Resonances of the anthracenyl anion probed by frequency-resolved photoelectron imaging of collision-induced dissociated anthracene carboxylic acid

Abstract

Resonances in polyaromatic hydrocarbon (PAH) anions are key intermediates in a number of processes such as electron transfer in organic electronics and electron attachment in the interstellar medium. Here we present a frequency- and angle-resolved photoelectron imaging study of the 9-anthracenyl anion generated through collision induced dissociation (CID) of its electrosprayed deprotonated anthracene carboxylic acid anion. We show that a number of π* resonances are active in the first 2.5 eV above the threshold. The photoelectron spectra and angular distributions revealed that nuclear dynamics compete with autodetachment for one of the resonances, while higher-lying resonances were dominated by prompt autodetachment. Based on electronic structure calculations, these observations were accounted for on the basis of the expected autodetachment rates of the resonances. Virtually no ground state recovery was observed, suggesting that the smallest deprotonated PAH that leads to ground state recovery is the tetracenyl anion, for which clear thermionic emission has been observed. The use of CID and photodissociation of organic carboxylic acid anions is discussed as a route to studying the dynamics of resonances in larger PAH anions.

---

Introduction

Polyaromatic hydrocarbons (PAHs) are present in many different chemical environments, where they play critical roles such as in coal¹ and as a nucleation point in soot formation relevant to atmospheric chemistry.^2,3 As PAHs are both carcinogenic and mutagenic,^4,5 it is prudent to establish an understanding of their part in anthropogenic pollution. From a technological perspective, PAHs have been considered as a basis for new organic electronics such as OLEDs, photo-voltaic cells, and organic field effect transistors.^6,7 For such devices, a fundamental knowledge of the electronic and redox properties of PAHs is essential. This is also the case for gas-phase PAHs, which have long been considered to be a key source of carbon in the interstellar medium and their presence has been inferred from unidentified IR bands.^8–10 It has also been predicted that a significant fraction of interstellar PAHs are in an ionised state,^10,11 which is not surprising given that the overwhelming majority of normal matter in the universe is in a plasma state. As an example, C₆₀⁺ was recently identified to contribute to the IR bands.¹² In regions of increased electron density, anionic PAHs are also expected and other hydrocarbon based anions, such as C₆H⁻ and C₄H⁻, have been observed with abundances in the order of a few percent of that of their neutral precursors.^13–15 The mechanism of anion formation in the interstellar medium is still unclear. Low energy electron attachment via a non-valence dipole-bound state has been the most commonly invoked mechanism.^16,17 However, the minimum dipole moment required for a dipole-bound state is ∼2.5 D,¹⁸ so most neutral PAHs do not possess a sufficient dipole moment to support this mechanism. An alternative process involves electron attachment via the valence excited states of the anion embedded in the continuum. Although autodetachment is the most likely outcome following electron attachment into such a resonance, excited state processes including internal conversion can compete and may eventually form the anionic ground state.^19–22 Valence resonances are particularly prevalent in highly π-conjugated systems.^18,23,24 Electron impact studies have confirmed that such resonance-mediated decay processes occur²³ and ground state anions of tetracene (C₁₈H₁₂) and pentacene (C₂₂H₁₄) have been observed following electron impact energies of ∼3 eV.²⁵

All of the molecular anions that have been confirmed to exist in the interstellar medium are closed-shell^{13,15,26–31} and therefore, PAH anions that are believed to exist in the interstellar medium are also likely to be closed-shell. Hence, deprotonated PAH anions are likely candidates. Over the past few years, we have developed anion photoelectron spectroscopy as a probe for resonances in a range of radical^{20,20,21,32–34} and closed-shell anions,^35–37 including the tetracenyl anion, C₁₈H₁₁⁻.¹⁹ Following excitation to a resonance, C₁₈H₁₁⁻ showed a preference for internal conversion back to the ground state of the anion rather than autodetachment from the resonance. Hence, from an anion formation perspective, when neutral C₁₈H₁₁ is approached by a low-energy electron with a kinetic energy that matches the anionic resonance, then the electronic ground state of the anion can be formed despite the unbound nature of the resonance. As these dynamics are enabled by the presence of a large number of π* resonances, it is not unreasonable to expect that the same processes may also occur in larger polyacenes. However, for smaller polyacenes fewer low-lying resonances exist, which leads to the question – how small can a polyacene be to efficiently recover the ground electronic state from an anion resonance? Here we present a study of the frequency- and angle-resolved photoelectron imaging of the anthracenyl anion, C₁₄H₉⁻, and show that, although resonance dynamics can be clearly identified, ground state recovery is inefficient as evidenced by a low yield of thermionic emission.

