Discovery of the interstellar chiral molecule propylene oxide (CH3CHCH2O)

See allHide authors and affiliations

Science  17 Jun 2016:
Vol. 352, Issue 6292, pp. 1449-1452
DOI: 10.1126/science.aae0328

Chiral molecule discovered in space

A chiral molecule is one that has two forms that are mirror images of each other: enantiomers. Biological systems overwhelmingly use one enantiomer over another, and some meteorites show an excess of one type. The two forms are almost identical chemically, so how this excess first arose is unknown. McGuire et al. used radio astronomy to detect the first known chiral molecule in space: propylene oxide. The work raises the prospect of measuring the enantiomer excess in various astronomical objects, including regions where planets are being formed, to discover how and why the excess first appeared.

Science, this issue p. 1449


Life on Earth relies on chiral molecules—that is, species not superimposable on their mirror images. This manifests itself in the selection of a single molecular handedness, or homochirality, across the biosphere. We present the astronomical detection of a chiral molecule, propylene oxide (CH3CHCH2O), in absorption toward the Galactic center. Propylene oxide is detected in the gas phase in a cold, extended molecular shell around the embedded, massive protostellar clusters in the Sagittarius B2 star-forming region. This material is representative of the earliest stage of solar system evolution in which a chiral molecule has been found.

The origin of homochirality is a key mystery in the study of our cosmic origins (1). Although homochirality is itself evolutionarily advantageous (2), the mechanism for the selection of one iso-energetic enantiomer over another is uncertain. Many routes to homochirality have been proposed through the amplification and subsequent transfer of a small primordial enantiomeric excess (e.e.). Disentangling these possible mechanisms requires that we understand the potential sources from which an e.e. may arise. The oldest materials on which e.e. data have been taken in the laboratory are meteoritic samples (3), yet the provenance of this e.e. remains a matter of considerable debate (4). Material in molecular clouds, from which planetary systems form, is processed through circumstellar disks (5) and can subsequently be incorporated into planet(esimal)s (6). Thus, a primordial e.e. found in the parent molecular cloud may be inherited by the fledgling system. Constraining the origin of any e.e. found in meteorites therefore requires the determination of the possible contributions of a primordial e.e. and thus the detection of a chiral molecule in these environments.

For the past 50 years, radio astronomy has been the primary method for studying the gaseous, complex molecular content of interstellar clouds. In this regime, observed spectral features correspond to fine-structure transitions of atoms, or pure rotational transitions of polar molecules, that can uniquely identify their carrier. The observations presented here were taken toward the Sagittarius B2 North [Sgr B2(N)] molecular cloud, the preeminent source for new complex-molecular detections in the interstellar medium (ISM).

Propylene oxide (Fig. 1) was initially detected using data from the publicly available Prebiotic Interstellar Molecular Survey (PRIMOS) project at the Green Bank Telescope (GBT), which provides nearly frequency-continuous, high-resolution, high-sensitivity spectral survey data toward Sgr B2(N) from 1 to 50 GHz (7). Based on our model of rotationally cold propylene oxide absorbing against the Sgr B2(N) continuum (8), only three transitions are predicted to have appreciable intensity above the survey noise floor: the b-type Q-branch 11,0 – 10,1, 21,1 – 20,2, and 31,2 – 30,3 transitions at 12.1, 12.8, and 14.0 GHz (λ = 2.478, 2.342, and 2.141 cm), respectively (8). The 11,0 – 10,1 line at 12.1 GHz is obscured by radio frequency interference (RFI) at the GBT; however, clear absorption signatures are observed from the 21,1 – 20,2 and 31,2 – 30,3 transitions (Fig. 2).

Fig. 1 The molecular structure of S-propylene oxide and R-propylene oxide.

(A) S-propylene oxide. (B) R-propylene oxide. Carbon, hydrogen, and oxygen atoms are indicated by gray, small white, and red spheres, respectively.

Fig. 2 Observations of the 11,0 – 10,1 (Parkes), 21,1 – 20,2 (GBT), and 31,2 – 30,3 (GBT) transitions of propylene oxide, in absorption, toward the Galactic center.

The 64 km s–1 systematic velocity characteristic of Sgr B2(N) is indicated by a vertical red line. The 11,0 – 10,1 transition of propanal is also seen in the Parkes data. Data are given as antenna temperature (Embedded Image) as a function of shift from local standard of rest velocity (VLSR), where 0 km s–1 is the measured laboratory frequency of the transition (8), and have been Hanning smoothed.

These features may be sufficient for a detection on their own at these wavelengths; however, we endeavored to confirm the detection by observing the 11,0 – 10,1 line at 12.1 GHz using the Parkes Radio Telescope (see the supplementary materials for a detailed description of the Parkes observations), which does not suffer from RFI in the region of the line. The data confirm the presence of a feature at the same velocity (~64 km s–1) as the transitions from PRIMOS, as well as fortuitously detecting a nearby feature of propanal, a structural isomer of propylene oxide (Fig. 2). The far larger Parkes beam (~115′′ versus 60′′) encompasses a much larger sample of environments, inhomogeneously broadening the observed transition and incorporating a second, distinct ~46 km s–1 component not seen by the GBT beam but previously observed in the material surrounding Sgr B2 (Fig. 3) (9).

