Joint JWST-DECam Lensing Reveals That the Bullet Cluster Is a Minor Merger

Abstract

We present the first robust virial masses of the Bullet Cluster's three individual components from a joint weak+strong lensing analysis combining JWST/NIRCam and DECam observations. Despite its status as the benchmark system for dark matter and merger studies, inferred mass ratios for the Bullet Cluster have spanned a wide range from $\sim$2:1 to $\gtrsim$10:1 over more than two decades. We revisit this tension through three key advances: (1) JWST's exceptional data quality enables us to resolve three distinct halos, (2) DECam's wide-field coverage beyond its virial radius eliminates the need for extrapolation, and (3) high-fidelity strong-lensing priors mitigate weak-lensing model bias. We obtain $M_{200c} = 15.11^{+2.48}_{-2.10} \times 10^{14}M_{\odot}$ for the main cluster and $1.49^{+0.32}_{-0.25} \times 10^{14}M_{\odot}$ for the subcluster, yielding a mass ratio of $10.14^{+3.22}_{-2.47}$, definitively classifying the Bullet Cluster as a minor merger. This result reconciles the long-standing tension in the mass ratio and provides updated initial parameters for future modeling of this iconic system.

Figures (12)

• Figure 1: Wide-field view of the Bullet Cluster region. The background shows our DECam $g$+$r$+$i$ color composite, covering the $1^{\circ} \times 1^{\circ}$ ($\hbox{$\sim$} 16\text{ Mpc} \times 16\text{ Mpc}$ at $z=0.296$) region centered on the target. The red dashed rectangle marks the JWST/NIRCam F200W coverage ($\hbox{$\sim$}6\hbox{$^\prime$} \times 2.5\hbox{$^\prime$}$). The green circle marks the total system's virial radius ($R_{200c}^{\text{Total}}$ = 2.27 Mpc; §\ref{['sec:bestfit']}), while the blue circle indicates the boundary of the DECam WL analysis region (7 Mpc radius) used for mass estimation.
• Figure 2: PSF correction quality for JWST F200W (left) and DECam $i$-band (right). Blue points show observed stellar ellipticities while red points show the ellipticity residuals after PSF correction (i.e., observed minus modeled ellipticity components). Residuals centered at (0,0) with small scatter demonstrate accurate and precise PSF modeling. The lower-left inset in each panel displays PSF size residuals.
• Figure 3: PSF model diagnostics for DECam $i$-band. Left: $D_1$ and $D_2$ statistics Rowe_2010 showing auto-correlation of PSF model residuals ($D_1$; red circles) and cross-correlation between observed stellar ellipticities and residuals ($D_2$; black crosses). The $D_1$ ($D_2$) statistic remains below $10^{-6}$ ($10^{-5})$ across all angular scales, demonstrating exceptional PSF modeling accuracy. Right: Star-galaxy correlation functions following Jee_2013, measuring the normalized correlation $\xi_{\rm sys} = \langle e^* \gamma \rangle^2 / \langle e^* e^* \rangle$ decomposed into tangential ($\xi_{tt}$; red circles) and cross ($\xi_{\times\times}$; black crosses) components. All correlation amplitudes remain below $10^{-5}$; notably, $\xi_{tt}$ and $\xi_{\times\times}$ fall below $10^{-6}$ once the separation exceeds $3'$. These exceptionally low PSF-diagnostic levels demonstrate that any PSF-induced systematics are negligible in comparison to the WL signal.
• Figure 4: Source selection for JWST (upper) and DECam (lower) galaxies. Blue, red, and yellow points represent final WL sources, photometric member candidates, and spectroscopic members, respectively. For JWST, background source selection is based on photometric redshifts and a magnitude cut (F200W $>22$). For DECam, we determine the red sequence from a linear fit to the color-magnitude relation obtained from spectroscopic members (§\ref{['sec:specdata']}), adopting a color ($g-r$) width of $\pm$0.1 mag. We selected sources fainter than $i>21.5$, $\hbox{$\sim$}6$ mag below the brightest cluster members. To minimize potential contamination from faint red-sequence galaxies, we excluded objects lying within $\pm$0.1 mag of the red-sequence color. Foreground contamination in this selection was estimated using control fields and incorporated into the quantitative analysis of the shear signal (§\ref{['sec_decam_sources']}).
• Figure 5: Spectroscopic and photometric redshift comparison for JWST sources and galaxy number density validation for DECam. Top: Spectroscopic versus photometric redshifts for 103 JWST sources with spectroscopic data compiled from Foex_2017, Puccetti_2020, and Richard_2021 (§\ref{['sec:specdata']}). The solid line indicates perfect agreement, while dashed lines mark 10% deviation. Of the matched sources, 86 (83.5%) show photometric redshifts within 10% of spectroscopic values. The red vertical dashed line indicates the Bullet Cluster redshift ($z = 0.296$). Bottom: Galaxy number density comparison between the sources in the Bullet Cluster field and DESY3GOLD control field as a function of magnitude for deriving lensing efficiency $\beta$. The excellent agreement between the two fields at the bright end ($\lesssim22.3$) indicates negligible member contamination in our source selection, while offsets at fainter magnitudes are due to differences in depth.