- The paper combines JWST/NIRCam imaging with ALMA Band 6 mapping to resolve 11 dusty star-forming galaxies at z=2.67 within a compact, strongly lensed protocluster core.
- The Koi Pond hosts a highly diverse population, including barred spirals, clumpy disks, interacting galaxies, and mergers, showing that starbursts can arise through both secular disk growth and interactions.
- The observations reveal compact dust emission, patchy attenuation, and striking gas–stellar offsets of about 2.5 kpc in one system, highlighting how environment may reshape gas reservoirs and future cluster galaxies.
The strongly lensed field PJ0846+15 ("the Koi Pond") offers an unusually detailed view of a dusty star-forming galaxy (DSFG) protocluster core at Cosmic Noon. Combining new JWST/NIRCam imaging with ALMA Band 6 continuum mapping, the authors resolve the stellar and dust structure of 11 DSFGs at z=2.67, all within ΔV=800 km s−1 and spanning ≲300 kpc in projection, magnified by a foreground cluster at z=0.766. The central result is that this protocluster core hosts a strikingly heterogeneous DSFG population: interacting systems with tidal tails coexist with ordered disks exhibiting spiral arms, bars, bulges, and clumpy star formation, implying that multiple evolutionary pathways operate simultaneously within a single nascent cluster core.
Observations and data reduction
The NIRCam observations were obtained through GO Cycle 3 program PID 6782 in six filters (F090W–F444W), reaching 5σ point-source depths of roughly 28.5–28.7 AB mag. The data were reduced with the JWST pipeline (v1.16.1) using PEARLS-style customizations to suppress JUMP artifacts, JumProPe for $1/f$ noise and wisp removal, GAIA DR2 astrometric alignment, and drizzling to a 0.03″/pixel grid. Photometry was performed on PSF-matched images (all convolved to F444W resolution via WebbPSF and PyPHER kernels) using manually drawn apertures to handle crowding, lensing distortion, and complex merger morphologies. Intracluster light was not subtracted; consequently, photometry for the ID6.1/6.2ab images is omitted due to contamination, and only the ID6.2c component is reported.
The ALMA Cycle 10 Band 6 program (2023.1.00299.S) used a 27-pointing mosaic covering 60″×60″ around the critical curve, centered at 228 GHz (rest-frame 358 μm at z=2.67). The deepest configuration (TM1) achieves an rms of ~0.034 mJy beam−1 at 0.17″×0.17″ resolution, comparable to the NIRCam angular resolution of ~0.15″. Imaging employed multi-scale deconvolution with auto-multithresh masking down to 1.5σ. Flux densities from TM1 and TM2 agree to within 10%, which the authors interpret as evidence that little extended flux is resolved out—suggesting dust structures are more compact and clumpy than low-resolution studies might infer.
Stellar colors, masses, and attenuation
Using F277W−F444W as a Balmer-break-redward color proxy, the sample spans F444W = 20–23 mag (median 21.62 mag) and F277W−F444W = 0.5–1.6 mag (median 0.80 mag), all magnification-corrected under the assumption that lensing is achromatic across both filters. Comparison with the SCUBADive sample of 289 COSMOS DSFGs at comparable redshift indicates the members are massive (ΔV=8000) galaxies spanning more than 1 dex in stellar mass with dust attenuation differences of ΔV=8001 mag. This is a notable claim but rests on color-based inference rather than full SED fitting; the authors acknowledge that without mid-infrared photometry, stellar masses of high-redshift red galaxies can be overestimated by up to 1 dex, and a refined JWST-based lens model is still forthcoming.
Resolved color maps reveal pixel-to-pixel variations as large as F277W−F444W ≈ 2 mag within individual sources, particularly in mergers where tidal features are bluer than their redder central components. Patchy attenuation is evident where rest-frame UV/optical emission penetrates the ISM in several members (ID4, ID6, ID7, ID11). A key implication is that integrated photometry can misrepresent the reddening properties of such morphologically complex systems.
Dust continuum versus stellar emission
Intrinsic 1.3 mm flux densities span an order of magnitude, ΔV=8002–1.9 mJy, with lensing enabling detection of a faint (<0.3 mJy) dust-continuum population that would otherwise be missed. While aggregate F277W−F444W versus dust flux shows no correlation, resolved maps generally show color excess tracing dust surface brightness—for most members (ID1, ID2, ID4, ID5, ID7)—supporting dust reddening rather than stellar age as the dominant driver of red colors, though star-formation history and metallicity degeneracies are not accounted for. Exceptions include ID8.1 and ID3, whose millimeter peaks coincide with optically bright point-like cores, plausibly nuclear starbursts or AGN emerging from their dust envelopes.
