The unidentified infrared emission feature peaking near 20.1 μm — the "21 μm feature" — was discovered in 1989 in carbon-rich protoplanetary nebulae and remains one of the longest-standing carrier-identification problems in circumstellar dust research. Zhang, Jiang & Li (2009) tested nine mostly inorganic candidate materials against elemental abundance budgets and secondary-feature constraints, ruling out eight and leaving nano FeO as the sole viable candidate, but they explicitly could not test the organic candidates — urea/thiourea, polycyclic aromatic hydrocarbons (PAHs), hydrogenated amorphous carbon (HAC), and PAH clusters — because the required optical data were missing: 21 μm band strengths relative to other characteristic bands and ultraviolet/visible optical constants governing stochastic heating, with the paper further caveating that realistic grain shapes, clustering, and morphologies can change the effective band-integrated opacities. The quantitative abundance-budget and secondary-feature tests that Zhang, Jiang & Li (2009) applied to the inorganic candidates were never extended to the organic candidates (including the fulleranes proposed subsequently), and the optical data those tests require remain unavailable. Later work refuted FeO (2013), argued indirectly against PAH clusters and a thiourea-type common carrier (2015), and found fullerane spectral fits promising but explicitly inconclusive (2020, 2024); no accepted identification exists as of 2026.

Background

When a low- or intermediate-mass star exhausts its nuclear fuel, it ascends the asymptotic giant branch (AGB), an evolutionary phase dominated by pulsation and heavy mass loss, and then transitions through a short-lived (about a thousand-year) post-AGB or protoplanetary nebula (PPN) stage before the ejected envelope is ionized and becomes a planetary nebula. During this transition the detached circumstellar envelope of gas and dust reradiates the starlight it absorbs in the infrared. In 1989, Kwok, Volk & Hrivnak discovered a strong, narrow dust emission band peaking at about 20.1 μm with a full width at half maximum of about 2.2–2.3 μm in four carbon-rich PPNe. This "21 μm feature" — the name is historical, the peak lies near 20.1 μm — is now known in about two dozen PPNe in the Milky Way and the Magellanic Clouds; the sources are metal-poor, carbon-rich F and G supergiants with strong infrared excesses, and the strongest source, HD 56126, emits roughly 8% of its total infrared power in this one feature. The feature shows remarkably little shape variation from source to source, and it is always accompanied by the unidentified infrared (UIR) emission bands at 3.3, 6.2, 7.7, 8.6 and 11.3 μm — generally attributed to polycyclic aromatic hydrocarbon (PAH) molecules, large carbon-bearing ring structures — and by an equally mysterious 30 μm feature.

More than a dozen carrier materials have been proposed. Among inorganics: titanium carbide (TiC) nanocrystals, whose laboratory-measured resonance profile matches the observed feature almost perfectly (von Helden et al. 2000); silicon disulfide (SiS2_2); silicon carbide (SiC) doped with impurities; SiC cores with SiO2_2 mantles; and the oxides FeO, Fe2_2O3_3, Fe3_3O4_4 (Posch, Mutschke & Andersen 2004). Among organics: urea or thiourea, PAHs, hydrogenated amorphous carbon (HAC, a disordered carbonaceous solid whose optical properties depend on its hydrogen content and on the ratio of sp2^2 to sp3^3 bonded carbon), and clusters of PAH molecules. None has been accepted: proposals mostly rest on matching the spectral profile, which is fragile because the profile of any candidate can be altered by grain size, shape, clustering, and porosity.

Zhang, Jiang & Li (2009) introduced a complementary, more robust test. For a candidate dust material they used the observed total power EtotE_{\rm tot} emitted in the 21 μm band and the material's mass absorption coefficient κabs(λ)\kappa_{\rm abs}(\lambda) (opacity per unit dust mass), in particular its integral over the band, κabsint\kappa_{\rm abs}^{\rm int}, to compute the minimum abundance [X/H]min⁡[\mathrm{X/H}]_{\min} of the key element X (e.g. Ti, S, Si, Fe) that must be locked in that dust to supply the observed power, at a plausible dust temperature TdT_{\rm d}. This is compared with the observationally available abundance [X/H]⋆[\mathrm{X/H}]_{\star} in the source's atmosphere and shell (taken from detailed abundance analyses; for HD 56126, [Ti/H] ≈ 0.013 ppm, [S/H] ≈ 4.07 ppm, [Fe/H] ≈ 3.24 ppm). A candidate whose [X/H]min⁡[\mathrm{X/H}]_{\min} exceeds the available [X/H]⋆[\mathrm{X/H}]_{\star} fails on abundance grounds; a candidate that also emits strong features at other wavelengths not seen in the 21 μm sources fails on secondary features. Crucially, this abundance-budget test is insensitive to profile-shape uncertainties from grain morphology. For very small grains (nanoparticles), the temperature entering the budget is not a single equilibrium value: the grains are stochastically heated by single stellar photons, undergoing temperature spikes and possessing a temperature distribution whose shape depends on the material's absorption at ultraviolet, visible, and near-infrared wavelengths.

