Title: Weak-lensing mass bias in merging galaxy clusters

Abstract: Although weak lensing (WL) is a powerful method to estimate a galaxy cluster mass without any dynamical assumptions, a model bias can arise when the cluster density profile departs from the assumed model profile. In a merging system, the bias is expected to become most severe because the constituent halos undergo significant structural changes. In this study, we investigate WL mass bias in binary cluster mergers using a suite of idealized hydrodynamical simulations. Realistic WL shear catalogs are generated by matching the source galaxy properties, such as intrinsic shape dispersion, measurement noise, source densities, etc., to those from Subaru and {\it Hubble Space Telescope} observations. We find that, with the typical mass-concentration ($M$-$c$) relation and the Navarro-Frenk-White (NFW) profile, the halo mass bias depends on the time since the first pericenter passage and increases with the mass of the companion cluster. The time evolution of the mass bias is similar to that of the concentration, indicating that, to first order, the mass bias is modulated by the concentration change. For a collision between two $\sim10^{15}~M_{\odot}$ clusters, the maximum bias amounts to $\sim60\%$. This suggests that previous WL studies may have significantly overestimated the mass of the clusters in some of the most massive mergers. Finally, we apply our results to three merger cases: Abell 2034, MACS J1752.0+4440, and ZwCl 1856.8+6616, and report their mass biases at the observed epoch, as well as their pre-merger masses, utilizing their merger shock locations as tracers of the merger phases.