WorksPhysics and Astronomy
Gravitational Lensing of the Bullet Cluster in the CCC+TL Cosmology: A Controlled Proof-of-Concept Study
Abstract
A proof-of-concept study of gravitational lensing in the Bullet Cluster under Covarying Coupling Constants plus Tired Light (CCC+TL) cosmology. Using a JWST spectroscopic catalogue of 219 cluster members and 135 gold multiple images in 49 systems from Rihtarsic et al. (2026), we construct convergence maps for CCC+TL and a simplified CDM model using their respective angular-diameter distances and critical surface densities. The CCC+TL model adopts a two-parameter effective-density prescription anchored to the same virial-scale mass budget as CDM, rather than fitted to multiple-image positions. The CCC+TL and CDM mean-convergence bands in radial rings overlap within the propagated virial-mass uncertainty across all rings containing multiple images (θ\leq80\arcsec). On a source-plane back-projection diagnostic ( with ), the unoptimised two-parameter CCC+TL pipeline yields \mathrm{RMS}=5.22\arcsec (4.63\arcsec–5.91\arcsec), compared with 6.82\arcsec (5.75\arcsec–8.09\arcsec) for the CDM reference. The optimised 32-parameter Lenstool reconstruction achieves an image-plane RMS of \sim\!0.41\arcsec, which is not directly comparable to these source-plane values. Within the uncertainty band, the source-plane scatter ranges partially overlap. The CCC+TL effective extra-gravity distribution follows the galaxy distribution rather than X-ray gas, reproducing the qualitative morphology of the Bullet Cluster, though this agreement is partly structural as profiles center on member galaxies by construction. For the same mass, Einstein-radius scaling differs by only 1.7% between cosmologies. These results are a necessary first step toward a full image-plane likelihood test of CCC+TL.
Introduction
Gravitational lensing provides one of the most direct probes of mass in the Universe because it is sensitive to the total projected gravitating matter along the line of sight, independent of its dynamical or radiative state. In the weak-lensing regime, coherent distortions of background galaxy shapes trace the projected surface density of foreground structures, while in the strong-lensing regime multiple images, giant arcs and Einstein rings constrain the inner projected potential with high precision. For galaxy clusters, the combination of strong and weak lensing has therefore become one of the principal routes to reconstructing two-dimensional mass distributions and testing the relation between baryons and the dominant gravitating component [1, 2, 3, 4].
Cluster lensing has played a particularly important role in the modern dark-matter paradigm. The Bullet Cluster, 1E 0657-56, remains the best-known example because its merger geometry produces a spatial separation between the collisional X-ray-emitting gas and the approximately collisionless galaxy component. Chandra observations identified a bow shock with Mach number of order 2–3, implying a recent, nearly plane-of-sky merger, while weak- and strong-lensing reconstructions showed that the main lensing peaks are displaced from the dominant baryonic gas component and lie much closer to the galaxy concentrations. These results established the Bullet Cluster as a stringent empirical test for any theory seeking to explain cluster lensing without invoking conventional non-baryonic dark matter [5, 6, 7, 8, 9].
Recent JWST observations have sharpened this test considerably. Using JWST/NIRCam imaging and NIRSpec spectroscopy, Rihtaršiˇc et al. [10] constructed an updated Bullet Cluster lens model constrained by 135 secure multiple images from 27 background galaxies with spectroscopic redshifts, and explicitly modelled the system with several large-scale haloes, cluster members, intracluster gas and surrounding group-scale haloes using Lenstool. These data now make it possible to revisit cluster lensing with alternative mass prescriptions while holding the observational constraints fixed [10, 11].
The present work is motivated by the Covarying Coupling Constants (CCC) framework and its extension with Tired Light (TL), CCC+TL, in which dark matter and dark energy are interpreted not as independent substances but as emergent effects of a modified cosmological framework. In earlier work, this programme was developed at the background level and then applied to galaxy rotation curves, cosmic expansion tests and cluster-scale phenomenology, including the suggestion that the apparent dark-matter component may arise from a density-dependent effective gravitating response rather than from a separate particle species [12, 13, 14, 15, 16].
A particularly useful feature of the CCC approach is that it naturally suggests a distinction between dense baryonic cores and lower-density outer regions. In the application developed here, the baryonic distribution of each cluster member galaxy is represented by a Hernquist profile, which is well suited to spheroidal stellar systems and has the practical advantage that both its three-dimensional density and projected quantities are analytically tractable in the standard case [17]. The CCC prescription is then introduced beyond a characteristic turn-off radius, where the effective gravitating density is taken to depart from the pure baryonic Hernquist form. This construction leads to an enhanced projected surface density without altering the basic lensing formalism itself.
The purpose of this paper is not to modify gravitational lensing theory itself, but to test whether the observed Bullet Cluster strong-lensing signal can be reproduced when the usual galaxy-scale halo prescription is replaced by a CCC effective density profile tied to the observed member-galaxy positions, fluxes and derived structural scales. This makes the comparison especially transparent: the redshifts, multiple-image constraints, gas component and lensing geometry remain standard, while only the assumed galaxy-member mass prescription is altered.
Accordingly, the present analysis should be viewed as a controlled substitution test rather than as a full optimised lens-model competition. We ask whether a CCC+TL effective-density prescription, applied to the same member catalogue and observational inputs, can generate convergence maps, enclosed projected masses and source-plane scatter values consistent with those of a simplified CDM reference model within current mass uncertainties.
We therefore determine whether the CCC+TL model, applied consistently to the 219-member JWST galaxy catalogue of Rihtaršić et al. [10] and without fitting to the multiple-image positions, produces a mass distribution compatible with the observed strong-lensing signal. § 1.1 presents the theoretical framework, § 2 describes the data and pipeline, § 3 presents the results, § 4 discusses their implications and limitations, and § 5 summarises the conclusions.
The CCC+TL Cosmological Model
Fundamentals of the CCC and CCC+TL model generating extra gravity, resulting in illusionary -matter, have been presented, discussed, and applied to galaxy rotation curves, cluster observations, and gravitational lensing in earlier papers [15, 16]. Therefore, we include here only what has direct relevance to this work.
A key structural ingredient of CCC is the invariant ratio , motivated by the form of the Einstein–Hilbert action, which requires the prefactor to be time-independent if the gravitational field equations are to retain their standard form. This motivates a scaling ansatz in which only quantities with non-zero length dimensionality vary: if a quantity has length dimension , then for some dimensionless microstructural scaling function . A simple realisation is , , , and , so that is constant, and Planck units acquire a natural microscopic interpretation. The length dimensionality rule is modified slightly for the expressions involving time differentiation or integration, such as for the Hubble rate with no length dimension scaling as .
An easily workable function is with a free parameter. The Friedmann equations then acquire additional terms that behave effectively as virtual dark energy (-energy) and virtual dark matter (-matter) without invoking new particles [12, 15]. The parameter may be determined by fitting observational data, such as supernovae type 1A (Pantheon+) or CMB (Planck) data. The angular diameter distances in CCC+TL substantially exceed their CDM counterparts, particularly at high source redshift.