Fig. 3 Illustration of source structure within the Sgr B2 region.

(A) The GBT and Parkes beams probe different portions of the cold molecular shell in front of the bright continuum sources/hot cores within Sgr B2. Molecules in the shell that are not backlit by continuum sources are not seen in absorption. (B) As the schematic of the sky view shows, the GBT (red) and Parkes (blue) beams probe different continuum sources, with the GBT beam probing only Sgr B2(N), whereas the Parkes beam also includes most of Sgr B2(M) to the south.

A fit to the observations using a single excitation temperature for propylene oxide finds a column density of NT = 1 × 1013 cm–2 and a rotational excitation temperature of Tex = 5 K (8). Although an excitation temperature of 5 K is indeed the best fit to the data, we note that the most rigorous constraint on Tex is from the nondetection at higher frequencies in PRIMOS, giving an upper limit of Embedded Image35 K. Changes in Tex greatly affect NT, and model parameters that fit the data nearly as well are possible for excitation conditions between Tex = 5 and 35 K. These models all reproduce the observed features from the GBT and Parkes and are consistent with the nondetection of propylene oxide at 3 mm; under these conditions, no transitions of propylene oxide would be detectable in the reported observations (10).

A search of spectral line catalogs reveals no reasonable interfering transitions from other molecular species. Propylene oxide is an asymmetric rotor with modest rotational constants and therefore has numerous (~450) transitions that fall within the PRIMOS data. For lower excitation temperatures (Tex ~ 10 to 35 K), at most 80 have an intensity ≥1% of the strongest predicted line. Of these, ~13% are unobservable due to a lack of available receivers at the GBT. Inspection of the entire PRIMOS data set showed no absorption or emission features attributable to propylene oxide at any of these frequencies except the three listed above, in good agreement with the model and the sensitivity of the survey.

This detection is complementary to the upper limit placed by (10) on the nondetection of warm, compact propylene oxide at Tex = 200 K toward Sgr B2(N) at mm wavelengths using the Mopra Telescope. This search was sensitive only to a warm population of propylene oxide, however, and resulted in a nondetection, with an upper limit column density of 6.7 × 1014 cm–2 for an excitation temperature of Tex = 200 K and compact source size (5′′), such as that expected for gas associated with the embedded protostellar clusters/hot cores in this cloud (10).

In sources with strong background continuum, of which Sgr B2(N) is a prominent example, many rotationally cold, high-dipole moment species are observed almost exclusively in absorption against the continuum source, as shown in Fig. 3. Because of the exceptionally low line densities, only two to five well-measured centimeter-wavelength spectral features are needed to securely claim a detection [see, e.g., (1113)]. This stands in stark contrast to mm-wave detections, particularly toward Sgr B2(N), where dozens of lines are typically required. Based on a statistical analysis of the line density in our observations of Sgr B2(N), we find the likelihood of three random features falling within three resolution elements of the propylene oxide transitions to be ≤6 × 10–8 (8).

Taken together, the GBT and Parkes observations provide strong evidence of cold, low-abundance propylene oxide toward the Sgr B2 cloud complex, in excellent agreement with previously established upper limits, as well as with previous observations of complex organic molecules. Indeed, many of the complex organics seen toward Sgr B2(N) are found not in or near the hot cores, but, like propylene oxide, in a cold, extended shell around the source. In these regions, molecules are often liberated into the gas phase via nonthermal, shock-driven desorption, resulting in colder, spatially extended gas-phase populations that are often more abundant than predicted by standard warm-up models (14). This is consistent with the observation that the structurally similar ethylene oxide is consistently found to have low excitation temperatures (11 to 35 K), well below the temperature of the surrounding grains (15), with the detections of glycolaldehyde (11), ethanimine (16), and propylene oxide’s structural isomers propanal (13) and acetone (8), in this region, and with the general pattern of shock-driven liberation of complex molecules in the so-called central molecular zone (14).

From a chemical perspective, the presence of propylene oxide in Sgr B2(N) is not surprising. Propylene oxide is the third species of the C3H6O family detected toward this source. Its structural isomers, propanal (CH3CH2CHO) (13) and acetone [(CH3)2CO] (17), are both seen toward Sgr B2(N), and propylene oxide is not the first epoxide found in the ISM. Ethylene oxide (CH2OCH2) is structurally similar to propylene oxide, differing by only a methyl group, and has been detected toward numerous massive star-forming regions, including Sgr B2(N) (15, 18). In the case of acetone, Belloche et al. (19) report a column density of NT = 1.49 × 1017 cm–2, but for a warm population with Tex = 100 K that peaks at the position of the hot core. In the detection of propanal, column densities were not determined (13).

