Revealing tensions in neutron star observations with pressure anisotropy

Abstract

Pressure isotropy, i.e., equality between radial and tangential pressure, is often assumed when studying neutron stars. However, mechanisms such as pion/kaon condensation, magnetic fields, and dark matter clustering can lead to pressure anisotropy. This work presents a comprehensive measurement of pressure anisotropy in neutron stars. Our analysis incorporates an extensive set of nuclear experimental constraints and multi-messenger astrophysical observations. We find that the Bayes factor for anisotropy against isotropy $\gtrsim 3 : 1$, when the anisotropy is allowed to vary between individual stars. The posterior indicates a population-wide preference for negative anisotropy, primarily driven by PSR J0740+6620. Due to the lack of radius measurements for $2M_\odot$ neutron stars, we cannot rule out density-scale-dependent anisotropy. Therefore, both phase transitions and density-scale-independent mechanisms, such as magnetic fields, dark matter clustering, or deviations from general relativity are viable explanations. While the evidence for anisotropy remains inconclusive, these results demonstrate that pressure anisotropy can be utilized as a tool for identifying missing physics in neutron star modeling or revealing novel physics in the era of multi-messenger astronomy.

Figures (4)

• Figure 1: Illustration of the impact of pressure anisotropy on neutron star structure by varying the level of anisotropy. A positive anisotropy, i.e., the radial pressure exceeds the tangential pressure, leads to smaller yet stiffer neutron stars. In contrast, a negative anisotropy tends to produce larger but softer stars. The maximum a posteriori equation of state from Ref. Huth:2021bsp is used for the isotropic baseline.
• Figure 2: Posterior on the nuclear empirical parameters conditioned on the experimental constraints tabulated in Tab. \ref{['tab:nuclear_constraints']}. The darker and lighter shading are indicating the $68\%$ and $95\%$ credible intervals, respectively
• Figure 3: Posterior of $\mu_{\gamma}$ and $\sigma_{\gamma}$ for the two equation-of-state models considered. The posteriors are highly consistent and indicate a general negative anisotropy. The values quoted for $\mu_\gamma$ are the maximum a posteriori (dot-dashed line), together with the $95\%$ credible interval (dashed line)as the uncertainty. The 2D contours are shown at $68\%$ (darker shade) and $95\%$ (lighter shade) levels.
• Figure 4: The posteriors of $\gamma$ for each individual observation, using the Metamodel + peak equation-of-state model are shown. The baseline, shown in black, is the predictive posterior on $\gamma$ based on the posterior of $\mu_\gamma$ and $\sigma_\gamma$ as shown in Fig. \ref{['fig:mu_gamma_posterior']}, which uses all the constraints. The posteriors of individual events are relatively uninformative across all observations, except for PSR J0740+6620 (shown in red), which has a stronger preference for negative anisotropy compared to others.