Fullerene Solves an Interstellar Puzzle

RESEARCH NEWS & VIEWS

ASTROCHEMISTRY were attributed to known effects caused by the neon matrix. Nonetheless, the interstellar- medium study2 estimated that about 0.6% of + interstellar carbon may be in the form of C60 , Fullerene solves an and proposed that fullerenes may have a signif- icant role in interstellar chemistry. Arguments + to support the C60 match to the two strongest interstellar puzzle near-infrared DIBs also include the species’ extreme resistance to ultraviolet radiation, Laboratory measurements confirm that a ‘buckyball’ ion is responsible for two that the strengths and widths of the two DIBs near-infrared absorption features found in spectra of the interstellar medium, were mutually correlated, and that their band casting light on a century-old astrochemical mystery. See Letter p.322 widths were consistent with the spectrum of a rotating fullerene-like molecule2,10,11. To provide unambiguous confirmation, laboratory + PASCALE EHRENFREUND & BERNARD FOING studies, theoretical modelling and astronomical gas-phase spectra of C60 were called for. observations4. As a result, a host of candidate In the current study, Campbell et al. report + ore than 400 unidentified absorption carriers has been proposed, ranging from dust the laboratory synthesis of a complex of C60 features have been recorded in grains to intriguing carbonaceous species, and helium at 5.8 kelvin, providing the best spectra of the interstellar medium of including fullerenes (a carbon allotrope)3,6. approximation so far of the gas-phase spec- + Mthe Milky Way and other galaxies. The nature High-resolution observations of fine spectro- trum of C60 . The laboratory measurements 7,8 + of these ‘diffuse interstellar bands’, which are scopic details of certain DIBs have contributed of C60 band maxima at 9,632.7 and 9,577.5 Å present in spectra of the interstellar medium to the consensus that polyatomic carbonaceous show a stunning match to the observed wave- from the near-ultraviolet to the near-infra- molecules in the gas phase are promising can- lengths of the corresponding interstellar red domains, has evaded elucidation since didate carriers. The discovery and synthesis of DIBs, with an accuracy of ±0.1 Å. Moreover, their discovery nearly 100 years ago — one fullerenes (including buckminsterfullerene) and the respective band widths of 2.2 and 2.5 Å are of the longest-standing mysteries of modern of graphene in laboratories, have revolution- consistent with the 2.8-Å widths measured for astronomy. Writing on page 322 of this issue, ized our understanding of carbon chemistry these DIBs10,11. Twenty-one years after the ini- Campbell et al.1 report laboratory spectra of and its applications ranging from astronomy and tial assignment, the authors’ superb laboratory the buckminsterfullerene, or ‘buckyball’, ion biotechnology to industrial processes. results, which were achieved with a technically + C60 . The spectra provide an excellent match In 1994, a search for the strongest signals of advanced set-up, have conclusively identified + + to two interstellar bands that we discovered C60 in interstellar-material spectra led to the C60 as the carrier of these two DIBs. 2 2 + and attributed to this ion , and securely iden- discovery of two absorption features (Fig. 1) With the two strongest C60 bands securely + tify C60 as a component of the interstellar at 9,577 and 9,632 ångströms. These coincided matched to DIBs, prospects are improving for medium. with independent measurements of electronic the identification of the carriers of the more + During the past 40 years, advances in transitions of C60 in laboratory spectra using than 400 remaining DIB features, which have astronomical instrumentation have enabled frozen-matrix techniques — in which the tar- been recorded in a variety of astrophysical set- the detection of many weak diffuse interstellar get species is diluted with an unreactive host tings. In the past decade, observations across bands (DIBs) (Fig. 1), and their various widths species (such as neon) at very low tempera- the entire optical range have established that have been accurately measured3–5. Numerous tures, and is then spectrally analysed9. This diffuse band carriers are ubiquitous in space. efforts have been made to identify the likely preliminary match of two interstellar DIBs DIBs have also recently been detected and + molecular group (the ‘carrier’) that causes with C60 showed wavelength shifts of 3–10 Å mapped in the Magellanic Clouds (the Milky the absorption bands through laboratory relative to the laboratory measurements; these Way’s satellite galaxies) and in other galaxies. The carrier molecules may therefore constitute an important reservoir of organic material 1.00 a b 1.10 throughout the Universe. The clear identification of additional DIB 0.95 1.05 carriers could have a far-reaching impact on our understanding of astrochemical reaction 0.90 + 1.00 networks. C60 was possibly the easiest and most conspicuous molecule to identify owing 0.85 0.95 to its signature of two strong bands in the * near-infrared range that is clearly evident in 0.80 0.90 interstellar-medium spectra, which have been 2,10,11

