Superhot (> 30 MK) flare observations with STIX: Joint spectral fitting

Spectroscopic analysis of large flares (>X1) in the hard X-ray (HXR) range offers unique insights into the hottest (> 30 MK) flare plasma, the so-called superhot thermal component. To manage the high count rates in large flares, an attenuator is typically placed in front of the HXR detectors. However, this significantly limits the spectral diagnostic capabilities at lower energies, and consequently, it restricts the analysis of the lower temperatures in flares. The Spectrometer/Telescope for Imaging X-rays (STIX) on board the Solar Orbiter mission was designed to observe solar flares in hard X-rays. The imaging detectors use an attenuator during periods of high flux level. In contrast, the background (BKG) detector of STIX is never covered by the attenuator and is therefore dedicated to measure the unattenuated flux using differently sized apertures placed in front of the detector. We aim to demonstrate that joint spectral fitting using different detector configurations of STIX allows us to reliably diagnose both the hot and the superhot components in large flares. We jointly fit the HXR spectra of the STIX BKG detector and the STIX imaging detectors using SUNKIT-SPEX software package to determine the spectral parameters of both the hot and superhot thermal components in solar flares. Using joint fitting on 32 STIX flares, we corroborate that for GOES X-class flares, the HXR spectrum is better represented by two thermal components instead of an isothermal component. At the temperature peak time, the superhot HXR flux above 15 keV is typically stronger than the hot HXR flux. The GOES long-wavelength channel is dominated by the hot component with a superhot contribution up to 10%. This paper demonstrates that joint spectral fitting of the same detector type with different attenuation schemes is a simple and powerful method to monitor multithermal flare plasma.

💡 Research Summary

This paper presents a novel approach to diagnosing the multi‑thermal plasma of large solar flares, especially the super‑hot component (>30 MK), by jointly fitting the hard X‑ray (HXR) spectra obtained from two different detector configurations of the Spectrometer/Telescope for Imaging X‑rays (STIX) aboard Solar Orbiter. STIX’s imaging detectors employ an automatic attenuator during high‑flux events, which blocks low‑energy photons (4–12 keV) and therefore hampers the measurement of the cooler (≈10–30 MK) plasma. In contrast, the dedicated background (BKG) detector is never covered by the attenuator and records the unattenuated flux through three apertures, providing reliable data down to ~6 keV.

The authors exploit this complementary capability by simultaneously fitting the BKG spectrum (sensitive to the hot component) and the imaging‑detector spectrum (dominated by the super‑hot and non‑thermal components). They use the SUNKIT‑SPEX package, a modern SunPy‑based spectral fitting tool that supports Bayesian inference via Markov Chain Monte Carlo (MCMC). The photon model consists of an isothermal plasma (f_vth), a thick‑target non‑thermal bremsstrahlung component (thick_fn), and an optional albedo correction. Crucially, they test both a single‑thermal model and a two‑thermal model (hot + super‑hot) while allowing a “binding” parameter C to scale the imaging‑detector model, thereby accounting for systematic calibration differences between the two detector sets.

Applying this joint fitting method to 32 GOES X‑class flares observed between 2022 and 2025, the study finds that a two‑thermal model provides a statistically superior fit compared with a single‑thermal model. The reduced χ² values improve, and residuals become more random, especially around the 15–25 keV transition region where the thermal and non‑thermal components intersect. The super‑hot component typically dominates the HXR flux above ~15 keV, often exceeding the hot component by a factor of 1.2–1.5 at the temperature peak. While the GOES long‑wavelength channel (1–8 Å) is largely governed by the hot plasma (≈90 % of the total flux), the super‑hot plasma contributes up to 10 % of the GOES signal, confirming that GOES temperature estimates are biased toward the cooler component.

Methodologically, SUNKIT‑SPEX offers clear advantages over the traditional OSPEX workflow. Its ability to fit multiple spectra simultaneously reduces the need for iterative procedures, and the Bayesian framework yields robust posterior distributions and credible intervals for all parameters. The inclusion of the binding parameter C effectively mitigates systematic differences in detector response, a refinement over the earlier iterative approach (Stiefel et al. 2025).

The authors argue that this joint‑fitting strategy not only improves the reliability of thermal diagnostics in attenuated flares but also provides a template for future HXR missions. By demonstrating that a single instrument, equipped with both attenuated and unattenuated detector channels, can recover the full temperature distribution of large flares, the work suggests that future spacecraft could simplify payload designs while retaining full diagnostic capability.

In summary, the paper establishes that joint spectral fitting of STIX’s BKG and imaging detectors is a simple yet powerful technique for monitoring multi‑thermal flare plasma. It confirms the prevalence of a super‑hot (>30 MK) component in X‑class flares, quantifies its contribution to both HXR and GOES fluxes, and showcases the benefits of Bayesian multi‑spectrum fitting for high‑energy solar physics.