Neutron-capture Elements in Planetary Nebulae: First Detections of Near-infrared [Te iii] and [Br v] Emission Lines^{* †}

Nucleosynthetic processes in asymptotic giant branch (AGB) stars produce a substantial fraction of the trans-iron (Z > 30) elements present in the Solar System and the universe. During the late thermally pulsing AGB evolutionary stage of low- and intermediate-mass (∼1.5–8 M_⊙) stars, iron peak nuclei in the intershell region between the He- and H-burning shells experience sequential neutron(n)-captures interlaced with β decays, forming heavier elements via the so-called slow neutron capture or "s-process." The free neutrons are released as a result of alpha captures by ¹³C or ²²Ne nuclei during phases of the thermal pulse cycle (Käppeler et al. 2011). Convective dredge-up from the intershell region (third dredge-up (TDU)) may convey s-processed material, along with C, to the stellar surface, which is released into the interstellar medium by stellar winds and planetary nebula (PN) ejection (Karakas & Lattanzio 2014, and references therein).

PNe are valuable laboratories for studying the s-process because they represent the final envelope compositions of their progenitor stars after the last thermal pulse and cessation of nucleosynthesis and TDU. The past few years have seen rapid development of near-infrared (NIR) spectroscopy as a tool for such investigations. The first NIR emission lines of n-capture elements discovered, [Kr iii] 2.199 μm and [Se iv] 2.287 μm (Dinerstein 2001), have been widely observed in Milky Way and Magellanic Cloud PNe (Sterling & Dinerstein 2008; Sterling et al. 2015; hereafter SPD15; Mashburn et al. 2016). Additional NIR line detections include [Rb iv], [Cd iv], [Ge vi], [Se iii], and [Kr vi] (Sterling et al. 2016, 2017).

In this Letter we present the discovery of NIR lines from the n-capture elements Br and Te, and the atomic data needed to derive chemical abundances from the line intensities. Although optical lines of Te²⁺ and Br²⁺ have previously been reported in some PNe (Péquignot & Baluteau 1994; Sharpee et al. 2007; hereafter SH07; Madonna et al. 2017; hereafter MGR17), they are faint and potentially blended with lines of other species. The discovery of relatively bright NIR lines of these n-capture elements provides an important new opportunity for determining the contributions of AGB stars to galactic enrichment.

2.1. Espectrógrafo Multiobjeto Infra-Rojo (EMIR) Observations

2.2. Immersion GRating INfrared Spectrometer (IGRINS) Observations

Table 3 presents the intensities of NIR n-capture element lines as ratios to nearby H i lines, and Figure 1 shows excerpts of the spectral regions near the features that we identify as [Te iii] and [Br v]. While the lines are unresolved in the EMIR data, their profiles as seen with IGRINS reflect actual velocity structure and are consistent with the profiles of other ionic lines. Because the ionization potential (IP) of Te⁺ is 18.55 eV and that of Te²⁺ is 27.96 eV, a significant fraction of the Te in PNe with cool central stars, such as IC 418, should be in the form of Te²⁺. Despite its hotter central star, the bright, s-process enriched PN NGC 7027 also shows [Te iii]. Because stellar photons of at least 47.24 eV are required to produce Br⁴⁺, it is not surprising that we do not see [Br v] in IC 418.

Zoom In Zoom Out Reset image size

We identify the line near 2.1019 μm as the 5s²5p^{2 3}P₁–³P₀ transition of Te²⁺. Although the most recently reported vacuum laboratory value is 2.1021 μm (Tauheed & Naz 2011), the IGRINS observations indicate that the actual wavelength is 2.1019 μm, in good accord with the experimental data of Joshi et al. (1992). The older energy level values of Moore (1958) correspond to a wavelength of 2.1048 μm. This highlights the importance of ongoing atomic physics research, especially laboratory spectroscopy, for identifying n-capture element transitions, as well the value of high-spectral resolution astronomical observations for measuring precise wavelengths.

We considered alternate identifications within ±10Å of the detected features by querying the Atomic Line List v2.05b21.¹⁰ One of the most plausible alternate identifications is [Fe ii] 2.1017 μm. However, other [Fe ii] lines of the same multiplet are not detected, although they are predicted to be brighter than this transition according to calculations with PyNeb v1.0.26 (Luridiana et al. 2015). Additionally, the feature at 2.1017 μm is too strong by several orders of magnitude to be consistent with the Fe⁺ ionic abundance derived from [Fe ii] 1.2570, 12946, and 1.6440 μm. An even closer wavelength coincidence with the observed feature is 4s ²P_1/2–⁴P_1/2 [Ni ii] 2.1019 μm, one of several transitions that can be enhanced over the strength expected for pure collisional excitation by UV continuum fluorescence (Bautista et al. 1996). We eliminated this possibility as several other [Ni ii] transitions that are expected to be strong under such excitation conditions (1.7249, 1.7651, 2.0492, and 2.0811 μm) are not detected.