Normalized Intensity * corrected for atmospheric absorptions . 0.75 0.85 The species’ cage-like symmetry and geome- 9,580 9,620 9,660 try6 (Fig. 1), as well as its chemical properties, 0.70 increase its chances of survival in space and 4,000 5,000 6,000 7,000 8,000 9,000 A, JAN CAMI; ADAPTED FROM REF. 4/B, PASCALE EHRENFREUND 4/B, PASCALE FROM REF. CAMI; ADAPTED A, JAN its expected natural formation in astrophysi- Wavelength (Å) 4,10

& BERNARD FOING; REF. 10/ESO/JAMES HEDBERG/(CC BY-NC-SA 3.0) HEDBERG/(CC BY-NC-SA 10/ESO/JAMES REF. & BERNARD FOING; cal environments . One possible approach to identify more DIB carriers would be to detect Figure 1 | Buckminsterfullerene in space. a, High-resolution absorption spectra of the interstellar and quantify the presence in space of other medium, such as that shown here, include more than 400 unidentified diffuse interstellar9,600 bands (DIBs, white lines) spread across the optical and near-infrared domains3,5,10. Fullerenes and other fullerene compounds, which could be formed carbon-bearing molecules have been proposed as the absorbing species causing some of the interstellar by the addition of hydrogen atoms, metals or DIBs, but conclusive identification had been elusive. b, Campbell et al.1 now prove that the charged other radicals in different charge or reactive + 10 buckminsterfullerene ion C60 is indeed responsible for the two strongest interstellar DIBs in the states. near-infrared at 9,577 and 9,632 Å (indicated by stars) discovered2 in 1994. In previous analyses of mid-infrared

