EDP Sciences logo
Subscriber Authentication Point
Search
Menu
Home All issues Volume 690 (October 2024) A&A, 690 (2024) A397 Full HTML
Press Release
Open Access
Issue
A&A
Volume 690, October 2024
Article Number A397
Number of page(s) 9
Section Extragalactic astronomy
DOI https://doi.org/10.1051/0004-6361/202449621
Published online 25 October 2024

© The Authors 2024

Licence Creative CommonsOpen Access article, published by EDP Sciences, under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

This article is published in open access under the Subscribe to Open model. Subscribe to A&A to support open access publication.

1. Introduction

The circumgalactic medium (CGM) is the bound gas halo surrounding galaxies outside their interstellar medium (ISM) and inside their virial radius, extending out to a few hundred kiloparsecs. The CGM may be the key regulator of the galactic gas supply. Gas flows occurring between the CGM and the interstellar medium (ISM) are thought to shape galaxies and drive their evolution, via feedback, accretion, and recycling of gas. Thus, investigating the metallicity, structure, and kinematics of the different gas phases will help us to understand how galaxies gain, eject, and recycle the gas during their existence (see Tumlinson et al. 2017 for a review).

Because the CGM is very diffuse and therefore almost invisible in emission, its physical properties remain largely unconstrained. Our understanding of it has so far come mostly from studies based on absorption lines produced by the CGM around galaxies, imprinted on the spectrum of background objects such as quasars and radio galaxies. The CGM of a few non-active nearby galaxies has also been studied in emission (e.g. Hayes et al. 2016), but the procedure is extremely challenging. The presence of a powerful active galactic nucleus (AGN) can render the CGM gas observable in emission around galaxies up to many tens of kiloparsecs, well into the CGM. Giant emission line nebulae (≳60 kpc in size and sometimes > 100 kpc) associated with quasars and radio galaxies at different redshifts have been studied since the 1980s (e.g. Baum et al. 1988; McCarthy et al. 1990; McCarthy 1993; van Ojik et al. 1996; Villar-Martín et al. 2003, 2018; Borisova et al. 2016; Fossati et al. 2021; Balmaverde et al. 2022; Wang et al. 2023). The discovery of giant (≥70 kpc), widely spread reservoirs of molecular gas (e.g. Emonts et al. 2016; Falkendal et al. 2021) associated with several high-z (z ≳ 2) radio galaxies reveals a multi-phase CGM that was chemically enriched when the Universe was 3 Gyr old and that supplies the gas from which galaxies grow.

Wide-field integral field spectroscopic instruments on 8–10 meter telescopes, such as the Multi Unit Spectroscopic Explorer (MUSE) on the Very Large Telescope (VLT), have created excellent opportunities to detect and study in great detail the elusive material from the warm (T < 105 K) ionised CGM around powerful active galaxies at different redshifts thanks to the illumination by the active nucleus. If it were not for the excitation by the AGN continuum, this gas might not be detected or would only be observable through absorption line studies.

The subject of this work is the well-known Teacup radio-quiet type-2 quasar at z = 0.085. It shows a ∼10 kpc loop of ionised gas resembling the handle of a teacup (hence its nickname), which was discovered by volunteers of the Galaxy Zoo project (Keel et al. 2012) and which has been widely studied in the context of AGN feedback and its potential impact on galaxy evolution. The system has been proposed to be the scenario of a giant outflow generated either by an AGN wind or induced by a 1 kpc radio jet whose effects are noticed up to at least ∼10 kpc from the AGN and that might be responsible for the bubble-like morphology (Gagne et al. 2014; Harrison et al. 2015; Keel et al. 2015; Ramos Almeida et al. 2017; Villar-Martín et al. 2018; Moiseev & Ikhsanova 2023; Venturi et al. 2023, hereafter V23).

Villar-Martín et al. (2018) discovered a > 100 kpc ionised nebula associated with this object (see also Villar Martín et al. 2021, which could be the product of a merger that occurred 1–2 Gyr ago (Keel et al. 2015). This rich gas reservoir, which extends into the CGM, has been rendered visible due to the activity of the quasar nucleus. The AGN photoionisation dominates the excitation of the spatially resolved gas emission up to its boundary (Gagne et al. 2014; Villar-Martín et al. 2018; V23; Moiseev & Ikhsanova 2023), except at some locations outside the putative quasar ionisation cones where evidence for shock excitation has been found (V23). Stellar photoionisation could also contribute to ionising the gas locally in some tidal features. The large-scale kinematics are strongly reminiscent of rotation (Villar-Martín et al. 2018; Moiseev & Ikhsanova 2023) and tentative results suggest subsolar nebular abundances (∼0.5 Z⊙, Villar-Martín et al. 2018). The well-known bubble appears to be expanding from the nucleus and out into the nebula.

We present in this paper a detailed optical spectroscopic study of the Teacup nebula based on VLT-MUSE archival data, with the main goal of mapping the gas abundances in two spatial dimensions. The ultimate scientific goal is to establish whether the giant outflow has an impact on the distribution of heavy elements from the nucleus on large spatial scales. We discuss the results in the context of other studies of the CGM and its role in the evolution of galaxies.

The paper is organised as follows. The data and analysis method are presented in Sects. 2 and 3, respectively. The latter includes the description of the methods used to derive the gas chemical abundances and physical properties of the gas. The results are presented in Sect. 4 and discussed in Sect. 5. The main conclusions are summarised in Sect. 6.

