Gas-phase formation of the resonantly stabilized 1-indenyl (C₉H₇^•) radical in the interstellar medium

Abstract

The 1-indenyl (C₉H₇^•) radical, a prototype aromatic and resonantly stabilized free radical carrying a six- and a five-membered ring, has emerged as a fundamental molecular building block of nonplanar polycyclic aromatic hydrocarbons (PAHs) and carbonaceous nanostructures in deep space and combustion systems. However, the underlying formation mechanisms have remained elusive. Here, we reveal an unconventional low-temperature gas-phase formation of 1-indenyl via barrierless ring annulation involving reactions of atomic carbon [C(³P)] with styrene (C₆H₅C₂H₃) and propargyl (C₃H₃^•) with phenyl (C₆H₅^•). Macroscopic environments like molecular clouds act as natural low-temperature laboratories, where rapid molecular mass growth to 1-indenyl and subsequently complex PAHs involving vinyl side-chained aromatics and aryl radicals can occur. These reactions may account for the formation of PAHs and their derivatives in the interstellar medium and carbonaceous chondrites and could close the gap of timescales of their production and destruction in our carbonaceous universe.

The aromatic and resonantly stabilized free 1-indenyl radical is prepared in gas-phase under low-temperature conditions.

INTRODUCTION

For the past decade, resonantly stabilized free radicals (RSFRs), organic open-shell transients in which the unpaired electron is delocalized over the carbon skeleton, have emerged as fundamental molecular building blocks to polycyclic aromatic hydrocarbons (PAHs) and carbonaceous nanoparticles in combustion processes and in the interstellar medium (ISM) (1–6). Reaction pathways from RSFRs such as propargyl (C₃H₃^•), cyclopentadienyl (C₅H₅^•), benzyl (C₇H₇^•), and 1-indenyl (C₉H₇^•) (fig. S1A) to complex PAHs like pyrene (C₁₆H₁₀), fluoranthene (C₁₆H₁₀), and chrysene (C₁₈H₁₂) have been investigated (5, 6). However, the presence of PAHs along with their derivatives such as ionized, hydrogenated, and side-chain substituted counterparts in the ISM and carbonaceous chondrites such as Allende, Orgueil, and Murchison still represents a critical paradox in our Universe (7–9). Timescales of PAH injection from carbon-rich circumstellar envelopes exceed a few 10⁹ years, but the lifetimes of PAHs in deep space are limited by degradation through interstellar shocks, ultraviolet photons, and galactic cosmic rays to only a few 10⁸ years (10–12). In essence, PAHs are destroyed faster than synthesized in carbon-rich asymptotic giant branch stars (12) along with planetary nebulae as their descendants (13). Hence, PAHs should not be observable via, e.g., unidentified infrared emissions (14), in the range from 3 to 14 μm, and diffuse interstellar bands’ discrete emission lines covering the visible (400 nm) to the near-infrared region (1.2 μm) (15) of the electromagnetic spectrum.

Rapid molecular mass growth processes involving RSFRs signify a solution to this puzzle through the synthesis of PAHs in cold molecular clouds at temperatures as low as 10 K (2, 8, 16). Supported by recent observations of the propargyl radical in the Taurus molecular cloud (TMC-1) (17, 18), RSFRs can reach high fractional abundances of up to a few 10⁻⁸ with respect to molecular hydrogen. RSFRs hold lifetimes longer than nonresonantly stabilized hydrocarbon radicals due to the delocalization of the unpaired electron over (part of) the carbon backbone (19) and can undergo rapid radical-radical recombination reactions (20, 21). However, whereas molecular beam experiments (22, 23) and astrochemical modeling (17, 18) provided compelling evidence on the formation of the simplest RSFR in TMC-1, propargyl (C₃H₃^•), via the bimolecular reaction of atomic carbon [C(³P)] with ethylene (C₂H₄), reaction pathways to more complex RSFRs are largely unknown (24). This knowledge is of particular importance for the 1-indenyl (C₉H₇^•) radical (25). As a prototype of an aromatic and RSFR, the bicyclic, fused C5-C6 species represents the key building block of nonplanar PAHs such as corannulene (C₂₀H₁₀), nanobowls (C₄₀H₁₀), and fullerenes (C₆₀ and C₇₀) (fig. S1B) with the five-membered ring in the carbon backbone curving the carbon skeleton out of plane (26, 27). Therefore, an unraveling of the synthetic routes of the 1-indenyl (C₉H₇^•) radical is critical to constrain the population of PAHs delivered from reactions of RSFRs in our Universe.

