Tracing the Quenching History in Galaxy Clusters in the EAGLE Simulation
Diego Pallero, Facundo A. Gómez, Nelson D. Padilla, S. Torres-Flores, R. Demarco, P. Cerulo, D. Olave-Rojas
MMNRAS , 1–14 (2015) Preprint 25 June 2019 Compiled using MNRAS L A TEX style file v3.0
Tracing the Quenching History of Cluster Galaxies in theEAGLE Simulation
Diego Pallero (cid:63) , Facundo A. G´omez , , Nelson D. Padilla , , S. Torres-Flores ,R. Demarco , P. Cerulo and D. Olave-Rojas Departamento de F´ısica y Astronom´ıa, Universidad de La Serena, Av. Juan Cisternas 1200 Norte, La Serena, Chile Instituto de Investigaci´on Multidisciplinar en Ciencia y Tecnolog´ıa, Universidad de La Serena, Ra´ul Bitr´an 1305, La Serena, Chile Instituto de Astrof´ısica, Pontificia Universidad Cat´olica de Chile, Santiago, Chile Centro de Astro-Ingenier´ıa, Pontificia Universidad Cat´olica de Chile, Santiago, Chile Departamento de Astronom´ıa, Facultad de Ciencias F´ısicas y Matem´aticas, Universidad de Concepci´on, Concepci´on, Chile Departamento de F´ısica, Facultad de Ciencias, Universidad del B´ıo B´ıo, Collao 1202, Concepci´on, Chile
Accepted XXX. Received YYY; in original form ZZZ
ABSTRACT
We use the EAGLE hydrodynamical simulation to trace the quenching history ofgalaxies in its 10 most massive clusters. We use two criteria to identify moments whengalaxies suffer significant changes in their star formation activity: i) the instantaneousstar formation rate (SFR) strongest drop, Γ SDSFR , and ii) a“quenching”criterion based ona minimum threshold for the specific SFR of (cid:46) − yr − . We find that a large fractionof galaxies ( (cid:38) ) suffer their Γ SDSFR outside the cluster’s R . This “pre-processed”population is dominated by galaxies that are either low mass and centrals or inhabitlow mass hosts ( . M (cid:12) (cid:46) M host (cid:46) . M (cid:12) ). The host mass distribution is bimodal,and galaxies that suffered their Γ SDSFR in massive hosts ( . M (cid:12) (cid:46) M host (cid:46) . M (cid:12) )are mainly processed within the clusters. Pre-processing mainly limits the total stel-lar mass with which galaxies arrive in the clusters. Regarding quenching, galaxiespreferentially reach this state in high-mass halos ( . M (cid:12) (cid:46) M host (cid:46) . M (cid:12) ). Thesmall fraction of galaxies that reach the cluster already quenched has also been pre-processed, linking both criteria as different stages in the quenching process of thosegalaxies. For the z = satellite populations, we find a sharp rise in the fraction ofquenched satellites at the time of first infall, highlighting the role played by the densecluster environment. Interestingly, the fraction of pre-quenched galaxies rises with fi-nal cluster mass. This is a direct consequence of the hierarchical cosmological modelused in these simulations. Key words: galaxies:clusters: general – galaxies: evolution – galaxies: formation –galaxies: star formation – galaxies: haloes
Since the first half of the twentieth century, it has beenknown that colors reflect the predominant stellar popula-tions in galaxies and that they are related to their mor-phology (Morgan & Mayall 1957). The colour-morphology(Roberts & Haynes 1994) and color-magnitude relations(Chester & Roberts 1964; Faber 1973) are now widely usedto study the properties of galaxies. As a result, rather thanselecting objects according to their early- or late-type mor- (cid:63) E-mail:[email protected] phology, galaxies can be separated between red and blue ,which naturally relates with their star formation and metal-enrichment history. Studies in the local universe show that,in general, galaxies present a strong bimodal color distri-bution (Strateva et al. 2001; Baldry et al. 2006; Cassataet al. 2008), regardless of the environment in which they re-side (Hogg et al. 2004; Baldry et al. 2006). Reproducing thisbimodality, and understanding the role played by the envi-ronment, has become an important goal for galaxy-evolutiontheories (Trayford et al. 2015; Nelson et al. 2018).One of the first indications that the environment playsa fundamental role in driving the evolution of galaxies was c (cid:13) a r X i v : . [ a s t r o - ph . GA ] J un D. Pallero et al. the morphology-density relation (Dressler 1980, 1984). Ob-servational studies have shown that in high density environ-ments there is a greater fraction of galaxies with early-typemorphology than in low density environments, and that thefraction of early-type galaxies in clusters rises toward thecluster’s center (Brough et al. 2017; Cava et al. 2017).In addition, several studies during the last decades haveshown that dense environments can also affect the star for-mation history of galaxies (Gunn & Gott 1972; Dressler1980; Moore et al. 1996; Poggianti et al. 2001; Boselli et al.2005). Naturally, the cores of galaxy clusters are an ideallaboratory to study how the environment affects the evolu-tion of galaxies in dense regions and at different redshifts(Cayatte et al. 1990; Smail et al. 1997; Bravo-Alfaro et al.2000; Boselli et al. 2005). Evidence of global transformationsfor galaxies over time is given by the increasing fraction ofspiral in clusters up to z ∼ ∼ − of galaxies belongingto a massive cluster (M halo ∼ . − . [M (cid:12) ] ) at z = R clusters arericher in passive galaxies than the field. These results canonly be explained if star formation was quenched in galax-ies prior to their accretion on to clusters, when they werestill members of in-falling groups (pre-processing Zabludoff& Mulchaey 1998; Fujita 2004). The fact that galaxies are“preprocessed” in groups before their accretion on to clustersshows that groups of galaxies constitute an important piecein the physics of galaxy formation and evolution, becausethe processes that take place within them may significantlyalter star formation and change the structural and chemicalproperties of galaxies. Groups provide, then, further labora-tories to study the environmental drivers of galaxy evolution(e.g. Dressler et al. 2013; Bianconi et al. 2018; Olave-Rojaset al. 2018).At z ∼
0, the specific star-formation rate of galaxiesin dense environments is significantly smaller than in lowerdensity regions (Hashimoto et al. 1998; Lewis et al. 2002;Kauffmann et al. 2004; Gray et al. 2004; Balogh et al. 2007).Additionally, higher fractions of quiescent or passive galax- ies are found in dense regions (Poggianti et al. 1999; Baldryet al. 2006; van den Bosch et al. 2008; Gavazzi et al. 2010;Haines et al. 2013). These studies also provide evidence thatthe star-forming activity and galaxy morphology can be cor-related with z ∼ “environmentalquenching” , it is important to take into account internal pro-cess that can drive galaxy quenching. This process, knownas “mass quenching” or “internal quenching” , can arise as aresult of, e.g. internal gas consumption, supernova and AGNfeedback, star formation feedback or halo gas heating (seee.g. Peng et al. 2010, Efstathiou 2000, Croton et al. 2006,Dekel & Birnboim 2008, Cantalupo 2010). The dominanceof one way over the other is where the dichotomy of “natureversus nurture” was born, and has been one of the main sub-jects of study for extragalactic astronomy in the last years.According to Oesch et al. (2016), quenching may startshortly after the first appearance of the galaxies, at roughlyz ∼
11, but the environment does not play an importantrole until z ∼ ∼ ∼ ∼ ∼ > ∼
0. On the one hand, at high redshift, the cessationof star formation is mainly driven by a combination of gasconsumption (due to an enhancement of star formation) andgas outflows as a result of supernovae and AGN feedback.On the other hand, at low redshift, dynamical mass removalmechanisms (due to environment) may be the main driverfor the quenching of galaxies in clusters.A detailed description of the main mechanisms that leadto environmental quenching is provided by Boselli & Gavazzi(2006) and Jaff´e et al. (2016). They separate these mecha-nisms in three broad categories: • Gravitational interactions between galaxies: Mergerscan change drastically the star formation history of galax-ies, as well as their morphology and kinematics. This phe-nomenon is usually observed in low-density environmentsuch as groups of galaxies. In high-density environments,galaxies can experience harassment from other cluster mem-bers, through fast and aggressive encounters (Toomre &Toomre 1972; Barnes & Hernquist 1996; Walker et al. 1996;Moore et al. 1999); • Interactions between galaxies and the intra-clustermedium: Ram-pressure from the intracluster medium canstrip the gas of the galaxies and remove their interstellarmedium (Gunn & Gott 1972; Abadi et al. 1999; Quilis et al.
