Weak-Lensing Detection of Intercluster Filaments in Three Nearby Cluster Systems

TL;DR

Direct detection of intercluster filaments via weak lensing is challenging due to their weak deflection signal. The authors develop a matched-filter approach tailored to filament geometry and apply it to wide-field DECam data from LoVoCCS across three nearby clusters, achieving robust detections of intercluster bridges. Using MCMC, they infer filament properties with maximum convergence $\\kappa_0 \sim 0.015$–$0.053$ and widths $h_c \sim 0.11$–$0.45$ Mpc, and find orientations consistent with spectroscopic and red-sequence structures. These results support the cosmic web picture and show that current and upcoming wide-field weak-lensing surveys can directly map dark matter filaments around massive clusters.

Abstract

Direct detection of intercluster filaments is challenging due to their low surface density, resulting in a weak deflection field. We present weak-lensing detections of intercluster filaments using wide-field Dark Energy Camera (DECam) observations from the Local Volume Complete Cluster Survey (LoVoCCS). A matched-filter method was applied to identify filamentary structures in three nearby ($z < 0.1$) systems centered on Abell 401, Abell 2029, and Abell 3558. We discover two prominent filaments ($\geq 4σ$) in each system, with the strongest detections ($6.4σ- 7.3σ$) around Abell 401 and Abell 2029. In particular, we report the first robust weak-lensing detections $(>5 σ)$ of the intercluster bridges connecting the cluster pairs Abell 401/399, Abell 2029/2033, Abell 2029/SIG, and Abell 3558/3556. Adopting a filament convergence model motivated by numerical simulations, we infer the maximum convergence ($κ_0$) and characteristic width ($h_{\mathrm{c}}$) for all six filaments, yielding $κ_0 \sim 0.015 - 0.053$ and $h_{\mathrm{c}} \sim 0.11 - 0.45 \ \mathrm{Mpc}$. The performance of the matched-filter technique is validated using mock shear catalogs and further tested on a null field around Abell 2351. We also explore the potential of using the B-mode lensing signal of filaments to suppress cluster-induced shear contamination. These results demonstrate the feasibility of directly mapping dark matter filaments with current and future wide-field weak-lensing datasets.

Figures (15)

• Figure 1: Convergence field, $\kappa(\boldsymbol{x})$, and binned shear pattern, $\langle\gamma(\boldsymbol{x})\rangle_{\Delta x}$, for the mock catalog. Left: Convergence map of the mock model overlaid with the binned shear pattern, computed using a pixel size of $\Delta x = 1.53 \ \mathrm{arcmin}$. White circles indicate the radial cutoffs $(r_1 = 0.91 \ \mathrm{Mpc}$, $r_2 = 3.64 \ \mathrm{Mpc})$ used to restrict the filter. The legend in the bottom-left corner provides a mapping between the shear length in the plot to the physical shear value. Right: Same as left panel, but with shape noise ($\sigma_\gamma = 0.45$) added to the mock shear field. In both panels, convergence values are truncated at $\kappa < 0.2$ to emphasize filamentary structures.
• Figure 2: Schematic of the coordinate system and reference frames used in this work. The original frame (black) $\hat{x}_1$--$\hat{x}_2$ is centered on the primary cluster. The decomposition axis or rotated frame (green) $\hat{x}_{1'}$--$\hat{x}_{2'}$ subtends a CCW angle $\theta$ with the positive $\hat{x}_1$ axis. The filament (red) is oriented at angle $\theta_{\mathrm{f}}$. Any point in the field (blue) can be represented by its Cartesian angular position $\boldsymbol{x} = \{x_1, x_2\}$ or corresponding polar coordinates $(r, \phi)$. The perpendicular distance of the point from the filament axis is marked by $h = r |\sin(\phi - \theta_{\mathrm{f}})|$. The entire coordinate system is right-handed. The compass in the lower-right corner provides the cardinal directions in the image frame.
• Figure 3: Constructed matched filter and optimal matched filter in real space. Left: Matched filter modeled after a template filament with characteristic width $h_{\mathrm{c, filter}} = 0.15 \ \mathrm{Mpc}$ and normalization $\kappa_0 = 1$. We note that the choice of $\kappa_0$ does not affect the analysis. White dashed circles indicate the radial cutoffs $(r_1 = 0.91 \ \mathrm{Mpc}$, $r_2 = 3.64 \ \mathrm{Mpc})$, and the red arrow marks the filter orientation $(\theta = 135^\circ)$. Right: Corresponding filter optimized using $\hat{W}(k)$ with a cutoff frequency of $k_{\mathrm{cut}} = 0.21 \ \mathrm{arcmin}^{-1}$.
• Figure 4: Filament detection results for the mock catalog. Left: Matched-filter statistics $\Gamma_{\times}(\theta)$ and $\Gamma_{+}(\theta)$ as a function of the search angle $\theta$. The blue and red solid lines correspond to the tangential and cross components, respectively, with the light blue shade indicating the $1\sigma$ uncertainty in the tangential signal. The uncertainty in the cross signal is omitted for visual clarity. The dotted blue and red lines denote the corresponding statistics in the absence of noise. The violet and pink vertical dashed lines represent detections associated with filaments oriented at $\theta_{\mathrm{f_1}} = 10^{\circ}$ and $\theta_{\mathrm{f_2}} = 70^{\circ}$, respectively ; the green dashed line marks the orientation of the secondary cluster $(\phi = 130^{\circ})$. Right: Polar representation of the left panel
• Figure 5: Orthogonal decomposition of the negative cross gradient. The blue, red and green solid lines represent the tangential statistic, $A(\theta)$, negative cross gradient, $B(\theta)$, and the residual, $R(\theta)$. The corresponding shaded regions represent the $1\sigma$ uncertainty in each statistic.