Massive Star Cluster Formation and Destruction in Luminous Infrared Galaxies in GOALS. II. An ACS/WFC3 Survey of Nearby LIRGs

We present the results of a Hubble Space Telescope ACS/HRC FUV, ACS/WFC optical study into the cluster populations of a sample of 22 Luminous Infrared Galaxies in the Great Observatories All-Sky LIRG Survey. Through integrated broadband photometry, we have derived ages and masses for a total of 484 star clusters contained within these systems. This allows us to examine the properties of star clusters found in the extreme environments of LIRGs relative to lower luminosity star-forming galaxies in the local universe. We find that by adopting a Bruzual & Charlot simple stellar population model and Salpeter initial mass function, the age distribution of the clusters declines as , consistent with the age distribution derived for the Antennae Galaxies, and interpreted as evidence for rapid cluster disruption occurring in the strong tidal fields of merging galaxies. The large number of young clusters identified in the sample also suggests that LIRGs are capable of producing more high-mass clusters than what is observed to date in any lower luminosity star-forming galaxy in the local universe. The observed cluster mass distribution of is consistent with the canonical −2 power law used to describe the underlying initial cluster mass function (ICMF) for a wide range of galactic environments. We interpret this as evidence against mass-dependent cluster disruption, which would flatten the observed CMF relative to the underlying ICMF distribution.

Context. Despite the nearly hundred gravitational-wave detections reported by the LIGO-Virgo-KAGRA Collaboration, the question of the cosmological origin of merging binary black holes (BBHs) remains open. The two main formation channels generally considered are from isolated field binaries or via dynamical assembly in dense star clusters. Aims. Here we focus on understanding the dynamical formation of merging BBHs within massive clusters in galaxies of different masses. Methods. To this end, we applied a new framework to consistently model the formation and evolution of massive star clusters in zoom-in cosmological simulations of galaxies. Each simulation, taken from the FIRE project, provides a realistic star formation environment, with a unique star formation history, that hosts realistic giant molecular clouds that constitute the birthplace of star clusters. Combined with the code for star cluster evolution CMC, we are able to produce populations of dynamically formed merging BBHs across cosmic time in different environments. Results. As the most massive star clusters preferentially form in dense massive clouds of gas, we find that, despite their low metallicities favouring the creation of black holes, low-mass galaxies contain few massive clusters and therefore make a limited contribution to the global production of dynamically formed merging BBHs. Furthermore, we find that massive clusters can host hierarchical BBH mergers with clear, identifiable physical properties. Looking at the evolution of the BBH merger rate in different galaxies, we find strong correlations between BBH mergers and the most extreme episodes of star formation. Finally, we discuss the implications for future LIGO-Virgo-KAGRA gravitational wave observations.

Young massive star clusters (YMCs, with M $\geq$10$^4$ M$_{\odot}$) are proposed modern-day analogues of the globular clusters (GCs) that were products of extreme star formation in the early universe. The exact conditions and mechanisms under which YMCs form remain unknown -- a fact further complicated by the extreme radiation fields produced by their numerous massive young stars. Here we show that GC-sized clusters are naturally produced in radiation-hydrodynamic simulations of isolated 10$^7$ M$_{\odot}$ Giant Molecular Clouds (GMCs) with properties typical of the local universe, even under the influence of radiative feedback. In all cases, these massive clusters grow to GC-level masses within 5 Myr via a roughly equal combination of filamentary gas accretion and mergers with several less massive clusters. Lowering the heavy-element abundance of the GMC by a factor of 10 reduces the opacity of the gas to radiation and better represents the high-redshift formation conditions of GCs. This results in higher gas accretion leading to a mass increase of the largest cluster by a factor of ~4. When combined with simulations of less massive GMCs (10$^{4-6}$ M$_{\odot}$), a clear relation emerges between the maximum YMC mass and the mass of the host GMC. Our results demonstrate that YMCs, and potentially GCs, are a simple extension of local cluster formation to more massive clouds and do not require suggested exotic formation scenarios.