Here, by exploiting molecular beams experiments (crossed molecular beams; chemical reactor) along with electronic structure calculations and astrochemical modeling, we provide persuasive evidence on the gas-phase formation of the 1-indenyl (C₉H₇^•; X²A₂) radical through the elementary reaction of atomic carbon (C; ³P) with styrene (C₈H₈; X¹A′) (Reaction 1) and via the radical-radical reaction of propargyl (C₃H₃^•; X²B₁) with phenyl (C₆H₅^•; X²A₁) (Reaction 2) (fig. S2). These bimolecular reactions are barrierless and exoergic with all transition states residing lower than the energies of the separated reactants and could proceed rapidly even at molecular cloud conditions of 10 K. These reactions may provide facile routes to 1-indenyl (C₉H₇^•) via cyclization and aromatization reactions at astronomically observable fractional abundances of (6.1 ± 3.5) × 10⁻¹¹ with respect to molecular hydrogen in the TMC-1. As a prototype of an aromatic RSFR, 1-indenyl can undergo molecular mass growth processes leading to more complex aromatic structures such as naphthalene (C₁₀H₈) (3), phenanthrene (C₁₄H₁₀) (4), pyrene (C₁₆H₁₀), and fluoranthene (C₁₆H₁₀) (5). Whereas in low-temperature interstellar environments, single-collision conditions dominate (28), elevated pressures such as in hydrocarbon-rich atmospheres of planets and their moons like Titan facilitate a three-body collisional stabilization of the reaction intermediate(s) with the buffer molecules like molecular nitrogen (N₂) (29, 30). Thus, the route to indene (C₉H₈) formation is opened. These third-body stabilization processes also become increasingly important in high-pressure combustion settings with indene (C₉H₈) ubiquitously identified in hydrocarbon flames of, e.g., ethane (31), n-butane (32), and benzene (33). This investigation fills a critical gap in our knowledge of the underlying elementary mechanisms leading to (aromatic) RSFRs in extreme environments by exploiting the 1-indenyl (C₉H₇^•) radical as a benchmark. These reactions eventually lead to nonplanar PAHs and carbonaceous nanostructures (carbonaceous grains) acting as molecular factories for biorelevant complex organics like amino acids and sugars in deep space (16), thus fundamentally enhancing our understanding of the chemical evolution of carbonaceous matter in the Universe (27)

$$\mathrm{C}(12\mathrm{amu})+{\mathrm{C}}_8{\mathrm{H}}_8(104\mathrm{amu})\to {{\mathrm{C}}_9{\mathrm{H}}_7}^{\bullet }(115\mathrm{amu})+{\mathrm{H}}^{\bullet }(1\mathrm{amu})$$ (1)

$${{\mathrm{C}}_3{\mathrm{H}}_3}^{\bullet }(39\mathrm{amu})+{{\mathrm{C}}_6{\mathrm{H}}_5}^{\bullet }(77\mathrm{amu})\to {{\mathrm{C}}_9{\mathrm{H}}_7}^{\bullet }(115\mathrm{amu})+{\mathrm{H}}^{\bullet }(1\mathrm{amu})$$ (2)

RESULTS

Carbon-styrene system: Laboratory frame

The crossed molecular beam studies of ground state atomic carbon (C; 12 amu) with styrene (C₆H₅C₂H₃; 104 amu) (Fig. 1, A and B, and Reaction 1) revealed reactive scattering signal at mass-to-charge (m/z) ratios of 116 (C₉H₈⁺/¹³CC₈H₇⁺), 115 (C₉H₇⁺/¹³CC₈H₆⁺), and 114 (C₉H₆⁺/¹³CC₈H₅⁺). Ion counts at m/z = 116 and 114 were collected at a level of 9 ± 2% and 78 ± 4% with respect to m/z = 115. Because the TOFs recorded at these m/z values are indistinguishable after scaling, we conclude the existence of a single H-loss reaction channel (Reaction 1). Ion counts at m/z = 116 originate from naturally abundant ¹³CC₈H₆⁺, while m/z = 114 can be accounted for through dissociative electron impact ionization of the C₉H₇ neutral product in the electron impact ionizer to C₉H₆⁺. The laboratory angular distribution (LAD) was, therefore, collected at m/z 115 from 28° to 66° in 2.5° intervals. This distribution spans about 40° within the scattering plane as defined by both beams and displays a forward-backward symmetry with a maximum at the center-of-mass (CM) angle of 52.8 ± 0.6°. This finding suggests indirect reaction dynamics via long-lived C₉H₈ intermediate(s) (Fig. 1, A and B). Because the hydrogen atom can be eliminated from the phenyl (C₆H₅^•) group and/or from the vinyl moiety (C₂H₃^•), crossed beam experiments were also conducted with D3- and D5-styrene, i.e., C₆H₅C₂D₃ and C₆D₅C₂H₃, respectively. These studies reveal that the atomic hydrogen loss can originate from both the vinyl moiety and the phenyl group (Materials and Methods).

Fig. 1. Results for the reaction of atomic carbon (C) with styrene (C₈H₈).

LAD (A), TOF spectra (B), CM translational energy (C), and angular flux distributions (D). Data are recorded at m/z = 115. The black circles with their error bars indicate the normalized experimental distribution, the open circles delineate the experimental data, the red lines represent the best fits obtained, and the shaded areas depict the error limits of the best fit.

CM frame

The interpretation of the laboratory data delivers conclusive evidence of the atomic hydrogen emission from both the vinyl moiety and the aromatic ring of styrene. It is our goal now to elucidate the underlying reaction mechanism(s) and to assign the C₉H₇ product isomer(s) formed in the bimolecular gas-phase reaction. This is accomplished by transforming the laboratory data into the CM frame (34), resulting in the CM translational energy distribution [P(E_T)] and the angular distribution [T(θ)] (Fig. 1, C and D). Best fits of the laboratory data are achieved by exploiting a single-channel fit forming products with a mass combination of 115 amu (C₉H₇) and 1 amu (H). Crucial information on the reaction channel(s) and dynamics is derived through a detailed inspection of these CM functions. First, energy conservation dictates that the high energy cutoff of the P(E_T) of 389 ± 24 kJ mol⁻¹ presents the sum of the collision energy (35.4 ± 1.4 kJ mol⁻¹; table S1) plus the reaction exoergicity for those molecules born without internal excitation. Therefore, a reaction energy of 353 ± 25 kJ mol⁻¹ is revealed. This experimentally derived exoergicity agrees well with the reaction exoergicity of 360 ± 5 kJ mol⁻¹ to form the 1-indenyl radical (C₉H₇^•) plus atomic hydrogen. Second, the P(E_T) displays a pronounced distribution maximum of 20 ± 2 kJ mol⁻¹, revealing a tight exit transition state. Third, the T(θ) has an intensity over the complete angular range and reveals a forward-backward symmetry with a maximum at 90° (sideway scattering); these findings support indirect reaction dynamics involving C₉H₈ complex(es) with a lifetime longer than the rotational period, which undergoes unimolecular decomposition through atomic hydrogen loss nearly parallel to the total angular momentum vector (34).

