KMT-2025-BLG-1314 and KMT-2025-BLG-1392: two microlensing planetary/brown-dwarf candidates analyzed with differentiable code

Abstract

Analysis of binary-lens microlensing events typically requires intensive computation because of the multimodal and complex posterior distributions. With the recent development of the JAX-based differentiable binary-lensing modeling package microlux, we present an analysis of two microlensing events with planet/brown-dwarf candidates, KMT-2025-BLG-1314 and KMT-2025-BLG-1392. Both events exhibit the "Close/Wide" degeneracy, and KMT-2025-BLG-1314 suffers from the "Planet/Binary" degeneracy and a recently recognized "Point/Finite" degeneracy among the planetary solutions. For KMT-2025-BLG-1314, the binary mass ratio is $\log q \sim -3.5$ for the planetary solutions and $\log q > -1.5$ for the binary solutions, while for KMT-2025-BLG-1392, we find $\log q \sim -1.3$. We show that for the analysis of KMT-2025-BLG-1314, Hamiltonian Monte Carlo (HMC), enabled by microlux, provides robust parameter inference and outperforms traditional Markov chain Monte Carlo (MCMC) methods in the presence of bimodal posteriors.

Figures (7)

• Figure 1: Light curve and the best-fit model of the microlensing event KMT-2025-BLG-1314. Different data sets are shown in different colors. The "Close Binary BC" solution has the lowest $\chi^2$. The bottom panel shows the residual distribution for four different solutions relative to the "Close Binary BC" solution. Different data sets are aligned to the KMTC11 $I$-band data.
• Figure 2: The $\chi^2$ distribution projected onto the four planes of $(\log s, \log q, \alpha, \log w)$ space from the grid search result of KMT-2025-BLG-1314. The grid points with $\Delta \chi^2 > 49$ are left blank. The different degenerate solutions are marked with their names. The gray dashed lines in the $(\log s, \log w)$ and $(\log s, \log q)$ panels represent the envelopes with constant value of $\log q$ and $\log w$, respectively. The red and blue dashed lines mark the boundaries between resonant and close/wide caustics geometries Dominik1999. In total, ten local mimina are identified.
• Figure 3: Trajectory configuration of 10 solutions for KMT-2025-BLG-1314. The red curves denote the caustic geometries, the blue dots represent the position of the primary lens, and the purple circles indicate the source radii. The black lines show the source-lens relative trajectory with the arrow indicating the source motion direction. The coordinate origin is set at the magnification center. The 2L1S lensing parameters are shown in Table \ref{['tab:1314_2L1S_parameters']}.
• Figure 4: Light curve and the best-fit model of the microlensing event KMT-2025-BLG-1392. The bottom panel shows the residual distribution for the 2L1S close/wide solutions and the 1L2S solution. Note that the multi-site data in the top panel are aligned to the KMTA scale using the source and blend fluxes ($f_s, f_b$) of the"Close" model. Because the 1L2S model has a significantly different $\chi^2$ and requires different flux parameters, its model curve shows an apparent visual offset from the aligned data. The residuals in the bottom panels, however, are computed independently using each model's respective best-fit flux parameters.
• Figure 5: Trajectory configuration of the event KMT-2025-BLG-1392.