GOALS-JWST: Revealing the Buried Star Clusters in the Luminous Infrared Galaxy VV 114

Studies of star formation in nearby galaxies across a wide variety of environments reveal that the majority of star formation occurs in stellar clusters, groups, and associations (e.g., Lada & Lada 2003). Star formation activity is particularly intense within starburst and merging galaxies, where thousands of compact clusters (R_eff ∼ 3–5 pc), many with masses M ≥ 10⁴ M_⊙, are readily observed (e.g., Whitmore et al. 2010; Mulia et al. 2015; Randriamanakoto et al. 2019; Adamo et al. 2020). At the earliest phases of their evolution, these young massive star clusters (YMCs) will experience strong feedback from O-star winds, and may become unbound as the remaining gas in the parent giant molecular cloud (GMC) is expelled within a crossing time (Krause et al. 2020). However, extremely high-pressure regions of the interstellar medium (ISM) will increase the difficulty in dispersing gas from the cluster and may actually result in a higher cluster formation efficiency (CFE; Kruijssen et al. 2011).

Observations of YMCs in nearby galaxies provide evidence for fundamental differences in how feedback operates in different environments. For example, Whitmore & Zhang (2002) identified a small population of "very red clusters" in the Antennae merger whose optical colors suggest that they are young (t = 4–6 Myr), massive (M = 10^5–6 M_⊙), and very dusty (A_V = 5–7). In contrast, observations of the low-metallicity dwarf NGC 4449 reveal that many of the youngest clusters (M ∼ 10⁴ M_⊙, t = 3–5 Myr) have relatively low dust extinction (A_V ≤ 1.5), suggesting efficient removal of the remaining gas and dust (Reines et al. 2008).

Simulations suggest that YMCs formed within the densest regions of the ISM may host ionized stellar winds that slow down due to critical radiative cooling, ultimately allowing the cluster to retain fuel for ongoing (or future) generations of star formation (Silich et al. 2020). Constraining this short-lived phase in YMC evolution remains a critical missing piece for understanding how star cluster formation proceeds in different environments.

Before the onset of stellar feedback, YMCs are deeply embedded within their birth clouds of gas and dust and suffer large internal extinction (A_V ∼ 7–10), rendering them effectively unobservable at UV-optical wavelengths. Therefore, these objects are consistently missed in studies of early cluster feedback and evolution that rely on UV-optical-based diagnostic tools (Hollyhead et al. 2016; Hannon et al. 2022). At longer wavelengths, near- and mid-infrared emission can be used to trace the dust cocoons surrounding newly formed YMCs (Corbelli et al. 2017; Messa et al. 2021). However, the resolution of Spitzer at λらむだ = 3–5μみゅーm (∼2'') only allowed individual clusters to be resolved for galaxies at very close distances, making it difficult to build meaningful samples of embedded YMCs at higher-redshifts and in different environments.

Luminous IR galaxies (LIRGs: defined as L_IR [8–1000 μみゅーm] ≥10^11.0 L_⊙) host the most extreme stellar nurseries in the local universe. The activity in LIRGs is primarily triggered by gas-rich galaxy interactions, and at their peak, local LIRGs have star formation rates (SFRs) ∼100 times higher than the Milky Way, placing them well above the star formation main sequence (SFMS) (Speagle et al. 2014). Many local LIRGs have individual star-forming clumps with sizes of 50–100pc and SFR surface densities (Σしぐま_SFR) of 0.1–10 M_⊙ yr⁻¹ kpc⁻²) that rival those of high-z galaxies (Larson et al. 2020). Therefore, local LIRGs represent the ideal laboratory for studying YMC formation and evolution in extreme environments at high resolution.

With the unprecedented sensitivity and resolving power of the James Webb Space Telescope (JWST), we are now able to identify and characterize individual YMCs forming in the densest and dustiest regions of the ISM in galaxies out to ∼100 Mpc in the near-infrared (NIR). In this Letter, we present JWST NIR observations of the embedded YMC population in the starburst LIRG VV 114 (VV 114 = IC 1623). VV 114 is classified as an early-stage merger with the two galaxies aligned east–west and a projected nuclear separation of ∼8 kpc (VV 114E and VV 114W, respectively). Previous HST observations of the galaxy have been used to characterize the UV-bright population of clusters found almost exclusively in VV 114W (Linden et al. 2021). The observations described in this Letter represent the most comprehensive census to date of YMC formation and evolution in VV 114 in the NIR and sets the stage for future JWST studies of YMC formation in nearby LIRGs.

