Exploring the observed properties of JWST Little Red Dots with cosmological simulations and dust radiative transfer

abstract · abstract

JWST has uncovered compact high-redshift (z>=4) sources, dubbed Little Red Dots (LRDs), characterised by a distinct V-shaped continuum. Their nature remains debated since broad-line AGN signatures coexist with non-AGN continua, marked by weak X-ray and mid-IR emission. We investigate which properties of LRDs naturally emerge within simulations of massive AGN host galaxies and which require additional physics. We post-process two cosmological zoom-in simulations of the same massive halo, one including black hole growth and AGN feedback (AGNfid) and the other with only star formation and stellar feedback (SFonly). We analyse snapshots spanning z=6-9 to probe different evolutionary stages. Using SKIRTv8, we develop a forward-modelling framework exploring various dust models, dust-to-metal ratios, and intrinsic AGN spectra. We generate mock spectra and images to apply JWST observational LRD selection criteria. Not designed to model LRDs, our AGNfid simulations naturally reproduce their properties: V-shaped continua, compact morphologies with rest-UV complexity, UV luminosities, attenuations, overmassive BHs, sub-Eddington accretion, and Balmer break strengths (~1-2) typical of most LRDs. Spectral decomposition reveals that the rest-UV continuum is dominated by stellar emission, whereas the optical continuum stems from a dust-attenuated AGN. Dust geometry and AGN-to-stellar contribution drive photometric differences between LRDs and Little Blue Dots. Our results suggest that LRDs do not necessarily represent a distinct class of objects. Instead, their defining properties can arise in massive AGN host galaxies from the interplay between stellar emission, an obscured central engine, and dust geometry. Super-Eddington accretion, dust-free gas columns, evolved stellar populations, and photoionisation modelling are required to test whether this framework reproduces the diversity of observed LRDs.

S6: 6 Summary and conclusions / block 1 · html

Mock NIRCam observations reveal extended rest-frame UV structures, including host galaxy emission, merging companions, and stellar clumps, surrounding a compact red core. This morphology arises because the host remains visible in the UV while the central AGN dominates at longer wavelengths. These features closely resemble those observed in a substantial fraction of LRDs, with extended rest-UV emission reported in ∼ \sim 40% of the sample ( Rinaldi et al., 2025 ; Baggen et al., 2026a ) .

S6: 6 Summary and conclusions / block 1 · html

The simulated LRDs reproduce the UV-bright portion of the observed population ( − 22.5 < M UV < − 19.5 -22.5<M_{\rm UV}<-19.5 ) and exhibit moderate V-band attenuation ( A V < 3.5 A_{\rm V}<3.5 ). Their stellar populations produce moderate Balmer breaks, F 4000 / F 3600 ∼ 1 F_{4000}/F_{3600}\sim 1 – 2 2 , consistent with approximately 40 40 – 50 % 50\% of observed LRDs. The strongest observed discontinuities require additional ingredients, such as more evolved stellar populations or absorption and reprocessing by dense excited neutral hydrogen.

S3: 3 Mock Observations / block 1 · html

The resulting F444W photometric maps are utilised to identify and select LRD candidates. To avoid systems in which the BH has only recently been seeded, we restrict our analysis to AGN-host galaxies located at the center of the most massive halo in the AGNfid simulation. For each selected system, we then identify the brightest object in the F444W map within a projected radius of 5 5 kpc from the center of the most massive halo and use it for all subsequent analyses. For consistency, we apply the same criterion to the star-forming galaxies in the SFonly simulation.

S4.SS1: 4.1 Synthetic Spectral Energy Distributions / block 1 · html

Figure 2 : Synthetic spectral energy distributions (SEDs) extracted within an aperture diameter of D ap = 0.4 ′′ D_{\text{ap}}=0.4^{\prime\prime} centered on the brightest F444W source within the central 5 5 kpc of the most massive halo. Panels display results for the SFonly ( top ) and AGNfid ( bottom ) simulation suites. Shaded vertical bands indicate the bandpasses of the JWST/NIRCam filters utilised in our photometric selection pipeline (see Sec. 4.2 ). Individual spectra are colour-coded by their emergent absolute UV magnitude ( M UV M_{\text{UV}} ) to demonstrate the population variance in UV luminosity. To highlight the differences in dust absorption and thermal re-emission across different dust extinction models from 193 , a representative case is highlighted in distinct colours: with SMC-like extinction (cyan), MW-like extinction (red), and the modified SMC featuring a small-grain cutoff of a min > 0.1 ​ μ ​ m a_{\text{min}}>0.1\,\mu\text{m} (orange).

