses and SFR from the MPA-JHU catalogue (Brinchmann et al., 2004). We note that the photometry used here is k-corrected and corrected for Galactic extinction, but not corrected for attenuation intrinsic to the galaxy. We define the GV in and using the following cuts respectively (1) (same as adopted in this work) and (2) (Mendel et al., 2013). Fig. 8, left panels, shows the position of the GV defined in this way in the colour-mass diagram.

We then consider the position of GV galaxies, selected using these two different cuts, in the -SFR plane (Fig. 8, right panels). The SFMS definition of Renzini & Peng (2015) is adopted (dashed red lines in Fig. 8, right panels, correspond to the SFMS dex). We find that, using the definition of the GV, galaxies are preferentially selected either on the SFMS and its lower envelope or in the passive population, with very few intermediate sSFR galaxies actually selected (green contours in Fig. 8, right panels corresponds to the position of the GV-selected galaxies). Therefore the GV selection is not only biased by the inclusion of star forming galaxies, but also avoids intermediate sSFR systems (the same point is illustrated in Salim 2014). Using the definition of the GV, on the other hand, we select preferentially galaxies below the SFMS.

We quantify the contamination of star forming galaxies using these two definitions of the GV by calculating the fraction of GV galaxies that lie within 0.3 dex of the SFMS. Using the green valley definition, this contamination fraction is 30%, while using the definition the contamination decreases to 7%. We conclude that selection is sufficient to select intermediate sSFR systems, while one should avoid selecting GV galaxies using colours.

Figure 8: Top left: The rest-frame colour - mass diagram for the MaNGA parent sample. The definition of the GV in colours of Lackner & Gunn (2012) is shown by the dashed green lines. Top right: The SFR- diagram for the MaNGA parent sample, showing the position of the galaxies selected using the -GV cuts of Mendel et al. (2013) using green contours, enclosing respectively 50% and 70% of the -GV galaxies. The plot demonstrates that the adopted cut in select a large number of galaxies on the SFMS, while also selecting red sequence objects. Overall the colour cut does not preferentially select galaxies between the SFMS and the upper limits representing passive galaxies. Bottom: Same as top panels, but using rest-frame colours instead of . The GV is defined to have . This selection criteria leads to a sample of galaxies located preferentially below the SFMS, as shown by the green contours in the bottom right panel.

Appendix B The properties of quiescent regions in central LIER galaxies

In this work we have argued that cLIER galaxies host quiescent central regions, and that LIER excitation is a useful classifier to select these centrally quiescent objects. In the scenario where LIER emission is due to hot evolved stars, the EW(H) in emission is expected to be low, as indeed observed in the centres of cLIER galaxies. A complementary measure of quiescence may be derived from the stellar population properties, by analysing, for example, an age-sensitive tracer like . In this appendix we explicitly show the equivalence of these different definitions of quiescence, in accordance with the detailed analysis already presented in Belfiore et al. (2016) and Belfiore et al. (2017b).

First, we focus on the properties of the central regions of galaxies. For each galaxy in our sample we derive the mean EW(H) and in a central aperture 3 in diameter. In Fig. 9 we show the position of star forming (blue) and cLIER with (red) galaxies in the plane defined by the central EW(H) and . We see that cLIER galaxies lie at the high and low EW(H) end of the distribution. An EW(H) of 3Å is found to be a good empirical division between star forming and passive galaxies in SDSS (Cid Fernandes et al., 2011). In the current sample, 84% of high-mass cLIER have EW(H) < 3 Å and 96% have EW(H) < 6 Å. We conclude that the central regions of high-mass cLIERs are classified as quiescent using both the EW(H) and .

The position of the low-mass cLIERs , which have not been used in this work, in the EW(H) - plane is also shown in Fig. 9 (green dots). These galaxies lie in the region of the EW(H) - occupied by the star forming galaxies, although they occupy the lower EW(H) tail of the distribution. Visual inspection of the emission line maps of these objects shows that these galaxies are primarily low-mass systems with very clumpy line emission. In cases where the galaxy centre does not correspond to an area of high star formation, it may be dominated by diffuse ionised gas, thus making these galaxies appear as central LIER objects. More detailed analysis of these objects is beyond the scope of this appendix. We conclude, however, that we are justified in their exclusion in this work, which aimed at studying the cessation of star formation in the central regions of galaxies.