Throughout this Letter, we adopt a WMAP cosmology of H₀ = 69.3 km s⁻¹ Mpc⁻¹, Ωおめが_matter = 0.286, and Ωおめが_Λらむだ = 0.714 (e.g., see Hinshaw et al. 2013).

JWST imaging of VV 114 was obtained as part of the Early Release Science (ERS) program "A JWST Study of the Starburst-AGN Connection in Merging LIRGs" (PID 1328; co-PIs: Armus, Evans). Additional studies focusing on the medium-resolution spectrometer (MRS) and mid-infrared instrument (MIRI) observations of VV 114 are presented in Rich et al. (2022) and Evans et al. (2022), respectively.

Near-infrared camera (NIRCam; Rieke et al. 2005) observations of VV 114 were taken on 2022 July 29 and retrieved from the Mikulski Archive for Space Telescopes (MAST). The galaxy was imaged using the F150W (λらむだ = 1.5 μみゅーm) and F200W (λらむだ = 2.0 μみゅーm) short-wavelength (SW) and F356W (λらむだ = 3.5 μみゅーm) and F444W (λらむだ = 4.4 μみゅーm) long-wavelength (LW) filters for 2319 s and 580 s, respectively, with module B. These observations utilized the "FULL" array with the "RAPID" readout mode and the "INTRAMODEL" three-point dither pattern.

The raw data have been reduced using the JWST calwebb pipeline v1.5.3 (CRDS 11.16.3). After in-flight calibration (Rigby et al. 2022), updated reference files were included in the reduction pipeline (CRDS0942). In particular, this update takes into account the higher sensitivity of NIRCam in the LW filters. To account for additional changes to the zero-point calibration (CRDS0989), we use the PHOTMJSR corrections published for module B in Boyer et al. (2022), which are derived from a resolved stellar population analysis of M92 (PID 1334). In our analysis, we focus on the two SW filters as well as the F356W to constrain the spectral energy distribution (SEDs) of individual star clusters in the NIR. These filters are the least affected by potential contributions from either a hot dust component or circumstellar dust from AGB stars, which can have a significant effect on the integrated flux of star clusters at λらむだ ≥ 4 μみゅーm (Villaume et al. 2015), and are not included in the models used to interpret the derived cluster colors (see Section 3.2)

The level-3 products (.i2d) were then aligned to the Gaia DR2 reference frame using the Drizzlepac module TweakReg (Gaia Collaboration et al. 2018). Finally, our F150W, F200W, and F356W science frames were drizzled to a common scale of 005 pixel⁻¹, which was chosen to match the angular resolution of existing HST WFC3 and ACS UV-optical observations of VV 114, and result in a physical resolution of ∼20 pc at the distance of VV 114 (84.4 Mpc). We note that this resolution is several times larger than the typical size of YMCs observed in nearby galaxies; however, it is still well below the average size of cluster complexes and massive OB associations (R_eff ∼ 100–200 pc; Bastian et al. 2006). We further discuss the effects of resolution when interpreting our results in Section 4.

In Figures 1 and 2, we compare HST optical and JWST NIR false-color images of VV 114. At optical wavelengths, we see that VV 114E and the "overlap" region between the east and west nuclei are enshrouded by dust lanes and that detections of bright point sources at wavelengths λらむだ < 0.4 μみゅーm in this region are extremely limited. In contrast, the western nucleus has significantly less dust obscuration, allowing for bright blue point sources to be easily identified throughout the disk. From the NIRCam images, it is clear that at longer wavelengths our observations are sensitive to many bright, red sources, several of which were previously hidden behind the dust lane in the "overlap" region of the merger.

Zoom In Zoom Out Reset image size

3.1. Cluster Identification and Photometry

Star cluster candidates in all three NIRCam filters were selected using the program Source Extractor (Bertin & Arnouts 1996). We considered all sources with local signal-to-noise ratio (S/N) thresholds of ≥3, 5, and 5 in five continuous pixels at F150W, F200W, and F356W, respectively, utilizing 64 deblending subthresholds. We increased the detection threshold for the F200W and F356W filters due to the increased background levels and the crowding of faint blue sources between bright star-forming regions at these wavelengths (right panel, Figure 2). This results in 1837, 1689, and 1860 detections for each filter, respectively. Finally, requiring sources to be detected in all filters, and spatially matched to within 1 pixel, results in a final candidate list of 511 sources.

