Fragmentation of Multiply Charged C₁₀H₈ Isomers Produced in keV Range Proton Collision

Abstract

The dissociation of multiply charged C₁₀H₈ isomers produced in fast proton collisions (velocities between 1.41 and 2.4 a.u.) is discussed in terms of their fundamental molecular dynamics, in particular the processes that produce different carbon clusters in such a collision. This aspect is assessed with the help of a multi-hit analysis of daughter ions detected in coincidence with the elimination of H${}^{+}$ and CH_n⁺ (n = 0 to 3). The elimination of H${}^{+}$/C${}^{+}$ is found to be significantly different from CH₃⁺ loss. The loss of CH₃⁺ proceeds through a cascade of momentum-correlated dissociations with the formation of heavy ions such as C₉H₅⁺, C₉H₅²⁺ and C₇H₃⁺. The structure of such large fragment ions is predicted with the help of their calculated ground state electronic energies and the multi-hit time-of-flight (ToF) correlation between the second and third hit fragments if detected. Furthermore, we report experimentally the super-dehydrogenation of naphthalene and azulene targets, with evidence of complete dehydrogenation in a single collision.

1. Introduction

The dissociation dynamics of a variety of polycyclic aromatic hydrocarbons (PAHs) have been investigated under irradiation with ultraviolet (UV), vacuum UV (VUV) or energetic charged particles in several works in the past [1,2,3,4,5,6,7,8,9,10,11,12]. Such measurements provide fundamental insight into the quantum chemistry of the organic molecules with PAHs as a convenient model system. A considerable body of this research domain has focused on the dissociation of monocations. The measurements of monocations are often accomplished with high levels of computational effort, and investigating the dissociation dynamics of highly charged ions is a challenge. The excess Coulomb energy stored in the di- or trication can radically change the nature of the dissociation dynamics. In addition, the complexities of the data assimilation, analysis and interpretations make the investigations on multiply charged ions a relatively uncommon endeavour.

The dissociative multiple ionisation of PAHs can occur in hostile astronomical regions by a number of excitation processes, such as multi-stage VUV absorption, interaction with X-rays or the low-energy component of cosmic rays, stellar wind protons and/or ions and other energetic particles in the interstellar medium [13,14,15,16]. The laboratory research in this domain is mainly conducted using extreme UV or X-ray photon, while charged particle collisions remain less explored [16,17,18]. In the last two decades, the study of the upper atmosphere of the jovian planets and their moons have led to the substantial understanding of how energetic charged particle radiation from the Sun can play a crucial role in their atmospheric composition and evolution [19,20,21,22]. In situ measurements by the Cassini–Huygens mission have demonstrated the importance of the 10 to 100 keV energy range of protons in the dynamics of Titan’s ionosphere [21,22]. Such interactions are known to be more efficient in producing highly charged ions and diverse energetic fragment ions (C⁺, CH₃⁺ and C₂H₂⁺ etc.), which may eventually induce a very rich and complex chemistry in Titan’s ionosphere, atmosphere and haze.

Here, we compare the multi-fragmentation of highly charged azulene (C₁₀H₈) and naphthalene (C₁₀H₈) ions produced in fast proton collisions. Very little difference was observed between the multi-fragmentation mass spectrum of these two isomeric targets. We mainly focus on the emission of light fragments H${}^{+}$, C${}^{+}$, CH₃⁺. The former two channels occur via violent multi-fragmentation, while the latter proceeds via an intermediate isomer that is often overlooked due to its relatively low intensity. Also, complete or partial de-hydrogenation events are observed in this work under a single collision condition.

