Helium abundances and its radial gradient from the spectra of HII regions and ring nebulae of the Milky Way
J. E. Méndez-Delgado, C. Esteban, J. García-Rojas, K. Z. Arellano-Córdova, M. Valerdi
aa r X i v : . [ a s t r o - ph . S R ] J un MNRAS , 1–15 (2015) Preprint 12 June 2020 Compiled using MNRAS L A TEX style file v3.0
Helium abundances and its radial gradient from thespectra of H ii regions and ring nebulae of the Milky Way J. E. M´endez-Delgado , ⋆ , C. Esteban , , J. Garc´ıa-Rojas , , K. Z. Arellano-C´ordova , and M. Valerdi Instituto de Astrof´ısica de Canarias, E-38200 La Laguna, Tenerife, Spain Departamento de Astrof´ısica, Universidad de La Laguna, E-38206, La Laguna, Tenerife, Spain Instituto Nacional de Astrof´ısica, ´Optica y Electr´onica. Apdo. Postal 51 y 216, Puebla, Mexico Instituto de Astronom´ıa, Universidad Nacional Aut´onoma de M´exico, Apdo. Postal 70-264 Ciudad Universitaria, Mexico
Accepted XXX. Received YYY; in original form ZZZ
ABSTRACT
We determine the radial abundance gradient of helium in the disc of the Galaxy frompublished spectra of 19 H ii regions and ring nebulae surrounding massive O stars. Werevise the Galactocentric distances of the objects considering Gaia
DR2 parallaxes anddetermine the physical conditions and the ionic abundance of He + in a homogeneousway, using between 3 and 10 He i recombination lines in each object. We estimatethe total He abundance of the nebulae and its radial abundance gradient using fourdifferent ICF(He) schemes. The slope of the gradient is always negative and weaklydependent on the ICF(He) scheme, especially when only the objects with log( η ) < − − − , consistent withthe predictions of chemical evolution models of the Milky Way and chemodynamicalsimulations of disc galaxies. Finally, we estimate the abundance deviations of He, Oand N in a sample of ring nebulae around Galactic WR stars, finding a quite similarHe overabundance of about +0.24 ± Key words:
ISM: abundances – H ii regions – Galaxy: abundances – Galaxy: disc –Galaxy: evolution – ISM: bubbles – stars: massive– stars: Wolf-Rayet Although rare on Earth, helium is the second most abun-dant element in the Universe and constitutes about 24-25% of its baryonic mass. The vast majority of the cos-mic helium was produced during the primordial nucleosyn-thesis phase just after the Big Bang. The fraction of pri-mordial mass in helium, Y P , has been determined follow-ing three different techniques. The results of WilkinsonMicrowave Anisotropy Probe (WMAP) and Planck satel-lites devoted to the study of the Cosmic Microwave Back-ground (CMB) anisotropies have obtained values between0.245 and 0.247 assuming the Standard Big Bang Nucleosyn-thesis (Coc et al. 2005; Planck Collaboration et al. 2018).Studies of intergalactic clouds near pristine absorption sys-tems along the line of sight of bright distant quasars de-termine upper limits of Y P in the range 0.225 and 0.283(Cooke & Fumagalli 2018). The last technique, based on theanalysis of spectra of H ii regions of metal-poor galaxies gives ⋆ E-mail: [email protected] values of Y P between 0.238 and 0.257 in the most recentworks (Izotov et al. 2014; Aver et al. 2015; Peimbert et al.2016; Fern´andez et al. 2018; Valerdi et al. 2019).After the Big Bang, helium is produced by hydrostaticnucleosynthesis in the interior of stars of all initial masses.Low-mass stars produce this element through the proton-proton chain while intermediate mass and massive ones viathe CNO cycle. Helium can be also efficiently destroyed instellar interiors by the triple-alpha process. The amount ofthis element that is actually ejected by a given star and en-rich the ISM depends on its initial mass and the importanceof stellar winds.The analysis of Galactic H ii region spectra indi-cates the presence of radial gradients of the abundancesof heavy elements – such as O, N, Ne, S, Ar or Cl –along the disc of the Milky Way (e.g. Shaver et al. 1983;Deharveng et al. 2000; Rudolph et al. 2006; Balser et al.2011; Esteban & Garc´ıa-Rojas 2018). The form of such gra-dients reflects the action of stellar nucleosynthesis, the dis-tribution and history of star formation and gas flows in thechemical evolution of the Galaxy. Although the helium abun- © J. E. M´endez-Delgado et al. dance should increase with the metallicity, there is not aclear evidence of the presence of a radial gradient of this el-ement in the Milky Way. Some authors (e.g. Peimbert et al.1978; Talent & Dufour 1979) find a slight or marginal evi-dence of a negative gradient but others find a flat distribu-tion of the helium abundance along the Galactic disc (e.g.Shaver et al. 1983; Fern´andez-Mart´ın et al. 2017, see § ii regions only shows recombinationlines of He + . Since the ionization potential of He is 24.6 eV,we expect the presence of neutral helium in the H ii region,but it cannot be observed. To determine the total heliumabundance from the measured He + / H + ratio, we have torely on an ionization correction factor, ICF. In the absenceof a tailored photoionization model for the object, we haveto assume a particular ICF scheme based on the results ofgrids of photoionization models or on the similarity of ion-ization potentials of other particular ions. All ICF schemesare parameterized by the ionization degree measured in thenebular spectrum. There are several ICF schemes for heliumavailable in the literature and based on different ionic ra-tios (e.g. Peimbert & Torres-Peimbert 1977; Peimbert et al.1992; Kunth & Sargent 1983; Zhang & Liu 2003).There are two more additional sources of uncertaintyin the determination of the total abundance of helium andthey are related to deviations from the pure recombinationspectrum of He i . The He i atom has two different level statesdepending on its total spin quantum number, singlets andtriplets. While – in principle – the intensity ratios of sin-glet lines should follow the predictions of recombination the-ory, the triplet spectrum is affected by the metastabilityof the lowest triplet level, the 2 S (Osterbrock & Ferland2006). This metastable level produces important collisionaland and self-absorption effects. Collisions with free electronsmay pump electrons from 2 S to upper levels, enhancingthe intensities of the lines coming from those excited lev-els with respect to the predictions of recombination the-ory. Although triplet lines are the most affected, some sin-glets lines can be also enhanced (Sawey & Berrington 1993;Kingdon & Ferland 1995). Photons of transitions ending on2 S can be reabsorbed, and, eventually, emitted as other He i lines. This self-absorption process increases or decreases lineintensities of triplet lines with respect to recombination ac-cording to the line considered. The strength of collisionaleffects on the He i spectrum depends on the electron densityand temperature of the ionized gas, being higher in denserand hotter nebulae. The calculations of the recombinationspectrum of He i performed by Porter et al. (2012) includethe collisional effects in the level populations of He . Self-absorption effects depend mostly on density, and are not ex-pected to be very important in H ii regions (Benjamin et al.2002).The aim of this paper is to explore the existence of aradial gradient of helium in the Milky Way using the bestdataset of deep spectra of H ii regions available to date andconsidering revised distances for the nebulae. The selectiongives priority to deep spectra that have been treated homo-geneously in their reduction process. The structure of thispaper is as follows. In § ii region spectra used and the determination of the revised distances.In § n e and T e – adopted for each ob-ject; the He + and He abundance recalculations, as well asdiscuss the different ICFs used. In § § § We have used reddening corrected intensity ratios of severalemission lines of 38 spectra corresponding to 24 Galactic H ii regions and ring nebulae around massive stars (Wolf-Rayet,WR or O-type stars). The data have been obtained, reducedand analyzed by our group and published in Esteban et al.(2004, 2013, 2016, 2017), Esteban & Garc´ıa-Rojas (2018)and Garc´ıa-Rojas et al. (2004, 2005, 2006, 2007). We willrefer this group of publications as the ”set of source pa-pers”. The spectra were obtained with the Ultraviolet Vi-sual Echelle Spectrograph (UVES, D’Odorico et al. 2000) atthe Very Large Telescope (VLT); the OSIRIS (Optical Sys-tem for Imaging and low-Intermediate-Resolution IntegratedSpectroscopy) spectrograph (Cepa et al. 2000, 2003) at the10.4 m Gran Telescopio Canarias (GTC) and the MagellanEchellette (MagE) spectrograph at the 6.5m Clay Telescope(Marshall et al. 2008). The details of the observations andthe instrument configurations used are described in the setof source papers.In the set of source papers the assumed distance of theSun to the Galactic Centre was R = 8.0 kpc (Reid 1993).Most of the recent determinations do not provide substan-tially different values of this parameter, but the precision hasincreased considerably. Bland-Hawthorn & Gerhard (2016)present a compilation of direct (primary), model-based andsecondary determinations, proposing a best estimate for thedistance to the Galactic Center of 8.2 ± R G ) of the objects. The GRAVITYcollaboration has published a very recent geometric distancedetermination of the Galactic centre black hole with 0.3%uncertainty (Gravity Collaboration et al. 2019), which is en-tirely consistent with 8.2 kpc.The set of source papers give the Galactocentric dis-tances, R G for each object. In the case of the H ii regions,Esteban et al. (2017) and Esteban & Garc´ıa-Rojas (2018)adopted the mean values of kinematic and stellar distancesgiven in different published references and an uncertaintycorresponding to their standard deviation. In order to ob-tain an improved set of R G values, we have made a revi-sion taking into account new distance measurements basedon Gaia parallaxes of the second data release (DR2). ForM8, M16, M17, M20, M42 and NGC 3576 we have adoptedthe distances derived by Binder & Povich (2018) from
Gaia parallaxes, with typical uncertainties in the range 0.1-0.3kpc. For the rest of the objects we have searched the dis-tances of the ionizing and/or associated star (or stars) in-ferred by Bailer-Jones et al. (2018) from
Gaia
DR2 data
MNRAS , 1–15 (2015) elium abundances and gradient in Galactic nebulae and a distance prior that varies smoothly as a functionof the Galactic coordinates of the objects according to aGalaxy model. In general, we find that distances based on Gaia parallaxes are very similar to those assumed previ-ously in Esteban et al. (2017) and Esteban & Garc´ıa-Rojas(2018). The mean difference for the 23 objects for whichthe two kinds of determinations can be compared is about9% (the median is 5%). The uncertainties in the distanceof the
Gaia parallaxes tend to be larger than about 1 kpcand larger than those quoted by Esteban et al. (2017) andEsteban & Garc´ıa-Rojas (2018) for objects at heliocentricdistances about or larger than 5 kpc. The largest differencescorrespond to the the most external object of the sample(Sh 2-209, 38%), but also for the objects located at R G ≤ Gaia
DR2 data (Bailer-Jones et al. 2018) with those adopted byEsteban et al. (2017) or Esteban & Garc´ıa-Rojas (2018), as-suming one or the other depending on which determinationgives the smallest uncertainties. We describe the details foreach object below.