To determine the relative populations of these molecules in the cold shell around Sgr B2(N), we have used the same procedure as for propylene oxide (8). We find that a column density of NT = 6 × 1013 cm–2 with Tex = 6.2 K reproduces the 11 propanal transitions observed in the full PRIMOS data set (see the supplementary materials) to within a factor of Embedded Image2. Similarly, using 18 detected lines of acetone in PRIMOS (see the supplementary materials), we find that a column density of NT = 2.1 × 1014 cm–2 with Tex = 6.2 K reproduces these features within a factor of ~2. The best-fit Tex = 5 K for propylene oxide is in remarkably good agreement with these values, which, due to the larger number of observed transitions over a wider frequency range for propanal and acetone, are much more rigorously constrained. Tex up to 35 K for propylene oxide is formally allowed in the propylene oxide fit, due to loose constraints stemming from the narrow range of energy levels covered in a narrow frequency window. However, the best-fit, Tex = 5 K, is greatly bolstered by the similar conditions exhibited by the acetone and propanal populations.

All three members of the C3H6O family are then detected in absorption in the PRIMOS data at remarkably similar excitation conditions, suggesting that they likely occupy the same cold, shocked region surrounding Sgr B2(N). Propanal and acetone are thermodynamically favored over propylene oxide, residing 22.7 and 30.8 kcal mol–1 (0.98 and 1.33 eV) lower in energy, respectively (20). However, although the relative column densities derived here do roughly follow the pattern of increasing abundance with increasing stability, chemistry in molecular clouds is largely kinetically controlled, rather than thermodynamically, and relative abundances do not regularly follow thermodynamic patterns (21, 22). The recent detection of acetone and propanal at an abundance ratio of three to one in comet 67P/Churyumov-Gerasimenko shows that members of the C3H6O family also feature prominently in the volatile organic content of comet nuclei, and the remarkably similar ratios to those observed toward Sgr B2(N) suggest that such kinetically controlled routes to both species are widespread and not isolated to extraordinary interstellar sources (23).

The leading models for the production and enhancement of an e.e. in the interstellar medium likely act over time scales far longer than the delivery of complex organic material to the planet-forming region of disks (2426). A number of mechanisms have been proposed for gas-phase routes in the ISM to create such a primordial e.e. Although beta decay–related chemistry has been proven to generate slight chiral asymmetries (25) that would be universal in nature, perhaps the most intriguing route, astronomically, is enantiomerically selective photochemistry induced by circularly polarized light (CPL) (24). Here, the chirally sensitive chemical reaction networks would be stochastically driven on the spatial scales of giant molecular cloud complexes. Toward the Orion Nebula cluster, for example, significant CPL patterns capable of producing e.e. do not extend over the entire protostellar cluster but have been detected over regions that are large compared with individual protoplanetary disks (26). We have rigorously examined the possible mechanisms for determining an e.e. (see the supplementary materials) and concluded that the standard, total power observations shown here cannot determine whether such an e.e. exists in the case of propylene oxide but that high-precision, full-polarization-state measurements can, in principle. Critically, the detection of gas-phase propylene oxide toward the Galactic center provides a molecular target for such observations and demonstrates that interstellar chemistry can reach sufficient levels of complexity to form chiral species in environments with the physical conditions required to produce an enantiomeric excess.

Supplementary Materials

Materials and Methods

Figs. S1 to S5

Tables S1 to S4

References (2747)

References and Notes

  1. Materials and methods are available as supplementary materials on Science Online.
Acknowledgments: We thank the PRIMOS team, GBT, and Parkes staff for their ongoing support in acquiring the GBT and Parkes data and S. Breen, S. Mader, and J. Reynolds for assistance with Parkes data reduction. We acknowledge the support of L. Snyder, J. M. Hollis, and F. Lovas. B.A.M. thanks J. Mangum and J. Corby for helpful discussions. P.B.C. and B.A.M. acknowledge the support of a NASA Astrobiology Institute Early Career Collaboration Award. B.A.M. is funded by a National Radio Astronomy Observatory Jansky Postdoctoral Fellowship. R.A.L. and I.A.F. are funded by a National Science Foundation Graduate Research Fellowship. P.B.C., I.A.F., and G.A.B. acknowledge support from the NASA Astrobiology Institute through the Goddard Team (M. J. Mumma, PI) under Cooperative Research Agreements NNX09AH63A and NNX15AT33A (NNX09AH63A), and the NSF Astronomy and Astrophysics (AST-1109857) grant program. Access to the entire PRIMOS data set, specifics on the observing strategy, and overall frequency coverage information is available at The spectra obtained with Parkes are available through this website as well. Data from project AGBT06B-006 are available in the NRAO Archive at The National Radio Astronomy Observatory is a facility of the National Science Foundation operated under cooperative agreement by Associated Universities, Inc. The Australia Telescope Compact Array (Parkes Radio Telescope/Mopra Radio Telescope/Long Baseline Array) is part of the Australia Telescope National Facility, which is funded by the Australian government for operation as a National Facility managed by the Commonwealth Scientific and Industrial Research Organisation (CSIRO).

Stay Connected to Science

Navigate This Article