We compare the observed mass functions and age distributions of star clusters in six well-studied galaxies: the Milky Way, Magellanic Clouds, M83, M51, and Antennae. In combination, these distributions span wide ranges of mass and age: $10^2\lea M/M_{\odot}\lea10^6$ and $10^6\lea\tau/yr \lea10^9$. We confirm that the distributions are well represented by power laws: $dN/dM\propto M^{\beta}$ with $\beta \approx-1.9$ and $dN/d\tau\propto\tau^{\gamma}$ with $\gamma\approx -0.8$. The mass and age distributions are approximately independent of each other, ruling out simple models of mass-dependent disruption. As expected, there are minor differences among the exponents, at a level close to the true uncertainties, $\epsilon_{\beta}\sim\epsilon_{\gamma}\sim$~0.1--0.2. However, the overwhelming impression is the similarity of the mass functions and age distributions of clusters in these different galaxies, including giant and dwarf, quiescent and interacting galaxies. This is an important empirical result, justifying terms such as "universal" or "quasi-universal." We provide a partial theoretical explanation for these observations in terms of physical processes operating during the formation and disruption of the clusters, including star formation and feedback, subsequent stellar mass loss, and tidal interactions with passing molecular clouds. A full explanation will require additional information about the molecular clumps and star clusters in galaxies beyond the Milky Way.

The Great Observatories All-Sky LIRG Survey (GOALS) is a comprehensive,\nmultiwavelength study of luminous infrared galaxies (LIRGs) in the local\nuniverse. Here we present the results of a multi-component, spectral\ndecomposition analysis of the low resolution mid-IR Spitzer IRS spectra from\n5-38um of 244 LIRG nuclei. The detailed fits and high quality spectra allow for\ncharacterization of the individual PAH features, warm molecular hydrogen\nemission, and optical depths for silicate dust grains and water ices. We find\nthat starbursting LIRGs, which make up the majority of GOALS, are very\nconsistent in their MIR properties (i.e. tau_9.7um, tau_ice, neon line and PAH\nfeature ratios). However, as their PAH EQW decreases, usually an indicator of\nan increasingly dominant AGN, LIRGs cover a larger spread in these MIR\nparameters. The contribution from PAHs to the total L(IR) in LIRGs varies from\n2-29% and LIRGs prior to their first encounter show higher L(PAH)/L(IR) ratios\non average. We observe a correlation between the strength of the starburst\n(IR8) and the PAH fraction at 8um but not with the 7.7 to 11.3 PAH ratio,\nsuggesting the fractional PDR emission, and not the overall grain properties,\nis associated with the rise in IR8 for galaxies off the starburst main\nsequence. We detect crystalline silicate features in 6% of the sample but only\nin the most obscured sources (s_9.7um < -1.24). Ice absorption features are\nobserved in 11% (56%) of GOALS LIRGs (ULIRGs). Most GOALS LIRGs have\nL(H2)/L(PAH) ratios elevated above those observed for normal star-forming\ngalaxies and exhibit a trend for increasing L(H2)/L(PAH) ratio with increasing\nL(H2). While star formation appears to be the dominant process responsible for\nexciting the H2 in most of the GOALS galaxies, a subset of LIRGs (10%) show\nexcess H2 emission that is inconsistent with PDR models and may be excited by\nshocks or AGN-induced outflows.\n

We have used previously published observations of the CO emission from the Antennae (NGC 4038/4039) to study the detailed properties of the supergiant molecular complexes with the goal of understanding the formation of young massive star clusters. Over a mass range from 5 × 106 to 9 × 108 M☉ the molecular complexes follow a power-law mass function with a slope of -1.4 ± 0.1, which is very similar to the slope seen at lower masses in molecular clouds and cloud cores in the Galaxy. Compared with the spiral galaxy M51, which has a similar surface density and total mass of molecular gas, the Antennae contain clouds that are an order of magnitude more massive. Many of the youngest star clusters lie in the gas-rich overlap region, where extinctions as high as AV ~ 100 mag imply that the clusters must lie in front of the gas. Young clusters found in other regions of the galaxies can be as far as 2 kpc from the nearest massive cloud, which suggests that either young clusters can form occasionally in clouds less massive than 5 × 106 M☉ or that these young clusters have already destroyed their parent molecular clouds. Combining data on the young clusters, thermal and nonthermal radio sources, and the molecular gas suggests that young massive clusters could have formed at a constant rate in the Antennae over the last 160 Myr and that sufficient gas exists to sustain this cluster formation rate well into the future. However, this conclusion requires that a very high fraction of the massive clusters that form initially in the Antennae do not survive as long as 100 Myr. Furthermore, if most young massive clusters do survive for long periods, the Antennae must be experiencing a relatively short burst of cluster formation to prevent the final merger remnant from exceeding the observed specific frequency of star clusters in elliptical galaxies by a wide margin. Finally, we compare our data with two models for massive star cluster formation and conclude that the model in which young massive star clusters form from dense cores within the observed supergiant molecular complexes is most consistent with our current understanding of this merging system.