Throughout this paper, we assume a flat ΛCDM cosmology following Planck Collaboration VI (2020), with H0 = 67.4 km s−1 Mpc−1, and Ωm = 0.31. This gives a spatial scale of 1.65 kpc arcsec−1 at z = 0.085.

2. Data

The data were collected for the 0102.B-0107 programme (principal investigator, PI: L. Sartori; see V23 for details) with the European Southern Observatory (ESO) Very Large Telescope (VLT) and the Multi Unit Spectroscopic Explorer (MUSE, Bacon et al. 2010). This instrument covers a 1′ × 1′ field of view (FoV) in the Wide Field Mode (WFM), with a spatial sampling of 0.2″ pix−1. The wavelength coverage is ∼4650–9300 Å, with a 1.25 Å pix−1 spectral sampling and a resolving power of R = λ/Δλ ∼ 1700–3400 (ΔV ∼ 176–88 km s−1).

The observations were performed in March 2019. The processed archive data cube was used for this study. The Teacup ionised nebula fills a large fraction of the MUSE FoV. Separate sky cubes are not available from these observations. Sky oversubtraction due to contamination of the sky spectrum by object emission was identified at a few spatial locations of very low surface brightness, which appeared as artefacts that mimic absorption features adjacent to the strongest emission lines (especially [OIII]λ5007). Because our attempts to improve the sky subtraction did not achieve significant improvements, we finally used the archive cube. To further evaluate the potential impact of this effect, a comparison with the processed archive datacube from programme 0103.B-0071 (PI: C. Harrison) was also performed. Although separate sky cubes were obtained for this programme, the resulting sky subtraction was not significantly better. The comparison was in any case valuable to characterise the impact of the sky subtraction on the data quality. The artefacts are in general faint in comparison with the emission lines and are shifted in wavelength so that they can be efficiently identified and isolated. We confirm that imperfections on the sky subtraction do not affect our results and conclusions. These are all based on the 0102.B-0107 datacube because of the significantly better seeing (FWHM = 0.74″ vs. 1.3″ for the 0103.B-0071 program) and the somewhat better signal-to-noise ratio (S/N).

3. Analysis

3.1. Spatially resolved emission line flux measurements

Our main goal is to map the electron temperature, Te, and the oxygen abundance, O/H, in two spatial dimensions to investigate whether the giant outflow has an impact on the distribution of heavy elements across the galaxy and out into the CGM. For the Te determination, it is essential to measure the flux ratio of a nebular to an auroral emission line, such as the [OIII]λ5007/λ4363 (or [NII]λ6583/λ5755). For this, we defined spatial apertures at different locations through the nebula and extracted the integrated spectrum from each one.

The apertures were selected based on the visual inspection of the [OIII]λ5007 morphology at different wavelengths (i.e. velocities) scanned through the line profile. These scans reveal striking morphological changes with velocity and are especially useful to identify some faint structures, such as tidal tails, arcs, knots, and filaments, which have similar velocities (Fig. 1). By restricting the central λ and the width, Δλ, of the narrow-band images, the contrast of the lowest-surface-brightness features is enhanced and the spatial apertures covering them can be defined with greater accuracy to maximise the S/N of the integrated spectra. This was useful to detect faint emission lines through the largest possible extension, while preserving at the same time the morphological information on the diversity of nebular structural elements.

thumbnail Fig. 1.

[OIII] continuum subtracted images covering different spectral windows that were selected to highlight the diversity of nebular morphological features. Each image covers a different velocity (i.e. spectral) range relative to the nuclear systemic velocity, as is indicated on top. The nebular morphology strongly varies with velocity. The left panel shows the total [OIII] flux narrow-band image. The well-known ∼10 kpc ionised bubble is marked with a tiny yellow star in the left panel. The green lines indicate the position angles of the radio axis to the northeast and to the west (Harrison et al. 2015). To guide the reader, the letters ‘A’–‘F’ mark some emission line features that can also be identified in the mask map of Fig. 2.

The map of the positions of the resulting 64 selected apertures is shown in Fig. 2 (left panel). We also show in adjacent panels of the same figure the W80 and Vs maps based on [OIII]λ5007. W80 is the velocity width that encloses 80% of the total line flux and Vs is the velocity shift relative to the narrow core of the nuclear [OIII] line, considered here to be an indicator of the systemic velocity (Greene & Ho 2005). These maps can be directly compared with those of V23, which were based on a spaxel-by-spaxel analysis. The huge dimensions of the nebula both along and perpendicular to the radio axis are apparent. Although some prominent features stand out (knots, filaments, etc), gas emission seems to fill the entire area within the nebular outer boundaries. The large-scale rotation pattern (Gagne et al. 2014; Villar-Martín et al. 2018; Moiseev & Ikhsanova 2023) and the enhanced width, W80, of the lines in the direction perpendicular to the radio axis are clear (V23).

thumbnail Fig. 2.

Map of the masks used in our analysis (left) and [OIII] kinematic maps (middle and right panels). The colours in the first map have no particular meaning but help to differentiate the apertures. A 1D spectrum was extracted from each one, so that a single W80 and Vs (middle and left panels) value is associated with each aperture. Vs is the velocity shift relative to the narrow core of the nuclear [OIII] line. The maps cover the total MUSE FoV (∼1′×1′). W80 and Vs are in km s−1. Letters ‘A’–‘F’, the yellow star, and the solid lines have the same meanings as in Fig. 1.

