Colector pluvial en Blvd. Cañaveral

The first consequence of mergers is the alteration of the number density of quiescent galaxies with stellar

masses above M?,lim(Sect. 4.6.1.1), as the accretion of less massive systems efficiently increases the stellar

mass of galaxies below M?,lim. Different merger channels can contribute at different levels, as well as

the mass ratios will have a different impact or weight on the massive quiescent population. As commented above, the presence of a stellar mass–age and –metallicity correlation suggests that mergers are a mechanism to modify the stellar content of massive quiescent galaxies. Our goal is to determine what is the impact due to the observational rate of mergers. These predictions actually set, for the first time, observational constraints to the variations of the average stellar populations of the massive quiescent population, ages and

4.6. The individual effects of mergers, "frosting", and "progenitor" bias on the evolution of the massive quiescent sample 143

metallicities, via mergers (Sects. 4.6.1.2 and 4.6.1.3). We subsequently compare whether these predictions agree with the observations of quiescent stellar populations until z = 0.2 (Chapter 3).

In brief, Figs. 4.1 and 4.2 illustrate the impact of mergers on the number density of massive quiescent galaxies for BC03 and EMILES SSP models, see also Table 4.4. While in Figs. 4.3 and 4.4, we present the impact of mergers on the mass-weighted age and metallicity PDFs respectively, see also Tables 4.5 and 4.6. Details are given in captions and subsections below.

4.6.1.1 Evolution of the number density by mergers

The total balance between the new quiescent massive galaxies (channel "new", see Sect. 4.2.2) owing to mergers and those that leave the massive quiescent sample (channel "out", see Sect. 4.2.4) will shed light on whether mergers play a role in the number density increment observed in the quiescent massive sample since z = 1 (see Sect. 3.6). It is worth recalling again that the merger rates are observationally constrained and they do not follow any theoretical assumption.

In general, there is a good agreement amongst the three SSP model predictions assumed in this work (i. e. BC03 and EMILES with both BaSTI and Padova00 isochrones) in which the channel "new" presents a remarkable contribution to the increase in number of massive quiescent galaxies (see top panels in Figs. 4.1 and 4.2). The largest contribution to the massive quiescent sample of "new" galaxies correspond to major mergers between two quiescent galaxies, where the number density variation varies among SSP models around 0.1–0.2 dex since z = 1. This result is not surprising since the merger rate of quiescent galaxies (principal galaxy in the pair, see Eqs. 4.7 and 4.9) is the largest at z < 1. In addition, the fraction of red com- panions is also larger than the blue ones (respectively 65 and 35 %, see Sect. 4.2.1.3), which also contributes to be the most efficient scenario for the inclusion of new massive quiescent galaxies. Major mergers (mass

ratio 1 ≤ µMM < 1/4) between two quiescent galaxies, whose final stellar mass exceeds the mass limit of

the massive quiescent sample, are more frequent than minor mergers (1/4 ≤ µmm ≤ 1/10), whose number

density variation does not surpass 0.05 dex.For the "new" channel and mergers between galaxies of different spectral types (quiescent as principal and star-forming as companion, and vice versa), the contribution to the number of massive quiescent galaxies is more subtle, lower than 0.03 dex, independently whether this is via major or minor merger (see top panels in Figs. 4.1 and 4.2). Only for EMILES and Padova00 isochrones, major mergers between star-forming and quiescent galaxies show a non-negligible contribution.

The channel "out", or the channel responsible for the decrease in the number density of the massive

quiescent sample (mergers between quiescent galaxies with stellar masses above M?,lim and mergers with

star-forming galaxies with bluer intrinsic colours than mF365 − mF551 < 1.5 as a result), present a non-

negligible contribution to the variation in its number density (see top panels in Figs. 4.1 and 4.2). Major mergers in channel "out" play an important role for diminishing the number density for both merger scenar- ios: between massive quiescent galaxies and with a different spectral type. For both cases, major mergers reduce the number density of the massive quiescent sample ∼ 0.07 dex. As expected, minor mergers in

channel "out" do not modify drastically the number density of quiescent galaxies above M?,lim, as these

variations are below 0.05 dex since z ∼ 1. This is due to the fact that minor mergers with star-forming galaxies do not largely modify the colours of massive quiescent galaxies, and consequently these are still red. On the other hand, minor mergers between massive quiescent galaxies are restricted to the massive-end of the MF, that is, the more restricted in number quiescent galaxies. In fact, minor mergers between massive quiescent galaxies is the least efficient scenario for reducing the number of massive quiescent galaxies.