Zoom In Zoom Out Reset image size

We fit 2D elliptical Gaussian profiles to all 511 candidates in order to measure the major- and minor-axis FWHM at F150W. We then remove from our catalog sources with a major-axis FWHM that is ≥2σしぐま_FWHM,maj larger than the FWHM of the F150W instrumental point-spread function (1.628 pixels =005) determined from WebbPSF (Perrin et al. 2014). This requirement removes 137 candidates, including 63 objects identified by eye as background galaxies with extended structure and 74 additional sources that are likely larger clumps and complexes of clusters blended to appear as single objects. Here we focus on identifying compact star clusters, which can be best modeled as single stellar populations (SSPs), and the derived properties can be compared directly to catalogs where strict size cuts have been applied. A full accounting of clumps and larger star-forming regions, which are likely a combination of multiple single stellar populations, is outside the scope of this Letter.

In order to understand the overlap between the cluster catalog presented in Linden et al. (2021), which was produced using HST UVIS F336W, and the ACS F435W, and F814W observations, we cross-reference the 179 HST sources with the 374 compact cluster candidates identified with JWST. We find that 75% (135/179) of the sources identified with HST overlap with our NIRCam catalog within 2 pixels (01). For the 25% of the HST sources that we are not selecting with our JWST detection thresholds, we find that several of these sources have a JWST/F200W S/N ∼ 1–3 and FWHM slightly larger than the JWST/F150W PSF. Importantly, the majority of these clusters were determined to have intermediate ages (t ≥ 20 Myr), and thus, our selection method is not biased against detecting the youngest (t < 5 Myr) clusters in VV 114. To push the detection thresholds lower such that we can detect all faint sources with JWST will require a detailed analysis of the completeness as a function of cluster location within the galaxy, age, mass, and size. These issues will be addressed in a follow-up paper including NIRCam imaging from all ERS targets in PID 1328.

On the other hand, when we compare the cluster candidates that were originally identified in the HST/F814W image (4060 sources with an S/N ≥3 in 5 pixels), we find that 298/374 (80%) of the compact cluster candidates identified with NIRCam were detected with HST/F814W. This fraction decreases to 60% and 40% for HST/F435W and HST/F336W, respectively. This comparison demonstrates that at least 20%–60% of the sources identified in the NIR are completely missed at optical and UV wavelengths, where clusters were identified down to the limit of m_F336W(Vega) ∼ 26 mag (Linden et al. 2021).

In Figure 2, we zoom in on the string of massive star-forming regions in the spiral arm extending from the western galaxy, which contains several UV- and optically bright clusters (identified as the bright white sources—left panel), as well as sources that appear very red, possibly still embedded in their birth clouds (white arrows). In the right panel, it is clear that these faint red sources are much brighter in the longer-wavelength JWST images, further suggesting that they are highly obscured, and possibly young. Compact sources identified in our JWST NIRCam images are circled in gray, sources identified in our HST images alone are circled in black, and sources that appear in both catalogs are circled in purple. We note that the handful of sources that are only detected in HST images are all faint blue sources embedded in regions of diffuse emission and are therefore not retained due to our higher S/N threshold (≥5) in the NIRCam LW filters. Overall, the right panel of Figure 2 demonstrates that many massive star-forming regions in VV 114 contain a mix of very young (red) and intermediate-age (blue) stellar clusters, many of which are detected with HST. This mixture is reminiscent of one of the most well-studied star-forming regions, 30 Doradus in the Large Magellanic Cloud, which contains both a UV-bright YMC, R136, as well as embedded YSOs in the NW region of the complex (Walborn & Blades 1997).

Photometry for 374 compact clusters identified across the three NIRCam filters was then calculated using the IDL package APER. We used an aperture of radius 2 pixels (01), with an annulus from 4 to 6 pixels to measure the 3σしぐま-clipped mean local background surrounding each cluster. Errors are estimated by varying the inner radius of the sky annulus from 2 to 5 pixels. Aperture corrections for F150W, F200W, and F356W were calculated based on the encircled energy values presented in Perrin et al. (2014) using the WebbPSF models. We additionally applied a correction for foreground Galactic extinction, using the Schlafly & Finkbeiner (2011) dust model and the empirical reddening law of Fitzpatrick (1999). We present our measurements using the AB magnitude photometric system. Finally, we repeat this procedure for the 179 HST-identified clusters in Linden et al. (2021) so that direct comparisons of the two catalogs can be made.