Additional supporting evidence for the identification of the 2.1019 μm line as [Te iii] is the consistency of Te²⁺ ionic abundances derived from this line with values based on [Te iii] λ7933.31. From the flux of this line in IC 418 reported by Sharpee et al. (2003), we find a Te²⁺ abundance within ∼0.2 dex of that from the NIR line (see Table 3). There is only an upper limit on the intensity of [Te iii] λ7933.31 in NGC 7027 (SH07), but this is consistent with our results as the corresponding limit on the Te²⁺ abundance is larger than the value that we derive from [Te iii] 2.1019 μm (Table 3).

We identify the line at 1.6429 μm seen in NGC 7027 as the 4s²4p ${{}^{2}{{\rm{P}}}^{{\rm{o}}}}_{3/2}-{}^{2}{{{\rm{P}}}^{{\rm{o}}}}_{1/2}$ transition of Br⁴⁺, for which the energy levels provided by Riyaz et al. (2014) yield a vacuum wavelength of 1.6429 μm. We searched the Atomic Line List and found no plausible alternate atomic lines. H₂ 6 − 4 Q(5), which can be strong in sources with highly fluorescent H₂ spectra (K. F. Kaplan et al. 2018, in preparation), is a potential contaminant, but the IGRINS data demonstrate that it is not present at a detectable intensity.

The good agreement between EMIR and IGRINS results and the high resolution of IGRINS data, which eliminates the possibility of blends, give us confidence that [Te iii] and [Br v], respectively, are the correct identifications for the 2.1019 μm and 1.6429 μm lines.

The atomic structures, A-values, and radial wavefunctions for Br v and Te iii were computed with the autostructure code Badnell (1986, 1997). We diagonalized the Breit–Pauli Hamiltonian within a statistical Thomas–Fermi–Dirac–Amaldi model potential V(λ_nl) (Badnell 1997). The potential for each orbital was characterized by scaling the radial parameter by a quantity λ_nl that is optimized variationally by minimizing a weighted sum of the LS term energies. The LS terms are represented by configuration-interaction (CI) wavefunctions.

The CI expansion for the Br v system that we used includes the 3d¹⁰4s²4p, 3d¹⁰4s4p², 3d¹⁰4s²5s, 3d¹⁰4s²5p, $3{{\rm{d}}}^{10}4{{\rm{s}}}^{2}5{\rm{d}}$, $3{{\rm{d}}}^{9}4{{\rm{s}}}^{2}4{{\rm{p}}}^{2}$, $3{{\rm{d}}}^{9}4{{\rm{s}}}^{2}4{\rm{p}}4{\rm{d}}$, and $3{{\rm{d}}}^{9}4{{\rm{s}}}^{2}4{\rm{p}}5{\rm{s}}$ configurations, while that for Te iii includes $4{{\rm{d}}}^{10}5{{\rm{s}}}^{2}5{{\rm{p}}}^{2}$, $4{{\rm{d}}}^{10}5{\rm{s}}5{{\rm{p}}}^{3}$, $4{{\rm{d}}}^{10}5{{\rm{s}}}^{2}5{\rm{p}}5{\rm{d}}$, and $4{{\rm{d}}}^{9}5{{\rm{s}}}^{2}5{{\rm{p}}}^{3}$.

The quality of our atomic structure representations was evaluated by comparing predicted term energies with the experimental values of Riyaz et al. (2014) and Joshi et al. (1992) for Br v and Te iii, respectively. Our calculated energies are found to differ from experimental values by less than 5%. Additional small semi-empirical corrections to the orbitals were calculated through the coupling of energy terms that minimized the relative energy differences. Finally, the level energies were shifted to the experimental values in order to compute monopole, dipole, and quadrupole transition probabilities.

The scattering calculations were done with the BP R-matrix program (Berrington et al. 1997), using the orbitals from our autostructure calculation, retaining CI from 39 LS terms and 88 fine structure levels for Br v and 39 LS terms and 75 levels for Te iii. The calculations explicitly include partial waves from states with L ≤ 9 and multiplicities 1, 3, and 5 for Br v, and 2 and 4 for Te iii. Contributions from the higher partial waves were estimated using a top-up procedure.

