LoCuSS: exploring the connection between local environment, star formation and dust mass in Abell 1758
Matteo Bianconi, Graham P. Smith, Chris P. Haines, Sean L. McGee, Alexis Finoguenov, Eiichi Egami
MMNRAS , 1–16 (2019) Preprint January 15, 2020 Compiled using MNRAS L A TEX style file v3.0
LoCuSS: exploring the connection between localenvironment, star formation and dust mass in Abell 1758
M. Bianconi ,(cid:63) , G. P. Smith , C. P. Haines , , S. L. McGee , A. Finoguenov ,E. Egami School of Physics and Astronomy, University of Birmingham, Edgbaston, Birmingham, B15 2TT, UK Instituto de Astronom´ıa y Ciencias Planetarias de Atacama, Universidad de Atacama, Copayapu 485, Copiap´o, Chile Department of Physics, University of Helsinki, Gustaf H¨allstr¨omin katu 2a, FI-0014 Helsinki, Finland Steward Observatory, University of Arizona, 933 North Cherry Avenue, Tucson, AZ 85721, USA
Accepted XXX. Received YYY; in original form ZZZ
ABSTRACT
We explore the connection between dust and star formation, in the context of environ-mental effects on galaxy evolution. In particular, we exploit the susceptibility of dust toexternal processes to assess the influence of dense environment on star-forming galax-ies. We have selected cluster Abell 1758 from the Local Cluster Substructure Survey(LoCuSS). Its complex dynamical state is an ideal test-bench to track dust removaland destruction in galaxies due to merger and accretion shocks. We present a sys-tematic panchromatic study (from . µ m with GALEX to µ m with Herschel ) ofspectroscopically confirmed star-forming cluster galaxies at intermediate redshift. Weobserve that the main subclusters (A1758N and A1758S) belong to two separate large-scale structures, with no overlapping galaxy members. Star-forming cluster membersare found preferentially outside cluster central regions, and are not isotropically dis-tributed. Rather, these galaxies appear being funneled towards the main subclustersalong separate accretion paths. Additionally, we present the first study of dust-to-stellar (DTS) mass ratio used as indicator for local environmental influence on galaxyevolution. Star-forming cluster members show lower mean values (32 % at . σ ) ofDTS mass ratio and lower levels of infrared emission from birth clouds with respectto coeval star-forming field galaxies. This picture is consistent with the majority ofstar-forming cluster members infalling in isolation. Upon accretion, star-formation isobserved to decrease and warm dust is destroyed due to heating from the intraclustermedium radiation, ram-pressure stripping and merger shocks. Key words: galaxies: clusters: individual: Abell 1758 – galaxies: evolution – galaxies:star formation
Dust plays an important role in shaping the evolution ofgalaxies. It acts as a catalyst for the formation of molec-ular gas, which accumulates in the dense and cold cloudsthat become the birthplace of stars (Galliano et al. 2018, fora review). Dust is also responsible for reprocessing UV ra-diation from newly-born stars, resulting in an extinction oflight from galaxies at short wavelengths, and a re-emitting ofthat energy at infrared wavelengths. It is thought that dustformation occurs predominantly via the growth of grainsin external layers of AGB star atmospheres and super- (cid:63) [email protected] novae ejecta, which are later distributed into the interstellarmedium by stellar winds.As it traces the creation of galaxy’s stellar content, andis mixed through the interstellar medium, measurements ofthe dust content are crucial for understanding why the starformation rate density of universe has declined since z (cid:39) and what drives the quenching of star formation. Observa-tions have shown that while star-forming galaxies have highdust content (particularly as a fraction of their stellar mass),passive galaxies do not (Smith et al. 2012). This has beenextended by showing that the dust mass of a galaxy directlycorrelates with the star formation rate (SFR), at least forgalaxies in the field (da Cunha et al. 2010). It is not clearwhat happens to the dust created during star formation,such that it is no longer detected in massive and passive © a r X i v : . [ a s t r o - ph . GA ] J a n M. Bianconi et al. N E A1758NA1758S A1758-g8
Figure 1.
The merging cluster A1758. Top panel: zoom-in mosaic of A1758N observed with HST-ACS in F435W, F606W and F814Wfilters from the RELICS survey (Coe et al. 2019). Bottom panel: the underlying map in shades of black shows the surface mass densitysignal-to-noise (SNR) ratio based on the weak-lensing analysis of (Okabe & Smith 2016, see end of Section 3). Black contours trace theextended X-ray emission measured with
XMM-Newton , detected above a 4 σ threshold in the wavelet analysis (Haines et al. 2018), andare logarithmically spaced between . × − and . × − count s − . Star-forming cluster and field galaxies are plotted as red circles andblue squares, respectively. Yellow hexagons mark the brightest galaxy position of the three cluster subclumps, referred to as NW, NE andS according to their coordinates. The X-ray contours encompassed by the green dot-dashed circle (with radius equal to r ) correspondto the X-ray group A1758-g8 discovered in Haines et al. (2018). galaxies. It has been proposed that it is destroyed via mech-anisms internal to the galaxy, such as supernovae shocks(Jones 2004, for a review), or is driven out of the galaxy byan outflow or consumed less efficiently due to heating fromactive galactic nuclei (Gobat et al. 2018).Even more uncertain is what role dust plays in the en-vironmentally driven suppression, or quenching, of star for- mation. Environmental processes have been shown to affectatomic gas content, resulting in truncated density profiles inthe outskirts of galaxies (Davis et al. 2013). Environmentaleffects on molecular gas, and consequently on dust, are still asubject of debate (Cortese et al. 2012; Koyama et al. 2017),but there are measurements of the spatial distribution ofdust in cluster galaxies that are consistent with it having MNRAS000
XMM-Newton , detected above a 4 σ threshold in the wavelet analysis (Haines et al. 2018), andare logarithmically spaced between . × − and . × − count s − . Star-forming cluster and field galaxies are plotted as red circles andblue squares, respectively. Yellow hexagons mark the brightest galaxy position of the three cluster subclumps, referred to as NW, NE andS according to their coordinates. The X-ray contours encompassed by the green dot-dashed circle (with radius equal to r ) correspondto the X-ray group A1758-g8 discovered in Haines et al. (2018). galaxies. It has been proposed that it is destroyed via mech-anisms internal to the galaxy, such as supernovae shocks(Jones 2004, for a review), or is driven out of the galaxy byan outflow or consumed less efficiently due to heating fromactive galactic nuclei (Gobat et al. 2018).Even more uncertain is what role dust plays in the en-vironmentally driven suppression, or quenching, of star for- mation. Environmental processes have been shown to affectatomic gas content, resulting in truncated density profiles inthe outskirts of galaxies (Davis et al. 2013). Environmentaleffects on molecular gas, and consequently on dust, are still asubject of debate (Cortese et al. 2012; Koyama et al. 2017),but there are measurements of the spatial distribution ofdust in cluster galaxies that are consistent with it having MNRAS000 , 1–16 (2019) oCuSS: environment, star formation and dust in A1758 Name Centre (RA, Dec) N Redshift L X Mass M Radius r Velocity dispersion(RA, Dec) (cid:104) z (cid:105) ( . − . ) [ erg s − ] [ M (cid:12) ] [Mpc] [ km s − ]A1758N 203.18848, 50.54294 176 0.27879 ± . a ± c ± ± ± . b ± d ± ± ± . b ± d ± ± Table 1.
