The abundance discrepancy in H ii regions
In this paper we discuss some results concerning the abundance discrepancy problem in the context of H ii regions. We discuss the behavior of the abundance discrepancy factor (ADF) for different objects and ions. There are evidences that stellar abundances seem to agree better with the nebular ones derived from recombination lines in high-metallicity environments and from collisionally excited lines in the low-metallicity regime. Recent data point out that the ADF seems to be correlated with the metallicity and the electron temperature of the objects. These results open new ways for investigating the origin of the abundance discrepancy problem in H ii regions and in ionized nebulae in general.
Instituto de Astrofísica de Canarias, E-38200 La Laguna, Tenerife, Spain
Departamento de Astrofísica, Universidad de La Laguna, E-38206, La Laguna, Tenerife, Spain
The emission-line spectrum of H ii regions can be detected and analyzed even at very large distances. They are essential probes to measure distances of extragalactic objects, the intensity and properties of star formation processes and the chemical composition of the interstellar medium. In fact, H ii regions provide most of our current knowledge about the chemical composition of the Universe. The spectra of nebulae contain emission lines of several ionization species of chemical elements that are present in the ionized gas. The analysis of emission-lines permits to determine the abundance – the ratio of the number of atoms of a given element with respect to those of H – of important elements as He, C, N, O, Cl, -elements as Ne, S and Ar and iron-peak elements as Fe and Ni. In the next decade, the use of 30-50m aperture telescopes will probably permit to derive abundances of some neutron-capture s-elements as Se or Kr in the brightest H ii regions.
The O abundance determined from nebular spectra is the most widely used proxy of metallicity for galaxies at different redshifts. A proper determination of the physical conditions – electron temperature, , and density, – is necessary for obtaining reliable and accurate metallicities. Many of the emission lines in nebular spectra are excited by collisions with free electrons – the so-called collisionally excited lines (CELs) – and are forbidden by the selection rules. CELs can be very bright due to the high electron temperatures and low densities of ionized nebulae. In addition, there are also bright recombination lines of H i – Balmer series in the optical – and He i. In the last two decades, our group has developed a long-term research project devoted to measure very faint pure recombination lines (RLs) of heavy-element ions in H ii regions from high- and intermediate-resolution spectroscopical data. The optical-NIR spectral range contains RLs of O i, O ii, C ii and Ne ii, which intensities are of the order of 0.0001 to 0.001 times that of H. Due to their faintness, these lines can only be detected and measured in bright nebulae using large aperture telescopes.
For several decades it has turned out that abundances of heavy-element ions determined from the standard method based on the intensity ratios of collisionally excited lines (CELs) are systematically lower than those derived from the faint RLs emitted by the same ions. This fact, usually known as the abundance discrepancy (AD) problem is quantified by the abundance discrepancy factor (hereafter ADF), defined as the difference between the logarithmic abundances derived from RLs and CELs:
In Galactic and extragalactic H ii regions, the ADF of O – ADF(O) – is between 0.10 and 0.35 dex (e.g. García-Rojas & Esteban 2007; López-Sánchez et al. 2007; Esteban et al. 2009, 2014; Toribio San Cipriano et al. 2016, 2017) and can be even much larger in some planetary nebulae (PNe, see paper by García-Rojas in this proceedings). It is important to remark that other ions with optical RLs: C, Ne, and O give values of their ADFs similar to those obtained for the ADF(O) for the same object (García-Rojas & Esteban 2007; Toribio San Cipriano et al. 2017), although the data for those ions are more limited and difficult to obtain (see Table 1). It is clear that changes or uncertainties in the face value of metallicity we adopt for celestial bodies may have a major impact in many fields of Astrophysics as the ingredients of chemical evolution models and predicted stellar yields (e.g. Carigi et al. 2005), the luminosity –and mass– metallicity relations for local and high-redshift star-forming galaxies (e.g. Tremonti et al. 2004), the calibration of strong-line methods for deriving the abundance scale of extragalactic H ii regions and star-forming galaxies at different redshifts (e.g. Peña-Guerrero et al. 2012, López-Sánchez et al. 2012) or the determination of the primordial helium (e.g. Peimbert 2008), among others.
|Ion||Orion Neb.||M8||30 Dor||N66C||NGC 5253|
1- Esteban et al. (2004); 2- García-Rojas et al. (2007); 3- Peimbert (2003);
4- Toribio San Cipriano et al. (2017); 5- López-Sánchez et al. (2007).