MNRAS000
MNRAS000 , 1–14 (2015) uenching history of Cluster Galaxies Figure 1.
Examples of the selection criteria used in this work to determine whether a galaxy was processed or quenched, for a randomgalaxy in our sample. Panel (a) shows the star formation rate against redshift. The red ellipse highlights the strongest drop in the starformation activity, and the blue dashed line corresponds to the time of the first infall into the final cluster’s R . Panel (b) shows thespecific star formation rate of the galaxy against redshift. The red solid line shows the critical star formation rate imposed by our selectioncriterion to define quenched galaxies. The blue dashed line corresponds to the first infall into R . Panel (c) shows the growth of thegalaxy stellar mass through cosmic time. The blue star indicates the moment when the processing started. We can see that the growthof the stellar mass is suppressed after the strongest drop, and that the specific star formation rate decreases abruptly after passing R . • Gravitational interactions between clusters and galax-ies: The tremendous gravitational potential of the clustercan perturb some observable properties of the members, in-ducing gas inflows, forming bars, compressing the gas or con-centrating the star formation (Miller 1986; Byrd & Valtonen1990; Boselli & Gavazzi 2006).It is precisely because of this complex nature of theenvironmental quenching that it is difficult to separate theaforementioned processes. It is expected that, at least, someof these processes act simultaneously and that they are ef-fective in different overlapping regions of the cluster. Somestudies show that the effectiveness of these processes islinked to the galaxy’s cluster-centric distance (Moran et al.2007).A good approach to study the mechanisms that im-pact galaxy evolution is through cosmological models (Fu-jita 2004; Wetzel et al. 2013; Vijayaraghavan & Ricker 2013;Schaye et al. 2015; Nelson et al. 2018). Several works haveused simulations to understand the properties of galaxies indifferent environments, and how their evolutionary historyresults in changes of their properties such as colors, stel-lar mass and star formation rate (Trayford et al. 2015, 2016;Katsianis et al. 2017; Tescari et al. 2018; Nelson et al. 2018).Hydrodynamical simulations can be used to define and testdifferent criteria that can be used to understand the pro-cesses that drive galaxies to be quenched. Simulations alsoallow us to follow the evolution of galaxies in different en-vironments and the evolution of their properties from z ∼
20 to z = 0. Since clusters at z > . are difficult to detect,due to the fact that they are still in an assembling process, simulations are a helpful tool to study the role that the envi-ronment plays at such high redshift (see e.g. Overzier 2016).In this paper, we use the public database from the state-of-the-art EAGLE hydrodynamic simulations (Schaye et al.2015; Crain et al. 2015; McAlpine et al. 2016) to trace theevolution history of the satellite galaxies that belong to theten most massive clusters at z ∼
0. We aim at identifyingthe environment in which galaxies preferentially cease theirstar formation and signatures that could be used to deter-mine the main physical mechanism leading to the cessationof star formation of cluster satellite galaxies. We comparethe results obtained from two different criteria to identifywhen star formation in galaxies significantly drops. The hy-drodynamic simulations of the EAGLE project are perfectlysuited for this study since they provide the possibility tostudy the evolution of galaxies and their properties.This paper is organized as follows: in Section 2 we de-scribe the EAGLE simulation, its main characteristics andthe main potentialities that it provides for this study. In Sec-tion 3 we define the two criteria used in this work to locatethose moments when galaxies suffer an important variationin their star formation; in Section 4 we describe the resultsobtained using our two approaches, putting special intereston the environment where these events take place. Finally, inSection 5 we summarize our main conclusions and compareour results with both observational and theoretical works.A brief discussion of some future projects are presented inthis section as well.
MNRAS , 1–14 (2015)
D. Pallero et al.
Figure 2.
Distribution of Γ SDSFR (normalized instantaneous strongest drop of the SF activity), as a function of the redshift at which ittakes place. Blue dots correspond to galaxies that suffer their Γ SDSFR outside the cluster’s R (pre-processed), while red dots to thosethat suffer their Γ SDSFR inside R (processed in-situ). The medians for both samples are indicated by the dashed lines. The panels areorganized from left to right as high-mass clusters ( . < log M host [ M (cid:12) ] < . ), intermediate-mass clusters ( . < log M host [ M (cid:12) ] < . )and low-mass clusters ( . < log M host [ M (cid:12) ] < . ), respectively. The EAGLE project, is a suite of cosmological hydrodynam-ical N-body simulations. These simulations were run with amodified version of the GADGET-3 code, wich is an im-proved version of GADGET-2 (Springel 2005). All the sim-ulations adopt a flat Λ CDM cosmology whose parameterswere calibrated with the data obtained by the Planck mis-sion (Planck Collaboration et al. 2014); Ω Λ = 0.693, Ω m =0.307, Ω b = 0.04825, σ = 0.8288, Y = 0.248 and H = 67.77km s − .In particular, for this work we select our sample of galaxiesfrom the main simulation, referred to as L100N1504, whichconsists of a periodic box with a volume of (100cMpc) , ini-tially containing 1,504 gas particles with an initial mass of1.81 × M (cid:12) , and the same amount of dark matter particleswith a mass of 9.70 × M (cid:12) .Each simulation counts with 29 discrete snapshots from red-shift 20 to 0, with a time span between consecutive snapshotsranging from 0.3 to 1Gyr. Radiative cooling and photoheat-ing are implemented following Wiersma et al. (2009a), as-suming an optically thin X-Ray/UV background (Haardt& Madau 2001). Star formation is implemented stochasti-cally following Schaye & Dalla Vecchia (2008), and usingthe metallicity-dependent density threshold shown in Schaye(2004). This reproduces the observed Kennicutt-Schmithlaw (Kennicutt 1998). Each particle is assumed to be asingle-age stellar population, with a Chabrier initial massfunction in the range 0.1 M (cid:12) - 100M (cid:12) (Chabrier 2003).Stellar evolution is modelled as shown in Wiersmaet al. (2009b), and chemical enrichment is followed forthe 11 elements that most contribute to radiative cool-ing from massive stars (Type II supernovae and stellarwinds) and intermediate-mass stars (Type Ia supernovaeand AGB stars). Following Dalla Vecchia & Schaye (2012),the thermal-energy product of stellar feedback is stochas-tically distributed among the gas particles surrounding theevent without a preferential direction.The EAGLE project calibrated the free parameters as- sociated with stellar feedback to match the observations forthe stellar mass function in a range of 10 M (cid:12) - 10 M (cid:12) andthe size-mass relation for galaxies in a range of 10 M (cid:12) -10 M (cid:12) (Schaye et al. 2015; Furlong et al. 2015, 2017). Theappropriate calibration of the subgrid physics and the goodagreement with the observational data make these simula-tions our best tool to study the evolution in the star forma-tion of galaxies in these mass ranges for different environ-ments.The halo catalogues provided in the public database(used in this work) were built using a friend-of-friends (FoF)algorithm which identifies dark matter overdensities follow-ing Davis et al. (1985), considering a linking length of 0.2times the average inter-particle spacing. Baryonic particlesare assigned to the FoF halo of their closest dark matter par-ticle. Subhalo catalogues were built using the subfind algo-rithm (Springel et al. 2001; Dolag et al. 2009), which iden-tifies local overdensities using a binding energy criterion forparticles within a FoF halo. We will define as galaxies thosestructures recognized by the subfind algorithm (Springelet al. 2001; Dolag et al. 2009) which posses a total stellarmass greater than 10 M (cid:12) and a total mass greater than 10 M (cid:12) . These masses are obtained by direct summation of thecorresponding particles; i.e., particles bound to the subhaloaccording to the subfind . Since the simulations consideredin this work have a baryonic mass resolution of . × M (cid:12) and dark matter mass resolution of . × M (cid:12) , we en-sure that we have at least 100 baryonic and dark matterparticles in each galaxy, thus avoiding spurious results andnon-physical detections. The analyzed clusters correspondto the 10 most massive in the simulation at z = . They allposses M > M (cid:12) . A galaxy is defined as satellite if itcan be found inside the host R at z = . MNRAS000
