The gamma-ray glow known as the Galactic Center ridge appears to bend away from a simple power-law pattern, according to an analysis of very-high-energy observations. The curvature is significant at more than three standard deviations, but the fit does not select one description: broken-power-law, logarithmic-parabola and exponential-cutoff models were similarly preferred across the 0.4–100 TeV range.
The analysis also places the preferred cosmic-ray injection site close to Sgr A*. Its best-fit position was −14.0 parsecs relative to the reference position, with statistical uncertainties of +4.2 and −9.7 parsecs. After an estimated ±7-parsec systematic uncertainty, the allowed interval ran from −19 to +1.2 parsecs around Sgr A*.
A detailed map of the central region
The study set out to build a multi-component description of the Galactic Center region and use it to estimate how cosmic rays are distributed in space and energy, as well as where they may originate. It used H.E.S.S. observations collected from March 2004 to June 2020. After restricting the observations to zenith angles below 40 degrees, the selected data represented about 365 hours over the central molecular zone.
The researchers fitted spatial information and reconstructed energy together in a three-dimensional analysis. The model included four previously reported point sources, a template for gamma rays produced when cosmic rays interact with material in the central molecular zone, large-scale Galactic emission and a residual hadronic background. Nuisance parameters were used to represent major instrumental and background uncertainties.
The systematic-error simulation procedure was repeated 500 times for each spectral-model shape and model component.
The bend is real in the fit, but its shape is not
A curved spectrum means that the relative gamma-ray intensity changes with energy in a way a single unbroken power law does not capture. The analysis found this departure from a simple power law to be statistically significant. Yet its three curved alternatives could not be distinguished from one another with the available ridge data. Depending on the chosen description, the corresponding transition in the parent-proton population was estimated at roughly 15–50 TeV for a break or 60–160 TeV for a cutoff.
That ambiguity matters because the inferred parent-proton transition depends on the chosen shape. The study therefore identifies a statistically significant bend in the gamma-ray data, but not a single transition scale or spectral formula.
A pattern consistent with a central source
The fitted cosmic-ray density fell with distance in a way close to the assumed inverse-distance form. Its radial index was α = 1.10, with a statistical uncertainty of ±0.05 and a systematic uncertainty of ±0.10. The authors interpret that profile as consistent with continuous injection near the Galactic Center.
The spectrum also showed no significant change across the central molecular zone. A fit with a constant spectral index gave Γ = 2.322 ± 0.002 and a p-value of 0.19, while a linear trend with distance was not statistically preferred, with a p-value of 0.61. In plain terms, the analysis did not find evidence that the fitted spectrum systematically hardened or softened across the zone.
Taken together, the inverse-distance-like density profile and the lack of a detected spectral trend are consistent with the study’s continuous-injection interpretation near the Galactic Center.
Nearby clusters are not favored as the main source
The analysis tested whether the Arches cluster could make a major additional contribution. Its best-fit amplitude was close to zero, and, when its spectrum was tied to that of the central region, its cosmic-ray luminosity contribution had a 24% upper limit at two standard deviations. The Quintuplet cluster was not separately tested, so the result does not show that either cluster contributes nothing.
The study also used the fitted emission to examine the possible line-of-sight placement of molecular clouds. Allowing a maximum 30% variation in gamma-ray flux produced the strongest qualitative constraints for MC20, MC50, Sgr C and the Dust Ridge, and challenged some positions proposed by dynamical models. These are qualitative line-of-sight constraints, not a direct set of exact cloud distances.
What the analysis leaves open
The result remains dependent on the model used to separate overlapping components. The analysis combines spatial and spectral information in three dimensions, including a cosmic-ray-interaction template, large-scale Galactic emission and residual hadronic background. The systematic-error procedure was repeated 500 times for each model shape and component, but the competing curved ridge spectra remained similarly preferred.
The available fit cannot distinguish the broken-power-law, logarithmic-parabola and exponential-cutoff options. It also leaves the parent-proton transition as a range rather than a single value, so the ridge’s spectral bend is a constrained feature of the model rather than a settled physical story.
The manuscript is a version-one preprint on arXiv, identified as arXiv:2608.25946v1. Its front matter reports receipt on 25 March 2026 and acceptance on 11 August 2026. The acknowledgements list support from national and institutional funders and H.E.S.S. operational partners, including BMBF, the Max Planck Society, the Helmholtz Association, CNRS, CEA, STFC, ARC and JSPS.
Paper data and sources
Original title: Evidence for a spectral steepening of the gamma-ray emission from the Galactic Center ridge
Authors: A. Acharyya, F. Aharonian, H. Ashkar et al.
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-26
DOI: Not available
Original paper · Full text