The 1D spectra extracted from the individual apertures were used to measure the fluxes of the lines involved in our analysis: [OIII]λλ4959,5007 and λ4363, Hγ, Hβ, Hα, [NII]λλ6548,6583, and λ5755 and [SII]λ6716,6731. For a subset of spectra the stellar continuum was relatively strong compared with the emission line fluxes – that is, the equivalent width of the lines was low – so that it was first necessary to fit it and subtract it. The effects of stellar absorption can affect the Balmer emission lines and need to be accounted for to obtain accurate reddening estimates. It was also necessary in several spectra to reconstruct the Hγ and [OIII]λ4363 baseline and recover both line fluxes. The continuum was fitted for each spectrum with the PYPIPE3D tool (e.g., Lacerda et al. 2022), which performs a decomposition of the observed stellar spectra into multiple simple stellar populations, each of a different age and metallicity. As a test, we also subtracted the continuum with the method described in Cazzoli et al. (2022). The results were consistent. We also checked the reddening inferred from Hγ/Hβ and Hα/Hβ in the final, corrected spectra. Case B Hα/Hβ = 2.87 and Hγ/Hβ = 0.47 were assumed. Reddening estimations from both ratios are in general consistent within the errors. Changing these values within the range Hα/Hβ = 2.76–3.05 and Hγ/Hβ = 0.45–0.47 (Osterbrock 1989) has no significant impact. The EB − V values implied by Hα/Hβ were used for extinction correction since they have smaller errors. Using the EB − V implied by Hγ/Hβ instead also has a negligible impact on the results.

3.2. Derivation of temperatures and gas chemical abundances

We used the Python code PYNEB (Luridiana et al. 2015) to estimate Te of [OIII] for spectra with detected [OIII]λ4363 using [OIII]λ5007/λ4363, and, when possible, of [NII] using [NII]λ6583/λ5755. This code was also used to infer the electron density, ne, using [SII]λλ6717,6731. We calculated the corresponding errors by applying to the calculations a Monte Carlo iteration using the nominal fluxes perturbed with the observational errors.

For the derivation of abundances, the direct method (i.e. the derivation of total chemical abundances from their ionic fractions and the measured electron temperature) cannot be satisfactorily applied in the case of the narrow line regions (NLRs) in AGNs (e.g. Dors et al. 2015). Instead, we used the code HII-CHI-MISTRY (hereafter HCM, Pérez-Montero 2014). The approach of this code consists of a Bayesian-like comparison between certain reddening-corrected ratios of emission line fluxes relative to a recombination H line (Hβ in our case), which are sensitive to the total oxygen abundance (O/H), the nitrogen-to-oxygen ratio (N/O), and the ionisation parameter (U), with the predictions from a large grid of photoionisation models. In particular, we used version 5.3 of HCM, which considers models calculated under the most usual conditions in the NLR of active galaxies. As a result, the most probable values and their uncertainties for O/H, N/O and log(U) were obtained.

The grid of models is explained in detail in Pérez-Montero et al. (2019). The gas is assumed to be distributed homogeneously with a filling factor of 0.1 and a constant electron density of ne = 500 cm−3. The densities across the Teacup nebula are, however, significantly lower (ne ≲ 150 cm−3, see Table A.1 and V23). Since collisional de-excitation effects are not relevant for the line ratios under consideration in this density regime, the inferred abundances are not affected by the comparatively high density used in the models (Pérez-Montero et al. 2019).

The spectral energy distribution (SED) was considered to be composed of two components: one representing the big blue bump peaking at 1 Ryd, and the other a power law with a spectral index of αOX = −1.2 representing the non-thermal X-ray radiation (see Pérez-Montero et al. 2019 for additional details). The stopping criterion to measure the resulting emergent spectrum is that the proportion of free electrons in the ionised gas is lower than 2%. The models consider a Mathis et al. (1977) grain size distribution and a dust-to-gas mass ratio of 7.5 × 10−3 (Rémy-Ruyer et al. 2014).

The grid of models varies the oxygen and nitrogen abundances within 12 + log(O/H) = 6.9–9.1 in bins of 0.1 dex and log(N/O) = − 2.0–0.0 in bins of 0.125 dex, respectively. The rest of the chemical abundances are scaled in the models with respect to oxygen following the solar proportions, 12 + log(O/H) = 8.69 ± 0.04, given by (Asplund et al. 2021). In accordance with these authors, we also assumed 12 + log(N/H) = 7.83 ± 0.07 for the Sun.

The original grid of models in Pérez-Montero et al. (2019) covers values of log(U) from −4.0 to −0.5 in bins of 0.25 dex, although we only assumed log(U) >  − 2.5 to break the degeneracy with U for the ratios of high- to low-excitation lines in the optical range predicted by the models. This restriction does not imply any significant change in the derivation of the chemical abundances (see Pérez-Montero et al. 2019 for a better clarification). The code also provides uncertainties for the derived quantities calculated as the quadratical addition of the standard deviation of the obtained Bayesian distribution with the uncertainty obtained from a Monte Carlo iteration through the given input nominal flux of each line perturbed with its corresponding observational error.