In the present work, the CCC+TL framework enters only through (i) the effective density prescription used for the galaxy and cluster components, and (ii) the cosmology-dependent angular-diameter distances entering .
Gravitational Lensing
In the thin-lens approximation, the observable quantity relevant for lensing is the projected surface mass density at angular vector coordinates and radial coordinate ,
where is the density and is the angular-diameter distance to the deflector (lens). The dimensionless convergence is then
with critical surface density
Here, , , and are the angular-diameter distances to the deflector at redshift , to the source at redshift , and from deflector to source, respectively, with the observer at redshift 0. These are geometric lensing factors rather than Euclidean separations, and must be computed independently from the cosmological distance relation rather than approximated by [1, 3, 4]. For a flat background, one may write .
For an axisymmetric lens, the projected mass interior to projected radius is
and the corresponding deflection angle is proportional to . In practice, standard cluster lens models represent the mass distribution as a superposition of large-scale smooth haloes, galaxy-member sub haloes and, when data permit, an explicit gas component. Modern implementations combine strong-lensing constraints in the core with weak-lensing information at larger radii, and public software such as Lenstool now supports such joint modelling in a self-consistent framework [5, 10, 11].
In the present CCC implementation, we retain this standard lensing framework and alter only the mass prescription assigned to the cluster-member galaxies. The baryonic density of each member galaxy is taken to follow a Hernquist profile,
where is the baryonic, or in practice stellar, mass assigned to galaxy , and is its characteristic radius [17].
Following the CCC prescription used in our earlier work [15, 16], we define a turn-off radius , and hence a turn-off density
For , the effective gravitating density is taken to be purely baryonic,
For , the effective gravitating density is replaced by the CCC form
While and are expected to be galaxy specific due primarily to their varying morphologies, in practice, we define a single turnoff density for all galaxies and determine for each galaxy of stellar mass using Equation (6). This computation simplification eliminates the need to know the characteristic radius of each galaxy at the cost of losing details of the lensing profile. This construction preserves the baryonic inner profile while generating a more slowly declining effective outer density derived from CCC, thereby enhancing the projected lensing signal without introducing an explicit particle-dark-matter halo. The physical motivation and the associated density-threshold picture were developed in [15, 16].
For the cluster-scale component, a pure Hernquist form is too centrally concentrated for the present application. We therefore adopt a phenomenological modified-Hernquist profile with an dependence, chosen as a simple spherical ansatz that is shallower in the outskirts and qualitatively closer to NFW-like behaviour. We emphasise that this is an empirical cluster-scale prescription, universality of which is yet to be verified:
Using the prescription similar to galaxies above, we define for clusters turn-off radius and turn-off density .
As is the case of galaxies, for , the effective gravitating density for clusters is taken to be purely baryonic,
For , the effective gravitating density is replaced by the CCC form
As we will see below, the turn-off density at the galaxy scale is substantially higher than at the cluster scale.
Now we can consider the projection of the density profiles to lensing quantities. For a spherically symmetric galaxy profile, the projected surface density of each galaxy is obtained via the Abel integral,
Here, for a galaxy centred at ,
Then the total galaxy contribution on the cluster map is
When the gas map is available, it is added as a separate layer:
The convergence is defined as
Because the lensing formalism remains unchanged, the test of the CCC prescription is reduced to a comparison of the resulting convergence field and image-plane predictions against those obtained from the standard parametric halo model.
The convergence test used in this work is, therefore, straightforward. First, a baseline lens model is constructed using the standard flux-scaled galaxy prescription employed in the observational study. Secondly, the galaxy-member contribution is replaced by the CCC effective profile while the observational constraints, redshifts, gas component, and lensing geometry are kept fixed. The resulting , aperture masses and, where relevant, image residuals can then be compared directly. If the CCC prescription is viable, it should reproduce the principal convergence peaks and lensing strengths without requiring the same galaxy-associated dark-halo component as in the standard model. This type of controlled substitution is especially natural for the recent Bullet Cluster JWST analysis, where the member catalogue and F277W-based scaling information are already available from the modelling framework of Rihtaršič et al. [10].
At the Bullet Cluster geometry (, representative source redshift ), the key quantities are compared in Table 1.
| Quantity | CDM | CCC+TL | (CCC+TL/CDM) |
| (Mpc) | 910.3 | 1090.3 | 1.198 |
| (Mpc) | 1726.7 | 4408.4 | 2.553 |
| (Mpc) | 1333.4 | 3937.4 | 2.953 |
| 0.7723 | 0.8932 | 1.156 | |
| ( kpc) | 0.722 | ||
| kpc per arcsec | 4.4131 | 5.2859 | 1.198 |
Table 1. Cosmological distances and critical surface densities at the Bullet Cluster geometry (, ). The CCC+TL and CDM distances are determined following Gupta [12].
The CCC+TL (CTL) distances are larger than CDM (LC) by a factor of , so all physical lengths in the CCC+TL pipeline scale as their CDM values. The critical surface density, , is 27.8% lower in CCC+TL than in CDM at . These two effects—lower and larger physical scales ( suppression of projected surface density)—almost exactly cancel, as shown in § 1.4.
Hereafter, since the density prescription described above is an integral component of the CCC+TL model, we refer to all derived quantities—effective density profiles, -matter distributions, and convergence maps—simply as CCC+TL (abbreviated CTL in figures and tables), without further reference to CCC alone unless necessary.
Two-Parameter CCC+TL Solution
The CCC+TL implementation used here contains two calibration parameters. These are not fitted to the multiple-image positions, but are fixed by requiring consistency with the adopted virial-scale mass budget of the cluster.
Galaxy-scale turn-off (): The turn-off radius of each galaxy follows the density prescription [16]
where is the reference density calibrated from galaxy rotation curves, an average of all seven galaxies of different morphologies that Gupta [15] studied. The galaxy-scale normalization parameter is calibrated such that the total CCC+TL projected mass at the virial scale matches the adopted reference-model mass budget, i.e., the CCC+TL and reference CDM pipelines are normalised to have the same total projected mass within . Notably, the projected mass inside a finite aperture is not the same as the total 3D virial mass within .
The present work does not establish a universal value of ; rather, it tests whether values consistent with earlier galaxy-scale work can also reproduce the Bullet Cluster lensing signal.
Cluster-scale turn-off (): The cluster-scale -matter is seeded by the total cluster stellar mass, , where is the stellar mass fraction [18]. Specifically, is the total mass enclosed within a sphere of radius , within which the mean mass density is exactly 200 times the critical density of the universe at the cluster’s redshift. Its turn-off parameter is set by requiring the cluster-scale -matter mass to equal the CDM dark matter mass within :
where is the gas fraction. This gives , corresponding to a turn-off radius for the main cluster () calculated using
Both parameters are derived without reference to the multiple-image positions, so the CCC+TL pipeline has effective degrees of freedom.