2. Computational Details

Multiply charged ions in the keV range collision are produced with a substantial amount of internal energy which assists in crossing various transition state barriers involved in isomerisation or dissociation processes. When a parent ion eliminates smaller neutral or ionic fragments like CH₃⁺ or C₂H₂⁺ from a precursor isomer, the smaller fragment will carry a substantial amount of vibrational energy per degree of freedom and the larger fragment will be produced close to the final equilibrium configuration of precursor isomer in the ground electronic state. Ground-state energies are therefore used here as a measure to predict the structure of large fragment ions. Various isomers of C₉H₅⁺ and C₉H₅²⁺, the large fragment ions formed after CH₃⁺ loss, were optimised by density functional theory (DFT) calculations using the Dunning basis set cc-pVDZ to predict the total ground state electronic energy. This basis was chosen due to its larger set of basis functions compared to the one used for the complete C₁₀H₈ ions, since the respective fragment species are not stable or naturally occurring equilibrium structures. The energies of the doubly and triply charged C₁₀H₈ isomers that can emit CH₃⁺ are also calculated. The relatively simple and stable ions of C₁₀H₈ were calculated using much simpler nonlocal hybrid B3LYP with 6-31G(d) basis, incorporated in the GAUSSIAN 09 package [23]. The numerical details of the structure and energies of all the relevant species are listed in the Supplementary Material.

3. Experimental Details and Analysis

A proton beam of energy ranging from 50 keV to 150 keV, in steps of 25 keV, was extracted from the electron cyclotron resonance ion source at the Low energy Ion Beam Facility (LEIBF) at the Inter-University Accelerator Centre (IUAC), New Delhi. These ions were made to interact with the target molecules in a ToF mass spectrometry system. The emitted electron (or electrons) as well as the possibly neutralised projectile in a given collision event were separately detected using channeltron detectors, one at the opposite direction of the recoil ToF tube and other at about 1.5 m distance from the interaction region. The recoil ions were accelerated and recorded using a 40 mm active diameter position-sensitive micro-channel plate detector (PSD) with a delay line anode. The data was recorded in common stop multi-hit mode. The secondary electron signal and the neutralised projectile signals were logically ORed and then delayed to be used as a common stop signal, with a relative delay adjusted to give priority to the projectile detection. The target molecules, naphthalene (nph) and azulene (az), were injected into the vacuum system using a long hypodermic needle connected to an external reservoir with a valve. The target samples were kept at room temperature and the effusion of the target vapour due to their vapour pressure at room temperature was found to be adequate to perform the experiment with an event rate of about 1 to 2 kHz. Entirely different target masses (not reported here) were used in between the two experiments and totally different supply lines were used for the two targets to avoid any cross contamination.

The recorded data were classified based on the electron emission and electron capture process and background noise was reduced using various sum conditions on the PSD. We report here the data sorted based on the electron emission mode. Although up to eight-hit events could be recorded, data up to the third hit were sufficient to evaluate the necessary fragmentation channels. The positive counts beyond the third hit were negligibly small. Typically, 76% and 19% of events recorded for the 50 keV proton beam collisions were purely single and double hits, respectively, and less than 4% of events were triple hits. The exact percentage depends strongly on the collection and detection efficiencies for the ion signal. The extraction field was high enough to collect all the ions onto the detector, However, the detection efficiency varies from ∼45% for a single carbon ion to ∼20% for an intact molecular ion. The detection efficiency of the micro-channel plate detector was calculated using an empirical model reported in the literature [24].

4. Results and Discussion

4.1. Normalisation Process

The single ionisation cross section of az and nph are known to be identical for the studied collision energies [11]. The single hit data of two isomers are normalised to the sum of singly charged parent ions (m/q 128 and 129), neutral H loss (m/q 127) and C₂H₂ loss (m/q 102) ions from parent monocation of a given isomer. The normalisation procedure was chosen for two reasons. Firstly, because of the neural H loss contribution, although negligible, wait s difficult to separate from the parent peak. Secondly, it was seen that both isomers have same single ionisation cross sections, but az${}^{+}$, being a high-energy isomer, experiences higher fraction of C₂H₂ emission than nph, the details of which have been published elsewhere [11]. All the one-dimensional (1D) spectra shown henceforth are normalised by this procedure and multiplied by an arbitrary factor of 1000 (arbitrary common factor). The two-dimensional (2D) spectra are shown as counts without normalisation.