4. Results

We show in Table A.1 the O/H and N/O abundances predicted by photoionisation models for different apertures, as well as the [OIII] and [NII] electron temperatures inferred with PYNEB for the densities estimated with [SII]λλ6716,6731 (see Sect. 3.2). The observed extinction corrected line ratios relative to Hβ and measured with the 1 dim spectra are also shown. This analysis was performed for a set of 23 (including the nucleus) apertures where [OIII]λ4363 has been detected (see left panel in Fig. 3). The top (lower) part of the table comprises the apertures in which [NII]λ5755 is detected (undetected). This does not affect the estimated chemical abundances, but simply indicates whether T[NII] could be inferred or not. The O/H, N/O, and T[OIII] information in Table A.1 is also shown as 2 dim maps in Figs. 3 and 4.

thumbnail Fig. 3.

Maps of (O/H)/(O/H)⊙ ratio and T[OIII]. The masks used in this analysis are shown in the left panel. They correspond to the apertures for which [OIII]λ4363 is detected. Notice that the FoV is smaller than in Figs. 1 and 2. The exact values of O/H and T[OIII] are in Table A.1. T[OIII] is in units of 104 K.

thumbnail Fig. 4.

Map of (N/O)/(N/O)⊙. The FoV is the same as in Fig. 3. The nuclear value is 1.99, which is outside the colour bar range to enhance the contrast for visualisation purposes. The (N/O)/(N/O)⊙ values are in Table A.1.

The main result that stands out in Fig. 3 is that the giant ionised bubble shows significantly higher O/H and lower T[OIII] than the gas at most locations across the giant nebula. This is confirmed in all apertures along the bubble edge, where 12 + log(O/H) is in the range 8.68 ± 0.09–8.77 ± 0.12. As in the nucleus (8.66 ± 0.11), these abundances are consistent with the solar value or slightly higher.

For the rest of the nebula (that is, the extended gas outside and beyond the bubble), except in Ap. 13, the oxygen abundances are subsolar everywhere (as was tentatively found by Villar-Martín et al. 2018). Considering all of the apertures across the nebula, the log of the median is 12 +log(O/H) = 8.49 (or 63% solar), with values as low as 8.37 ± 0.08–8.44 ± 0.13 (48%–56% solar) in Ap. 15, 16, 19, 21, and 22. Interestingly, the gas encircled by the bubble’s edges (Ap. 11), with 12 +log(O/H) = 8.53 ± 0.17 (69% solar), shows an intermediate abundance between the most enriched gas (at the bubble’s edges) and the more metal-poor gas in the rest of the nebula.

It is important to highlight that the abundance predictions are based on the assumption that the gas is photoionised by the AGN everywhere. This is reasonable for most apertures except, possibly, at a few positions located in the direction perpendicular to the main axis of the nebula, where shocks might be contributing to the excitation of the gas (V23). These apertures are labelled Ap. 12, 18, and 19 in the left panel of Fig. 3. In spite of this uncertainty, the fact that (except for Ap. 13) this gas follows the general behaviour of the O/H and T[OIII] maps suggests that the derived abundances are reliable.

In general, the T[OIII] map (right panel in Fig. 3) mimics the behaviour of the O/H maps in an anti-correlation. The nucleus shows the minimum temperature, T4 = T[OIII]/104 = 1.27 ± 0.08 K, in comparison with the rest of the gas, which shows T4 > 1.5 everywhere. As before, the bubble stands out in this map, with the edges being significantly colder (T4 ∼ 1.5–1.6 in Ap. 3–9) than the giant nebula, which is very hot, with T4 ∼ 1.7–1.9 almost everywhere. V23 present a 2 dim T[OIII] map tracing the nucleus and the ionised bubble. They infer a narrow range of significantly lower T[OIII] ∼ (1.3–1.4) ∼ 104 K compared to our estimations. The reason for this discrepancy remains unknown.

In regard to the N/O abundance (Fig. 4 and Table A.1), the most obvious result is the much higher nuclear value (log(N/O) = − 0.56 ± 0.08, i.e. roughly twice the solar abundance ratio of −0.86 ± 0.07), compared with the solar or somewhat below solar N/O everywhere else (−1.03 ± 0.09 ≤ log(N/O)≤−0.84 ± 0.20), including the bubble edge. There is not such a striking trend of N/O with location as was found for O/H. The bubble edge shows one of the lowest N/O ratios (∼68–85% N/O⊙), although this is not unique in comparison with other regions of the giant extended nebula. The bubble-edge N/O values are lower than the gas encircled by it (Ap. 11), the nucleus, and the gas in between (Ap. 1 and 2).

T[NII] could be measured only for the nucleus and up to the bubble edge (Table A.1). The temperature difference between these two regions is less pronounced than for T[OIII]: T 4 [ NII ] ∼ 0.983 ± 0.066 $ T_{4\rm [NII]}\sim 0.983\pm0.066 $ K in the nucleus and somewhat higher in the bubble (∼1.01–1.12 K).

5. Discussion

With a maximum extension traced by the MUSE data of ∼126 kpc (this is a lower limit, since the gas fills the FoV in this direction), the Teacup giant nebula traces part of the CGM. Its properties are strongly conditioned by the AGN, but it still provides valuable information about the CGM. If it were not for the nuclear activity, most (if not all) of this gas would remain invisible.