NGC 2579 . We analyzed the
Gaia
DR2 data for starslocated within 1 arcmin around the center of the H ii region.The parallaxes of the stars give a large and rather uncertainheliocentric distance for the object (5.31 ± R G of about 10.88 ± R G = 12.40 ± α velocity field as well as theprevious photometrical and kinematical determinations byRusseil et al. (2007). We assume the distance determined byCopetti et al. (2007) for this object. Sh 2-83 . There are not specific studies investigating theionizing sources of this nebula. The only studies that givethe distance of Sh 2-83 are those by Caplan et al. (2000)and Anderson et al. (2015) based on kinematic data of thenebular gas, and both give the same heliocentric distanceof 18.4 kpc, the largest value of the whole sample. If theobject is at such large distance, its parallax would be toosmall to be well measured in
Gaia
DR2. Therefore, it is notsurprising that we do not find stars at so large distances inour analysis for stars located within 1 arcmin around thecenter of the H ii region. We adopted the distance obtainedby Caplan et al. (2000) and Anderson et al. (2015), that wasthe one assumed by Esteban et al. (2017) for this object. Sh 2-209 . Chini & Wink (1984) identified 3 ionizingstars for this nebula, indicating their position in a photo-graph. We have searched for those stars in the
Gaia
DR2database obtaining R G about 10.5 kpc for all of them. Thesevalues are very much lower than the 17.00 ± R G ∼ R G = 17 kpc, it would imply aheliocentric distance of about 9 kpc, but we do not find starslocated at such large distances within 3 arcmin around thecentre of the H ii region in Gaia
DR2. Such a high heliocen-tric distance may imply a parallax too small to be measuredin
Gaia
DR2 and this may be the explanation of not findingsuitable distant stars in the catalogue. We finally adopt thedistance of 17.00 ± Sh 2-311 . This H ii region is ionized by a single star:HD 64315, but it is a multiple stellar system (Lorenzo et al.2017) and its Gaia parallax is negative. We obtained andrepresented the parallaxes and proper motions of all starslocated within 5 arcmin around HD 64315 (1810 stars) find-ing 3 defined peaks in the distribution. We calculated thedistance corresponding to the center of each peak and theuncertainty corresponding to 1 σ . We assumed the distanceof 11.22 ± ± Gaia data for this objectThe rest of the sample objects are located at heliocen-tric distances about or larger than 5 kpc and their
Gaia
DR2parallaxes give rather uncertain distances. We have finallyassumed the distance adopted by Esteban et al. (2017) orEsteban & Garc´ıa-Rojas (2018) for all these nebulae. As ithas been said before, this distance corresponds to the meanof several independent photometrical and kinematical deter-minations (obtained mainly from Caplan et al. 2000; Russeil2003; Russeil et al. 2007; Quireza et al. 2006; Balser et al.2011; Foster & Brunt 2015) that provide a standard devia-tion lower than the distance uncertainty obtained from the
Gaia parallaxes.
NGC 3603 . This distant object has been cited as aGalactic giant H ii region and compared with 30 Doradusin the Large Magellanic Cloud. Drew et al. (2019) compileda list of almost 300 candidate O stars of the associated clus-ter and estimate an heliocentric distance of 7.0 ± Gaia
DR2 data, that implies a R G = 8.6 ± ± Sh 2-100 . Samal et al. (2010) found that Sh 2-100 is in amolecular complex containing 7 H ii regions. Those authorsidentify several ionizing stars inside the complex. We haveobtained their Gaia parallaxes that give a very large helio-centric distance of 9.5 + . − . kpc for the object, correspondingto R G = 10.2 + . − . kpc, which is consistent with the valueadopted by Esteban et al. (2017) but much more uncertain. Sh 2-127 . Rudolph et al. (1996) conclude that this H ii region is ionized by two O-type stars. Only one of them isvisible in the optical, that seems to be associated to the Gaia
DR2 source 217611164377605888. The parallax of thissource is very uncertain and gives a very large heliocentricdistance of 9.5 + . − . kpc, that corresponds to R G = 13.2 + . − . kpc. This value is consistent with the distance of 14.2 ± MNRAS000
DR2 source 217611164377605888. The parallax of thissource is very uncertain and gives a very large heliocentricdistance of 9.5 + . − . kpc, that corresponds to R G = 13.2 + . − . kpc. This value is consistent with the distance of 14.2 ± MNRAS000 , 1–15 (2015)
J. E. M´endez-Delgado et al.
Figure 1.
Spatial distribution of the sample nebulae (blue tri-angles) onto the Galactic plane with respect to the centre of theMilky Way. The concentric circles indicate increasing Galactocen-tric distances (in kpc). The Sun position is given by the orangesmall circle.
Sh 2-128 . This H ii region is ionized by the star ALS19702 (Bohigas & Tapia 2003). Its parallax is very uncer-tain, due to its heliocentric distance of 8.6 + . − . kpc, corre-sponding to R G = 12.7 + . − . kpc, but entirely consistent withthat of 12.50 ± Sh 2-152 . We use the star no. 4 of Russeil et al. (2007)for obtaining the
Gaia
DR2 parallax (the stars no. 1 to3 are more likely associated with Sh 2-153), obtaining a R G of 10.9 + . − . kpc, in agreement but slightly more un-certain than the distance of 10.3 ± Sh 2-212 . Moffat et al. (1979) catalogued several starsassociated to this star-forming region. Their
Gaia parallaxesare very uncertain and provide an heliocentric distance of6.0 + . − . kpc and R G = 13.9 + . − . kpc, which is consistent withthat of 14.6 ± Sh 2-288 . Avedisova & Kondratenko (1984) found thatthis nebulae is ionized by the star GSC 04823-00146. Its veryuncertain parallax gives R G = 12.7 + . − . kpc. We use the moreprecise distance of 14.10 ± ii region or ring nebulaaround WR or O-type star), adopted R G , instrument andtelescope with which they were observed and the referenceto their published spectra. In Fig. 1 we show the spatialdistribution of the sample nebulae onto the Galactic plane. Although the set of source papers present calculations ofphysical conditions – electron temperature, T e , and density, n e – and ionic and total abundances of several elements –including helium in some cases– for the nebulae, we decided to re-calculate all the relevant quantities in order to have ahomogeneous set of values using the same methodology andupdated atomic data. The values of n e and T e used for the calculation of theionic abundances of the H ii regions have been taken fromArellano-C´ordova et al. (2020). In the case of the ring neb-ulae – G2.4 + n e and T e usingthe same methodology and atomic data as for the rest of theobjects. Following Arellano-C´ordova et al. (2020), we usethe version 1.0.26 of pyneb (Luridiana et al. 2015) in com-bination with the atomic data for collisionally excited lineslisted in Table 2 and the reddening corrected line-intensityratios published in the source papers. The n e was calcu-lated using the ratio [S ii ] λ / λ or its average with[O ii ] λ / λ when both density diagnostics were avail-able. We obtain low densities for most of the objects, andassumed the value n e = 100 cm − (Osterbrock & Ferland2006) to determine T e and ionic abundances when n e <
100 cm − . We used a two-zone scheme characterized by T e ([N ii ]) and T e ([O iii ]) derived from the line intensity ratios[N ii ] ( λ + λ )/ λ and [O iii ] ( λ + λ )/ λ ,respectively. Although the intensity of the [N ii ] λ ii regions – to estimate T e ([N ii ]) or T e ([O iii ]) in those objects where one of these temperatureindicators was not available. We have used Monte Carlo cal-culations to estimate the uncertainties associated to eachvalue of n e and T e . We generated 500 random values for eachdiagnostic line ratio assuming a Gaussian distribution witha standard deviation equal to the associated uncertainty ofthe line intensities involved in the diagnostic. With these dis-tributions, we calculated new simulated values of n e and T e .Their associated errors correspond to a deviation of 68 per-cent – equivalent to one standard deviation – centred in themode of the distribution. The final results for n e , T e ([N ii ])and T e ([O iii ]) and their associated uncertainties are includedin Table 3. + abundances All the spectra of our sample show several He i recombina-tion lines. Although the set of source papers present calcula-tions of the He + / H + and He/H ratios of some of the nebulae,we have re-calculated them using the physical conditions in-dicated in Section 3.1. For each object, we use all or severalof the following list of He i lines: λ λ λ λ λ λ λ λ λ λ λ λ λ λ λ λ λ + abundancehas been determined using pyneb , the aforementioned lineintensity ratios taken from the set of source papers, the val-ues of n e and T e ([O iii ]) given in Table3 and the effective re-combination coefficient computations by Porter et al. (2012, MNRAS , 1–15 (2015) elium abundances and gradient in Galactic nebulae Table 1.
Sample of objects which spectroscopical data are used in this paper.Object Type a R G (kpc) Telescope Instrument ReferenceG2.4+1.4 WR-RN . + . − . GTC OSIRIS Esteban et al. (2016)MTC MagEM8 H ii ± b VLT UVES Garc´ıa-Rojas et al. (2007)M16 H ii ± b VLT UVES Garc´ıa-Rojas et al. (2006)M17 H ii ± b VLT UVES Garc´ıa-Rojas et al. (2007)M20 H ii ± b VLT UVES Garc´ıa-Rojas et al. (2006)M42 H ii ± b VLT UVES Esteban et al. (2004)NGC 2579 H ii ± c VLT UVES Esteban et al. (2013)NGC 3576 H ii ± b VLT UVES Garc´ıa-Rojas et al. (2004)NGC 3603 H ii ± c VLT UVES Garc´ıa-Rojas et al. (2006)NGC 6888 WR-RN 7.94 ± b GTC OSIRIS Esteban et al. (2016)NGC 7635 O-RN 9.43 + . − . GTC OSIRIS Esteban et al. (2016)RCW 52 O-RN 7.83 + . − . MTC MagE Esteban et al. (2016)RCW 58 WR-RN 7.63 + . − . MTC MagE Esteban et al. (2016)Sh 2-83 H ii ± c GTC OSIRIS Esteban et al. (2017)Sh 2-100 H ii ± c GTC OSIRIS Esteban et al. (2017)Sh 2-127 H ii ± c GTC OSIRIS Esteban et al. (2017)Sh 2-128 H ii ± c GTC OSIRIS Esteban et al. (2017)Sh 2-152 H ii ± c GTC OSIRIS Esteban & Garc´ıa-Rojas (2018)Sh 2-209 H ii ± c GTC OSIRIS Esteban et al. (2017)Sh 2-212 H ii ± c GTC OSIRIS Esteban et al. (2017)Sh 2-288 H ii ± c GTC OSIRIS Esteban et al. (2017)Sh 2-298 WR-RN 11.56 + . − . VLT UVES Esteban et al. (2017)Sh 2-308 WR-RN 9.67 + . − . MTC MagE Esteban et al. (2016)Sh 2-311 H ii ± b VLT UVES Garc´ıa-Rojas et al. (2005) a H ii : H ii region; WR-RN: Wolf-Rayet ring nebula; O-RN: ring nebula around O-type star. b Distance determined from
Gaia
DR2 parallaxes (see text for details). c Distance taken from Esteban et al. (2017) or Esteban & Garc´ıa-Rojas (2018).
Table 2.
