GOALS-JWST: Unveiling Dusty Compact Sources in the Merging Galaxy IIZw096

IIZw096 (CGCG448-020, IRAS20550+1656 ) is a merging, luminous infrared galaxy (LIRG) at z = 0.0361 with an infrared (IR) luminosity of L_{IR,8−1000μm} = 8.7 × 10¹¹ L_⊙, one of the more than 200 LIRGs in the Great Observatories All-sky LIRGs Survey (GOALS; Armus et al. 2009). Previous imaging with the Spitzer Space Telescope revealed that the majority (up to 80%) of the IR luminosity of the entire system comes from a region outside of the merging nuclei (Inami et al. 2010), making it an even more extreme case than the well known Antennae Galaxies (Mirabel et al. 1998; Brandl et al. 2009).

The system consists of regions A, C, and D (Figure 1(c)) along with a merging spiral galaxy to the northwest (Goldader et al. 1997). Regions C and D are not detected or have extremely low signal-to-noise ratios (S/Ns) at ultraviolet and optical wavelengths with the Hubble Space Telescope (HST), but only at near-IR and longer wavelengths (Inami et al. 2010; Barcos-Munoz et al. 2017; Wu et al. 2022; Song et al. 2022). Although the Spitzer/MIPS 24 μm image suggested that a single compact source in region D dominates the emission, the large beam size made it impossible to resolve the exact location of the immense far-IR emission. The superior sensitivity and resolving power of the James Webb Space Telescope (JWST) lets us pinpoint the source that is responsible for the intense IR emission and study its complex environment on sub-kpc scales in the mid-IR for the first time.

Zoom In Zoom Out Reset image size

Here, we present high spatial resolution mid-IR imaging of IIZw096 taken with the JWST Mid-InfraRed Instrument (MIRI; Rieke et al. 2015; Bouchet et al. 2015). Throughout this paper, we adopt a cosmology with H = 70 km s⁻¹ Mpc⁻¹, Ω_M = 0.28, and Ω_Λ = 0.72. The redshift of IIZw096 (z =0.0361) corresponds to a luminosity distance of 160 Mpc and a projected physical scale of $725\,\mathrm{pc}\,{\mathrm{arcsec}}^{-1}$.

The JWST observations were performed under the Directors Discretionary Time Early Release Science (ERS) program 1328 (co-PIs: L. Armus and A.S. Evans). Images of IIZw096 were obtained on 2022 July 2 with a MIRI subarray (SUB128) using the F560W (λ₀ = 5.6 μm), F770W (7.7 μm), and F1500W (15 μm) filters. The pointing was centered at 31435167, 1712769 (J2000), where the prominent mid-IR emission was identified with Spitzer. The observations were dithered and the exposure times set to avoid saturation (46, 48, and 48 s, respectively). The data were reduced with the standard JWST calibration pipeline (data processing software ver.2022_2a, calibration software ver.1.5.3; Gordon et al. 2015; Bushouse et al. 2022) and up-to-date reference files from the Calibration References Data System. The images in this work, including HST, are aligned to the Gaia Data Release 3 catalog (Gaia Collaboration et al. 2016, 2021).

The ∼10× improvement in the spatial resolution of JWST compared to Spitzer resolves the mid-IR emission into individual clumps down to scales of ≲100–200 pc, enabling measurements of the mid-IR color and L_IR surface density to study the nature of IIZw096.

3.1. Mid-IR Clumps in the Disturbed Region

Individual sources are identified with the DAOFIND algorithm (Stetson 1987) in the F770W SUB128 subarray image, which have the highest S/N, using the Python photutils package (Bradley et al. 2021). The detection threshold is 5σ. The sources detected in the F770W image are used as priors for photometry in all three bands (Figure 1). We assign identification (ID) numbers for the detected sources in ascending order of R.A. For a subset of sources shown by Wu et al. (2022), we also refer to their source names (Figure 1(c)). The same 12 clumps are detected in the F560W image, while in the F1500W image, four of them (IDs 4, 9, 10, and 11/C0) lie on the structure of the point-spread function (PSF) of the brightest source (ID 8/D1), making confident detections and flux measurements difficult. Thus, we only report their upper limits.

