A model-based sulphur count
A new model-based analysis estimates more sulphur in active galactic nuclei than the comparison method does. Using HCm version 6.2, the researchers obtained sulphur abundances about 0.3 dex higher than determinations that combine direct measurements with an ionisation-correction factor. A dex is a logarithmic abundance unit. An ionisation-correction factor is intended to account for ionisation stages not directly represented in the observed lines.
HCm compares optical emission-line ratios after reddening correction, together with their measurement errors, against large photoionisation-model grids. In ordinary terms, it searches sets of computer models of ionised gas for patterns that match the spectrum. The method is designed to estimate sulphur abundance without inserting an empirical correction for unobserved ionisation stages, which are central to the authors’ explanation of the higher result.
To set the configuration used later, the researchers compared different model grids. The comparison showed a clear minimum at αox = −1.2 and a 98% free-electron stopping criterion, the setting used for the subsequent MaNGA analysis.
The first check looks promising
The first check used the main D23 control sample of 33 nearby objects, plus 12 additional Seyfert 2 objects. At the preferred setting, HCm’s ionic-equivalent abundances—estimates for the individual ionisation stages represented in the lines—differed from direct-method measurements by an average of −0.03 dex for oxygen and +0.03 dex for sulphur. The authors describe that as essentially no systematic offset at the ionic level.
That close match does not settle the total-abundance question. HCm total oxygen values had a small mean offset from D23, but the differences were not uniform across the full abundance range. For sulphur, the HCm solutions were approximately 0.3 dex above the direct-method-plus-ICF determinations. D23 total oxygen, and especially total sulphur, depend on the ICF prescriptions used, so those values are not independently known total-abundance references.
The authors interpret the sulphur excess as a warning that classical H II-region ionisation-correction factors may underestimate higher-ionisation sulphur stages in AGN. In plain language, a correction developed for H II-region spectra may miss some forms of sulphur around an active nucleus. The study presents this as an interpretation of the comparison, not a direct measurement of all sulphur stages.
Different galaxy regions, shared trend
The preferred setup was then applied to 6,225 MaNGA AGN spectra. The sample contained 5,359 Seyfert 2 spectra and 866 LINER-like spectra, giving the analysis a broad comparison between two classes of AGN emission region.
Across the MaNGA regions, Seyferts and LINER-like objects preserved a common positive relationship between S/O, the sulphur-to-oxygen ratio, and metallicity. At high oxygen abundance, Seyferts tended to have a higher median S/O than star-forming regions, while LINER-like regions were intermediate. The pattern describes an association in the observed spectra; it does not show that AGN class differences caused the abundance differences.
When sulphur-to-hydrogen abundance, S/H, replaced oxygen abundance as the metallicity tracer, the increasing S/O sequences remained but the differences from star-forming regions became weaker. Seyferts still tended to sit along the upper envelope. The authors say this supports the possibility that sulphur is a more robust tracer of total gas-phase metallicity, while noting that ionisation, depletion, geometry and aperture effects could also contribute to the remaining separation.
The LINER-like result also shifted with the assumed ionising spectrum, or spectral energy distribution. Softer AGN continua produced larger S/O values, whereas pAGB models produced narrower distributions near the solar ratio. The result leaves the physical meaning of the LINER pattern open: different assumptions about the ionising source produce different sulphur-to-oxygen distributions.
What the comparison cannot settle
That model dependence is the main reason to treat the total abundances cautiously. The control comparison supports close agreement for ionic oxygen and sulphur, but it does not establish that HCm’s total S/H scale is the true one, because the comparison reference is itself ICF-dependent. Nor does the small mean oxygen offset establish one-to-one agreement across the abundance range.
The document remains an arXiv preprint: its front matter identifies it as version 1, dated 26 Aug 2026. Its immediate contribution is a model-based route for reading sulphur from AGN optical spectra, together with evidence that both the correction scheme and the assumed ionising spectrum can shape the answer.
Paper data and sources
Original title: HII-CHI-mistry for active galactic nuclei: optical sulphur abundances without empirical ionisation correction factors
Authors: E. Pérez-Montero, O. L. Dors, I. A. Zinchenko et al.
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-26
DOI: Not available
Original paper · Full text