Photofragmentation of corannulene (C₂₀H₁₀) and sumanene (C₂₁H₁₂) cations in the gas phase and their astrophysical implications

Abstract

Aromatic infrared bands (AIBs) dominate the mid-infrared spectra of many galactic and extragalactic sources. These AIBs are generally attributed to fluorescence emission from aromatic molecules. Unified efforts from experimentalists and theoreticians to assign these AIB features have recently received additional impetus with the launch of the James Webb Space Telescope (JWST) as the Mid-InfraRed Instrument (MIRI) delivers a mid-IR spectrum with greatly increased sensitivity and spectral resolution. PAHs in space can exist in either neutral or ionic form, absorb UV photons and undergo fragmentation, becoming a rich source of small hydrocarbons. This top-down mechanism of larger PAHs fragmenting into smaller species is of utmost importance in photo-dissociation regions (PDR) in space. In this work, we experimentally and theoretically investigate the photo-fragmentation pathways of two astronomically significant PAH cations – corannulene (C₂₀H₁₀) and sumanene (C₂₁H₁₂), which are structural motifs of fullerene C₆₀, to understand their sequential fragmentation pathways. The photo-fragmentation experiments exhibit channels that are significantly different from planar PAHs. The breakdown of the carbon skeleton is found to follow different pathways for C₂₀H₁₀ and C₂₁H₁₂ because of the number and positioning of pentagon rings, yet the most abundant low mass cations produced by these two species are found to be similar. The low mass cations showcased in this work could be of interest due to their possible astronomical detections. For completeness, the qualitative photofragmentation behaviour of dicationic corannulene and sumanene has also been investigated, but the potential energy surface of these dications is beyond the scope of this paper.

---

1 Introduction

AIBs are discrete emission features observed in the interstellar medium (ISM), circumstellar regions, and galactic and extragalactic sources.¹ The AIBs are very characteristic of the C–C and C–H vibrational modes of polycyclic aromatic hydrocarbons (PAHs) with prominent features centered in the 3.3, 6.2, 7.7, 8.6, 11.3, and 12.7 μm regions. These bands are attributed to the IR fluorescence of UV pumped PAH species that after internal conversion and intramolecular vibrational energy distribution relax through vibrational emission. PAHs can appear in the ISM either in neutral, cationic, protonated, hydrogenated, dehydrogenated or substituted (with one or more carbon atoms replaced with atoms like nitrogen, oxygen, etc.) form.² The wavelength and relative intensity of these astronomical AIBs are found to vary with the physical environment.³ In spite of numerous theoretical and experimental investigations for several decades to deduce the emitters of AIBs, the exact form of PAHs or the group of PAHs responsible for the emission continues to puzzle astronomers. Recent identifications of small PAHs in dark clouds through their rotational spectra have confirmed the presence of this important class of molecules in space.⁴

PAHs play an important role in the ionization and energy balance of interstellar gas.⁵ In addition, observations indicate that larger PAHs (with 20 to 100 carbon atoms) harbour 10–15% of the elemental carbon in the ISM. Besides PAHs, buckminsterfullerene, C₆₀, has also been identified in the ISM through its infrared bands using the Spitzer Space Telescope,⁶ and this was confirmed through its electronic transitions in the far-red.^7,8 It is found that, close to stars, the abundance of C₆₀ increases rapidly while the abundance of PAHs decreases away from stars.^9,10 This has been attributed to photochemical fragmentation and isomerization processes under the influence of the strong stellar ultraviolet (UV) radiation field but the details of these processes are not yet understood. A detailed study of the bright reflection nebula, NGC 7023, has provided evidence that the evolution of the profile of the mid-IR bands is related to the chemical evolution under the effect of UV photons.³

PAHs are assumed to be formed in the envelopes of evolved stars and then to be injected in the ISM.^11,12 In the past, astronomical models generally postulated that the abundance of specific PAHs in the ISM was mainly controlled by their thermodynamic properties and followed the pattern of stabilomers originally identified by Stein and Fahr (1985).¹³ However, under the severe conditions of space where there are strong radiation fields, photochemical pathways may control the abundances of interstellar PAHs. In this scenario, it is considered that stars inject a rich variety of PAHs – formed through processes akin to those in sooting flames – into space where they are quickly weeded down into a small set of extremely stable species by the strong radiation field. PAHs in space typically absorb 10⁸ UV photons over their lifetime (100 Myr) and undergo destruction through a photo-fragmentation process.⁹ The composition of the interstellar PAH family is greatly influenced by the fragmentation process initiated by UV photon absorption as highly excited PAHs can lose H atoms, CH, and C₂/C₂H₂ groups rather than relax through IR emission. This competition between fragmentation and radiative relaxation is largely controlled by the size of PAH.^10,14–17

Recent laboratory experiments demonstrate that upon photo-excitation the fragmentation of PAHs offers a new chemical paradigm of complex, interstellar molecule formation – instead of bottom–up, where reacting smaller species form larger ones, a top-down mechanism, where fragmentation products of larger precursors occur.^17,18 The observed anti-correlation between PAHs and C₆₀ abundances in astronomical environments strongly suggests that one species forms from the other and this photochemical relationship has the potential to give us direct clues on the processes that are dominant in the evolution of interstellar PAHs. The fullerenes contain fused hexagon and pentagon rings, and pentagon formation may provide the key.¹⁹ In an intense UV radiation environment, this photochemical breakdown may also contribute to the diversity of small hydrocarbon radicals commonly observed in these regions.²⁰ Understanding the dissociation of PAHs through photochemical excitation requires dedicated laboratory studies to identify the detailed photo-fragmentation of PAHs in space.

Neutral and cationic C₆₀ – identified through their unique IR and visible spectrum^6,7 – are the largest molecular species identified in space. The presence of pentagons in a planar graphene-flake leads to bending of the molecular structure enabling closure upon itself and the formation of a 3D structure.^21,22 Pentagon formation is thus a key step in the photochemical transformation of PAHs into fullerenes. Spectroscopic evidence for pentagon formation upon PAH photo-dissociation has been reported in the laboratory for small PAH species, containing up to three aromatic rings.^23–25 These experiments suggest that PAHs with only hexagonal rings could be the parent molecules (Parent-PAHs) for the formation of PAHs with pentagonal ring(s) under adequate conditions. In addition, some PAHs with pentagons may be more photochemically stable than PAHs of similar size²⁶ and therefore represent likely candidates for the interstellar molecular inventory. However, the photochemical evolution of PAHs with pentagons is largely unexplored and the generality of this conclusion is unknown. Given the presence of C₆₀ in space, the likely astrochemical connection of C₆₀ and PAHs in space, and the potential photochemical stability of a pentagon containing PAHs, a focused study to improve the understanding of such species is portrayed in this work.

The PAHs with pentagons chosen for this work are: corannulene (C₂₀H₁₀) and sumanene (C₂₁H₁₂) as shown in Fig. 1. The coordinates of the optimized geometry have been provided in Fig. S1 and S2 in the ESI.† These PAHs have been selected because they represent “pentagon-variations” on the structure of the well-studied compact PAH, coronene. Moreover, corannulene and sumanene are molecular building blocks/fragments of C₆₀ as the pentagon is connected to five hexagons in corannulene and the pentagons in sumanene are separated from each other through (partial) rings of hexagons.²⁷ When corannulene has ten H atoms which are all sp² hybridized, sumanene has six H atoms that are sp² and six more H atoms that are sp³ hybridized (Fig. 1 bottom panel). This makes the chemistry of these two molecules very different from each other. The photochemistry of the Parent-PAH, coronene as well as the fullerenes C₆₀ and C₇₀, has been studied in detail in the gas phase, however, the buckybowls (C₂₀H₁₀ and C₂₁H₁₂) are the least explored.^28,29

Fig. 1 Geometry of cationic (a) corannulene, C₂₀H₁₀⁺ and (b) sumanene, C₂₁H₁₂⁺ optimized using the M062x/6-311++G(3df,2pd) method.

Several spectroscopic studies have been performed on corannulene to look for its presence in the interstellar medium.^30–32 In particular, the diffuse interstellar bands at 6614 and 6196 Å have been proposed to arise from the electronic states of corannulene.³³ Moreover, the matrix-isolated IR spectra of hub-protonated corannulene (hub-H+C₂₁H₁₀⁺) showed close resemblance with the astronomical UIR bands,³⁴ but gas-phase spectra are required for positive astronomical detection. Because of its non-planar structure,³⁵ corannulene possesses a large dipole moment (2.0 Debye) and its microwave spectrum has been recorded.³⁶ A detailed model suggests that this species could be present in the Red Rectangle nebula, yet the hunt for this molecule still continues.³⁷ The spectroscopy of sumanene is not so well constrained. The molecule has a slightly larger bowl depth (1.11 Å) than corannulene (0.87 Å) and it is closer to the curvature of the C₆₀ structure.^27,38 The presence of three sp³-hybridized benzylic sites is a unique feature in sumanene, which makes its chemistry very different from that of corannulene.³⁹ Sumanene is expected to be an important precursor for the formation of C₆₀via a bottom-up mechanism.

To date, a number of laboratory experiments have been dedicated to the study of the photochemistry of smaller PAHs (C < 20).^20,40–42 Experimental studies of PAHs relevant to astrochemistry have widely used a time-of-flight, photo-ionization mass spectrometry (PIMS) method, focusing on reaching internal energy at which the dissociation rate is ∼10⁴ s⁻¹.⁴³

In these experiments, loss channel(s) mainly involve sequential hydrogen loss followed by the opening up of the C₂H₂ loss channel at slightly higher internal energies. However, different molecular dynamical timescales are showcased in the ISM. A competition takes place between fragmentation of PAHs, isomerization, and relaxation through IR emission that occurs on time scales of the order of ∼0.01–1 s⁻¹, depending on the size of the species. This work studies in detail the photofragmentation pattern of the buckybowls, the influence of the presence of pentagon rings on the photo-dissociation pattern, and the isomerization process involved in the formation of smaller fragment cations with the aid of the mass spectroscopic data obtained in the ‘instrument for photo-dynamics of PAHs’ (i-PoP) system.