Summary of the principal properties of the two main subclusters, A1758N and S, and the X-ray group A1758-g8 from Haineset al. (2018). From left to right, halo name, central coordinate, number of spectroscopically confirmed members, X-ray luminosity L X ( a from ROSAT , b from XMM-Newton , Haines et al. 2018), mass M ( c from the combined Chandra-XMM analysis of Martino et al. 2014,whereas d is computed using the scaling relation between X-ray luminosity L x and M from Leauthaud et al. 2010), radius r andvelocity dispersion, which is estimated from the velocity distribution of member galaxies. M is defined as the mass contained within r , which encompasses an overdense region presenting an average density 200 times higher than the Universe critical density at thecluster redshift ρ crit ( z ) , i.e. M = π r ρ crit ( z ) (Voit 2005). log(Flux density [Jy]) l o g ( N u m b e r ) m m m ClusterSFmembersFieldSFgalaxies100 m160 m250 m
Figure 2.
Number of star forming galaxies detected in our far-IR observations as a function of flux density at , , and µ m . At bright fluxes, the number counts of star-forming clus-ter galaxies falls more steeply than the number of star-formingfield galaxies, suggesting a dearth of IR-bright galaxies in clus-ters relative to the field. The vertical lines correspond to the fluxlimits in each infrared band. been stripped from the galaxy (Gomez et al. 2010; Walteret al. 2011). Further evidence for differential dust content inclusters and the field were found in the first systematic dustsurveys of the Local Universe (Cortese et al. 2012). Whilethis has shown that the dust content of galaxies in clusters isdifferent from that of galaxies in the field, the physical mech-anism causing this could be any of: ram pressure stripping(Gunn & Gott 1972; Jablonka et al. 2013), galaxy harass-ment (Moore et al. 1996), strangulation (Larson et al. 1980)or heating from the intracluster medium (ICM) (Mok et al.2016).Clearly, a powerful method to test models of dust for-mation and destruction, and their relation to star formation,is to examine how key scaling relations, such as between dustand stellar mass and dust mass and star-formation rate, vary in different environments. Furthermore, the cluster’s dynam-ical state has to be taken into account. Merger events areaccompanied by shock fronts, expanding through the ICM,which in turn can affect the gas and dust content in clustergalaxies.In this paper, we concentrate on a single cluster Abell1758 (A1758) at z = . , which is known for its complexformation history forged by recent and ongoing mergers ofseparate clusters (Table 1 and Figure 1). A1758 is thereforean ideal laboratory in which to study the impact of localenvironment within clusters and cluster dynamics on dustcontent in member galaxies.X-ray analysis by David & Kempner (2004) first ev-idenced the absence of excess X-ray emission betweenA1758N and S, which originates from merger shocks com-pressing the ICM. This suggested that A1758N and S haveyet to interact with each other. Further analysis of Chan-dra images revealed that the broadly-peaked X-ray emissionto the North is associated with two prominent subclumpsA1758NW and A1758NE separated by 800 kpc, and cur-rently receding from each other, being observed some 300Myr after the first core-passage (David & Kempner 2004).Recently, Schellenberger et al. (2019) has confirmed this sce-nario, and additionally identified a shock front on the Northside of the sub-cluster A1758NW, best-fit by a supersoniccollision with Mach number 1.6, indicating a relative veloc-ity of 2100 km s − . David & Kempner (2004) hypothesizedthat the South subcluster is further divided into two sub-structures, that will merge perpendicularly to the plane ofthe sky (Monteiro-Oliveira et al. 2017; Schellenberger et al.2019). The scenario of multiple clumps at different stagesof merging is also corroborated by numerical simulationsby Durret et al. (2011) and Machado et al. (2015). Thezoomed-in image on A1758N with Hubble -ACS (Coe et al.2019) confirms that the majority of cluster members presentspheroidal/elliptical morphology. Nevertheless, disc galaxiesemerge at increasing distance from the cluster cores. Distinctspiral arms, together with signatures of ram-pressure strip-ping (Ebeling & Kalita 2019), indicate that these galaxiesare undergoing first encounter with the cluster environment.This paper is structured in the following manner. In Sec-tion 2, we present the datasets used. In Section 3, we presentthe methods and main results of the data analysis. In Section4, we discuss the results and future prospects of the project.Throughout this work, we assume H =
70 km s − Mpc − , MNRAS , 1–16 (2019)
M. Bianconi et al.
RA (J2000) +50.550°+50.552° D e c ( J ) / m [observer frame] L o g ( L / L ) Log(SFR) = 0.92Log(Mstar) = 11.039Log(Mdust) = 8.731 = 5.600 RA (J2000) +50.538°+50.540° D e c ( J ) / m [observer frame] L o g ( L / L ) Log(SFR) = 0.465Log(Mstar) = 10.715Log(Mdust) = 7.873 = 1.47 Figure 3.