3.2. Stellar Population Synthesis in the NIR

For all confirmed clusters, the measured three-band magnitudes are compared against SSP evolutionary models generated using the isochrone synthesis code yggdrasil (Zackrisson et al. 2011). This code computes the evolution of an instantaneous SF burst based on a Kroupa IMF and the Padova-AGB stellar evolution tracks over an age range of 1 Myr–10 Gyr (e.g., Bertelli et al. 1994). Additionally, these models use CLOUDY (Ferland et al. 2017) to add the contribution from nebular emission lines by varying the fraction of ionizing photons that ionize the surrounding gas (f_cov) of 50%–100%. We also choose to adopt both solar and subsolar metallicity models, as suggested based on the metallicity ranges observed in nearby LIRGs (Rich et al. 2012). Our model choices allow us to make consistent comparisons with results from previous studies of YMCs in LIRGs (Linden et al. 2017; Adamo et al. 2020) as well as the Legacy Extragalactic UV Survey (LEGUS) of nearby normal star-forming galaxies, which adopt the same yggdrasil models to determine cluster ages, masses, and extinctions (e.g., Ryon et al. 2017; Messa et al. 2018; Cook et al. 2019).

In the left panel of Figure 3, we see that the yggdrasil model tracks decrease vertically from 1–5 Myr, approximately perpendicular to the direction of the extinction vector shown in the lower right. The direction and magnitude of this vector are determined using Equation (9) in Salim & Narayanan (2020). Following the discussion in Section 3.2.3, we adopt a NIR power-law slope (βべーた) of −1.6. Studies that sample a variety of Milky Way sight lines find values for βべーた ranging from −0.9 to −2.3, with a slight preference for finding flatter NIR slopes in regions of higher extinction (Fitzpatrick & Massa 2009).

Zoom In Zoom Out Reset image size

The vertical feature in the SSP model occurs when massive stars evolve from blue supergiants into the red supergiant (RSGs) evolutionary phase. RSGs are young, massive, luminous stars that rapidly exhaust their core hydrogen. The luminosities of RSGs peak at ∼1 μみゅーm with absolute J-band magnitudes of M_J = −8 to −11, rivaling the integrated light of Milky Way globular clusters (Larsen et al. 2011). Thus, RSGs, when present, can dominate the NIR flux of a YMC. Therefore the F150W – F200W NIR colors can be used as an absolute age indicator for YMCs younger or older than 5 Myr due to the strong effects RSGs have on the overall continuum shape and the sharp vertical jump in the NIR model colors (Gazak et al.2013).

3.3. The Color–Color Diagram

The age distribution of stellar clusters in galaxies is the combined result of the cluster formation history and cluster disruption. In a galaxy with a constant cluster formation rate, the slope of the age distribution (γがんま) then becomes a measure of how long clusters survive, or at what rate they are being destroyed. If we consider YMCs at a single age, our measured slope implies that their numbers will decrease with time as ${dN}/{dT}=-N/\gamma$, where a steeper value for γがんま corresponds to more rapid cluster destruction.

Studies of the age distribution slope for clusters in nearby galaxies tend to focus on the most massive clusters at ages 10 < τたう(yr) < 300 Myr, which are the easiest sources to observe at UV-optical wavelengths (Krumholz et al. 2019). Therefore, comparisons between different galaxies have been restricted to a limited region of the full cluster age and mass distributions thus far.

In Figure 4, we show the cluster age distribution for VV 114, which contains 90/179 (∼50%) confirmed clusters detected at ≥3σしぐま in HST/F336W, HST/F435W, and HST/F184W (black points) with masses 10^5.2 < M(M_⊙) < 10⁸. This cut represents the mass-completeness limit for star clusters with ages t ≤ 10^8.5 yr determined from the HST catalog using artificial source injection and recovery tests for clusters drawn from a distribution of absolute magnitudes (Linden et al. 2021). This limit is adopted specifically to study the cluster age distribution over a large range in cluster ages, due to the fact that clusters dim as they age, and therefore, we increase the lower-mass limit as a function of increasing age in order to retain a mass-complete cluster sample. Thus, our choice of a very conservative mass limit (sources with masses of 10⁴ M_⊙ are detected in both HST and JWST imaging) allows us to be confident that the measured slope is not affected by issues related to completeness. We previously derived an age distribution slope from 10 to 400 Myr of γがんま = −1.02 ± 0.17 (solid black line) with little evidence for significant deviations as a function of age. We note that this value is steeper than the slopes observed in both LMC (blue) and M83 (green) (Adamo & Bastian 2015), as well as observations of the Antennae merger galaxy (teal) (Fall et al. 2005), providing evidence that the slope of the age distribution tends to be steep in starburst and actively merging galaxies in the local universe.