2 Experiment

The experiments presented in this paper were performed on the ‘instrument for photo-dynamics of PAHs’ (i-PoP), housed at the Laboratory for Astrophysics (LfA) at Leiden observatory. The i-PoP has been described in detail in a previous paper so only a brief description is given here.¹⁸ The setup consists of two differentially pumped chambers; a source chamber that has a commercially available quadrupole ion trap (QIT, Jordan C-1251), and a detection chamber which encompasses a reflectron time-of-flight spectrometer (Jordan D-850). The molecules used in this work are commercially available from Tokyo Chemical Industry; corannulene (C₂₀H₁₀, >97.0% pure), and sumanene (C₂₁H₁₂, 99.0% pure). The samples were evaporated in the QIT chamber using a Heat Wave Labs built oven held at around 40 °C and 70 °C for corannulene and sumanene respectively. The sample vapours were ionized by electron impact ionization at 70 eV with an electron gun (EGUN, Jordan C-950), which is integrated into the QIT chamber and the resulting cations after ionization were then guided into the ion trap through an electrostatic ion gate.

A 1600 V top-top RF signal was used for these experiments on the ring electrode at an operating frequency of 1.25 MHz in the ion trap to hold ions. This enables the trap to retain masses from approx. 98 amu up to several hundred amu. In an attempt to probe a much lower mass range to look for smaller cationic fragments an RF of 1200 V was also used in some experiments which pushed the lower observable limit to 73 amu. In addition, the experiments with dications uses an RF of 1000 V for observations until reaching 61 amu. Helium buffer gas was released into the center of the ion trap up to a static pressure of 1–2 × 10⁻⁶ mbar in the QIT chamber. The PAH cations were confined to the center of the ion trap through collisions with the He buffer gas and persisted there until the cations were directed from the ion trap into the time-of-flight (TOF) detection chamber.^44,45 The ions in the ion trap were irradiated with a nanosecond pulsed Quanta-Ray Nd:YAG laser (DCR2A-3235) pumping a dye laser (LIOP-TEC, Quasar2-VN) which was set to deliver a particular wavelength from 610–630 nm photons, chosen depending on the molecule to be studied. The laser was horizontally guided through the ion trap and was operated at 10 Hz to irradiate the trapped ions.

The reasoning for the choice of 610–630 nm (red light) laser radiation was to scan over the potential energy surface of corannulene and sumanene cations which was the main aim of this experiment. The choice of red light from dye laser radiation allows minimization of the multiple ionization in competition with fragmentation that are known to occur with VUV photons.^46,47 For instance, using a green light for photolysis would give rise to the same set of product cations. However, there are higher chances of the cations undergoing multiple ionization, making the analysis complicated. Moreover, the breakdown analysis to deduce the steps of fragmentation would not be possible as several fragment cations are produced in higher intensities at the very beginning of the laser pulses. Hence, using longer wavelength photons in a multi-photon process allowed us to obtain more information about the fragmentation pattern of the molecules in a gentler manner to explicate additional nuances in the fragmentation patterns.

The data acquisition cycle is achieved through timing sequences controlled by a high-precision delay generator (SRS DG535), which is triggered by the Q-switch timing of the Nd:YAG laser to ensure synchronization at the start of each measurement cycle. Each operation cycle involves capturing ions and filling of the ion trap, mass isolation of the parent ion, irradiation of the ion cloud with a laser, and extraction of the ions into the TOF tube. The scan cycle begins with an empty ion trap and was initiated with the opening of the ion gate which is achieved using a normal gating procedure of applying a DC voltage to a metal lens with a circular slit in the middle as an inlet for the ions, and the ion trap fills for a duration of 2.7 s. To isolate the masses of the parent PAH ions, a 25 ms (from 2.7 s to 2.95 s) long Stored Waveform Inverse Fourier Transform (SWIFT) pulse was applied to one of the end caps of the ion trap to isolate the parent species.⁴⁸ Shortly afterward a re-thermalization period was given for 0.05 s (from 2.95 s to 3.00 s). It should be noted that the intense electron impact ionization source not only produced parent PAH cations but fragments due to H/H₂ loss as well. For the purposes of this work, the SWIFT pulse was applied to isolate the parent cations as well as the cations corresponding to H/H₂ losses from the parent (i.e., the 247–252 amu range) and efficiently filter out other spurious signals or contaminations. After the SWIFT pulse was employed, the laser beam shutter was opened, and the ion cloud was irradiated. Typically, 40 pulses and 30 pulses from the dye laser were used for corannulene and sumanene cations, respectively. At the end of each irradiation time with the specified number of pulses, the ions were accelerated out of the ion trap and into the field-free TOF region at the end of which the ions were detected using a multichannel plate (MCP) detector and digitized using a 8-bit GaGe Cobra card. The data acquisition was carried out using LABVIEW software to obtain and calibrate the mass spectrum.

For each mass spectrum, an average of 50 scans was used and before each measurement, a normalization spectrum of 15 scans was taken with the laser under the off condition. These normalization spectra were used to normalize the peak areas of the mass spectra measured with the laser on, which resulted in errors of 5%. The number of laser pulses showcased in the experiments was 40 pulses for corannulene cations and 30 pulses for sumanene cations, in steps of 2 pulses (e.g. 0, 2, 4, 6, … 30). These numbers of pulses were chosen such that most of the parent cations would be depleted in the ion trap and converted to fragments. In each case, the total cycle duration was kept constant at 6 s and 7 s for corannulene and sumanene cations respectively to ensure that all datasets are cross comparable and the only differing parameter is pulse energy and/or number of pulses.

In short, ion gate opening → trapping ions in the ion trap → mass isolation using SWIFT → photolysis by a specified number of laser pulses in a sequence (0 or 2 or 4…) → extraction of fragment ions → recording the TOF mass spectrum, is a full cycle that repeats for each measurement. The recorded mass spectrum was calibrated from the time domain to m/z using a MATLAB code. The same software was also used to deduce the peak area of each cationic peak using a Pearson type IV fit as this gave a more reliable result for peak area compared to a Gaussian fit. The obtained peak areas were normalized using the normalized scans of mass spectrum as described above, and these normalized values of peak area were used to make the breakdown diagrams presented in this paper.

3 Theory

Density functional theory (DFT) was used to explore the potential energy surfaces (PES) of corannulene and sumanene cations. The Minnesota functional M06-2x which includes dispersion in combination with a triple-zeta quality basis set – 6-311++G(3df,2pd) – was chosen, including polarisation and diffuse functions which are needed to describe accurately reaction energies and barrier heights with density functional theory.⁴⁹ For example, the energy difference between vinylidene (CCH₂) and acetylene (C₂H₂) is predicted to be 1.85 eV with this method while is calculated to be 1.86 eV at the CCSD(T)/cc-pVTZ level.⁵⁰ All the calculations were performed with Gaussian C.02⁵¹ and the results were visualized with the software Molden.⁵² Intrinsic Reaction Coordinate (IRC) calculations were performed to check that the transition states connected the intended reactants and products. The expectation value of the total spin of the system 〈S²〉 was used as a proxy in the presence of higher spin states. In the large majority of the cases the structure with the lowest spin multiplicity (e.g. doublet for the isomers of C₂₀H₁₀⁺) is also the more stable structure and that is the one we included in the PES. Special cases are indicated in the discussion. Throughout the paper, energies are zero-point corrected and expressed in eV. Due to the large number of possible photodissociation channels, the exploration of the PES with static DFT calculations is not exhaustive, yet it can shed light on some of the fragmentation behaviours seen in the experiments.

4 Results and discussion

4.1 Corannulene cation (C₂₀H₁₀⁺)

Fig. 2 shows the mass spectrum of C₂₀H₁₀⁺ obtained with the i-PoP system after applying the SWIFT pulse. The m/z = 250 corresponds to the parent cation C₂₀H₁₀⁺. The less intense peaks at m/z = 251 and 252 are due to the ¹³C components present in the sample. It is also observed that the electron impact on gaseous C₂₀H₁₀⁺ to ionize the neutral parent species also induces H abstraction to produce C₂₀H₉⁺ and C₂₀H₈⁺.

Fig. 2 Mass spectrum of C₂₀H₁₀⁺ obtained upon electron impact using the Re-TOF instrument integrated in the i-PoP system.

An electron energy of 70 eV and emission of 0.6 mA was used to produce a suitable amount of corannulene cations, C₂₀H₁₀⁺. Fig. 3a presents the TOF-MS of C₂₀H₁₀⁺ for the pulse energies of 2.23 mJ per pulse. The usage of red light with a pulse energy of 2.23 mJ per pulse allows to decode a number of interesting fragmentation pathways. The fragments produced upon laser irradiation shown in these figures are cationic, and their neutral counterpart fragments would not be detectable through ion trap TOF mass spectrometry. In these experiments, 50 scans were taken for each mass spectrum. Such mass spectra were obtained after 2 pulses, 4 pulses, etc. By accumulation of the fragments upon sequential laser induced fragmentation, the mass spectra are stacked to visualize the increase in fragmentation yield, as presented in Fig. 3a. The RF value used for the experiment in Fig. 3a is 1600 V which is efficient for detecting mass peaks from 98–260 m/z, which can display the CH/C₂H₂ loss channels in good detail.

Fig. 3 (a) TOF mass spectra for corannulene radical cations irradiated with 0, 2, 4, … and 40 laser pulses (the mass spectra are presented in steps of 4 pulses for clarity) with a pulse energy of 2.23 mJ per pulse with RF = 1600 V i.e. spectra displaying the high-mass cations with prominent features in the m/z = 260–140 range; (b) breakdown diagrams for the fragmentation of C₂₀H₁₀⁺. The normalized mass peak areas as a function of the number of pulses are plotted, corresponding to the most prevalent hydrocarbon compounds formed with the irradiation with a 610 nm laser, with 2.23 mJ per pulse and RF of 1600 V.

4.1.1 Photodissociation of C₂₀H₁₀⁺ – the high mass range. As known for any PAH, dehydrogenation is the first step in the photo-fragmentation process. C₂₀H₁₀⁺ first loses H and H₂/2H, and it can be observed that the peak corresponding to C₂₀H₈⁺ is more intense than C₂₀H₉⁺. This means that the second H atom is easily abstracted compared to the first H atom. Likewise, the fourth H atom loss is also observed more prominently than the third H atom loss. The even-numbered favouring for H loss channels was also noticed in other PAH cations like coronene (C₂₄H₁₂⁺), hexabenzocoronene (C₄₂H₁₈⁺) and dibenzopyrene (C₂₄H₁₄⁺)⁵³ and seems to be independent of the molecular structure or planarity.