Postage stamps and SED of two star-forming cluster members. In the left column, we see typical spiral features and colours ofstar-forming galaxies (30 kpc radius HST-ACS cutouts from the RELICS survey, Coe et al. 2019). In the right column, photometry (redopen circles) and SED best-fit model (black curves) plots display the different properties of each galaxy and the extended wavelengthcoverage of the LoCuSS dataset. Salient properties obtained for each galaxy from SED fit are highlighted in text:
Log ( SFR [ M (cid:12) yr − ]) , Log ( M dust [ M (cid:12) ]) and Log ( M ∗ [ M (cid:12) ]) and χ of the fit. Ω M = . and Ω Λ = . , and will not explicitly write thebase (always 10) of logarithms. A1758 is among the clusters selected for the LoCuSS survey(Smith et al. 2010). As a result, it benefits from the ex-tensive coverage in both wavelength and area with GALEX(FUV and NUV), Subaru/Suprime-Cam ( g + and R bands),UKIRT/WFCAM( J and K band), mid-infrared 24 µ m withSpitzer/MIPS (reaching 90% completeness at µ Jy) andfar-infrared with Herschel (Haines et al. 2010; Smith et al.2010; Pereira et al. 2010), both covering (cid:48) × (cid:48) fields. Inparticular, as part of the LoCuSS Open Time Key Programon Herschel
A1758 was observed at 100 and 160 µ m withPACS and at , , and µ m with SPIRE (Smith et al.2010). Herschel flux limits are 13.0, 17.0, 14.0, 18.9, 20.4mJy from 100 to 500 µ m at 3 σ (Rawle et al. 2012a). Ad-ditionally, A1758 is part of the the volume-limited high- L x LoCuSS sub-sample of 50 clusters and has
XMM -Newtonimaging (see Martino et al. 2014 for further observationaldetails). These observations were utilised to detect 39 newinfalling galaxy groups surrounding 23 LoCuSS clusters, captured at their first encounter with the cluster environ-ment (Haines et al. 2018). Furthermore, wide-field ( ≈ µ m with Spitzer down to µ Jy . Archival SDSS andWISE photometry was added to the data pool. In partic-ular, we used the AllWISE Source Catalog, reaching fluxlimits (at SNR 5) of 54, 71, 730 µ Jy for 3.4, 4.6 and 12 µ m respectively .In this work, we focus on star-forming galaxies, bothas cluster members and field galaxies. The sample of coevalfield galaxies is included as a benchmark to allow the studyof environmental effects on star formation. In particular, weconsider those spectroscopically-confirmed cluster membergalaxies that are detected at 24 µ m and also lie within theHerschel-PACS footprint. Field galaxies are selected fromobservations of 5 additional clusters from the LoCuSS sur-vey at z < . , that were observed with Herschel PACS andSPIRE instruments in the exact same way as A1758, that is http:// wise2 . ipac . caltech . edu / docs / release / allwise / expsup / sec2 1 . html MNRAS000
XMM -Newtonimaging (see Martino et al. 2014 for further observationaldetails). These observations were utilised to detect 39 newinfalling galaxy groups surrounding 23 LoCuSS clusters, captured at their first encounter with the cluster environ-ment (Haines et al. 2018). Furthermore, wide-field ( ≈ µ m with Spitzer down to µ Jy . Archival SDSS andWISE photometry was added to the data pool. In partic-ular, we used the AllWISE Source Catalog, reaching fluxlimits (at SNR 5) of 54, 71, 730 µ Jy for 3.4, 4.6 and 12 µ m respectively .In this work, we focus on star-forming galaxies, bothas cluster members and field galaxies. The sample of coevalfield galaxies is included as a benchmark to allow the studyof environmental effects on star formation. In particular, weconsider those spectroscopically-confirmed cluster membergalaxies that are detected at 24 µ m and also lie within theHerschel-PACS footprint. Field galaxies are selected fromobservations of 5 additional clusters from the LoCuSS sur-vey at z < . , that were observed with Herschel PACS andSPIRE instruments in the exact same way as A1758, that is http:// wise2 . ipac . caltech . edu / docs / release / allwise / expsup / sec2 1 . html MNRAS000 , 1–16 (2019) oCuSS: environment, star formation and dust in A1758 log(L IR / L ) N u m b e r M p c Cluster SF membersField SF galaxies
Figure 4.
Comoving number density of star-forming clustermembers and field galaxies as a function of their total infraredluminosity. The shaded areas are (cid:112) N / Volume . covering the same sized fields ( (cid:48) × (cid:48) ) to the same depthsin all five far-IR bands. From these data we select field star-forming galaxies within the redshift range 0.23 < z < − of the mean red-shift of cluster members (Haines et al. 2013). In Figure 2 we plot the number of spectroscopically con-firmed star-forming galaxies in the cluster and field samplesas a function of flux density in the , , and µ m bands, without any attempt to normalize the counts to thevolume surveyed. The number counts of star-forming fieldgalaxies are flatter than the number counts of star-formingclusters galaxies, especially at µ m and µ m , suggestinga dearth of IR-bright galaxies in clusters relative to the fieldand/or differences in typical spectral shape at these wave-lengths between the two samples. Radiation at these wave-lengths is characteristic of warm dust surrounding regions ofstar formation: spiral galaxies discs and arms are the locusof new episodes of star formation, resulting in UV emissionfrom newly born stars which is reprocessed by surroundingdust in the IR, both of which are lacking in typical ellipticalgalaxies. Typical morphologies of star-forming galaxies canbe seen in the left panels of Figure 3 and in the Appendix(Figure 11) confirming the expected spiral morphology. In order to search further discrepancies between cluster andfield object, we combined the entire photometric coverage toobtain the spectral energy distribution (SED) of each galaxy.The public code MAGPHYS (Multi-wavelength Analysis of Galaxy Physical Properties) by da Cunha et al. (2008) allowsus to derive salient physical parameters of galaxies by fittingmock SED templates to the observed multi-wavelength pho-tometric data.For each mock template, the radiation of stars, assumedto be the only heating source, is reprocessed by dust througha two-phase grey body model. Stellar emission is computedusing the population synthesis model of Bruzual & Char-lot (2003) and by assuming a Chabrier (2003) initial massfunction, which is restricted to . < M / M (cid:12) < . Thedust model includes both dense molecular clouds and a dif-fuse interstellar medium, following Charlot & Fall (2000),re-emitting the stellar radiation through four separate com-ponents at different temperatures. In particular, these mod-els comprise cold (15-25 K) and warm (30-60 K) dust, a hotcontinuum (130-250 K) from interstellar grains and poly-cyclic aromatic hydrocarbon emission, the total radiation ofwhich spans between 3 and 1000 µ m (da Cunha et al. 2008,Clemens et al. 2013). Stellar population and dust models arejoined to produce mock SED if they yield similar ISM lumi-nosities, with respect to the total dust emission. The mockSED are fit to the observational data and the fit is evalu-ated using χ . The best-fit physical parameters are selectedas the median of the probability density of the parametervalues weighted by the probability exp (− χ / ) of each fittedmock SED (da Cunha et al. 2008, Clemens et al. 2013).In this work, we utilise estimates of stellar mass M ∗ ,dust mass M dust , total 3-1000 µ m IR luminosity L IR , therelative contributions to L IR from birth clouds L BC , andstar formation rate SFR (see Table A1 and A2 for a listof properties of star-forming cluster and field galaxies, re-spectively). The flux limits quoted in Section 2 (up to 160 µ m , which encloses the dust emission bump at the redshiftconsidered here) are used as upper limits for the fluxes ofthe sources with no detection. This helps in constrainingthe models, in particular at longer wavelength, and avoid-ing non-physical dust masses. Figure 4 shows that the co-moving