We do not assume that the Bullet Cluster is dynamically relaxed or in hydrostatic equilibrium. The use of virial-scale quantities such as serves here only as a normalization convention for the projected lensing comparison, while the actual observable test is carried by the convergence field and source-plane scatter.
Einstein Radius Scaling Between Cosmologies
For an axisymmetric lens, the Einstein radius satisfies , or equivalently . For the same physical mass distribution expressed in angular units, the change in physical scale per arcsecond between the two cosmologies modifies the projected surface density by a factor , while the lensing threshold changes through . Hence for the same cluster mass the Einstein radius in arcseconds rescales approximately as (CTL is abbreviation for CCC+TL and LC is for CDM)
Substituting the values of Table 1:
The lower in CCC+TL (which enlarges ) is almost exactly cancelled by the larger physical scales (), giving a net difference of only 1.7%. This near-perfect cancellation is not a coincidence: it reflects the same compensation found for the convergence itself (), and means that for the same mass, CCC+TL and CDM predict essentially identical Einstein radii in arc-seconds.
The uncertainty in propagates as , giving a spread in that is common to both models.
Our framework follows the observational decomposition adopted by Rihtaršiˇc et al. [10], who model the Bullet Cluster with multiple large-scale haloes, cluster members, intracluster gas, and surrounding group-scale structure. The present work does not attempt to reproduce that full optimised model, but instead performs a controlled substitution test in which the standard galaxy-halo prescription is replaced by a CCC+TL effective-density prescription while retaining the same observational inputs as far as possible. Group-scale substructure and full dynamical modelling remain beyond the scope of the present pilot calculation.
Data and Pipeline
Observational Data
We use three observational inputs from Rihtaršić et al. [10].
Member-galaxy catalogue. This comprises 219 spectroscopically confirmed cluster members with ICRS (International Celestial Reference System) coordinates, F277W AB magnitudes ( mag), and spectroscopic redshifts. Stellar masses are derived from the calibration of Rihtaršić et al.,
giving . A physical scale of kpc is used for CDM [10]; the CCC+TL pipeline uses kpc.
Gold multiple-image catalogue. This comprises 135 secure multiple images from 49 lensed systems (27 distinct galaxies), all with spectroscopic redshifts spanning (Figure 1).
Figure 1. JWST member-galaxy catalogue. (Top): projected sky positions of the 219 JWST spectroscopically confirmed cluster members relative to the main cluster BCG (red star), colour-coded by F277W
AB magnitude. The bullet subcluster is to the west (right in this plot and left in the sky; east is left). (Bottom): F277W magnitude distribution of the 219 member galaxies. The catalogue spans both the main cluster and the bullet subcluster.
3. Cluster gas. We use the Chandra-derived gas mass map from Clowe et al. [6], as adopted in the Lenstool model of Rihtaršić et al. [10]. In our pipeline, the gas is represented by a calibrated dual elliptical -model [19], with total gas mass of within in each subcluster.
Pipeline Architecture
Two fully independent pipelines are run simultaneously—one for CDM and one for CCC+TL—using their respective angular diameter distances and . The pipelines share the same galaxy catalogue but differ in all physical length and mass scales. In the CCC+TL pipeline, all projected positions and profile radii scale by relative to their CDM values.
Galaxy component. For each of the 219 catalogue members, the surface density is computed numerically via the Abel integral (Equation (12)) for the CCC+TL effective density profile (Equation (8)), with determined by Equation (17). The total galaxy surface density is . In the CDM pipeline, the galaxy contribution is computed using the analytic Hernquist profile [17].
Cluster -matter component (CCC+TL only). A cluster-scale CCC+TL profile is centred on each BCG (main cluster and bullet subcluster), seeded by . This component has no CDM counterpart; it is the functional replacement for the NFW dark matter halo. The galaxy-scale and cluster-scale -matter profiles are additive because they are seeded by distinct stellar masses at different physical scales: the cluster profile uses as its seed (absorbing uncatalogued members into the cluster-scale component, paralleling how CDM absorbs them into the cluster NFW halo), while individual galaxy profiles use from the photometric catalogue.
NFW dark matter (CDM only). Two NFW halos [20] are centred on the BCG positions, with masses and concentrations from Springel and Farrar [21]; see also [22]: , ; , .
Convergence maps. The convergence map for each model is
where is model specific (Table 1) and is a single value per model, evaluated at the reference source redshift (see Equation (24) for the per-source rescaling of the deflection field). Maps are computed on a pixel grid covering (pixel scale ); convergence has been verified against a grid at /pixel (Appendix § 6).
In projection, the galaxy-scale and cluster-scale CCC+TL components overlap spatially on the convergence map, but they are not intended as duplicate mass assignments; rather, they represent effective gravitating responses seeded by distinct baryonic structures at different physical scales.
Deflection angles. The deflection angle field is computed from each map via discrete Fourier-transform convolution,
where is the pixel size of the grid over used to produce Table 4 and Figure 7; convergence against a grid at /pixel is better than 1.5% on all three scatter metrics (Appendix § 6). We refer to the right-hand side of Equation (23) as . Since above is built with evaluated at a single reference source redshift , the field is the reduced deflection for a source at . For an image at a source redshift , the physical reduced deflection is obtained by rescaling relative to the reference:
| Model | RMS All Images | Mean per System | Median per System | |
| CDM unoptimised () | 2 | 6.82″ (5.75″–8.09″) | 5.93″ (5.00″–7.02″) | 6.71″ (5.78″–8.27″) |
| CCC+TL () | 2 | 5.22″ (4.63″–5.91″) | 4.44″ (3.92″–5.05″) | 4.92″ (4.55″–5.59″) |
| CTL/unopt-ΛCDM | – | 0.77× | 0.75× | 0.73× |
Table 4. Source-plane back-projection scatter: RMS spread of back-projected source positions across all 135 images, and mean/median per lensed system ( with ). This diagnostic requires no assumed and is not optimised by any model; it is not a goodness-of-fit statistic (see § 4.2). All values are computed under Equation (24) with on the grid specified in Appendix § 6. Ranges in parentheses span the full uncertainty band; the central value is quoted first. Both rows are reproducible from the input catalogues using the released pipeline (Appendix § 6). The Lenstool-optimised reconstruction of Rihtaršiˇc et al. [10] achieves an image-plane RMS of ∼0.41″ on this gold catalogue; this is a different quantity than the source-plane scatter tabulated here and the two cannot be compared directly.
Figure 7. Source-plane back-projection scatter, computed under Equation (24) with on the pixel grid described in Appendix § 6. (Top): aggregate RMS, mean-per-system, and median-per-system scatter for the two simplified two-parameter models; error bars span the full uncertainty band. (Middle): per-system CCC+TL versus unoptimised-CDM scatter, colour-coded by source redshift; points below the diagonal correspond to systems where CCC+TL yields smaller scatter, which is the case for the great majority of systems in the sample. (Bottom): ratio of CCC+TL to unoptimised-CDM scatter for each system, ordered by ratio value; system identifiers are staggered on two rows for readability.
with computed in each cosmology independently. Equation (24) is cosmology-consistent and reduces to the identity at by construction; for the Bullet Cluster gold sample (–), the ratio varies between 1.128 and 1.163, so the per-source rescaling contributes a sub-5% differential between the two cosmologies on the source-plane back-projection diagnostic. The pipeline implementing Equations (23) and (24) is available at https://github.com/rgupta-uottawa/bullet-ccctl URL (accessed on 14 October 2026).