The giant nebula shows subsolar abundances, with O/H ∼ (48%–84%) × (O/H)⊙ almost everywhere, and a median of 63%, well below the nuclear, roughly solar abundance. For comparison, different works based on absorption line studies have shown that the dense gas in the CGM of z ≲ 1 galaxies has a bimodal metallicity distribution function, with an equal number of absorbers in the low-metallicity (Z ≲ 0.03 Z⊙) and high-metallicity (Z ∼ 0.4 Z⊙) branches (e.g. Lehner et al. 2013; Wotta et al. 2016). The abundance of the Teacup ionised nebula falls in the latter group (clearly, this does not discard the presence of lower-metallicity gas). The high-metallicity branch has been proposed to trace galactic winds, recycled outflows, and tidally stripped gas (Lehner et al. 2013).

Lower CGM metallicities compared with the ISM have been found for different galaxy types, based on studies of absorption line systems at ≲200 kpc from their host galaxies (e.g. Kacprzak et al. 2019). These authors found a large range of abundances in the CGM of isolated star-forming galaxies, 0.01 ≲ Z/Z⊙ < 1 and an offset of log(dZ) =  − 1.17 ± 0.11 between the CGM and ISM, which shows no dependence on stellar mass. The relation has a large scatter of 1σ = 0.72. The offset for the Teacup is smaller for O/H (log(dZ)∼ − 0.17 using the median nebular O/H), although still within the scatter. Different processes may be at work in this system, related to the nuclear activity and/or its merger history.

We have shown that the ∼10 kpc Teacup bubble to the east of the nucleus, which is known to be driven by an AGN wind or the small nuclear radio jet, is associated with obvious changes in the gas abundance. The bubble edge shows significantly higher O/H (solar or slightly solar, similar to the nucleus) in comparison with the subsolar O/H across the rest of the nebula. Most likely as a consequence, the bubble edge is also significantly less hot (T4 ∼ 1.5–1.6) than the rest of the nebula (1.7 ≲T4≲ 1.9). Therefore, the outflow appears to be causing a change in the gas metal content from the nucleus up to ∼10 kpc.

This mechanism may also explain the nuclear deficit of O/H. For the Teacup values of log(M*/M⊙) = 11.15 ± 0.05 and star-forming rate (SFR) of ∼10 M⊙ yr−1 (Jarvis et al. 2020; Ramos Almeida et al. 2022), 12 + log(O/H) ∼ 8.85 is expected, according to the mass-metallicity-SFR (M*-Z-SFR) relation by Pérez-Montero et al. (2013) or ∼8.8 for the extrapolation of the M*-Z-SFR relation by Andrews & Martini (2013) to high M*. Seyfert 2 galaxies with similar M* tend to show O/H consistent with these predictions (Pérez-Díaz et al. 2021). The Teacup nucleus, on the contrary, shows lower than expected O/H (8.66 ± 0.05), while the nuclear N/O (−0.56 ± 0.08, Table A.1) is consistent with that expected for its M* (log(N/O) = − 0.54, Pérez-Montero et al. 2013; Andrews & Martini 2013).

The dilution produced by inward flows of low-metallicity gas (for instance, from the outskirts of the two merging galaxies) could explain the nuclear O/H and N/O (Edmunds 1990; Rupke et al. 2010). The outflow is also an interesting possibility. This scenario is supported by the similar O/H of the bubble edge and the nuclear gas, which is moreover enhanced in comparison with the rest of the nebula. On the one hand, radiative outflows can couple more efficiently with metals via resonance line scattering (e.g. Pauldrach et al. 1994; Arav et al. 1994; Higginbottom et al. 2024; see also Edmunds 1990). On the other, galaxies show metallicity gradients such that the inner regions have higher abundances (Searle 1971; Kewley et al. 2019; Maiolino & Mannucci 2019). Based on this, a gradient is expected to exist within the ∼4 kpc diameter nuclear aperture used in the Teacup analysis. The outflow has been generated in the inner regions (≲1 kpc) close to the AGN (Harrison et al. 2015; Ramos Almeida et al. 2017; V23), where the gas is expected to be more metal-rich. If it drags gas out to large distances, as is proposed in this scenario, the global metallicity of the residual gas within the nuclear aperture would be lower as a consequence.

An implication of this scenario is that the outflow has been capable of ejecting gas from the galaxy’s centre and has dragged it up to ∼10 kpc. If the outflow expanded without displacing significant amounts of gas to large distances, the bubble abundance would be similar to the rest of the nebula. The implications are important. This supports the view that metal-enriched galactic outflows (driven by an AGN in this particular case) shape the mass-metallicity relationship, by removing metals from galaxy potential wells and ejecting them to large distances, possibly out into the CGM (Tremonti et al. 2004; Peeples et al. 2014; Chisholm et al. 2018; Tortora et al. 2022).

The behaviour of N/O remains to be explained. An outflow could preserve N/O (Edmunds 1990 specific modelling would be valuable for AGN-generated winds). On the contrary, this ratio is depleted at the bubble’s edge in comparison with the nucleus and its gradient shows no obvious spatial correlation with the bubble, but just a tentative trend to show among the lowest N/O. This is not necessarily a discrepancy. Given the complexity of the N/O behaviour in terms of secondary and primary stellar production processes, and the fact that the ejected gas would mix with gas across the nebula with a non-uniform N/O distribution, it is difficult to predict how this ratio would behave as the bubble expands and mixes with the pre-existing reservoir.