Atomic dataset used for collisionally excited lines of selected heavy-element ions.Transition probabilitiesIon and energy levels Collisional strengthsN + Froese Fischer & Tachiev (2004) Tayal (2011)O + Froese Fischer & Tachiev (2004) Kisielius et al. (2009)O + Wiese et al. (1996); Storey & Zeippen (2000) Storey et al. (2014)S + Podobedova et al. (2009) Tayal & Zatsarinny (2010)S + Podobedova et al. (2009) Grieve et al. (2014) i lines and Storey & Hummer (1995) for H i lines. The calculations by Porter et al. (2012) include colli-sional effects in the level populations of He . A Monte Carlosimulation similar to the one described in Section 3.1 – in-cluding random distributions of n e , T e and the line intensities– was applied to estimate the uncertainty of the He + / H + ra-tios derived for each individual line of each spectrum. The He + / H + ratios obtained for each of the He i lines of the spec-tra of the sample objects are shown in Table A1. Lines withintensity uncertainties greater than 40% and those affectedby blending with telluric lines or other spectral features werenot considered for abundance determinations. Moreover, the He + / H + ratios calculated from the bright He i λ λ S level on the predicted recombination spectrumof He i (Benjamin et al. 2002), that only affect noticeablyto triplet lines. Moreover, He i λ i λ He + / H + ratios of each spectrum. The weight of each line hasbeen taken as the inverse of the square of the error associ-ated with its He + / H + value. For a given spectrum, we havecalculated means taking all possible combinations of the in-dividual lines without repetition, leaving at least 3 lines toaverage (for those spectra with more than 3 He i lines). Oncethis is done, of all the means, we consider those within the5th percentile of uncertainty and, from this subset, we finallyselect the mean for which more He i lines have been used forits calculation. This method excludes the most discrepantindividual lines but maintain the largest possible numberof them giving consistent values. The finally adopted mean He + / H + ratio of each spectrum is also included in Table A1.The uncertainty associated to that mean corresponds to theweighted standard deviation. In Table A2, we give the list MNRAS000
Atomic dataset used for collisionally excited lines of selected heavy-element ions.Transition probabilitiesIon and energy levels Collisional strengthsN + Froese Fischer & Tachiev (2004) Tayal (2011)O + Froese Fischer & Tachiev (2004) Kisielius et al. (2009)O + Wiese et al. (1996); Storey & Zeippen (2000) Storey et al. (2014)S + Podobedova et al. (2009) Tayal & Zatsarinny (2010)S + Podobedova et al. (2009) Grieve et al. (2014) i lines and Storey & Hummer (1995) for H i lines. The calculations by Porter et al. (2012) include colli-sional effects in the level populations of He . A Monte Carlosimulation similar to the one described in Section 3.1 – in-cluding random distributions of n e , T e and the line intensities– was applied to estimate the uncertainty of the He + / H + ra-tios derived for each individual line of each spectrum. The He + / H + ratios obtained for each of the He i lines of the spec-tra of the sample objects are shown in Table A1. Lines withintensity uncertainties greater than 40% and those affectedby blending with telluric lines or other spectral features werenot considered for abundance determinations. Moreover, the He + / H + ratios calculated from the bright He i λ λ S level on the predicted recombination spectrumof He i (Benjamin et al. 2002), that only affect noticeablyto triplet lines. Moreover, He i λ i λ He + / H + ratios of each spectrum. The weight of each line hasbeen taken as the inverse of the square of the error associ-ated with its He + / H + value. For a given spectrum, we havecalculated means taking all possible combinations of the in-dividual lines without repetition, leaving at least 3 lines toaverage (for those spectra with more than 3 He i lines). Oncethis is done, of all the means, we consider those within the5th percentile of uncertainty and, from this subset, we finallyselect the mean for which more He i lines have been used forits calculation. This method excludes the most discrepantindividual lines but maintain the largest possible numberof them giving consistent values. The finally adopted mean He + / H + ratio of each spectrum is also included in Table A1.The uncertainty associated to that mean corresponds to theweighted standard deviation. In Table A2, we give the list MNRAS000 , 1–15 (2015)
J. E. M´endez-Delgado et al.
Table 3.
Physical conditions and some heavy-element ionic abundances a for the sample objects. n e T e ([O iii ]) T e ([N ii ])Object Zone (cm − ) (K) (K) O + O + S + S + N + G2.4+1.4 A1 110 ±
50 8180 ± b ±
700 8.10 ± ± ±
160 10370 ±
990 8980 ±
350 7.90 ± ± ± ± ± ±
70 13520 ± ± ± ± ±
100 10850 ± ± ± ± ± ± ±
40 9300 ± b ±
920 7.96 ± ± ±
480 8040 ±
90 8360 ±
110 8.35 ± ± ± ± ±
220 7600 ±
210 8320 ±
140 8.46 ± ± ± ± ±
130 7970 ±
120 8900 ±
210 7.79 ± ± ± ± ±
60 7760 ±
250 8260 ±
130 8.48 ± ± ± ± ± ±
50 10140 ±
280 7.74 ± ± ± ± ±
220 9360 ±
170 8640 ± c ± ± ± ± ±
280 8440 ±
60 8760 ±
200 8.04 ± ± ± ± ±
910 9020 ±
140 11190 ±
550 7.35 ± ± ± ± ±
100 19040 ± ±
700 8.51 ± ± ± ±
80 12650 ±
410 7410 ±
90 8.31 ± ± ± ± ±
80 9970 ±
660 7580 ±
110 8.13 ± ± ± ± ± ±
90 10130 ±
970 8870 ±
140 7.46 ± ± ± ± ± ±
70 9970 ±
550 8740 ±
260 7.47 ± ± ± ± ± ±
110 9570 ±
780 8720 ±
180 7.54 ± ± ± ± ± ±
90 7240 ± b ±
160 8.32 ± ± ± ± ±
40 8030 ±
520 8040 ±
520 7.98 ± ± ± ± ±
50 7570 ±
610 8200 ±
380 8.08 ± ± ± ± ±
150 8710 ±
420 9050 ±
90 8.14 ± ± ± ± ±
100 7280 ± b ±
440 8.20 ± ± ± ± ±
100 7020 ± b ±
500 8.21 ± ± ± ± ±
70 5910 ± b ±
430 8.87 ± ± ± ± ±
100 4770 ± b ±
180 8.36 ± ± ± ± ± ±
90 10370 ±
370 11960 ±
640 7.15 ± ± ± ± ±
240 8250 ±
150 8610 ±
250 7.73 ± ± ± ± ±
100 9660 ± b ±
150 8.20 ± ± ± ± ±
100 9970 ±
340 10550 ±
240 7.84 ± ± ± ± ±
80 7360 ±
120 8200 ±
70 8.46 ± ± ± ± ±
290 10760 ± b ±
800 7.67 ± ± ± ± ±
100 11250 ±
970 8350 ±
770 8.16 ± ± ± ± ±
270 9200 ±
530 9430 ±
340 8.20 ± ± ± ± ±
100 11720 ±
210 11650 ±
490 8.15 ± ± ± ± ± ±
100 16600 ± ± ± ± ± ± ± ±
80 8940 ±
110 9270 ±
180 8.28 ± ± ± ± a In units of 12 + log( X n + / H + ). b T e ([O iii ]) estimated from T e ([N ii ]) using equation 3 of Esteban et al. (2009). c T e ([N ii ]) estimated from T e ([O iii ]) using equation 3 of Esteban et al. (2009). of individual He i lines used to derive the average value ofthe He + / H + ratio adopted for each spectrum.The usual methodology for calculating chemical abun-dances in H ii regions is to consider a zone of high and lowionization and adopt a representative T e for each zone. How-ever, this may be inappropriate for several ions, because theycan emit radiation in an intermediate zone or in both ioniza-tion zones. This is the case of He + , since it can emit in thehigh and low ionization zone. To calculate He + optimally,the characteristic temperature of the ion – T e (He i ) in thiscase – should be used or the precise geometry of the differentionization zones of each nebula should be known. However,the difference of He + / H + considering T e (He i ) and T e ([O iii ])is expected to be small given the small dependence on tem-perature of the intensity ratios of recombination lines. Thedifference between T e (He i ) and T e ([O iii ]) is important forthose interested in determining the primordial helium abun- dance, with the least possible uncertainty ( ≤ He + / H + ratio is accept-able, even more taking into account that statistically thoseuncertainties will not be reflected in the determination of agradient based on data of several objects and whose mainsource of uncertainty is not He + / H + but the He / H + ratio.Nevertheless, we have explored the effects that T e and op-tical depth effects of the metastable 2 S level – τ ( S ) butparametrized by τ (3889) – would have on the He + / H + ratios.Figure 2 presents the comparison between the He + / H + ratios obtained from the method followed in this work andthose obtained using Helio14 with the list of lines pre-sented in Table A2 and T e (O ii +O iii ), derived followingPeimbert et al. (2002). T e (O ii +O iii ) takes into account theemission of He i in the low ionization zone. Helio14 code isan updated version of
Helio10 presented by Peimbert et al.
MNRAS , 1–15 (2015) elium abundances and gradient in Galactic nebulae (2012). The code uses the effective recombination coefficientsof Storey & Hummer (1995) for H i , and Porter et al. (2007)for He i , the collisional contribution of He i lines calculatedby Sawey & Berrington (1993) and the optical depth effectsin the triplets estimated by Kingdon & Ferland (1995). He-lio14 calculates the most likely values for n e (He i ), τ ( S ) and the He + / H + ratio from the theoretical He i /H β ratiosusing as input a set of parameters, atomic data, and up to20 He i /H β line intensity ratios along with their uncertain-ties. Then, it compares the observed ratios to the theoreticalones minimising χ : χ = Õ λ nh I ( λ ) I ( H β ) i obs − h I ( λ ) I ( H β ) i theo o σ I ( λ ) . (1)Where σ I ( λ ) is the uncertainty in the intensity of λ .The comparison of Figure 2 shows that all the objects areentirely consistent with the 1:1 relation. The nebulae withthe largest deviations with respect the 1:1 line are RCW 52,Sh 2-209 and Sh 2-308, although the values agree with suchrelation within the uncertainties. These three objects areamong the ones with fewer He i lines in their spectra: 4, 4 and2, respectively, so the determination of the abundance of He i is very sensitive to the variations of the fitted parametersin the minimisation of χ . Helio14 determines the availableparameter space within χ min < χ < χ min + to estimate the1 σ error bars. In any case, from these 3 objects, just RWC 52was taken into account in the determination of radial He i gradients. The rest are ring nebula associated Wolf-Rayetstars and they present an overabundance of helium. Thisis discussed in greater detail in §
5. The general deviationbetween the values of He + / H + ratios obtained with Helio14 and those obtained following the procedure described in thissection is less than 0.008 dex even including the 3 objectsshowing the largest differences.From all the spectra in our sample, we have deconvolvedHe i λ i λ i λ i λ i λ β with uncertainties smaller than 3%. In these last cases, westudy the possible effect of τ (3889) on the determination ofthe He + / H + ratios. A small uncertainty in the measurementof He i λ τ (3889) and consequently in the abundanceof He + / H + .In the nebulae where both lines cannot be deblendedwith multiple Gaussian fitting, the contribution of He i λ i λ He + / H + ratios for Sh 2-152, NGC 3576, NGC 7635-A4, M16, M17, M20 and M42 using Helio14 with the sameinput parameters as in Figure 2 but taking into account τ (3889) and the results we obtain with our methodology.The general dispersion is smaller than 0.007 dex. From thiscomparison, we can conclude that our results can be consid-ered virtually independent of the effects of τ ( S ) . In fact,this is minimised by averaging He + / H + ratios from a large + /H + ) This work + l og ( H e + / H + ) H e li o Figure 2.