As shown in Figure 1, the most prominent mid-IR source is ID 8 (source D1), lying 047 northeast of ID 7 (source D0). Although ID 8 is fainter than ID 7 in the HST/NICMOS 1.6 μm image, its emission exceeds ID 7 at 5.6 μm and 7.7 μm by a factor of three. At 15 μm, the ID 8-to-ID 7 flux ratio increases to about five. Although ID 7 was previously speculated to be associated with the bulk of the total IR emission due to its prominence at 1.6 μm (Inami et al. 2010), the majority of the mid-IR emission in fact originates from ID 8.

There are five bright mid-IR clumps (IDs 1, 4, 5, 6, and 12) that are either not detected or have an extremely low S/N in the HST 1.6 μm image. ID 1 is in the less dusty region A to the southwest. Three sources, IDs 4, 5, and 6, are south of ID 7. The remaining one, ID 12, is located 227 northeast of ID 8. Interestingly, these new mid-IR selected sources are not concentrated in the dustiest region, but spread throughout the perturbed region.

Additional structure is evident in the MIRI image, outside of the main power source in the IIZw096 system (Figure 2). Region A, which accounts for most of the optical emission, hosts a number of clumps in the MIRI data. Around region C, the emission peaks at 1.6 μm but fades toward the mid-IR.

Zoom In Zoom Out Reset image size

3.2. Mid-IR Colors of the Clumps

Aperture photometry was employed to measure the fluxes of the detected clumps, except for IDs 7 and 8 due to their relative proximity (see below). The aperture radii used for F560W, F770W, and F1500W are 027, 028, and 030, respectively, with aperture corrections of 0.65, 0.65, and 0.50. ³⁰ Aperture photometry in the F1500W image was performed after subtracting ID 8, due to its prominent PSF pattern (note that the PSF has not been subtracted in Figure 1(f)). The source subtraction was performed using a PSF generated by WebbPSF (Perrin et al. 2012, 2014). ³¹ To account for the extremely red color of ID 8, a power-law spectrum with a spectral slope of 3, resembling the mid-IR color of IIZw096, was used to generate the PSF. This provides a more accurate flux measurement of the sources around ID 8.

To extract the fluxes of IDs 7 and 8, a simultaneous two-dimensional Gaussian fit was performed in the F560W and F770W images. In the F1500W image, the flux of ID 8, which dominates the emission, was also extracted via a simultaneous two-dimensional Gaussian fit. However, we did not adopt the flux of ID 7 from this fit because this source lies on the Airy ring of ID 8. Instead, the flux of ID 7 was derived from a single Gaussian fit ³² to the image after subtraction of ID 8 to minimize contamination from the PSF.

The local background was measured using a 3σ-clipped median of various annuli. They have a minimum inner radius 2.4× the half-width at half-maximum (HWHM) of the PSF and a maximum radius that is 14× (for F560W and F770W) and 7× (F1500W) the PSF HWHM in steps of 005 around each source. During this process, the other detected sources were masked out to avoid background overestimation. Each of the measured background levels was separately subtracted from the measured flux, providing a distribution of fluxes for each source. The median value of this distribution is reported as the final flux density in Table 1. The 16th and 84th percentiles were adopted as the flux uncertainties.

Table 1. Flux Density of the Clumps Detected in F560W, F770W, and F1500W

ID	R.A.	Decl.	F560W	F770W	F1500W
	degree	degree	mJy	mJy	mJy
1	314.350210	17.126449	${0.51}_{-0.03}^{+0.03}$	${3.17}_{-0.23}^{+0.16}$	${8.56}_{-0.18}^{+0.30}$
2	314.350342	17.126498	${0.35}_{-0.05}^{+0.03}$	${2.62}_{-0.27}^{+0.17}$	${7.10}_{-0.25}^{+0.36}$
3	314.350638	17.126226	${0.13}_{-0.01}^{+0.03}$	${0.87}_{-0.08}^{+0.06}$	${2.79}_{-0.09}^{+0.14}$
4	314.351233	17.127361	${0.13}_{-0.02}^{+0.03}$	${0.97}_{-0.15}^{+0.05}$	<1.87
5	314.351263	17.127169	${0.30}_{-0.01}^{+0.02}$	${1.60}_{-0.06}^{+0.05}$	${3.46}_{-0.03}^{+0.18}$
6	314.351444	17.127124	${0.13}_{-0.01}^{+0.03}$	${0.86}_{-0.04}^{+0.03}$	${2.27}_{-0.04}^{+0.13}$
7/D0	314.351423	17.127521	${1.79}_{-0.05}^{+0.05}$	${8.60}_{-0.30}^{+0.30}$	${32.16}_{-1.74}^{+1.74}$
8/D1	314.351552	17.127566	${6.17}_{-0.04}^{+0.04}$	${25.40}_{-0.21}^{+0.21}$	${155.39}_{-1.62}^{+1.62}$
9	314.351688	17.127728	${0.31}_{-0.03}^{+0.04}$	${1.91}_{-0.19}^{+0.14}$	<4.07
10	314.351945	17.127588	${0.15}_{-0.02}^{+0.03}$	${0.75}_{-0.18}^{+0.09}$	<2.80
11/C0	314.351972	17.127732	${0.39}_{-0.03}^{+0.02}$	${1.09}_{-0.18}^{+0.08}$	<0.73
12	314.352156	17.127821	${0.11}_{-0.01}^{+0.02}$	${0.67}_{-0.05}^{+0.04}$	${2.33}_{-0.07}^{+0.14}$
Total			${10.45}_{-0.33}^{+0.36}$	${48.52}_{-1.94}^{+1.39}$	${214.06}_{-4.04}^{+4.63}$