296 | NATURE | VOL 523 | 16 JULY 2015 | CORRECTED 24 JULY 2015 © 2015 Macmillan Publishers Limited. All rights reserved NEWS & VIEWS RESEARCH emission spectra of a young planetary nebula Astronomical surveys of DIB strengths and 3. Herbig, G. H. Annu. Rev. Astron. Astrophys. 33, (the ionized, ejected envelope of an old star) profiles in different interstellar and circumstel- 19–73 (1985). 4. Cox, N. L. J. & Cami, J. (eds) Proc. Int. Astron. Un. taken with NASA’s Spitzer Space Telescope, lar environments are needed to elucidate the Vol. 9 (2014). researchers for the first time identified12 spe- formation and fate of carriers. All these research 5. Hobbs, L. M. et al. Astrophys. J. 705, 32–45 (2009). cific molecular-vibration bands that were disciplines must be combined to provide further 6. Kroto, H. W. et al. Nature 318, 162–163 (1985). 7. Ehrenfreund, P. & Foing, B. H. Astron. Astrophys. assigned to compounds of the fullerenes C60 insights into the photochemistry, thermal con- 307, L25–L28 (1996). and C70. Vibrational spectroscopy cannot per- ditions and integration in protoplanetary-disk 8. Sarre, P. et al. Mon. Not. R. Astron. Soc. 277, L41–L43 fectly distinguish between molecules of similar material that affect the inventory and evolution (1995). 9. Fulara, J., Jakobi, M. & Maier, J. P. Chem. Phys. Lett., chemical bonds or between molecular excita- of organic molecules from molecular clouds to 211, 227–234 (1993). tion states, but it can provide a global inventory star- and planet-forming regions. ■ 10. Foing, B. H. & Ehrenfreund, P. Astron. Astrophys. of molecular groups from the pool of organic 317, L59–L62 (1997). 11. Galazutdinov, G. A. et al. Mon. Not. R. Astron. Soc. materials in space. Electronic spectroscopy, Pascale Ehrenfreund is at the Space Policy 317, 750–758 (2000). such as that used by Campbell and colleagues, Institute, George Washington University, 12. Cami, J., Bernard-Salas, J., Peeters, E. & Malek, S. E. can pinpoint the fingerprints of individual Washington DC 20052, USA. Bernard Foing Science 329, 1180–1182 (2010). 13. Leger, A. & Puget, J. L. Astron. Astrophys. 137, L5–L8 molecules. Additional observations of cir- is at the European Space Research and (1984). cumstellar nebulae have shown that, under the Technology Centre of the European Space 14. Allamandola, L. J., Tielens, A. G. G. M. & Barker, J. R. right conditions, fullerenes can and do form Agency, ESA-ESTEC, 2200 Noordwijk, Astrophys. J. 290, L25–L28 (1985). 4 15. Vuong, M. H. & Foing, B. H. Astron. Astrophys. 363, efficiently in space . the Netherlands. L5–L8 (2000). Other carbon-bearing molecules known e-mail: [email protected]; 16. Salama, F. et al. Astrophys. J. 728, 154–162 (2011). as polycyclic aromatic hydrocarbons (PAHs) [email protected] 17. Zhen, J., Castellanos, P., Paardekooper, D. M., — fused benzene rings — have also been pro- Linnartz, H. & Tielens, A. G. G. M. Astrophys. J. 797, 13,14 1. Campbell, E. K., Holz, M., Gerlich, D. & Maier, J. P. L30–L34 (2014). posed as carriers of vibrational bands that Nature 523, 322–328 (2015). 18. Berné, O. & Tielens, A. G. G. M. Proc. Natl Acad. Sci. are observed in the mid-infrared domain. 2. Foing, B. H. & Ehrenfreund, P. Nature 369, 296 (1994). USA 109, 401–406 (2012). These are estimated to be the most abundant complex organic molecules in space, compris- ing more than about 15% of cosmic carbon, SYSTEMS BIOLOGY and have also been proposed as potential + DIB carriers. However, in contrast to C60 , PAHs can undergo partial loss of hydrogen atoms (dehydrogenation), and would there- Network evolution fore exist in a large number of molecular isomers, diluting their spectral signatures15. This might explain why, despite several dedicated hinges on history efforts16, no specific PAH electronic feature has been clearly associated with DIBs. Some The effects of mutations in proteins can depend on the occurrence of previous metal–PAH complexes may eventually be mutations. It emerges that such historical contingency is also important during shown to have special molecular configura- the evolution of gene regulatory networks. See Letter p.361 tions that have few isomers and strong oscil- lator strengths — these would have undiluted spectral features, which would permit their AARON M. NEW & BEN LEHNER evolution of gene regulatory networks has not electronic spectroscopic identification and been well studied. matching to astrophysical DIBs4. nyone who has dealt with bureaucracy Budding yeasts, a group of fungi that The relationships between PAHs and fuller- knows that the order in which events includes the model organism Saccharomyces enes in space represent an intriguing puzzle. occur can be vital to achieving a given cerevisiae (brewer’s yeast), have two mating 17 + Laboratory studies have shown that C60 Atask. This is also true for evolution, in which types, ‘a’ and ‘α’. When exposed to mating can form following sequential light-induced certain mutations are sometimes beneficial, pheromone, both mating types switch on genes fragmentation of large PAH cations. Infra- or even viable, only if other mutations have necessary for mating, and a cells also switch on red observations obtained by the Spitzer and occurred first1 — a phenomenon known as genes specifically required for mating with α the European Space Agency’s Herschel Space historical contingency. In this issue, Sorrells cells, referred to as a-specific genes (asgs). The 18 2 Observatory have also shown that C60 can et al. (page 361) show how historical contin- identity of these asgs has changed little over form in the tenuous and cold environment gency constrained the evolution of an entire hundreds of millions of years of evolution2,6,7, of an interstellar cloud illuminated by strong gene regulatory network. but the way in which their expression is ultraviolet starlight — possibly as the product Historical contingency has long been controlled has altered considerably. of the conversion of PAHs to graphene, and thought3 to have a role in evolution, but only Sorrells et al. reconstructed the evolution then to fullerenes. since the advent of large-scale DNA sequen of asg regulation in S. cerevisiae. They showed As Campbell and colleagues’ findings show, cing have we been able to peer back in time that a regulatory protein, Ste12, initially acti- the identification of space molecules can be and replay the evolutionary tape to establish vated asgs indirectly by binding to a DNA- achieved only by measuring laboratory spectra how the process works in biological systems. binding protein called a2, and then evolved to under astrophysically relevant conditions, and A wealth of DNA-sequence information bind directly to DNA. By constructing plau- the spectra must unambiguously match the means that we can now infer the sequences sible evolutionary intermediates, the authors astronomical data. This task requires theoretical of extinct genes, allowing us to resurrect the provide evidence that this change in Ste12 modelling to determine other plausible fuller- proteins that they encoded4. Studying the role function was made possible by another change, ene- and PAH-candidate DIB carriers and a sys- of these ancestral proteins in living cells has in the way the two mating types are defined. tematic programme of laboratory spectroscopy revealed that only certain paths can be taken Previous work from this group6,7 found that, of these in frozen neon matrices (for screen- during the evolution of protein functions5. But in ancestral species, a cells were defined by the ing) and in the gas phase (for identification). the part played by historical contingency in the presence of a2. But in S. cerevisiae, a cells are