The origin of the AD is still unknown, but several hypotheses have been proposed. In this paper we will limit our discussion to the case of H ii regions. For PNe the situation is more complex because they can contain material with different physical conditions and composition (see Peimbert & Peimbert 2006; García-Rojas, this proceedings). The first hypothesis for explaining the AD was formulated by Torres-Peimbert et al. (1980). They propose that the AD is produced by spatial fluctuations of electron temperature, , in the ionized gas, in the form originally proposed by Peimbert (1967). However, these temperature variations cannot be reproduced by chemically homogeneous photoionization models and different mechanisms have been invoked to explain their presence in ionized nebulae (see Esteban 2002; Peimbert & Peimbert 2006). According to this scenario, RLs should provide the true abundances because their emissivities are much less dependent on temperature that those of CELs. In fact, in the typical range of of H ii regions, the emissivities of the RLs of heavy-element ions are almost identical to that of H and therefore abundances derived from RLs are practically independent on . Nicholls et al. (2012) have proposed that a -distribution of electron energies – departure from the Maxwell-Boltzmann distribution – can be operating in nebulae and produce the AD, being also the determinations based on RLs the most reliable ones. However, some recent works have seriously questioned this hypothesis (e. g. Zhang et al. 2016; Ferland et al. 2016). A third hypothesis was proposed by Stasińska et al. (2007) and is based on the existence of semi-ionized clumps embedded in the ambient gas of H ii regions. These hypothetical clumps – unmixed supernova ejecta – would be strong RL emitters, denser, cooler and more metallic than the ambient nebular gas. Assuming this scenario, abundances derived from RLs and CELs should be upper and lower limits, respectively, to the true ones, though those from CELs should be more reliable. There is even a last scenario outlined by Tsamis et al. (2011), which involves the presence of high-density clumps but without abundance contrast between the components.
Until now, the only work where the AD problem in a sample of H ii regions has been discussed in length is that by García-Rojas & Esteban (2007). They found that the ADF is fairly constant and of order 2 in a limited sample of Galactic and extragalactic H ii regions. In addition, they did not find correlations between the ADF(O) and the O/H, O/H ratios, the ionization degree, , FWHM of several bright emission-lines, and the effective temperature of the main ionizing stars within the observational uncertainties. On the contrary, they found that ADF seems to be slightly dependent on the excitation energy of the levels that produce the RLs, a fact that is consistent with the predictions of the classical temperature fluctuations or -distributions hypotheses.
2 Comparisons between nebular and stellar abundances
If CELs and RLs give different abundances, in which lines shall we trust? There are some observational evidences that can give us clues about this important question. In Figure 1 we represent the abundances of 4 representative elements – C, N, O, and Ne – for which abundances have been derived from different kinds of lines in the Orion Nebula: CELs in the ultraviolet (UV, Walter et al. 1992; Tsamis et al. 2011), optical (Esteban et al. 2004) and far infrared (FIR, Simpson et al. 1986; Rubin et al. 2011), and optical RLs (Esteban et al. 2004). Except in the case of C – for which there are two discrepant independent abundance determinations from UV CELs – there is a tendency to increase the abundance when we move to the right in the diagrams, in the sense that the lines which intensity is less dependent on give higher abundances. This tendency agrees qualitatively with the predictions if the AD is produced by the temperature fluctuations or -distribution hypotheses.
In Figure 2 we compare the abundances of C, O and Ne of the Orion Nebula – using UV or optical CELs and RLs – with those of other objects of the Solar Neighbourhood for which absolute abundances can be derived. We include the Sun (Asplund et al. 2009; Caffau et al. 2008, 2009), young F&G stars (Sofia & Meyer 2001), and B-type stars (Nieva & Simón-Díaz 2011). The size of the bars is proportional to their uncertainty. For C/H we can see that the solar and stellar values are quite consistent with the values obtained for the Orion Nebula using RLs, a similar trend is also found for O/H. However, we would expect some depletion into dust for both elements in the Orion Nebula, of the order of about 0.1 dex according to Esteban et al. (2004). Ne is a noble gas and no dust depletion is expected for this element and the results also indicate that the stellar and solar Ne abundances are more consistent with the Ne/H ratios determined from RLs in the Orion Nebula. It is interesting to note that the accurate determination of Ne/H obtained by Rubin et al. (2011) from FIR CELs observed with Spitzer is consistent with the RLs values. This result is also consistent with the predictions of the temperature fluctuations or -distribution hypotheses considering the low temperature dependence of the emissivity of FIR CELs. From the results presented in figures 1 and 2, one can conclude that nebular abundances determined from RLs are more consistent with those of the Sun and other stellar objects of the Solar Neighbourhood than those determined from CELs.