Figure 9: The position of star forming (blue) and cLIER galaxies used in this work (, red points) in the EW(H) - plane. EW(H) and are computed from the MaNGA data using a central aperture 3 in diameter. The dotted line represents the empirical division between star forming and passive galaxies (EW(H) 3Å) suggested in previous SDSS work (Cid Fernandes et al., 2011). Low-mass cLIER galaxies () are also shown as green points. These galaxies are low-mass systems where the central aperture falls in an area dominated by diffuse ionised gas and have different properties from the high-mass cLIER galaxies, where the central regions are classified as quiescent using both EW(H) and .

Footnotes

  1. pubyear: 2017
  2. pagerange: SDSS IV MaNGA - sSFR profiles and the slow quenching of discs in green valley galaxiesB
  3. pubyear: 2016

References

  1. Abramson L. E., Kelson D. D., Dressler A., Poggianti B., Gladders M. D., Oemler A., Vulcani B., 2014, ApJ, 785, L36
  2. Abramson L. E., Gladders M. D., Dressler A., Oemler A., Poggianti B., Vulcani B., 2016, ApJ, 832, 7
  3. Albareti F. D., et al., 2016, arXiv:1608.02013v1
  4. Baldry I. K., Glazebrook K., Brinkmann J., Ivezic Z., Lupton R. H., Nichol R. C., Szalay A. S., 2004, ApJ, 600, 681
  5. Baldwin J. A., Phillips M. M., Terlevich R., 1981, PASP, 93, 5
  6. Belfiore F., et al., 2016, MNRAS, 461, 3111
  7. Belfiore F., et al., 2017a, MNRAS, 170, 151
  8. Belfiore F., et al., 2017b, MNRAS, 466, 2570
  9. Bell E. F., de Jong R. S., 2000, MNRAS, 312, 497
  10. Binette L., Magris C. G., Stasińska G., Bruzual A. G., 1994, A&A, 292, 13
  11. Blanton M. R., Moustakas J., 2009, ARA&A, 47, 159
  12. Blanton M. R., Kazin E., Muna D., Weaver B. A., Price-Whelan A., 2011, AJ, 142, 31
  13. Blanton M. R., et al., 2017, AJ, 154, 28
  14. Boissier S., Prantzos N., 1999, MNRAS, 307, 857
  15. Bouché N., et al., 2010, ApJ, 718, 1001
  16. Brinchmann J., Charlot S., White S. D. M., Tremonti C., Kauffmann G., Heckman T., Brinkmann J., 2004, MNRAS, 351, 1151
  17. Bundy K., et al., 2010, ApJs, 719, 1969
  18. Bundy K., et al., 2015, ApJ, 798, 7
  19. Calzetti D., 2001, PASP, 113, 1449
  20. Calzetti D., Kinney A. L., Storchi-Bergmann T., 1994, ApJ, 429, 582
  21. Calzetti D., Armus L., Bohlin R. C., Kinney A. L., Koornneef J., Storchi-Bergmann T., 2000, ApJ, 533, 682
  22. Cappellari M., 2016, ARA&A, 54, 1
  23. Cappellari M., 2017, MNRAS, 466, 798
  24. Cappellari M., Emsellem E., 2004, PASP, 116, 138
  25. Chabrier G., 2003, PASP, 115, 763
  26. Cherinka B., Sánchez-Gallego J., Andrews B., 2017, sdss/marvin: Marvin Beta 2.1.0, doi:10.5281/zenodo.292632, https://doi.org/10.5281/zenodo.292632
  27. Cheung E., et al., 2013, ApJ, 779, 162
  28. Chiappini C., Matteucci F., Romano D., 2001, ApJ, 20, 41
  29. Cid Fernandes R., Stasińska G., Mateus A., Vale Asari N., 2011, MNRAS, 413, 1687
  30. Dressler A., Abramson L., 2014, Proc. IAU, 311, 140