However, unlike other medium-sized PAHs, C₂₀H₁₀⁺ loses only four H atoms before the carbon skeleton starts opening up through CH loss. In fact, the CH loss channel is one of the most dominant channels for the fragmentation of C₂₀H₁₀⁺. The CH loss fragments C₁₉H₉⁺, C₁₈H₈⁺ and C₁₇H₇⁺ are clearly observed in the mass spectrum:

C₂₀H₁₀⁺(–CH) → C₁₉H₉⁺(–CH) → C₁₈H₈⁺(–CH) → C₁₇H₇⁺.

We note that there could be a competition between the CH loss and C₂H₂ loss channel as the first C₂H₂ loss channel (220–228 m/z) is also getting more intense at pulse 12 and then decreases. This means that the C₂H₂ loss channel is possible from C₂₀H₁₀⁺ with either single or a two-step fragmentation:

C₂₀H₁₀⁺ (–CH) → C₁₉H₉⁺ (–CH) → C₁₈H₈⁺

C₂₀H₁₀⁺ (–C₂H₂) → C₁₈H₈⁺

After three sequential CH losses, i.e. after the parent, C₂₀H₁₀⁺, becomes C₁₇H₇⁺, it is not possible to lose any more CH. Fig. 3b demonstrates that after this point the C₂H₂ loss channel becomes dominant. The consecutive C₂H₂ loss can be described as:

C₁₇H₇⁺(–C₂H₂) → C₁₅H₅⁺(–C₂H₂) → C₁₃H₃⁺.

Alternately, it is also possible for C₁₇H₇⁺ to undergo subsequent loss of two C₂ from the carbon skeleton:

C₁₇H₇⁺(–C₂) → C₁₅H₇⁺(–C₂) → C₁₃H₇⁺.

It is noticeable from the mass spectrum (Fig. 3a) that the cations C₁₇H₇⁺, C₁₅H₇⁺ and C₁₃H₇⁺ are sequentially produced after the pulses 4, 8 and 12 respectively, which demonstrates that the C₂ loss is probably sequential.

Fig. 3b depicts the breakdown diagrams of the parent and product cations: the top panel shows the increase of C₂₀H₈⁺ upon laser irradiation up to pulse 10 and then a decrease in peak area, which could mean that this species fragments further into smaller secondary cations. Similarly in the middle panel C₁₈H₈⁺ and C₁₈H₆⁺ increases up to pulse 10 and then decreases. At this juncture (from pulse 10), the smaller cations C₈H₅⁺, C₉H₅⁺ and C₁₁⁺ start to form and increase steadily. The breakdown equations predicted in this section for the origin of low mass cations (below 150 m/z) are purely based on the data obtained in the mass spectrum and the breakdown diagram. The breakdown of this corannulene cation is expected to involve several intermediate steps including H roaming, C–C bond cleavage followed by opening of the C skeleton, etc. before the formation of the observed low mass cations. The breakdown channels given below are those predicted based on the end products and do not include any intermediate or transition states. Therefore, caution must be taken in perceiving the breakdown predictions presented in this paper.

Possible breakdown mechanisms for C₁₈H₈⁺ could be:

C₂₀H₁₀⁺(–C₂H₂) → C₁₈H₈⁺ → C₈H₅⁺ + C₁₀H₃,

C₂₀H₁₀⁺(–C₂H₂) → C₁₈H₈⁺ → C₉H₅⁺ + C₉H₃,

C₂₀H₁₀⁺(–C₂H₂) → C₁₈H₈⁺ → C₁₁⁺ + C₇H₈.

The C₁₈H₆⁺ product can be produced after two hydrogen losses followed by two consecutive CH losses, e.g.

C₂₀H₁₀⁺(−H₂/2H) → C₂₀H₈⁺ (–CH)

→ C₁₉H₇⁺(–CH) → C₁₈H₆⁺.

Or with two hydrogen losses followed by a C₂H₂ loss to eventually produce C₈H₅⁺, C₉H₅⁺ and C₁₁⁺ with the mechanism:

C₂₀H₁₀⁺(–H₂) → C₂₀H₈⁺(–C₂H₂) → C₁₈H₆⁺

→ C₈H₅⁺ + C₁₀H;

(or)

→ C₉H₅⁺ + C₉H;

(or)

→ C₁₁⁺ + C₇H₆.

It is evident from Fig. 3b that the parent C₂₀H₁₀⁺ and its ¹³C component starts decreasing immediately after the laser pulses are applied; and there are few fragment cations produced right from the beginning of laser irradiation. Nevertheless, there are several fragment cations that are visible only after 6 to 10 laser pulses. This could either be because the cations visible after 6 laser pulses are initially less abundant, or, they are secondary cations (i.e. they are fragmented from the cations after H/CH/C₂H₂ loss of C₂₀H₁₀). Hence, these large cations, after CH/C₂H₂ losses, further fragment into smaller species upon secondary fragmentation to produce the observed low mass cations (C ≤ 11).

C₁₁⁺ could also be produced directly from the parent cation upon two or more dehydrogenations of the parent, followed by ring opening and isomerization:

C₂₀H₁₀⁺ → C₂₀H_10−n⁺ → C₁₁⁺ + C₉H_10−n.

This is because the radical sites are necessary for the ring opening and isomerization process to take place. Similar to C₁₁⁺, another carbon cation C₁₀⁺ is also observed as a secondary fragment only after pulse 16 (Fig. 3a):

C₂₀H₁₀⁺(–H₂/2H) → C₂₀H₈(–C₂H₂) → C₁₈H₆⁺

→ C₁₀⁺ + C₈H₆

(or)

C₁₈H₈⁺ → C₁₈H_8−n⁺ → C₁₀⁺ + C₈H_8−n.

The other intense low mass fragments like C₈H₃⁺, C₈H₅⁺, C₉H₃⁺ and C₁₀H₅⁺ are observed from the beginning of laser irradiation and are observed to be primary cationic fragments from the parent cation immediately after H loss:

C₂₀H₁₀⁺ → C₂₀H_10−n⁺ → C₉H₃⁺ + C₁₁H_7−n

C₂₀H₁₀⁺ → C₂₀H_10−n⁺ → C₁₀H₅⁺ + C₁₀H_5−n

4.1.2 Photodissociation of C₂₀H₁₀⁺ – the low mass range. Fig. 4a shows the mass spectrum upon irradiation with sequential laser pulses using RF = 1200 V, illustrating a striking difference in the nuances of the low mass region (80–130 m/z) compared to Fig. 3a (i.e. experiment with RF = 1600) because of the reduced RF value to 1200 V in this case. However, the CH and C₂H₂ losses that are prominently observed in the experiment with RF = 1600 V, are also observed in Fig. 4a.⁵⁴ In this case, smaller photofragments like C₇H₃⁺, C₈H₅⁺, C₉H₃⁺, C₁₀⁺ and C₁₁⁺ are observed at higher intensity. We specifically note that in Fig. 3b, C₈H₃⁺ dominates over C₈H₅⁺, while in Fig. 4b, C₈H₅⁺ is dominant over C₈H₃⁺. We consider it is possible that, besides C₈H₅⁺, the 101 m/z peak has a contribution by the C₇H⁺–H₂O reaction that gives rise to C₇HO⁺ (or C₇H⁺–H₂O complex) due to trace contamination of H₂O in the ion chamber. More experiments are underway at our laboratory to investigate this further. For that reason, we refrain from further analysis of this low mass range.

Fig. 4 (a) TOF mass spectra for corannulene radical cations irradiated with 0, 2, 4, … and 40 laser pulses (mass spectra is presented in steps of 10 pulses for clarity) with a pulse energy of 2.65 mJ per pulse with RF = 1200 V i.e. spectra displaying the low-mass cations with prominent features in the m/z = 140–80 range; (b) breakdown diagrams for the fragmentation of C₂₀H₁₀⁺. The normalized mass peak areas as a function of the number of pulses are plotted, corresponding to the most prevalent hydrocarbon compounds formed in with the irradiation with a 610 nm laser, with 2.65 mJ per pulse and RF of 1200 V.

It is evident from the breakdown diagrams in Fig. 4b (right panel) that C₇H₃⁺ and C₈H₅⁺ are formed from the very beginning of laser irradiation, e.g.

C₂₀H₁₀⁺ → C₇H₃⁺ + C₁₃H₇

C₂₀H₁₀⁺ → C₈H₅⁺ + C₁₂H₅

Again, C₁₀⁺ and C₁₁⁺ appear to be secondary products arising from pulse 10 and pulse 6 respectively (Fig. 4b). The other cationic products are in relatively lower abundance and are also produced right after the irradiation with the first two laser pulses. This could mean that an energy of 5.3 mJ is already sufficient to break the carbon skeleton of C₂₀H₁₀⁺ into two or more fragments.

4.1.3 Theoretical results for the fragmentation of corannulene cation. In this section the results from the DFT calculations on the corannulene cation and its relevance with the experimental results are discussed. The first focus is on the H/2H loss and CH/C₂H₂ loss, which will then proceed to proposing some pathways for the further loss of CH and C₂H₂, observed in the experiments. H/2H/H₂, CH and CCH₂ loss channels. Fig. 5 shows the potential energy surface for the corannulene cation which includes proposed mechanisms leading to loss of H/2H/H₂, CH and C₂H₂.

Fig. 5 Potential energy surface (PES) for the C₂₀H₁₀⁺, obtained at the M06-2X/6-311++G(3df,2pd) level. Energies are given relative to the energy of the corannulene cation. Channels are colour-coded to help visualization. All the structures in the PES are at their lowest multiplicity, except for C₂₀H₉⁺ (triplet), int3 and int6 (quadruplet).