Since 2002, our group has measured C ii and O ii RLs in several tens of extragalactic H ii regions (Esteban et al. 2002, 2009, 2014; López-Sánchez et al. 2007; Toribio San Cipriano 2016, 2017). In several of those papers we have compared our nebular abundances derived from CELs and RLs with those determined in early B-type stars located in their vicinity. The O abundance of early B-type stars – which is not expected to be affected by stellar evolution effects – should reflect the present-day chemical composition of the interstellar material in the regions where they are located. The results for extragalactic objects are not always consistent with the aforementioned ones discussed for the Orion Nebula. Toribio San Cipriano et al. (2016) compared O abundances derived from CELs and RLs in H ii regions of the spiral galaxies M33 and NGC300 with the O abundances derived in B supergiants by Urbaneja et al. (2005a, 2005b). These authors found that the stellar O/H ratios are in better agreement with the nebular abundances calculated using RLs in the case of M33, while in the case of NGC 300 the agreement is better with CELs. More recently, Toribio San Cipriano et al. (2017) make a similar comparison in the case of the Magellanic Clouds. They find that the O abundances determined by Hunter et al. (2009) for B-type stars in young clusters of the LMC and SMC are more consistent with the values obtained from CELs for the associated H ii regions.
Bresolin et al. (2016) suggested that nebular abundances determined either from CELs or RLs can display different levels of agreement with the B supergiant determinations in different galaxies, depending on the metallicity regime. They proposed that the RL-based nebular metallicities agree with the stellar metallicities better than the CELs in the high-abundance regime, but that this situation reverses at low abundance values. In Figure 3 we have made a similar exercise than in figure 11 of Bresolin et al. (2016) but including data for individual H ii regions with high quality nebular CELs and RLs determinations and compare with O/H ratios determined from young supergiant stars located in the same star-forming regions or galactocentric distance in the same galaxy. The squares correspond to data for the Orion Nebula and B-type stars of its associated cluster (Esteban et al. 2004; Simón-Díaz & Stasińska 2011). For the MCs we include data for the ionized gas (Toribio San Cipriano et al. 2017) and B-type stars (Hunter et al. 2009) of the star- forming regions N11 (LMC, circles) and N66 (SMC, stars). We have considered data points of several H ii regions of M33 (triangles) and NGC300 (diamonds) taken from Toribio San Cipriano et al. (2016) and Esteban et al. (2009), the stellar abundances have been estimated from the radial O abundance gradients determined by Urbaneja et al. (2005a, 2005b) from the spectra of B-type supergiants, evaluated at the galactocentric distances of each H ii region. The down-facing triangles represent values for the dwarf irregular galaxy NGC 6822, the nebular data are taken from Esteban et al. (2014) and the stellar ones from the spectral analysis performed by Venn et al. (2001). Nebular data have been increased 0.1 dex to correct for dust depletion in all the objects. From the figure, one can conclude that while in the case of M33 and NGC 6822 both kinds of lines give nebular abundances consistent with the stellar ones, in the other cases we obtain contradictory results. The nebular O abundances determined from RLs in the Orion Nebula are the ones consistent with stellar determinations while in the MCs and NGC 300 the determinations based on CELs are the only ones consistent with the stellar abundances. This seems to be qualitatively consistent with the result obtained by Bresolin et al. (2016). We need further high-quality data – specially in the higher- and lower-metallicity regimes – to confirm this tendency.