S4.SS2.SSS0.Px2: Compactness / block 1 · html

We then evaluate whether our mock sources satisfy the compactness criterion 12 12 12 We caution that flux extraction from fixed apertures has been found to introduce up to 30-40% of noise ( Rinaldi et al., 2026a , e.g.,) . The resulting candidate fractions are therefore likely biased low and should be regarded as conservative lower limits. A morphology-based analysis would be required to quantify and correct for this incompleteness. proposed by Labbe et al. 2024a :

S6: 6 Summary and conclusions / block 1 · html

We have investigated whether the characteristic properties of Little Red Dots, namely rest-optical compactness and a spectral V-shape, can arise naturally within cosmological simulations of massive AGN-host galaxies. We performed 3D dust continuum radiative transfer calculations using SKIRTv8 to post-process the cosmological zoom-in simulations of galaxy and supermassive black hole (SMBH) formation presented in 182 . Specifically, we analyzed snapshots at z = 6 − 9 z=6-9 from two simulation suites: AGNfid , which includes SMBH formation, accretion, and feedback; SFonly , which only includes the physics of star formation. We further explored the impact of varying the dust-to-metal ratio ( f d f_{\rm d} ), grain size distribution and composition, and the intrinsic AGN spectral energy distribution. We developed a mock observation pipeline to convert the radiative transfer outputs into synthetic JWST /NIRCam photometric maps. Focusing on the brightest sources within the central 5 ​ kpc 5\,\text{kpc} region of the main halos, we applied the photometric selection criteria of 94 alongside the F444W compactness criterion of Labbe et al. 2024a to identify LRD candidates.

S6: 6 Summary and conclusions / block 1 · html

Our main findings can be summarised as follows:

S6: 6 Summary and conclusions / block 1 · html

Both the AGNfid and SFonly populations can exhibit the typical "V-shape" of LRDs. However, only systems containing an accreting BH simultaneously satisfy the photometric and compactness criteria used to identify LRDs. Within our simulations, an AGN is therefore required to reproduce the combination of red optical colours and compact rest-frame optical morphology characteristic of the observed population.

S6: 6 Summary and conclusions / block 1 · html

The simulated LRDs host SMBHs that lie systematically above the local M BH − M ⋆ M_{\rm BH}-M_{\star} relations. In contrast, some compact, non-V-shaped AGNfid objects lie closer to the local scaling relations, mirroring the spectroscopically inferred differences between LRD and Little Blue Dot (LBD) populations. The simulations also reproduce the continuous relation between L 5100 L_{5100} and β opt \beta_{\rm opt} inferred for LRDs and LBDs. Together, these results support an interpretation in which the two populations are not intrinsically distinct, but occupy different regions of a continuous distribution governed by the relative AGN contribution and line-of-sight attenuation.

S3: 3 Results / block 1 · html

4 of 14 objects show a significant deviation from a smooth rollover at > 3 ​ σ >3\sigma significance with the Paschen continuum extrapolation lying consistently above the Brackett continuum (see legend in Fig. 1 or Fig. 5 ). Notably, this includes a 12 ​ σ 12\sigma outlier in a relatively modest sample size of objects. This is the Rosetta Stone LRD, which has the highest continuum SNR in the sample with SNR 1.2 ​ μ ​ m ≈ 100 \mathrm{SNR_{1.2\,\mu m}}\approx 100 , so we highlight this in the top panel of Fig. 1 where, for comparison, multiple blackbodies of varying temperature and subject to differing levels of dust reddening are shown. Although simple blackbody emission can reproduce the Brackett continuum or the very bluest Paschen continuum emission, they cannot smoothly fit both wavelength regimes. Most objects consistent with smooth rollovers comes from the lower SNR part with none of the 5 objects with SNR 1.2 ​ μ ​ m > 10 \mathrm{SNR_{1.2\,\mu m}}>10 (useful to constrain the Brackett continuum) is consistent with smooth rollover within 1 σ \sigma , and in all cases because the Paschen extrapolation overshoots the Brackett continuum.