The lowest dissociation product is the dehydrogenated corannulene via H loss, that can be abstracted directly from the corannulene cation with a bond dissociation energy (BDE) of 4.96 eV or after isomerization reactions involving H roaming.§ The H roaming reactions require up to 4 eV, and each of the intermediates created (int1, Int2a and int2b, Fig. 5) can lose an H atom, all of which leads to the same final structure, i.e. singly-dehydrogenated curved corannulene (C₂₀H₉⁺). The BDEs for the H-shifted intermediate can be derived from the difference in energy between corannulene and singly dehydrogenated corannulene and are:

3.19 eV for int1 (loss from the aliphatic group),

2.63 eV for int2a (loss from the tertiary carbon) and

2.45 eV for int2b (loss from the vinylidene CCH₂ chain).

H loss from int4 can proceed from the CCH₂ chain at slightly higher BDE (3.28 eV) than for the other H-shifted isomers; this is because, in this case losing one of the aliphatic hydrogens from the chain will create a high-lying isomer of C₂₀H₉⁺ rather than the more stable singly-dehydrogenated corannulene structure. For int5, all the H atoms present in the molecule are aromatic (Fig. 5) and their BDE will be around 5 eV. The loss of a second H atom from the singly-dehydrogenated corannulene cation has a BDE of 3.24 eV. We also investigated the H₂-loss channel from int1. The transition state is at 4.85 eV above corannulene. This behaviour and the BDEs presented in this work agree very well with previous studies on planar PAHs.^55–58

Isomerization is also the first step toward the C-loss channels, which needs comparable energies of 7.4–7.5 eV (Fig. 5). Once int1 is created, the bond between two carbon atoms can be broken with a transition state (ts2b) of 3.40 eV, leading to the formation of a vinylidene group (int2b, 2.51 eV) on an almost flat molecule. From there a CCH₂ unit can be released through a C–C bond dissociation with 7.44 eV of energy with respect to corannulene (4.93 eV from int2b). Interestingly, analysis of the optimized structures along the C–C dissociation curve shows that the loss of a CCH₂ unit triggers the rearrangement of 2 carbon bonds in the molecule. This leads to a change in the position of the pentagon in the structure, which is now on the outside and thus leads to an almost flat structure for the C₁₈H₈⁺ isomer. An additional path leading to the release of a CCH₂ unit goes through intermediate int2a (2.33 eV) where the H moves to the tertiary carbon atom. The breaking of a C–C bond and migration of the H atom (ts4) leads to the formation of another C₂₀H₁₀⁺ isomer with a vinylidene group (int4)¶ at 2.55 eV with respect to C₂₀H₁₀⁺. With additional 4.90 eV, int4 can lose a CCH₂ unit while undergoing a carbon rearrangement similar to what was observed for CCH₂ loss from int2b.

The formation of int2a can also lead to the creation of an ethynyl group (int5) through a lower transition state (ts5, 2.92 eV). Int5 is only 1.11 eV above the corannulene cation. From there, CH can be released from the molecules with 6.44 eV, leading to the formation of an almost flat isomer of C₁₉H₉⁺ that contains a seven-membered ring. As observed for other intermediates analysed before, the structure rearrangement (the formation of a seven-membered ring from the insertion of the remaining C atom into the six-membered ring, leading to int7) happens toward the last stages of the CH bond cleavage. Calculations show that the structure with a 7-5 membered ring is thermodynamically preferred over a 6-6 structure because the heptagon facilitates the formation of a triple bond—r = 1.23 Å—between the two dehydrogenated carbon atoms in the 7-membered ring.

4.1.4 Further CH and C₂H₂ losses. Fig. 6 shows two possible pathways toward further CCH₂ and CH losses from the daughter molecules of corannulene cations, namely C₁₉H₉⁺, C₁₈H₈⁺ and C₁₇H₇⁺, informed from the results obtained for the fragmentation of the parent molecule. Since the limiting step for both H- and C-loss channels in the corannulene cation is the direct dissociation of a C–H or C–C bond, we investigate here only the final structures and not the H-roaming/C–C breaking steps leading to fragmentation. Rearrangements of the carbon structure, even if likely happening, are not considered here. As explained above, the corannulene cation can lose a CCH₂ group, leading to the formation of a C₁₈H₈⁺ isomer (orange solid arrow, top part of Fig. 6). From there, H roaming isomerization, similar to what is observed in coronene, promotes the opening of the bottom left hexagonal ring (see Fig. 5) and can lead to C₁₇H₇⁺, thanks to a CH loss (with a BDE of 8.50 eV) or to C₁₆H₆⁺ through a CCH₂ (BDE of 9.26 eV). It is expected that this set of reactions can proceed also from the bottom right ring of C₁₈H₈⁺ with similar energetics. For C₁₈H₈⁺ the BDEs for CH and CCH₂ have a larger difference (0.76 eV) than for the corannulene cation (0.10 eV). It is interesting to notice that the loss of CH from C₁₈H₈⁺ does not promote the rearrangement of the pentagon into a hexagon but requires an additional reaction step. The isomer of C₁₇H₇⁺ considered here already possesses two isolated radical sides and likely the formation of a new bond due to rearrangement of the pentagon-C group into a hexagon does not lower the energy of the molecule significantly.

Fig. 6 DFT optimized structures and reaction energies involved in the sequential loss of 2 CCH₂ (orange solid lines), CCH₂ and CH, and 3 CH loss (green dashed lines) from the corannulene cation.

The bottom sequence of Fig. 5 illustrates the structures related to sequential loss of 3 CH groups from C₂₀H₁₀⁺ (green dashed arrow). Starting from C₁₉H₉⁺, a CH loss from the bottom right hexagon will lead to the formation of a C₁₈H₈⁺ isomer with two heptagons. This requires 8.15 eV of energy, just 0.7 eV higher than the BDE for the first CH loss. A third CH loss can proceed from the only hexagon left in C₁₈H₈⁺ with 10.28 eV, leading to the formation of a C₁₇H₇⁺ isomer (bottom of Fig. 6). The energy needed for this last CH loss is 2 eV higher than for the previous steps: this is likely because the ethynyl chain on C₁₇H₇⁺ cannot be inserted into the heptagon to create a 9-membered ring that stabilizes the structure.

4.1.5 Discussion – fragmentation of the corannulene cation. The experiments on corannulene fragmentations show H and 2H/H₂ loss as the lowest dissociation channel (Fig. 2 and 3a). Looking at the PES, the H loss channel very likely happens from H-shifted isomers of C₂₀H₁₀⁺ such as int1, int2a and int2b rather than directly from the corannulene structure. The H₂ loss can also happen from isomer int1 and the height of the barrier is comparable to that of H-loss. It is not possible, based on energetics only, to constrain if the peak at m/z = 128 is due to sequential hydrogen loss or H₂ loss. Monte Carlo simulation⁵⁶ based on DFT-derived RRKM rates can be used to disentangle between the two contributions but is out of the scope of this paper.

Regarding the C-loss channels, the DFT calculations show that the C₁₈H₈⁺ peak results from CCH₂ rather than acetylene loss. The investigation of the PES did not lead to any viable routes toward direct or indirect acetylene loss, even considering higher multiplicities. See the ESI† (Fig. S1) and the footnote in the previous section. The competition between the CH and CCH₂ channels deduced from the experiments can be explained looking at Fig. 5, where the two channels go through int1 and have similar reaction energies. At lower internal energy the CCH₂ loss happens likely from int2b, while the CH loss starts from int5. The relative importance of the two channels at different internal energies will depend on the number of possible channels (including back-reactions and further isomerization) and also on the enthalpy of formation of the two C–H cleavage reactions, that influences the shape of the reaction rate.⁵⁹ Moreover, the irradiation of corannulene cations with the dye laser produces nanosecond pulses leading to a multi-photon excitation process that eventually leads to fragmentation of the molecule involving virtual energy state(s). As the laser pulses are applied the internal energy of the molecular cation increases which leads to fragmentation following the potential energy surface. This being a non-linear photochemical process, both CH and CCH₂ loss channels are produced in competition in spite of the minor energy difference of 1.5 eV (int 3 and int 7 compared to int 2a and int 5 respectively) after the H roaming and opening of the C skeleton.

Considering further fragmentation, Fig. 6 shows that the sequential loss of two CH units or that of one CCH₂ from C₂₀H₁₀⁺ leads to a population of different C₁₈H₈⁺ isomers. From there only CH loss is observed, leading to C₁₇H₇⁺, while there is no trace of C₁₆H₆⁺ peaks. This is different from what is observed in the fragmentation of corannulene cation, where the two channels are in competition. Calculations of the reaction energies for CH and CCH₂ loss (Fig. 6 top sequence) show that CH-loss of C₁₈H₈⁺ requires indeed lower energy. The presence of partly dehydrogenated sites in different rings affects the energy of this reaction, because it prevents rearrangement of the structure leading to isomers with lower energy. As mentioned above, there might be a mixture of C₁₈H₈⁺ isomers and they will likely not follow the same photodissociation pathway. Also, we cannot exclude that additional isomerization (e.g. H-migration) might take place on some or all of these structures. The presence of two groups of C₁₈H₈⁺ isomers, behaving differently, could be the reason why the experiments show that this peak increases upon increasing the laser pulses and starts decreasing on further irradiation (Fig. 3a). The isomer of C₁₈H₈⁺, formed by two consecutive CH losses (as shown in the bottom part of Fig. 6) is expected to undergo further fragmentation to form C₁₁⁺. This behaviour could be inferred from Fig. 3b and 4b, by the fact that the intensity of the C₁₈H₈⁺ trace does not completely drop to zero.

Finally, the mass spectra obtained with an RF value of 1200 V (e.g.Fig. 4a) show the presence of strong peaks corresponding to C₇H₃⁺ and C₈H₅⁺. These peaks appear very early in the fragmentation process, and it is postulated they might come directly from the parent molecule, likely after the isomerization reaction. Indeed, it has been proposed that carbon skeleton rearrangement (e.g. formation of 7-membered rings) in PAHs can lead to dissociation to form long hydrocarbon chains or carbon rings.⁵⁵ Looking at Fig. 6, we note that the C₁₇H₇⁺ isomer at the bottom of the figure, and its parent molecule C₁₈H₈⁺, have a 7-membered ring with 3 H atoms. The cleavage of two carbon bonds could release a C₇H₃ fragment from the molecules, e.g. (without considering the charge):

C₁₈H₈⁺ → C₇H₃⁺ + C₁₁H₅

C₁₇H₇⁺ → C₇H₃⁺ + C₁₀H₄

In this type of experiment, it is expected that the charge of the parent molecule would remain with the largest fragment, yet a strong peak is observed for C₇H₃⁺ but not for C₁₀H₄⁺ or C₁₁H₅⁺. A possible explanation is that the large fragments mentioned above quickly break down into smaller neutral hydrocarbons in the release of C₇H₃⁺. Molecular dynamics simulations based on DFTB on PAHs have shown these multiple hydrocarbon fragmentation channels if enough energy is available.⁶⁰ It could also be the case that the ionization potential of C₇H₃ is smaller than the remaining larger fragment. A previous study on both the cation and radical neutral version of C₇H₃ shows that, among the possible isomers, a three-membered carbon ring fused with the linear C₄H chain is the local minimum of the PES and its ionization energy is only 1.6 eV.