An alternative scenario to gas ejection from the centre is that local chemical enrichment has been produced by young stars. This is supported by the detection of blue-coloured continuum emission co-spatial with the bubble edge due to a population of stars that are younger (≲100–150 Myr) than in the rest of the galaxy (≳0.5–1 Gyr, V23). According to these authors, widespread star formation has been triggered at the edge of the bubble due to the compressing action of the jet and outflow (positive feedback). The timescale could be long enough to enrich the local gas with O, but not with secondary N (Mollá et al. 2006; Kumari et al. 2018). This would explain the enhanced O/H in comparison with the rest of the nebula, while N/O is not clearly different except in the nucleus, where N/O is significantly higher.

Yet another possible scenario to explain the abundance values in the bubble relates to the depletion by dust of metals from the gas-phase ISM. If shocks destroyed dust as the bubble expands (e.g. Dopita et al. 2018), metals could be released to the gas. Since oxygen is more sensitive to depletion than nitrogen (Jones & Ysard 2019), this could explain the higher O/H of the bubble in comparison with the rest of the nebula, and the bubble’s tentatively lower N/O.

Whether due to nuclear gas ejection, local star formation, or dust destruction, in all three scenarios the AGN-induced outflow is responsible for the metal enrichment of the gas at distances as large as ∼10 kpc. The implications are different. In the ejection scenario, the behaviour of O/H provides observational evidence of how AGN-induced outflows can deprive the central regions of galaxies from metals and transfer them up to very large distances, possibly out of the galaxy and into the CGM. In the second scenario, the behaviour of O/H provides evidence of how AGN-induced outflows can produce local metal enrichment (this is also the case in the third scenario, the dust depletion one) at very large distances from the nucleus, with a delay between the quasar onset and the induced metal enrichment of ∼100–150 Myr.

6. Conclusions

The giant (≳126 kpc) nebula associated with the Teacup QSO2 at z = 0.085 traces part of its CGM. Its properties are strongly influenced by the nuclear activity up to the outer detected emission line regions, where it still provides valuable information about the CGM surrounding the quasar host galaxy. If it were not for the nuclear activity, most (if not all) of this gas would remain invisible. This study is an example of the great potential of studying giant nebulae to investigate in emission the CGM around active galaxies at all redshifts.

The widely studied AGN-driven outflow responsible for the well-known ionised bubble is enhancing the gas metal content (O/H) up to ∼10 kpc from the AGN. The giant nebula shows subsolar metallicity almost everywhere, except the bubble, which has an approximately solar or slightly super-solar metallicity.

This could be a consequence of the ejection of metal-rich gas from the nucleus. In such a scenario, the Teacup provides observational evidence for how AGN feedback can deprive the central regions of galaxies of gas and displace metals out to very large distances, possibly out of the galaxy. It supports the view that metal-enriched AGN outflows can shape the mass-metallicity relationship of galaxies. Alternatively, the O/H enrichment could have been produced locally by the young stellar population formed at the bubble’s edge, possibly as a consequence of positive feedback ∼100–150 Myr (V23). A third possibility is the release of oxygen to the gas phase as a consequence of dust destruction in the bubble by shocks triggered by the expanding outflow. In any of the scenarios considered, the nuclear activity is the ultimate mechanism responsible for the metal enrichment of the gas at large extranuclear distances (∼10 kpc).

Acknowledgments

We are grateful to Luc Binette, Bjorn Emonts and Bruno Rodríguez for useful feedback. We also thank the referee for the careful revision of the manuscript and valuable suggestions. MVM and ACL research has been funded by grant Nr. PID2021-124665NB-I00 by the Spanish Ministry of Science and Innovation/State Agency of Research MCIN/AEI/ 10.13039/501100011033 and by “ERDF A way of making Europe”. SC acknowledges financial support from the Severo Ochoa grant CEX2021-001131-S and the Ministry of Science, Innovation and Universities (MCIU) under grants PID2019-106027GB-C41. EPM acknowledges financial support by project Estallidos8 PID2022-136598NB-C32 (Spanish Ministry of Science and Innovation). The Cosmology calculator by Wright (2006) has been used.

