A carbon-rich atmosphere on a windy pulsar planet

TL;DR

This study presents JWST spectroscopy of PSR J2322-2650b, a Jupiter-mass pulsar companion, revealing a carbon-rich, hydrogen-depleted atmosphere dominated by C$_2$ and C$_3$ with extreme C/O and C/N ratios. Through a combination of low- and high-resolution spectra, radial-velocity measurements, and atmospheric modeling (HELIOS and 3D GCMs), the authors infer a helium-dominated bulk composition and detect a westward wind pattern, consistent with rapid rotation and deep gamma-ray heating. The findings challenge conventional black widow formation scenarios, requiring novel enrichment pathways, and showcase JWST’s ability to probe exotic exoplanetary atmospheres in pulsar systems with potential implications for neutron-star equation-of-state constraints. Overall, the work opens a new regime in exoplanet chemistry and dynamics under extreme irradiation and motivates further observations of Tidarren-like systems.

Abstract

A handful of enigmatic Jupiter-mass objects have been discovered orbiting pulsars. One such object, PSR\,J2322-2650b, uniquely resembles a hot Jupiter exoplanet due to its minimum density of 1.8 g/cm^3 and its ~1900 K equilibrium temperature. We use JWST to observe PSR J2322-2650b's emission spectrum across an entire orbit. In stark contrast to every known exoplanet orbiting a main-sequence star, we find an atmosphere rich in molecular carbon (C3, C2) with strong westward winds. Our observations open up new exoplanetary chemical (ultra-high C/O and C/N ratios of >100 and >10,000, respectively) and dynamical regimes (ultra-fast rotation with external irradiation) to observational study. The extreme carbon enrichment poses a severe challenge to the current understanding of ``black widow'' companions, which were expected to consist of a wider range of elements due to their origins as stripped stellar cores.

Figures (11)

• Figure 1: PSR J2322-2650 compared to the black widows from SpiderCat koljonen_2025 and to PSR J1719-1438 bailes_2011, showing it is in a distinct area of parameter space. PSR J1719-1438, though similar in mass, has a much higher mean density of 21 g/cm$^3$; its bulk composition is likely very different. The temperature of all companions was very roughly estimated by assuming the spindown luminosity is isotropically radiated and all of the energy hitting the companion is converted to heat. The minimum density was derived from the orbital period via $\rho_{\rm min} = 3\pi/(0.46^3 G P^2)$, combining Paczyński's approximation for the Roche lobe ($R_L = 0.46^3 q^{1/3} a$; paczynski_1971) with Kepler's third law.
• Figure 2: The observed emission flux and brightness temperature of PSR J2322-2650b. a, observed PRISM spectrum of the dayside, the average of the two quadratures, and nightside at native resolution. The time-averaged G235H spectrum is also plotted, binned in wavelength by a factor of 16. The two shaded wavelength regions are those of the light curves in Fig. \ref{['fig:wlc']}. b, PRISM spectra expressed in brightness temperatures by assuming R/D = (1.1 $R_J$) / (630 pc). We plot HELIOS radiative-convective equilibrium forward models in grey. A linear function of the log of C$_3$'s absorptions cross sections is plotted in green, showing that C$_3$ absorption explains the sudden dip at 3.014 um, the recovery at 4 um, and the slow decline toward the red edge.
• Figure 3: C$_2$ detection significance and radial velocities obtained from the JWST NIRSpec/G235H data. (a) The detection significance from cross correlation of C$_2$ in the G235H data, as a function of projected orbital velocity and barycentric radial velocity. C$_2$ is detected at 21$\sigma$ at V$_{\rm sys} \approx 0$ and $K_{\rm comp}=190$ km/s. (b) In black, the companion radial velocity measured from each integration of the G235H observations using a data-derived template; in yellow, a sinusoidal fit to the data. The vertical offset is arbitrary. In the colored bars, we show the phases we define to be the nightside, quadrature, and dayside for the purposes of calculating the average spectra in Fig. \ref{['fig:spectra']}.
• Figure 4: The relative abundance of C$_2$ to CO in chemical equilibrium. (a) Abundance ratios for C/He = $10^{-2}$ and P=10 mbar, for an atmosphere with no other elements. (b) Ratios in an atmosphere with H/He=0.01 as well as solar values of N/He, P/He, S/He. The solid white curves represent the 3$\sigma$ constraint from our cross correlation test (C$_2$/CO $>$ 0.17).
• Figure 5: The companion light curve in two continuum bands interpreted with a pulsar direct heating model. (a) The models are shown by the curves and the data are shown by the points. The model does not capture the strong orbit-to-orbit variations at minimum ($\phi \sim 0$ and $\phi \sim 1$), but provides a good description of the flux from the heated face and constrains the binary parameters (see Appendix \ref{['sec:prism_phase_curve']}). Schematics of the Earth-view of the heated companion's surface temperature are shown below the light curve. (b),(c) The residuals of the models compared to their corresponding light curves.