CMB Birefringence from Axion String Networks Calibrated to an AMR Simulation
📄 arXiv:2608.24456 · 📥 PDF · 2026-08-25 · astro-ph.CO
Authors: Mustafa A. Amin [arXiv · scholar] , Mudit Jain [arXiv · scholar] , Andrew J. Long [arXiv · scholar] , Aden Pugsley [arXiv · scholar] , Moira Venegas [arXiv · scholar] , Magdalena Whelley [arXiv · scholar]
📄 Abstract
A cosmological network of axion strings may exist in the Universe today. If axion-like particles couple to electromagnetism, such a network induces spatially varying birefringence in the polarization of the cosmic microwave background (CMB), which can be probed by current and next-generation CMB experiments. We calibrate a loop-crossing model against a large-scale adaptive-mesh-refinement (AMR) simulation of axion-string network dynamics in the early Universe and use the calibrated model to predict CMB birefringence from recombination to today. We find that the non-detection of anisotropic birefringence in CMB observations places a strong upper bound on the electromagnetic anomaly coefficient $\mathcal{A}$ that enters the axion-photon coupling $g_{aγγ} = - \mathcal{A} α_\mathrm{em} / πf_a$. A joint analysis of available anisotropic birefringence measurements constrains $|\mathcal{A}| < 0.24$ at 95% C.L., which is independent of the Peccei-Quinn scale $f_a$, assuming that the axions are hyperlight so that the network survives until today. This limit strongly restricts the high-energy embedding of hyperlight axions, excluding the minimal Grand Unified Theory prediction for the electromagnetic anomaly coefficient at high significance. In addition, we discuss the implications of an axion-string origin for the recently reported evidence of isotropic birefringence.