Propargyl-phenyl system: Mass spectra

The reaction of propargyl (C₃H₃^•) with the phenyl radical (C₆H₅^•) was conducted in a chemical microreactor (3). The microreactor consists of a heated silicon carbide (SiC) tube and is incorporated within the source chamber of a molecular beam apparatus equipped with a Wiley-McLaren reflectron time-of-flight mass spectrometer (Re-TOF-MS). Thermally labile propargyl bromide (C₃H₃Br) and nitrosobenzene (C₆H₅NO) precursor molecules were seeded in helium carrier gas and pyrolyzed in situ through cleavage of the weak carbon-bromine and carbon-nitrogen bonds, respectively, producing selectively helium-seeded propargyl (C₃H₃^•) (20) and phenyl radicals (C₆H₅^•) (35) at a reactor temperature of 1373 ± 10 K (Materials and Methods). The reaction products were expanded supersonically and probed isomer-selectively via tunable vacuum ultraviolet (VUV) light from a synchrotron (3). Control experiments of only helium-seeded nitrosobenzene and only propargyl bromide under identical experimental conditions were conducted (35) too (“blank experiments”) (Fig. 2A). Likewise, a representative mass spectrum recorded at a photon energy of 9.50 eV for the reaction of the phenyl (C₆H₅^•) with propargyl (C₃H₃^•) radical is displayed in Fig. 2B. A careful analysis of these mass spectra provides compelling evidence that the signal at m/z = 115 (C₉H₇⁺) and 116 (C₉H₈⁺/¹³CC₈H₇⁺) originates from the phenyl–propargyl radical–radical reaction but is absent in the control experiments. Accounting for the molecular weight of the reactants (C₆H₅^•, 77 amu; C₃H₃^•, 39 amu) and the products (C₉H₇, 115 amu; C₉H₈, 116 amu), molecules with the formula C₉H₇ are the reaction products of the reaction of propargyl (C₃H₃^•) with phenyl (C₆H₅^•) (Reaction 2) (fig. S2); the formation of C₉H₈ molecules can be linked to adducts and/or their isomerization products of the recombination of phenyl with propargyl radicals, which are then stabilized by a third-body reaction with helium (20).

Fig. 2. Results for the phenyl (C₆H₅^•) with propargyl (C₃H₃^•) reaction.

Photoionization mass spectra recorded at a photon energy of 9.50 eV with mass values of 107 (C₆H₅NO⁺), 108 (¹³CC₅H₅NO⁺), 118 (C₃H₃⁷⁹Br⁺), and 120 (C₃H₃⁸¹Br⁺); the inset highlights ion signal from m/z = 110 to 127 (A). Photoionization mass spectra recorded at 9.50 eV and 1373 ± 10 K (B). Photoionization efficiency (PIE) curves recorded at 116 (C) and 115 (D) for products of interest formed at a reactor temperature of 1373 ± 10 K. The black/blue/green/orange/magenta lines in (C) and (D) represent experimental and reference PIE curves, and the red line resembles the overall fit. The error bars consist of two parts: ±10% based on the accuracy of the photodiode and a 1 σ error of the PIE curve averaged over the individual scans.

Photoionization efficiency spectra

To unravel the nature of the isomer(s) formed, we turn to photoionization efficiency (PIE) curves, which report the intensity of a well-defined ion of a specific m/z ratio as a function of photon energy that is 7.50 to 10.0 eV in this case. The PIE curves for m/z = 116 (C₉H₈⁺/¹³CC₈H₇⁺) (Fig. 2C) and 115 (C₉H₇⁺) (Fig. 2D) can be reproduced via fitting with a linear combination of known reference curve(s) (36) of distinct isomer(s). First, the PIE curve at m/z = 116 can be replicated by the sum of four channels: 3-phenyl-1-propyne [(C₆H₅)CH₂CCH], phenylallene [(C₆H₅)HCCCH₂], indene (C₉H₈), and ¹³C-indenyl (¹³CC₈H₇⁺) (Fig. 2C). Second, the experimentally derived PIE curve at m/z = 115 (Fig. 2D, black) can be reproduced by a single contributor, the 1-indenyl radical (C₉H₇^•, blue) (37, 38). ¹³C-indenyl ion counts contribute to a level of 9.9% with respect to that of 1-indenyl as shown in Fig. 2D. The onset of the ion counts at 8.10 ± 0.05 eV correlates well with the National Institute of Standards and Technology–evaluated adiabatic ionization energy (IE) of indene at 8.14 ± 0.01 eV (39); the adiabatic IE of 3-phenyl-1-propyne [(C₆H₅)H₂CCCH] and phenylallene [(C₆H₅)HCCCH₂] were determined to 8.75 ± 0.05 eV and 8.30 ± 0.05 eV, respectively (40). Therefore, we may conclude that, within our error limits, 3-phenyl-1-propyne, phenylallene, indene, and ¹³C-indenyl contribute to signal at m/z = 116 with the branching ratios of the ion counts of 34 ± 2%, 24 ± 3%, 36 ± 2%, and 6 ± 3%, respectively. It is possible that 3-phenyl-1-propyne and phenylallene can undergo unimolecular decomposition via H loss to form the corresponding C₉H₇ radicals (m/z = 115), namely, 1-phenylpropargyl (1PPR; C₆H₅CHCCH) and 3-phenylpropargyl (3PPR; C₆H₅CCCH₂) radicals. From our fitting results, the match of the reference indenyl PIE with our experimental PIE curve of m/z = 115 demonstrates that indenyl should be the dominant product for m/z = 115 at our high-temperature experimental condition of 1373 ± 10 K. These results are supported by the recent kinetic calculations for the branching ratios for the C₆H₅ plus C₃H₃ reaction at different pressure and temperatures, which reveal that 1PPR and 3PPR are relatively minor products compared to 1-indenyl (41). Until now, there is very limited experimental PIE data for 1PPR and 3PPR (37). A comparison of the available experimental PIE of each reference radical with our experimental PIE curve of 115 is shown in fig. S3. Because the experimental PIE of 1PPR and 3PPR is recorded in a very short range of photon energy (7.5 to 7.8 eV for 3PPR and 7.5 to 9.0 eV for 1PPR), it is not possible to precisely calculate the branching ratios for the indenyl, 1PPR, and 3PPR.