C₁₈H₁₁⁻ was previously generated using electrospray ionisation of C₁₈H₁₂ dissolved in toluene. However, this produced the deprotonated anion in low yield and the overall ion signal suffered from instability. Furthermore, electrospray ionisation offers no selectivity over the isomeric form produced and generally forms the most stable species. In applying the same methodology to produce C₁₄H₉⁻, we found that the yield was too low and so here we employ an alternative method for producing the anthracenyl anion. Within the mass-spectrometry community, two main methods have been developed to generate carbanions.^38,39 The first is based on the reaction of F⁻ with a trimethyl-silanated molecule.^40,41 This was recently used by the Neumark group to study the neutral ground and first excited states of the three isomers n-C₁₄H₉ (with n = 1, 2, and 9) by slow-electron velocity map imaging (SEVI) of C₁₄H₉⁻ generated through molecular beam co-expansion of NF₃ and n-(trimethylsilyl)-anthracene in He carrier gas in the presence of excess electrons.⁴² A second commonly employed method in mass spectrometry is decarboxylation of a deprotonated carboxylic acid anion through collision induced dissociation (CID) in a buffer cell.⁴³ Because CO₂ is very stable and has a negative electron affinity, decarboxylation is often the lowest energy path and leaves the charge at the location of the CO₂ loss. Here, we employ this method, starting with the three isomeric forms of anthracene carboxylic acid (n-C₁₄H₉–CO₂H, where n = 1, 2, or 9 as shown in Fig. 1), and employ frequency- and angle-resolved photoelectron imaging to probe the resonance dynamics of the 9-C₁₄H₉⁻ fragment ion produced through CID.

Fig. 1 Structures of (a) anthracene and (b)–(d) the three isomers of anthracene carboxylic acid. The three locations are numbered according to the labels in (a). Decarboxylation of (b)–(d) would form n-C₁₄H₉⁻, deprotonated at the same location as the departed carboxylic acid.

Methodology

Experimental

A detailed description of the experiment has been presented elsewhere.^44–46 Here, a brief overview is presented of the instrument used and we detail the changes that have been implemented to enable CID prior to photoelectron imaging. A schematic of the new elements of the instrument is shown in Fig. 2. Anions were produced by electrospray ionisation of a 1 mM solution of n-C₁₄H₉–CO₂H (procured from Tokyo Chemical Industry and used without further purification) in methanol. The electrospray plume was drawn into the first of several vacuum regions through a long (20 cm) capillary tube (0.6 mm inner diameter). This first region is connected to the next regions by pin-holes which incrementally reduce the pressure from 1 Torr in the first, to 10⁻² Torr in the second, and 10⁻⁴ Torr in the third differentially pumped region. Each region houses sequential ring-electrode ion guides, consisting of both radially-confining RF fields and longitudinal DC ramps. This effectively forms a tube where potential energy guides the ions along by means of the DC ramps. An additional potential can be applied to encourage the ions to pass through the differential apertures, while at the end of the ion guide in the third region, a potential barrier was used to stop and trap the ions. The arrangement allows for a considerable increase in the transmission from the capillary to the trap as well as a much increased trapping volume.⁴⁶ The trapping voltage was rapidly switched to extract the ions in the trap and these were then injected into a collinear Wiley-McLaren time-of-flight mass spectrometer.⁴⁷ Mass-selected ion packets then entered a continuous mode velocity map imaging spectrometer,^44,48 where the ions were subject to a tuneable ∼6 ns laser pulse from a Nd:YAG pumped OPO (Continuum Surelite II-10, Horizon I) and the resultant photoelectrons were collected on a position sensitive detector.⁴⁹ Raw images were analysed using the Polar Onion Peeling algorithm to determine both the photoelectron spectrum and the angular distribution.⁴⁹ Photoelectron spectra were calibrated using the known photoelectron spectrum of I⁻ and the resolution of the spectrometer was around 5%.

Fig. 2 Schematic of the electrospray ionisation source and vacuum interface. A series of ion guides is used to drive the ions towards a trap located in region 3. The expanded view shows the region where CID is induced and the potential used to control the extent of CID. The pressures in the differentially pumped regions are given in parenthesis and are in units of Torr.

Under normal operating conditions, anions experience only a small DC voltage drop from the last electrode of the first ion guide to the pinhole, and the same on the other side. By increasing the magnitude of this voltage drop into the second ion guide, significant CID can be induced as evidenced by fragments appearing in the time-of-flight mass spectrum. The most prominent fragment was an ion with a m/z 44 less than the parent anion n-C₁₄H₉–CO₂⁻, corresponding to CO₂ loss.