  31. Drory N., et al., 2015, AJ, 149, 77
  32. Emsellem E., Renaud F., Elmegreen B., Gabor J. M., 2015, MNRAS, 446, 2468
  33. Fang J. J., Faber S. M., Salim S., Graves G. J., Rich R. M., 2012, ApJ, 761, 23
  34. Fang J. J., Faber S. M., Koo D. C., Dekel A., 2013, ApJ, 776, 63
  35. Gavazzi G., et al., 2015, A&A, 580, A116
  36. Genzel R., et al., 2015, ApJ, 800, 20
  37. Gladders M. D., Oemler A., Dressler A., Poggianti B., Vulcani B., Abramson L., 2013, ApJ, 770, 64
  38. Goddard D., et al., 2017, MNRAS, 466, 4731
  39. Gonçalves T. S., Martin D. C., Menéndez-Delmestre K., Wyder T. K., Koekemoer A., 2012, ApJ, 759, 67
  40. González Delgado R. M., et al., 2015, A&A, 581, A103
  41. González Delgado R. M., et al., 2016, A&A, 590, A44
  42. Gunn J. E., et al., 2006, AJ, 131, 2332
  43. Hubble E. P., 1936, Realm of the Nebulae. Yale University Press
  44. Kauffmann G., et al., 2003a, MNRAS, 341, 33
  45. Kauffmann G., et al., 2003b, MNRAS, 346, 1055
  46. Kauffmann G., et al., 2007, ApJs, 173, 357
  47. Kennicutt R. C., 1998, ARA&A, 36, 189
  48. Kewley L. J., Dopita M. A., Sutherland R. S., Heisler C. A., Trevena J., 2001, ApJ, 556, 121
  49. Kewley L. J., Groves B., Kauffmann G., Heckman T., 2006, MNRAS, 372, 961
  50. Lackner C. N., Gunn J. E., 2012, MNRAS, 421, 2277
  51. Law D. R., et al., 2015, AJ, 150, 19
  52. Law D. R., et al., 2016, AJ, 152, 83
  53. Lilly S. J., Carollo C. M., Pipino A., Renzini A., Peng Y., 2013, ApJ, 772, 119
  54. Madau P., Dickinson M., 2014, ARA&A, 52, 415
  55. Martig M., Bournaud F., Teyssier R., Dekel A., 2009, ApJ, 707, 250
  56. Martin D. C., et al., 2007, ApJs, 173, 342
  57. Masters K. L., et al., 2010, MNRAS, 405, 783
  58. McDermid R. M., et al., 2015, MNRAS, 448, 3484
  59. Mendel T. J., Simard L., Ellison S. L., Patton D. R., 2013, MNRAS, 429, 2212
  60. Mendez A. J., Coil A. L., Lotz J., Salim S., Moustakas J., Simard L., 2011, ApJ, 736, 110
  61. Mollá M., Díaz A. I., 2005, MNRAS, 558, 521
  62. Morselli L., Popesso P., Erfanianfar G., Concas A., 2017, A&A, 597, A97
  63. Muñoz-Mateos J. C., Gil de Paz A., Boissier S., Zamorano J., Jarrett T., Gallego J., Madore B. F., 2007, ApJ, 658, 1006
  64. Muñoz-Mateos J. C., Boissier S., Gil de Paz A., Zamorano J., Kennicutt, Jr. R. C., Moustakas J., Prantzos N., Gallego J., 2011, ApJ, 731, 10
  65. Nelson E. J., et al., 2016, ApJ, 828, 27
  66. Noeske K. G., et al., 2007, ApJ, 660, L43
  67. O’Connell R. W., 1999, ARA&A, 37, 603
  68. Osterbrock D. E., Ferland G. J., 2006, Astrophysics of gaseous nebulae and active galactic nuclei. University Science Books, Mill Valley, CA
  69. Pan Z., Kong X., Fan L., 2013, ApJ, 776, 14
  70. Pandya V., et al., 2017, arXiv:1611.03869v1
  71. Peng Y.-j., Maiolino R., 2014, MNRAS, 443, 3643
  72. Peng Y.-j., et al., 2010, ApJ, 721, 193
  73. Peng Y.-j., Maiolino R., Cochrane R., 2015, Nature, 521, 192
  74. Pezzulli G., Fraternali F., Boissier S., Munoz-Mateos J. C., 2015, MNRAS, 451, 2324