S3: 3 Results / block 1 · html

While a Paschen jump signature is inconsistent with the null hypothesis of cold thermalised blackbody continuum, we emphasise that an apparently smooth rollover (i.e. the absence of a clear Paschen jump) is not inconsistent with direct recombination emission in the partially optically thick regime seen in Sirocco models. However, plotting the significance of a Paschen spectral break versus the best-fit blackbody temperature in Fig. 5 , four T BB ≳ 5000 ​ K T_{\mathrm{BB}}\gtrsim 5000\,\mathrm{K} LRDs are consistent with a smooth rollover. These resemble the ‘Cliff’-like LRDs ( de Graaff et al. 2025a ) . If interpreted literally as blackbody temperatures, these T BB T_{\rm BB} values fall in a regime where H - -opacity can be efficient in stellar atmospheres, potentially producing pseudo-thermalised continua. Reproducing blackbody-adjacent continuum shapes may motivate the inclusion of H - -opacity within Sirocco in the future.

S2.SS2: II.2 Synthetic continuum observations / block 1 · html

The synthetic SEDs are generated using the methods described in [ 38 , 37 , 36 ] . In summary, the Monte Carlo dust radiative transfer code SKIRT [ 8 ] is used to make predictions for the ultraviolet to far-infrared SED. Dust is modeled as tracing metal-rich gas, with a dust-to-metals mass ratio of 0.4 0.4 and a [ 159 ] Milky Way prescription for its composition. Gas above 10 6 K is assumed to be dust-free to account for thermal sputtering of grains. The radiation transfer is performed on an octree grid, the cell sizes of which are adjusted so that no more than 0.0001 per cent of the total dust mass is contained in one cell. In order to probe a range of sightlines, we generate synthetic SEDs for each simulated galaxy along 7 7 lines of sight (rotating the galaxy 0 , 30 , 60 , 90 , 120 , 150 0,30,60,90,120,150 and 180 180 degrees with respect to the “face-on” orientation, which is determined by the angular momentum vector of the gas particles). Further details of the dust modeling and dust attenuation curve properties of these simulated galaxies are provided in [ 36 ] . In this paper, we explore SEDs for snapshots where simulated galaxies have sizes consistent with observed LRDs. For each selected snapshot, we compute the SED including all stars and dust within < 3 <3 kpc, which is larger than the stellar half-mass radius R eff R_{\mathrm{eff}} of the LRD candidates that we explore. We find that photons from the stellar population within R eff R_{\mathrm{eff}} dominate the SEDs in the rest frame UV to Near-IR wavelength range, so this is a sufficiently large aperture. We compute Balmer break strengths from the SEDs of FIRE galaxies using a method following previous work [ 17 , 138 ] , where the strength is defined as the ratio of fluxes F ν F_{\upnu} at 4200 Å (red flux) and 3500 Å (blue flux):

S3: III Results / block 1 · html

Figure 2: Evolutionary trajectories of simulated galaxies in the circular velocity ( V circ V_{\mathrm{circ}} ; measured within 100 pc)–stellar half mass radius ( R eff R_{\mathrm{eff}} ) plane, color coded by redshift. The shaded region indicates the parameter space occupied by observed LRDs: R eff < 300 ​ pc R_{\rm eff}<300\,\rm pc and broad Balmer line width > 1000 >1000\, km ​ s − 1 \mathrm{km\,s^{-1}} (interpreted as galaxy-scale dynamics with V circ > 500 V_{\rm circ}>500 km ​ s − 1 \mathrm{km\,s^{-1}} ). Galaxies in dark matter halos selected to reach M halo ≳ 10 13 M_{\mathrm{halo}}\,\gtrsim\,10^{13} M ⊙ at z = 2 z=2 (halos B1, B2, C1, and C2) exhibit ultra-compact, LRD-like phases in the redshift range z ≈ 4 − 8 z\approx 4-8 , while halos below this mass (A1, A2, A4, and A8; selected to reach M halo ≈ 10 12.5 M_{\mathrm{halo}}\approx 10^{12.5} M ⊙ at z = 2 z=2 ) do not become compact enough at early times. In our model, ultra-compact galaxies that feature LRD-like properties are central galaxies residing in the progenitors of group-size halos at z = 0 z=0 ( M halo M_{\mathrm{halo}} ≳ \gtrsim 10 13.5 M ⊙ ).