3 Adf versus some nebular parameters
García-Rojas & Esteban (2007) explored the dependence of the ADF with different properties of H ii regions. Those authors reported that the ADF seems to be independent of metallicity however, their sample was rather limited in both, the number of objects and metallicity. In Figure 4, we present the behavior of the ADF(O) with respect to O/H ratio, , and ionization degree (O/O ratio). Green squares correspond to Galactic H ii regions (García-Rojas & Esteban 2007; Esteban et al. 2004, 2013); red circles and blue stars represent objects in the LMC and SMC, respectively (Toribio San Cipriano et al. 2017); orange triangles correspond to H ii regions in M33 (Toribio San Cipriano et al. 2016); cyan pentagons to objects in M101 (Esteban et al. 2009); and pink down-facing triangles to star-forming dwarf galaxies (López-Sánchez et al. 2007; Esteban et al. 2014). Let’s focus our attention in panel a) of Figure 4, although the ADF(O) values of many objects have large uncertainties, the distribution of the points indicates a complex behavior in the ADF(O) versus O/H diagram. The ADF distribution shows a minimum at 12 + log(O/H) 8.5 and seems to increase when the O/H becomes higher or lower. This apparent “seagull” shape, if real, is difficult to explain. In the case of Galactic H ii regions we obtain a high dispersion of ADFs and a tendency to larger values when O/H becomes higher. Considering the closeness of the Galactic nebulae, their dispersion may be due to aperture effects because we tend to observe bright small areas in these objects and their ADF could not be representative of the whole nebula. As it has been proven in several works (e.g. Mesa-Delgado et al. 2012, and references therein) the presence of localized high-velocity flows and/or high density clumps can produce large ADFs. The trend of the low-metallicity wing of the “seagull” shape of Figure 4, seems to be clearer than the high-metallicity wing. The low-metallicity objects correspond to more distant objects, their spectra are obtained from apertures that encompass a large fraction of the nebula and, therefore, their observed areas are more representative of the global emission. Toribio San Cipriano et al. (2017) also find that the ADF(C) seems to increase toward lower metallicities, reinforcing the validity of the apparent correlation between ADF(O) and O/H shown in panel a).
Panel b) of Figure 4 shows the ADF(O) versus for the high-ionization species, basically ([O iii]). We can see a similar seagull shape in this relation but inverted with respect to that shown in panel a) –. This result is not unexpected, as and metallicity are related because O is one of the most important coolants in ionized nebulae. The trend of a decrease of ADF with increasing of ([O iii]) in H ii regions of near-solar metallicity and ([O iii]) 10 K was previously reported by Rodríguez & Manso Sainz (2014).
Panel c) of Figure 4 represents the ADF(O) versus O/O ratio – ionization degree – of the objects. The panel shows two main groups of objects. Most of them lie around O/O 0.8 and a small band is located at O/O 0.5, these last objects – most of them Galactic H ii regions – show the lowest ionization degrees as well as ADFs higher than the average. It is interesting to note that, in panels a) and b), this group of low-ionization objects lie precisely in the zones of high O/H ratio and low where we find the apparent trend of increasing ADFs. Panels d) and e) of Figure 4 show the same relations than panels a) and b) but removing the low-ionization objects (those with O/O 0.5). In these new panels we can see that the trends of the ADF towards high metallicity and low are practically washed out. The interpretation of this behavior is not easy. In low-ionization nebulae with O/O 0.5, the O/H ratio is dominated by O, and therefore the physical conditions of the O zone – specially – may be not representative of the whole H ii region. We will further investigate this interesting result.
Another interesting aspect of Figure 4 is the large ADF(O) values reported in some giant H ii regions belonging to star-forming dwarf galaxies (pink down-facing triangles). Esteban et al. (2016) speculate that large ADFs in H ii regions might be produced by the presence of shocks due to large-scale interaction between the ionized gas and of stellar winds. A process that may dominate the kinematics of these complex and huge nebulae. This suggestion is based on the results obtained by Mesa-Delgado et al. (2014) and Esteban et al. (2016), who determined the ADF(O) in several Galactic ring nebulae associated with evolved massive stars. These objects show mean ADF(O) between 0.38 and 0.50 dex, values larger than the ADFs of normal H ii regions with the same O abundance but similar to the high ADF(O) values found in star-forming dwarf galaxies.
CE thanks Oli Dors Jr. and the rest of the members of the LOC for their kind invitation and financial support to participate in this enjoyable workshop. LTSC is supported by the FPI Program by MINECO under grant AYA2011-22614. This work was also partially funded by MINECO under grant AYA2015-65205-P.