Applying this to HD 56126, Zhang, Jiang & Li (2009) ruled out five candidates on abundance grounds (TiC nanoclusters, Ti-coordinated fullerenes, SiS2_2, doped SiC, SiO2_2-mantled SiC), ruled out three on secondary features (carbon-silicon mixtures, Fe2_2O3_3, Fe3_3O4_4), and found nano FeO viable. However, they explicitly could not apply the test to the organic candidates — urea/thiourea, PAHs, and HAC — because the necessary laboratory data do not exist or are too uncertain: the optical properties of urea/thiourea are unknown; the 21 μm band strength of PAHs relative to their UIR bands is poorly determined; and HAC is ill-characterized, its spectra varying with hydrogenation and sp2^2/sp3^3 ratio. They noted that PAH clusters "probably deserve a detailed investigation," and that the stochastic-heating temperatures (hence the derived [X/H]min⁡[\mathrm{X/H}]_{\min}) of several candidates depend on ultraviolet/visible optical constants that have not been measured, while realistic grain shapes, clustering, and morphologies can change the effective κabsint\kappa_{\rm abs}^{\rm int}.

Subsequent work transformed the landscape but did not close the question. Li, Liu & Jiang (2013) refuted the 2009 paper's favored candidate: modeling stochastically heated FeO around HD 56126, they found FeO emits too broad a feature and requires more Fe than is available. Mishra, Li & Jiang (2015) showed the fluxes of the 21 μm, 30 μm, and UIR features do not correlate with one another, arguing against a common carrier (such as thiourea) for the 21 and 30 μm features and not supporting large PAH clusters as the 21 μm carrier — an indirect, not quantitative, constraint. Mishra, Li & Jiang (2016) found both features correlate with the AGB-phase mass-loss rate, suggesting the carriers form during the AGB and that the 21 μm carrier may require ultraviolet photons for excitation. Zhang (2020) and Liao, Zhang & Sadjadi (2024) investigated fulleranes (hydrogenated fullerenes, C60_{60}Hm_m) through theoretically simulated vibrational spectra: fits to nine 21 μm sources favor moderately hydrogenated fulleranes over PAHs, but the authors state explicitly that no firm conclusion is possible at present. Molecular-line surveys of the 21 μm sources (Qiu et al. 2024, 2025; Sun et al. 2026) characterize their gas environments (e.g. CS is extremely scarce) without identifying the carrier. Thus, as of 2026 the carrier is unidentified, no candidate from the 2009 inorganic tests survives, and the quantitative abundance-budget and secondary-feature testing of the organic candidates — the explicit gap left by Zhang, Jiang & Li (2009) — has not been carried out.

Problem Statement

Determine which, if any, of the organic candidate carriers of the 21 μm emission feature in carbon-rich post-AGB stars — urea/thiourea, polycyclic aromatic hydrocarbons and PAH clusters, hydrogenated amorphous carbon, and the fulleranes (C60_{60}Hm_m) — is consistent with the observed 21 μm emission of those sources given the available elemental abundances and the observed secondary-feature constraints. For each candidate the relevant properties are its 21 μm band-integrated mass absorption coefficient κabsint\kappa_{\rm abs}^{\rm int} and its band strengths relative to its other characteristic emission bands, which set its predicted secondary features, together with the ultraviolet/visible/near-infrared absorption properties that determine the temperature distributions of these stochastically heated nanoparticles in post-AGB radiation fields. For the strongest 21 μm sources (at least HD 56126), the minimum required elemental abundance [X/H]min⁡[\mathrm{X/H}]_{\min} is compared with the observationally available [X/H]⋆[\mathrm{X/H}]_{\star} and any predicted secondary features with the observed spectra, establishing for each candidate the quantitative grounds (abundance margin, predicted-but-unobserved features, profile match) bearing on its viability.