Download table as:  ASCIITypeset image

Based on the measured fluxes in all three MIRI bands, we show a F1500W/F560W−F770W/F560W color–color diagram in Figure 3. These colors are a sensitive measure of the mid-IR continuum slope, and the F770W/F560W color is also sensitive to emission of polycyclic aromatic hydrocarbon (PAH) at 7.7 μm. The individual clumps detected in the MIRI SUB128 images are shown in the left panel. We also present the same diagram for local LIRG nuclei using Spitzer/IRS low-resolution spectra taken with Short-Low (SL; 5.5–14.5 μm) and Long-Low (LL; 14−38 μm) spectroscopy (Stierwalt et al. 2013, 2014). We generated synthetic photometry from the spectra using the MIRI filter curves and the Python synphot package (STScI Development Team 2018). The SL and LL slits fully cover regions C and D. The LL slit also covers region A. As expected from the IRS spectra (Inami et al. 2010; Stierwalt et al. 2013), the F1500W/F560W color from the synthetic photometry of the dust-obscured region in IIZw096 is an outlier with a much redder color than the rest of the local LIRGs. With JWST/MIRI, we are now able to decompose the emission into individual clumps to study the distribution of their mid-IR colors.

Zoom In Zoom Out Reset image size

The summed flux of all the clumps in color space agrees well with the color measured with the much larger beam of Spitzer/IRS (Figure 3 right). All MIRI-detected clumps show redder colors (F1500W/F560W ≳ 15) than most of the local LIRG nuclei, except for ID 5, which has F1500W/F560W = 11 ±0.7. ID 8 (D1) is the reddest, while ID 7 (D0) has a comparable F1500W/F560W color to the other clumps. The newly detected source ID 12 is the second reddest, along with IDs 2 and 3 having excess emission at 15 μm and 7.7 μm compared to most of the other sources.

The F770W/F560W colors of the clumps are spread over the range 4 ≲ F770W/F560W ≲ 8. As expected, since they dominate the mid-IR flux, the F770W/F560W colors of the two brightest sources, IDs 7 and 8, agree well with the color of IIZw096 derived from the synthetic photometry of the IRS spectrum. The remaining sources, except for IDs 5, 10, and 11, are redder, with F770W/F560W ≳ 6. These red sources have F770W/F560W colors consistent with strong PAH emission. The bluer sources suggest weaker PAH emission or an excess of hot dust. The F770W/F560W colors of IDs 7 and 8 are comparable to sources with 6.2 μm PAH equivalent widths (EQWs) of about half those seen in pure starburst nuclei (Figure 3 right), which are consistent with the direct measurement in the IRS spectrum (6.2 μm PAH EQW = 0.26 μm; Inami et al. 2010) and may indicate an excess of very hot dust.

3.3. IR Luminosity Surface Density

Our JWST imaging has revealed the complexity of the dustiest region of the merging galaxy, IIZw096. The three main components in the perturbed region show a wide variety of optical–IR colors and morphologies, with a mix of bright, unresolved clumps and diffuse emission. This suggests a range in properties, such as extinction, SFR, age, and dust temperature in this ongoing merger.