Distribution of Γ SDSFR (normalized instantaneous strongest drop of the SF activity), as a function of the redshift at which ittakes place. Blue dots correspond to galaxies that suffer their Γ SDSFR outside the cluster’s R (pre-processed), while red dots to thosethat suffer their Γ SDSFR inside R (processed in-situ). The medians for both samples are indicated by the dashed lines. The panels areorganized from left to right as high-mass clusters ( . < log M host [ M (cid:12) ] < . ), intermediate-mass clusters ( . < log M host [ M (cid:12) ] < . )and low-mass clusters ( . < log M host [ M (cid:12) ] < . ), respectively. The EAGLE project, is a suite of cosmological hydrodynam-ical N-body simulations. These simulations were run with amodified version of the GADGET-3 code, wich is an im-proved version of GADGET-2 (Springel 2005). All the sim-ulations adopt a flat Λ CDM cosmology whose parameterswere calibrated with the data obtained by the Planck mis-sion (Planck Collaboration et al. 2014); Ω Λ = 0.693, Ω m =0.307, Ω b = 0.04825, σ = 0.8288, Y = 0.248 and H = 67.77km s − .In particular, for this work we select our sample of galaxiesfrom the main simulation, referred to as L100N1504, whichconsists of a periodic box with a volume of (100cMpc) , ini-tially containing 1,504 gas particles with an initial mass of1.81 × M (cid:12) , and the same amount of dark matter particleswith a mass of 9.70 × M (cid:12) .Each simulation counts with 29 discrete snapshots from red-shift 20 to 0, with a time span between consecutive snapshotsranging from 0.3 to 1Gyr. Radiative cooling and photoheat-ing are implemented following Wiersma et al. (2009a), as-suming an optically thin X-Ray/UV background (Haardt& Madau 2001). Star formation is implemented stochasti-cally following Schaye & Dalla Vecchia (2008), and usingthe metallicity-dependent density threshold shown in Schaye(2004). This reproduces the observed Kennicutt-Schmithlaw (Kennicutt 1998). Each particle is assumed to be asingle-age stellar population, with a Chabrier initial massfunction in the range 0.1 M (cid:12) - 100M (cid:12) (Chabrier 2003).Stellar evolution is modelled as shown in Wiersmaet al. (2009b), and chemical enrichment is followed forthe 11 elements that most contribute to radiative cool-ing from massive stars (Type II supernovae and stellarwinds) and intermediate-mass stars (Type Ia supernovaeand AGB stars). Following Dalla Vecchia & Schaye (2012),the thermal-energy product of stellar feedback is stochas-tically distributed among the gas particles surrounding theevent without a preferential direction.The EAGLE project calibrated the free parameters as- sociated with stellar feedback to match the observations forthe stellar mass function in a range of 10 M (cid:12) - 10 M (cid:12) andthe size-mass relation for galaxies in a range of 10 M (cid:12) -10 M (cid:12) (Schaye et al. 2015; Furlong et al. 2015, 2017). Theappropriate calibration of the subgrid physics and the goodagreement with the observational data make these simula-tions our best tool to study the evolution in the star forma-tion of galaxies in these mass ranges for different environ-ments.The halo catalogues provided in the public database(used in this work) were built using a friend-of-friends (FoF)algorithm which identifies dark matter overdensities follow-ing Davis et al. (1985), considering a linking length of 0.2times the average inter-particle spacing. Baryonic particlesare assigned to the FoF halo of their closest dark matter par-ticle. Subhalo catalogues were built using the subfind algo-rithm (Springel et al. 2001; Dolag et al. 2009), which iden-tifies local overdensities using a binding energy criterion forparticles within a FoF halo. We will define as galaxies thosestructures recognized by the subfind algorithm (Springelet al. 2001; Dolag et al. 2009) which posses a total stellarmass greater than 10 M (cid:12) and a total mass greater than 10 M (cid:12) . These masses are obtained by direct summation of thecorresponding particles; i.e., particles bound to the subhaloaccording to the subfind . Since the simulations consideredin this work have a baryonic mass resolution of . × M (cid:12) and dark matter mass resolution of . × M (cid:12) , we en-sure that we have at least 100 baryonic and dark matterparticles in each galaxy, thus avoiding spurious results andnon-physical detections. The analyzed clusters correspondto the 10 most massive in the simulation at z = . They allposses M > M (cid:12) . A galaxy is defined as satellite if itcan be found inside the host R at z = . MNRAS000 , 1–14 (2015) uenching history of Cluster Galaxies According to Peng et al. (2010), the quenching of a galaxyis the result of a process with two different components.A continuous component associated with internal galacticprocesses such as star formation and AGN feedback, anda “once-only” component due to environmental processes.Note, however, that other mechanisms like mergers may alsohave an important effect on the star formation activity.To determine the moment when the star formation ac-tivity in a galaxy drops in a significant way, two differentcriteria are introduced: one based on the maximum dropof the SFR between two consecutive snapshots of the sim-ulation, and the other based on a minimum threshold forspecific star formation rate (sSFR). The first criterion seeksto identify those mechanisms that abruptly reduce star for-mation in galaxies, while the second one is meant to definewhen a galaxy is actually quenched, that is, it is no longerforming stars (e.g. Weinmann et al. 2010; De Lucia et al.2012; Wetzel et al. 2012). The aim of using these two cri-teria is to determine and understand the different stagesof quenching and how they are affected by the environment.From now on, we will refer to a galaxy as “processed” when itsuffers its strongest drop, whereas we will refer to a galaxy as“quenched” when it reaches the imposed threshold in sSFR.
One of our goals is to identify the mechanisms that canabruptly reduce the star formation in galaxies. For this pur-pose, we first calculate for each galaxy the variation of thestar formation rate between two consecutive snapshots in thesimulation, normalized by the star formation in the earliestsnapshot. This is Γ SFR = SFR i + − SFR i SFR i , (1)where the subscript i indicates the simulation snapshot, and i + is at a lower redshift than i . Γ SFR is computed only ifthe difference between the SFR value in the two snapshots islarger than × − M (cid:12) yr − . This constraint was imposedto avoid measures of Γ SFR for galaxies already quenched. fWethen define Γ SDSFR as the fraction of star formation lost at themoment when the strongest drop occurs, i.e. Γ SDSFR = | min Γ SFR | (2)We refer to this method as the “Strongest drop selec-tion criterion”. Γ SDSFR takes into account those episodes whena “once-only” event affects the star formation activity of thegalaxies but does not take into account any rejuvenationscenario that could take place afterwards. For this reason,it is not a good tracer of definitive quenching. However, theinformation gathered by this criterion allows us to find theepochs at which the galaxy suffers a “processing” event, inparticular the most significant one. An example of this selec-tion criterion is shown in Figure 1, panel (a), where we plot,as a function of time, the star formation rate of a randomgalaxy in our sample. The red ellipse highlights the momentwhen Γ SDSFR takes place. In particular for this galaxy, the Γ SDSFR is the result of several processes that heat and remove its cold gas content, producing a stagnation in the evolution ofthe stellar mass and a small decrease in the total gas massof the galaxy. Unfortunately, we cannot isolate the differentmechanisms that produce this processing event due to thelack of temporal resolution. We will further explore this ina future work using a better suited simulation.