Results
Convergence Maps and Radial Profiles
Figure 2 shows the convergence maps for both models. The CCC+TL map is dominated by the 219 individual galaxy profiles at small projected radii and by the extended cluster -matter envelope at larger radii; there is no dark matter halo. The CDM map is dominated by the NFW halo at all radii. Qualitatively, both models produce mass peaks that trace the galaxy distribution rather than the gas—this spatial coincidence is automatic in CCC+TL because the -matter density profile is centred on individual galaxies, not on the gas.

Figure 2. Convergence maps for ΛCDM and CCC+TL (CTL). Each model uses its own and angular-diameter distances. Positive points east; west is to the right following standard astronomical orientation. (Top left): ΛCDM reference model comprising NFW dark matter haloes, calibrated -model gas, and Hernquist stellar profile. (Top right): CCC+TL total convergence, comprising galaxy-scale and cluster-scale CTL profiles plus gas, with no explicit dark matter halo. (Bottom left): Cluster-scale CCC+TL -matter component only (seeded by centred on each BCG), shown with CTL ; the main-cluster turn-off radius is kpc. (Bottom right): Galaxy-scale CCC+TL profiles only (219 individual member galaxies) shown with CTL ; the granular structure reflects the discrete positions of the catalogue members. In all panels, symbols mark the main BCG (yellow star), bullet BCG (green star), main gas peak (red cross), and bullet gas peak (orange cross).
Figure 3 shows the ΛCDM and CCC+TL mass component peaks superimposed on the CCC+TL contours. Peaks are very close to each other and hard to decipher. Peak positions are tabulated in Table 2.

Figure 3. Mass component peak positions overlaid on the CCC+TL total-convergence contours. The panels show the full field and zoomed views of the main-cluster core and bullet subcluster. The plotted symbols identify the BCGs, gas peaks, ΛCDM NFW peaks, and the total peaks for both models.
Table 2. Angular positions of mass and gas component peaks relative to the main cluster BCG (, ). Positive RA offset is east and positive Dec offset is north. Physical radii in the fourth and fifth columns use the scale appropriate to each model ( kpc for CDM and kpc for CCC+TL). The CDM total peak is dominated by the bullet-subcluster NFW halo, whereas the CTL total peak lies near the main BCG.
| Component | (arcsec) | (arcsec) | (kpc) | (kpc) | Model |
| Main BCG (origin) | 0.0 | 0.0 | 0 | 0 | Both |
| Bullet BCG | Both | ||||
| Main gas peak (Chandra) | 0.0 | 0 | Both | ||
| Bullet gas peak (Chandra) | Both | ||||
| CDM NFW DM peak (main) | 0.0 | 0.0 | 0 | 0 | CDM |
| CDM NFW DM peak (bullet) | CDM | ||||
| CDM total peak | CDM | ||||
| CCC+TL total peak | CCC+TL |
Table 2.
Figure 4 shows the azimuthally averaged radial profiles in arcseconds—the natural common observable for both cosmologies. The CCC+TL profile is more extended at small radii () due to the sharply peaked galaxy profiles, and falls off more steeply than CDM at where the NFW halo dominates.
Figure 4. Azimuthally averaged radial profiles. (Top): the CDM reference model. (Bottom): the CCC+TL (CTL) model. Profiles are plotted against projected angular radius in arcseconds, which is the common observable for both cosmologies. The upper axis gives the corresponding physical radius for each model. The horizontal dashed line marks .
Mass Peak Positions and Morphological Consistency
The locations of the total peaks are useful morphological diagnostics, but they are not themselves the primary observables. The strong-lensing constraints are determined by the full deflection field sampled at the multiple-image positions. Accordingly, the peak comparisons below are interpreted qualitatively, while quantitative model performance is assessed in Sections 4.3–4.5 through enclosed projected mass, ring-averaged convergence and source-plane scatter.
A defining observational constraint of the Bullet Cluster is the spatial separation between the gravitational lensing mass peaks and the X-ray gas peaks: in the CDM framework, the mass peaks follow the collisionless galaxy distributions, while the gas lags behind, offset by – kpc [7]. In the CCC+TL model, this morphological pattern emerges automatically, without any tuning, because the -matter density profiles are seeded by and centred on individual galaxies—not on the gas.
Figure 3 shows all component peak positions overlaid on the CCC+TL convergence contours. Table 2 gives their angular coordinates relative to the main BCG (the coordinate origin), with east to the left following the standard astronomical convention adopted throughout this work. Several features of Table 2 are physically informative: First, the CDM total peak falls at (″, ″), within ″ of the bullet subcluster BCG and NFW halo centre. This is expected: the bullet subcluster has a higher NFW concentration. ( vs. for the main cluster), making its NFW profile more centrally peaked per unit mass. Second, the CCC+TL total peak is at (, )— ( kpc in CCC+TL) from the origin. This position is displaced from the main gas peak (, ) by in the north direction, and from the main BCG by . The displacement reflects the combined effect of the galaxy CCC+TL profiles (which individually peak at their host galaxies) and the cluster -matter profile (centred on the BCG). The CCC+TL mass distribution is displaced from the main gas peak and is instead associated with the observed galaxy distribution, reproducing the qualitative Bullet Cluster signature that the effective gravitating mass need not trace the dominant X-ray gas component. We stress, however, that the exact location of the total maximum is not by itself the decisive lensing observable. Third, both the CCC+TL and CDM mass peaks are substantially offset from the main gas peak (, ). In CDM the offset is because the total is dominated by the bullet subcluster NFW halo; in CCC+TL the main cluster galaxies and cluster -matter produce a peak north of the gas, consistent with the spatial distribution of the brightest cluster members. Both reproduce the qualitative result that peak does not coincide with the X-ray gas centroid.
Enclosed Projected Mass
Figure 5 and Table 3 compare the enclosed projected mass as a function of angular radius. At (near the observed Einstein radius of the main cluster), CCC+TL encloses compared to for CDM, a ratio of 1.19. Table 3 shows that the CCC+TL-to-CDM enclosed-mass ratio remains above unity at and falls below unity by , implying a crossing between those radii, close to , where CCC+TL and CDM enclose essentially equal projected mass ( and , respectively). These differences are well within the uncertainty on (§ 3.4).
| (arcsec) | (kpc) | (kpc) | () | () | Ratio (CTL/CDM) |
| 20 | 88 | 106 | 0.45 | 0.51 | 1.15 |
| 30 | 132 | 159 | 0.96 | 1.14 | 1.19 |
| 50 | 221 | 264 | 2.28 | 2.94 | 1.29 |
| 100 | 441 | 529 | 7.64 | 8.69 | 1.14 |
| 160 | 706 | 846 | 14.98 | 14.36 | 0.96 |
Table 3. Enclosed projected mass for the central mass scenario (). Physical radii are given in CDM kpc ( kpc) and CTL kpc ( kpc).