4.2. Single Hit Analysis

The single hit mass spectrum of az and nph after normalisation are found to be similar, as shown in Figure 1. The mass spectra includes only the data wherein a single ionic species is detected in a given event; thus, the data shown is a mix of pure single ion events as well as events wherein only one of the fragment ions produced is detected. The peak at m/q 45 is acetone that was used for cleaning the target lines. The sharp peaks at m/q 20 and 40 are due to the addition of Ar to the background gas, and all the mass spectra analysed here were calibrated using these reference peaks. One striking difference between the two mass spectra is the 60% excess yield at m/q 102 peak (loss of neutral C₂H₂ from monocation) in the az mass spectrum. This was explained on the basis of the ground-state energy difference between singly charged az and nph in our previous work [11]. Similarly, hydrocarbon fragments with six or fewer carbon atoms are produced in a slightly higher yield for the target az. This may again be due to the less stable configuration of az in higher charge states as suggested in a recent work (Figure 1 in reference [25]). However, the fragments in the region m/q 98 (C₈H_m⁺ where 0 < m < 10) and 88 (C₇H_m⁺, where 0 < m < 10) have similar yields for two isomers in the mass spectrum. Overall, there is little difference between the fragmentation mass spectra of az and nph in high-energy collisions, which may be due to the lower energy difference between multiply charged isomers compared to singly charged or neutral isomers. In a recent paper, Lee et al. [26] demonstrated that even at much lower internal energies than in the present study, an equilibrium can be established between nph⁺ and az⁺ before dissociation. In fact, the energy transfer to the molecule is expected to be very high (typically few tens of eV) in collisional excitation with fast proton, so that the intact molecular ion can explore all available isomers prior to dissociation, independently from the initially chosen isomer.

4.3. Coincidence Analysis of Double Hit Data

A 2D ToF coincidence map is plotted with first-hit ToF on the horizontal axis and second-hit ToF on the vertical axis as shown in Figure 2. The correlation patterns found are similar for az and nph. The coincidence map is rich in several ToF correlation patterns, most of which are broad and lack momentum correlation, representing multi-fragmentation or charge separation of highly charged ions with more than two fragment ions produced in a single event. The islands in coincidence with H${}^{+}$ are the strongest channels in the 2D map. Most of the islands in the 2D spectrum are poorly resolved due to high momentum release, limiting our investigation to a few sets of channels. A few islands in the map are binary fragmentation channels, which are momentum-correlated, and most of them have an extended tail, indicating a fragmentation inside the extraction field of the ToF.

The carbon-conserving binary fragmentation islands that appear in the double-hit coincidence map (see Figure 2) are summarised below in Equation (1). Here, n = 8, 7 and 6 are major islands. An island with n = 5 also exists, but it is difficult to separate channels in this island due to the effect of pulse pair resolution. The island with n = 9 is not as intense as other channels. Some of the channels with n = 8 and n = 7 have an extended tail representing a dissociation within the extraction timescale of ToF. Meanwhile, no strong tail is observed for channels with n = 6, indicating fast dissociation before extraction.

$$\begin{array}{c}\hfill C_{10}{H}_8^{x+}⟹C_n{H}_m^{y+}+C_{10-n}{H}_p^{(x-y)+}+(8-m-p)H\end{array}$$

(1)

Fragments with mainly few carbon atoms accompany with dehydrogenation. For instance, as one goes from n = 9 to n = 6, the channels shift from hydrogenated to dehydrogenated channels. Furthermore, there is a clear evidence of loss of H in multiples of two as observed in the previous studies [27,28]. One such example is the emission of C₂H_n⁺, where the main channel follows 4H or 2H₂ loss, not even 2H/H₂loss. The loss of C₂H₃⁺, on the other hand, follows no such restriction. Channels on other islands cannot be summarised as Equation (1) because of the large incorporation of neutral and cationic hydrocarbon carbon fragments. Some interesting features are observed in such channels as discussed below in some examples.