Dust morphology relative to stars varies widely: compact nuclear-only dust components (ID10, ID11) contrast with galaxy-wide emission (ID1, ID2). In all members the dust remains more compact than the stellar continuum, consistent with other JWST–ALMA studies. Notably, the reddest, most dust-emitting members (ID1, ID2) are smooth clumpy disks lacking bars or spiral arms, whereas the dynamically stable barred spirals (ID9, ID10, ID11) sit in the outskirts with lower gas content and modest reddening.
Mergers, interactions, and displaced gas
Four members show interaction signatures at different stages. ID8 is an early-to-intermediate stage prograde merger with two kinematically distinct CO(3–2) components (~100 km sΔV=8003 offset), ~8 kpc tidal "antennae" tails, and a UV-bright nucleus in 8.1 with possible PSF signature suggestive of an AGN or intense nuclear starburst. ID3 shows a candidate tidal tail hosting more than two dozen point-like sources, interpreted as candidate tidally ejected or induced stellar clusters. ID4 presents a tension worth emphasizing: its stellar morphology clearly shows two merging components, yet CO(3–2) reveals an ordered velocity gradient and single-component dust continuum. The authors note that current resolution cannot reliably distinguish merger kinematics from disk rotation, though a surviving post-merger disk remains possible. ID6 is a pair of spheroidal, bulge-hosting galaxies ~5 kpc apart (magnification-corrected) with diffuse tidal shells rather than extended tails, consistent with a radially dominated collision of dynamically hot progenitors.
The most striking result concerns ID6: in both galaxies, the molecular gas and dust are offset by ~0.8″ (~2.5 kpc corrected) from the stellar emission, coinciding instead with tidal debris imaged at magnification ΔV=8004 on the lensing caustic. The authors argue against self-gravitating newly formed reservoirs—the CO velocity fields suggest both components retain coherent rotation inherited from their progenitor disks—but concede it is unclear how gas can be displaced several kpc while maintaining rotation, and raise ram-pressure stripping in the protocluster environment as an alternative mechanism. This interpretation remains open.
Secular starbursts and evolutionary pathways
A substantive claim of the paper is that secular processes alone can drive starburst-level activity in protocluster cores. The clumpiest, most widely dust-obscured members are disks (ID1, ID2) with spatially aligned kpc-scale clumps in both ALMA TM1 continuum and F277W−F444W maps, consistent with in-situ disk instability sustaining galaxy-wide obscured star formation. Meanwhile, candidate (super-)stellar clusters at tens-of-parsec scales are resolved in the tidal emission of ID3 and ID6 (ΔV=8005 point-like clumps in ID6), potentially surviving as globular cluster progenitors. Bulge-like nuclei in the merger systems (ID3, ID6, ID8) are flagged as early signatures of transition toward quiescent ellipticals, with tidal disruption of molecular gas possibly accelerating quenching.
Placed in context, J0846 contrasts with the Spiderweb protocluster—where massive quiescent galaxies already occupy the core and disk DSFGs reside in the outskirts—and resembles other compact DSFG-rich cores (e.g., SPT2349–56 at ΔV=8006) expected to coalesce into proto-BCGs of ΔV=8007 within a few hundred Myr. The authors caution that cross-protocluster comparisons are complicated by differing formation stages, and that environmental effects on bar/spiral formation at ΔV=8008 remain unconstrained.
Limitations and open questions
Several caveats bear directly on the quantitative results. Magnification corrections derive from a pre-JWST lens model, with differential lensing across extended sources not yet applied; a refined model is deferred to future work. Stellar mass and attenuation inferences are color-based proxies pending full SED modeling, vulnerable to the known degeneracies noted above. Merger classifications rely partly on visual identification, and the ID4 case demonstrates that kinematic evidence for rotation cannot exclude an ongoing merger at available resolution. The displaced-gas configuration in ID6 lacks a settled physical explanation between tidal displacement and ram-pressure stripping. Finally, whether the apparent survival of rotating disks through major interactions (ID4, ID6) is generic or a projection effect remains unresolved.
Conclusion
This paper delivers the first resolved rest-frame near-infrared stellar and 358 μm dust continuum view of the J0846 protocluster core, detecting all 11 CO-confirmed members and revealing a population spanning more than 1 dex in stellar mass, over an order of magnitude in dust luminosity, and a full morphological range from barred spirals to tidally disrupted mergers. Its principal contribution is demonstrating that DSFG-rich protocluster cores are not uniformly merger-dominated starburst environments: secular, clumpy disk growth can produce the reddest and dustiest members, while mergers concurrently build bulges and displace cold gas reservoirs. The diversity observed within a single ΔV=8009 kpc core implies that the assembly of quiescent elliptical cluster galaxies proceeds along multiple concurrent channels, with future source-plane reconstructions and resolved stellar-mass and gas analyses required to quantify their relative contributions.