Although ID 7 (D0) is the brightest source at 1.6 μm, the JWST mid-IR data demonstrate that ID 8 (D1) generates the bulk of the total IR emission in IIZw096. The location of ID 8 also coincides with two OH megamasers (Migenes et al. 2011; Wu et al. 2022). ³⁴ Megamasers are often found in merging (U)LIRGs, in close proximity to the nuclei (e.g., Roberts et al. 2021), marking regions of extremely high gas density and strong far-IR radiation.

The emission we have targeted with JWST is clearly responsible for the bulk of the luminosity in IIZw096 and it arises from outside of the prominent two merging galaxies, one of which is the spiral galaxy to the northwest that lies outside of the MIRI SUB128 field of view (FOV) (Figure 1(a)). However, it is possible that ID 8 is a third nucleus in this system. Given the observed mid-IR morphology, region A + C + D could be a single disrupted galaxy or it could be two galaxies with source A0 being one nucleus and ID 8 (D1) being the other. The diffuse, extended emission around ID 8 would then be the remnants of the third galaxy's disk. The estimated stellar mass of ID 8 is ∼10⁹ M_⊙ and there is a similarly large mass of gas in this region (Inami et al. 2010; Wu et al. 2022). These estimates might indicate ID 8 is a partially stripped third nucleus. Although the current MIRI SUB128 images do not provide evidence for either the two- or three-body merger scenario, our upcoming JWST observations (ERS program 1328) may elucidate this question. The deeper MIRI and NIRCam full-array imaging may detect a more pronounced disk-like morphology around ID 8. A detection of a rotation curve around ID 8 by the planned spectroscopic observations could suggest that ID 8 is a third nucleus.

ID 8 (D1) generates 40%–70% of the total IR emission of the IIZw096 system, corresponding to an L_IR surface density of >3–5 × 10¹² L_⊙ kpc⁻². This is ∼10× the characteristic surface brightness of starbursts, but comparable to super star clusters (Meurer et al. 1997) including some in the Antennae Galaxies (Brandl et al. 2009). The L_IR surface density limit of ID 8 is also consistent with the ULIRG nuclei studied at 12.5 μm with Keck (Soifer et al. 2000, 2001). In addition, the 33 GHz continuum imaging (01 resolution) taken with the Very Large Array shows that the peak emission is located at ID 8 (Song et al. 2022). These authors estimated a SFR surface density of 470 ± 60 M_⊙ yr⁻¹ kpc⁻², corresponding to an L_IR surface density of (3.9 ± 0.5) × 10¹² L_⊙ kpc⁻², which agrees with the value derived from the mid-IR. Given the L_IR surface density limit from the mid-IR and the column density of ∼10²⁵ cm⁻² obtained via the molecular gas mass (within an aperture of 02 × 016; Wu et al. 2022), ID 8 appears to be below the Eddington limit if its size is 175 pc. (e.g., Pereira-Santaella et al. 2021; Barcos-Munoz et al. 2015).

The clumps in the disturbed region, including ID 8 (D1), are much redder in F1500W/F560W than local LIRG nuclei. Based on the 9.7 μm silicate optical depth (τ_9.7μm ∼ 1) derived from Spitzer/IRS spectroscopy, the V-band extinction is estimated to be ≥19 mag (Inami et al. 2010). The unusually red colors of IIZw096 could be due to an extremely young, highly obscured starburst or active galactic nucleus (AGN), triggered by the recent merger. The other LIRG with a very red F1500W/F560W color, IRAS22491-1808, is also an ongoing merger with bright clumps of star formation (Surace 1998; Surace et al. 1998).

The F770W/F560W color traces the 7.7 μm PAH emission and is also a good proxy for the 6.2 μm PAH EQW (Figure 3, right). Using the 6.2 μm PAH EQW as a diagnostic of starbursts and AGNs (e.g., Brandl et al. 2006; Armus et al. 2007; Petric et al. 2011), the clumps with F770W/F560W ≳ 5 are consistent with pure star formation. However, the color of ID 8 (D1) is in the range where an AGN cannot be excluded. Although an analysis of the X-ray spectra of IIZw096 obtained with Chandra, XMM-Newton, and NuSTAR also favors star formation, the non-detection of ID 8 by NuSTAR does not rule out a Compton-thick AGN if the column density exceeds 10²⁵ cm⁻² (Iwasawa et al. 2011; Ricci et al. 2021). In fact, the estimated column density of ID 8 is ∼10²⁵ cm⁻² (Wu et al. 2022). The evidence may be consistent with the presence of a Compton-thick AGN but it is equally consistent with a starburst, and neither is conclusive. Our upcoming mid-IR and near-IR spectroscopic data (ERS program 1328) are expected to shed additional light on the underlying energy source of this heavily obscured source, perhaps through detection of one or more coronal emission lines.