Young star clusters like R136 in the Large Magellanic Cloud and NGC 3603, Westerlund 1, and 2 in the Milky Way are dynamically more evolved than expected based on their current relaxation times. In particular, the combination of a high degree of mass segregation, a relatively low central density, and the large number of massive runaway stars in their vicinity are hard to explain with the monolithic formation of these clusters. Young star clusters can achieve such a mature dynamical state if they formed through the mergers of a number of less massive clusters. The shorter relaxation times of less massive clusters cause them to dynamically evolve further by the time they merge, and the merger product preserves the memory of the dynamical evolution of its constituent clusters. With a series of $N$-body simulations, we study the dynamical evolution of single massive clusters and those that are assembled through merging smaller clusters together. We find that the formation of massive star clusters through the mergers of smaller clusters can reproduce the currently observed spatial distribution of massive stars, the density, and the characteristics (number and mass distribution) of the stars ejected as runaways from young dense clusters. We therefore conclude that these clusters and possibly other young massive star clusters formed through the mergers of smaller clusters.

We present a new implementation of star formation in cosmological simulations by considering star clusters as a unit of star formation. Cluster particles grow in mass over several million years at the rate determined by local gas properties, with high time resolution. The particle growth is terminated by its own energy and momentum feedback on the interstellar medium. We test this implementation for Milky Way-sized galaxies at high redshift by comparing the properties of model clusters with observations of young star clusters. We find that the cluster initial mass function is best described by a Schechter function rather than a single power law. In agreement with observations, at low masses the logarithmic slope is , while the cutoff at high mass scales with the star formation rate (SFR). A related trend is a positive correlation between the surface density of the SFR and fraction of stars contained in massive clusters. Both trends indicate that the formation of massive star clusters is preferred during bursts of star formation. These bursts are often associated with major-merger events. We also find that the median timescale for cluster formation ranges from 0.5 to 4 Myr and decreases systematically with increasing star formation efficiency. Local variations in the gas density and cluster accretion rate naturally lead to the scatter of the overall formation efficiency by an order of magnitude, even when the instantaneous efficiency is kept constant. Comparison of the formation timescale with the observed age spread of young star clusters provides an additional important constraint on the modeling of star formation and feedback schemes.

The Great Observatories All-Sky LIRG Survey (GOALS) is a multiwavelength study of luminous infrared galaxies (LIRGs) in the local universe. Here we present low resolution Spitzer spectra covering 5-38um and provide a basic analysis of the mid-IR spectral properties for nearby LIRGs. In a companion paper, we discuss detailed fits to the spectra. The GOALS sample of 244 nuclei in 180 luminous and 22 ultraluminous IR galaxies represents a complete subset of the IRAS RBGS and covers a range of merger stages, morphologies and spectral types. The majority (>60%) of GOALS LIRGs have high 6.2um PAH equivalent widths (EQW > 0.4um) and low levels of silicate absorption (s_9.7um >-1.0). There is a general trend among the U/LIRGs for silicate depth and MIR slope to increase with LIR. U/LIRGs in the late stages of a merger also have on average steeper MIR slopes and higher levels of dust obscuration. Together these trends suggest that as gas & dust is funneled towards the center of a coalescing merger, the nuclei become more compact and obscured. The sources that depart from these correlations have very low PAH EQW (EQW < 0.1um) consistent with their MIR emission being dominated by an AGN. The most heavily dust obscured sources are the most compact in their MIR emission, suggesting that the obscuring (cool) dust is associated with the outer regions of the starburst. As the merger progresses a marked decline is seen for the fraction of high EQW (star formation dominated) sources while the fraction of composite sources increases but the fraction of AGN-dominated sources remains low. When compared to the MIR spectra of submillimeter galaxies (SMGs) at z~2, the average GOALS LIRG is more absorbed at 9.7um and has more PAH emission. However, when the AGN contributions to both the local LIRGs and the high-z SMGs are removed, the average local starbursting LIRG closely resembles the starbursting SMGs.