We further wish to define a criterion that aims at identify-ing the moment when the galaxies reach a definitive stateof “quenching”. Several different definitions of “quenchedgalaxy” have been proposed in the literature. Here we usedthe criterion used in Wetzel et al. (2013). According to thiscriterion, a galaxy can be considered effectively quenchedonce it reaches a sSFR Q = − yr − . At this point thegalaxy is considered to be passive. From now on we willrefer to those galaxies with a sSFR lower than sSFR Q as“quenched galaxies”, and we will call this selection crite-rion the “Critical sSFR Selection Criterion”. When using thissemi-observational definition, we will only focus on galaxiesthat are quenched at redshift z = . This is to ensure thatthe selected galaxies will not suffer a rejuvenation processduring their evolution. From each of our quenched galaxies,we will extract information about the environment and thetime when the quenching state is reached.An example of this selection criterion is shown in Fig-ure 1, panel (b), where the sSFR is shown for the samegalaxy from the previous example as a function of time. Thered line indicates the sSFR threshold established in previousworks (Weinmann et al. 2010; De Lucia et al. 2012; Wetzelet al. 2012, 2013) for passive galaxies. In particular for thegalaxy shown in the example, the critical star formation isreached once it crosses the R of the cluster for the firsttime, showing the importance of dense environments in thequenching of star formation. We wish to study the dependencies of star formation quench-ing on environmental and internal processes focusing ondense environments such as those that can be found ingalaxy clusters. For this it is necessary to characterize theproperties of individual galaxies such as stellar mass, sSFRand total mass, as well as the overall properties of the hostcluster such as total mass and virial radius. We will studyhow these properties evolve as a function of time, and fo-cus on those moments where individual galaxies experiencesharp falls in their star formation rates.As previously discussed in Section 1, in this work wefocus on the population of galaxies associated with the 10most massive clusters of the EAGLE simulations. In orderto study properties of these galaxies as a function of clustersmass with better statistics, the clusters were stacked in threedifferent bins of z = total mass: • high mass: . < log M host [ M (cid:12) ] < . , • intermediate mass: . < log M host [ M (cid:12) ] < . , • low mass: . < log M host [ M (cid:12) ] < . .We will refer to these three categories as HMC, IMC andLMC, respectively. The numbers of cluster that fall in each MNRAS , 1–14 (2015)
D. Pallero et al.
Figure 3.
Mass distribution of galaxies and their hosts at key moments related to the Γ SDSFR . Each row shows the results obtained afterstacking the distribution of galaxies associated with clusters within different mass ranges. Column (a) shows the mass distribution ofthe host of each galaxy at the moment of their Γ SDSFR . Column (b) shows the total mass distribution of the galaxies at their Γ SDSFR . Column(c) shows the stellar mass fraction distributions at Γ SDSFR . Column (d) shows the stellar mass distribution of galaxies at the time of theirfirst infall into the cluster they belong to at z = . Blue, red and greed bars correspond to galaxies pre-processed, in-situ processed, andprocessed as centrals, respectively. The dashed lines indicate the median of each distribution. bin are 2 for the HMC, 5 for the IMC and 3 for the LMC. Inthis section we present our results based on the two previ-ously defined criteria to identify the time at which the starformation activity of a galaxy is significantly altered. Wewill use the terms pre and in-situ for galaxies that suffer thepreviously described processes inside or outside the clusterR , respectively. We first focus on abrupt changes in the SF activity. Westart by computing Γ SDSFR for all galaxies that belong to the10 most massive clusters at z = . Our goal is to assess whereand when they suffer their most significant processing event.The total number of galaxies in the HMC, IMC and LMCbins are N gal = 846, N gal = 1430 and N gal = 421, respectively.Note that the differences in the number of galaxies is mainlydue to the number of clusters that fall in each mass bin.In Figure 2 we show Γ SDSFR for all galaxies as a func-tion of the redshift at which this event takes place. Theblue and red dots correspond to pre- and in-situ processedgalaxies, respectively. The different panels show the resultsfor the different mass bins. We can clearly see that, for thepre-processed population, there is no preferential redshift for Γ SDSFR to take place. Note as well that there is no clearcorrelation between redshift and the typical value of Γ SDSFR for both populations. This indicates that these “once-only”events that significantly affect star formation activity arenot associated with any preferential epoch.In all mass bins, the majority of the galaxies have beenpre-processed. Interestingly, for in-situ processed galaxies, Γ SDSFR typically occurs at lower values of redshift than for pre-processed galaxies. This can be seen from the dashed verticallines, which indicate the median redshift for each population.Note as well that the pre-processed fraction grows with clus-ters mass, but the median in redshift for pre-processing re-mains the same regardless of the mass bin. This shows that,although more massive clusters accrete a greater number ofpre-processed galaxies, the redshift at which Γ SDSFR typicallytakes place is independent of the z = mass of the clustersin which galaxies reside.To understand how the processing affects the evolutionof galaxies and which is the role played by the environment,we characterize the mass distribution of the hosts in whichthese galaxies resided when they suffered their Γ SDSFR . In Fig-ure 3, panels a), we show the fraction of galaxies per bin ofhost halo mass, M host , at the time of Γ SDSFR . Fractions are ex-pressed with respect to the total galaxy sample. We split our
MNRAS000
MNRAS000 , 1–14 (2015) uenching history of Cluster Galaxies Figure 4.
Mass distribution of galaxies and their hosts at the moment when they reach their quenching state. Each row shows theresults obtained after stacking the distribution of galaxies associated with clusters within different mass ranges. Column (a) shows themass distribution of the host of each galaxy. Column (b) shows distribution of galaxies’ total mass. Column (c) shows the stellar massfraction distributions. Column (d) shows the distribution of times, in lookbacktime, at which galaxies become quenched. Blue, red andgreed bars correspond to galaxies pre-quenched, in-situ quenched, and quenched as centrals, respectively. The dashed lines indicate themedian of each distribution. sample in three populations: pre-processed central galaxies,pre-processed satellites, and in-situ processed satellites. Notethat for the pre-processed central population, the mass of thehost where the galaxies reside at the moment of processingis nearly the mass of the galaxy itself. From this panel wecan clearly see that the three populations are well separatedin the host mass distribution, regardless of the mass of thecluster. The median M host of each population is indicatedwith dashed lines. As we can see, according to the criterion Γ SDSFR , central pre-processed galaxies tend to suffer their Γ SDSFR in low-mass halos, preferentially in halos with total mass be-tween . (cid:46) M host [M (cid:12) ] (cid:46) . . For galaxies pre-processedas satellites, Γ SDSFR occurs in a large variety of halo masses,ranging between (cid:46) M host [M (cid:12) ] (cid:46) . , with a mediannear . M host [M (cid:12) ] regardless of the mass of the cluster(the typical mass of galaxy groups). On the other hand, forthe in-situ processed, it preferentially occurs in higher masshalos, with total masses larger than . M host [M (cid:12) ] .To explore the relation between Γ SDSFR and environmentwe compute, for the overall processed galaxy population, thedistribution of total mass (M galaxy ) and the stellar mass frac-tion (M (cid:63) /M galaxy ) at the time they suffer their Γ SDSFR . Theseare shown on panels b) and c) of Figure 3, respectively. Ingeneral we find that in-situ processed galaxies tend to have a marginally larger M galaxy than pre-processed galaxies. Inter-estingly, the difference in (M (cid:63) /M galaxy ) for these three pop-ulations is significantly more evident, with the central pre-processed galaxies showing the lowest stellar mass fractions.This is in agreement with the results shown in Figure 2,where we show that Γ SDSFR for the in-situ population occursat lower redshift, thus giving more time to these galaxiesto grow in stellar mass. Note as well that there is a prefer-ence for pre-processing to occur in galaxies when they stillremain as centrals, specially for the LMC bin, as shown bythe green bars. We found that, for the pre-processed pop-ulation, 54.07% in the HMC bin, 52.14% in the IMC, and69.81% in the LMC were pre-processed as centrals.As expected for central galaxies, the M host and M galaxy distributions are similar. In Figure 3, panel d), we show thedistribution of stellar mass, M (cid:63) , for all galaxies at the timeof the first R crossing. We can clearly see that the differ-ence in M (cid:63) between in-situ and pre-processed galaxies is notonly present at the time of Γ SDSFR , but pre-processed galaxiestend to arrive in the cluster with a significantly lower stellarmass. These results suggest that one of the strongest effectsassociated with this pre-processing is to limit the final stellarmass of satellites in galaxy clusters. As an example, in Fig-ure 1, Panel c, we show how the Γ SDSFR significantly affects the
MNRAS , 1–14 (2015)
D. Pallero et al.
Figure 5.