Electronic structure calculations and reaction mechanisms: Carbon-styrene system

The electronic structure calculations of relative energies of local minima, transition states, and products were conducted at an accuracy of ±5 kJ mol⁻¹ (Fig. 3) (42, 43) and reveal that the reaction of atomic carbon with styrene can be initiated on the triplet surface through a barrierless addition of the carbon atom to the π electron density of the carbon-carbon double bond of the vinyl moiety (C₂H₃). This addition leads to a phenyl-substituted triplet cyclopropylidene intermediate i1 stabilized by 225 kJ mol⁻¹ with respect to the separated reactants. In analogy to the reaction with the stem compound ethylene (C₂H₄) (44, 45), the computations also predict that insertion of atomic carbon into any carbon-hydrogen bond of the vinyl or phenyl functional groups is closed. i1 ring opens via a barrier of only 36 kJ mol⁻¹ to the thermodynamically preferred triplet phenylallene (i2). Four reaction pathways are opened from i2: a hydrogen migration from the ortho position of the phenyl ring to the C2 of the side chain forming i3, a 1,2-H migration in the side chain to phenyl-substituted triplet vinylmethylene (i5), and two unimolecular decompositions via H loss from the CH and CH₂ moieties forming 3-phenylpropargyl (p3) and 1-phenylpropargyl (p2) radicals via tight exit transition states. Commencing with i3, cis-trans isomerization of the side chain leads to i4; this is followed by hydrogen migration from the C1 carbon atom of the side chain to the ortho carbon atom of the aromatic ring yielding i5, which can further isomerize via cis-trans isomerization to i6. i6 can unimolecularly decompose via atomic hydrogen elimination to 1-phenylpropargyl (p2) or undergo cyclization to the bicyclic 3aH-indene intermediate (i7). This bicyclic intermediate can lose a hydrogen atom to form 1-indenyl (p1) or isomerize via hydrogen atom migration to triplet indene (i8) before hydrogen atom elimination from the CH₂ moiety to 1-indenyl (p1). Both pathways involve tight exit transition states of 35 and 60 kJ mol⁻¹ with respect to the separated products with the i7 → i8 → p1 + H pathway unfavorable compared to i7 → p1 + H due to the high barrier of 163 kJ mol⁻¹ for the hydrogen atom migration.

Fig. 3. Potential energy surfaces of the C + C₈H₈ reaction.

Potential energy surface for the reaction of ground-state atomic carbon (C; ³P) with styrene (C₈H₈; X¹A′). All energies are given in kilojoules per mole with respect to the energy of the separated reactants.

How do these results compare to our experimental findings? First, the experimentally determined reaction exoergicity for the hydrogen atom loss to C₉H₇ isomers of −353 ± 25 kJ mol⁻¹ agrees well with the computed data of −360 ± 5 kJ mol⁻¹ to yield 1-indenyl (p1) plus atomic hydrogen. Therefore, we can conclude that at least the indenyl radical (C₉H₇^•, p1) is formed in the bimolecular reaction of atomic carbon with styrene. The presence of the theromodynamically less stable isomers 1-phenylpropargyl (p2; −217 ± 5 kJ mol⁻¹) and 3-phenylpropargyl (p3; −208 ± 5 kJ mol⁻¹) can be masked in the low energy section of the CM translational energy distribution. This conclusion also gains support from statistical calculations with Rice-Ramsperger-Kassel-Marcus (RRKM) theory (Materials and Methods) to predict branching ratios within the limits of complete energy randomization (42). This treatment predicts that 1-indenyl (p1) contributes to 54.9% of all products, whereas 1-phenylpropargyl (p2) and 3-phenylpropargyl (p3) account for 41.6 and 3.5%, respectively. It shall be highlighted that alternative indenyl isomers, i.e., 2-indenyl (144 kJ mol⁻¹), 3-indenyl (145 kJ mol⁻¹), 4-indenyl (133 kJ mol⁻¹), 5-indenyl (132 kJ mol⁻¹), 6-indenyl (133 kJ mol⁻¹), and 7-indenyl (132 kJ mol⁻¹), are thermodynamically less stable than 1-indenyl with relative energies indicated in the parentheses (46). Therefore, the formation of 2- to 7-indenyl cannot account for the experimentally derived reaction energy.

Second, the P(E_T) reveals a tight exit transition state in the decomposition of C₉H₈ reaction intermediate(s). The presence of a tight exit transition state is verified computationally for the pathways i7 → i8 → p1 + H and i7 → p1 + H. The reverse reaction, i.e., the recombination of atomic hydrogen with 1-indenyl (p1), formally represents a radical-radical reaction and hence should proceed via atom-radical recombination without an entrance barrier if this reaction proceeds on the singlet surface (47, 48). However, the experimental evidence of a tight exit transition state supports our computations that the reaction proceeds on the triplet surface without intersystem crossing to the singlet manifold.