Maxwellian-averaged collision strengths were computed for both ions for selected temperatures between 5000 K and 30000 K. New atomic data files with the resulting values for Br v and Te iii were added to PyNeb. The A-values and Maxwellian-averaged collision strengths are presented in Table 2.

Table 2.  A-values and Collision Strengths

		ϒ(T)
Transition	Aij[s−1]	5000 K	8000 K	10000 K	15000 K	20000 K	30000 K
				Te2+
3P1–3P0	1.194	5.812	4.965	4.553	3.848	3.4	2.848
3P2–3P0	1.209E–02	2.565	2.597	2.62	2.652	2.642	2.561
1D2–3P0	1.062E–04	1.08	1.1	1.103	1.082	1.045	0.9653
1S0–3P0	...	0.1534	0.159	0.1626	0.1694	0.1721	0.1696
3P2–3P1	0.52	8.74	8.487	8.403	8.255	8.085	7.684
1D2–3P1	5.141	4.317	4.43	4.463	4.421	4.297	4.004
1S0–3P1	6.797E+01	0.6048	0.632	0.6414	0.6502	0.6452	0.6158
1D2–3P2	5.266	8.606	8.86	8.924	8.896	8.749	8.368
1S0–3P2	7.098	1.053	1.177	1.222	1.271	1.273	1.228
1S0–1D2	7.902	4.021	3.949	3.915	3.855	3.818	3.763
				Br4+
${{}^{2}{{\rm{P}}}^{0}}_{3/2}\mbox{--}{{}^{2}{{\rm{P}}}^{0}}_{1/2}$	0.991	5.39	5.014	4.908	4.892	5.055	5.507
${}^{2}{{\rm{D}}}_{3/2}\mbox{--}{{}^{2}{{\rm{P}}}^{0}}_{1/2}$	1.507E+06	0.932	0.771	0.7088	0.6186	0.568	0.5084
${}^{2}{{\rm{D}}}_{5/2}{-}^{2}{{{\rm{P}}}^{0}}_{1/2}$	4.920E+04	0.8377	0.7848	0.7679	0.7479	0.7355	0.7109
${}^{2}{{\rm{P}}}_{1/2}{-}^{2}{{{\rm{P}}}^{0}}_{1/2}$	...	0.8529	0.8508	0.8575	0.8684	0.8653	0.8371
${{}^{2}{\rm{P}}}_{3/2}{\mbox{--}}^{2}{{{\rm{P}}}^{0}}_{1/2}$	7.741E+08	3.292	3.208	3.179	3.138	3.112	3.069
${{}^{2}{{\rm{D}}}^{0}}_{3/2}{\mbox{--}}^{2}{{{\rm{P}}}^{0}}_{1/2}$	...	1.505	1.585	1.633	1.708	1.733	1.707
${{}^{2}{{\rm{D}}}^{0}}_{5/2}{\mbox{--}}^{2}{{{\rm{P}}}^{0}}_{1/2}$	2.694E+09	1.488	1.446	1.423	1.385	1.368	1.359
${{}^{2}{\rm{D}}}_{3/2}{\mbox{--}}^{2}{{{\rm{P}}}^{0}}_{3/2}$	4.842E+05	1.522	1.244	1.14	0.9867	0.8946	0.7781
${}^{2}{{\rm{D}}}_{5/2}\mbox{--}{{}^{2}{{\rm{P}}}^{0}}_{3/2}$	6.594E+05	1.601	1.549	1.536	1.502	1.458	1.368