Time evolution of cluster-centric distance for a subset of four galaxies in our sample. In panel a) we show a galaxy pre-quenchedas central, in panels b) and f) we have galaxies pre-quenched as satellite, and for panels c), d) and e), we have galaxies quenched in-situ.The color coding indicates the sSFR at each time. The dashed line shows the time evolution of the cluster R and the star shows themoment when the galaxies suffer their “processing” event. The label indicates the galaxy stellar mass at z = . subsequent growth of M (cid:63) in a galaxy. For the pre-processedpopulation, we have derived the time difference between theinfall time, t inf , and the pre-processing time, t proc . In gen-eral, we find that ( t proc − t inf ) is smaller for satellite galaxiesthan for centrals, and that this quantity grows with clustermass. This result explains the difference in stellar mass ra-tio at the moment of the processing seen in Figure 3 for thepre-processed population. Central galaxies suffer their pre-processing earlier than the satellite sample and, despite thefact that both populations shows similar M (cid:63) at the momentof the infall, those which had their strongest drop as centralsare more dark matter dominated.It is clear from Figure 3 that centrals represent an im-portant fraction of the pre-processed population, as theyconstitute (cid:38) of this population in any mass bin. In iso-lated and low-mass galaxies several mechanisms can signif-icantly affect the star formation history and current starformation activity. Examples are photo-reionization, whichlimits their gas reservoir to form stars (Hopkins et al. 2014;Chan et al. 2018), or supernova feedback which, thanks tothe injection of large amounts of kinetic energy into theintergalactic medium, can eject significant fractions of theavailable gas (Dekel & Silk 1986; Dav´e et al. 2011; Bier-nacki & Teyssier 2018). In addition, as shown by Ben´ıtez-Llambay et al. (2013), ram-pressure stripping from the gasdistribution within the cosmic web can efficiently remove thegas content of isolated low-mass galaxies. Pre-processing ingalaxies that were not centrals at the time of Γ SDSFR is gener- ally associated with ram-pressure stripping within the cor-responding host. However, Figure 3 shows that the environ-ment associated with a massive host galaxy plays a minorrole in the pre-processing of low-mass galaxies.As discussed before, the fraction of galaxies processedin-situ is rather low ( (cid:46) ), and these galaxies tend to bemore massive than the pre-processed population at the timeof their corresponding Γ SDSFR . Their most significant drop instar formation activity took place within the R of themain cluster. Thus, the main mechanisms acting are tidaland ram-pressure stripping within the cluster itself. Thishighlights the role played by the denser environment associ-ated with galaxy clusters.There is a small fraction ( < Γ SDSFR takes place in high-mass halos differ-ent from the main cluster. These halos correspond to objectsthat belong to massive galaxy-groups, in the mass range . (cid:46) M host [M (cid:12) ] (cid:46) . , that are later accreted into themain cluster. In Section 4.1 we focused on the the properties of galax-ies when they suffer their strongest drop in their star for-mation, Γ SDSFR . These drops do not necessarily result in thecessation of the star formation activity. Rather, as shownin Section 4.1, on average pre-processed galaxies arrive inthe cluster with a significantly lower stellar mass than those
MNRAS000
MNRAS000 , 1–14 (2015) uenching history of Cluster Galaxies Figure 6.
Time evolution of the cumulative fraction of quenched galaxies within the clusters’ R . The red, green and blue lines showthe results for the high, intermediate and low mass clusters, respectively. The shaded regions show Poissonian errors. galaxies processed in-situ. Thus, instead of ceasing the starformation activity, an early Γ SDSFR constrains the final galacticstellar mass.In this Section we will focus on the moment when galax-ies become effectively quenched. In R each cluster, wesearch for galaxies with sSFR values lower than sSFR Q de-fined in Section 3.2, and track their specific star forma-tion history to identify the moment when this threshold iscrossed. As before, we separate our galaxy sample in threebins according to cluster mass. The number of quenchedgalaxies in each bin is N gal = 780, 1282 and 374 for theHMC, IMC and LMC bins, respectively. Note that, in gen-eral, the number of quenched galaxies in each bin is (cid:46) smaller than the number of galaxies that have suffered sometype of processing.In the left panels of Figure 4 we show the host mass dis-tribution associated with each galaxy at the time in whichthey became quenched. As before, for galaxies that becamequenched while being centrals (green bars), M host ∼ M galaxy .Contrary to what is found with the Γ SDSFR criterion, we findthat, independently of the cluster mass bin, the vast major-ity of galaxies become quenched within massive hosts with . (cid:46) M host [M (cid:12) ] (cid:46) . . This highlights the importantrole played by the denser environment of massive clusterson the overall quenching of their galaxy members. As an ex-ample we show, in Figure 5, the time evolution of the sSFRof six galaxies in our sample as they approach the centralgalaxy of one of our clusters. The dashed lines show thetime evolution of the clusters R and the color bar thesSFR of each galaxy. The star denotes the moment whenthe Γ SDSFR takes place. Note that galaxies in panel a) reachtheir quenching state as centrals. Also, it is interesting tonote that the quenching state is reached as a consequence oftheir Γ SDSFR . For galaxies in panels b) and f), they reach theirquenching state as satellites before they were accreted bythe cluster, and galaxies in panels c) d) and e) are quenched inside the cluster R . Also, in any case, as galaxies ap-proach the cluster center, their sSFR slowly decreases. How-ever, the change in sSFR just after the first R crossing issignificantly more abrupt, in some cases rapidly resultingin quenching. On the other hand, galaxies that quenchedin low-mass halos, i.e. . (cid:46) M host [M (cid:12) ] (cid:46) . , did it ascentrals, highlighting the regime where internal quenchingprocesses are most relevant.The red bars on Figure 4 indicate the distributions ofthe in-situ quenched galaxies population. Interestingly, wefind that the fraction of galaxies that arrived in the clusteralready quenched (i.e., pre-quenched population) increaseswith cluster mass. For comparison we find of the galax-ies were quenched in-situ in the LMC bin, but only inthe HMC bin. This apparent relation between the fraction ofpre-quenched galaxies with cluster mass is further exploredbelow. As in the case of the Γ SDSFR criterion, we find the to-tal mass distribution of pre- and in-situ quenched galax-ies to be very similar (medium-left panels), but they showa significant offset on their stellar masses at the momentof quenching (medium-right panels). As expected, we findthat most pre-quenched galaxies ( ∼ ) have also been pre-processed, indicating the important role played by the pre-processing in the quenching of low-mass objects. Panel d) ofFigure 4 shows the distribution of ( t q − t inf ) , where t q repre-sents the galaxy quenching time. We find a relation between t q and cluster mass for the pre-quenched population, wherethe high mass bin presents bigger differences between bothtimes. This is a result of the hierarchical scenario; i.e., biggerclusters accrete bigger structures