abstract · abstract

Since the launch of James Webb Space Telescope (JWST) in late 2021, our understanding of high-redshift objects has faced several upheavals. JWST has discovered much more massive galaxies and supermassive black holes (SMBH) than cosmological models had expected. Furthermore, JWST observations have revealed an entirely novel population of high-redshift objects. Characterized by a dominant red rest-frame component and point-like morphology, these ``little red dots'' (LRD) have set off a flurry of observational and theoretical follow-up. The current identity of LRD is highly debated, yet falling into two main scenarios: active galactic nuclei (i.e., SMBH) or compact star-forming regions. If star-forming, LRD would represent the highest stellar densities ever observed. If SMBH, their high Eddington fractions, and already high masses, help elucidate the growth of the most massive SMBH found by JWST in the early Universe ($z \gtrsim4)$. In this mini-review, we present the observational evidence accumulated to date, including sub-millimeter probes of LRD dust masses, constraints on radio and X-ray emission from stacking, and rest-frame ultraviolet \& optical measurements provided by JWST. Furthermore, we highlight how identifying additional LRD that are truly primarily SMBH-driven may help to shed light on the formation of `overly massive' SMBH discovered by JWST within the first billion years since the Big Bang.

abstract · abstract

Little Red Dots (LRDs) are compact, red sources discovered by JWST at high redshift ($z \gtrsim 4$), marked by distinctive "V-shaped" spectral energy distributions (SEDs) and often interpreted as rapidly accreting AGNs. Their evolution remains unclear, as identifying counterparts at lower redshifts is challenging. We present WISEA J123635.56+621424.2 (here dubbed {\it the Saguaro}), a $z=2.0145$ galaxy in GOODS-North, as a possible analog of high-redshift LRDs and a potential missing link in their evolutionary path toward lower-redshift systems. It features a compact LRD-like nucleus surrounded by a face-on spiral host. Its connection to LRDs includes that: (1) its nuclear spectrum shows a clear "V-shaped" SED; and (2) when redshifted to $z=7$, surface brightness dimming makes the host undetectable, thus mimicking an LRD. This suggests that high-redshift LRDs may be embedded in extended hosts. To test this, we stack rest-frame UV images of 99 photometrically selected LRDs, revealing faint, diffuse emission. Stacking in redshift bins reveals mild radial growth, consistent with the expected galaxy size evolution. A simple analytic model confirms that surface brightness dimming alone can explain their compact appearance. Lastly, we show that {\it the Saguaro} is not unique by describing similar objects from the literature at $z\lesssim3.5$. Taken together, our results support a scenario in which LRDs may not be a distinct population, but could be the visible nuclei of galaxies undergoing a short-lived, AGN-dominated evolutionary phase, with their compact, red appearance driven largely by observational biases.

abstract · abstract

Previous Sirocco radiative-transfer models of gas-cocooned AGN predicted lower-column counterparts to little red dots (LRDs): compact, X-ray-weak sources with bluer continuum slopes and Balmer jumps rather than Balmer breaks. The recently identified population of little blue dots (LBDs) closely resembles this predicted phase. Here we explore these lower-column-density cocoons in which nebular recombination emission remains visible while strong Balmer-continuum absorption is avoided. We find that a sequence of increasing column density connects more classical AGN spectra, Balmer-jump LBD-like spectra at $N_{\rm H}\!\sim\!{\rm few}\times10^{24} \mathrm{cm^{-2}}$, and Balmer-break LRD-like spectra at higher columns. In this sequence, electron scattering produces exponential line wings and suppresses X-ray emission before strong Balmer absorption features, characteristic of higher column densities, appear. We therefore propose that LBDs are lower-column analogues of LRDs within a common gas-cocooned AGN sequence. This interpretation predicts that Balmer-jump emission, X-ray weakness, permitted lines with exponential wings, He II $λ$4686 emission, smaller H$α$ FWHM values and equivalent widths than in LRDs, and weak or absent absorption features are characteristic of LBDs. We compare to three example LBD spectra and identify Balmer-jump signatures in them.