${}^{2}{{\rm{P}}}_{1/2}\mbox{--}{{}^{2}{{\rm{P}}}^{0}}_{3/2}$	2.146E+06	2.128	2.291	2.336	2.341	2.287	2.152
${}^{2}{{\rm{P}}}_{3/2}{\mbox{--}}^{2}{{{\rm{P}}}^{0}}_{3/2}$	6.632E+07	2.793	2.618	2.571	2.514	2.476	2.397
${{}^{2}{{\rm{D}}}^{0}}_{3/2}{\mbox{--}}^{2}{{{\rm{P}}}^{0}}_{3/2}$	7.471E+08	7.551	7.202	7.062	6.833	6.68	6.461
${{}^{2}{{\rm{D}}}^{0}}_{5/2}{\mbox{--}}^{2}{{{\rm{P}}}^{0}}_{3/2}$	1.313E+09	1.586	1.525	1.482	1.402	1.357	1.313
${}^{2}{{\rm{D}}}_{5/2}{\mbox{--}}^{2}{{\rm{D}}}_{3/2}$	0.129	4.457	3.822	3.521	3.027	2.735	2.424
2P1/2–2D3/2	2.285E–04	2.666	2.34	2.215	2.026	1.912	1.785
2P3/2–2D3/2	0.666	0.6057	0.5456	0.5272	0.5109	0.5135	0.5339
${{}^{2}{{\rm{D}}}^{0}}_{3/2}{\mbox{--}}^{2}{{\rm{D}}}_{3/2}$	2.818E–03	0.3996	0.3728	0.3685	0.3783	0.4003	0.4478
${{}^{2}{{\rm{D}}}^{0}}_{5/2}{\mbox{--}}^{2}{{\rm{D}}}_{3/2}$	9.839	0.1252	0.1689	0.1908	0.2288	0.2531	0.282
${{}^{2}{\rm{P}}}_{1/2}{\mbox{--}}^{2}{{\rm{D}}}_{5/2}$	0.277	7.721	6.645	6.211	5.539	5.147	4.744
${}^{2}{{\rm{P}}}_{3/2}{\mbox{--}}^{2}{{\rm{D}}}_{5/2}$	2.092	0.7815	0.7616	0.763	0.7834	0.8142	0.8767
${{}^{2}{{\rm{D}}}^{0}}_{3/2}{\mbox{--}}^{2}{{\rm{D}}}_{5/2}$	1.531	1.294	1.16	1.119	1.09	1.106	1.174
${{}^{2}{D}^{0}}_{5/2}{\mbox{--}}^{2}{{\rm{D}}}_{5/2}$	38.414	0.2605	0.3401	0.38	0.448	0.4907	0.5414
${}^{2}{{\rm{P}}}_{3/2}{\mbox{--}}^{2}{{\rm{P}}}_{1/2}$	0.780	1.142	1.008	0.9629	0.9157	0.9126	0.9448
${{}^{2}{{\rm{D}}}^{0}}_{3/2}{\mbox{--}}^{2}{{\rm{P}}}_{1/2}$	5.646	2.394	2.123	2.032	1.929	1.903	1.924
${{}^{2}{{\rm{D}}}^{0}}_{5/2}{\mbox{--}}^{2}{{\rm{P}}}_{1/2}$	1.190	0.2343	0.3052	0.355	0.4641	0.547	0.6547
${{}^{2}{{\rm{D}}}^{0}}_{3/2}{\mbox{--}}^{2}{{\rm{P}}}_{3/2}$	0.0002	3.196	3.405	3.512	3.693	3.796	3.879
${{}^{2}{{\rm{D}}}^{0}}_{5/2}{\mbox{--}}^{2}{{\rm{P}}}_{3/2}$	21.699	1.437	1.402	1.383	1.351	1.334	1.323
${{}^{2}{{\rm{D}}}^{0}}_{5/2}{\mbox{--}}^{2}{{{\rm{D}}}^{0}}_{3/2}$	30.410	2.151	2.047	1.995	1.919	1.885	1.864