and, thus, environmentaleffects are more significant since earlier epochs. In generalfor the in-situ quenched population, we find no differencein ( t q − t inf ) , between the different mass bins, highlightingthe role of the virial-radius crossing in the star formationquenching of galaxies.In figure 6 we show the time evolution of the cumula- MNRAS , 1–14 (2015) D. Pallero et al. tive fraction of quenched galaxies, N q / N total , as a functionof cluster-centric distance. Here, N q represents the numberof quenched galaxies within a given radius, R , and N total thetotal number of galaxies within the same distance. The dif-ferent lines correspond to the different cluster mass bins.Interestingly, we see that at early times, between z ∼ and z ∼ . , the fraction of quenched galaxies grows towards thecluster outskirts. However, at later times this trend reverses,showing a decreasing fraction of quenched galaxies with dis-tance. During the last decade, surveys such as WINGS (Cavaet al. 2017) and SAMI (Brough et al. 2017) have shown that:(i) the fraction of quenched galaxies grows towards z = .This is attributed to the environment having more time toact on cluster galaxies;(ii) the fraction of quenched galaxies decreases withcluster-centric distance. Thanks to the denser environmentsthat can be found in the inner cluster region, galaxies, es-pecially those with lower masses, can be more efficiently de-pleted of their gas reservoir.Our results are in good agreement with these observations.We have previously highlighted a correlation between thefraction of pre-quenched galaxies and cluster mass. We fur-ther explore this correlation in Figure 7. Here we show howthe cumulative fraction of quenched galaxies, with respectto the total number of all galaxies that can be found withinR at z = , grows as a function of the normalized time, t − t infall . To obtain this plot, we first compute for each galaxywithin R at z = the time when it crossed R for thefirst time. Second, for each galaxy we define the variable t − t infall and identify the moment when it became quenchedon this new time scale. Finally, we compute the cumulativequenched galaxy fraction as a function of t − t infall . This fig-ure allows us to study how the fraction of quenched galaxieschanges as a function of the time they remain either outside(negative t − t infall ) or inside (positive t − t infall ) the cluster’sR . The different lines are associated with the galaxy pop-ulations of different clusters. The colors indicate the mass ofeach cluster at z = . Note that, in all clusters, the fractionof quenched galaxies slowly grows as galaxies approach thecluster’s R , again highlighting the role of pre-processing.Interestingly, there is a change in the slope of this cumula-tive function around the time of the first R crossing, i.e. − (cid:46) t − t infall (cid:46) . During this period, the fraction ofquenched galaxies raises more rapidly than during any otherepoch. This is in agreement with the behaviour of the sSFRobserved in Figure 5, and clearly displays the role playedby the cluster’s environment. We can also observe a largedispersion in the fraction of galaxies that arrive quenchedat the cluster’s R , with values that go from 20 to 60 % .More importantly, this fraction shows a dependency withfinal cluster mass, with larger values for more massive clus-ters.To study the origin of this trend we compute the massdistribution of the structures, M host , where the quenchedgalaxy population at z = were located at the snapshotbefore their first R crossing. This is shown in Figure 8,panels a). As before, each row corresponds to the resultsobtained from a different cluster mass bin. The blue bars in-dicate the fraction of pre-quenched galaxies, while the whitebars show all the quenched galaxies found within the clus-ter at z = . The dashed lines indicate the median for the Figure 7.
Cumulative fraction of quenched galaxies as a func-tion of the normalized time scale, t − t infall . The infall time t infall iscomputed for each individual galaxy. The color coding indicatesthe total mass of each cluster at z = . Negative (positive) t − t infall corresponds to periods of time when galaxies are located outside(inside) the cluster’s R . pre-quenched population. Interestingly, pre-quenched galax-ies on the LMC bin tend to arrive in lower mass structuresthan in the rest of the cluster mass bins. However, no signif-icant difference is observed in both the distribution of total(M galaxy ) and stellar masses (M (cid:63) ) of the pre-quenched galaxypopulations at infall, shown in panels b) and c), respectively.Our results indicate that the larger fraction of pre-quenched galaxies in larger mass clusters is the result ofthe hierarchical nature of the Λ CDM cosmological modelused in this work, in which larger mass object can accretemore massive substructures. These more massive substruc-tures are naturally more efficient in quenching their owngalaxy satellite population, thus resulting in a larger frac-tion of pre-quenched galaxies at z = . In this paper we have presented a study of the differentenvironmental-quenching and processing scenarios under-gone by the satellite galaxies of the ten most massive clustersin the state-of-the-art EAGLE hydrodynamical simulation.Two different criteria were defined to explore the differentprocesses that significantly affect the SFR of these galaxiesalong their history. Our goal is to quantify and characterizethe role played by the environment in these processes.For the instantaneous strongest drop in SFR, we findthat the majority of galaxies suffer their Γ SDSFR outside thecluster’s R (pre-processed fraction (cid:38) ). This fractiongrows with cluster mass. We find that there is no correlationbetween the strength of the Γ SDSFR and the time at which itoccurs, nor a preferential redshift for it to happen. Nonethe-less, for galaxies processed in-situ, Γ SDSFR tends to happenat lower redshift than for the pre-processed population. Interms of the environment, while in-situ processing mainlyoccurs in massive hosts, pre-processing shows a strong pref-
MNRAS000
MNRAS000 , 1–14 (2015) uenching history of Cluster Galaxies Figure 8.
Mass distribution of galaxies and their hosts at the moment before crossing the cluster’s R . Each row shows the resultsobtained after stacking the distribution of galaxies associated with clusters within different mass ranges. Column (a) shows the massdistribution of the host of each galaxy. Column (b) shows the distribution of galaxy total mass. Column (c) shows the distributions ofstellar mass. The blue bars correspond to the galaxies quenched before the first infall and the white bars correspond to all galaxies inour sample. Dashed lines correspond to the median of the pre-quenched population. erence to take place in galaxies that are either low massand central ( . (cid:46) M host [M (cid:12) ] (cid:46) . ) or that belong tolow-to-middle mass hosts ( (cid:46) M host [M (cid:12) ] (cid:46) . ). Ourresults are in good agreement with those published by Bian-coni et al. (2018), who observationally studied a sample of 23massive clusters (M = 10 . [M (cid:12) ]) with 34 infalling groups(log M (cid:63) [M (cid:12) ] = 10.75), located in outer cluster regions.They found that at cluster-centric distances R ∼ . R thefraction of star-forming galaxies in infalling groups is half ofthat in the clusters. According to this, Bianconi et al. (2018)suggest that the pre-processing in groups is the responsiblefor these results.Interestingly, for galaxies with similar total mass, atthe time of arrival in the main