number density of star-forming cluster galaxies isa factor ∼ higher than the comoving number density ofstar-forming field members. This is not surprising given thatgalaxy clusters such as A1758 are overdense regions, result-ing in high number counts of galaxies per unit volume, com-pared to the field sample. This holds true also when con-sidering star-forming galaxies, given that the volume stud-ied here includes the wider overdense infall region (see alsoHaines et al. 2015). In literature, luminous and ultra lumi-nous infrared galaxies (LIRG and ULIRG) are classified forhaving L IR > L (cid:12) and L IR > L (cid:12) , respectively. It isinteresting to notice the flatter trend of field star-forminggalaxies above LogL IR = . , suggesting a higher fraction ofluminous infrared galaxies in the field than in the cluster.We select a final sample of cluster and field star-forminggalaxies that satisfy M ∗ > M (cid:12) , L IR > . L (cid:12) and SFR > . (cid:12) yr − . The two samples comprise 90 and 68cluster and field star-forming galaxies, respectively. This se-lection extends beyond the level at which the samples can beconsidered complete; see for example the lowest luminositybin in Figure 4. The results described in the following sec-tions are insensitive to whether we restrict our samples tobe more complete. We interpret this as indicating that anyeffects of incompleteness at the faint limit of our samplesaffect both samples in the same manner. MNRAS , 1–16 (2019)
M. Bianconi et al.
Dissecting stellar and dust masses, and SFR aims at mea-suring the impact of infall onto cluster on the star for-mation cycle and how it echoes among these properties.This, when compared to the coeval reference field sample,helps in constraining quenching since dust and stars pro-duction/destruction cycle is susceptible to processes actingin different environments.Overall, both cluster and field star-forming galaxiesspan similar values of stellar and dust mass and SFR, asit can be seen in the different panels of Figure 5. As high-lighted by the average errors distributions in each panel ofFigure 5, dust masses have higher fractional errors ( ≈ )with respect to stellar masses and SFR ( ≈ ). Despitethe high values, such uncertainties are typical for dust massestimation. In the top-left panel of Figure 5, we can seethat our sample presents overall comparable values of stel-lar mass and star formation rates to the COSMOS low red-shift sample ( z ≈ . , Lee et al. 2015). A clear correlationcan be seen in the bottom panel of Figure 5 between SFRand M dust for cluster and field galaxies. This is not surpris-ing given that dust emission consists of re-processed lightfrom newly formed stars. Our sample shows consistent val-ues of star-formation rate and dust masses with respect tothe star-forming galaxies with z < M (cid:63) > M (cid:12) , the meanSFR per unit stellar mass of field and cluster star-forminggalaxies is . × − yr − and . × − yr − respectively,at . σ . This is consistent with the discrepancy in specificSFR between cluster and field star-forming galaxies foundin Haines et al. (2013), suggesting slow quenching (occur-ring on timescales between 0.7-2.0 Gyr) of star formation ingalaxies upon accretion on clusters.Despite the difference between the typical SFR of fieldand cluster star forming galaxies, both populations are moreclosely described by evolutionary models of spiral galaxiesthan by proto-spheroidal starburst galaxies. This can be seenin the better agreement of the Milky Way- and M101-likemodels with our data in Figure 5 than the starburst galaxymodels from Calura et al. (2017). These models describegalaxies spanning increasing range of masses. In particular,spheroidal galaxy models assume total baryonic masses of × , and M (cid:12) . Spiral-like models span approx-imately 13 Gyr of evolution, while spheroidals are limitedto approximately 1 Gyr. For this reason, we display time-steps of 1 Gyr only for spiral-like models. Each model evo-lution starts at low values of stellar mass and SFR in thetop and bottom panels, respectively. This comparison helpshighlighting the different evolutionary path between star-forming field galaxies and cluster members. In particular,we can notice the absence of recent starbursting galaxies inour sample, which are much better described by long-lastingstar formation history of typical spiral galaxies. In addition,we confirm that dust and stellar masses scale almost indis-tinguishably, even for vastly different object, and that SFR is a key parameter to better constrain the evolution of thesegalaxies. We have already seen in the skymaps shown Figure 1 thatstar-forming cluster galaxies appear to reside preferentiallyin the cluster outskirts. We now investigate this further usingthe redshifts from ACReS to disentangle the membershipsof A1758N, A1758S, and A1758-g8. Figure 6 shows the dis-tribution of cluster members in declination versus redshiftplane. This display helps to highlight the different compo-nents of A1758. To facilitate identification, the declinationof the main cluster components, A1758N and S togetherwith A1758-g8, are marked by vertical dashed lines. We canclearly notice a gap in the galaxy distributions belongingto the two clusters A1758N and A1758S, confirming thatthe two systems have yet to encounter each other and mixtheir galaxy populations. Furthermore, star forming galax-ies are funneled towards A1758N and S not isotropically, butrather along separate accretion paths. Overall, this suggeststhat A1758N and S belong to two separate virialized darkmatter halos, which are not yet connected.Colour map codes the dust mass of star-forming clus-ter members. The dearth of star-forming galaxies is evidentin the core of both subclusters, together with the increaseof dusty galaxies towards the cluster outskirts, which cor-respond to the outer edge of merger shocks presented inSchellenberger et al. (2019) (see also X-ray countour edgein Figure 1). North of A1758, an elongation in galaxy dis-tribution suggests channels of accretion, potentially filamen-tary structures, which are not traced by the X-ray emission.Similarly, group members are located perpendicular to theaxis connecting A1758N and S, and might suggest a fur-ther infall channel toward the northern clump. Star-formingcluster members do not follow the distribution of virializedpassive cluster members, as confirmed in Figure 7. In thisplot, both clusters A1758N and A1758S, and group mem-bers are stacked according to their projected distances fromtheir respective cluster/group centre and velocities with re-spect to their average halo redshift. In the case of A1758N,the centre is fixed to be half-way between the two northernBCGs, and for A1758S and A1758-g8 the centre correspondsto the coordinates of the southern BCG and X-ray emissionpeak, respectively. Projected distances and velocities are fur-ther scaled by their halo r and velocity dispersion σ , re-spectively. When compared to the rest of cluster members,star-forming galaxies present a flat distribution of velocities.Velocities displaying a Gaussian distribution peaked aroundlow values are typical of the old virialized population of clus-ter members. On the other hand, a flatter distribution of ve-locities is characteristic of infalling objects. The kurtosis ofthe velocity distribution of star-forming galaxies, which in-clude both clusters and group members, is γ = − . ± . .which is inconsistent at . σ with the γ = . of a Gaus-sian distribution. Jointly, this shows that star-forming galax-ies have been recently accreted onto the cluster potential(Haines et al. 2015). MNRAS , 1–16 (2019) oCuSS: environment, star formation and dust in A1758 Log M [M ] L o g S F R [ M y r ] Log M [M ] L o g M d u s t [ M ] Log SFR[M yr ] L o g M d u s t [ M ] M e, bar = 3 × 10 MM e, bar = 10 MM e, bar = 10 MMWM101Cluster SF membersField SF galaxies
Figure 5.