Figure 5. Enclosed projected mass . (Top): cumulative projected mass as a function of angular radius for the CDM reference model (blue) and CCC+TL (red), both at the central . Dashed curves show the individual components. (Middle): CTL/CDM ratio, showing the crossing near . (Bottom): mean convergence for each model.
Mass Uncertainty and Kappa Band Overlap
The virial mass carries a uncertainty for the main cluster and for the bullet subcluster [21, 22]. This propagates differently into the two models: in CDM, the dark matter halo dominates the budget and scales linearly with , giving – variation in mean across all radial rings (Figure 6). In the present implementation, the 219-galaxy contribution is held fixed by the photometric catalogue and does not vary when the virial-mass prior is perturbed; only the gas and cluster-scale -matter components are rescaled with . This is a property of the adopted pipeline rather than a statement that the galaxy component is physically decoupled from the virial mass definition. As a result, CCC+TL shows a smaller – spread, increasing toward larger radii where the cluster profile dominates.
Figure 6. Effect of the uncertainty on in radial rings. (Top): CDM mean per ring across the range. Middle: CCC+TL mean per ring. (Bottom): overlap of the CDM and CCC+TL bands; green ticks denote overlapping rings and red crosses non-overlapping rings.
Figure 6 shows the mean convergence in five radial rings for both models across the mass range. The key result is that the CDM and CCC+TL bands fully overlap within the strong-lensing regime (), i.e., in the three innermost rings that contain all multiple images. At larger radii the bands no longer overlap, but that region is probed primarily by weak lensing [2] and is not modelled here.
Source-Plane Scatter
Source-plane back-projection provides a self-consistency diagnostic of each deflection field that is independent of any assumed positional uncertainty and is not minimised by either model. For each lensed system with images at observed positions , we compute the back-projected source positions using the convention of Equation (24),
where is computed in the respective cosmology and . The scatter for system is
where is the mean source position. A perfect lens model maps all images of a given source to exactly the same source positions (); the RMS scatter over all 135 images measures the geometric self-consistency of each deflection field, but is not a goodness-of-fit statistic: it depends on the lens mapping and local magnification, and is not minimised by either simplified pipeline nor by the fully optimised Lenstool reconstruction.
The results are shown in Figure 7 and Table 4. The unoptimised two-parameter CCC+TL pipeline gives RMS = 5.22″ at the central (4.63″–5.91″ across the mass band), compared with 6.82″ (5.75″–8.09″) for the equivalent unoptimised two-parameter CDM reference. Both values are reproducible from the input catalogues using the released pipeline (Appendix § 6). For context, the fully optimised 32-parameter Lenstool reconstruction of Rihtaršiˇc et al. [10] achieves an image-plane RMS of ∼0.41″ on this gold catalogue, i.e., the two simplified two-parameter pipelines considered here are at least an order of magnitude from the image-plane performance of the optimised model. We stress that the Lenstool image-plane RMS and the source-plane back-projection RMS in Table 4 are different quantities, so a direct numerical comparison is not possible without rerunning the Lenstool model and back-projecting its deflection field under the convention of Equation (24); we therefore do not attempt such a comparison here.
Two further observations follow. First, within the ±30% uncertainty band the CCC+TL and CDM scatter ranges partially overlap (4.63″–5.91″ and 5.75″–8.09″, respectively); the two simplified pipelines produce source-plane scatter of comparable magnitude on this diagnostic, but this does not by itself establish that either model reproduces the observed multiple images—that question requires an image-plane analysis. Second, the cosmology-specific per-source rescaling of Equation (24) differs between the two models by less than 5% at every image redshift in the sample, so the model-to-model difference in Table 4 is driven by the different -field structures, not by the per-source rescaling itself.
The band broadens both scatter ranges to the point where they partially overlap, as noted above. We emphasise that this overlap is between two model predictions and does not test either model against the observed multiple-image positions themselves; a statement about whether the data distinguish the two cosmologies requires an image-plane likelihood analysis, which we identify as follow-up work (§ 4.6).
Discussion
Consistency Within Mass Uncertainties
The central result of this work is that the CCC+TL and CDM convergence bands overlap across the entire strong-lensing region (, corresponding to kpc in CCC+TL) when the uncertainty is propagated. This overlap is between two model predictions constructed under the common virial-scale mass-budget anchor described in § 1.3; it does not by itself establish that either model reproduces the observed multiple-image positions, nor that the data cannot distinguish the two models. The latter question requires a full image-plane analysis (§ 4.6. The divergence at reflects the different profile shapes at large radii: the CDM NFW halo has a shallower fall-off than the CCC+TL cluster -matter profile, but this regime is beyond the multiple-image region considered in this work and is accessible only to weak lensing.
An important asymmetry is that the CCC+TL mass uncertainty band is narrower than CDM’s (– vs. – in the inner rings), because the 219-galaxy CCC+TL contribution—which dominates the convergence at small radii—is anchored to the photometric catalogue and is completely independent of . This is a testable prediction: future measurements of with reduced uncertainty will shrink the CDM band faster than the CCC+TL band.
On the source-plane back-projection diagnostic (Table 4), the two-parameter CCC+TL pipeline gives RMS = at the central , below the corresponding of the two-parameter CDM reference. Both pipelines are an order of magnitude from the image-plane RMS achieved by the full 32-parameter Lenstool reconstruction of Rihtaršiˇc et al. [10], although as noted in § 3.5 that is a different quantity than the source-plane scatter tabulated here. Within the uncertainty band, the CCC+TL and CDM scatter ranges partially overlap. This overlap is between two model predictions and does not amount to a test against the observed multiple images themselves; whether the data distinguish the two cosmologies in the strong-lensing region therefore remains open and is deferred to a future image-plane analysis. Because the CCC+TL -matter components and itself are calibrated to the CDM virial-scale mass budget by construction, Table 4 also does not demonstrate that CCC+TL independently predicts the Bullet Cluster lensing signal; it demonstrates that, given the same virial-scale mass budget, the CCC+TL effective-density prescription produces a self-consistent deflection field of comparable magnitude to the simplified CDM pipeline on this diagnostic.
The Source-Plane Scatter as the Appropriate Metric
The source-plane back-projection diagnostic is informative at the present stage because it requires no assumed positional uncertainty, is not the objective function minimised by either of the two unoptimised reference pipelines considered here, and probes the geometric self-consistency of the inferred deflection field. It is not, however, a goodness-of-fit statistic: it depends on the lens mapping and local magnification, and cannot by itself discriminate between candidate deflection fields that reproduce the data equally well in the image plane. The decisive goodness-of-fit metric is the image-plane or RMS of back-projected images forward through the lens equation, which for the Lenstool-optimised reconstruction of Rihtaršiˇc et al. [10] on this gold catalogue is ; the source-plane values in Table 4 are not directly comparable to it. Within that limitation, the source-plane scatter nevertheless provides a transparent, optimisation-independent diagnostic for the present controlled substitution test. A statistically meaningful comparison between the CCC+TL and CDM pipelines on the Bullet Cluster multiple-image catalogue requires a full image-plane likelihood with marginalisation over the fixed inputs listed in § 4.6, which we identify as the subject of a dedicated follow-up paper.