cluster, the in-situ processedpopulation shows in general a larger stellar mass than thosepre-processed. This highlights the important role of pre-processing in limiting the star formation activity of low-massgalaxies. The origin of this pre-processing event can be ex- plained by a variety of different internal mechanisms suchas supernova feedback, photo-reionization, interactions andstarburst phases. Unfortunately, due to the poor time andspatial resolution available with this simulation, it is toohard to identify what the main mechanism acting on eachgalaxy is. In addition, ram-pressure from the cosmic web canalso cause an accelerated depletion of the gas reservoirs inlow mass galaxies, producing abrupt changes in their starformation (Ben´ıtez-Llambay et al. 2013).In the case of the Critical sSFR criterion, contrary toour results for Γ SDSFR , we find that quenching presents a strongpreference for high-mass halos to take place. This is a strongindicator that dense environments promote the definitivecessation of the star formation.Our results are in agreement with the observations pre-sented by Olave-Rojas et al. (2018), who find that the frac-tion of high-mass (M (cid:63) ≥ . [M (cid:12) ] ) red (i.e. passive) galax-ies in clusters (i.e quenched in-situ) is higher than the frac- MNRAS , 1–14 (2015) D. Pallero et al. tion of high-mass red galaxies in accreted groups (i.e. pre-quenched). We find that most of the pre-quenched galaxies( (cid:38) ) have also been pre-processed, evidencing the impor-tance of pre-processing in the quenching of low-mass galax-ies. In general we find a slight preference for pre-quenchingto take place at earlier times compared with quenching in-situ. The difference in the median of the quenching timedistribution is only of the order of 1 to 2 Gyr. As a func-tion of cluster-centric distance, close to z = the fractionof quenched galaxies grows toward the cluster center. Thisis in good agreement with the results obtained from obser-vational studies based on different surveys such as WINGS(Cava et al. 2017) and SAMI (Brough et al. 2017). However,at earlier times, between z ∼ and z ∼ . , this trend reverts,showing a fraction of quenched galaxies that grows towardsthe cluster outskirts.In general, we find that in comparison to the in-situquenched population, on average pre-quenched galaxies havelower stellar-masses. This result appears to be in disagree-ment with those presented by Hou et al. (2014) who foundthat, independent of galaxy mass, the fraction of quiescentgalaxies is higher in groups than in the clusters and field.However, we can reconcile our findings with those of Houet al. (2014) by noting that those authors only studied galax-ies with stellar masses in the range 9.5 < log M (cid:63) [ M (cid:12) ] < . ≤ M halo [ M (cid:12) ] ≤ . . In these massive anddense substructures the environmental quenching effects arestronger.We find a sharp rise in the fraction quenched satellitesat the time of the first infall, highlighting the role playedby the dense cluster environment. It is interesting to notethat, although galaxies prefer denser environments to reachtheir quenching state, the fraction of pre-quenched galaxiesin our sample grows with the total mass of the cluster at z = . We find that of galaxies were quenched in-situ inthe low-mass clusters, but only were quenched in-situfor the high-mass clusters. To explain why high-mass clustersshow higher fractions of pre-quenched galaxies, we explorethe mass distribution of the structures where the clustersatellite galaxies reside at the moment of accretion. We findthat high-mass clusters preferentially accrete their satellitesthrough structures and groups that are significantly moremassive than those accreted by low-mass clusters. This is adirect consequence of the hierarchical cosmological modelused in these simulations. More massive clusters tend toaccrete more massive substructures. Due to their own in-tracluster dense environments, these massive substructuresarrive in the clusters with their satellite population alreadyquenched.Cora et al. (2018a) explored the quenching time ofgalaxies, and the relevance of the environment on this pro-cess, using the semi-analytic model SAG (Cora et al. 2018b).A criterion similar to our sSFR threshold was imposed. Ac-cording to their results, environmental effects dominate thestar formation quenching of low-mass satellite galaxies (M (cid:63) < . [M (cid:12) ] . These results are in good agreement with ourresults. Panels a) and c) of Figure 4 show that we also findan important fraction of low-stellar mass galaxies that arequenched within the cluster’s R . Note that a significantfraction of the low-stellar mass galaxies that arrive in thecluster as quenched galaxies were actually quenched in thedense environments of massive groups. This exemplifies the relevance of the environment in the quenching of the clustersatellite population.We also find that there is a fraction of low-stellar massgalaxies that are quenched as centrals. According to Ben´ıtez-Llambay et al. (2013), this can be explained through a com-bination of different mechanisms that are acting simultane-ously on dwarf galaxies. Processes such as supernova feed-back and photo-reionization can reheat the cool gas of thesegalaxies inducing the quenching of their star formation ac-tivity, a scenario commonly referred to as mass quenching.In addition, as previously discussed, ram-pressure strippingtaking plance within the cosmic web filaments can also de-plete the gas reservoir of dwarf galaxies, producing a quench-ing state due to the environment.As we mentioned before, due to the limtied number ofsnapshots available in the simulation, we do not have the ca-pabilities to separate and distinguish the different overlapingprocesses that are influencing the star formation history ofthe galaxies. In a follow-up project we plan to explore thesedifferent mechanisms using more detailed hydrodynamicalsimulations from the C-EAGLE project. These simulationsprovide us with a great improvement in temporal resolution,with a temporal resolution for particles of ˜125 Myr and ˜25Myr for three particular intervals of redshift ([0-1],[4-5],[7-8]), and 500 Myr for group catalogues (Barnes et al. 2017;Bah´e et al. 2017). Since this simulation suite also counts witha sample of 30 clusters with a M in the range between of10 . < M [M (cid:12) ] < . , this study will also allow us toexplore in more detail the dependency between cluster massand fraction of pre-quenched galaxies. ACKNOWLEDGEMENTS
We thank Daniel Hernandez, Catalina Labayru, Ciria Lima,Antonella Monachesi and Catalina Mora for useful discus-sions and comments. We also want to thank the anony-mous referee for his/her many insightful comments, thatgreatly improved the quality of the manuscript. We ac-knowledge the Virgo Consortium for making their simu-lation data available. The EAGLE simulations were per-formed using the DiRAC-2 facility at Durham, managed bythe ICC, and the PRACE facility Curie based in France atTGCC, CEA, Bruyeres-le-Chatel. We also thank the sup-port given by the “Vicerrector´ıa de Investigaci´on de la Uni-vesidad de La Serena” for the support given by the pro-gram “Apoyo al fortalecimiento de grupos de investigaci´on”.DP also acknowledges financial support through the fel-lowship “Becas Doctorales Institucionales ULS”, granted bythe “Vicerrector´ıa de Investigaci´on y Postgrado de la Uni-versidad de La Serena. DP also thanks the hospitality ofPUC during the stay at the university. DP and FAG ac-knowledge financial support from the Max Planck Societythrough a Partner Group grant. FAG acknowledges financialsupport from CONICYT through the project FONDECYTRegular Nr. 1181264. D.O-R acknowledges the financial sup-port provided by CONICYT-PCHA through a PhD Scholar-ship, “Beca Doctorado Nacional A˜no 2015”, under contract2015-21150415. P.C. acknowledges the support provided byFONDECYT postdoctoral research grant no 3160375. NPacknowlodges the support from BASAL AFB-170002 CATA,CONICYT Anillo-1477 and Fondecyt Regular 1150300. R.D.