Top left: stellar mass plotted against SFR for cluster and field star-forming galaxies marked as red circles and blue squares,respectively. The dotted purple line marks the star-formation main sequence at z ≈ . from the COSMOS survey (Lee et al. 2015). Topright: stellar versus dust mass relation. Bottom panel: dust mass plotted versus SFR for cluster members and field star-forming galaxies.The dashed line shows the fit to a sample of star-forming galaxies with z < . from da Cunha et al. (2010). In each panel, the errorbars represents the dispersion associated to the median value of the plotted quantities, and connects the 16th to the 84th percentiles ofeach parameter distribution. Overplotted are the theoretical tracks computed in Calura et al. (2017), marking the evolution of stellarand dust mass and star-formation rate according to different galaxy evolution recipes. In particular, dashed lines and continuous linescorrespond to spheroidal starburst galaxies ( M e , bar ) and spirals of increasing baryonic mass, respectively (see Section 3.3). Spirals aremodelled in order to reproduce Milky Way (MW) and M101- type galaxies.MNRAS , 1–16 (2019) M. Bianconi et al.
Dec(J2000) R e d s h i f t L o g M d u s t [ M ] Figure 6.
Declination plotted with respect to redshift for the spectroscopically confirmed members (black points) of the cluster A1758,divided among the two main subsystems A1758N and A1758S (enclosed in the orange dashed polygon), and A1758-g8 (black crosses).Vertical dashed lines mark the declination of the X-ray centres of each subsystem. The colour-map codes dust mass of star-formingmembers in logarithmic scale.
A useful quantity to assess the efficiency of the dust duty-cycle in galaxies is the dust-to-stellar (DTS) mass ratio. Thisquantity accounts for the amount of dust per unit stellarmass contained in each galaxy. We present here the firststudy of the DTS ratio for a sample of cluster and field star-forming galaxies at intermediate redshift. We can see theglobal values of M dust and DTS for cluster, field and groupstar-forming galaxies in Figure 8. As previously mentioned,star-forming galaxies in both cluster and field present com-parable properties, within ≈ σ . More interestingly, subdi-viding further both samples in bins of stellar mass helps inhighlighting a discrepancy between them. Left panel of Fig-ure 9 shows the DTS ratio as a function of M ∗ . We see alinear trend of the DTS with respect to stellar mass, forboth cluster and field galaxies. More massive galaxies haveless fierce star formation with respect to smaller ones, perunit stellar mass. In addition, cluster galaxies present lowervalues of dust per unit stellar mass, suggesting further con-sumption/destruction due to environment, with respect tofield galaxies. Overall, we measure a shift towards lower val-ues of the DTS in cluster galaxies, with respect to field ob-jects, at ≈ . σ , when adding together the significance ineach bin. Interestingly, the displacement between DTS infield and cluster galaxies parallels that seen in Haines et al.(2013) when comparing specific-SFR of cluster and field star-forming galaxies.Right panel of Figure 9 shows the DTS of the thestacked population of cluster star-forming members accord-ing to their projected distance from both A1758N andA1758S centres, as computed for Figure 7. This plot showsthat among star-forming cluster members there is no strongradial trend of the DTS ratio. A dip in the DTS profile can be seen around ≈ . / r with respect to the northerncentre. This distance corresponds to the location of A1758-g8 (see Figure 6), in the case of A1758N. We notice thatthe DTS value for star-forming cluster members in this binwould be further decreased with the inclusion of group star-forming members, suggesting that group galaxies have lowervalues of DTS, compared to the average population of star-forming cluster members at that clustercentric distance. Wealso report that no significant mass segregation is evidentamong star-forming cluster members, as shown by the aver-age stellar masses in each bin in the right panel of Figure 9. Notwithstanding the large scatter among the cluster andfield star-forming galaxies, the mean SED of each galaxysample shows a clear difference at ∼ µ m , with clus-ter galaxies being fainter than field galaxies (left panel ofFigure 10). Emission at these wavelengths is dominatedby warm dust, polycyclic aromatic hydrocarbides and mid-infrared continuum, which are preferentially located in thesurroundings of birth clouds (da Cunha et al. 2008). There-fore emission from these components appears reduced incluster star-forming galaxies with respect to coeval field ob-jects. Furthermore, we overplot our SED of the ram-pressurestripped galaxy from Ebeling & Kalita (2019). The dip inthe SED of this galaxy at ∼ µ m shows that the emissionfrom birth clouds is reduced in this galaxy, suggesting ram-pressure stripping as a channel for dust removal/destruction.This is further supported by the comparison of IR lumi-nosity from stellar birth clouds L BC (dominated by the dif-fuse warm dust component) with stellar mass for cluster and MNRAS000
A useful quantity to assess the efficiency of the dust duty-cycle in galaxies is the dust-to-stellar (DTS) mass ratio. Thisquantity accounts for the amount of dust per unit stellarmass contained in each galaxy. We present here the firststudy of the DTS ratio for a sample of cluster and field star-forming galaxies at intermediate redshift. We can see theglobal values of M dust and DTS for cluster, field and groupstar-forming galaxies in Figure 8. As previously mentioned,star-forming galaxies in both cluster and field present com-parable properties, within ≈ σ . More interestingly, subdi-viding further both samples in bins of stellar mass helps inhighlighting a discrepancy between them. Left panel of Fig-ure 9 shows the DTS ratio as a function of M ∗ . We see alinear trend of the DTS with respect to stellar mass, forboth cluster and field galaxies. More massive galaxies haveless fierce star formation with respect to smaller ones, perunit stellar mass. In addition, cluster galaxies present lowervalues of dust per unit stellar mass, suggesting further con-sumption/destruction due to environment, with respect tofield galaxies. Overall, we measure a shift towards lower val-ues of the DTS in cluster galaxies, with respect to field ob-jects, at ≈ . σ , when adding together the significance ineach bin. Interestingly, the displacement between DTS infield and cluster galaxies parallels that seen in Haines et al.