A similar reasoning can be applied to the other abundant hydrocarbon, C₈H₅⁺, that can also be released from C₁₈H₈⁺ (Fig. 6, bottom) after consecutive cleavage of 3 C–C bonds, leaving behind a bicyclic structure made of one heptagon and one hexagon. Unfortunately, there are no studies on the properties of C₈H₅⁺ and a detailed investigation of its PES⁶¹ is beyond the scope of this paper.

4.2 Sumanene cation (C₂₁H₁₂⁺)

An electron energy of 70 eV and emission of 0.6 mA was used to produce a good amount of sumanene cations, C₂₁H₁₂⁺. Fig. 7 shows the mass spectrum of C₂₁H₁₂⁺ obtained with the i-PoP system. The m/z = 264 corresponds to the parent cation C₂₁H₁₂⁺. The moderately intense peaks at m/z = 265, 266, 267 and 268 are due to the ¹³C components present in the sample. The ¹³C isotopic abundance in the sample was estimated to be 20% by estimating their integrated band area with respect to the parent cation. It is to be noted that the natural abundance of ¹³C₂₀H₁₂ is slightly higher than that of other PAHs like corannulene or coronene. The reason behind this is still unknown as sumanene is one of the least studied PAHs to date. However, the abundance of the ¹³C components was reduced further to 15–17% in the photolysis experiments presented in this work by applying a narrow bandwidth SWIFT frequency. But caution must be taken in understanding the fragment cations after laser irradiation as the ¹³C isotopes could also contribute to them. It is also observed that the electron impact on gaseous C₂₁H₁₂⁺ to ionize the neutral parent species also induces H abstraction to produce C₂₁H₁₁⁺, C₂₁H₁₀⁺, C₂₁H₉⁺ and C₂₁H₈⁺. The mass of 263 and 261 could also arise due to the dehydrogenation of the ¹³C components producing ¹³CC₂₀H₁₀⁺ and ¹³CC₂₀H₈⁺ respectively. It is also observed that the double dehydrogenation of the sumanene cation is more favourable by comparing the peak intensity of C₂₁H₁₁⁺ and C₂₁H₁₀⁺.

Fig. 7 Mass spectrum of C₂₁H₁₂⁺ obtained upon electron impact using the Re-TOF instrument integrated into the i-PoP system.

4.2.1 Photodissociation of C₂₁H₁₂⁺ – the high mass range. Fig. 8a presents the TOF-MS of C₂₁H₁₂⁺ for the pulse energies of 3.365 mJ per pulse at 620 nm. The mass spectrum recorded at 0 pulse shows that –H and –2H loss peaks of the parent cation of sumanene are already present upon electron impact. There are also peaks of –3H and –4H present in trace amounts. But as the laser irradiation continues, these H loss peaks increase until 4 pulses and then start to decrease to form secondary cationic fragments as shown in Fig. 8a for the case of C₂₁H₁₀⁺ and C₂₁H₉⁺. In total, six consecutive H loss channels are observed for the parent cation. Sumanene does not completely dehydrogenate unlike hexabenzocoronene (C₄₂H₁₈⁺)^17,18 or dibenzopyrene (C₂₄H₁₄⁺)⁵³ that consecutively lost all H atoms.

Fig. 8 (a) TOF mass spectra for sumanene radical cations irradiated with 0, 2, 4, … and 30 laser pulses with a pulse energy of 3.365 mJ per pulse with RF = 1600 V i.e. spectra displaying the high-mass cations with prominent features in the m/z = 280–150 range; (b) normalized mass peak areas corresponding to the most prevalent hydrocarbon clusters formed in the fragmentation of the sumanene with RF = 1600 V.

In these experiments, 50 scans were taken for each mass spectrum. Such mass spectra were obtained after 2 pulses, 4 pulses, etc. By accumulation of the fragments upon sequential laser induced fragmentation of sumanene cation, the mass spectra are stacked to visualize the increase in fragmentation yield, as presented in Fig. 8a. The RF value used in Fig. 8a is 1600 V which is efficient to detect mass peaks from 98–260 m/z. This can display the CH/C₂H₂ loss channels in better detail than the low mass cations. It is obvious from the mass spectrum that the C₂H₂ is the most dominant channel in C₂₁H₁₂⁺ fragmentation.

C₂₁H₁₂⁺ (–C₂H₂) → C₁₉H₁₀⁺ (–C₂H₂) → C₁₇H₈⁺

C₁₇H₁₀⁺ (–H) & (–C₂H₂) → C₁₅H₇⁺ (–C₂H₂) → C₁₃H₅⁺

It is to be noted that the first two C₂H₂ losses have an even number of H atoms and the next two C₂H₂ losses have an odd number of H atoms. This could be explained by the isomerization of C₂₁H₁₂⁺ after two consecutive C₂H₂ loss channels, as isomerization becomes the key process in the formation of such rich carbon species especially the low mass cations with m/z < 140. A carbon loss channel could also be expected here, i.e.

C₁₇H₈⁺(–C₂) → C₁₅H₈⁺

and

C₁₅H₅⁺(–C₂) → C₁₃H₅⁺.

The C₁₇H₈⁺ that formed can either undergo further C₂H₂ loss and/or a H₂ loss as shown below:

C₁₇H₈⁺(–C₂H₂) → C₁₅H₆⁺ (–C₂H₂) → C₁₃H₄⁺

(and/or)

C₁₇H₈⁺(–H₂) → C₁₇H₆⁺

The C₁₉H₈⁺ cation which is observed in the mass spectrum can also be formed from the parent cation by two hydrogen losses followed by a C₂H₂ loss:

C₂₁H₁₂⁺(–H₂/2H) → C₂₁H₁₀⁺(–C₂H₂) → C₁₉H₈⁺.

The C₁₉H₈⁺ is further found to fragment in two possible ways:

C₁₉H₈⁺(–C₂H₂) → C₁₇H₆⁺

(and/or)

C₁₉H₈⁺(–H₂) → C₁₉H₆⁺.

4.2.2 Photodissociation of C₂₁H₁₂⁺ – the low mass range. In the low mass region C₉H₃⁺ is the most abundant product and happens to be one of the primary products of fragmentation, which is also noticeable from the breakdown diagram in Fig. 8b.

C₁₀H₅⁺ and C₁₁⁺ are the second most abundant products. To get better nuances in this region experiments were performed with a lower RF value of 1200 V, as shown in Fig. 9a, where the masses below 140 m/z are intense. Here C₇H₃⁺, C₈H₅⁺ and C₉H₃⁺ are the most abundant products of fragmentation of C₂₁H₁₂⁺, which is followed by C₉H₈⁺, C₁₀H₅⁺ and C₁₁⁺. It is evident from Fig. 9b that these low mass cations are formed from the very beginning of laser irradiation. However, the breakdown equations predicted in this section for the origin of low mass cations (below 150 m/z) are only based on the data obtained in the mass spectrum and the breakdown diagram just like the case of corannulene. The intermediate ions and transition states before the formation of the observed low mass cations are not showcased in the pathways given below.

C₂₁H₁₂⁺ → C₂₁H_12−n + → C₇H₃⁺ + C₁₄H_9−n

C₂₁H₁₂⁺ → C₂₁H_12−n + → C₈H₅⁺ + C₁₃H_7−n

C₂₁H₁₂⁺ → C₂₁H_12−n + → C₉H₃⁺ + C₁₂H_9−n

C₂₁H₁₂⁺ → C₂₁H_12−n⁺ → C₁₀H₅⁺ + C₁₁H_7−n

C₂₁H₁₂⁺ → C₂₁H_12−n⁺ → C₁₁⁺ + C₁₀H_12−n

C₁₇H₁₁⁺ → C₂₁H_12−n⁺ → C₁₀⁺ + C₇H_1−n

Fig. 9 (a) TOF mass spectra for sumanene radical cations irradiated with 0, 2, 4, … and 30 laser pulses with a pulse energy of 2.581 mJ per pulse with RF = 1200 V i.e. spectra displaying the low-mass cations with prominent features in the m/z = 80–150 range; (b) normalized mass peak areas corresponding to the most prevalent hydrocarbon clusters formed in the fragmentation of the sumanene with RF = 1200 V.

Since the mass visibility is more focused on the smaller cations it is clearly observed that C₇H₃⁺ and C₈H₅⁺ are the most favourable fragments of C₂₁H₁₂⁺. It is interesting to note that a laser pulse energy of 2.581 mJ per pulse is sufficient to fragment the C₂₁H₁₂⁺ to produce C₇H₃⁺ and C₈H₅⁺ just after 2 laser pulses. The other cationic fragments shown in Fig. 9b are in trace amounts; and most of them are produced at the beginning of the irradiation and keep increasing until pulse 30.

4.2.3 Theoretical results for the fragmentation of sumanene cation. Fig. 10 shows the PES of the sumanene cation C₂₁H₁₂⁺. For the theoretical section of sumanene, the different fragmentation channels identified are discussed followed by the explanation in the context of the experimental results. Because of the structure and symmetry of the molecule (C_3v), there are only two type of H atoms considered in the first fragmentation: H atoms belonging to pentagons (blue circles in Fig. 10) and H atoms belonging to hexagons (red circles). The BDEs for these two H types are 3.95 and 4.96 eV, respectively, with the removal from the pentagon being favoured. The difference in energy reflects the different character of the C–H bonds in the two cases, aromatic for the hexagons and aliphatic for the pentagons. A similar behaviour was reported both experimentally⁶² and theoretically⁶³ for the fluorene cation that also possesses a pentagon unit with 2H atoms. In those studies, the derived BDE for the H-loss from the pentagon was between 1.5 and 2 eV lower that one reported for sumanene here. H-loss can also happen after an isomerization reaction, in int1, where the most likely H to be removed is the one attached to the tertiary carbon atom with a BDE of 2.07 eV. In fact, the H roaming to the tertiary carbon is energetically the most preferred route compared to the direct H loss that needs 3.95 eV.