  75. Renzini A., Peng Y.-j., 2015, ApJ, 801, L29
  76. Saintonge A., et al., 2012, ApJ, 758, 73
  77. Saintonge A., et al., 2016, MNRAS, 462, 1749
  78. Salim S., 2014, Serbian Astron. J., 189, 1
  79. Salim S., Rich R. M., 2010, ApJ, 714, L290
  80. Salim S., et al., 2007, ApJs, 173, 267
  81. Salim S., Fang J. J., Rich R. M., Faber S. M., Thilker D. a., 2012, ApJ, 755, 105
  82. Salpeter E. E., 1955, ApJ, 121, 161
  83. Sánchez-Blázquez P., Rosales-Ortega F., Diaz A., Sanchez S. F., 2014, MNRAS, 437, 1534
  84. Sánchez S. F., et al., 2013, A&A, 554, A58
  85. Sarzi M., et al., 2006, MNRAS, 366, 1151
  86. Schawinski K., Thomas D., Sarzi M., Maraston C., Kaviraj S., Joo S. J., Yi S. K., Silk J., 2007, MNRAS, 382, 1415
  87. Schawinski K., et al., 2009, ApJ, 690, 1672
  88. Schiminovich D., et al., 2007, ApJs, 173, 315
  89. Smee S. A., et al., 2013, AJ, 146, 32
  90. Smethurst R. J., et al., 2015, MNRAS, 450, 435
  91. Strateva I., Ivezic Z., Knapp G. R., Narayanan V. K., Strauss M. a., 2001, AJ, 122, 1861
  92. Thilker D. a., et al., 2007, ApJs, 173, 538
  93. Trayford J. W., Theuns T., Bower R. G., Crain R. A., Lagos C. d. P., Schaller M., Schaye J., 2016, MNRAS, 460, 3925
  94. Trinchieri G., di Serego Alighieri S., 1991, AJ, 101, 1647
  95. Vazdekis A., Ricciardelli E., Cenarro a. J., Rivero-González J. G., Díaz-García L. a., Falcón-Barroso J., 2012, MNRAS, 424, 157
  96. Veilleux S., Osterbrock D. E., 1987, ApJs, 63, 295
  97. Wake D. A., et al., 2017, AJ, 154, 86
  98. White S. D. M., Frenk C. S., 1991, ApJ, 379, 52
  99. Wild V., Almaini O., Dunlop J., Simpson C., Rowlands K., Bowler R., Maltby D., Mclure R., 2017, MNRAS, 463, 832
  100. Wyder T. K., et al., 2007, ApJs, 173, 293
  101. Yan R., Newman J. a., Faber S. M., Konidaris N., Koo D., Davis M., 2006, ApJ, 648, 281
  102. Yan R., et al., 2016a, AJ, 151, 8
  103. Yan R., et al., 2016b, AJ, 152, 197
  104. de Jong R. S., 1996, A&A, 313, 377
Comments 0
Request Comment
You are adding the first comment!
How to quickly get a good reply:
  • Give credit where it’s due by listing out the positive aspects of a paper before getting into which changes should be made.
  • Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements.
  • Your comment should inspire ideas to flow and help the author improves the paper.

The better we are at sharing our knowledge with each other, the faster we move forward.
""
The feedback must be of minimum 40 characters and the title a minimum of 5 characters
   
Add comment
Cancel
Loading ...
102551
This is a comment super asjknd jkasnjk adsnkj
Upvote
Downvote
""
The feedback must be of minumum 40 characters
The feedback must be of minumum 40 characters
Submit
Cancel

You are asking your first question!
How to quickly get a good answer:
  • Keep your question short and to the point
  • Check for grammar or spelling errors.
  • Phrase it like a question
Test
Test description