Regarding those massive quiescent galaxies that suffer a merger and they still belong to the massive quiescent sample, or channel "in", there is a larger number of galaxies that experience mergers with other quiescent galaxies (see middle panel in Figs. 4.1 and 4.2) than those with a star-forming one, as expected from the fraction of blue and red galaxy companions. This means that the variations in the stellar populations of massive quiescent galaxies owing to mergers in channel "in" are mainly driven by the accretion of other less massive quiescent galaxies. Analysing the percentage of massive quiescent galaxies that suffer a merger event since z = 1, we find that minor mergers are slightly more frequent than major mergers. After taking

144 CHAPTER4. The impact of mergers, "frosting" and the "progenitor" bias on the global populations of quiescent galaxies

the predictions of BC03 and EMILES SSP models into account for the merger channel "in", 20–25 % (15–20 %) of massive quiescent galaxies suffer a minor (major) merger with another quiescent galaxy, whereas this percentage reduces down to 4–12 % (3–10 %) when the companion is a star-forming galaxy. EMILES+Padova00 and BC03 present the largest and smallest percentages respectively. In both scenarios and model predictions massive quiescent galaxies suffer a larger number of major mergers at z & 0.6, however at lower redshifts, the number of minor mergers dominates.

Focusing in the overall contribution of mergers to the increase in number of massive quiescent galaxies, see bottom panel in Figs. 4.1 and 4.2 and Table 4.4, we observe that the main mass ratios responsible for the changes in density number are major merger. The impact of minor mergers in the number density is less remarkable since z ∼ 1. Notice that minor mergers do not largely alter the number densities of massive quiescent galaxies, since the number of new galaxies compensates those going out from the massive quiescent sample. This means that there a mild flow of galaxies entering and exiting that can modify the global stellar populations of this massive sample (not accounting for the effects of channel "in") without altering their number density. In major mergers, the channel "new" is more dominant than the "out" one, driving to the increase in the number density. For BC03 and EMILES+Padova00, the evolution in number density is entirely explained by the merger rates observed in the Universe within the uncertainties. For EMILES+BaSTI, there are not enough mergers to explain the evolution in their number density.

4.6.1.2 Variations in the age PDFs by mergers

Minor and major mergers are expected to have a non-negligible impact in the global stellar populations of massive quiescent galaxies. For the first time, we explore this effect, which partly remain unnoticed owing to the continuous flow of galaxies with low alterations in the number density. In fact, the influence of the use of different SSP models for retrieving stellar population parameters via SED-fitting will be more remarkable in the ages and metallicities than in the number density (Sect. 4.6.1.1), where the dependence of the use of different SSP models was limited to the retrieved MFs.

As expected from the stellar mass-age relation, the new members coming from the merge of two less massive systems with younger stellar populations produces a general rejuvenation of massive quiescent galaxies, as long as the number of mergers is large enough. We confirm from the three sets of SSP models, see Table 4.5 and Fig. 4.3, that there are enough mergers as to modify the median age of the whole massive quiescent population by 0.5–1.2 Gyr since z = 1.0 up to z = 0.2. In addition, the variety of merger channels (channels "new", "out", and "in") with diverse stellar masses and spectral types also induces an increase of the width of the distributions of mass-weighted ages, and therefore, the distribution of ages becomes wider at lower redshifts. Both effects, a systematic rejuvenation and a wider range of ages, are determined in this thesis using the ALHAMBRA data and the results provided by MUFFIT (Sects. 3.7 and 3.8), although the delay in the ageing (respect a passive evolution) is greater than the predicted by our merger model. While the analysis of the stellar populations of these galaxies in the same redshift range, 0.2 ≤ z ≤ 1.0, presents stellar populations younger by 3.7, 1.6, and 2.8 Gyr than the expectations for a passive evolution (BC03, EMILES+BaSTI, and Padova00 respectively), these are much larger than the predicted by the merger model, which are around ∼ 1 Gyr. Therefore, the merger mechanism by itself does not seem to be able to explain the rejuvenation of massive quiescent galaxies according to the three SSP model predictions. However, the evolution of the width of the mass-weighted age PDF owing to mergers, 1–2 Gyr, is compatible with the one measured by MUFFIT in ALHAMBRA, 1 Gyr.