Third, the sideway scattering is evident from the computed structure of the transition state associated with the fragmentation of i7 to p1 + H (Fig. 3) with the atomic hydrogen leaving nearly perpendicularly to the rotational plane of the decomposing i7 intermediate. Note that the RRKM calculations also predict that 99% of 1-indenyl (p1) originates from i7 → p1 + H and less than 1% from i7 → i8 → p1 + H.

Last, we would like to address the experimental findings of the isotopic substitution studies in light of the computed surface (Fig. 3 and figs. S4 and S5). Branching ratios of the four pathways—(i) i2 → p3 + H (3.5%), (ii) i2 → p2 + H (29.5%), (iii) i6 → p2 + H (12.1%), and (iv) i7 → p1 + H (54.7%)—along with the dominance of 1-indenyl (p1; 54.9%) and 1-phenylpropargyl (p2; 41.6%) are predicted via the RRKM calculations (table S2). Considering these pathways, the predicted ratio of the hydrogen atom loss from the phenyl group versus the vinyl moiety is 1.22 ± 0.03:1. The isotopic experiments also determined the extent of the hydrogen loss in the C/C₆H₅C₂D₃ (from the phenyl group; fig. S4) and C/C₆D₅C₂H₃ systems (from the vinyl group; fig. S5) detected at m/z = 120 and 118, respectively, to be 1.18 ± 0.02:1 (Supplementary Materials). This result is in excellent agreement with the RRKM calculations, suggesting that, to eventually eliminate atomic hydrogen, the carbon atom preferentially attacks initially the vinyl moiety, but not the aromatic ring.

Considering the larger cone of acceptance of the aromatic ring compared to the vinyl group, these findings seem counterintuitive. Furthermore, atomic carbon was found to react with the benzene molecule (C₆H₆) (49) forming the cyclic 1,2-didehydrocycloheptatrienyl radical (C₇H₅^•) plus atomic hydrogen. This reaction has neither an entrance nor an exit barrier and is only weakly exoergic by 8 kJ mol⁻¹. In the present situation, the reaction of atomic carbon with the aromatic ring of styrene has two competing exit channels: a vinyl radical (C₂H₃^•) loss forming the cyclic 1,2-didehydrocycloheptatrienyl radical (C₇H₅^•) (fig. S6, p4) as the heavy cofragment (Reaction 3) and an atomic hydrogen loss leading to three distinct vinyl-1,2-didehydrocycloheptatrienyl radicals (Reaction 4) (fig. S6, p5 to p7). The reactions are only slightly exoergic, by 19 kJ mol⁻¹ (C₇H₅^• plus C₂H₃^•) and by 40, 6, and 15 kJ mol⁻¹ for the vinyl-1,2-didehydrocycloheptatrienyl (C₇H₄C₂H₃^•) plus atomic hydrogen channel depending on the position of the de facto carbon atom insertion via addition – ring opening into the six-membered ring of styrene with respect to the vinyl group. Therefore, we may conclude that, although Reactions 3 and 4 are feasible, they are much less favorable thermodynamically than the formation of 1-indenyl (p1) and 1-/3-phenylpropargyl (p2/p3). Reaction 4 cannot be responsible for the experimentally observed translational energy distribution.

$$\mathrm{C}({}^3\mathrm{P})+{\mathrm{C}}_6{\mathrm{H}}_5{\mathrm{C}}_2{\mathrm{H}}_3\to {\mathrm{C}}_7{{\mathrm{H}}_5}^{\bullet }+{{\mathrm{C}}_2{\mathrm{H}}_3}^{\bullet }$$ (3)

$$\mathrm{C}({}^3\mathrm{P})+{\mathrm{C}}_6{\mathrm{H}}_5{\mathrm{C}}_2{\mathrm{H}}_3\to {\mathrm{C}}_7{\mathrm{H}}_4{\mathrm{C}}_2{{\mathrm{H}}_3}^{\bullet }+{\mathrm{H}}^{\bullet }$$ (4)

