VIS
8.934 ± 0.311[O I] 6300 Å forbidden line (blended with Ni I) · 1D LTE · HARPS · Codex Grade
O · products, publication readiness and reproducible evidence
Solar oxygen: 8.755 ± 0.040
Asplund, Amarsi & Grevesse 2021: 8.69 ± 0.04 · this measurement +0.065 dex
Combined the way Asplund, Amarsi & Grevesse (2021) combine it: weighted mean and error of the indicator families [O I], O I, OH (Asplund+2021 Sect. 2.3); 1D results moved to the 3D scale by Asplund+2021 Table 3 (3D minus 1D MARCS for the same indicator). Each indicator is represented by our best analysis of it.
Not measurable in our spectra: OH.
[O I] 6300 Å forbidden line (blended with Ni I) · 1D LTE · HARPS · Codex Grade
O I · Asplund et al. 2021 lines · 3D NLTE (Amarsi) · IAG FTS · Reference Grade
O I · Asplund et al. 2021 lines · 3D NLTE (Amarsi) · Kitt Peak (Kurucz 2005) · Reference Grade
Every product is measured on one of three grades of spectral lines and the grades are never mixed in one number.
Scroll horizontally to read all columns · focus the table and use the arrow keys
| Grade | What it means | This element |
|---|---|---|
| Reference Grade | lines from a published solar line set (Asplund et al. 2021 and similar), measured with the oscillator strengths the authors used | 18 products -- O I · Asplund et al. 2021 lines |
| Codex Grade | lines chosen by the Codex whose oscillator strength carries a published uncertainty (laboratory measurement, NIST accuracy class A-C, or the paper's own stated error), between 5% and 60% deep | 10 products -- O I · Codex-graded lines; [O I] 6300 Å forbidden line (blended with Ni I) |
| Deep Grade | the same oscillator-strength rule on lines deeper than 60% -- stronger lines, more sensitive to damping and the model atmosphere | none for the Sun: every line in this grade is excluded as saturated or an outlier (see Excluded lines) |
Green band: Asplund, Amarsi & Grevesse 2021 photospheric reference, A(O) = 8.69 ± 0.04, drawn behind every row. Yellow band: Lodders, Bergemann & Palme 2025, 8.76 ± 0.05 (present-day Sun).
Green band: Asplund, Amarsi & Grevesse 2021 photospheric reference, A(O) = 8.69 ± 0.04, drawn behind every row. Yellow band: Lodders, Bergemann & Palme 2025, 8.76 ± 0.05 (present-day Sun).
Green band: Asplund, Amarsi & Grevesse 2021 photospheric reference, A(O) = 8.69 ± 0.04, drawn behind every row. Yellow band: Lodders, Bergemann & Palme 2025, 8.76 ± 0.05 (present-day Sun).
How to read this forest. Each row is one independent product, preserving its band, holding, engine, indicator and line set. Engines and indicator families are never averaged. Each bar is the uncertainty as Asplund et al. 2021 quote it: line-to-line scatter, the 3D-model term and the NLTE term (at least 0.03 dex). For a molecule the NLTE term is replaced by its dependence on the other element, and the 3D term comes from the 3D-to-1D difference Amarsi et al. 2021 publish for the same band. Atomic-data and observational terms are not in the bar, as in Asplund; they stay in the full error budget attached to the row. Grades describe the supplied line-pool provenance, not engine quality; evaluated theoretical CNO oscillator strengths are not relabelled as laboratory evidence. Quarantined, failed and unconstrained fits have no abundance mark. Experimental products remain explicitly not adopted. Literature references provide context, never tuning targets.
Every oxygen value on this page comes from fitting synthetic spectra to the observed solar spectrum, line by line or band by band, with the oxygen abundance as the free parameter. The synthesis includes the atomic lines and the molecular bands (CN, CH, C₂, OH and others) in every window, so blended absorption is modelled rather than attributed to the line being measured. The continuum is placed locally around every line by comparing the observation with the synthesis on its own highest pixels, and ground-based spectra are telluric-corrected across their full range.
Each indicator is measured independently on each solar spectrum and reported separately; nothing is averaged across indicators or instruments. Where a 3D or NLTE correction is published for the lines, it is applied as its own product (1D-LTE, 1D-NLTE and 3D-NLTE are separate rows). Every value carries a full uncertainty budget: fit, line strengths, microturbulence, continuum, fit window, model atmosphere, telluric residual, blends and coupling to the other C/N/O abundances, and the spread between spectra, each measured by refitting with that one input changed. The headline at the top is the Asplund-grade product: a published line set, the most complete treatment, then the most lines.