MNRAS000
MNRAS000 , 1–14 (2015) uenching history of Cluster Galaxies gratefully acknowledges support from the Chilean Centrode Excelencia en Astrof´ısica y Tecnolog´ıas Afines (CATA)BASAL grant AFB-170002. REFERENCES
Abadi M. G., Moore B., Bower R. G., 1999, MNRAS, 308, 947Bah´e Y. M., et al., 2017, MNRAS, 470, 4186Baldry I. K., Balogh M. L., Bower R. G., Glazebrook K., NicholR. C., Bamford S. P., Budavari T., 2006, MNRAS, 373, 469Balogh M. L., et al., 2007, MNRAS, 374, 1169Balogh M. L., et al., 2016, MNRAS, 456, 4364Barnes J. E., Hernquist L., 1996, ApJ, 471, 115Barnes D. J., et al., 2017, MNRAS, 471, 1088Ben´ıtez-Llambay A., Navarro J. F., Abadi M. G., Gottl¨ober S.,Yepes G., Hoffman Y., Steinmetz M., 2013, ApJ, 763, L41Bianconi M., Smith G. P., Haines C. P., McGee S. L., FinoguenovA., Egami E., 2018, MNRAS, 473, L79Biernacki P., Teyssier R., 2018, MNRAS, 475, 5688Boselli A., Gavazzi G., 2006, PASP, 118, 517Boselli A., et al., 2005, ApJ, 629, L29Bravo-Alfaro H., Cayatte V., van Gorkom J. H., Balkowski C.,2000, AJ, 119, 580Brough S., et al., 2017, ApJ, 844, 59Butcher H., Oemler Jr. A., 1984, ApJ, 285, 426Byrd G., Valtonen M., 1990, ApJ, 350, 89Cantalupo S., 2010, MNRAS, 403, L16Cassata P., et al., 2008, A&A, 483, L39Cava A., et al., 2017, A&A, 606, A108Cayatte V., van Gorkom J. H., Balkowski C., Kotanyi C., 1990,AJ, 100, 604Chabrier G., 2003, PASP, 115, 763Chan T. K., Kereˇs D., Wetzel A., Hopkins P. F., Faucher-Gigu`ereC.-A., El-Badry K., Garrison-Kimmel S., Boylan-Kolchin M.,2018, MNRAS, 478, 906Chester C., Roberts M. S., 1964, AJ, 69, 635Cora S. A., Hough T., Vega-Mart´ınez C. A., Orsi ´A., 2018a,preprint, ( arXiv:1801.03884 )Cora S. A., et al., 2018b, MNRAS, 479, 2Crain R. A., et al., 2015, MNRAS, 450, 1937Croton D. J., et al., 2006, MNRAS, 365, 11Cybulski R., Yun M. S., Fazio G. G., Gutermuth R. A., 2014,MNRAS, 439, 3564Dalla Vecchia C., Schaye J., 2012, MNRAS, 426, 140Dav´e R., Oppenheimer B. D., Finlator K., 2011, MNRAS, 415,11Davis M., Efstathiou G., Frenk C. S., White S. D. M., 1985, ApJ,292, 371De Lucia G., Weinmann S., Poggianti B. M., Arag´on-SalamancaA., Zaritsky D., 2012, MNRAS, 423, 1277Dekel A., Birnboim Y., 2008, MNRAS, 383, 119Dekel A., Silk J., 1986, ApJ, 303, 39Desai V., et al., 2007, ApJ, 660, 1151Dolag K., Borgani S., Murante G., Springel V., 2009, MNRAS,399, 497Dressler A., 1980, ApJ, 236, 351Dressler A., 1984, ARA&A, 22, 185Dressler A., et al., 1997, ApJ, 490, 577Dressler A., Oemler Jr. A., Poggianti B. M., Gladders M. D.,Abramson L., Vulcani B., 2013, ApJ, 770, 62Efstathiou G., 2000, MNRAS, 317, 697Faber S. M., 1973, ApJ, 179, 731Fasano G., Poggianti B. M., Couch W. J., Bettoni D., KjærgaardP., Moles M., 2000, ApJ, 542, 673Fujita Y., 2004, PASJ, 56, 29Furlong M., et al., 2015, MNRAS, 450, 4486Furlong M., et al., 2017, MNRAS, 465, 722 Gavazzi G., Fumagalli M., Cucciati O., Boselli A., 2010, A&A,517, A73Gray M. E., Wolf C., Meisenheimer K., Taylor A., Dye S., BorchA., Kleinheinrich M., 2004, MNRAS, 347, L73Gunn J. E., Gott III J. R., 1972, ApJ, 176, 1Haardt F., Madau P., 2001, in Neumann D. M., Tran J. T. V., eds,Clusters of Galaxies and the High Redshift Universe Observedin X-rays. p. 64 ( arXiv:astro-ph/0106018 )Haines C. P., et al., 2013, ApJ, 775, 126Haines C. P., et al., 2015, ApJ, 806, 101Hashimoto Y., Oemler Jr. A., Lin H., Tucker D. L., 1998, ApJ,499, 589Hogg D. W., et al., 2004, ApJ, 601, L29Hopkins P. F., Kereˇs D., O˜norbe J., Faucher-Gigu`ere C.-A.,Quataert E., Murray N., Bullock J. S., 2014, MNRAS, 445,581Hou A., Parker L. C., Harris W. E., 2014, MNRAS, 442, 406Jaff´e Y. L., Smith R., Candlish G. N., Poggianti B. M., SheenY.-K., Verheijen M. A. W., 2015, MNRAS, 448, 1715Jaff´e Y. L., et al., 2016, MNRAS, 461, 1202Just D. W., Zaritsky D., Sand D. J., Desai V., Rudnick G., 2010,ApJ, 711, 192Katsianis A., et al., 2017, MNRAS, 472, 919Kauffmann G., White S. D. M., Heckman T. M., M´enard B.,Brinchmann J., Charlot S., Tremonti C., Brinkmann J., 2004,MNRAS, 353, 713Kennicutt Jr. R. C., 1998, ApJ, 498, 541Kodama T., Smail I., Nakata F., Okamura S., Bower R. G., 2001,ApJ, 562, L9Lewis I., et al., 2002, MNRAS, 334, 673McAlpine S., et al., 2016, Astronomy and Computing, 15, 72McGee S. L., Balogh M. L., Bower R. G., Font A. S., McCarthyI. G., 2009, MNRAS, 400, 937Miller R. H., 1986, A&A, 167, 41Moore B., Katz N., Lake G., Dressler A., Oemler A., 1996, Nature,379, 613Moore B., Lake G., Quinn T., Stadel J., 1999, MNRAS, 304, 465Moran S. M., Ellis R. S., Treu T., Smith G. P., Rich R. M., SmailI., 2007, ApJ, 671, 1503Morgan W. W., Mayall N. U., 1957, PASP, 69, 291Muzzin A., et al., 2012, ApJ, 746, 188Nantais J. B., et al., 2016, A&A, 592, A161Nelson D., et al., 2018, MNRAS, 475, 624Oesch P. A., et al., 2016, ApJ, 819, 129Olave-Rojas D., Cerulo P., Demarco R., Jaff´e Y. L., Mercurio A.,Rosati P., Balestra I., Nonino M., 2018, MNRAS, 479, 2328Overzier R. A., 2016, A&ARv, 24, 14Peng Y.-j., et al., 2010, ApJ, 721, 193Peng Y.-j., Lilly S. J., Renzini A., Carollo M., 2012, ApJ, 757, 4Planck Collaboration et al., 2014, A&A, 566, A54Poggianti B. M., Smail I., Dressler A., Couch W. J., Barger A. J.,Butcher H., Ellis R. S., Oemler Jr. A., 1999, ApJ, 518, 576Poggianti B. M., et al., 2001, ApJ, 562, 689Poggianti B. M., et al., 2006, ApJ, 642, 188Quilis V., Moore B., Bower R., 2000, Science, 288, 1617Roberts M. S., Haynes M., 1994, in Meylan G., Prugniel P., eds,European Southern Observatory Conference and WorkshopProceedings Vol. 49, European Southern Observatory Con-ference and Workshop Proceedings. p. 197Schaye J., 2004, ApJ, 609, 667Schaye J., Dalla Vecchia C., 2008, MNRAS, 383, 1210Schaye J., et al., 2015, MNRAS, 446, 521Smail I., Ellis R. S., Dressler A., Couch W. J., Oemler A., SharplesR. M., Butcher H., 1997, ApJ, 479, 70Springel V., 2005, MNRAS, 364, 1105Springel V., White S. D. M., Tormen G., Kauffmann G., 2001,MNRAS, 328, 726Strateva I., et al., 2001, AJ, 122, 1861MNRAS , 1–14 (2015) D. Pallero et al.
Tescari E., et al., 2018, MNRAS, 473, 380Toomre A., Toomre J., 1972, in Bulletin of the American Astro-nomical Society. p. 214Trayford J. W., et al., 2015, MNRAS, 452, 2879Trayford J. W., Theuns T., Bower R. G., Crain R. A., LagosC. d. P., Schaller M., Schaye J., 2016, MNRAS, 460, 3925Treu T., Ellis R. S., Kneib J.-P., Dressler A., Smail I., Czoske O.,Oemler A., Natarajan P., 2003, ApJ, 591, 53Vijayaraghavan R., Ricker P. M., 2013, MNRAS, 435, 2713Vollmer B., Cayatte V., Balkowski C., Duschl W. J., 2001, ApJ,561, 708Walker I. R., Mihos J. C., Hernquist L., 1996, ApJ, 460, 121Weinmann S. M., Kauffmann G., von der Linden A., De LuciaG., 2010, MNRAS, 406, 2249Wetzel A. R., Tinker J. L., Conroy C., 2012, MNRAS, 424, 232Wetzel A. R., Tinker J. L., Conroy C., van den Bosch F. C., 2013,MNRAS, 432, 336Wiersma R. P. C., Schaye J., Smith B. D., 2009a, MNRAS, 393,99Wiersma R. P. C., Schaye J., Theuns T., Dalla Vecchia C., Tor-natore L., 2009b, MNRAS, 399, 574Zabludoff A. I., Mulchaey J. S., 1998, ApJ, 498, L5van den Bosch F. C., Aquino D., Yang X., Mo H. J., Pasquali A.,McIntosh D. H., Weinmann S. M., Kang X., 2008, MNRAS,387, 79This paper has been typeset from a TEX/L A TEX file prepared bythe author. MNRAS000