General comparison of the average He + / H + ratios ob-tained for the sample objects using our methodology described inthe text and using the Helio14 code (Peimbert et al. 2012). Thecontinuous line indicates the 1:1 relation. + /H + ) This work + l og ( H e + / H + ) H e li o Figure 3.
Comparison between He + / H + determinations after es-timating the effects of optical depth ( τ (3889)) in 7 objects: Sh 2-152, NGC 3576, NGC 7635-A4, M16, M17, M20 and M42. Thecontinuous line indicates the 1:1 relation. number of emission lines, including several singlets, whichintensities are practically independent of τ ( S ) .Among the brightest He i lines, λ λ λ i lines are λ λ λ λ λ λ λ i lines that have a greater ten-dency to provide less consistent values with respect to theaverage He + / H + ratio are: λ λ λ λ λ λ λ He + abundances ob-tained from them are similar to those obtained from othermore intense lines and they are considered to compute theaverage value of the He + / H + ratio. The He i λ iii ] λ MNRAS , 1–15 (2015)
J. E. M´endez-Delgado et al. the UVES ones. Another bright He i line is λ He + / H + ratios lower than the mean in G2.4+1.4,NGC 6888 and NGC 7635, the objects with the largest num-ber of spectra. The whole optical OSIRIS spectra of theseobjects were taken with two grisms, R1000B and R2500V.Since the He i λ The presence of He inside the H + zone may be signif-icant in low ionization nebulae. We need to assume anICF(He) to determine the total helium abundance fromthe measured He + / H + ratio, in the form: He/H = ICF(He) · He + / H + . However, the estimation of the ICF(He) is acontroversial issue and there is no a consensus amongthe different authors. Traditional ICF(He) schemes involveabundance ratios of different heavy-element ions such as O + , O + , S + or S + . For the H ii regions of our sam-ple, we have used the values of O + , O + , S + and S + abundances determined by Arellano-C´ordova et al. (2020),who use pyneb and the atomic data given in Table 2.In addition, we have re-calculated the N + abundancesof the WR ring nebulae using the same methodology,their N abundances will be discussed in §
5. The line in-tensities used to derive the heavy-element ionic abun-dances have been [O ii ] λλ , , [O ii ] λλ , , [S ii ] λλ , , [S iii ] λ iii ] λλ ii ] λλ , . Their reddening-corrected values have been taken from the set of sourcepapers. Arellano-C´ordova et al. (2020) use T e ([N ii ]) as rep-resentative of the O + , S + and N + emitting regions and T e ([O iii ]) as representative of the O ++ zone. Followingthe recommendation of Dom´ınguez-Guzm´an et al. (2019),Arellano-C´ordova et al. (2020) adopt the mean value of T e ([N ii ]) and T e ([O iii ]) to calculate the S ++ abundance.In the sake of homogeneity and as we did for determiningthe physical conditions, we used the same methodology andatomic data as Arellano-C´ordova et al. (2020) for derivingthe O + , O + , S + , S + or N + abundances for the ring nebulae.The heavy-element ionic abundances determined for eachspectrum and object are included in Table 3.Photoionization models (e.g. Stasi´nska & Schaerer1997) predict that helium is completely ionized inside the H + zone when the exciting star is hotter than about 39 000K(earlier than O6.5V), independently of the ionization pa-rameter. Most of the ionizing stars of our sample are ionizedby stars that should be colder, so one would expect a cer-tain amount of He in most of our nebulae. On the otherhand, observations in different directions inside H ii regionscarried out by Deharveng et al. (2000) show that even anH ii region excited by a star of spectral type earlier thanO6.5 can contain a significant amount of He . We have useddifferent ICF(He) schemes because, as we have said before,there is not a standard method to correct for the fractionof He , although all schemes consider that the higher theionization degree the higher the He + /He ratio of the nebu-lae. ICF(He) schemes usually consider the similarity betweenthe ionization potential of He (24.59 eV) and those of S + (23.33 eV) and/or O + (34.97 eV) assuming that the relation O + /O > He /He > S + /S should be applicable but using dif-ferent parameters proportional to the ionization degree ofthe nebula. One of the first ICF(He) schemes was proposedby Peimbert & Torres-Peimbert (1977) in their work on thechemical composition of the Orion Nebula that uses a linearcombination of O + /O and S + /S ratios, ICF ( He ) PTP77 = (cid:20) − γ · O + O − ( − γ ) · S + S (cid:21) − . (2)The value of γ depends on the density distribution(Peimbert et al. 1974). It can be determined when spectraat different positions covering different ionization conditionsare available for the nebula, but this is not the case for mostof our sample objects. The only objects with several slitpositions are all ring nebulae and bubbles, that are bet-ter represented by relatively thin ionized shells rather thantypical Str ˜A˝umgren spheres and, therefore, that methodfor estimating the parameter γ may not be appropriate.Peimbert & Torres-Peimbert (1977) estimated γ = 0.35 inthe case of the Orion Nebula and Peimbert et al. (1978)found γ = 0.20 for η Carina nebula. Those last authors as-sume the same value of 0.20 for other H ii regions with n e values similar to that of η Carina nebula (300-1000 cm − ).Lequeux et al. (1979) use equation 2 for determining the Heabundance for a sample of metal-poor H ii galaxies with n e in the 10-100 cm − range using γ = 0.15. In our case, follow-ing the prescriptions given in the cited works, we have usedequation 2 using values of γ interpolating between 0.15 to0.35 as a function of the n e of the nebulae.Other ICF(He) schemes only use O + /O or S + /S ra-tios as indicators of the ionization degree of the nebulae.Kunth & Sargent (1983), in their study of a sample of metal-poor galaxies, proposed an ICF(He) dependent on the O + /Oratio deduced from Peimbert et al. (1974) empirical models, ICF ( He ) KS83 = (cid:20) − . · O + O (cid:21) − . (3)Peimbert et al. (1992) presented a detailed spectroscopicalstudy of the highly-ionized Galactic H ii region M17 and pro-pose an ICF(He) parameterized only by the S + /S ionizationfraction. Zhang & Liu (2003) use a very similar scheme intheir study of the relatively low excitation planetary nebulaM 2-24 but depending on the S + / S + ratio, which – in con-trast to S + /S – is a parameter that can be obtained directlyfrom optical spectra: ICF ( He ) ZL03 = + S + S + . (4)Delgado-Inglada et al. (2014) do not recommend theuse of traditional ICF(He) schemes – as those given in equa-tions 3 and 4 – because the He / He + ratio is more dependenton the effective temperature of the ionizing source, whereasheavy-element abundance ratios depend essentially on theionization parameter. Vilchez (1989) proposed an ICF(He)scheme based on the radiation softness parameter defined as η = ( O + / S + ) · ( S + / O + ) (V´ılchez & Pagel 1988), which issensitive to the effective temperature of the ionizing sourceand a good indicator of the ionization structure of the neb-ula. Vilchez (1989) and Pagel et al. (1992) pointed out thatfor log( η ) < He inside the H + zone isnegligible for a large variety of models, but becomes very MNRAS , 1–15 (2015) elium abundances and gradient in Galactic nebulae Table 4.
Total He abundances a for the sample objects using different ICF schemes.ICF(He) scheme b Object PTP77 ZL03 KS83 B09G2.4+1.4 11.038 ± c M8 10.963 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± c NGC 7635 11.051 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± a In units of 12 + log(He/H). b PTP77: Peimbert & Torres-Peimbert (1977); ZL03: Zhang & Liu (2003); KS83: Kunth & Sargent (1983); B09: Bresolin et al. (2009). c He/H = He + / H + + He + / H + for this object. No ICF(He) applied. model dependent for log( η ) >
1. Izotov et al. (1994) andBresolin et al. (2009) have obtained approximate analyticalexpressions of ICF(He) as a function of η parameter basedon photoionization models. In our case, we have used the re-lation given by Bresolin et al. (2009) in their study of H ii re-gions in the spiral galaxy NGC 300 and based on the predic-tions of photoionization models by Stasi´nska et al. (2001), ICF ( He ) B09 = . + log ( η ) · [ . · log ( η ) − . ] . (5)In the case of the ring nebulae for which we have spectraat several slit positions and, therefore, several independentdeterminations of the He abundance, we have selected therepresentative He/H ratio of the nebula as a whole attend-ing to different considerations in each object. In the case ofG2.4 + T e and may be contaminated by shock excitation(see Esteban et al. 2016). For NGC 7635 we have taken theweighted mean value of zones A2, A3, A5 and A6, which cor-respond to the relatively highly-ionized gas of the expand-ing bubble. Zone A1 of NGC 7635 corresponds to one of thebrightest knots of Sh 2-162, a large H ii region that encom-passes the ring nebula NGC 7635. Sh 2-162 is ionized by thesame O-type star but shows a lower ionization degree. ZoneA4 covers the brightest of a group of several high-density cometary knots just at the southwest of the ionizing starthat show rather higher densities and quite lower ionizationdegrees.The WR ring nebulae G2.4 + ii lines in their spectra (Esteban et al. 2016). G2.4 + ii λ He + abundance using pyneb , T e ([O iii ])and the effective recombination coefficient calculated byStorey & Hummer (1995), obtaining 12 + log( He + / H + ) =10.45 ± + He + abundances we determine for the dif-ferent zones is very much lower, with values ranging from8.14 to 8.51, which are clearly negligible in comparison tothe He + / H + ratio determined in this object. We did not usean ICF(He) for G2.4 + He + / H + + He + / H + .In Table 4 we show the He/H ratios obtained for eachnebulae using the different ICF(He) schemes considered. Inthe case of the ICF(He) PTP77 , there is an explicit dependenceof the S/H ratio, which has been derived from the sum of the S + and S + abundances given in Table 3 and an appropriateICF(S) using the relations: SH = ICF ( S ) · (cid:20) S + H + + S + H + (cid:21) , (6)where ICF ( S ) = " − (cid:18) − O + O (cid:19) − / , (7) MNRAS000