Scroll horizontally to read all columns · focus the table and use the arrow keys
| Band | What was measured | Solar spectra | Treatments | Lines | Result A(O) | Largest uncertainty terms (dex) |
|---|---|---|---|---|---|---|
| VIS | [O I] 6300 Å forbidden line (blended with Ni I) | HARPS | 1D-LTE | 1 | 8.934 (1D-LTE) | fit / line-to-line scatter 0.31, continuum placement 0.17 |
| red-optical | O I · Codex-graded lines | IAG FTS, Kitt Peak (1984, molecfit), Kitt Peak (Kurucz 2005) | 1D-LTE, 1D-NLTE, 3D-NLTE | 2 | 8.672–8.680 (3D-NLTE) | fit / line-to-line scatter 0.11, NLTE / 3D correction 0.07 |
| red-optical | O I · Asplund et al. 2021 lines | IAG FTS, Kitt Peak (1984, molecfit), Kitt Peak (Kurucz 2005) | 1D-LTE, 1D-NLTE, 3D-NLTE | 4 | 8.751–8.754 (3D-NLTE) | continuum placement 0.05, NLTE / 3D correction 0.03 |
| NIR | O I · Asplund et al. 2021 lines | IAG FTS, Kitt Peak (1984, molecfit), Kitt Peak (Kurucz 2005) | 1D-LTE, 1D-NLTE, 3D-NLTE | 3 | 8.651–8.817 (3D-NLTE) | continuum placement 0.14, fit / line-to-line scatter 0.03 |
Oxygen is traced by forbidden [O I], permitted O I and molecular OH indicators where available. Blends, departures from LTE and atmospheric structure affect these families differently, so their results retain separate identities.
O I 615.8nm: 🔴 DISAGREES — reported to Ryan, not resolved. . Store log gf -0.409 (); summed pull -0.2966 (B+).
O I 630.0nm: 🔴 DISAGREES — reported to Ryan, not resolved. 🔴 store claims A, pull publishes B+. Store log gf -9.717 (A); summed pull -9.7744 (B+).
OPEN. Tachiev and NIST differ systematically by -0.016 dex on the 777 triplet and -0.005 dex on 8446. AGSS21 names Tachiev for N i only and states no O i gf source. Both values are reported in oi_tachiev_vs_nist.csv; NIST is what the store already carried and this script does not change it. Ryan decides.
Generated from the pinned science snapshot. Source commit e76c06c6173928306b433a805509e63a1a206c71
Feed O.json v1.102 · updated 2026-10-10T15:46:17Z
Generated 2026-11-05T21:50:45-07:00 (source commit timestamp)
Feed SHA-256 6093a4771418bb4cf714974a885f3281b3d4feb452a2af1e8d4cbd2610becf32
Generator scripts/generate_cno_publication.py v1.0.0
Sources are separated by their role. Registered grid lineage and quarantined diagnostics do not imply an adopted abundance or an applied correction.
See the source notes for the distinction between reference lineage and the treatment actually recorded in each feed entry.
Solar abundance and indicator analyses. Solar O atomic/molecular indicator lineage and atmospheric sensitivity; comparison literature, not an adopted Codex abundance. Sources: [1], [2].
Atomic data and gf authorities. The recorded O I/O II base pedigree is Wiese–Fuhr–Deters / Opacity Project theory. Per-transition exceptions and the open O I source decisions remain explicit. Sources: [3], [4], [5].
Atomic data and gf authorities. O I triplet source record: NIST — Tachiev NOT adopted; OPEN — under review Sources: [3], [6].
Model and grid sources. Registered O I line/grid lineage: Amarsi, Nissen & Skuladottir 2019, A&A 630, A104 — line-by-line 3D-NLTE grids built on the Amarsi+2018 O I model atom. The current feed treatment remains authoritative; a grid reference does not imply a correction was applied. Sources: [7], [8].
Atomic data and gf authorities. Atomic O I entry; source-family pedigree only, not a substitute for a per-transition gf join. Sources: [3], [4], [5].
Model and grid sources. Feed treatment: 1D-LTE; no NLTE/3D correction is claimed for this entry. Turbospectrum is the synthesis implementation. Sources: [9].
Model and grid sources. Registered C/O model lineage; per-product grid provenance must confirm the applied treatment. Run provenance: Amarsi, Nissen & Skuladottir 2019, A&A 630, A104 — 3D non-LTE - 1D LTE on STAGGER (table2 C I / table3 O I); PER-LINE additive correction, grid lines only (match <= 0.05 A) Sources: [7], [8].
Model and grid sources. Registered C/O model lineage; per-product grid provenance must confirm the applied treatment. Run provenance: Amarsi, Nissen & Skuladottir 2019, A&A 630, A104 — 1D non-LTE - 1D LTE (table5 C I / table6 O I); PER-LINE additive correction, grid lines only (match <= 0.05 A) Sources: [7], [8].
Special diagnostics. The synthesis sidecar names Johansson’s Ni I blend data, Storey & Zeippen’s forbidden-line gf lineage and Caffau’s CO5BOLD comparison anchor. The feed entry remains 1D-LTE and quarantined; the sidecar’s cited anchor is not promoted to a measured abundance. Sources: [10], [11], [12].
O-products-2f26fcc4de61O-products-ee093b5cd43bO-products-dcccc1bf145aO-products-7570ea595567O-products-34b4e514053bO-products-f76024e80acaO-products-97dff9c24866O-products-4ce6582718b9O-products-dbd116adb0feO-products-951619c74bb3O-products-400d2fb23109O-products-b43c168af108O-products-14585017d015O-products-4b0ab56d8136O-products-b35970fb8af7O-products-9f2918c00938O-products-b87159d4fc51O-products-1a0fed49853dO-products-30cc32d77032O-products-5dd7a576c91eO-products-fabf2e672c03O-products-b90f07a8459cO-products-192c0fab28e1O-products-882b5f8ac19fO-products-6fccf43002bdO-products-e1b890ab2006O-products-d115064cbf77O-products-0d45f4f81b47