References

  1. Andrews, B. H., & Martini, P. 2013, ApJ, 765, 140 [NASA ADS] [CrossRef] [Google Scholar]
  2. Arav, N., Li, Z.-Y., & Begelman, M. C. 1994, ApJ, 432, 62 [NASA ADS] [CrossRef] [Google Scholar]
  3. Asplund, M., Amarsi, A. M., & Grevesse, N. 2021, A&A, 653, A141 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  4. Bacon, R., Accardo, M., Adjali, L., et al. 2010, Proc. SPIE, 7735, 773508 [Google Scholar]
  5. Balmaverde, B., Capetti, A., Baldi, R. D., et al. 2022, A&A, 662, A23 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  6. Baum, S. A., Heckman, T. M., Bridle, A., et al. 1988, ApJS, 68, 643 [NASA ADS] [CrossRef] [Google Scholar]
  7. Borisova, E., Cantalupo, S., Lilly, S. J., et al. 2016, ApJ, 831, 39 [Google Scholar]
  8. Cazzoli, S., Hermosa Muñoz, L., Márquez, I., et al. 2022, A&A, 664, A135 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  9. Chisholm, J., Tremonti, C., & Leitherer, C. 2018, MNRAS, 481, 1690 [NASA ADS] [CrossRef] [Google Scholar]
  10. Dopita, M. A., Vogt, F. P. A., Sutherland, R. S., et al. 2018, ApJS, 237, 10 [NASA ADS] [CrossRef] [Google Scholar]
  11. Dors, O. L., Cardaci, M. V., Hägele, G. F., et al. 2015, MNRAS, 453, 4102 [Google Scholar]
  12. Edmunds, M. G. 1990, MNRAS, 246, 678 [NASA ADS] [Google Scholar]
  13. Emonts, B. H. C., Lehnert, M. D., Villar-Martín, M., et al. 2016, Science, 354, 1128 [NASA ADS] [CrossRef] [Google Scholar]
  14. Falkendal, T., Lehnert, M. D., Vernet, J., et al. 2021, A&A, 645, A120 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  15. Fossati, M., Fumagalli, M., Lofthouse, E. K., et al. 2021, MNRAS, 503, 3044 [NASA ADS] [CrossRef] [Google Scholar]
  16. Gagne, J. P., Crenshaw, D. M., Kraemer, S. B., et al. 2014, ApJ, 792, 72 [Google Scholar]
  17. Greene, J. E., & Ho, L. C. 2005, ApJ, 627, 721 [NASA ADS] [CrossRef] [Google Scholar]
  18. Harrison, C. M., Thomson, A. P., Alexander, D. M., et al. 2015, ApJ, 800, 45 [Google Scholar]
  19. Hayes, M., Melinder, J., Östlin, G., et al. 2016, ApJ, 828, 49 [NASA ADS] [Google Scholar]
  20. Higginbottom, N., Scepi, N., Knigge, C., et al. 2024, MNRAS, 527, 9236 [Google Scholar]
  21. Jarvis, M. E., Harrison, C. M., Mainieri, V., et al. 2020, MNRAS, 498, 1560 [NASA ADS] [CrossRef] [Google Scholar]
  22. Jones, A. P., & Ysard, N. 2019, A&A, 627, A38 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  23. Kacprzak, G. G., Pointon, S. K., Nielsen, N. M., et al. 2019, ApJ, 886, 91 [NASA ADS] [CrossRef] [Google Scholar]
  24. Keel, W. C., Chojnowski, S. D., Bennert, V. N., et al. 2012, MNRAS, 420, 878 [Google Scholar]
  25. Keel, W. C., Maksym, W. P., Bennert, V. N., et al. 2015, AJ, 149, 155 [NASA ADS] [CrossRef] [Google Scholar]
  26. Kewley, L. J., Nicholls, D. C., & Sutherland, R. S. 2019, ARA&A, 57, 511 [Google Scholar]
  27. Kumari, N., James, B. L., Irwin, M. J., et al. 2018, MNRAS, 476, 3793 [NASA ADS] [CrossRef] [Google Scholar]
  28. Lacerda, E. A. D., Sánchez, S. F., Mejía-Narváez, A., et al. 2022, New Astron., 97, 101895 [NASA ADS] [CrossRef] [Google Scholar]
  29. Lehner, N., Howk, J. C., Tripp, T. M., et al. 2013, ApJ, 770, 138 [NASA ADS] [CrossRef] [Google Scholar]
  30. Luridiana, V., Morisset, C., & Shaw, R. A. 2015, A&A, 573, A42 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  31. Maiolino, R., & Mannucci, F. 2019, A&ARv, 27, 3 [Google Scholar]
  32. Mathis, J. S., Rumpl, W., & Nordsieck, K. H. 1977, ApJ, 217, 425 [Google Scholar]
  33. McCarthy, P. J. 1993, ARA&A, 31, 639 [NASA ADS] [CrossRef] [Google Scholar]
  34. McCarthy, P. J., Spinrad, H., van Breugel, W., et al. 1990, ApJ, 365, 487 [NASA ADS] [CrossRef] [Google Scholar]
  35. Mannucci, F., Cresci, G., Maiolino, R., et al. 2010, MNRAS, 408, 2115 [NASA ADS] [CrossRef] [Google Scholar]
  36. Martin, D. C., Matuszewski, M., Morrissey, P., et al. 2015, Nature, 524, 192 [Google Scholar]
  37. Martin, D. C., O’Sullivan, D., Matuszewski, M., et al. 2019, Nat. Astron., 3, 822 [NASA ADS] [CrossRef] [Google Scholar]
  38. Mollá, M., Vílchez, J. M., Gavilán, M., et al. 2006, MNRAS, 372, 1069 [CrossRef] [Google Scholar]
  39. Moiseev, A. V., & Ikhsanova, A. I. 2023, Universe, 9, 66 [NASA ADS] [CrossRef] [Google Scholar]
  40. Osterbrock, D. E. 1989, Astrophysics of Gaseous Nebulae and Active Galactic Nuclei (University Science Books) [Google Scholar]
  41. Pauldrach, A. W. A., Kudritzki, R. P., Puls, J., et al. 1994, A&A, 283, 525 [NASA ADS] [Google Scholar]
  42. Peeples, M. S., Werk, J. K., Tumlinson, J., et al. 2014, ApJ, 786, 54 [NASA ADS] [CrossRef] [Google Scholar]
  43. Pérez-Díaz, B., Masegosa, J., Márquez, I., et al. 2021, MNRAS, 505, 4289 [CrossRef] [Google Scholar]
  44. Pérez-Montero, E. 2014, MNRAS, 441, 2663 [CrossRef] [Google Scholar]
  45. Pérez-Montero, E., Contini, T., Lamareille, F., et al. 2013, A&A, 549, A25 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  46. Pérez-Montero, E., Dors, O. L., Vílchez, J. M., et al. 2019, MNRAS, 489, 2652 [CrossRef] [Google Scholar]
  47. Planck Collaboration VI. 2020, A&A, 641, A6 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  48. Prescott, M. K. M., Martin, C. L., & Dey, A. 2015, ApJ, 799, 62 [NASA ADS] [CrossRef] [Google Scholar]
  49. Ramos Almeida, C., Piqueras López, J., Villar-Martín, M., et al. 2017, MNRAS, 470, 964 [Google Scholar]
  50. Ramos Almeida, C., Bischetti, M., García-Burillo, S., et al. 2022, A&A, 658, A155 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  51. Rémy-Ruyer, A., Madden, S. C., Galliano, F., et al. 2014, A&A, 563, A31 [Google Scholar]
  52. Rupke, D. S. N., Kewley, L. J., & Barnes, J. E. 2010, ApJ, 710, L156 [NASA ADS] [CrossRef] [Google Scholar]
  53. Searle, L. 1971, ApJ, 168, 327 [Google Scholar]
  54. Tortora, C., Hunt, L. K., & Ginolfi, M. 2022, A&A, 657, A19 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  55. Tremonti, C. A., Heckman, T. M., Kauffmann, G., et al. 2004, ApJ, 613, 898 [Google Scholar]
  56. Tumlinson, J., Peeples, M. S., & Werk, J. K. 2017, ARA&A, 55, 389 [Google Scholar]
  57. van Ojik, R., Roettgering, H. J. A., Carilli, C. L., et al. 1996, A&A, 313, 25 [NASA ADS] [Google Scholar]
  58. Venturi, G., Treister, E., Finlez, C., et al. 2023, A&A, 678, A127 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  59. Villar-Martín, M., Vernet, J., di Serego Alighieri, S., et al. 2003, MNRAS, 346, 273 [CrossRef] [Google Scholar]
  60. Villar Martín, M., Emonts, B., Humphrey, A., et al. 2014, MNRAS, 440, 3202 [CrossRef] [Google Scholar]
  61. Villar-Martín, M., Cabrera-Lavers, A., Humphrey, A., et al. 2018, MNRAS, 474, 2302 [Google Scholar]
  62. Villar Martín, M., Emonts, B. H. C., Cabrera Lavers, A., et al. 2021, A&A, 650, A84 [CrossRef] [EDP Sciences] [Google Scholar]
  63. Wang, W., Wylezalek, D., Vernet, J., et al. 2023, A&A, 680, A70 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  64. Wotta, C. B., Lehner, N., Howk, J. C., et al. 2016, ApJ, 831, 95 [NASA ADS] [CrossRef] [Google Scholar]
  65. Wright, E. L. 2006, PASP, 118, 1711 [NASA ADS] [CrossRef] [Google Scholar]