Propargyl-phenyl system

Having elucidated the formation of 1-indenyl (C₉H₇^•) in the reaction of atomic carbon with styrene, we now turn our attention to the propargyl-phenyl system (Fig. 4). According to the previous calculations (48), the reaction is initiated by the barrierless addition of the propargyl radical's CH₂ or CH terminal moieties to the radical site of phenyl producing intermediates i9 (3-phenyl-1-propyne) and i10 (phenylallene), respectively. These isomers reside in deep potential energy wells and are stabilized by 381 and 400 kJ mol⁻¹ with respect to the separated reactants. i9 isomerizes via a hydrogen shift and ring closure forming the bicyclic intermediate i11. Two pathways are open from i10: (i) isomerization to i11 via hydrogen atom migration from the terminal carbon at the CH₂ moiety and closure to a five-membered ring and (ii) isomerization to i12 through atomic hydrogen shift from the phenyl ring to the side chain. Once formed, i11 can further isomerize to indene (i13) via hydrogen migration or fragments to the 1-indenyl radical (p8) in an overall exoergic reaction. Alternatively, i12 can undergo ring closure to indene (C₉H₈) with the latter ejecting an atomic hydrogen barrierlessly forming the 1-indenyl radical (C₉H₇^•). A close inspection of the PES reveals five reaction sequences to 1-indenyl (C₉H₇^•) with barriers located below the energy of the separated reactants. These pathways are, in principle, accessible even at ultralow temperatures of 10 K in the cold molecular cloud: i9 → i11 → i13 → p8 + H, i10 → i11 → i13 → p8 + H, i9 → i11 → p8 + H, i10 → i11 → p8 + H, and i10 → i12 → i13 → p8 + H. It should be stressed that these calculations support the experimental identification of the 1-indenyl radical (C₉H₇^•) and of three adducts—3-phenyl-1-propyne (i9), phenylallene (i10), and indene (i13)—with three-body collisions with the buffer gas (helium) required to stabilize these initial collision complexes. Recently, Couch et al. (50) and Selby et al. (41) studied the propargyl − phenyl reaction. Couch et al. (50) conducted the experiment in a continuous flow over the temperature range of 800 to 1600 K and pressures of 25 torr. The authors detected ions attributable to C₉H₇⁺ and C₉H₈⁺. However, the electron impact ionization could neither reveal the source of C₉H₇⁺, i.e., nascent reaction product or dissociative electron impact ionization of C₉H₈, nor elucidate the nature of the structural isomers of C₉H₈. Selby et al. (41) studied this reaction over 300 to 1000 K and 4 to 10 torr using time-resolved multiplexed photoionization mass spectrometry. The isomer-resolved branching ratios for the C₉H₈ products are reported from 300 to 1000 K. On the basis of their photoionization spectra, at 300 K and 4 torr, only the direct C₉H₈ adducts 3-phenyl-1-propargyl and phenylallene are observed with the branching ratios of 77 ± 5% and 23 ± 3%, respectively. As the temperature and pressure are increased to 1000 K and 10 torr, respectively, two additional C₉H₈ isomers, 1-phenyl-1-propyne and indene, along with the C₉H₇ + H channel are observed. The corresponding branching fractions of C₉H₈ isomers are 61 ± 6%, 18 ± 3%, 6 ± 1%, and 15 ± 5% for 3-phenyl-1-propargyl, phenylallene, 1-phenyl-1-propyne, and indene, respectively. It should be noted that the C₉H₇ isomer distributions are not determined experimentally although it supplies 16 ± 5% of total C₉H₈ signal at 1000 K. Here, our experiment is conducted at much higher temperature and pressure of 1373 ± 10 K and 200 ± 10 torr exploiting a chemical microreactor and tunable synchrotron VUV light coupled with a Re-TOF-MS. Compared with the results from (41) under lower-temperature conditions, additional insights into the high-temperature chemistry of C₃H₃ + C₆H₅ reaction are derived. First, only three C₉H₈ isomers are observed at 116, namely, 3-phenyl-1-propyne (34 ± 2%), phenylallene (24 ± 3%), and indene (36 ± 2%). Second, 1-indenyl (C₉H₇) is determined to be the dominant product at 115 with 1PPR and 3PPR possibly being the minor products.

Fig. 4. Potential energy surfaces of the C₆H₅ + C₃H₃ reaction.

Dominant reaction pathways for the phenyl (C₆H₅^•) with propargyl (C₃H₃^•) reaction extracted from (48). The values on top/bottom of the arrows indicate the energies of the transition states. All energies are given in kilojoules per mole with respect to the energy of the separated reactants.

Astrochemical modeling

The identification of the 1-indenyl radical (C₉H₇^•) via the barrierless elementary reactions (Reactions 1 and 2) lends to an opportunity of incorporating these findings into astrochemical models of cold molecular clouds. It is vital to underline that both reactions have no entrance barriers, all barriers of the isomerization processes are below the energy of the separate reactants, and the overall reactions are exoergic. These findings denote key prerequisites for reactions to proceed in low-temperature environments such as in molecular clouds (10 K). To investigate the implications of our findings to the chemistry of 1-indenyl in cold molecular clouds, we untangled the viability of the formation of 1-indenyl through the neutral-neutral reactions (Reactions 1 and 2) using the University of Manchester Institute for Science and Technology Database (RATE2012) (51) operated with the two-step single-point time-dependent astrochemical model described in (52) and (53). The physical parameters of the models were updated according to Markwick et al. (54) and McElroy et al. (51); reaction rate coefficients and (preexponential) factor α (table S3) are parameterized according to McElroy et al. (51). The physical parameters of the cold molecular cloud are triggered by a temperature of 10 K, a cosmic ray ionization rate of 1.3 × 10⁻¹⁷ s⁻¹, a visual extinction of 10 Mag, and a number density of molecular hydrogen of 10⁴ cm⁻³. The chemical network was also enhanced with recently explored barrierless and rapid reactions related to the formation of propargyl (C₃H₃) (22), benzene (C₆H₆) (52), and indene (C₉H₈) (27). The key bimolecular reactions and photodissociation processes newly incorporated into the astrochemical model leading to the formation of 1-indenyl are shown in Fig. 5, and the details are depicted in table S3. In the simulations, the fiducial gas-phase model was first operated until the chemistry evolved to the steady state; then, the obtained molecular abundances at this point are used as the initial conditions for the second-step model where the ice mantle species are injected into the gas phase until the model reaches the steady state again. Last, the predictive capabilities of the model are verified by comparing the relevant species observed with modeled fractional abundances.

Fig. 5. Reaction network forming 1-indenyl.

Compilation of key bimolecular reactions and photodissociation processes newly incorporated into the astrochemical model leading to the formation of 1-indenyl.