Download table as:  ASCIITypeset image

We recalculated the physical conditions, and ionic and elemental abundances of O, S, and Ar that are used in our abundance analysis, using PyNeb and line fluxes from multi-wavelength studies of IC 418 (Bernard Salas et al. 2001; Pottasch et al. 2004) and NGC 7027 (Zhang et al. 2005). Multi-wavelength observations provide access to multiple ions per element, which minimizes systematic uncertainties due to the contributions of unobserved ions. We take the total elemental abundances of O in IC 418, and of Ar and S in both PNe, to be equal to the sum of all observed ionic abundances. For the high-excitation PN NGC 7027, we correct for the presence of unobserved O⁴⁺ using the formula from Delgado-Inglada et al. (2014). The atomic data used and the treatment of uncertainties are the same from the literature. The two sets of n-capture abundances agree to within 0.1 dex in all cases.

Table 3.  n-capture Ionic and Elemental Abundances

Line (μm) Ratio	Flux Ratio	log(Xi+/H+)+12	Ionization Correction Factora	log(X/H) + 12a	log(X/H) + 12b
				This Work
NGC 7027 (${T}_{{\rm{e}}}$ = 12450 ± 600 K, ne = ${50500}_{-14600}^{+25700}$ cm−3)
EMIR
[Se iv] 2.2864/Brγ	0.0919 ± 0.0055	3.20 ± 0.04	2.52 ± 1.12	3.60 ± 0.16	3.56 ± 0.14
[Br v] 1.6429/Br11	0.0295 ± 0.0021	2.23 ± 0.05	2.80 ± 0.56	2.67 ± 0.06	2.72 ± 0.06
[Kr iii] 2.1986/Brγ	0.0329 ± 0.0023	3.40 ± 0.04	1.32 ± 0.25	4.13 ± 0.08	4.11 ± 0.08
[Kr vi] 1.2330/Pβ	0.0078 ± 0.0005	2.55 ± 0.04	⋯	⋯	⋯
[Rb iv] 1.5973/Br11	0.0335 ± 0.0023	2.86 ± 0.05	1.58 ± 0.33	3.06 ± 0.09	3.03 ± 0.09
[Te iii] 2.1019/Brγ	0.0030 ± 0.0008	1.50 ± 0.11	17.28 ± 2.94	2.74 ± 0.12	2.64 ± 0.12
IGRINS
[Se iv] 2.2864/Brγ	0.0835 ± 0.0083	3.16 ± 0.05	2.52 ± 1.12	3.56 ± 0.17	3.52 ± 0.14
[Br v] 1.6429/Br11	0.0245 ± 0.0049	2.15 ± 0.08	2.80 ± 0.56	2.59 ± 0.09	2.64 ± 0.09
[Kr iii] 2.1986/Brγ	0.0307 ± 0.0031	3.38 ± 0.05	1.32 ± 0.25	4.12 ± 0.09	4.10 ± 0.08
[Rb iv] 1.5973/Br11	0.0290 ± 0.0029	2.80 ± 0.06	1.58 ± 0.33	3.00 ± 0.10	2.97 ± 0.10
[Te iii] 2.1019/Brγ	0.0031 ± 0.0003	1.51 ± 0.05	17.28 ± 2.94	2.75 ± 0.07	2.65 ± 0.07
Literaturec
[Te iii] 0.7933/Hβ	<0.002	<2.15	⋯	⋯	⋯
IC418 (${T}_{{\rm{e}}}$ = 8670 ± 250 K, ne = ${13650}_{-3400}^{+8200}$ cm−3)
EMIR
[Se iv] 2.2864/Brγ	0.0019 ± 0.0008	1.72 ± 0.15	5.81 ± 2.14	2.48 ± 0.19	2.46 ± 0.18
[Kr iii] 2.1986/Brγ	0.0258 ± 0.0015	3.51 ± 0.03	1.82 ± 0.32	3.77 ± 0.06	3.77 ± 0.06
[Te iii] 2.1019/Brγ	0.0122 ± 0.0011	2.29 ± 0.04	2.58 ± 0.31	2.70 ± 0.07	2.67 ± 0.07
IGRINS
[Se iv] 2.2864/Brγ	0.0014 ± 0.0001	1.57 ± 0.04	5.81 ± 2.14	2.33 ± 0.14	2.31 ± 0.14
[Kr iii] 2.1986/Brγ	0.0262 ± 0.0026	3.51 ± 0.04	1.82 ± 0.32	3.77 ± 0.07	3.77 ± 0.07
[Te iii] 2.1019/Brγ	0.0126 ± 0.0013	2.31 ± 0.04	2.58 ± 0.31	2.72 ± 0.07	2.69 ± 0.07
Literatured
[Te iii] 0.7933/Hβ	0.0017	2.53	⋯	⋯	⋯

Notes.

^aResults from our abundance re-calculations. For details of the adopted ionization correction factors (ICFs), see Section 5.1. ^bCalculated using values for physical conditions and for ionic and total abundances of light elements from the literature (see Section 5.1). ^cSharpee et al. (2007). ^dSharpee et al. (2003).

Download table as:  ASCIITypeset image

5.1. Elemental Abundances of Te and Br

5.2. Abundances of Other n-capture Elements

5.3. Comparison to Theoretical Models of AGB Evolution

5.4. Abundances and Origins of Br and Te in Astronomical Sources

S.M. and J.G.R. acknowledge support from the Spanish Ministerio de Economía y Competividad under projects AYA2015-65205-P and AYA2017-83383-P. M.A.B. and N.C.S. acknowledge support from NSF award AST-1412928, and H.L.D. from AST-1715332. J.G.R. acknowledges support from the Severo Ochoa excellence program (SEV-2015-0548). This work used IGRINS, developed by the University of Texas at Austin and the Korea Astronomy and Space Science Institute (KASI) with the financial support of NSF grant AST-1229522, the University of Texas at Austin, and the Korean GMT Project of KASI. This work has made use of the NASA Astrophysics Data System.