Physical Interpretation of the Alpha-Matter Profile
In the CCC+TL model, -matter is not a substance but a scale-dependent gravitational effect: in dense galaxy cores it is negligible, while in the low-density cluster outskirts it provides the additional gravitational attraction that in CDM would be attributed to dark matter. Critically, the -matter distribution inherits its spatial structure entirely from the stellar mass distribution of individual cluster members. Because galaxies and gas decouple during the Bullet Cluster merger, -matter automatically follows the galaxies—not the gas—reproducing the spatial offset that Clowe et al. [7] identified as the defining evidence for dark matter. No parameter tuning is required for this morphological agreement; it is a structural feature of the CCC+TL prescription.
We stress that this morphological agreement is partly structural: both the galaxy-scale CCC profile (one component per catalogue galaxy) and the cluster-scale CCC profile (two components centred on the main BCG and the bullet BCG) are, by construction, anchored to the galaxy distribution rather than to the gas. The observational statement that the Bullet Cluster lensing mass follows the galaxies rather than the gas is therefore reproduced in CCC+TL by the same mechanism as in CDM—mass components anchored to the collisionless component. It is not an independent CCC+TL prediction, but a consequence of how the CCC+TL effective-density prescription is attached to the baryonic structure.
Recently, Hernandez [23] has shown, within a QUMOND framework, that the standard objection that the Bullet Cluster lensing peaks must necessarily track the dominant X-ray gas is too simplistic, since compact galaxy-scale baryonic concentrations can generate a much stronger effective projected gravitating signal than an extended gas component. This reinforces the broader point that the Bullet Cluster constrains not only the total baryon budget, but also how a given theory maps baryonic structure into the inferred lensing mass distribution.
The work of Hernandez [23] is useful in demonstrating that the conventional Bullet Cluster argument against non-standard gravity is less decisive than is often claimed, since a galaxy-centred effective mass map can arise even when the X-ray gas dominates the baryonic mass budget. Nevertheless, that treatment is intentionally exploratory: it adopts plausible baryonic parameters and does not attempt a full fit of either the baryonic model or the QUMOND transition-function parameters. The present CCC+TL study goes beyond that level by constructing the effective gravitating density from an explicitly specified CCC prescription tied to the observed cluster-member properties and by propagating that prescription through the full projected-density and convergence calculation. The similarity between the two works is therefore conceptual, whereas the difference is methodological: Hernandez [23] provides a qualitative proof of principle, while the present analysis is a quantified substitution test within an observationally anchored lensing framework.
Famaey [24] has recently emphasised that the Bullet Cluster remains problematic for MOND in the same sense as other massive clusters: although the effective lensing signal can be associated primarily with the galaxy distribution rather than the X-ray gas, a residual missing-mass component is still required, and this component must be mostly collisionless and galaxy-centred. The present CCC+TL analysis shares the first of these conclusions but not the second. As in MOND, the effective gravitating signal in CCC+TL need not coincide with the dominant gas component. Unlike MOND, however, CCC+TL is not being tested here as a pure force-law modification with no additional effective mass contribution. Instead, the theory is implemented through an explicit -matter density prescription tied to the observed baryonic structure at galaxy and cluster scales, so that the relevant question is not whether a residual missing mass remains in the MOND sense, but whether this internally specified effective mass prescription can reproduce the observed Bullet Cluster lensing field in a controlled comparison.
Comparison with Earlier CCC+TL Results
The galaxy rotation curve study using the CCC framework [15] yields averaged over seven galaxies. Thus, we have taken . The present cluster-lensing analysis requires for the galaxy-scale component, corresponding to , consistent with the earlier study on the galaxy rotation curves. The cluster-scale -matter has (), which is substantially lower because the cluster envelope operates on a much larger scale () than individual galaxy halos (–) and uses a different density profile than for the galaxies.
Merger Dynamics and Bow Shock
An important complementary issue, not addressed in the present work, is the merger dynamics of the Bullet Cluster itself. In particular, the observed bow shock has often been associated with a very high collision or shock speed, of order 4700 , which has long been discussed as a potential challenge for standard CDM interpretations if taken too directly as the subcluster speed. However, hydrodynamical modelling has also shown that the shock speed need not be identical to the mass-centroid velocity of the bullet subcluster, thereby reducing the severity of this tension. More generally, extreme cluster collisions continue to motivate discussion about whether very high infall velocities are fully natural in CDM. The present paper is concerned only with the projected strong-lensing mass distribution, and therefore does not attempt a dynamical treatment of the merger or of the bow-shock kinematics. A full analysis of Bullet Cluster merger dynamics in the CCC+TL framework would provide an important complementary test, but it lies beyond the scope of the present lensing-focused study.
Limitations and Future Work
Fixed modelling inputs and effective model complexity. Although CCC+TL is described above as a two-parameter effective-density prescription, the full pipeline depends on a set of fixed inputs that are shared with the simplified CDM reference pipeline: the adopted cluster masses and [21, 22]; the stellar-mass fraction [18]; the gas fraction ; the [10] FZ77W calibration; the inner Hernquist and outer CCC galaxy-profile forms with prefactor fixed by continuity; the phenomenological cluster-scale CCC profile with prefactor fixed by continuity; the dual- gas model [19]; and the BCG and gas-peak positions. Only the two normalisation parameters and are free; the free-parameter count in Table 4 refers to these two only. The CDM reference pipeline is substantially simpler than the full JWST Lenstool reconstruction, which also includes group-scale and large-scale components. A fair confrontation with the fully optimised Lenstool model would require either expanding the CCC+TL pipeline to include its group-scale and large-scale components, or applying the same two-parameter restriction to the Lenstool model. We identify both as subjects of the follow-up paper.
Gas representation. In the present implementation, the gas component is represented by a calibrated dual -model approximation to the Chandra gas map. This provides a practical way to include the dominant X-ray gas contribution in the projected mass budget while keeping the analysis analytically transparent. A future implementation based directly on the FITS gas maps used in the observational modelling framework of Rihtaršič et al. [10] would allow a more detailed treatment of gas substructure, but is not required for the present strong-lensing comparison.
Adopted density prescriptions. The CCC+TL implementation uses a Hernquist baryonic profile for individual galaxies and a phenomenological modified-Hernquist form for the cluster-scale component. These choices provide a simple and physically motivated first framework for testing the projected lensing signal, but they are not intended to exhaust the range of possible profile choices. In particular, future work may explore alternative galaxy and cluster parameterisations, including profiles calibrated more directly to observed light distributions, in order to test how sensitive the inferred convergence field is to these modelling assumptions.