References

  1. Zhang, K., Jiang, B. W., Li, A. (2009). On the carriers of the 21 μm emission feature in post-asymptotic giant branch stars. Monthly Notices of the Royal Astronomical Society 396(3), 1247–1256. DOI: 10.1111/j.1365-2966.2009.14808.x. https://doi.org/10.1111/j.1365-2966.2009.14808.x
  2. Kwok, S., Volk, K. M., Hrivnak, B. J. (1989). A 21 micron emission feature in four proto-planetary nebulae. The Astrophysical Journal 345, L51–L54. DOI: 10.1086/185550. https://doi.org/10.1086/185550
  3. von Helden, G., Tielens, A. G. G. M., van Heijnsbergen, D., Duncan, M. A., Hony, S., Waters, L. B. F. M., Meijer, G. (2000). Titanium carbide nanocrystals in circumstellar environments. Science 288(5464), 313–316. DOI: 10.1126/science.288.5464.313. https://doi.org/10.1126/science.288.5464.313
  4. Posch, T., Mutschke, H., Andersen, A. (2004). Reconsidering the origin of the 21 micron feature: oxides in carbon-rich protoplanetary nebulae? The Astrophysical Journal 616, 1167–1180. DOI: 10.1086/425113. https://doi.org/10.1086/425113
  5. Li, A., Liu, J. M., Jiang, B. W. (2013). On iron monoxide nanoparticles as a carrier of the mysterious 21 μm emission feature in post-asymptotic giant branch stars. The Astrophysical Journal 777(2), 111. DOI: 10.1088/0004-637X/777/2/111. https://doi.org/10.1088/0004-637X/777/2/111
  6. Mishra, A., Li, A., Jiang, B. W. (2015). A tale of three mysterious spectral features in carbon-rich evolved stars: the 21 μm, 30 μm, and "unidentified infrared" emission features. The Astrophysical Journal 802(1), 39. DOI: 10.1088/0004-637X/802/1/39. https://doi.org/10.1088/0004-637X/802/1/39
  7. Mishra, A., Li, A., Jiang, B. W. (2016). On the relation between the mysterious 21 μm emission feature of post-asymptotic giant branch stars and their mass-loss rates. The Astrophysical Journal 825(1), 68. DOI: 10.3847/0004-637X/825/1/68. https://doi.org/10.3847/0004-637X/825/1/68
  8. Koike, C., Matsuno, J., Chihara, H. (2017). Variations in the infrared spectra of wüstite with defects and disorder. The Astrophysical Journal 845(2), 115. DOI: 10.3847/1538-4357/aa7deb. https://doi.org/10.3847/1538-4357/aa7deb
  9. Zhang, Y. (2020). Are fulleranes responsible for the 21 micron feature? Chinese Journal of Chemical Physics 33(1), 101–106. DOI: 10.1063/1674-0068/cjcp1911203. https://doi.org/10.1063/1674-0068/cjcp1911203
  10. Liao, X.-X., Zhang, Y., Sadjadi, S. (2024). Exploring the possibility of polycyclic aromatic hydrocarbons and fulleranes as the carrier of the 21 micron emission feature. Frontiers in Astronomy and Space Sciences 11, 1489982. DOI: 10.3389/fspas.2024.1489982. https://doi.org/10.3389/fspas.2024.1489982
  11. Qiu, J.-J., Zhang, Y., Nakashima, J., Zhang, J.-S., Li, F., Lu, D.-R., Tang, X.-D., Yu, X.-L., Jia, L.-W. (2024). Gas-phase molecules in protoplanetary nebulae with the 21 μm emission feature. The Astronomical Journal 167(3), 91. DOI: 10.3847/1538-3881/ad198f. https://doi.org/10.3847/1538-3881/ad198f
  12. Qiu, J.-J., Zhang, Y., Lu, D.-R., Chang, Z.-X., Zhang, J.-S., Li, X.-H., Tang, X.-D., Qiu, Y., Nakashima, J., Jia, L.-W. (2025). Gas-phase molecules in protoplanetary nebulae with the 21 μm emission feature. II. Carbon monosulfide. The Astronomical Journal 170, 230. DOI: 10.3847/1538-3881/adfdda. https://doi.org/10.3847/1538-3881/adfdda
  13. Sun, H.-M., Zhang, Z., Liu, J.-Y., Qin, S.-L., Zhang, Y. (2026). Interferometric molecular line observations toward the 21 μm protoplanetary nebula IRAS 06530-0213. arXiv:2607.25304. https://arxiv.org/abs/2607.25304