The mode of star formation that results in the formation of globular clusters and young massive clusters is difficult to constrain through observations. We present models of massive star cluster formation using the TORCH framework, which uses the Astrophysical MUltipurpose Software Environment (AMUSE) to couple distinct multi-physics codes that handle star formation, stellar evolution and dynamics, radiative transfer, and magnetohydrodynamics. We upgraded TORCH by implementing the N-body code PETAR, thereby enabling TORCH to handle massive clusters forming from 106 M⊙ clouds with ≥105 individual stars. We present results from TORCH simulations of star clusters forming from 104, 105, and 106 M⊙ turbulent spherical gas clouds (named M4, M5, M6) of radius R = 11.7 pc. We find that star formation is highly efficient and becomes more so at a higher cloud mass and surface density. For M4, M5, and M6 with initial surface densities 2.325 × 101,2,3 M⊙ pc−2, after a free-fall time of tff = 6.7,2.1,0.67 Myr, we find that ∼30%, 40%, and 60% of the cloud mass has formed into stars, respectively. The end of simulation-integrated star formation efficiencies for M4, M5, and M6 are ϵ⋆ = M⋆/Mcloud = 36%, 65%, and 85%. Observations of nearby clusters similar in mass and size to M4 have instantaneous star formation efficiencies of ϵinst ≤ 30%, which is slightly lower than the integrated star formation efficiency of M4. The M5 and M6 models represent a different regime of cluster formation that is more appropriate for the conditions in starburst galaxies and gas-rich galaxies at high redshift, and that leads to a significantly higher efficiency of star formation. We argue that young massive clusters build up through short efficient bursts of star formation in regions that are sufficiently dense (Σ ≥ 102 M⊙ pc−2) and massive (Mcloud ≥ 105 M⊙). In such environments, stellar feedback from winds and radiation is not strong enough to counteract the gravity from gas and stars until a majority of the gas has formed into stars.

Context. The formation of globular clusters remains an open debate. Dwarf starburst galaxies are efficient at forming young massive clusters with similar masses as globular clusters and may hold the key to understanding their formation. Aims. We study star cluster formation in a tidal debris, including the vicinity of three tidal dwarf galaxies, in a massive gas-dominated collisional ring around NGC 5291. These dwarfs have physical parameters that differ significantly from local starbursting dwarfs. They are gas rich, highly turbulent, their gas metallicity is already enriched up to half solar values, and they are expected to be free of dark matter. The aim is to study massive star cluster formation in this as yet unexplored type of environment. Methods. We used imaging from the Hubble Space Telescope using broadband filters that cover the wavelength range from the near-ultraviolet to the near-infrared. We determined the masses and ages of the cluster candidates by using the spectral energy distribution-fitting code CIGALE. We considered age-extinction degeneracy effects on the estimation of the physical parameters. Results. We find that the tidal dwarf galaxies in the ring of NGC 5291 are forming star clusters with an average efficiency of ∼40%, which is similar to blue compact dwarf galaxies. We also find massive star clusters for which the photometry suggests that they were formed at the very birth of the tidal dwarf galaxies. These clusters have survived for several hundred million years. Therefore our study shows that extended tidal dwarf galaxies and compact clusters may be formed simultaneously. In the specific case observed here, the young star clusters are not massive enough to survive for a Hubble time. However, it may be speculated that similar objects at higher redshift, with a higher star formation rate, might form some of the long-lived globular clusters.

We present the first results of a high-resolution Karl G. Jansky Very Large Array imaging survey of luminous and ultra-luminous infrared galaxies (U/LIRGs) in the Great Observatories All-sky LIRG Survey. From the full sample of 68 galaxies, we have selected 25 luminous infrared galaxies (LIRGs) that show resolved extended emission at sufficient sensitivity to image individual regions of star formation activity beyond the nucleus. With wideband radio continuum observations, which sample the frequency range from 3 to 33 GHz, we have made extinction-free measurements of the luminosities and spectral indicies for a total of 48 individual star-forming regions identified as having deprojected galactocentric radii (r G ) that lie outside the 13.2 μm core of the galaxy. The median 3–33 GHz spectral index and 33 GHz thermal fraction measured for these “extranuclear” regions is −0.51 ± 0.13 and 65% ± 11%, respectively. These values are consistent with measurements made on matched spatial scales in normal star-forming galaxies, and suggests that these regions are more heavily dominated by thermal free–free emission relative to the centers of local U/LIRGs. Further, we find that the median star formation rate derived for these regions is ∼1 M ⊙ yr−1, and when we place them on the sub-galactic star-forming main sequence of galaxies (SFMS), we find they are offset from their host galaxies’ globally averaged specific star formation rates. We conclude that while nuclear starburst activity drives LIRGs above the SFMS, extranuclear star formation still proceeds in a more extreme fashion relative to what is seen in local spiral galaxies.

