Exploring the effects of diffuse ionised gas in two local analogues of high-redshift star-forming galaxies

TL;DR

This paper tackles how diffuse ionised gas (DIG) biases emission-line diagnostics and gas-phase metallicities in two local analogues of high-redshift star-forming galaxies, UM 462 and IIZw40. Using spatially resolved MUSE data, the authors separate DIG from HII regions via the metric $\log(\Sigma(H\alpha))$, and derive oxygen abundances with the direct method and N2/O3N2 calibrators to study DIG effects. They find that DIG-dominated regions exhibit higher $12+\log(O/H)$ by roughly $0.2$ dex in UM 462 and $0.1$ dex in IIZw40 when using strong-line calibrators, with inter-bin differences up to about $0.4$ dex and $0.3$ dex, indicating that apparent metallicity gradients can arise from DIG ionisation rather than true abundance changes. The results imply smoothly varying ionisation conditions across the ISM, with DIG primarily sustained by photon leakage from HII regions and augmented by shocks, underscoring the need to account for DIG and aperture effects when interpreting metallicities in both local and distant galaxies.

Abstract

Aims. We investigate the impact of diffuse ionised gas (DIG) on the determination of emission line ratios and gas-phase metallicities in two local analogues of high-redshift star-forming galaxies: UM 462 and IIZw 40. Understanding how DIG affects these quantities is essential for interpreting unresolved observations of distant galaxies, where integrated spectra are often used to trace their chemical evolution. Methods. Using archival Very Large Telescope, Multi-Unit Spectroscopic Explorer (MUSE) data, we spatially resolved the warm ionised medium of both galaxies. We derived oxygen abundances through the direct method and several HII-based strong-line calibrators, and we used the H$α$ surface brightness ($Σ$(H$α$)) to distinguish regions dominated by HII or DIG emission. Results. Oxygen abundances derived from the N2 and O3N2 indices show an inverse correlation with $Σ$(H$α$), ionisation parameter, and EW(H$α$), with DIG-dominated regions exhibiting higher 12+log(O/H) than the galaxy mean by $\sim$0.2 dex in UM 462 and $\sim$0.1 dex in IIZw 40. The metallicity differences between HII-dominated and DIG-dominated $Σ$(H$α$) bins reach $\sim$0.4 dex and $\sim$0.3 dex in UM 462 and IIZw 40, respectively. The observed trends with $Σ$(H$α$), metallicity, EW(H$α$), and ionisation parameter indicate smoothly varying ionisation conditions rather than true abundance variations. These effects reflect different ionisation sources and levels, and can produce spurious metallicity gradients in galaxies with extended DIG structures, potentially mimicking signatures of metal-poor gas infall. In our sample, DIG ionisation is most likely dominated by photon leakage from H II regions, with additional contributions from feedback-driven shocks.

Figures (9)

• Figure 1: F(H$\alpha$) maps (left panels), EW(H$\alpha$) maps (middle), and $\sigma$(H$\alpha$) maps (right). Pixels/spaxels with a S/N < 5 were removed. The colour scale at the right of each panel indicates the F(H$\alpha$), EW(H$\alpha$) and $\sigma$(H$\alpha$) levels, respectively. North is up and east is left.
• Figure 2: Histograms of $\Sigma$(H$\alpha$) and EW(H$\alpha$). The HII-dominated spaxel values are shown in red, while DIG-dominated spaxel values are in blue. The dotted cyan lines indicate the classification by Lacerda2018, where DIG-dominated regions have EW(H$\alpha$) < 3 $\AA$ and HII region-dominated areas have EW(H$\alpha$) > 14 $\AA$. The mean spaxel values of our DIG-dominated areas are indicated by the vertical black lines.
• Figure 3: H$\alpha$ Surface brightness maps. The colour-code represents the regions of the galaxies divided into different $\Sigma$(H$\alpha$) bins of 37--$<$38 (purple), 38--$<$39 (blue), 39--$<$40 (green), 40--$<$41 (cyan), 41--$<$42 (orange), and 42--$<$43 (red) in units of log erg s$^{-1}$ kpc$^{-2}$. North is up, and east is left.
• Figure 4: Emission line ratios log([OIII]$\lambda$5007/H$\beta$), log([SII]$\lambda\lambda$6717,6731/H$\alpha$), and log([NII]$\lambda$6584/H$\alpha$) versus log($\Sigma$(H$\alpha$)). The vertical line indicates the threshold distinguishing DIG-dominated regions, defined as areas with $\Sigma$(H$\alpha$) $<$ 10$^{39}$ erg s$^{-1}$ kpc$^{-2}$. The black data points correspond to the mean values within the different $\Sigma$(H$\alpha$) bins defined in Sect. \ref{['sec_bins']}. There is a clear increasing pattern of [SII]/H$\alpha$ and [NII]/H$\alpha$ ratios and a decreasing trend of [OIII]/H$\beta$ as $\Sigma$(H$\alpha$) decreases. The colour scale at the right of each panel indicates the EW(H$\alpha$).
• Figure 5: BPT diagrams log([OIII]/H$\beta$ versus log([NII]/H$\alpha$) and log([OIII]/H$\beta$) versus log([SII]/H$\alpha$ for UM 462 (upper panels) and IIZw40 (lower panels). The colour code represents the spaxels of the galaxies divided into the different log($\Sigma$(H$\alpha$)) bins of $<$38--37 (purple), $<$39--38 (blue), $<$40--39 (green), $<$41--40 (cyan), $<$42--41 (orange), and $<$43--42 (red) in units of log erg s$^{-1}$ kpc$^{-2}$. In the BPT diagrams, we include the Kewley2001 (solid lines), Kauffmann2003 (dashed lines), and Schawinski2007 (dotted lines) model boundary lines, which divide regions dominated by star-forming HII, composite, and AGN/LINERs.