4.4. H⁺ Coincidence

As shown in Figure 3, the second-hit spectrum of nph with H⁺ in the first hit is compared to its single-hit spectrum. All hydrocarbon fragments correlated to the first-hit ToF of H⁺ are observed with higher intensity compared to the similar fragments in the single-hit spectrum, which can be related to the increased detection efficiency of H⁺. Two quick observations can be made here: (i) none of the peaks in the second hit spectrum are sharp. This implies that all partner fragments are produced with some kinetic energy. Also, H${}^{+}$ islands have no momentum correlation in the 2D. Both of these observations indicate that H${}^{+}$ originates mainly from highly charged parent ions, with charge state greater than 2+, which is consistent with a previous theoretical results for multiply charged naphthalene [29]. (ii) the intensities of fragments containing an odd number of carbon atoms, especially CH_m⁺ and C₃H_m⁺, are substantially larger than the peaks in the single hit. This may be due to the oscillating binding energy of carbon clusters produced. The propensity of such cluster ions is usually decided by the energy of formation and the ionisation potential [30].

The second-hit mass spectra of az and nph obtained in coincidence with H${}^{+}$ are shown in Figure 4. For the peaks around n = 6, 7 and 8, there is no difference between az and nph. The intensity of all other fragments are slightly higher for az, which might be due to the inherently less-stable configuration of highly charged az compared to nph. Another striking observation is the presence of super-dehydrogenated az⁺ and nph⁺ (m/q 120–127) as shown in Figure 5. For the first time, we report here the evidence of losing all hydrogen atoms from az/nph (m/z 120 in Figure 5a) in a single collision. The two isomers of C₁₀H₈ studied here prefer to lose 2, 6, 7 or 8 hydrogens, at least one of which is lost as H⁺. These super-dehydrogenations and the fact that 3, 4 and 5 H loss channels are not favoured warrants a detailed investigation, which may be useful to understand the trend of atomic/molecular hydrogen in the astrophysical region where PAHs are irradiated by stellar wind protons.

4.5. C⁺, CH⁺, CH₂⁺ Coincidence

The second-hit mass spectrum with C⁺ in the first hit of az and nph is shown in Figure 6. The second-hit mass spectrum of C⁺ fragment is identical to the second-hit mass spectrum of H⁺, except the fragments with 8 carbon atoms are missing. The second-hit mass spectrum of C⁺ is similar to the second-hit mass spectrum of CH⁺ and CH₂⁺, which are mainly populated with lighter fragments. This indicates that these fragments are produced in the multi-fragmentation of a highly charged ion. The similarity between the second-hit mass spectrum of H⁺, C⁺, CH⁺ and CH₂⁺ may also suggest that these ions are produced in a single event, but H⁺ is not detected; instead, C⁺ is observed in the first hit.

4.6. CH₃⁺ Coincidence

The ring-opening and ring-expansion mechanisms in smaller PAHs are pivotal to the understanding of formation of larger PAHs in ISM. To this end, the HACA (Hydrogen-Abstraction/acetylene-Addition) process [31] has been in the discussion for a quite some time and has been considered as a possible mechanism for the growth of larger PAHs. But, in recent years, laboratory experimental results have conclusively demonstrated that the HACA mechanism is not a favourable option, since it fails to account for the possible rate of formation of PAHs in the ISM [32]. A more plausible hypothesis for ring expansion and PAH growth has been proposed by Zhao et al. [32]. This mechanism involves the addition of a methyl group to a pentagon ring in the PAH so as to expand the ring to a hexagon structure and thus successively grow PAH size. In a laboratory setup, it is much easier to study the dissociation process to understand the reverse barriers associated with dissociative/associative reactions, than to experimentally look for the associative reaction itself. In this context, we looked at the second hit mass spectrum of CH₃⁺ channel and compared it with C⁺ and CH⁺. We believe that an investigation of the parent conformers and associated reverse barriers will add significant value to the understanding of the methyl addition reactions in PAHs.