Fig. 10 Potential energy surface (PES) for the C₂₁H₁₂⁺, obtained at the M06-2X/6-311++G(3df,2pd) level. Energies are given relative to the energy of the sumanene cation. All the structures are in their lower multiplicity state (doublet), except for the structure resulting from the H-loss from the hexagon (triplet) and int6 (quadruplet). Dissociation channels are color-coded to help visualisation. In addition, red indicates the H atom is removed from a hexagon, and blue from a pentagon.

The removal of the second H atom from the pentagon where the first H was removed requires 4.25 eV, while the removal of a second hydrogen from the partly dehydrogenated hexagon requires slightly less, 4.00 eV (not in the figure). If we consider that removing a H from any of the remaining pentagons requires also around 3.95 eV (see Fig. 10), this means that there is strong competition between the different H-loss channels and the peak corresponding to C₂₁H₁₀⁺ will be made of the isomer with both the H atoms abstracted from the sp³ hybridized pentagon site. H₂ can also be released from the aliphatic side with an energy barrier of 3.86 eV. As for corannulene, energetics are not enough to constrain if in the experiments we are witnessing H₂ or 2H loss.

4.2.4 Discussion – fragmentation of sumanene cation. For the sumanene cations, C-loss channels pass through isomerization (Fig. 10), in particular H moving from the pentagon ring to a tertiary carbon (int1, 0.89 eV) through a low barrier (ts1, 1.5 eV). From there the breaking of a C–C bond (ts2, 2.19 eV with respect to int1) leads to the formation of int2 (2.51 eV), where the hexagon opens up, creating a C₂H₂ unit that is attached to the nearby pentagon. Direct CH loss can happen from the C₂H₂ unit with a BDE of 5.68 eV (with respect to int2), leaving behind a C₂₀H₁₁⁺ structure (int7) where the remaining CH is inserted in the pentagon and forms a hexagon. Alternatively, only additional 3.33 eV are needed to release C₂H₂; the resulting int3 (C₁₉H₁₀⁺) can isomerize to the more stable int4 (1.68 eV with respect to int3). Further C₂H₂ loss is expected to occur from int3 and int4 with a similar mechanism and almost the same BDE (5.74 eV for 2nd C₂H₂versus 5.84 for the 1st). Int5 (C₁₇H₈⁺) is the first daughter of sumanene that is quasi-planar, and further loss of C₂H₂ from there requires almost 8 eV. The presence of several radical site in different rings is responsible for this high BDE value. We cannot exclude that additional structure rearrangement can take place, possibly lowering the BDE of the third C₂H₂ loss. This potential energy surface is in very good agreement with the experimental results where only C₂H₂ loss channels are observed.

4.2.5 Comparison of the fragmentation pattern of buckybowls. Sumanene fragmentation behaves very differently to that of coronene and corannulene because of the presence of three sp³ hybridized carbons associated with the pentagons in its peripheral structure, whereas each hexagon carbon in the periphery is bonded to only one hydrogen atom. The presence of pentagons affects the type of isomerization, leading sumanene to lose C₂H₂ rather than CCH₂ as seen in the case of coronene and corannulene. Furthermore, the presence of a pentagon seems to facilitate Stone–Wales-like rearrangement (i.e. two neighbouring hexagons isomerizing to one pentagon and one 7-membered ring), in the structure of both corannulene and sumanene, which leads to more stable products and thus lower BDEs.

4.2.6 Fragmentation pattern of C₂₀H₁₀⁺ and C₂₁H₁₂⁺ compared to C₂₄H₁₂⁺. Coronene and corannulene have very similar periphery (hexagons with duo group) and similar isomerization mechanisms leading to C-loss channels but reaction barriers and intermediates are in general lower for corannulene. The experiments show different behaviour, such as mostly dehydrogenation for coronene and CH/C₂H₂ competition for corannulene. The different behaviour, when it comes to the competition at lower energy might be ascribed to the curvature of corannulene. H-roaming reaction facilitates ring-opening (see Fig. 5), which in turn releases the strain in the C–C bond of the corannulene structure. We expect daughter molecules of corannulene to be mostly flat as revealed by the predicted molecular structures. In principle, once the first fragmentations take place, the presence of a planar structure should reflect in the experiments which could be very similar to the initial fragments observed in coronene. It is interesting to note that corannulene loses CCH₂ as shown in Fig. 5, whereas sumanene loses C₂H₂ as shown in Fig. 10, though both are fragmented from a peripheral hexagon ring. This demonstrates that H-roaming plays a crucial role in the fragmentation process.

Similarities and differences between fragmentation of corannulene and sumanene cations: the initial H losses were similar for both corannulene and sumanene – more favourable even-numbered H losses, and losses up to six H atoms upon photodissociation. Corannulene cation displayed a competition between CH and C₂H₂ losses, whereas sumanene cation displayed only C₂H₂ losses. The most intense low mass cations produced by photodissociation of corannulene and sumanene were mostly similar (C₇H₃⁺, C₈H₅⁺, C₉H₃⁺, C₁₀H₅⁺). C₁₁⁺ was observed in moderate intensity in these experiments.

The experimental and theoretical results demonstrate that peripheral H loss is more important than curvature in the photo-fragmentation process. It is evident that the amount of H loss that eventually creates two or more radical sites is crucial to determine the H roaming mechanisms. The H roaming process in corannulene and sumanene is one of the main reasons why CCH₂ loss is observed in corannulene, whereas C₂H₂ loss is observed in sumanene.

4.3 Corannulene dication (C₂₀H₁₀⁺⁺)

4.3.1 Photodissociation of C₂₀H₁₀⁺⁺ with 630 nm. The electron energy when increased to 90 eV and emission of 0.65 mA produced an intense peak of the corannulene dication, C₂₀H₁₀⁺⁺ (m/z = 125). The SWIFT pulses were used to isolate the dication peaks and this was further subject to laser irradiation with red light from the dye laser at 630 nm, which provides the maximum energy of 3.664 mJ per pulse at a Q-switch delay of 163.8 μs. This provides the optimum laser power for the fragmentation of dications (Fig. 11a). The photo-fragments produced after irradiation were all found to be cations and not dications. This means that the C₂₀H₁₀⁺⁺ tends to fragment into two different cations. The parent peak (C₂₀H₁₀⁺⁺) at 125 m/z could possibly have a contribution from C₁₀H₅⁺, as the mass cannot be resolved to distinguish the two cations. Similarly, the C₂₀H₆⁺⁺ peak could have contributions from C₁₀H₃⁺.

Fig. 11 (a) TOF mass spectra for corannulene radical dications (C₂₀H₁₀⁺⁺) irradiated with 0, 2, 4, … and 40 laser pulses with a pulse energy of 3.664 mJ per pulse with RF = 1000 V i.e. spectra displaying the low-mass cationic fragments with prominent features in the m/z = 150–70 range; (b) normalized mass peak areas corresponding to the most prevalent hydrocarbon clusters formed in the fragmentation of the corannulene dication (C₂₀H₁₀⁺⁺).

The major products observed in the experiments were C₆H₃⁺, C₇H₃⁺, C₈H₅⁺, C₉H₃⁺, C₉H₈⁺ and C₁₁⁺. It was not possible to observe any cations beyond 150 m/z because of the low RF value of 1000 V. The very high abundance of C₆H₃⁺ and C₇H₃⁺ suggests that C₂₀H₁₀⁺⁺ is capable of fragmenting into more than two components, following these mechanisms:

C₂₀H₁₀⁺⁺ → C₇H₃⁺ + C₈H₅⁺ + C₅H₂

C₂₀H₁₀⁺⁺ → C₆H₃⁺ + C₈H₅⁺ + C₆H₂

C₂₀H₁₀⁺⁺ → C₆H₃⁺ + C₇H₃⁺ + C₇H₄

C₂₀H₁₀⁺⁺ → C₉H₃⁺ + C₈H₅⁺ + C₃H₂

C₂₀H₁₀⁺⁺ → C₉H₃⁺ + C₇H₃⁺ + C₄H₄

C₂₀H₁₀⁺⁺ → C₉H₃⁺ + C₆H₃⁺ + C₅H₄

C₂₀H₁₀⁺⁺–H₂ → C₂₀H₈⁺⁺ → C₉H₈⁺ + C₁₁⁺

The breakdown equations predicted above are based on the data obtained in the mass spectrum and the breakdown diagram. The breakdown of this dication into the low mass monocations is expected to involve several intermediate steps. The equations provided do not include any intermediate or transition states, as molecular dynamics simulations alone can elucidate them. Moreover, the corannulene dications were not found to have any observable dicationic fragment (though dicationic fragments could be present in trace amounts) based on the m/z analysis. This means that the majority of the fragments are more likely to be cationic species. The structure identification, (except the structure of C₇H₃⁺ which is already known⁶¹) and the elucidation of the breakdown sequence requires sophisticated molecular dynamics calculations, which is a future scope of this research.