These modeling studies reveal captivating findings (Fig. 6 and fig. S7). First, the exceptional performance of the astrochemical model for TMC-1 can be well benchmarked for the propargyl (C₃H₃), cyanobenzene (C₆H₅CN), and indene (C₉H₈) recently detected at fractional abundances of 0.8 × 10⁻⁸ to 1.2 × 10⁻⁸ (17, 18), 2.4 × 10⁻¹¹ to 1.2 × 10⁻¹⁰ (55, 56), and 7.6 × 10⁻¹⁰ to 1.0 × 10⁻⁹ (57, 58) toward TMC-1. Prepared principally via the barrierless reactions of atomic carbon with ethylene (22) and of benzene with cyano radicals (59), the predicted peak fractional abundances of propargyl and cyanobenzene at 1.3 × 10⁵ and 7.9 × 10⁴ years of 2.9 × 10⁻⁹ to 1.1 × 10⁻⁸ and 4.1 × 10⁻¹¹ to 4.7 × 10⁻¹¹ with respect to molecular hydrogen replicate the astronomical observations exceptionally well (Fig. 6). A match between the modeled (1.0 × 10⁻¹⁰ to 9.0 × 10⁻¹⁰) and observed fractional abundances (7.6 × 10⁻¹⁰ to 1.0 × 10⁻⁹) has also been achieved for indene (C₉H₈) with dominating formation routes involving the reactions of methylidyne (CH) with styrene (C₈H₈) (27) and o-benzyne (C₆H₄) with allyl (C₃H₅) (60). Second, barrierless molecular mass growth processes via ring annulation were predicted to form 1-indenyl (C₉H₇^•) peaking at fractional abundances of 2.4 × 10⁻¹¹ to 9.7 × 10⁻¹¹ at 2.5 × 10⁵ years; Reactions 1 and 2 contribute nearly equally with 55 and 45%, respectively. These fractional abundances suggest that the 1-indenyl radical (C₉H₇^•) might be detectable by radio telescopes such as the Green Bank Observatory (61) and Yebes Radio Telescope (17).

Fig. 6. Results of the astrochemical model for TMC-1.

The fractional abundances of the gas-phase cyanobenzene (C₆H₅CN), indene (C₉H₈), propargyl (C₃H₃), and 1-indenyl (C₉H₇) are plotted as a function of time after the injection event. Astronomically observed fractional abundances, along with the uncertainties of cyanobenzene, indene, and propargyl, are visualized by the green-, blue-, and pink-colored horizontal bars.

DISCUSSION

Our combined experimental, computational, and astrochemical modeling study provided compelling evidence on the very first gas-phase preparation of the aromatic and resonantly stabilized 1-indenyl (C₉H₇^•) radical via elementary reactions of ground state atomic carbon (C) with styrene (C₆H₅C₂H₃) and of the propargyl radical (C₃H₃^•) with the phenyl radical (C₆H₅^•). These pathways provide unique access to 1-indenyl via barrierless entrance and exoergic reactions involving unconventional isomerization steps of the collision complexes via hydrogen shifts, ring closure/openings, and aromatization. These prerequisites represent fundamental criteria for both reactions to operate in cold molecular clouds such as TMC-1 and offer an unusual source of 1-indenyl (C₉H₇^•) as precursors to a low-temperature route of PAHs in deep space, which may assist in deciphering the “lifetime paradox” of PAHs in the ISM. These findings also suggest that, upon reaction with atomic carbon and propargyl radicals, any vinyl side-chained six-membered aromatic ring or any aryl radical can be converted to a cyclopentadienyl moiety under low-temperature conditions (fig. S2), thus providing a facile and barrierless route of annulation of a benzene ring embedded in complex PAHs. These reaction routes are of particular importance as a critical sink of propargyl radicals detected with fractional abundances of (1.0 ± 0.2) × 10⁻⁸ in TMC-1, which are unreactive with closed shell hydrocarbons under cold molecular cloud conditions revealing barriers to addition between 31 and 60 kJ mol⁻¹ for, e.g., alkynes, alkenes, and aromatics (62–64). In low-temperature atmospheres of planets and their moons such as Titan and Triton, the outcome of these reactions is influenced by the atmospheric pressures. Whereas higher altitudes with pressures less than 10⁻³ torr support bimolecular reactions leading to 1-indenyl (C₉H₇^•), third-body stabilization of the reaction intermediates in lower atmospheric layers leads to their stabilization and generation of distinct C₉H₈ isomers of 3-phenyl-1-propyne, phenylallene, and indene, which may act as fundamental molecular growth species in Titan’s organic aerosol layers (65). Overall, our investigations provide a very first step to systematically elucidate the fundamental reaction pathways to RSFRs in low temperature environments as a critical prerequisite to their molecular mass growth processes to PAHs in extreme environments. Although the C_2v symmetric 1-indenyl (C₉H₇^•) has not been detected yet in TMC-1, its resonance stabilization might lead to substantial fractional abundances, thus calling for future astronomical searches for 1-indenyl’s astronomical identification. Combined with our extended astrochemical model, this identification would provide critical constraints on the largely elusive low-temperature molecular mass growth processes to aromatics in TMC-1 and act as a catalyst initiating advanced RSFR-triggered chain-propagating reactions eventually culminating in the rapid formation of carbonaceous nanostructures in deep space.