Catalogue coverage and larger-scale structure. The present analysis is tied to the currently available 219-member JWST spectroscopic catalogue and should therefore be understood as a catalogue-based realization of the CCC+TL prescription in the strong-lensing core. Additional member identification, wider-field substructure, and explicit group-scale components may refine the outer projected mass distribution, as also suggested by the observational framework of Rihtaršiˇc et al. [10]. In the present work, such larger-scale contributions are effectively absorbed into the cluster-scale term, which is sufficient for the controlled comparison carried out here.
Cluster-scale normalization. The adopted stellar mass fraction is used here as a representative normalization for the cluster-scale stellar component, whose role is to seed the large-scale CCC+TL -matter contribution. The present analysis is therefore not intended as a precision determination of , but as a test of whether a plausible cluster-scale stellar normalization, combined with the observed galaxy catalogue, can account for the strong-lensing observables within the current mass uncertainties. Because the CCC+TL critical density differs substantially from its CDM counterpart—owing to the additional term in the effective Hubble rate—the conventional virial mass does not carry the same physical meaning across the two models. We therefore treat as a label for the total cluster mass inferred from CDM lensing analyses [21, 22], and use it as a common normalisation anchor for both pipelines.
No image-plane optimisation. The present paper does not perform a full image-plane optimisation of the CCC+TL model against the multiple-image positions. Instead, it is formulated as a controlled substitution test, in which the standard galaxy-halo prescription is replaced by the CCC+TL effective-density prescription while the observational inputs are otherwise kept fixed as far as possible. A future optimised implementation would provide a more direct comparison with the optimised Lenstool solution, but the absence of such an optimisation does not affect the interpretation of the present results as a first strong-lensing consistency test.
Component overlap in projection. The galaxy-scale and cluster-scale CCC+TL contributions overlap in projection on the convergence map, but they represent effective gravitating responses seeded by baryonic structure at different physical scales rather than duplicate assignments of the same component. This additive construction is sufficient for the present pilot calculation, while a more hierarchical treatment of galaxies, groups, gas, and cluster-scale structure may be explored in future work.
Restriction to the strong-lensing regime. The present work is restricted to the projected strong-lensing region, where the multiple-image constraints lie and where the comparison between the CCC+TL and reference CDM convergence bands is most direct. At larger projected radii, weak-lensing shear is expected to provide the more discriminating observable, since it is more sensitive to the outer slope and spatial extent of the projected mass distribution. A joint strong- plus weak-lensing treatment in the CCC+TL framework is therefore a natural next step, but lies beyond the scope of the present paper.
Merger dynamics. The present paper addresses the projected lensing mass distribution and does not attempt a dynamical treatment of the Bullet Cluster merger itself. Questions related to the collision velocity, bow-shock structure, and detailed merger geometry provide an important complementary test of any cosmological framework, but they involve additional hydrodynamical and dynamical considerations beyond those required for the present lensing analysis. A corresponding treatment in the CCC+TL framework would therefore be valuable, but is deferred to future work.
Conclusions
We have performed a controlled proof-of-concept comparison between the CCC+TL cosmology and a simplified CDM reference pipeline on the JWST Bullet Cluster strong-lensing catalogue, using a two-parameter prescription whose two calibration parameters—the galaxy-scale turn-off normalisation and the cluster-scale turn-off normalisation —are set from physical reference values together with mass-budget consistency against the common CDM virial-scale mass used by both pipelines, not from the multiple-image positions themselves. The principal findings are:
Convergence band overlap. Within the uncertainty, the CCC+TL and CDM mean-convergence bands fully overlap in all radial rings containing multiple images (). This is a statement about two model predictions evaluated on a common virial-scale mass-budget anchor; it does not by itself establish that either model reproduces the observed multiple-image positions, nor that the data cannot distinguish the two models.
Einstein radius equivalence. For the same cluster mass, the CCC+TL and CDM Einstein radii in arcseconds differ by only 1.7%, because CCC+TL’s lower is almost exactly offset by its larger physical scales ().
Source-plane back-projection scatter. On the geometric source-plane back-projection diagnostic—which requires no assumed positional uncertainty and is not minimised by either model—the two-parameter CCC+TL pipeline gives RMS = at the central (– across the mass band) and the equivalent two-parameter CDM reference gives (–). The two ranges partially overlap within the mass-uncertainty band. Both values are an order of magnitude from the image-plane RMS achieved by the full 32-parameter Lenstool reconstruction of Rihtaršić et al. [10], although these are different quantities and the comparison is not direct. Source-plane RMS, as discussed in § 4.2, is not a goodness-of-fit statistic.
Automatic morphological consistency. The CCC+TL -matter distribution naturally traces the galaxy positions rather than the gas peaks, reproducing the defining morphological signature of the Bullet Cluster without any positional tuning; we stress, however, that this agreement is partly structural, since the CCC+TL -matter components are anchored to the galaxy distribution by construction (§ 4.3).
Within the scope of a controlled normalisation-anchored substitution, this work demonstrates that a two-parameter CCC+TL effective-density prescription can be constructed to reproduce the projected mass distribution of a simplified CDM reference model in the strong-lensing region of the Bullet Cluster, within mass uncertainty. The present results are a necessary first step toward a full image-plane likelihood test of CCC+TL on the Bullet Cluster, but not a sufficient one: a quantitative falsification test requires a full image-plane fit to the multiple-image catalogue with nuisance-parameter marginalisation and with group-scale and large-scale components matched between the two models. We identify this as the natural subject of a subsequent paper. The full pipeline producing the numerical results of this paper, together with the input catalogues, is available at https://github.com/rgupta-uottawa/bullet-ccctl (accessed on 14 October 2026).
Author Contributions: Conceptualization, R.P.G. and N.S.; methodology, R.P.G.; software, R.P.G.; validation, R.P.G. and N.S.; formal analysis, R.P.G.; investigation, R.P.G.; resources, R.P.G.; data curation, R.P.G.; writing—original draft preparation, R.P.G.; writing—review and editing, R.P.G. and N.S.; visualization, R.P.G. and N.S.; supervision, R.P.G.; project administration, R.P.G.; funding acquisition, R.P.G. All authors have read and agreed to the published version of the manuscript.
Funding: This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Data Availability Statement: The JWST member-galaxy catalogue and multiple-image catalogue used in this study are from Rihtaršić et al. [10] and will be publicly released on the CANUCS website(https://niriss.github.io/lensing.html). The pipeline reproducing the numerical results of this paper, together with the input catalogues (JWST member galaxies and 135 gold multiple images from Rihtaršić et al. 2026, and the CCC+TL angular-diameter distance table), is available at https://github.com/rgupta-uottawa/bullet-ccctl.
Acknowledgments: We thank Gregor Rihtaršić and co-authors for making the JWST member-galaxy catalogue and multiple-image catalogue available in advance of the journal publication. We are grateful to the reviewers for their critical comments that helped improve the quality, consistency, and clarity of the paper.
Conflicts of Interest: The authors declare no conflicts of interest.