4.4 Sumanene dication (C₂₁H₁₂⁺⁺)

4.4.1 Photodissociation of C₂₁H₁₂⁺⁺ with 630 nm. The electron energy when increased to 80 eV and emission of 0.65 mA produced an intense peak of the sumanene dication, C₂₁H₁₂⁺⁺ (m/z = 132). The SWIFT technique were used to isolate the dication peaks and was further subject to laser irradiation at 630 nm, which provides an energy of 2.573 mJ per pulse at a Q-switch delay of 140.0 μs (Fig. 12a). C₆H₂⁺, C₇H₃⁺, C₈H₅⁺, and C₉H₃⁺ are the most abundant products of fragmentation of C₂₁H₁₂⁺⁺. Again, the mass beyond 140 m/z could not be observed because of the constraint of the RF value (1100 V). The very high abundance of C₆H₂⁺ and C₇H₃⁺ suggests that C₂₁H₁₂⁺⁺ is also capable of fragmenting into more than two components as follows:

C₂₁H₁₂⁺⁺ → C₇H₃⁺ + C₈H₅⁺ + C₆H₄

C₂₁H₁₂⁺⁺ → C₆H₃⁺ + C₈H₅⁺ + C₇H₄

C₂₁H₁₂⁺⁺ → C₆H₃⁺ + C₇H₃⁺ + C₈H₆

C₂₁H₁₂⁺⁺ → C₉H₃⁺ + C₈H₅⁺ + C₄H₄

C₂₁H₁₂⁺⁺ → C₉H₃⁺ + C₇H₃⁺ + C₅H₆

C₂₁H₁₂⁺⁺ → C₉H₃⁺ + C₆H₃⁺ + C₆H₆

Fig. 12 (a) TOF mass spectra for sumanene radical dications irradiated with 0, 2, 4, … and 30 laser pulses with a pulse energy of 2.573 mJ per pulse with RF = 1100 V i.e. spectra displaying the low-mass cationic fragments with prominent features in the m/z = 140–70 range; (b) normalized mass peak areas corresponding to the most prevalent hydrocarbon clusters formed in the fragmentation of the sumanene dication (C₂₁H₁₂⁺⁺).

The sumanene dications were not found to have any observable dicationic fragment. So, the majority of the fragments are more likely to be cationic species just like in the case of corannulene. Fig. 12b shows the breakdown diagram of the cations produced upon irradiation of C₂₁H₁₂⁺⁺. The parent dication has the same mass as that of C₁₁⁺ (m/z = 132) which cannot be resolved with the mass resolution of this TOF-MS, but is expected to be produced as one of the significant products in almost all the PAHs experimented with the i-PoP system with C < 14. Hence the trace of the peak area of the C₂₁H₁₂⁺⁺ does not drop to zero but keeps growing after pulse 24.

5 Astrophysical implications

A clue to understanding the relationship between C₆₀ and PAHs in space could be established by studying the molecules that could be intermediate steps leading to the formation of C₆₀ from PAHs.¹⁷ This is where the buckybowls gain importance in astrochemistry. Theoretical calculations and molecular dynamics simulations have shown that a fully benzenoid planar PAH can undergo dehydrogenation and possess one or more radical sites, which in turn pave the way for isomerization under intense radiation fields to form pentagon rings through Stone–Wales like defects.^10,13 This effect was also experimentally observed in a previous study in our laboratory where photodissociation of C₆₆H₂₈ produced the fullerene, C₆₀. Such PAHs with pentagons are also suggested to be more stable under interstellar conditions.²⁷ Alternatively, pentagons could be formed in space through CO loss from methoxy-functionalized PAHs,²⁹ produced by photolysis of PAHs trapped in interstellar ices. In view of this, molecular dynamics of pentagon containing PAHs, such as corannulene and sumanene, becomes imperative to link the bridge between the C₆₀ and PAHs. Moreover, these are medium-sized PAHs with C ≥ 20 that are symmetric (C_5v and C_3v for corannulene and sumanene respectively) and stable. Moreover, these are similar in size and structure to the fully benzenoid PAH, coronene (C₂₄H₁₂), which is commonly studied as a typical PAH representative in molecular astrophysics.

AIBs are particularly bright in the so-called photodissociation regions (PDRs) in space. PDRs are regions in space where far-UV photons with energies between 6 and 13.6 eV penetrate molecular clouds, photo-dissociating molecules and ionizing atoms such as C and heating the gas to temperatures of 500 K. Spectroscopic observations have revealed that the characteristics of the emitting interstellar PAH family changes with location in the PDR. This has been attributed to the photochemistry and photo-isomerization produced by the penetrating far-UV photons,^10,64,65 leading to the loss of functionalized groups such as CH₃ and of aromatic H, and the isomerization to C₆₀ and other fullerenes and cages. The surfaces of these PDRs also contain abundant small hydrocarbon radicals, and it has been suggested that these are the photochemically daughter products of PAHs exposed to the strong far-UV radiation field.²⁰

This study on the photofragmentation of corannulene and sumanene demonstrates that the buckybowl cations in the photodissociation region can undergo fragmentation to produce a variety of high and low mass fragment cations, which may contribute to the rich diversity of the interstellar molecular inventory. These experiments reveal the top-down approach where large PAHs break down to form smaller hydrocarbons or carbon skeletons. Interestingly, the fragment cations like C₁₀⁺, C₁₁⁺, etc., produced in these experiments are expected to have ring structures and are more stable than the fragments with linear or branched structures. It is to be noted that the carbon clusters like C₁₀⁺, C₁₁⁺, C₁₂⁺, C₁₃⁺, C₁₄⁺, etc., were produced predominantly for PAHs like coronene, dibenzopyrene, etc. But the buckybowl cations produce only C₁₀⁺ and C₁₁⁺, and only in trace amounts. This indicates the important influence of planarity on the photofragmentation of a PAH. The number of radical sites and the ring opening mechanism in the carbon skeleton are crucial in deciding the formation of such ring-structured carbon cations. The low mass cations C₇H₃⁺ and C₈H₅⁺ were produced in high abundance for both cationic and dicationic buckybowls. These small cations could be present in high abundance in the ISM. Given the intense radiation fields in the ISM, the cationic fragments C₇H₃⁺, C₈H₅⁺, C₉H₃⁺, C₁₀H₅⁺, C₁₀⁺, C₁₁⁺, etc., could potentially be the proxies to hunt for PAHs including the buckybowls (and other mid- to large-sized PAHs) in space.

The absorption spectrum of the interstellar medium contains a set of 500 absorption bands, the so-called diffuse interstellar bands (DIBs), in the near-UV to far-red range. These DIBs are generally attributed to molecular electronic absorptions.^66,67 Indeed, five of these bands have been identified with the fullerene, C₆₀⁺.⁷ The photochemical products of the PAHs in the experiments reported here could also contribute to the DIBs. Further laboratory studies are warranted to characterize the rotational, vibrational, and electronic spectroscopic properties of these species and to enable a search for them in space.

Finally, we note that, though single photon excitations dominate in space, multi-photon processes have a non-negligible probability to occur as some hundred million absorption events will happen over the lifetime of an interstellar PAH.⁹ As a consequence of these processes, the PAHs either emit IR fluorescence, fragment, or isomerize. This rich inventory of pentagon containing buckybowl cations, dications and their photo-fragmented low mass cations are important candidates to search for with the James Webb Space Telescope.

6 Conclusion

This work demonstrates that a diverse variety of fragment cations can be produced by photodissociation of a PAH cation in an ion trap. The TOF mass spectrum provides information about the various high and low mass cations that are produced by irradiation of corannulene and sumanene cations with a red light from a dye laser. Corannulene was found to lose up to six H atoms initially, where even H losses were found to be more favourable than the odd ones. This is followed by a competition between CH and CCH₂ losses. The competition between the CH and CCH₂ losses is evident in the comparable energy difference of 0.09 eV, as estimated from the calculated potential energy diagram (7.45 eV and 7.54 eV respectively). This means that the energy to lose CH and CCH₂ can almost be considered equal considering the errors in the predicted energies.

Sumanene also lost up to six H atoms which was clearly followed by only C₂H₂ losses. The potential energy diagram displays an energy difference of 2.71 eV between the C₂H₂ and CH losses (5.84 and 8.19 eV respectively for the C₂H₂ and CH losses) which is the reason why only C₂H₂ losses are observed in the case of sumanene. Interestingly, the most intense low mass cations produced by photodissociation of corannulene and sumanene cations were mostly similar (C₇H₃⁺, C₈H₅⁺, C₉H₃⁺, C₁₀H₅⁺). The photodissociation of corannulene and sumanene dications were also found to produce intense C₇H₃⁺, C₈H₅⁺, C₉H₃⁺, and C₁₀H₅⁺ cations, which indicates that the fragmentation process is irrespective of the charge state of a PAH. The high abundance of these small hydrocarbons in the experiments described in this paper is interesting and calls for similar experiments, down to low mass fragments, also for other PAH cations to understand if these species can indeed be the end point of the top-down interstellar chemistry of PAHs. In addition, experimental and theoretical characterisation of the spectroscopy of these species (e.g. C₇H₃^+ 61) is fundamental to searching for them in the ISM.

These results provide an insight into the evolution of PAHs with pentagon(s) in space. These experiments determine the vulnerability of pentagon-bearing PAHs under intense radiations under space conditions and will aid in identifying possible fragments of PAHs with pentagon rings that could be present in the ISM. The future scope of this work is to provide the IR characteristics of such species that can be searched for in the JWST data.

Author contributions

PS – conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, validation, visualization, writing – original draft, writing – review and editing. AC – conceptualization, formal analysis, investigation, methodology, validation, visualization, writing – original draft, writing – review and editing. JK – investigation, methodology, validation, software, writing – review and editing. HL – supervision, investigation, visualization, validation. AT – supervision, conceptualization, funding acquisition, investigation, validation, writing – review and editing.

Data availability

The data supporting this article have been included as part of the ESI.†

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

P. S. acknowledges the European Union and Horizon 2020 Postdoctoral funding awarded under the Marie Skłodowska-Curie action (grant number 101062984) H. L. acknowledges the Netherlands Research School for Astronomy (Nederlandse Onderzoekschool Voor Astronomie, NOVA) and the Netherlands Organisation for Scientific Research (Nederlandse Organisatie voor Wetenschappelijk Onderzoek, NWO). A. G. G. M. T. acknowledges the NWO for a Spinoza Prize (Spinozapremie). We thank SURF (https://www.surf.nl) for the support in using the National Supercomputer Snellius.