1. Izotov et al. (1994) andBresolin et al. (2009) have obtained approximate analyticalexpressions of ICF(He) as a function of η parameter basedon photoionization models. In our case, we have used the re-lation given by Bresolin et al. (2009) in their study of H ii re-gions in the spiral galaxy NGC 300 and based on the predic-tions of photoionization models by Stasi´nska et al. (2001), ICF ( He ) B09 = . + log ( η ) · [ . · log ( η ) − . ] . (5)In the case of the ring nebulae for which we have spectraat several slit positions and, therefore, several independentdeterminations of the He abundance, we have selected therepresentative He/H ratio of the nebula as a whole attend-ing to different considerations in each object. In the case ofG2.4 + T e and may be contaminated by shock excitation(see Esteban et al. 2016). For NGC 7635 we have taken theweighted mean value of zones A2, A3, A5 and A6, which cor-respond to the relatively highly-ionized gas of the expand-ing bubble. Zone A1 of NGC 7635 corresponds to one of thebrightest knots of Sh 2-162, a large H ii region that encom-passes the ring nebula NGC 7635. Sh 2-162 is ionized by thesame O-type star but shows a lower ionization degree. ZoneA4 covers the brightest of a group of several high-density cometary knots just at the southwest of the ionizing starthat show rather higher densities and quite lower ionizationdegrees.The WR ring nebulae G2.4 + ii lines in their spectra (Esteban et al. 2016). G2.4 + ii λ He + abundance using pyneb , T e ([O iii ])and the effective recombination coefficient calculated byStorey & Hummer (1995), obtaining 12 + log( He + / H + ) =10.45 ± + He + abundances we determine for the dif-ferent zones is very much lower, with values ranging from8.14 to 8.51, which are clearly negligible in comparison tothe He + / H + ratio determined in this object. We did not usean ICF(He) for G2.4 + He + / H + + He + / H + .In Table 4 we show the He/H ratios obtained for eachnebulae using the different ICF(He) schemes considered. Inthe case of the ICF(He) PTP77 , there is an explicit dependenceof the S/H ratio, which has been derived from the sum of the S + and S + abundances given in Table 3 and an appropriateICF(S) using the relations: SH = ICF ( S ) · (cid:20) S + H + + S + H + (cid:21) , (6)where ICF ( S ) = " − (cid:18) − O + O (cid:19) − / , (7) MNRAS000 , 1–15 (2015) J. E. M´endez-Delgado et al. which was proposed by Stasi´nska (1978). The uncertaintyof the total He abundance has been determined propagat-ing the errors of the He + / H + ratio and that of the ICF(He)– and ICF(S) – used, which comes from the derivation ofeach equation of the ICF and the quoted errors of the ionicabundances given in Table 3. In Fig. 4 we show the radial distribution of the He abun-dances of the H ii regions of our sample – a total of 19 ob-jects – using the different ICF(He) schemes introduced inequations 2 to 5. We have also included the ring nebulaearound O-type objects in this group because they do notcontain chemically enriched material from the central star(see Esteban et al. 2016). In these diagrams we have sepa-rated the nebulae in two groups, those with log( η ) < He – rep-resented with full black squares and the rest of them (greyempty squares). This separation is made in order to explorethe radial distribution of the He abundances minimizing theeffect of the ICF(He) and maintaining a reasonable numberof objects distributed along a significant part of the Galacticdisc. The least-quares linear fits to the objects with log( η ) < R G and the He/H ratios for each observational pointassuming a Gaussian distribution with a sigma equal to theuncertainty of each quantity. We performed a least-squareslinear fit to each of these random distributions. It is impor-tant to remark that the uncertainty of R G has been consid-ered in the fittings, which is not usually taken into accountin most works except, for example, in Esteban et al. (2017)or Esteban & Garc´ıa-Rojas (2018). The uncertainties asso-ciated to the slope and intercept correspond to the standarddeviation of the values of these two quantities obtained fromthe fits. From the radial gradients represented in Fig. 4 andtheir parameters collected in Table 5, we can see that al-though the uncertainties of the slopes are larger than theslope values in most cases, they show pretty similar valuesin all cases, especially when only the objects with log( η ) < ii regions withlog( η ) < He + / H + .Each data point is represented with a colour proportional toits log( η ), and the diagram indicates that the value of the η parameter does not seem to determine the behaviour of thegradient. In fact, the objects with the lowest values of log( η )(i.e. those with the lowest expected fraction of He ) are pre-cisely those located at larger R G showing lower He + / H + ra-tios. In Fig. 6, we can compare the least-squares linear fits ofthe He abundances obtained for the H ii regions with log( η ) < − − − , with adispersion of values smaller than the typical uncertainty ofeach individual determination of the slope. The differencesare larger when we compare the values of the He/H givenby each fit for a given R G . The no-ICF(He) case gives thelowest values of the He/H ratio – as expected – and theICF(He) by Peimbert & Torres-Peimbert (1977) gives thelargest ones. It is interesting to note that the other ICF(He)schemes give He/H values that differ less than 0.02 dex. It isremarkable that the ICF scheme from Bresolin et al. (2009)is the most sensitive to the propagation of the errors. Its userequires very precise determinations of O + , S + , O + and S + .The presence of abundance gradients of heavy el-ements – such as O, N, Ne, S, Ar or Cl – alongthe disc of the galaxy is a well established fact fromthe observational point of view (e.g. Shaver et al. 1983;Deharveng et al. 2000; Rudolph et al. 2006; Balser et al.2011; Esteban & Garc´ıa-Rojas 2018). However, althoughtheoretical models predict its existence, this has not beenthe case for He. Very recent cosmological chemodynamicalsimulations for galaxy formation and evolution predict neg-ative radial gradients of He/H due to the inside-out growthof galaxy discs as a function of time (Vincenzo et al. 2019).Previous chemical evolution models of the Milky Way disc byMatteucci & Chiappini (1999) and Chiappini et al. (2002)predicted He/H gradients with slopes between − − − in the R G range from 4 to 18 kpc.Kubryk et al. (2015) obtain a similar range of values (fromabout − − − ) from models includingradial motions of gas and stars in the Milky Way disc anddifferent sets of stellar yields. These values are pretty consis-tent with our results given in Table 5. From the observationalpoint of view, the evidence of a Galactic radial gradient ofthe He/H has been elusive. Peimbert et al. (1978) found anegative gradient of − ± − from the anal-ysis of the abundance obtained from 3 He i lines in a smallsample of 5 Galactic H ii regions. Talent & Dufour (1979)obtain a marginal evidence of a negative gradient, − ± − for a larger sample of objects based solely onthe intensity of He i λ λ i lines in samples with different numberof objects. In their work, Fern´andez-Mart´ın et al. (2017) in-clude some of the H ii regions of our sample (Sh2-83, Sh2-212and NGC 7635), whose data are based on other spectroscopicobservations. The He + / H + ratios per line shown in theirTable A.2. are consistent with our results, although thereare slight differences attributable to the different recombi-nation coefficients used. However, the methodology followedby Fern´andez-Mart´ın et al. (2017) to adopt a final He + / H + value differs from ours, among other things, by the use offewer He i lines and the inclusion of He i λ ii regionsnot considered in the present paper, we decided not to in-clude them for the sake of having a group of nebulae witha homogeneous data reduction process; furthermore, the R G of those objects are well covered by the H ii regions of oursample.Other determinations of the radial He/H gradient in theMilky Way come from the analysis of the emission-line spec- MNRAS , 1–15 (2015) elium abundances and gradient in Galactic nebulae Table 5.
Radial He/H gradients using different combinations of H ii regions and ICF schemes.All H ii regions Only with log( η ) < − ) Intercept Slope (dex kpc − ) InterceptPTP77 − ± ± − ± ± − ± ± − ± ± − ± ± − ± ± − ± ± − ± ± −− −− − ± ± R G (Kpc) + l og ( H e / H ) R G (Kpc) + l og ( H e / H ) R G (Kpc) + l og ( H e / H ) R G (Kpc) + l og ( H e / H ) Figure 4.
Radial distribution of the He abundances of the H ii regions of our sample using different ICF(He) schemes. Upper left: usingthe ICF(He) proposed by Peimbert & Torres-Peimbert (1977, PTP77); upper right: ICF(He) by Zhang & Liu (2003, ZL03); lower left:ICF(He) by Kunth & Sargent (1983, KS83); lower right: ICF(He) by Bresolin et al. (2009, B09). Black full squares indicate H ii regionswith log( η ) < ii regions with log( η ) ≥ η ) < tra of Planetary Nebulae (PNe). Early works such as thoseby D’Odorico et al. (1976), Peimbert & Serrano (1980) orFa´undez-Abans & Maciel (1987) found slopes between − − − , while others reported negligible gradi-ents (e.g. Pasquali & Perinotto 1993). At any rate, PNe arenot confident probes of the abundances of the ISM becausethey are composed of stellar ejecta material and can be con-taminated by the products of nucleosynthesis. Trying to al-leviate this drawback, Maciel (2001) introduced correctionsto the measured He abundance owing to the contaminationfrom the progenitor star obtaining an essentially flat radialHe/H gradient for a large sample of PNe. Ring nebulae are bubbles of gas swept-up by the mechan-ical action of the mass loss episodes experienced by theirmassive stellar progenitors. They are rather common aroundWolf-Rayet (WR) stars and luminous blue variables (LBVs). There have been several spectroscopical works devoted tostudy the chemical composition of the ionized gas containedin ring nebulae around Galactic WR stars (e.g. Kwitter 1984;Esteban et al. 1992, 2016; Stock et al. 2011). Those workshave found that some ring nebulae (NGC 6888, RCW 58,Sh 2-308 and M 1-67) show clear chemical enrichment pat-terns, in general they are He and N enriched and someof them show some O deficiency, indicating that they arecomposed by ejecta material that has suffered contami-nation by the CNO cycle. Esteban & Vilchez (1992) andEsteban et al. (1992) compared the abundance pattern ofGalactic ejecta ring nebulae with the surface abundancespredicted by evolutionary models of massive stars, findingthat the He, N and O abundances of the nebulae are con-sistent with the expected surface composition of stars withinitial masses between 25 and 40 M ⊙ at the red supergiant(RSG) phase, prior to the onset of the WR stage. Later stud-ies by Mesa-Delgado et al. (2014) and Esteban et al. (2016)using stellar evolution models by Ekstr¨om et al. (2012) andGeorgy et al. (2012) confirmed that non-rotational modelsof stars of initial masses between 25 and 40 M ⊙ seem to MNRAS000
Radial distribution of the He abundances of the H ii regions of our sample using different ICF(He) schemes. Upper left: usingthe ICF(He) proposed by Peimbert & Torres-Peimbert (1977, PTP77); upper right: ICF(He) by Zhang & Liu (2003, ZL03); lower left:ICF(He) by Kunth & Sargent (1983, KS83); lower right: ICF(He) by Bresolin et al. (2009, B09). Black full squares indicate H ii regionswith log( η ) < ii regions with log( η ) ≥ η ) < tra of Planetary Nebulae (PNe). Early works such as thoseby D’Odorico et al. (1976), Peimbert & Serrano (1980) orFa´undez-Abans & Maciel (1987) found slopes between − − − , while others reported negligible gradi-ents (e.g. Pasquali & Perinotto 1993). At any rate, PNe arenot confident probes of the abundances of the ISM becausethey are composed of stellar ejecta material and can be con-taminated by the products of nucleosynthesis. Trying to al-leviate this drawback, Maciel (2001) introduced correctionsto the measured He abundance owing to the contaminationfrom the progenitor star obtaining an essentially flat radialHe/H gradient for a large sample of PNe. Ring nebulae are bubbles of gas swept-up by the mechan-ical action of the mass loss episodes experienced by theirmassive stellar progenitors. They are rather common aroundWolf-Rayet (WR) stars and luminous blue variables (LBVs). There have been several spectroscopical works devoted tostudy the chemical composition of the ionized gas containedin ring nebulae around Galactic WR stars (e.g. Kwitter 1984;Esteban et al. 1992, 2016; Stock et al. 2011). Those workshave found that some ring nebulae (NGC 6888, RCW 58,Sh 2-308 and M 1-67) show clear chemical enrichment pat-terns, in general they are He and N enriched and someof them show some O deficiency, indicating that they arecomposed by ejecta material that has suffered contami-nation by the CNO cycle. Esteban & Vilchez (1992) andEsteban et al. (1992) compared the abundance pattern ofGalactic ejecta ring nebulae with the surface abundancespredicted by evolutionary models of massive stars, findingthat the He, N and O abundances of the nebulae are con-sistent with the expected surface composition of stars withinitial masses between 25 and 40 M ⊙ at the red supergiant(RSG) phase, prior to the onset of the WR stage. Later stud-ies by Mesa-Delgado et al. (2014) and Esteban et al. (2016)using stellar evolution models by Ekstr¨om et al. (2012) andGeorgy et al. (2012) confirmed that non-rotational modelsof stars of initial masses between 25 and 40 M ⊙ seem to MNRAS000 , 1–15 (2015) J. E. M´endez-Delgado et al.