The second-hit mass spectrum of two isomers with CH₃⁺ in the first hit are shown in Figure 7. Once again we see the similarity between az and nph. There is a clear difference between the second-hit mass spectrum of C⁺ and CH₃⁺ channels. C₇H₃⁺, C₉H₅⁺ and C₉H₅²⁺ are the major fragments formed after CH₃⁺ loss. These channels appear with a tail in the 2D ToF mass spectrum (see Figure 2). We were able to determine the number of H atoms in each of these fragments using ToF correlations in the 2D. The mass spectrum of az below m/q of 46 was heavily contaminated by the acetone, which is used for cleaning the target lines before introducing this isomer target. The second-hit mass spectrum of az is subtracted from nph and shown in Figure 7b. The differential mass spectrum is dominated by fragments from acetone (below m/q of 46). We further analysed the CH_m⁺ (m/q 12 to 17) region in coincidence with H${}^{+}$ in the 2D ToF map (Figure 8). It is observed that masses m/q 12, 13 and 14 (C⁺, CH⁺, CH₂⁺) are clearly visible, but only a few counts at m/q 15 (CH₃⁺) are present due to chance coincidence with H⁺. This once more supports our argument that the production of H⁺, C⁺, CH⁺ and CH₂⁺ are related, whereas CH₃⁺ is produced exclusively by a different mechanism.

The relative ground state energy of C₁₀H₈²⁺/C₁₀H₈³⁺ isomers, which can emit CH₃⁺ are shown in

Table 1. Ground state energies of C₁₀H₈²⁺ and C₁₀H₈³⁺ isomers relative to the most stable isomer ($\Delta$Es, eV), as calculated using DFT method with 631G(d) basis.

	Structure	$\mathbf{\Delta }$E of Dication	$\mathbf{\Delta }$E of Trication
Nph		0	0
Az		0.41	0.29
A		3.07	2.20
B		1.96	1.14

. Structure A (see Table 1) was proposed as the precursor ion for CH₃⁺+C₉H₅⁺ channel by Kingston et al. [33]. This was further investigated by Leach et al. [18] based on the kinetic energy release (KER) of CD₃⁺ loss in dication of naphthalene-d${}_8$ (C₁₀D₈²⁺) [18]. The experimental KER reported by them was approximately 1 eV, whereas the structure A would correspond to about 2 eV KER and therefore Leach et al. [18] suggested that a linear geometry may be preferred over structure A. The most probable value of KER for this channel obtained in this work is 2.9 eV for nph and az targets. This value matches with the value measured recently by Reitsma et al. [34]. It suggests that structure A may be a common isomer of az and nph that can emit CH₃⁺, as originally suggested by Kingston et al. [33]. But, when we performed DFT calculations for other possible structures, we observed that the dication structure A was about 3 eV higher in the energy with respect to nph²⁺. On the other hand, structure B shown in Table 1 is found to be more favourable with 2.0 eV higher energy than nph²⁺. Hence, we propose that the elimination of CH₃⁺ might occur via a common dication conformer with structure B. Obviously, the exact transition of nph or az to the parent isomer of the CH₃⁺ loss channel would occur through various complex intermediate as well as transition states. Such calculations have been carried out for some important dissociation channels of monocations in the past [35,36], but rarely attempted for di- or trications of nph and az. Calculations of such complexity are presently beyond the scope of this work. But it does not impede the present work in explaining the CH₃⁺ elimination in az²⁺ and nph²⁺ via a common isomer, because the formation of a dication in the high energy proton collision proceeds via double plasmon excitation process, as suggested in our previous work [11]. This process is found to deposit an internal energy of about 13 eV in the resulting dication [11]. Thus, this internal energy will be sufficient to cross any possible transition state barrier the species may encounter on the way to producing CH₃⁺ eliminating parent structure.