Stellar winds of massive ($\gtrsim 9\, \mathrm{M_\odot }$) and very massive ($\gtrsim 100\, \mathrm{M_\odot }$) stars may play an important role in the metal-enrichment during the formation of star clusters. With novel high-resolution hydrodynamical griffin-project simulations, we investigate the rapid recycling of stellar wind-material during the formation of massive star clusters up to $M_\mathrm{cluster}\sim 2\times 10^5\, \mathrm{M_\odot }$ in a low-metallicity dwarf galaxy starburst. The simulation realizes new stars from a stellar initial mass function (IMF) between $0.08$ and $\sim 400\, \mathrm{M_\odot }$ and follows stellar winds, radiation and supernova-feedback of single massive stars with evolution tracks. Star clusters form on time-scales less than ∼5 Myr, and their supernova-material is very inefficiently recycled. Stellar wind-material, however, is trapped in massive clusters resulting in the formation of stars self-enriched in Na, Al, and N within only a few Myr. Wind-enriched (second population, 2P) stars can be centrally concentrated in the most massive clusters ($\gtrsim 10^4\, \mathrm{M_\odot }$) and the locked wind-material increases approximately as $M_\mathrm{cluster}^{2}$. These trends resemble the characteristics of observed 2P stars in globular clusters (GCs). We fit scaling relations to the lognormal distributed wind-mass fractions and extrapolate to possible GC progenitors of $M_\mathrm{cluster}=10^7\, \mathrm{M_\odot }$ to investigate whether a dominant 2P could form. This can only happen if the IMF is well-sampled, single massive stars produce at least a factor of a few more enriched winds, for example, through a top-heavy IMF, and a significant fraction of the first population (unenriched) stars is lost during cluster evolution.

To study the formation of star clusters and their properties in a dwarf–dwarf merging galaxy, we have performed a numerical simulation of a dwarf–dwarf galaxy merger by using the Tree+GRAPE $N$-body/SPH code ASURA. In our simulation, 13 young star clusters are formed during the merger process. We show that our simulated star clusters can be divided into two types: with and without [Fe$/$H] abundance variations. The former is created by a seed star cluster (the first-generation stars) formed in compressed gas. These stars contaminate the surrounding gas by Type II supernovae. At that time, the energy injection is insufficient to induce an outflow of the surrounding gas. After that, the contaminated gas falls into the seed, thereby forming a new generation of stars from the contaminated gas. We also show that most star clusters are formed in the galactic central region after the second encounter and fall into the galactic center due to dynamical friction within several hundred Myr. As a result, close encounters and mergers between the clusters take place. Although the clusters with shallower gravitational potential are tidally disrupted by these close encounters, others survive and finally merge at the center of the merged dwarf galaxies to create a nuclear star cluster. Therefore, the nuclear star cluster comprises various stellar components in ${[\rm Fe/H]}$ abundance and age. We discuss our work in the context of observations and demonstrate the diagnostic power of high-resolution simulations in the context of star cluster formation.

The Great Observatories All-sky LIRG Survey (GOALS) is combining imaging and spectroscopic data from the Herschel, Spitzer, Hubble, GALEX, Chandra, and XMM-Newton space telescopes augmented with extensive ground-based observations in a multiwavelength study of approximately 180 Luminous Infrared Galaxies (LIRGs) and 20 Ultraluminous Infrared Galaxies (ULIRGs) that comprise a statistically complete subset of the 60μm-selected IRAS Revised Bright Galaxy Sample. The objects span the full range of galaxy environments (giant isolated spirals, wide and close pairs, minor and major mergers, merger remnants) and nuclear activity types (Seyfert 1, Seyfert 2, LINER, starburst/HII), with proportions that depend strongly on the total infrared luminosity. I will review the science motivations and present highlights of recent results selected from over 25 peer-reviewed journal articles published recently by the GOALS Team. Statistical investigations include detection of high-ionization Fe K emission indicative of deeply embedded AGN, comparison of UV and far-IR properties, investigations of the fraction of extended emission as a function of wavelength derived from mid-IR spectroscopy, mid-IR spectral diagnostics and spectral energy distributions revealing the relative contributions of AGN and starbursts to powering the bolometric luminosity, and quantitative structure analyses that delineate the evolution of stellar bars and nuclear stellar cusps during the merger process. Multiwavelength dissections of individual systems have unveiled large populations of young star clusters and heavily obscured AGN in early-stage (II Zw 96), intermediate-stage (Mrk 266, Mrk 273), and late-stage (NGC 2623, IC 883) mergers. A recently published study that matches numerical simulations to the observed morphology and gas kinematics in mergers has placed four systems on a timeline spanning 175-260 million years after their first passages, and modeling of additional (U)LIRGs is underway. A very brief description of ongoing work with Herschel and ALMA will be given. The talk will conclude with a discussion of the demographics of dual AGN (kpc-scale orbits) in the GOALS sample, including the difficulty of their detection and confirmation, a proposed sequence representing a progression from dual AGN to binary AGNs (sub-pc scale orbits), and implications for the scarcity of confirmed binary QSOs in recent large surveys. Grant support from NASA is gratefully acknowledged.