Chlorine and Potassium Enrichment in the Cassiopeia A Supernova Remnant

TL;DR

The study addresses the long‑standing challenge of understanding odd‑Z nucleosynthesis by directly measuring P, Cl, and K in the Cassiopeia A remnant with high‑resolution X‑ray spectroscopy from XRISM. Using spatially resolved spectroscopy and two‑component NEI plasma modeling, the authors detect Cl and K at near‑solar abundance levels in O‑rich ejecta, with P also present but less precisely constrained. The observed Cl/Ar and K/Ar ratios favor stellar processes such as shell mergers or binary interactions over standard single‑star SN yields, implying that pre‑SN stellar activity significantly boosts odd‑Z production. These results provide a crucial empirical benchmark, linking stellar evolution, explosive nucleosynthesis, and galactic chemical evolution, and motivating refined theoretical models that incorporate complex pre‑SN interiors.

Abstract

The elements in the universe are synthesized primarily in stars and supernovae, where nuclear fusion favors the production of even-Z elements. In contrast, odd-Z elements are less abundant and their yields are highly dependent on detailed stellar physics, making theoretical predictions of their cosmic abundance uncertain. In particular, the origin of odd-Z elements such as phosphorus (P), chlorine (Cl), and potassium (K), which are important for planet formation and life, is poorly understood. While the abundances of these elements in Milky Way stars are close to solar values, supernova explosion models systematically underestimate their production by up to an order of magnitude, indicating that key mechanisms for odd-Z nucleosynthesis are currently missing from theoretical models. Here we report the observation of P, Cl, and K in the supernova remnant Cassiopeia A using high-resolution X-ray spectroscopy with XRISM, with the detection of K at above the 6$σ$ level being the most significant finding. Supernova explosion models of normal massive stars cannot explain the element abundance pattern, especially the high abundances of Cl and K, while models that include stellar rotation, binary interactions or shell mergers agree closely with the observations. Our observations suggest that such stellar activity plays a significant role in supplying these elements to the universe.

Figures (9)

• Figure 1: Three-color image of the core-collapse SNR Cas A. Red, green, and blue represent Chandra X-ray images of O- (0.6 to 0.85 keV), Si- (1.76 to 1.94 keV), and Fe-enhanced (6.54 to 6.92 keV) regions, respectively. The white grids indicate the fields of view of our two-pointing observations with XRISM Resolve, which has a $6\times6$ pixel array: a pixel at a corner is only used for calibration. Pixels we selected for our spectral analysis are highlighted in yellow color. In the southeast (SE) and north (N) regions, we examined the spectrum of each pixel and selected those exhibiting large equivalent widths of K emission; the spatial distribution of these pixels closely matches that of the O and Si emission lines. In contrast, the west (W) region exhibits weak O line intensity, and the K emission line in its spectrum is faint (see Figure \ref{['fig:CasA_spec']}).
• Figure 1: Elemental mass ratios in Cas A compared to supernova nucleosynthesis models across a wider range of model parameters. Top: Comparison with 15$M_\odot$ models with the several initial metalicities ([Fe/H]=$-2, -1, 0$) and rotation velocities ($v=0, 150, 300$ km s$^{-1}$) provided by Limongi & Chieffi limongi2018. Middle: Comparison with 13, 14, 15, 16$M_\odot$ models of the single star cases and the binary cases provided by farmer et al. farmer2023. In each binary case, the companion star mass is set to 0.8 times of the progenitor. Bottom: Comparison with non-rotating single star models with several masses provided by Sukhbold et al. sukhbold2016. The 14.9$M_\odot$ and 19.3$M_\odot$ models shows onion-like layers constituted with O-, Ne-, and C-burning layers. In contrast, in the 19.8$M_\odot$ and 20.1$M_\odot$ models, the O-burning layers are merging with the outer layers like shell mergers (see Supplementary Figure 3). To facilitate comparison between the observed values and the model values, only the model values are connected with lines.
• Figure 1: The Resolve spectrum and the best-fit model of Region SE. The supectrum is shown as the black points and the total best-fit model of AtomDB is shown as the red line. The purple and blue dashed lines show the two-NEI components, respectively. The grey line shows the Gaussian component adopted to the 3.2 keV structure.
• Figure 2: Resolve X-ray spectra of Cas A. The best-fit models are overplotted as a solid line with colors representing the spectral regions in Figure \ref{['fig:CasA_image']} (orange: region N, red: region SE, green: region W). In the top panel, the labels indicate the positions of the He-$\alpha$ emission lines for each element. The vertical broken lines correspond to the centroid energies of the resonance lines of P ($\sim2.152$ keV), Cl ($\sim2.790$ keV), and K ($\sim3.511$ keV) in the rest frame. The bottom three panels show the zoom-in spectra around He-$\alpha$ P (left), Cl (middle), and K (right) line with the best fit models with or without the K emission line. In the spectra of the P band, models without P emission show residuals near 2.13 keV, particularly in the SE and W regions, indicating the signature of P emission in the spectra, while no significant residual is seen in the N region. In the Cl and K bands, models lacking the corresponding lines show residuals in the N and SE regions, but not in the W region. The intermediate-mass elements, particularly in region N, exhibit complex velocity structures (see Figure 4 in Suzuki et al. 2025 suzuki2025), which introduces modeling uncertainties. For clarity, the spectrum of region SE has been shifted slightly along the energy axis in the zoomed-in panels.
• Figure 2: Comparison between observed and predicted Cl/S and K/Ar ratios across multiple stellar models. The hatched light blue regions represent the 1$\sigma$ confidence intervals for the observed mass ratios in region SE of Cas A. Sukhbold et al. sukhbold2016 models (12–30 $M_\odot$) are shown with different marker shapes for progenitor mass ranges (circle: 12–15 $M_\odot$, square: 15–20 $M_\odot$, star: 20–30 $M_\odot$). Shell merger candidates are highlighted in red; non-merger models are shown in black. Models from Limongi & Chieffi limongi2018 (13–25 $M_\odot$) are shown with squares and circles representing solar and sub-solar metallicity ([Fe/H] = $-1$ and 0), respectively. Rotation is indicated by color: purple for non-rotating and green for rotating models at 300 km s$^{-1}$. Roberti et al. roberti2024 models extend LC18 to super-solar metallicity ([Fe/H] = 0.3), and are plotted in a similar format (non-rotating and rotating cases). Farmer et al. farmer2023 models (10–30 $M_\odot$) are shown as stars: blue for single-star models and orange for binary models. Nomoto et al. nomoto2013 models (10–30 $M_\odot$) are indicated as grey crosses. The figure shows that models involving shell mergers or binary interactions can produce enhanced Cl and K abundances, consistent with our observational constraints.