Zoom In Zoom Out Reset image size

Using our HST-derived cluster catalog, we found four clusters with t < 2.5 Myr and 26 with 2.5–6 Myr. The selection of YMC candidates in Figure 3 is meant to isolate sources whose F150W – F200W and F200W – F356W colors suggest they must be young, massive, and heavily dust-enshrouded sources. In red, we show the updated age distribution including all 17 YMCs with F150W – F200W >0.75 identified in our NIRCam imaging (corresponding to an SSP model at 3 Myr) assigned to the youngest age bin. We additionally include the four sources (purple points in the selection box with F200W – F356W >0.2), whose previously derived ages with HST suggested they were slightly older (with large uncertainty), assigned to the second youngest age bin. If we assume the 17 newly identified sources all have ages between 1 and 3 Myr, the average stellar mass of these sources is 10^5.9 M_⊙, which is well above our HST-derived completeness limit of 10^5.2 M_⊙.

Given the conservative nature of our color selection method, we are likely excluding sources with F150W – F200W <0.75, which are also young (t < 6 Myr), and therefore, the values in Figure 4 for the youngest age bins represent lower limits. One additional source of uncertainty is the effect blending has on the derived stellar masses of individual YMC candidates. At the resolution of our observations, multiple lower-mass clusters may appear as a single massive source. In Linden et al. (2017) we examined the effect cluster blending has on the number of extracted sources in the Antennae merger (D_L = ∼ 24 Mpc; Whitmore et al. 2010) by progressively smoothing HST F435W, F555W, and F814W images. At the distance of VV 114 (84.4Mpc) we find that ∼25% of sources with stellar mass ≥10⁶ M_⊙ would be identified as two or more clusters at the distance of the Antennae. However, we note that reducing the stellar mass of all YMCs with F150W – F200W >0.75 by half results in an average of 10^5.6 M_⊙, which is still above our completeness limit.

For sources that are only detected with NIRCam and have 0.1 < F150W − F200W < 0.75, the possible range of ages is much larger (5–100 Myr) due to the degeneracy of the SSP models in the NIR and, therefore, the contamination from sources at older ages. Taking the median age of sources in this cloud to be 10 Myr, we derive an average stellar mass of 10^5.1 M_⊙. We stress that because our completeness limit is sufficiently high, this population of JWST sources excluded in our analysis, due to degeneracy in the derived ages, will have even lower stellar masses and ultimately should not impact the results presented in Figure 4. The age distribution slope which includes the reddest YMCs is γがんま = −1.30 ± 0.39 for 10⁶ < τたう(yr) < 10⁷ yr. This value is consistent with the previously determined slope within the uncertainties on both measurements and confirms that VV 114 has a steep age gradient for all massive star clusters across the entire range of cluster ages observed. This result provides further evidence that steep age distribution slopes are found predominantly in starburst and merging galaxies.

We note that the steep slope observed in VV 114 may actually be due to an increase in YMC formation as opposed to cluster destruction over the last 10^8.5 yr. However, CALIFA IFU observations of VV 114 suggest that the extinction-corrected enhancement of the star formation rate over the last 300 Myr relative to normal star-forming galaxies is a only factor of ∼2–3 in the central half-light radius (Cortijo-Ferrero et al. 2017), which would result in a change in γがんま of ∼ − 0.2. We, therefore, attribute the much steeper slope observed in VV 114 to rapid cluster destruction as opposed to changes in the YMC formation rate.