MATERIALS AND METHODS

Experimental methods: Carbon-styrene system

The gas phase reaction of ground-state atomic carbon (C; ³P) with styrene (C₆H₅C₂H₃; X¹A′) was carried out under single-collision conditions using a crossed molecular beam apparatus (34). The primary atomic carbon reactant was generated in situ via laser ablation of graphite exploiting the 266-nm output of an Nd:YAG laser at 8 to 10 mJ pulse⁻¹ and seeding the ablated atoms in helium. A chopper wheel selected a part of the pulsed supersonic beam defined by the peak velocity v_p of 2534 ± 61 m s⁻¹ and a speed ratio S of 2.2 ± 0.3. Under these experimental conditions, the carbon atoms produced are only in the ground electronic state (³P) (66). The primary beam crossed a secondary beam of xenon-seeded (99.999%; Matheson) styrene (C₈H₈; >99%; TCI) defined by v_p = 389 ± 6 m s⁻¹ and S = 28.0 ± 0.4 with a seeding fraction of 1% at a backing pressure of 550 torr. This resulted in a CM angle of 52.8 ± 0.6° and a collision energy of 35.4 ± 1.4 kJ mol⁻¹. The D3- and D5-styrene (98% D; CDN Isotopes) were used in the experiment to extract information on the position of the atomic hydrogen loss (table S1). After electron-impact ionization of the neutral products (80 eV, 2 mA), the ions were monitored using a quadrupole mass spectrometer operating in the TOF mode under ultrahigh-vacuum conditions (6 × 10⁻¹² torr). After the normalization to CM angle and integration of the TOFs, the LAD distribution at well-defined m/z can be derived. A forward-convolution method based on Jacobian transformation is used to convert the data collected in the laboratory frame (LAD; TOFs) into the CM frame (28, 34).

Experimental methods: Propargyl-phenyl system

The experiments were conducted at the Chemical Dynamics Beamline (9.0.2.) of the Advanced Light Source using a high-temperature chemical microreactor consisting of a resistively heated SiC tube of 20-mm heating length and 1-mm inner diameter. This device is located inside the source chamber of a molecular beam setup equipped with a Re-TOF-MS (21). The molecular beam apparatus is designed to study elementary chemical reactions ultimately leading to PAH growth in situ. In detail, phenyl (C₆H₅^•) and propargyl (C₃H₃^•) radials were prepared in situ via pyrolysis of nitrosobenzene (C₆H₅NO; ≥ 97%; Sigma-Aldrich) (35) and propargyl bromide (C₃H₃Br; > 98%; Sigma Aldrich) (20) precursors, respectively. The reactants were seeded in helium carrier gas at total pressures of 200 ± 10 torr at the reactor inlet. The propargyl bromide precursor was kept in a stainless-steel bubbler at a temperature of 199 K obtained in a dry ice–ethanol bath; the nitrosobenzene precursor was stored in a Swagelok particulate filter (Swagelok SS-6F-15) and cooled to 273 K via an ice bath. Both vessels were arranged in series: The helium carrier gas first passed through the liquid propargyl bromide sample; the resulting helium-seeded propargyl bromide gas mixture was then passed over the solid nitrosobenzene sample before entering the reactor. The temperature of the SiC tube was determined using a type C thermocouple to be 1373 ± 10 K. The products formed in the reactor are entrained in a molecular beam, which passed a 2-mm skimmer located 10 mm downstream to the reactor; the beam then entered the main chamber, which houses the Re-TOF-MS. The neutral products within the supersonic molecular beam were photoionized in the extraction region of the mass spectrometer using quasi-continuous tunable synchrotron VUV light. VUV single-photon ionization represents essentially a fragment-free ionization technique and is a soft ionization method compared to the harsher conditions of electron impact ionization (67). The ions formed via soft photoionization were extracted and detected by a microchannel plate detector. Under our experimental condition, the residence time in the reactor tube is a few tens of microseconds (68, 69). PIE curves, which report ion counts as a function of photon energy with a step interval of 0.05 eV at a well-defined m/z, were produced by integrating the signal recorded at the specific m/z for the species of interest. Control experiments were also proceeded by expanding neat helium carrier gas with each precursor separately into the resistively heated silicon carbide tube, but neither indene nor indenyl was detected. Last, reference PIE curves of helium-seeded 1-bromoindene (C₉H₇Br) (46) and 3-phenyl-1-propyne (C₉H₈; 97%; Sigma Aldrich) were recorded in the present work within the same experimental setup, respectively, whereas the PIE curve of phenylallene was provided by Kukkadapu et al. (40).

Computational methods: Electronic structure calculations

Electronic structure calculations in the present work were carried out on the triplet C₉H₈ potential energy surface accessed by the carbon-styrene reaction. In particular, geometries of the reactants, products, intermediates, and transition states involved in this reaction were optimized at the density functional theory B3LYP/6-311G(d,p) level (table S4) (70, 71), with vibrational frequencies computed using the same theoretical approach. Single-point energies were further refined within the explicitly correlated CCSD(T)-F12/cc-pVTZ-F12 method (43, 72), which normally provides the energetic parameters of local minima and transition states within “chemical accuracy” of ~5 kJ mol⁻¹ for hydrocarbons in terms of average absolute deviations (42). The Gaussian 16 (73) and MOLPRO 2021 (74) program packages were used for the ab initio calculations. RRKM theory together with the steady-state approximation (75, 76) implemented in our Unimol code (77) was applied to compute energy-dependent rate constants of all unimolecular reaction steps on the C₉H₈ triplet surface after the addition of the atomic carbon to styrene and to assess product branching ratios at zero-pressure limit under single-collision conditions.

Astrochemical modeling

To explore the chemistry leading to the 1-indenyl radical in the cold molecular cloud TMC-1, we constructed the chemical model by expanding the RATE12 network (51) with barrierless neutral-neutral reactions forming the 1-indenyl radical and their precursors. First, a time-dependent gas-phase model was operated until the chemistry evolves to the steady state, i.e., after typically 10⁷ years. Simulations were also conducted in a separate model with the ice mantle species injected into the gas phase through reactive desorption until it reaches the steady state; this strategy was exploited successfully to demonstrate the key role of neutral-neutral reactions in the formation of benzene in TMC-1 (52). These results were then benchmarked with astronomical observations for cyanobenzene (C₆H₅CN), propargyl (C₃H₃), and indene (C₉H₈) to verify the predictive capabilities of the chemical network (Supplementary Materials).

Supplementary Materials

This PDF file includes:

Supplementary results

Figs. S1 to S7

Tables S1 to S4