JWST/MIRI imaging has demonstrated uncharted aspects of the dust emission from the extremely luminous merging galaxy IIZw096. The high spatial resolution and high-sensitivity mid-IR imaging of this work yields the following findings:

• 1.  
  For the first time, we have spatially resolved the mid-IR emission of the merger-induced heavily dust-obscured region of IIZw096. We identify the source (ID 8/D1) that is responsible for the bulk of the mid-IR emission, accounting for 40%−70% of the total IR emission of the system.
• 2.  
  In total, 12 clumps are detected in the F770W (and F560W) image, five of which are newly identified and were not detected or had low S/N detections at 1.6 μm with HST/NICMOS. Most of the clumps have similar F1500W/F560W colors, ranging from ∼15 to 25. These colors are about twice as red as local LIRG nuclei, but agree with the colors derived from synthetic photometry of Spitzer spectra of this system. Among LIRG nuclei, the F770W/F560W colors roughly correlate with the 6.2 μm PAH EQW, and therefore the clumps have colors indicative of 6.2 μm PAH EQWs from ∼0.3 μm to 0.6 μm, slightly lower than but including pure star formation.
• 3.  
  The estimated L_IR of ID 8 (D1) is 3–5 × 10¹¹ L_⊙, which corresponds to a SFR of 40–60 M_⊙ yr⁻¹ if it is star forming. As the source is unresolved, we estimate its L_IR surface density to be >3–5 × 10¹² L_⊙ kpc⁻² or a SFR surface density of >400–600 M_⊙ yr⁻¹ kpc⁻². Such high surface densities put source D1 in a range comparable to young super star clusters and ULIRG nuclei.

The JWST mid-IR imaging described in this Letter has revealed a hidden aspect of IIZw096, and has opened a door toward identifying heavily dust-obscured sources which cannot be found at shorter wavelengths. Future planned spectroscopic observations of IIZw096 will provide additional information on the nature of the dust, ionized gas, and warm molecular gas in and around the disturbed region of this luminous merging galaxy.

The authors would like to thank the referee whose constructive comments helped improve the manuscript. The JWST data presented in this paper were obtained from the Mikulski Archive for Space Telescopes (MAST) at the Space Telescope Science Institute. The specific observations analysed can be accessed via https://doi.org/10.17909/8c47-wb74. STScI is operated by the Association of Universities for Research in Astronomy, Inc., under NASA contract NAS526555. Support to MAST for these data is provided by the NASA Office of Space Science via grant NAG57584 and by other grants and contracts. This work is based on observations made with the NASA/ESA/CSA James Webb Space Telescope. These observations are associated with program 1328. The data were obtained from the Mikulski Archive for Space Telescopes at the Space Telescope Science Institute, which is operated by the Association of Universities for Research in Astronomy, Inc., under NASA contract NAS 5-03127 for JWST. This research has made use of the NASA/IPAC Extragalactic Database (NED), which is funded by the National Aeronautics and Space Administration and operated by the California Institute of Technology. H.I. and T.B. acknowledge support from JSPS KAKENHI grant No. JP21H01129 and the Ito Foundation for Promotion of Science. Y.S. is supported by the NSF through grant AST 1816838 and the Grote Reber Fellowship Program administered by the Associated Universities, Inc./ National Radio Astronomy Observatory. The Flatiron Institute is supported by the Simons Foundation. Vivian U acknowledges funding support from NASA Astrophysics Data Analysis Program (ADAP) grant 80NSSC20K0450. A.M.M. acknowledges support from the National Science Foundation under grant No. 2009416. S.A. gratefully acknowledges support from an ERC Advanced grant 789410, from the Swedish Research Council and from the Knut and Alice Wallenberg (KAW) foundation. K.I. acknowledges support by the Spanish MCIN under grant PID2019-105510GB-C33/AEI.

Facilities: JWST (MIRI) - , HST (ACS - , NICMOS) NED - .

Software: astropy (Astropy Collaboration et al. 2013, 2018), photutils (Bradley et al. 2021), synphot (STScI Development Team 2018), WebbPSF (Perrin et al. 2012, 2014).