Deflection pipeline and reproducibility
The lensing pipeline implementing Equations (23) and (24) is available as open-source Python3 code at https://github.com/rgupta-uottawa/bullet-ccctl. We document here its three essential numerical ingredients and the convergence tests supporting the quoted figures.
Convergence grid. The convergence map is built on a pixel grid covering on the sky, giving a pixel scale of . The map is identical in angular coordinates for both cosmologies; the physical scale is set per cosmology by the corresponding (see Table 1). Numerical convergence has been tested against a grid (/pixel); the aggregate source-plane scatter is stable to better than 1.5% on all three metrics (RMS, Mean, Median).
Deflection by DFT convolution. The deflection field is computed from Equation (23) by zero-padded two-dimensional DFT convolution of with the kernel . Zero-padding to twice the grid size suppresses wrap-around contamination. For an idealised singular isothermal sphere with , the pipeline recovers to better than 2% in all rings between and from the SIS centre.
Per-source rescaling. The convention of Equation (24) is derived from :
so the per-source factor can be computed from the angular-diameter distances alone and tends to unity at by construction.
Mass-uncertainty band. The band in Table 4 is constructed by rebuilding the NFW haloes , the gas model (normalisation rescaled to preserve inside the enlarged ), and the cluster-scale CCC turn-off radius consistently at and . The member-galaxy Hernquist (ΛCDM) and galaxy-scale CCC (CCC+TL) components depend only on the catalogue values and are therefore identical across the three scenarios.
References
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
References
- [1]Bartelmann, M.; Schneider, P. Weak gravitational lensing. Phys. Rep. 2001, 340, 291–472.
- [2]Cho, B. Y.; Jee, M. J.; Joo, H.; Cha, S.; HyeongHan, K. Joint JWST-DECam Lensing Reveals that the Bullet Cluster Is a Minor Merger. arXiv 2025, arXiv:2512.03150.arxiv.org/abs/2512.03150
- [3]Narayan R., Bartelmann M., 1996, arXiv, astro-ph/9606001. doi:10.48550/arXiv.astro-ph/9606001DOI
- [4]Umetsu, K. luster-galaxy weak lensing. Astron. Astrophys. Rev. 2020, 28, 7.
- [5]Bradač, M.; Clowe, D.; Gonzalez, A.H.; Marshall, P.; Forman, W.; Jones, C.; Markevitch, M.; Randall, S.; Schrabbäck, T.; Zaritsky, D. Strong and Weak Lensing United. III. Measuring the Mass Distribution of the Merging Galaxy Cluster 1ES 0657-558. Astrophys. J. 2006, 652, 937–947.DOI
- [6]Clowe, D.; Gonzalez, A.; Markevitch, M. Weak-Lensing Mass Reconstruction of the Interacting Cluster 1E 0657-558: Direct Evidence for the Existence of Dark Matter. Astrophys. J. 2004, 604, 596–603.DOI
- [7]Clowe, D.; Bradač, M.; Gonzalez, A.H.; Markevitch, M.; Randall, S.W.; Jones, C.; Zaritsky, D. A Direct Empirical Proof of the Existence of Dark Matter. Astrophys. J. 2006, 648, L109–L113.DOI
- [8]Markevitch, M.; Gonzalez, A.H.; David, L.; Vikhlinin, A.; Murray, S.; Forman, W.; Jones, C.; Tucker, W. A Textbook Example of a Bow Shock in the Merging Galaxy Cluster 1E 0657-56. Astrophys. J. 2002, 567, L27–L31.DOI
- [9]Markevitch M. Chandra Observation of the Most Interesting Cluster in the Universe. In Proceedings of the X-Ray Universe 2005 (ESA SP-604); Wilson A., Ed.; European Space Agency: Noordwijk, The Netherlands, 2006; p. 723.DOI
- [10]Rihtaršič, G.; Bradacxc, M.; Desprez, G.; Harshan, A.; Martis, N.S.; Willott, C.J.; Asada, Y.; Sarrouh, G.T.E.; Cornil-Baiotto, C.; Biviano, A.; et al. Mapping dark matter in the Bullet Cluster using JWST imaging and spectroscopy. arXiv 2026, arXiv:2601.22245.DOI
- [11]Niemiec, A.; Jauzac, M.; Jullo, E.; Limousin, M.; Sharon, K.; Kneib, J.-P.; Natarajan, P.; Richard, J. hybrid-LENSTOOL: A self-consistent algorithm to model galaxy clusters with strong- and weak-lensing simultaneously. Mon. Not. R. Astron. Soc. 2020, 493, 3331–3340.DOI
- [12]Gupta, R.P. JWST early Universe observations and ΛCDM cosmology. Mon. Not. R. Astron. Soc. 2023, 524, 3385–3395.DOI
- [13]Gupta, R.P. On Dark Matter and Dark Energy in CCC+TL Cosmology. Universe 2024, 10, 266.DOI
- [14]Gupta, R.P. Testing CCC+TL Cosmology with Observed Baryon Acoustic Oscillation Features. Astrophys. J. 2024, 964, 55.DOI
- [15]Gupta, R.P. Testing CCC+TL Cosmology with Galaxy Rotation Curves. Galaxies 2025, 13, 108.DOI
- [16]Gupta, R.P. The Origin of Dark Matter and Dark Energy: Covarying Coupling Constants? Symmetry 2026, 18, 300.DOI
- [17]Hernquist, L. An Analytical Model for Spherical Galaxies and Bulges. Astrophys. J. 1990, 356, 359.DOI
- [18]Paraficz, D.; Kneib, J.-P.; Richard, J.; Morandi, A.; Limousin, M.; Jullo, E.; Suyu, S.H. The Bullet cluster at its best: Weighing stars, gas, and dark matter. Astron. Astrophys. 2016, 594, A121.DOI
- [19]Brownstein, J.R.; Moffat, J.W. The Bullet Cluster 1E0657-558 evidence shows modified gravity in the absence of dark matter. Mon. Not. R. Astron. Soc. 2007, 382, 29–47.DOI
- [20]Navarro, J.F.; Frenk, C.S.; White, S.D.M. A Universal Density Profile from Hierarchical Clustering. Astrophys. J. 1997, 490, 493–508.
- [21]Springel, V.; Farrar, G.R. The speed of the ‘bullet’ in the merging galaxy cluster 1E0657–56. Mon. Not. R. Astron. Soc. 2007, 380, 911–925.DOI
- [22]Randall, S.W.; Markevitch, M.; Clowe, D.; Gonzalez, A.H.; Bradač, M. Constraints on the Self-Interaction Cross Section of Dark Matter from Numerical Simulations of the Merging Galaxy Cluster 1E 0657-56. Astrophys. J. 2008, 679, 1173–1180.DOI
- [23]Hernandez, X. A detailed MOND modelling of the Bullet Cluster. arXiv 2026, arXiv:2604.10811.arxiv.org/abs/2604.10811
- [24]Famaey, B. On the residual missing mass of the Bullet Cluster. arXiv 2026, arXiv:2605.10022.arxiv.org/abs/2605.10022