Finally, the lack of significant deviations in the shape of the cluster age distribution suggests that the cluster disruption rate in VV 114 has not changed dramatically over the last 0.5 Gyr. For YMCs with ages t < 5 Myr, supernovae will not have had a sufficient amount of time to unbind the cluster, whereas feedback from clusters with ages t > 10 Myr will be dominated by supernovae. In Figure 4, we see that the age distribution slope in these two regimes is similar, indicating that the important feedback channel for cluster disruption is the early winds/radiation pressure at t < 5 Myr, and/or the overall galactic environment sets the destruction rate (as parameterized by γがんま) throughout the merger. Future JWST observations of YMCs in LIRGs, which span a range in SFR, merger stages, and distance from the SFMS, will be crucial to uncovering what truly sets the balance between YMC formation and evolution in extreme star-forming galaxies.

We have presented the results of a James Webb Space Telescope NIRCam investigation into the young massive star cluster population in the luminous infrared galaxy VV 114. We identify 374 compact YMC candidates with FWHM_F150W − FWHM_PSF < 2σしぐま_FWHM,maj as well as S/N ≥ 3, 5, and 5 at F150W, F200W, and F356W, respectively, for which we perform aperture photometry. Based on the derived NIR magnitudes and colors of our new YMC catalog, we reach the following conclusions:

(1). We find that 20% of the JWST-identified star clusters are undetected at optical wavelengths. This fraction increases to 60% at λらむだ = 0.336 μみゅーm and demonstrates that although UV observations are necessary for HST-based studies of star clusters in LIRGs, this approach does miss large fractions of highly embedded YMCs.

(2). Based on the yggdrasil SSP models, we newly identify 17 YMCs in our JWST catalog with F150W – F200W and F200W – F356W colors, suggesting they are all very young, dusty (A_V ∼ 5−15), and massive (10^5.8 < M(M_⊙) < 10^6.1). The inclusion of these "hidden" sources quadruples the number of t < 3 Myr clusters and nearly doubles (×1.7) the number of t < 6 Myr clusters previously detected with HST in VV 114. Finally, we identify a group of NIRCam-detected clusters with F200W – F356W > −0.4 and F150W – F200W ∼0.4, which may represent a more evolved, yet still dust-enshrouded, population of YMCs in VV 114.

(3). The resulting cluster age distribution slope of ${dN}/d\tau \propto {\tau }^{-1.30+/-0.39}$ for 10⁶ < τたう(yr) < 10⁷ is consistent with the previously determined value of γがんま = −1.02 ± 0.17 for 10⁷ < τたう(yr) < 10^8.5, and confirms that VV 114 has a steep age gradient for all massive star clusters across the entire range of cluster ages observed. The consistency between our JWST- and HST-derived age distribution slopes may imply that feedback from stellar winds and supernovae are not able to significantly impact the overall balance of formation and destruction of YMCs in VV 114 over the last 0.5 Gyr.

All of the data presented in this paper were obtained from the Mikulski Archive for Space Telescopes (MAST) at the Space Telescope Science Institute. The specific observations analyzed can be accessed via:

• 1  
  JWST NIRCam F150W data 10.17909/4cgv-j876
• 2  
  JWST NIRCam F200W data 10.17909/kd6r-2b69
• 3  
  JWST NIRCam F356W data 10.17909/m3gb-9322
• 4  
  HST WFC3/UVIS F336W data 10.17909/b77r-bd89
• 5  
  HST ACS/WFC F435W data 10.17909/30vk-9f12
• 6  
  HST ACS/WFC F814W data 10.17909/wer3-ds83

S.T.L was partially supported through NASA grant HST-GO16914. A.S.E. was supported by NSF grant AST 1816838 and by NASA through grants HST-GO10592.01-A, HST-GO11196.01-A, and HST-GO13364 from the Space Telescope Science Institute, which is operated by the Association of Universities for Research in Astronomy, Inc., under NASA contract NAS5-26555. V.U. acknowledges funding support from NASA Astrophysics Data Analysis Program (ADAP) grant 80NSSC20K0450. H.I. and T.B. acknowledge support from JSPS KAKENHI grant No. JP21H01129 and the Ito Foundation for Promotion of Science. F.K. acknowledges support from the Spanish program Unidad de Excelencia Maria de Maeztu CEX2020-001058-M, financed by MCIN/AEI/10.13039/501100011033. This work was also partly supported by the Spanish program Unidad de Excelencia Mariá de Maeztu CEX2020-001058-M, financed by MCIN/AEI/10.13039/501100011033. This research has also made use of the NASA/IPAC Extragalactic Database (NED) which is operated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration. Finally, the National Radio Astronomy Observatory is a facility of the National Science Foundation operated under cooperative agreement by Associated Universities, Inc.