Table 6.
He, O and N abundance deviations with respect to abundance gradients in WR ring nebulae.Object 12+log(He/H) ∆ (He/H) 12+log(O/H) ∆ (O/H) 12+log(N/H) ∆ (N/H)G2.4+1.4 11.04 ± + ± ± − ± ± + ± ± + ± ± − ± ± + ± ± + ± ± + ± ± + ± ± + ± ± + ± ± + ± ± + ± ± − ± + Figure 5.
Radial distribution of the He abundances of the H ii regions with log( η ) < He + / H + ). The colour of the squares indicates the value of log( η )of each object. The solid lines represent the least-quares linearfits to the data. R G (Kpc) + l og ( H e / H ) ZL03KS83PTP77B09ICF=1.0
Figure 6.
Comparison of the least-quares linear fits of the Heabundances obtained for the H ii regions with log( η ) < reproduce the abundance patters in most of the ejecta neb-ulae, being this range wider in the case of Sh 2-308, from 25to 50 M ⊙ . Only rotational models of 25 M ⊙ show agreementwith the data for NGC 6888, RCW 58 and Sh 2-308.The availability of new Gaia distances, the results ofthis paper and the recent reassessment of the Galactic ra-dial gradients of O/H and N/H by Esteban & Garc´ıa-Rojas(2018), allow us to perform a better estimate of the chemicalenrichment pattern in He, N and O of the WR ring nebu-lae observed by our group (Esteban et al. 2016). In Table 6we give the total He/H, O/H and N/H ratios and the dif-ference with the values expected from the radial abundance gradients for the sample of WR ring nebulae included in Ta-ble 1. The He abundances of each object have been takenfrom Table 4. In the case of RCW 58, Sh 2-298 and Sh 2-308, we have considered the mean and standard deviation ofthe 4 values obtained using the different ICF(He) schemes.The O abundances have been calculated simply adding the O + / H + and O + / H + ratios given in Table 3, except in thecase of G2.4 + He + / H + . In this case, we have considered theICF(O) proposed by Delgado-Inglada et al. (2014) for deter-mining the O/H ratio. The N/H ratio has been calculatedusing the N + abundance given in Table 3 and the classicalICF(N) by Peimbert & Costero (1969), that assumes N/O= N + / O + .In Table 6 we can see that G2.4 + O + / H + ratio. The overabundance ofHe/H is very similar in the three objects, of about +0.24 ± We determine the radial abundance gradient of helium ofthe Milky Way from published spectra of H ii regions. Thedata set is the largest collection of deep spectra of GalacticH ii regions available. We also include data of similar qual-ity for several ring nebulae surrounding massive O and WRstars. The total number of nebulae included in the sampleis 24. We have revised the Galactocentric distances of theobjects, R G , considering Gaia
DR2 parallaxes and previouskinematic and spectroscopical determinations. We have de-termined the physical conditions – n e and T e – and the ionicabundance of He + and other selected ions in a homogeneousway and using the most recent atomic data sets. We havedetermined the He + abundance using several He i recombi-nation lines. In the case of the H ii regions, we have usedbetween 3 and 10 individual He i lines depending on the ob-ject, selecting only those well-measured lines that are notaffected by line-blending, telluric contamination or impor-tant self-absorption effects. The total He abundance of theobjects have been estimated using four different ICF(He) MNRAS , 1–15 (2015) elium abundances and gradient in Galactic nebulae schemes based on different ionic ratios. We have determinedthe Galactic radial He abundance gradient using the resultsof 19 objects, including H ii regions and ring nebulae aroundO-type stars. We find that although the uncertainties of theslopes are larger than the slope values obtained using severalof the four ICF(He) schemes, they show consistent values inall cases, especially when only the objects with log( η ) < − − − , con-sistent with the predictions of different chemical evolutionmodels of the Milky Way and chemodynamical simulationsof galactic discs. We have estimated the abundance devia-tions of He, O and N of the ring nebulae around WR starsincluded in our sample with respect to the radial gradients,finding that only NGC 6888, RCW 585 and Sh 2-308 canbe considered ejecta nebulae. These objects show He andN overabundances. The degree of enrichment of He is verysimilar and about +0.24 ± ACKNOWLEDGEMENTS
We thank the referee ´Angeles I. D´ıaz for a constructivereport. We acknowledge support from the State ResearchAgency (AEI) of the Spanish Ministry of Science, Innova-tion and Universities (MCIU) and the European RegionalDevelopment Fund (FEDER) under grant with referenceAYA2015-65205-P. JG-R acknowledges support from anAdvanced Fellowship from the Severo Ochoa excellenceprogram (SEV-2015-0548). The authors acknowledge sup-port under grant P/308614 financed by funds transferredfrom the Spanish Ministry of Science, Innovation andUniversities, charged to the General State Budgets andwith funds transferred from the General Budgets of the Au-tonomous Community of the Canary Islands by the MCIU.KZA-C acknowledges support from Mexican CONACYTposdoctoral grant 364239. JEM-D thanks the Fundaci´onCarolina for the support provided for his Master’s stud-ies. JEM-D also thanks the support of the Instituto deAstrof´ısica de Canarias under the Astrophysicist ResidentProgram and acknowledges support from the MexicanCONACyT (grant CVU 602402). This work has made useof data from the European Space Agency (ESA) mission
Gaia ( ), processed bythe Gaia
Data Processing and Analysis Consortium (DPAC, ).Funding for the DPAC has been provided by national insti-tutions, in particular the institutions participating in the
Gaia
Multilateral Agreement.
DATA AVAILABILITY
This research is based on public data available in the refer-ences. The whole dataset can be obtained from the authorsby request. All the results are available in the tables and inthe appendix of this article.
REFERENCES
Anderson L. D., Armentrout W. P., Johnstone B. M., BaniaT. M., Balser D. S., Wenger T. V., Cunningham V., 2015,ApJS, 221, 26Arellano-C´ordova K. Z., Esteban C., Garc´ıa-Rojas J., M´endez-Delgado J. E., 2020, arXiv e-prints, p. arXiv:2005.11372Avedisova V. S., Kondratenko G. I., 1984, Nauchnye Informatsii,56, 59Aver E., Olive K. A., Skillman E. D., 2015,J. Cosmology Astropart. Phys., 2015, 011Bailer-Jones C. A. L., Rybizki J., Fouesneau M., Mantelet G.,Andrae R., 2018, AJ, 156, 58Balser D. S., Rood R. T., Bania T. M., Anderson L. D., 2011,ApJ, 738, 27Benjamin R. A., Skillman E. D., Smits D. P., 2002, ApJ, 569, 288Binder B. A., Povich M. S., 2018, ApJ, 864, 136Bland-Hawthorn J., Gerhard O., 2016, ARA&A, 54, 529Bohigas J., Tapia M., 2003, AJ, 126, 1861Bresolin F., Gieren W., Kudritzki R.-P., Pietrzy´nski G., UrbanejaM. A., Carraro G., 2009, ApJ, 700, 309Brocklehurst M., 1971, MNRAS, 153, 471Caplan J., Deharveng L., Pe˜na M., Costero R., Blondel C., 2000,MNRAS, 311, 317Cepa J., et al., 2000, in Iye M., Moorwood A. F., eds, Society ofPhoto-Optical Instrumentation Engineers (SPIE) ConferenceSeries Vol. 4008, Optical and IR Telescope Instrumentationand Detectors. pp 623–631Cepa J., et al., 2003, in Iye M., Moorwood A. F. M., eds, Societyof Photo-Optical Instrumentation Engineers (SPIE) Confer-ence Series Vol. 4841, Instrument Design and Performance forOptical/Infrared Ground-based Telescopes. pp 1739–1749Chiappini C., Renda A., Matteucci F., 2002, A&A, 395, 789Chini R., Wink J. E., 1984, A&A, 139, L5Coc A., Angulo C., Vangioni-Flam E., Descouvemont P., Adah-chour A., 2005, Nuclear Phys. A, 752, 522Cooke R. J., Fumagalli M., 2018, Nature Astronomy, 2, 957Copetti M. V. F., Oliveira V. A., Riffel R., Casta˜neda H. O.,Sanmartim D., 2007, A&A, 472, 847D’Odorico S., Peimbert M., Sabbadin F., 1976, A&A, 47, 341D’Odorico S., Cristiani S., Dekker H., Hill V., Kaufer A., KimT., Primas F., 2000, in Bergeron J., ed., Society of Photo-Optical Instrumentation Engineers (SPIE) Conference SeriesVol. 4005, Society of Photo-Optical Instrumentation Engi-neers (SPIE) Conference Series. pp 121–130Deharveng L., Pe˜na M., Caplan J., Costero R., 2000, MNRAS,311, 329Delgado-Inglada G., Morisset C., Stasi´nska G., 2014, MNRAS,440, 536Dom´ınguez-Guzm´an G., Rodr´ıguez M., Esteban C., Garc´ıa-RojasJ., 2019, arXiv e-prints, p. arXiv:1906.02102Drew J. E., Mongui´o M., Wright N. J., 2019, MNRAS, 486, 1034Ekstr¨om S., et al., 2012, A&A, 537, A146Esteban C., Garc´ıa-Rojas J., 2018, MNRAS, 478, 2315Esteban C., Vilchez J. M., 1992, ApJ, 390, 536Esteban C., Vilchez J. M., Smith L. J., Clegg R. E. S., 1992,A&A, 259, 629Esteban C., Peimbert M., Garc´ıa-Rojas J., Ruiz M. T., PeimbertA., Rodr´ıguez M., 2004, MNRAS, 355, 229Esteban C., Bresolin F., Peimbert M., Garc´ıa-Rojas J., PeimbertA., Mesa-Delgado A., 2009, ApJ, 700, 654Esteban C., Carigi L., Copetti M. V. F., Garc´ıa-Rojas J., Mesa-Delgado A., Casta˜neda H. O., P´equignot D., 2013, MNRAS,433, 382Esteban C., Mesa-Delgado A., Morisset C., Garc´ıa-Rojas J., 2016,MNRAS, 460, 4038Esteban C., Fang X., Garc´ıa-Rojas J., Toribio San Cipriano L.,2017, MNRAS, 471, 987MNRAS000