abstract · abstract

The James Webb Space Telescope recently uncovered a population of massive black holes (BHs) in the first billion years after the Big Bang. Among these high-redshift BH candidates, observations have identified a class of active galactic nuclei candidates, dubbed Little Red Dots (LRDs), with extraordinarily compact gas reservoirs and peculiar spectral features. LRDs clearly emerge at redshift z<8 and their abundance declines by z<5. Recent theoretical studies have explored the link between LRDs and the formation of heavy BH seeds in the early Universe, such as direct-collapse BHs (DCBHs). Here we present results from preliminary runs for the MELIORA cosmological hydrodynamical simulations, where we implement an accurate model for DCBH formation, accounting for the Lyman-Werner radiation field and mass-inflow rates in the target host haloes. We aim to test whether or not DCBH formation could lead to systems resembling those hypothesized for LRDs. We find that the population of newly formed DCBHs in the simulations exhibits a steep decline at z<6, akin to the emergence of LRDs, primarily driven by reduced inflows. The birth of DCBHs is associated with a significant gas compaction event, followed by a phase of intense luminosity in the 200 Myr after their birth, and subsequently by the formation of the first PopIII stars in these very haloes. If these DCBHs nurseries are associated with LRDs, then it could explain their weak emission from X-rays and hot dust.

abstract · abstract

The recent discovery of Little Red Dots (LRDs) by the James Webb Space Telescope has challenged traditional models of early galaxy and black hole co-evolution. The nature of these highly compact objects remains heavily debated, with explanations divided between dust-reddened active galactic nuclei (AGN) and extremely dense stellar populations. We perform high-resolution cosmological simulations to model the formation of LRD precursors. Motivated by recent high-redshift observations and theoretical results, we specifically explore environments characterized by high star formation efficiencies (30\% and 100\%) and confined feedback. Our simulations naturally produce highly compact galaxies with stellar masses of $10^7-6 \times 10^8 $\,M$_\odot$, with most of the mass concentrated within $200-300$ pc. We find that, in these dense environments, gas inflows, gravitational torques, and stellar dynamical friction operate on highly efficient timescales. Over a 10 Myr timescale, gas inflows can accumulate $\rm \sim 10^7 M_\odot$ at the galactic center, while gravitational torques and dynamical friction can contribute an additional $10^5-10^9$\,M$_\odot$ and $10^3-10^4$\, M$_\odot$ through the inward migration of massive stars. Assuming a conservative 10\% efficiency to account for feedback, this rapid mass accumulation can lead to the formation of a $\sim 10^6$\,M$_\odot$ central black hole, naturally giving rise to an AGN in these dense systems. Therefore, stellar and AGN interpretations of LRDs may not be mutually exclusive; rather, dense stellar systems are likely precursors to AGN.

abstract · abstract

Observations with the James Webb Space Telescope (JWST) have uncovered a substantial population of high-redshift, broad-line active galactic nuclei (AGNs), whose properties challenge standard models of black hole growth and AGN emission. We analyze a spectroscopic sample of 34 Type 1 AGNs from the JWST Advanced Deep Survey (JADES) survey, spanning redshifts 1.7 < z < 9, to constrain the physical nature of the accretion flows powering these sources with broad-line diagnostics statistically for the first time. At z > 5, we find a marked suppression of high-ionization emission lines (HeII, CIV, NV) relative to prominent broad Halpha and narrow [OIII] features. This contrast places strong constraints on the shape of the ionizing spectral energy distribution (SED) and on the physical conditions in the broad-line region (BLR). By comparing the observations to photoionization models based on SEDs of black holes accreting at sub-Eddington ratios, we show that standard AGN continua struggle to reproduce the observed broad line ratios and equivalent widths across a wide ionization parameter range. These results suggest the need for modified SEDs -- either intrinsically softened due to super-Eddington accretion or radiative inefficiencies in the innermost disk, or externally filtered by intervening optically thick gas that absorbs or scatters the highest-energy photons before they reach the BLR.