4.7. Multihit Analysis of CH₃⁺ Channel

A ToF correlation between second and third hit fragments of C⁺ and CH₃⁺ channels are shown in Figure 9. Several broad islands are observed in the 2D coincidence map of the C⁺ channel. Only three binary dissociation channels of C₉H₅²⁺ are observed in the CH₃⁺ emission and are summarised below.

$$\begin{array}{c}\hfill C_9{H}_5^{2+}⟹C_n{H}_m^{+}+C_{9-n}{H}_{m-5}^{+}+(5-m-n)H\end{array}$$

(2)

The third-hit mass spectra of the CH₃⁺ channel for az and nph are shown in Figure 10. Both mass spectra are very similar, consisting of n = 2–9 fragments of moderate size. The formation of such neutral or ionic hydrocarbon species is of importance in the astronomical context [37]. For instance, the structural conformations of partially hydrogenated C₇ and C${}_9$ neutrals as well as ions are proposed as the possible carriers of some weak diffuse interstellar bands in the ISM [38,39]. Steglich et al. [38] computationally identified stable structures of C₉H₅ radicals, which are found to feature visible absorption bands that coincide with a few very weak diffuse interstellar bands. We consider here the same structures to identify stable isomers of mono- and dications of C₉H₅. As mentioned earlier, the di- and trications are produced at high internal energies in fast proton collisions. A significant fraction of this energy can be utilised to overcome various transition state barriers to be able to produce stable fragment geometries. The proposed isomers of C₉H₅⁺ and C₉H₅²⁺ and their ground state energies relative to the most stable isomer are given in

Table 2. Ground-state energies of C₉H₅⁺ and C₉H₅²⁺ isomers relative to the most stable isomer ($\Delta$Es, eV), as calculated using DFT method with cc-pVDZ basis.

	Structure	$\mathbf{\Delta }$E of Monocation	$\mathbf{\Delta }$E of Dication
A		0.9	0
B		1.11	0.45
C		1.33	0.52
D		0	0.49
E		0.27	0.02
F		0.22	1.17
G		1.54	0.2

. According to this calculation, the most stable isomers of C₉H₅²⁺ are A and E. The third hit mass spectrum of CH₃⁺ channel indicates that an ensemble of C₉H₅²⁺ isomers are formed, which can produce fragment partners, C₂H_m⁺−C₇H_m⁺, C₃H_m⁺−C₆H_m⁺ and C₄H_m⁺−C₅H_m⁺, in equal intensities. This observation combined with the energy of optimised structures suggests that structure A may predominate the ensemble of C₉H₅²⁺ isomers. We also propose that structure D (the lowest energy structure of C₉H₅⁺) might be an important fraction of C₉H₅⁺ isomers formed after eliminating CH₃⁺ from the final conformer of nph${}^2+$/az${}^2+$. The last two conclusions require more dedicated experimental and theoretical investigations.

5. Conclusions

Swift charged-particle-induced single and double ionisation of nph as well as az are known to be plasmon-dominated. These multiple ionisation processes deposit a high amount of excess energy in the resultant cations, which then can be considered to have a high probability for high-energy isomer formation, often followed by dissociation. One such mechanism of CH₃⁺ elimination is found to progress via a common isomer of nph and az. Moreover, a cascade of dissociation is observed in multihit analysis up to a third hit of this channel. New parent dication as well as trication isomers of nph and az are proposed here based on the ground-state energies calculated by DFT theory, which can eliminate CH₃⁺. C₉H₅⁺ formed after the loss of CH₃⁺ eliminates C₂H₂ to form C₇H₃⁺. The dication fragment C₉H₅²⁺ undergoes binary fragmentations such that the precursor ion is an isomer containing a long carbon chain. Various DFT optimised structural conformers of C₉H₅²⁺ were compared and the lowest-energy structure is found to have a structure of a pentagon with a long linear chain.

In addition to the decay cascade of the CH₃⁺ loss channel, a few other observations are worth noting. First, we report the super-dehydrogenation of nph and az in a single collision condition with a detectable intensity of total dehydrogenation. Second, the production of H⁺, C⁺, CH⁺ and CH₂⁺ are related, whereas CH₃⁺ is produced exclusively by a different mechanism. These observations warrant more theoretical investigations.