Anderson L. D., Armentrout W. P., Johnstone B. M., BaniaT. M., Balser D. S., Wenger T. V., Cunningham V., 2015,ApJS, 221, 26Arellano-C´ordova K. Z., Esteban C., Garc´ıa-Rojas J., M´endez-Delgado J. E., 2020, arXiv e-prints, p. arXiv:2005.11372Avedisova V. S., Kondratenko G. I., 1984, Nauchnye Informatsii,56, 59Aver E., Olive K. A., Skillman E. D., 2015,J. Cosmology Astropart. Phys., 2015, 011Bailer-Jones C. A. L., Rybizki J., Fouesneau M., Mantelet G.,Andrae R., 2018, AJ, 156, 58Balser D. S., Rood R. T., Bania T. M., Anderson L. D., 2011,ApJ, 738, 27Benjamin R. A., Skillman E. D., Smits D. P., 2002, ApJ, 569, 288Binder B. A., Povich M. S., 2018, ApJ, 864, 136Bland-Hawthorn J., Gerhard O., 2016, ARA&A, 54, 529Bohigas J., Tapia M., 2003, AJ, 126, 1861Bresolin F., Gieren W., Kudritzki R.-P., Pietrzy´nski G., UrbanejaM. A., Carraro G., 2009, ApJ, 700, 309Brocklehurst M., 1971, MNRAS, 153, 471Caplan J., Deharveng L., Pe˜na M., Costero R., Blondel C., 2000,MNRAS, 311, 317Cepa J., et al., 2000, in Iye M., Moorwood A. F., eds, Society ofPhoto-Optical Instrumentation Engineers (SPIE) ConferenceSeries Vol. 4008, Optical and IR Telescope Instrumentationand Detectors. pp 623–631Cepa J., et al., 2003, in Iye M., Moorwood A. F. M., eds, Societyof Photo-Optical Instrumentation Engineers (SPIE) Confer-ence Series Vol. 4841, Instrument Design and Performance forOptical/Infrared Ground-based Telescopes. pp 1739–1749Chiappini C., Renda A., Matteucci F., 2002, A&A, 395, 789Chini R., Wink J. E., 1984, A&A, 139, L5Coc A., Angulo C., Vangioni-Flam E., Descouvemont P., Adah-chour A., 2005, Nuclear Phys. A, 752, 522Cooke R. J., Fumagalli M., 2018, Nature Astronomy, 2, 957Copetti M. V. F., Oliveira V. A., Riffel R., Casta˜neda H. O.,Sanmartim D., 2007, A&A, 472, 847D’Odorico S., Peimbert M., Sabbadin F., 1976, A&A, 47, 341D’Odorico S., Cristiani S., Dekker H., Hill V., Kaufer A., KimT., Primas F., 2000, in Bergeron J., ed., Society of Photo-Optical Instrumentation Engineers (SPIE) Conference SeriesVol. 4005, Society of Photo-Optical Instrumentation Engi-neers (SPIE) Conference Series. pp 121–130Deharveng L., Pe˜na M., Caplan J., Costero R., 2000, MNRAS,311, 329Delgado-Inglada G., Morisset C., Stasi´nska G., 2014, MNRAS,440, 536Dom´ınguez-Guzm´an G., Rodr´ıguez M., Esteban C., Garc´ıa-RojasJ., 2019, arXiv e-prints, p. arXiv:1906.02102Drew J. E., Mongui´o M., Wright N. J., 2019, MNRAS, 486, 1034Ekstr¨om S., et al., 2012, A&A, 537, A146Esteban C., Garc´ıa-Rojas J., 2018, MNRAS, 478, 2315Esteban C., Vilchez J. M., 1992, ApJ, 390, 536Esteban C., Vilchez J. M., Smith L. J., Clegg R. E. S., 1992,A&A, 259, 629Esteban C., Peimbert M., Garc´ıa-Rojas J., Ruiz M. T., PeimbertA., Rodr´ıguez M., 2004, MNRAS, 355, 229Esteban C., Bresolin F., Peimbert M., Garc´ıa-Rojas J., PeimbertA., Mesa-Delgado A., 2009, ApJ, 700, 654Esteban C., Carigi L., Copetti M. V. F., Garc´ıa-Rojas J., Mesa-Delgado A., Casta˜neda H. O., P´equignot D., 2013, MNRAS,433, 382Esteban C., Mesa-Delgado A., Morisset C., Garc´ıa-Rojas J., 2016,MNRAS, 460, 4038Esteban C., Fang X., Garc´ıa-Rojas J., Toribio San Cipriano L.,2017, MNRAS, 471, 987MNRAS000 , 1–15 (2015) J. E. M´endez-Delgado et al.
Fa´undez-Abans M., Maciel W. J., 1987, A&AS, 129, 353Fern´andez-Mart´ın A., P´erez-Montero E., V´ılchez J. M., MampasoA., 2017, A&A, 597, A84Fern´andez V., Terlevich E., D´ıaz A. I., Terlevich R., Rosales-Ortega F. F., 2018, MNRAS, 478, 5301Foster T., Brunt C. M., 2015, AJ, 150, 147Froese Fischer C., Tachiev G., 2004,Atomic Data and Nuclear Data Tables, 87, 1Garc´ıa-Rojas J., Esteban C., Peimbert M., Rodr´ıguez M., RuizM. T., Peimbert A., 2004, ApJS, 153, 501Garc´ıa-Rojas J., Esteban C., Peimbert A., Peimbert M., Ro-dr´ıguez M., Ruiz M. T., 2005, MNRAS, 362, 301Garc´ıa-Rojas J., Esteban C., Peimbert M., Costado M. T., Ro-dr´ıguez M., Peimbert A., Ruiz M. T., 2006, MNRAS, 368, 253Garc´ıa-Rojas J., Esteban C., Peimbert A., Rodr´ıguez M., Peim-bert M., Ruiz M. T., 2007, Rev. Mexicana Astron. Astrofis.,43, 3Georgy C., Ekstr¨om S., Meynet G., Massey P., Levesque E. M.,Hirschi R., Eggenberger P., Maeder A., 2012, A&A, 542, A29Gravity Collaboration et al., 2019, A&A, 625, L10Grieve M. F. R., Ramsbottom C. A., Hudson C. E., Keenan F. P.,2014, ApJ, 780, 110Hawley S. A., 1978, ApJ, 224, 417Izotov Y. I., Thuan T. X., Lipovetsky V. A., 1994, ApJ, 435, 647Izotov Y. I., Thuan T. X., Guseva N. G., 2014, MNRAS, 445, 778Kingdon J., Ferland G. J., 1995, ApJ, 442, 714Kisielius R., Storey P. J., Ferland G. J., Keenan F. P., 2009,MNRAS, 397, 903Kubryk M., Prantzos N., Athanassoula E., 2015, A&A, 580, A127Kunth D., Sargent W. L. W., 1983, ApJ, 273, 81Kwitter K. B., 1984, ApJ, 287, 840Lequeux J., Peimbert M., Rayo J. F., Serrano A., Torres-PeimbertS., 1979, A&A, 500, 145Liu X.-W., Storey P. J., Barlow M. J., Danziger I. J., Cohen M.,Bryce M., 2000, MNRAS, 312, 585Lorenzo J., Sim´on-D´ıaz S., Negueruela I., Vilardell F., Garcia M.,Evans C. J., Montes D., 2017, A&A, 606, A54Luridiana V., Morisset C., Shaw R. A., 2015, A&A, 573, A42Maciel W. J., 2001, Ap&SS, 277, 545Marshall J. L., et al., 2008, in Society of Photo-Optical In-strumentation Engineers (SPIE) Conference Series. p. 54( arXiv:0807.3774 ), doi:10.1117/12.789972Matteucci F., Chiappini C., 1999, in Walsh J. R., Rosa M. R.,eds, Chemical Evolution from Zero to High Redshift. p. 83( arXiv:astro-ph/9812315 )Mesa-Delgado A., Esteban C., Garc´ıa-Rojas J., Reyes-P´erez J.,Morisset C., Bresolin F., 2014, ApJ, 785, 100Moffat A. F. J., Fitzgerald M. P., Jackson P. D., 1979, A&AS,38, 197Osterbrock D. E., Ferland G. J., 2006, Astrophysics of gaseousnebulae and active galactic nuclei. 2nd. ed. Sausalito, CA:University Science BooksPagel B. E. J., Simonson E. A., Terlevich R. J., Edmunds M. G.,1992, MNRAS, 255, 325Pasquali A., Perinotto M., 1993, A&A, 280, 581Peimbert M., Costero R., 1969, Boletin de los Observatorios To-nantzintla y Tacubaya, 5, 3Peimbert M., Serrano A., 1980, Rev. Mex. Astron. Astrofis., 5, 9Peimbert M., Torres-Peimbert S., 1977, MNRAS, 179, 217Peimbert M., Rodriguez L. F., Torres-Peimbert S., 1974, Rev.Mex. Astron. Astrofis., 1, 129Peimbert M., Torres-Peimbert S., Rayo J. F., 1978, ApJ, 220, 516Peimbert M., Torres-Peimbert S., Ruiz M. T., 1992, RMxAA,24, 155Peimbert A., Peimbert M., Luridiana V., 2002, ApJ, 565, 668Peimbert A., Pe˜na-Guerrero M. A., Peimbert M., 2012, ApJ,753, 39 Peimbert A., Peimbert M., Luridiana V., 2016, Rev. Mex. Astron.Astrofis., 52, 419Planck Collaboration et al., 2018, arXiv e-prints,p. arXiv:1807.06209Podobedova L. I., Kelleher D. E., Wiese W. L., 2009,Journal of Physical and Chemical Reference Data, 38, 171Porter R., Ferland G. J., MacAdam K., 2007, in American Astro-nomical Society Meeting Abstracts. p. 885Porter R. L., Ferland G. J., Storey P. J., Detisch M. J., 2012,MNRAS, 425, L28Porter R. L., Ferland G. J., Storey P. J., Detisch M. J., 2013,MNRAS, 433, L89Quireza C., Rood R. T., Bania T. M., Balser D. S., Maciel W. J.,2006, ApJ, 653, 1226Reid M. J., 1993, ARA&A, 31, 345Rudolph A. L., Brand J., de Geus E. J., Wouterloot J. G. A.,1996, ApJ, 458, 653Rudolph A. L., Fich M., Bell G. R., Norsen T., Simpson J. P.,Haas M. R., Erickson E. F., 2006, ApJS, 162, 346Russeil D., 2003, A&A, 397, 133Russeil D., Adami C., Georgelin Y. M., 2007, A&A, 470, 161Samal M. R., et al., 2010, ApJ, 714, 1015Sawey P. M. J., Berrington K. A., 1993, Atomic Data and NuclearData Tables, 55, 81Shaver P. A., McGee R. X., Newton L. M., Danks A. C., PottaschS. R., 1983, MNRAS, 204, 53Stasi´nska G., 1978, A&A, 66, 257Stasi´nska G., Schaerer D., 1997, A&A, 322, 615Stasi´nska G., Schaerer D., Leitherer C., 2001, A&A, 370, 1Stock D. J., Barlow M. J., Wesson R., 2011, MNRAS, 418, 2532Storey P. J., Hummer D. G., 1995, MNRAS, 272, 41Storey P. J., Zeippen C. J., 2000, MNRAS, 312, 813Storey P. J., Sochi T., Badnell N. R., 2014, MNRAS, 441, 3028Talent D. L., Dufour R. J., 1979, ApJ, 233, 888Tayal S. S., 2011, ApJS, 195, 12Tayal S. S., Zatsarinny O., 2010, ApJS, 188, 32Valerdi M., Peimbert A., Peimbert M., Sixtos A., 2019, ApJ,876, 98Vilchez J. M., 1989, Ap&SS, 157, 9V´ılchez J. M., Pagel B. E. J., 1988, MNRAS, 231, 257Vincenzo F., Miglio A., Kobayashi C., Mackereth J. T., Montal-ban J., 2019, A&A, 630, A125Wiese W. L., Fuhr J. R., Deters T. M., 1996,J.Phys.Chem:Ref.Data, Monograoh No. 7, Atomic tran-sition probabilities of carbon, nitrogen, and oxygen : acritical data compilation. American Chemical Society, Wash-ington, DC, and the National Institute of Standards andTechnology (NIST)Zhang Y., Liu X. W., 2003, A&A, 404, 545
APPENDIX A: IONIC ABUNDANCES FOREACH INDIVIDUAL He i LINE
In this Appendix we include four tables. Table A1 showsthe He + / H + ratio for each individual He i line of the spec-tra analyzed in this work. In the second column we indicatethe level state that produce each line, S for singlet and Tfor triplet. We also include the mean He + / H + ratio for eachspectra. In Table A2 we give the list of individual He i linesused to derive the mean value of the He + / H + ratio adoptedfor each spectrum. In table Table A3 we include the resultsof the abundances of He + / H + estimated with the code He-lio14 . In Table A4 we include the ICFs used in each totalabundance calculation.