Appendix A: Table of photoionisation model predictions

Table A.1.

Photoionisation model predictions

All Tables

Table A.1.

Photoionisation model predictions

In the text

All Figures

thumbnail Fig. 1.

[OIII] continuum subtracted images covering different spectral windows that were selected to highlight the diversity of nebular morphological features. Each image covers a different velocity (i.e. spectral) range relative to the nuclear systemic velocity, as is indicated on top. The nebular morphology strongly varies with velocity. The left panel shows the total [OIII] flux narrow-band image. The well-known ∼10 kpc ionised bubble is marked with a tiny yellow star in the left panel. The green lines indicate the position angles of the radio axis to the northeast and to the west (Harrison et al. 2015). To guide the reader, the letters ‘A’–‘F’ mark some emission line features that can also be identified in the mask map of Fig. 2.
In the text
thumbnail Fig. 2.

Map of the masks used in our analysis (left) and [OIII] kinematic maps (middle and right panels). The colours in the first map have no particular meaning but help to differentiate the apertures. A 1D spectrum was extracted from each one, so that a single W80 and Vs (middle and left panels) value is associated with each aperture. Vs is the velocity shift relative to the narrow core of the nuclear [OIII] line. The maps cover the total MUSE FoV (∼1′×1′). W80 and Vs are in km s−1. Letters ‘A’–‘F’, the yellow star, and the solid lines have the same meanings as in Fig. 1.
In the text
thumbnail Fig. 3.

Maps of (O/H)/(O/H)⊙ ratio and T[OIII]. The masks used in this analysis are shown in the left panel. They correspond to the apertures for which [OIII]λ4363 is detected. Notice that the FoV is smaller than in Figs. 1 and 2. The exact values of O/H and T[OIII] are in Table A.1. T[OIII] is in units of 104 K.
In the text
thumbnail Fig. 4.

Map of (N/O)/(N/O)⊙. The FoV is the same as in Fig. 3. The nuclear value is 1.99, which is outside the colour bar range to enhance the contrast for visualisation purposes. The (N/O)/(N/O)⊙ values are in Table A.1.
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.

OSZAR »