References

1. A. Candian, J. Zhen and A. G. Tielens, Phys. Today, 2018, 71, 38–43.
2. E. Peeters, C. Mackie, A. Candian and A. G. Tielens, Acc. Chem. Res., 2021, 54, 1921–1933.
3. C. Joblin, A. Tielens, O. Berné and P. Pilleri, EAS Publ. Ser., 2011, 46, 49–54.
4. B. A. McGuire, R. A. Loomis, A. M. Burkhardt, K. L. K. Lee, C. N. Shingledecker, S. B. Charnley, I. R. Cooke, M. A. Cordiner, E. Herbst and S. Kalenskii, et al. , Science, 2021, 371, 1265–1269.
5. A. G. Tielens, From Gas to Stars to Dust, 1995.
6. J. Cami, J. Bernard-Salas, E. Peeters and S. E. Malek, Science, 2010, 329, 1180–1182.
7. E. K. Campbell, M. Holz, D. Gerlich and J. P. Maier, Nature, 2015, 523, 322–323.
8. E. K. Campbell, Mol. Phys., 2020, 118, e1797918.
9. A. G. Tielens, The physics and chemistry of the interstellar medium, Cambridge University Press, 2005.
10. O. Berné and A. G. Tielens, Proc. Natl. Acad. Sci. U. S. A., 2012, 109, 401–406.
11. M. Frenklach and E. D. Feigelson, Astrophys. J., 1989, 341, 372–384.
12. I. Cherchneff, J. R. Barker and A. G. Tielens, Astrophys. J., 1992, 401, 269–287.
13. S. E. Stein and A. Fahr, J. Phys. Chem., 1985, 89, 3714–3725.
14. O. Berné, J. Montillaud and C. Joblin, Astron. Astrophys., 2015, 577, A133.
15. B. Croiset, A. Candian, O. Berné and A. Tielens, Astron. Astrophys., 2016, 590, A26.
16. B. J. West, L. Lesniak and P. M. Mayer, J. Phys. Chem. A, 2019, 123, 3569–3574.
17. J. Zhen, P. Castellanos, D. M. Paardekooper, H. Linnartz and A. G. Tielens, Astrophys. J. Lett., 2014, 797, L30.
18. J. Zhen, D. Paardekooper, A. Candian, H. Linnartz and A. Tielens, Chem. Phys. Lett., 2014, 592, 211–216.
19. N. Ota, arXiv, 2017, preprint, arXiv:1704.06197 DOI:10.48550/arXiv.1704.06197.
20. J. Pety, D. Teyssier, D. Fossé, M. Gerin, E. Roueff, A. Abergel, E. Habart and J. Cernicharo, Astron. Astrophys., 2005, 435, 885–899.
21. T. Chen and Y. Wang, Astron. Astrophys., 2020, 644, A146.
22. H. Kroto, Astrophys. J., 1987, 314, 352–356.
23. J. Bouwman, A. J. de Haas and J. Oomens, Chem. Commun., 2016, 52, 2636–2638.
24. A. J. de Haas, J. Oomens and J. Bouwman, Phys. Chem. Chem. Phys., 2017, 19, 2974–2980.
25. S. Banhatti, D. B. Rap, A. Simon, H. Leboucher, G. Wenzel, C. Joblin, B. Redlich, S. Schlemmer and S. Brünken, Phys. Chem. Chem. Phys., 2022, 24, 27343–27354.
26. S. P. Ekern, A. G. Marshall, J. Szczepanski and M. Vala, J. Phys. Chem. A, 1998, 102, 3498–3504.
27. G. NarahariáSastry, et al. , J. Chem. Soc., Perkin Trans. 2, 1993, 1867–1871.
28. J. Oomens, B. G. Sartakov, A. Tielens, G. Meijer and G. von Helden, Astrophys. J., 2001, 560, L99.
29. J. Zhen, P. Castellanos, H. Linnartz and A. G. Tielens, Mol. Astrophys., 2016, 5, 1–8.
30. M. Gatchell, P. Martini, F. Laimer, M. Goulart, F. Calvo and P. Scheier, Faraday Discuss., 2019, 217, 276–289.
31. G. Rouillé, C. Jäger, M. Steglich, F. Huisken, T. Henning, G. Theumer, I. Bauer and H.-J. Knölker, ChemPhysChem, 2008, 9, 2085–2091.
32. C. Pérez, A. L. Steber, A. M. Rijs, B. Temelso, G. C. Shields, J. C. Lopez, Z. Kisiel and M. Schnell, Phys. Chem. Chem. Phys., 2017, 19, 14214–14223.
33. L. Bernstein, R. Shroll, G. Galazutdinov and Y. Beletsky, Astophys. J., 2018, 859, 174.
34. P. Sundararajan, M. Tsuge, M. Baba and Y.-P. Lee, ACS Earth Space Chem., 2018, 2, 1001–1010.
35. I. Weber, M. Tsuge, P. Sundararajan, M. Baba, H. Sakurai and Y.-P. Lee, J. Phys. Chem. A, 2022, 126, 5283–5293.
36. F. J. Lovas, R. J. McMahon, J.-U. Grabow, M. Schnell, J. Mack, L. T. Scott and R. L. Kuczkowski, J. Am. Chem., 2005, 127, 4345–4349.
37. P. Pilleri, D. Herberth, T. Giesen, M. Gerin, C. Joblin, G. Mulas, G. Malloci, J.-U. Grabow, S. Brünken and L. Surin, et al. , Mon. Not. R. Astron. Soc., 2009, 397, 1053–1060.
38. R. G. Lawton and W. E. Barth, J. Am. Chem., 1971, 93, 1730–1745.
39. T. Amaya and T. Hirao, Chem. Rec., 2015, 15, 310–321.
40. C. Lifshitz, Int. Rev. Phys. Chem., 1997, 16, 113–139.
41. I. Garkusha, J. Fulara, P. J. Sarre and J. P. Maier, J. Phys. Chem. A, 2011, 115, 10972–10978.
42. C. Joblin, L. D'Hendecourt, A. Leger and D. Defourneau, The First Symposium on the Infrared Cirrus and Diffuse Interstellar Clouds, 1994, p. 291.
43. H. Jochims, H. Baumgärtel and S. Leach, Astrophys. J., 1999, 512, 500.
44. N. A. Sassin, S. C. Everhart, J. I. Cline and K. M. Ervin, J. Chem. Phys., 2008, 128, 234305.
45. K. Gulyuz, C. N. Stedwell, D. Wang and N. C. Polfer, Rev. Sci. Instrum., 2011, 82, 054101.
46. J. Zhen, S. R. Castillo, C. Joblin, G. Mulas, H. Sabbah, A. Giuliani, L. Nahon, S. Martin, J.-P. Champeaux and P. M. Mayer, Astrophys. J., 2016, 822, 113.
47. G. Wenzel, C. Joblin, A. Giuliani, S. R. Castillo, G. Mulas, M. Ji, H. Sabbah, S. Quiroga, D. Peña and L. Nahon, Astron. Astrophys., 2020, 641, A98.
48. V. M. Doroshenko and R. J. Cotter, Rapid Commun. Mass Spectrom., 1996, 10, 65–73.
49. B. J. Lynch, Y. Zhao and D. G. Truhlar, J. Phys. Chem. A, 2003, 107, 1384–1388.
50. N.-y Chang, M.-y Shen and C.-h Yu, J. Chem. Phys., 1997, 106, 3237–3242.
51. M. Frisch, G. Trucks, H. Schlegel, G. Scuseria, M. Robb and J. Cheeseman, et al., Gaussian 16, Rev. C.01, Gaussian Inc., Wallingford, CT, 2016.
52. G. Schaftenaar and J. H. Noordik, J. Comput.-Aided Mol. Des., 2000, 14, 123–134.
53. H. R. Hrodmarsson, J. Bouwman, A. G. Tielens and H. Linnartz, Int. J. Mass Spectrom., 2022, 476, 116834.
54. R. E. March and J. F. Todd, Quadrupole ion trap mass spectrometry, John Wiley & Sons, 2005.
55. S. Panchagnula, J. Kamer, A. Candian, H. Hrodmarsson, H. V. J. Linnartz, J. Bouwman and A. G. G. M. Tielens, Phys. Chem. Chem. Phys., 2024 10.1039/d4cp01301h.
56. P. Castellanos, A. Candian, J. Zhen, H. Linnartz and A. Tielens, Astron. Astrophys., 2018, 616, A166.
57. G. Trinquier, A. Simon, M. Rapacioli and F. X. Gadéa, Mol. Astrophys., 2017, 7, 37–59.
58. B. West, F. Useli-Bacchitta, H. Sabbah, V. Blanchet, A. Bodi, P. M. Mayer and C. Joblin, J. Phys. Chem. A, 2014, 118, 7824–7831.
59. T. Baer and P. M. Mayer, J. Am. Soc. Mass Spectrom., 1997, 8, 103–115.
60. A. Simon, M. Rapacioli, G. Rouaut, G. Trinquier and F. X. Gadéa, Philos. Trans. R. Soc., 2017, 375, 20160195.
61. A. Chakraborty, J. Fulara, R. Dietsche and J. P. Maier, Phys. Chem. Chem. Phys., 2014, 16, 7023–7030.
62. B. West, S. R. Castillo, A. Sit, S. Mohamad, B. Lowe, C. Joblin, A. Bodi and P. M. Mayer, Phys. Chem. Chem. Phys., 2018, 20, 7195–7205.
63. J. Szczepanski, J. Banisaukas, M. Vala, S. Hirata and W. R. Wiley, J. Phys. Chem. A, 2002, 106, 6935–6940.
64. C. Joblin, A. Tielens, L. Allamandola and T. Geballe, Astrophys. J., 1996, 458, 610.
65. R. Chown, A. Sidhu, E. Peeters, A. G. Tielens, J. Cami, O. Berné, E. Habart, F. Alarcón, A. Canin and I. Schroetter, et al. , Astron. Astrophys., 2024, 685, A75.
66. A. G. Tielens and T. P. Snow, The diffuse interstellar bands, Springer Science & Business Media, 2012, vol. 202.
67. N. Cox and J. Cami, Proc. Int. Astron. Union., 2013, 9, 412–415.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4cp01247j

‡ Prof. Harold Linnartz sadly passed away on 31st December 2023.

§ In the rest of this paper we use the term H roaming and H migration interchangeably.

¶ Optimising int4 with quadruplet multiplicity leads to an alternative structure with a –CH–CH group at 4.24 eV with respect to the corannulene (1.69 eV higher than int4 in the doublet spin state (not in the figure). From there we found a transition state leading to the loss of C₂H₂ with a barrier of 6.36 eV. However, the IRC calculation was inconclusive and the products (C₁₈H₈⁺ + C₂H₂) lie at 7 eV, higher than the transition state, making this path unlikely.

---