We are in a privilege position to discern the role of the mass ratio in the process of rejuvenation via mergers, because a priori more massive systems are more efficient to modify the stellar content of a massive galaxy but these contain more similar stellar populations. On the other hand, the accretion of a less massive systems has a lower impact in mass, but their stellar populations differ more from the ones of the massive galaxies owing to the stellar mass-age relation. We find that the variation of the median values of mass- weighted ages are weaker for the case of minor mergers (< 0.5 Gyr) than for the major ones (< 1 Gyr), although the differences are very subtle (see Table 4.5 and Fig. 4.3). EMILES+BaSTI predictions present

4.6. The individual effects of mergers, "frosting", and "progenitor" bias on the evolution of the massive quiescent sample 145

the lowest modifications in the median age, and the major and minor contribution are basically the same. Regarding the width of the distribution of age values, major mergers imprint larger variations than minor ones do, although these differences are minimal. The increase in width of the mass-weighted age PDF, splitting major from minor merger effects, are around ∼ 1.5 Gyr.

Thereby, we conclude that mergers drive a slightly process of rejuvenation of 1 Gyr from z = 1 to z = 0.2 and increase the range of mass-weighted age values in the population of massive quiescent galaxies. Despite this, the rejuvenation observed from ALHAMBRA data cannot be fully explained by mergers, and it is necessary the inclusion of other mechanism to reproduce its evolution. Nevertheless, the evolution in the width of the age PDF can be reproduced in an scenario in which mergers are present.

4.6.1.3 Variations in the metallicity PDFs by mergers

We study the global effects of major and minor mergers in the evolution of metallicities in massive quiescent galaxies. In this case, there are larger discrepancies in the metallicity predictions of quiescent galaxies owing to the use of different SSP models (see Sects. 3.7 and 3.8).

The effects of mergers on the mass-weighted metallicity PDF is a general decrement of the median val- ues (see Table 4.6 and Fig. 4.4) since z = 1.0. In particular, BC03 SSP models present the lower decrement in metallicity, whereas the EMILES predictions show a more prominent evolution due to mergers. In fact, the evolution of the median of the mass-weighted metallicity PDF derived from BC03 SSP models in Chap- ter 3, exhibits a maximum at z ∼ 0.6 that is at odds with the mild trend obtained for mergers. Notice that the presence of a maximum in the median metallicity at z . 0.6, makes the merger model to predict a very soft increase in metallicity. For EMILES, the decrement in the median of metallicity is more remarkable and continuous, ∼ 0.10 dex in both cases, than in BC03. In addition, the predictions about the evolution of the median of the mass-weighted metallicity PDF obtained with the merger model agree better with the observed ones by MUFFIT and ALHAMBRA data (a decrement of 0.1–0.2 dex). In particular, Padova00 isochrones ask for a variation in the median metallicity slightly larger than observed. Consequently, mergers are able to reproduce the decrement in the medians of the metallicity PDFs without the necessity of extra mechanisms. Except for BC03, in which the width of the mass-weighted metallicity PDF remains almost constant, mergers also increase the width of the distributions of metallicity values of massive quiescent galaxies since z = 1. Indeed, the increase in the width of the metallicity PDF predicted by our merger model (. 0.1 dex) is in good agreement with those obtained observationally by MUFFIT and ALHAMBRA (∼ 0.1 dex).

Treating the effects of major and minor mergers separately (see Table 4.6 and Fig. 4.4), we find that ma- jor mergers show similar contributions to modify both the medians and widths of mass-weighted metallicity PDFs. For the medians this is ∼ 0.1 dex for both EMILES isochrones, and almost negligibles for BC03. The same result is obtained for minor mergers. In the same sense, the effect of mergers with BC03 SSP models on the widths of the metallicity PDFs persists unnoticed, whereas for EMILES these are ∼ 0.1 dex for both isochrones and for major and minor mergers. Consequently, minor and major mergers contribute similarly to the global metallicity evolution since z = 1 to z = 0.2. Notice that the effects of minor mergers is more remarkable at lower redshifts, and at the larger ones major mergers played a slightly greater role.

To conclude, mergers can naturally explain the slight decrease in the observed metallicity distributions of massive quiescent galaxies in ALHAMBRA (not for BC03) since z = 1.