Asplund, M., Grevesse, N., Sauval, A. J. et al. 2009, Annual Review of Astronomy & Astrophysics, 47, 481 \referenceBresolin F., Kudritzki R.-P., Urbaneja M. A. 2016, \apj, 830, 64 \referenceCaffau, E., Ludwig, H. G., Steffen, M. 2008, \aap, 488, 1031 \referenceCaffau, E., Maiorca, E., Bonifacio, P. et al. 2009, \aap, 498, 877 \referenceCarigi, L., Peimbert, M., Esteban, C. et al. 2005, \apj, 623, 213 \referenceEsteban, C. 2002, RevMexAASC, 12, 56 \referenceEsteban, C., Bresolin, F., Peimbert, M. et al. 2009, \apj, 700, 654 \referenceEsteban, C., Carigi, L., Copetti, V. F. et al. 2013, \mnras, 433, 382 \referenceEsteban, C., García-Rojas, J., Carigi, L. et al. 2014, \mnras, 443, 624 \referenceEsteban, C., Mesa-Delgado, A., Morisset, C. et al. 2016, \mnras, 460, 4038 \referenceEsteban, C., Peimbert, M., García-Rojas, J. et al. 2004, \mnras, 355, 229 \referenceEsteban, C., Peimbert, M., Torres-Peimbert, S. et al. 2002, \apj, 581, 241 \referenceFerland, G. J., Henney, W. J., O’Dell, C. R. et al. 2016, RevMexAA, 52, 261 \referenceGarcía-Rojas, J., Esteban, C. 2007, \apj, 670, 457 \referenceGarcía-Rojas, J., Esteban, C., Peimbert, A. et al. 2007, RevMexAA, 43, 3 \referenceHunter, I., Lennon, D. J., Dufton, P. L. et al. 2009, \aap, 504, 211 \referenceLópez-Sánchez, A. R., Dopita, M. A., Kewley, L. J. et al. 2012, \mnras, 426, 2630 \referenceLópez-Sánchez, A. R., Esteban, C., García-Rojas, J. et al. 2007, \apj, 656, 168 \referenceMesa-Delgado, A., Esteban, C., García-Rojas, J. et al. 2014, \apj, 785, 100 \referenceMesa-Delgado, A., Núñez-Díaz, M., Esteban, C. et al. 2012, \mnras, 426, 614 \referenceNicholls, D. C., Dopita, M. A., Sutherland, R. S. 2012, \apj, 752, 148 \referenceNieva, M. F., Simón-Díaz, S. 2011, \aap, 532, 2 \referencePeimbert, A. 2003, \aap, 584, 735 \referencePeimbert, M. 1967, \apj,150, 825 \referencePeimbert, M. 2008, Current Science, 95, 1165 \referencePeimbert, M., Peimbert, A. 2006, in Planetary Nebulae in our Galaxy and Beyond, ed. M. J. Barlow, R. H. Méndez, IAU Symp. 234, Cambridge Univ. Press, p. 227 \referencePeña-Guerrero, M. A., Peimbert, A., Peimbert, M. 2012, \apj, 756, 14 \referenceRubin, R. H., Simpson, J. P., O’Dell, C. R. et al. 2011, \mnras, 410, 1320 \referenceSimón-Díaz, S., Stasińska, G. 2011, \aap, 526, 544 \referenceSimpson, J. P., Rubin, R. H., Erickson, E. F. et al. 1986, \apj, 311, 895 \referenceSofia, U. J., Meyer, D. M. 2001, \apj, 544, 221 \referenceStasińska, G., Tenorio-Tagle, G., Rodríguez, M. et al. 2007, \aap, 471, 193 \referenceRodríguez, M., Manso Sainz, R. 2014, RevMexAASC, 44, 22 \referenceToribio San Cipriano, L., Domínguez-Guzmán, G., Esteban, C. et al. 2017, \mnras, submitted \referenceToribio San Cipriano, L., García-Rojas, J., Esteban, C. et al. 2016, \mnras, 458, 1866 \referenceTorres-Peimbert, S., Peimbert, M., Daltabuit, E. 1980, \apj, 238, 133 \referenceTremonti, C. A., Heckman, T. M., Kauffmann, G. et al. 2004, \apj, 613, 898 \referenceTsamis, Y. G., Walsh, J. R., Vílchez, J. M. et al. 2011, \mnras, 412, 1367 \referenceUrbaneja, M. A., Herrero, A., Bresolin, F. et al. 2005a, \apj, 622, 862 \referenceUrbaneja, M. A., Herrero, A., Kudritzki, R.-P. et al. 2005b, \apj, 635, 311 \referenceVenn, K. A., Lennon, D. J.; Kaufer, A. et al. 2001,\apj, 547, 765 \referenceWalter, D. K., Dufour, R. J., Hester, J. J. 1992, \apj, 397, 196 \referenceZhang, Y., Zhang, B., Liu, X.-W. 2016, \apj, 817, 68