Effective collision strengths for excitation and de-excitation of nebular [O III] optical and infrared lines with kappa distributed electron energies

We present effective collision strengths for electron excitation and de-excitation of the ten forbidden transitions between the five lowest energy levels of the astronomically abundant doubly-ionised oxygen ion, O^{2+}. The raw collision strength data were obtained from an R-matrix intermediate coupling calculation using the Breit-Pauli relativistic approximation published previously by the authors. The effective collision strengths were calculated with kappa-distributed electron energies and are tabulated as a function of the electron temperature and kappa.

💡 Research Summary

The paper presents a comprehensive set of effective collision strengths (Υ) for electron excitation and de‑excitation among the five lowest fine‑structure levels of doubly‑ionised oxygen (O III). Using previously published raw collision strengths (Ω) from a high‑precision Breit‑Pauli R‑matrix calculation (72‑term target, UCL‑Belfast‑Strathclyde code), the authors perform thermal averaging over κ‑distributed electron energy distributions, rather than the traditional Maxwell‑Boltzmann (MB) distribution. The κ‑distribution, characterised by a parameter κ (3/2 < κ < ∞), introduces a high‑energy tail that becomes more pronounced as κ decreases, allowing a systematic exploration of non‑thermal electron populations that have been invoked to explain the long‑standing discrepancy between abundances derived from optical recombination lines (ORLs) and collisionally‑excited lines (CELs) in planetary nebulae (PNe) and H II regions.

The effective collision strength for excitation from level i to j is defined as
Υ_{i→j}= (√π/2) ∫₀^∞ Ω_{ij}(ε_i) ε_i e^{−ε_i/k_BT} f_κ(ε_i,T,κ) dε_i,
and for de‑excitation as
Υ_{j→i}= (√π/2) ∫₀^∞ Ω_{ij}(ε_j) ε_j f_κ(ε_j,T,κ) dε_j,
where f_κ is the κ‑distribution function. Unlike the MB case, Υ and its de‑excitation counterpart are not identical for κ‑distributions, especially for transitions with large energy separations.

The authors tabulate log₁₀ Υ and log₁₀ Υ for a grid of electron temperatures (log T = 2.0–4.3 in steps of 0.025 dex) and κ values ranging from 1.6 to 10⁶ (non‑uniform spacing). To aid practical use, they introduce a scaled quantity Υ_s = Υ exp(−ΔE/k_BT), which mitigates the steep rise of Υ at low T for high‑ΔE transitions. A cutoff is applied: any scaled value below 10⁻⁴ of its value at T = 10⁴ K for the same κ is replaced by the sentinel “99999999” to preserve table shape while signalling negligible contribution to excitation rates.

Conversion to collisional rate coefficients follows standard formulae:
q_{ij}=2√π c α a₀² ω_i R k_BT Υ_s · 8.629 × 10⁻⁶ / √T (cgs),
q_{ji}=2√π c α a₀² ω_j R k_BT Υ · 8.629 × 10⁻⁶ / √T.
Here ω_i/j are statistical weights, c the speed of light, α the fine‑structure constant, a₀ the Bohr radius, and R the Rydberg energy.

Key findings emerge from contour plots of Υ_s/√T for representative transitions. The 1–4 transition (³P → ¹D, responsible for the strong λλ 4959, 5007 lines) shows little sensitivity to κ for κ ≳ 3; the excitation rate remains essentially MB‑like. In contrast, the 1–5 transition (³P → ¹S, λ 4363 Å) is highly κ‑dependent: at κ ≈ 5–10 the excitation rate at T ≈ 6 300 K matches the MB rate at the canonical nebular temperature of 10⁴ K. Consequently, the classic temperature diagnostic ratio R = (I_{4959}+I_{5007})/I_{4363} can be reproduced at lower kinetic temperatures if a κ‑distribution is assumed. For example, R ≈ 194 (the standard MB value) occurs at T ≈ 6 300 K when κ = 10, implying that the O²⁺ ion density inferred from a given λ 5007 flux would be ≈3.2 times larger than under MB assumptions.

The paper also demonstrates that, despite the altered excitation rates, the internal level coupling within the five‑level atom remains essentially unchanged, allowing existing nebular modelling codes to incorporate the new Υ tables without structural modifications. The authors provide the full data set (ten Υ files and ten Υ files for de‑excitation) via the CDS repository, with detailed ReadMe documentation.

In summary, this work supplies the first publicly available, extensive κ‑averaged collision strength data for O III forbidden lines. It enables quantitative assessment of how non‑thermal electron energy distributions affect nebular temperature diagnostics and abundance determinations, offering a concrete tool for testing the κ‑distribution hypothesis in planetary nebulae and H II regions.