(2013) when comparing specific-SFR of cluster and field star-forming galaxies.Right panel of Figure 9 shows the DTS of the thestacked population of cluster star-forming members accord-ing to their projected distance from both A1758N andA1758S centres, as computed for Figure 7. This plot showsthat among star-forming cluster members there is no strongradial trend of the DTS ratio. A dip in the DTS profile can be seen around ≈ . / r with respect to the northerncentre. This distance corresponds to the location of A1758-g8 (see Figure 6), in the case of A1758N. We notice thatthe DTS value for star-forming cluster members in this binwould be further decreased with the inclusion of group star-forming members, suggesting that group galaxies have lowervalues of DTS, compared to the average population of star-forming cluster members at that clustercentric distance. Wealso report that no significant mass segregation is evidentamong star-forming cluster members, as shown by the aver-age stellar masses in each bin in the right panel of Figure 9. Notwithstanding the large scatter among the cluster andfield star-forming galaxies, the mean SED of each galaxysample shows a clear difference at ∼ µ m , with clus-ter galaxies being fainter than field galaxies (left panel ofFigure 10). Emission at these wavelengths is dominatedby warm dust, polycyclic aromatic hydrocarbides and mid-infrared continuum, which are preferentially located in thesurroundings of birth clouds (da Cunha et al. 2008). There-fore emission from these components appears reduced incluster star-forming galaxies with respect to coeval field ob-jects. Furthermore, we overplot our SED of the ram-pressurestripped galaxy from Ebeling & Kalita (2019). The dip inthe SED of this galaxy at ∼ µ m shows that the emissionfrom birth clouds is reduced in this galaxy, suggesting ram-pressure stripping as a channel for dust removal/destruction.This is further supported by the comparison of IR lumi-nosity from stellar birth clouds L BC (dominated by the dif-fuse warm dust component) with stellar mass for cluster and MNRAS000 , 1–16 (2019) oCuSS: environment, star formation and dust in A1758 Figure 7.
Left panel: projected distance from the halo centre plotted against velocity for the stacked sample of subclusters A1758N(up-triangles), A1758S (down-triangles) and A1758-g8 members (circles), in units of r and velocity dispersion of their parent halo. Thecolour-bar codes dust masses LogM dust . Empty black symbols mark the non star-forming spectroscopically confirmed cluster members.Right panel: distribution of the velocities of star-forming and non star-forming cluster members plotted as orange hatched and greyhistograms, respectively. field galaxies in Figure 10. At a fixed stellar mass, field star-forming galaxies have higher L BC than cluster star-forminggalaxies. We aim to progress the current knowledge on star-formationquenching by studying how dust in star-forming galaxies isaffected by the local environment within clusters. In par-ticular, this work is the first to explore dust evolution ina sample of star-forming galaxies in the dynamically com-plex cluster A1758 and across its main substructures. A1758allows us to assess the effect of merger-induced shocks onstar-formation and dust properties of clusters members.Uniquely, the LoCuSS dataset used here allows for a di-rect comparison between star-forming cluster members witha coeval sample of mass-matched star-forming field galaxies.Specifically for A1758, our data span far-UV to far-IR wave-lengths and include imaging with
GALEX , Subaru, SDSS,UKIRT,
WISE , Spitzer , and
Herschel , plus 96% completespectroscopic follow-up of candidate star-forming galaxiesfrom ACReS observations with Hectospec on MMT. Galax-ies considered here have been selected for having M star > M (cid:12) , SFR > . (cid:12) yr − .We find that star-forming galaxies, whether located inclusters or in field, span a similar range of stellar and dustmasses, whilst the IR luminosity and SFR of star-formingcluster galaxies are lower than star-forming field galaxies.The dust-to-stellar mass ratio (DTS) of cluster star-forming galaxies is a factor ≈ lower than that of field star-forming galaxies at . σ significance. This result impliesan effect of the cluster environment on both the dust con-tent, and the SFR of star-forming galaxies. Among clustermembers, DTS appears to vary little with respect to clus-tercentric distance in both the north and south part of thecluster, with the exception of galaxies within an infalling X-ray group which overall have lower values of dust per unitstellar mass.Galaxy members of A1758N and S are distributed sep-arately in position and redshift space, suggesting that thetwo subclusters belong to separate virialized dark matterhalos. This is further confirmed by analysing the distribu-tion of star-forming galaxies. Star-forming members are dis-tributed towards the cluster outskirts, distant from the ac-tively merging cores of A1758N, A1758S and the mergershocks. These galaxies are being accreted along separateaccretion paths, rather than isotropically. We verified theabsence of additional virialized substructures associated tostar-forming cluster member, via a comparison with diffuseX-ray emission and weak lensing mass maps from Haineset al. (2018) and Okabe & Smith (2016), respectively. Thissuggests that these galaxies are infalling in isolation. Isolatedgalaxies do not suffer from pre-processing in groups prior totheir infall, thus preserving field-like properties. The com-bination of position and velocity classifies the majority ofcluster star-forming members as recent or back-splash infall-ers, i.e. approaching or receding from their first encounterwith the cluster core (Haines et al. 2015). These galaxieshave spent limited time in the harsh ICM, whereas galaxies MNRAS , 1–16 (2019) M. Bianconi et al.