MNRAS , 1–15 (2015) elium abundances and gradient in Galactic nebulae Table A1. He + / H + ratios for each individual He i line of all the spectra used in this workHe i line Level G2.4+1.4(˚A) state A1 A2 A3 A4 A5 λ ± λ ± ± ± ± ± λ ± ± ± ± ± ± ± ± ± λ ± ± ± ± ± λ ± ± ± ± ± λ λ ± ± ± ± ± λ ± ± ± ± ± λ ± ± ± ± ± λ ± ± ± ± ± λ ± ± ± ± ± λ ± ± ± ± ± λ ± ± ± ± ± λ ± ± ± ± ± λ ± ± ± ± ± λ ± ± ± ± ± λ ± ± ± ± ± λ ± ± ± ± ± ± ± ± λ ± ± λ ± ± ± λ ± ± ± λ ± ± ± λ ± ± ± ± λ ± ± λ ± ± ± ± λ ± ± ± λ ± ± ± ± λ ± ± ± ± ± λ ± λ ± ± ± ± ± λ ± ± ± ± ± λ ± ± ± λ ± ± ± ± ± ± ± ± λ ± ± ± ± ± λ ± ± ± ± λ ± ± ± ± ± λ ± ± ± ± ± λ ± ± ± ± ± λ ± ± ± ± ± λ ± ± ± ± ± λ ± ± ± ± ± λ ± ± ± ± ± ± ± ± ± ± A TEX file prepared bythe author.MNRAS , 1–15 (2015) J. E. M´endez-Delgado et al.
Table A1. continuedHe i line Level NGC 7635(˚A) state A2 A3 A4 A5 A6 λ ± ± ± λ ± λ ± ± ± ± ± λ ± ± ± λ ± ± ± ± ± λ ± ± ± λ ± ± ± λ ± ± ± ± ± λ ± ± ± ± ± λ ± ± ± ± ± ± ± ± λ ± λ ± ± λ ± ± λ ± ± ± ± ± λ ± λ ± ± ± ± λ ± ± ± ± ± λ ± ± λ ± ± ± ± ± λ ± ± ± ± ± λ ± ± ± ± ± ± ± ± λ ± ± λ ± λ ± ± ± ± λ ± λ ± ± ± ± λ ± ± ± ± ± λ ± ± ± λ ± ± ± ± ± λ ± ± ± ± ± λ ± ± ± ± ± ± ± ± ± λ ± λ ± ± λ ± ± λ ± λ ± ± λ ± λ ± ± λ ± ± λ ± ± λ ± ± λ ± λ ± ± ± λ ± ± ± λ ± ± λ ± ± ± ± , 1–15 (2015) elium abundances and gradient in Galactic nebulae Table A2.
List of lines used for the calculation of the mean He + / H + ratio for each spectra.Object Zone Lines usedG2.4+1.4 A1 5876, 6678A2 4471, 5876, 6678A3 5876A4 5876, 6678A5 5876, 6678M8 3614, 3965, 4121, 4388, 4438, 4471, 4713, 4922, 5876, 6678M16 3614, 3965, 4388, 4438, 4471, 4713, 4922, 5876, 6678, 7281M17 3614, 3965, 4388, 4438, 4471, 4922, 5876, 6678M20 3614, 3965, 4388, 4438, 4471, 4922, 5016, 5876, 6678M42 3614, 3965, 4388, 4438, 4471, 4922, 5876, 6678, 7281, 9464NGC 2579 3965, 4026, 4388, 4438, 4471, 4922, 5016, 6678, 9464NGC 3576 3614, 3965, 4026, 4388, 4438, 4922, 6678, 7281, 9464NGC 3603 4121, 4388, 4471, 4922, 5876, 6678, 7281NGC 6888 A1 5016, 5876, 6678A2 4026, 4388, 4922, 5048, 5876, 6678A3 4026, 4388, 4922, 5876, 6678A4 4026, 4471, 4922, 5876, 6678A5 4026, 4471, 4922, 5876, 6678A6 4026, 4471, 4922, 5876, 6678RCW 52 4471, 5016, 5876RCW 58 4388, 4471, 4922, 5016, 5876NGC 7635 A1 4026, 4388, 4713, 5016, 6678A2 5048, 5876, 6678A3 4713, 4922, 5048, 5876, 6678A4 4026, 4388, 4471, 4922, 5016, 7281A5 4922, 5016, 5876, 6678A6 5016, 5876, 6678Sh 2-83 4471, 4922, 5016, 6678Sh 2-100 4026, 4471, 4922, 5016, 5048, 6678Sh 2-127 4922, 5016, 6678Sh 2-128 4471, 4922, 5016, 6678Sh 2-152 4026, 4922, 5048, 5876, 7281Sh 2-209 5876, 6678, 7281Sh 2-212 4922, 5016, 5876Sh 2-288 4471, 5016, 5048, 5876Sh 2-298 4026, 4388, 4471, 4713, 4922, 5876Sh 2-308 5876, 6678Sh 2-311 3614, 3965, 4026, 4121, 4388, 4471, 4713, 4922, 5876, 6678, 7281MNRAS000
List of lines used for the calculation of the mean He + / H + ratio for each spectra.Object Zone Lines usedG2.4+1.4 A1 5876, 6678A2 4471, 5876, 6678A3 5876A4 5876, 6678A5 5876, 6678M8 3614, 3965, 4121, 4388, 4438, 4471, 4713, 4922, 5876, 6678M16 3614, 3965, 4388, 4438, 4471, 4713, 4922, 5876, 6678, 7281M17 3614, 3965, 4388, 4438, 4471, 4922, 5876, 6678M20 3614, 3965, 4388, 4438, 4471, 4922, 5016, 5876, 6678M42 3614, 3965, 4388, 4438, 4471, 4922, 5876, 6678, 7281, 9464NGC 2579 3965, 4026, 4388, 4438, 4471, 4922, 5016, 6678, 9464NGC 3576 3614, 3965, 4026, 4388, 4438, 4922, 6678, 7281, 9464NGC 3603 4121, 4388, 4471, 4922, 5876, 6678, 7281NGC 6888 A1 5016, 5876, 6678A2 4026, 4388, 4922, 5048, 5876, 6678A3 4026, 4388, 4922, 5876, 6678A4 4026, 4471, 4922, 5876, 6678A5 4026, 4471, 4922, 5876, 6678A6 4026, 4471, 4922, 5876, 6678RCW 52 4471, 5016, 5876RCW 58 4388, 4471, 4922, 5016, 5876NGC 7635 A1 4026, 4388, 4713, 5016, 6678A2 5048, 5876, 6678A3 4713, 4922, 5048, 5876, 6678A4 4026, 4388, 4471, 4922, 5016, 7281A5 4922, 5016, 5876, 6678A6 5016, 5876, 6678Sh 2-83 4471, 4922, 5016, 6678Sh 2-100 4026, 4471, 4922, 5016, 5048, 6678Sh 2-127 4922, 5016, 6678Sh 2-128 4471, 4922, 5016, 6678Sh 2-152 4026, 4922, 5048, 5876, 7281Sh 2-209 5876, 6678, 7281Sh 2-212 4922, 5016, 5876Sh 2-288 4471, 5016, 5048, 5876Sh 2-298 4026, 4388, 4471, 4713, 4922, 5876Sh 2-308 5876, 6678Sh 2-311 3614, 3965, 4026, 4121, 4388, 4471, 4713, 4922, 5876, 6678, 7281MNRAS000 , 1–15 (2015) J. E. M´endez-Delgado et al.
Table A3.
Results of
Helio14 code: He + / H + ratios, minimal χ and T e (OII+OIII).Object Zone He + / H + χ T e (OII+OIII) (K)G2.4+1.4 A1 . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . , 1–15 (2015) elium abundances and gradient in Galactic nebulae Table A4.
ICF values. ICF(He) schemeObject PTP77 ZL03 KS83 B09G2.4+1.4 -M8 . ± .
09 1 . ± .
02 1 . ± .
07 1 . ± . M16 . ± .
21 1 . ± .
08 1 . ± .
07 1 . ± . M17 . ± .
02 1 . ± .
01 1 . ± .
01 1 . ± . M20 . ± .
16 1 . ± .
06 1 . ± .
07 1 . ± . M42 . ± .
04 1 . ± .
02 1 . ± .
02 1 . ± . NGC 2579 . ± .
05 1 . ± .
02 1 . ± .
04 1 . ± . NGC 3576 . ± .
04 1 . ± .
02 1 . ± .
02 1 . ± . NGC 3603 . ± .
01 1 . ± .
01 1 . ± .
01 1 . ± . NGC 6888 -NGC 7635 A1 . ± .
17 1 . ± .
09 1 . ± .
08 1 . ± . NGC 7635 A2 . ± .
08 1 . ± .
04 1 . ± .
07 1 . ± . NGC 7635 A3 . ± .
10 1 . ± .
05 1 . ± .
08 1 . ± . NGC 7635 A4 . ± .
10 1 . ± .
06 1 . ± .
05 1 . ± . NGC 7635 A5 . ± .
23 1 . ± .
13 1 . ± .
14 1 . ± . NGC 7635 A6 . ± .
28 1 . ± .
15 1 . ± .
18 1 . ± . RCW 52 . ± .
49 1 . ± .
17 1 . ± .
36 2 . ± . RCW 58 . ± .
10 1 . ± .
03 1 . ± .
12 1 . ± . Sh 2-83 . ± .
02 1 . ± .
02 1 . ± .
01 1 . ± . Sh 2-100 . ± .
02 1 . ± .
01 1 . ± .
01 1 . ± . Sh 2-127 . ± .
12 1 . ± .
04 1 . ± .
08 1 . ± . Sh 2-128 . ± .
07 1 . ± .
03 1 . ± .
03 1 . ± . Sh 2-152 . ± .
07 1 . ± .
02 1 . ± .
05 3 . ± . Sh 2-209 . ± .
30 1 . ± .
16 1 . ± .
14 1 . ± . Sh 2-212 . ± .
40 1 . ± .
15 1 . ± .
30 1 . ± . Sh 2-288 . ± .
25 1 . ± .
11 1 . ± .
12 1 . ± . Sh 2-298 . ± .
53 1 . ± .
20 1 . ± .
06 1 . ± . Sh 2-308 . ± .
08 1 . ± .
06 1 . ± .
04 1 . ± . Sh 2-311 . ± .
14 1 . ± .
05 1 . ± .
07 1 . ± . MNRAS000