Cluster SF Field SF Group SF L o g M d u s t [ M ] Cluster SF Field SF Group SF L o g D T S Figure 8. M dust and DTS for cluster, field and group star-forminggalaxies. The solid horizontal line shows the mean of each sam-ple, the coloured box encloses the upper and lower quartiles, andwhiskers extend to an additional of interquartile range, en-closing approximately 3 σ . infalling within the group have already being processed bythe hot intragroup medium. This has been proven effectivein reducing the SFR (Bianconi et al. 2018), and appears tosimilarly influence the dust content of galaxies.Our study extends to higher redshift recent findings ondust consumption and destruction from observational cam-paigns on local clusters (Cortese et al. 2010), which showedevidences of truncation in the radial distribution of multi-phase gas and dust harboured in discs of star-forming galax-ies. This suggests outside-in removal processes due to thecluster environment (Rawle et al. 2012b; Finn et al. 2018).Overall, the bulk of dust mass, which is locked in cold clumpspreferentially within the plane of galaxies, appears unaf-fected by the cluster environment on the timescales consid-ered here. We observe primarily the decrease emission fromwarm dust grains in cluster galaxies with respect to field ob-jects. This reflects the reduction of ionizing radiation fromnewly born stars, following the decrease of SFR, but alsohas been associated to the destruction of small, not shielded,dust grains due to ICM emission (Bocchio 2014; Gjergo et al.2018). Sputtering is proposed to destroy dust grains on veryshort ( yr ) timescales. For sputtering to be effective, dusthas to be removed from the galactic plane where it is shieldedfrom the surrounding ICM radiation. Dust in cluster galaxiesis prone to being removed or destroyed by ram-pressure or shocks passing through the galaxy, resulting in the dearthof infrared luminous galaxies in the cluster cores. Hence,the timescale of ram-pressure stripping ( > yr ) is dom-inating sputtering of dust particles. Interestingly, detailedthermal conduction formalism has been recently included inthe framework of cosmological simulations, showing its ef-ficiency in distributing the energy from the hot ICM intogalaxies and reducing their star formation (Kannan et al.2017).We present here a systematic panchromatic study ofspectroscopically confirmed star-forming cluster galaxies atintermediate redshift, which allows us to explore in detailthe connection between local environment, star formationand dust. In this work, we conclude that A1758 is confirmedto be an actively evolving cluster, composed of two mainsubsystems North and South which belong to dynamicallyseparated large-scale structures, and are accreting galaxiesalong separate paths, rather than isotropically. We observethe effect of the local cluster environment echoing amongthe properties of star-forming members, which include mor-phology, star formation rate, dust emission and masses. Inparticular, star-forming cluster members present diminishedvalues of star formation and dust masses with respect toa coeval sample of star-forming field galaxies. We measuredirectly a decrease in the emission of birth clouds in star-forming cluster members, with respect to field galaxies. Thissuggests that the timescale for transformation upon accre-tion to the cluster is slow. Among the proposed mechanismsresponsible for the transformation, we include ram-pressurestripping, harassment, strangulation, heat conduction fromthe ICM and ICM shocks removing/destroying dust in re-cently accreted galaxies. ACKNOWLEDGEMENTS
MB, GPS, and SLM acknowledge support from the Scienceand Technology Facilities Council through grant numberST/N021702/1. CPH acknowledges the hospitality of theUniversity of Birmingham while part of this work was com-pleted. MB acknowledges Francesco Calura for the helpfuldiscussions and for providing the theoretical evolutionarytracks from Calura et al. (2017).
APPENDIX A: PHYSICAL PROPERTIES OFSTAR-FORMING GALAXIES
In this section, we include additional postage stamps ofstar-forming cluster members covered by HST-ACS imagingfrom the RELICS survey (Coe et al. 2019) and tables withRA, Dec, redshift,
Log ( M ∗ ) , Log ( M ∗ ) , Log ( L IR ) , Log ( M dust ) , Log ( SFR ) of both cluster and field star-forming galaxies. MNRAS000
Log ( M ∗ ) , Log ( M ∗ ) , Log ( L IR ) , Log ( M dust ) , Log ( SFR ) of both cluster and field star-forming galaxies. MNRAS000 , 1–16 (2019) oCuSS: environment, star formation and dust in A1758 Log M [M ] L o g D T S ClusterSFmembersFieldSFgalaxies
Projected distance, R/r L o g D T S A1758NA1758S
Projected distance [Mpc]
Figure 9.
Left panel: DTS mass ratio in equi-numeric bins plotted with respect to stellar mass. Red circles and blue squares mark fieldand cluster star-forming galaxies respectively. Right panel: 1-d profile of the DTS of the stacked star-forming cluster members of bothA1758 N and S with respect to projected clustercentric distance in units of r . The top-axis is computed using a mean r = . .The green and yellow profiles are computed with respect to the centre of A1758N and A1758S, respectively. Each point corresponds tothe mean quantity per distance bin, each of which contains on average ≈
20 and ≈ σ confidence interval. The mean Log M (cid:63) in each projected distance bin is overplotted. Log M * [M ] L o g L B C [ L ] ClusterSFmembersFieldSFgalaxies
Figure 10.
Left panel: individual (dashed lines) and mean (solid line) SED for both clusters (red) and field (blue) star-forming galaxies.Overlaid in green, the SED of the ram-pressure stripped galaxy described in Ebeling & Kalita (2019). Right panel: infrared luminosityfrom birth clouds L BC versus stellar mass of both clusters (red circles) and field (blue squares) star-forming galaxies, which is emittedpreferentially around 100 µ m.MNRAS , 1–16 (2019) M. Bianconi et al.
RA (J2000) +50.554°+50.556° D e c ( J ) RA (J2000) +50.558°+50.560° D e c ( J ) RA (J2000) +50.542°+50.544° D e c ( J ) RA (J2000) +50.536°+50.538° D e c ( J ) RA (J2000) +50.518°+50.520° D e c ( J ) RA (J2000) +50.530°+50.532° D e c ( J ) RA (J2000) +50.564°+50.566° D e c ( J ) RA (J2000) +50.506°+50.508° D e c ( J ) RA (J2000) +50.522°+50.524° D e c ( J ) Figure 11.
Postage stamps ( ≈
30 kpc in radius) centered on star-forming cluster members from the HST-ACS RELICS survey (Coeet al. 2019), in F435W, F606W and F814W filters. The top-right galaxy is the ram-pressure stripping candidate from Ebeling & Kalita(2019). MNRAS000
30 kpc in radius) centered on star-forming cluster members from the HST-ACS RELICS survey (Coeet al. 2019), in F435W, F606W and F814W filters. The top-right galaxy is the ram-pressure stripping candidate from Ebeling & Kalita(2019). MNRAS000 , 1–16 (2019) oCuSS: environment, star formation and dust in A1758 Table A1: Star-forming cluster galaxies’ properties from SED fitting.From left to right: RA, Dec, redshift,
Log ( M ∗ ) , Log ( L IR ) , Log ( M dust ) , Log ( SFR ) .RA(J2000) Dec(J2000) Redshift Log ( M ∗ [ M (cid:12) )] Log ( L IR [ L (cid:12) )] Log ( M dust [ M (cid:12) )] Log ( SFR [ M (cid:12) yr − )] MNRAS , 1–16 (2019) M. Bianconi et al.
Log ( M ∗ ) , Log ( L IR ) , Log ( M dust ) , Log ( SFR ) .RA(J2000) Dec(J2000) Redshift Log ( M ∗ [ M (cid:12) )] Log ( L IR [ L (cid:12) )] Log ( M dust [ M (cid:12) )] Log ( SFR [ M (cid:12) yr − )] MNRAS000
Log ( M ∗ ) , Log ( L IR ) , Log ( M dust ) , Log ( SFR ) .RA(J2000) Dec(J2000) Redshift Log ( M ∗ [ M (cid:12) )] Log ( L IR [ L (cid:12) )] Log ( M dust [ M (cid:12) )] Log ( SFR [ M (cid:12) yr − )] MNRAS000 , 1–16 (2019) oCuSS: environment, star formation and dust in A1758 MNRAS , 1–16 (2019) M. Bianconi et al.
References
Bianconi M., Smith G. P., Haines C. P., McGee S. L., FinoguenovA., Egami E., 2018, MNRAS, 473, L79Bocchio M., 2014, PhD thesis, Institut d’Astrophysique Spatiale(IAS), UMR 8617, CNRS/Universit´e Paris-Sud, 91405 Orsay,FranceBruzual G., Charlot S., 2003, MNRAS, 344, 1000Calura F., et al., 2017, MNRAS, 465, 54Chabrier G., 2003, PASP, 115, 763Charlot S., Fall S. M., 2000, ApJ, 539, 718Clemens M. S., et al., 2013, MNRAS, 433, 695Coe D., et al., 2019, arXiv e-prints, p. arXiv:1903.02002Cortese L., et al., 2010, A&A, 518, L49Cortese L., et al., 2012, A&A, 540, A52David L. P., Kempner J., 2004, ApJ, 613, 831Davis T. A., et al., 2013, MNRAS, 429, 534Durret F., Lagan´a T. F., Haider M., 2011, A&A, 529, A38Ebeling H., Kalita B. S., 2019, arXiv e-prints, p. arXiv:1907.12781Finn R. A., et al., 2018, ApJ, 862, 149Galliano F., Galametz M., Jones A. P., 2018, ARA&A, 56, 673Gjergo E., Granato G. L., Murante G., Ragone-Figueroa C., Tor-natore L., Borgani S., 2018, MNRAS, 479, 2588Gobat R., et al., 2018, Nature Astronomy, 2, 239Gomez H. L., et al., 2010, A&A, 518, L45Gunn J. E., Gott III J. R., 1972, ApJ, 176, 1Haines C. P., Smith G. P., Pereira M. J., Egami E., Moran S. M.,Hardegree-Ullman E., Rawle T. D., Rex M., 2010, A&A, 518,L19Haines C. P., Pereira M. J., et al. S., 2013, ApJ, 775, 126Haines C. P., Pereira M. J., et al. S., 2015, ApJ, 806, 101Haines C. P., et al., 2018, MNRAS, 477, 4931Jablonka P., Combes F., Rines K., Finn R., Welch T., 2013, A&A,557, A103Jones A. P., 2004, in Witt A. N., Clayton G. C., Draine B. T.,eds, Astronomical Society of the Pacific Conference Series Vol.309, Astrophysics of Dust. p. 347Kannan R., Vogelsberger M., Pfrommer C., Weinberger R.,Springel V., Hernquist L., Puchwein E., Pakmor R., 2017,ApJ, 837, L18Koyama S., et al., 2017, ApJ, 847, 137Larson R. B., Tinsley B. M., Caldwell C. N., 1980, ApJ, 237, 692Leauthaud A., et al., 2010, ApJ, 709, 97Lee N., et al., 2015, ApJ, 801, 80Machado R. E. G., Monteiro-Oliveira R., Lima Neto G. B., Cypri-ano E. S., 2015, MNRAS, 451, 3309Martino R., Mazzotta P., Bourdin H., Smith G. P., BartalucciI., Marrone D. P., Finoguenov A., Okabe N., 2014, MNRAS,443, 2342Mok A., et al., 2016, MNRAS, 456, 4384Monteiro-Oliveira R., Lima Neto G. B., Cypriano E. S., MachadoR. E. G., Capelato H. V., Lagan´a T. F., Durret F., Bagchi J.,2017, MNRAS, 468, 4566Moore B., Katz N., Lake G., 1996, ApJ, 457, 455Okabe N., Smith G. P., 2016, MNRAS, 461, 3794Pereira M. J., et al., 2010, A&A, 518, L40Rawle T. D., et al., 2012a, ApJ, 747, 29Rawle T. D., et al., 2012b, ApJ, 756, 106Schellenberger G., David L. P., O’Sullivan E., Vrtilek J. M.,Haines C. P., 2019, arXiv e-prints, p. arXiv:1907.10581Smith G. P., et al., 2010, A&A, 518, L18Smith D. J. B., et al., 2012, MNRAS, 427, 703Voit G. M., 2005, Reviews of Modern Physics, 77, 207Walter F., et al., 2011, ApJ, 726, L11da Cunha E., Charlot S., Elbaz D., 2008, MNRAS, 388, 1595da Cunha E., Eminian C., Charlot S., Blaizot J., 2010, MNRAS,403, 1894 This paper has been typeset from a TEX/L A TEX file prepared bythe author. MNRAS000