Computational

Density functional theory (DFT) calculations were performed using the Gaussian 09 computational package⁵⁰ and the B3LYP functional with the 6-31+G** basis set, which has been shown to perform well on anionic PAHs.⁴² The geometries of 1-, 2- and 9-C₁₄H₉⁻ and their corresponding radical neutral species were optimized, and were confirmed to represent the geometric minima through vibrational frequency analysis. All of the geometries of n-C₁₄H₉⁻ were confirmed to have planar minimum energy structures, in agreement with the previous literature.^11,42,51 The calculated energetics have been corrected for zero-point energy. Calculations using time-dependent (TD) DFT employing the Tamm–Dancoff approximation⁵² provided energetics and information about the character of both the anionic and neutral excited states accessed experimentally for 9-C₁₄H₉⁻ which has C_2V symmetry.

Additional ab initio calculations were carried out to predict the photoelectron angular distributions using the coupled-cluster equations of motion (EOM) formalism⁵³ and employing the QChem 4.4 package.⁵⁴ Specifically, EOM-IP-CCSD calculations (using the same basis set as before) were used to determine the Dyson orbital with the anion ground state as the initial reference wavefunction, Ψ^N, and the neutral ground state as the final wavefunction, Ψ^N−1.^55,56 The photoelectron angular distribution for this direct detachment channel was modelled using the ezDyson program (version 3.2) developed by Krylov and coworkers.^57,58

Results and discussion

Production of 9-C₁₄H₉⁻ from the CID of 9-C₁₄H₉–CO₂⁻

By altering the voltage drop between regions 1 and 2 in Fig. 2, both the parent anion and a mass-peak at a m/z 44 below the parent anion could be produced. The ratio of the fragment to the parent anion could be increased by increasing the voltage drop, which corresponds to increasing the kinetic energy of the parent anion through a high pressure (∼1 Torr) region of the instrument causing CID. It is well known that the CID of an organic carboxylic acid generally removes the CO₂ group to produce the deprotonated carbanion.^38,39 Under low-pressure conditions and “soft” CID, the negative charge generally remains localised in the location where the CO₂ was removed. However, in the present case, CID occurs in a region of higher pressure than that of a typical CID mass-spectrometric experiment and the timescales in our experiment are also very different (with the time from CID to photoelectron spectroscopy being several milliseconds). The consequence of this “in-source” CID was that we could only produce the lowest energy isomer (9-C₁₄H₉⁻) in isolation, regardless of the starting n-C₁₄H₉–CO₂⁻ isomer. We also found it necessary to flow dry Ar over the inlet of the mass-spectrometer (see Fig. 2) as O₂ and H₂O in the air led to several additional fragments, presumably produced by in-source reactions of CID-produced reactive intermediates. Note that even though 9-C₁₄H₉⁻ is formed vibrationally hot by CID, the anion is eventually thermalized to room temperature within the second and third ion guide regions of the instrument. Under the mildest conditions using 1-C₁₄H₉–CO₂⁻ that resulted in sufficient CID to perform photoelectron spectroscopy, some 1-C₁₄H₉⁻ could be formed, but the ion packet contained significant contamination from the 9-isomer too (see the ESI†).

Although in the present experiment only a single isomer could be formed exclusively, the use of CID in forming decarboxylated fragment anions is appealing because it relies on electrosprayed anions for which there is essentially no size or volatility limit of the precursor anion. For larger PAHs, the vapour pressure rapidly reduces. For example, perylene has a vapour pressure of ∼4 × 10⁻² Pa at 400 K (ref. 59) and needs to be heated to very high temperatures for molecular beam studies. In contrast, electrospray ionisation of the perylene carboxylic acid is readily achieved.

Frequency- and angle-resolved photoelectron spectra of 9-C₁₄H₉⁻

The frequency-resolved photoelectron spectra are shown in Fig. 3 as a false-colour plot. It is composed of 21 photoelectron spectra at 0.1 eV intervals (the individual photoelectron spectra are given in the ESI†). Each spectrum has been normalised so that the highest energy peak has a unit maximum intensity. As the photon energy is increased, the eKE of the highest energy peak increases by the same amount indicating that this is a direct detachment feature. Extrapolation of the high eKE edge yields ADE = 1.7(1) eV, which is in agreement with the value of 1.7155(2) eV determined by Neumark and coworkers.⁴² At around hν ∼ 3.0 eV, a second channel opens up which also leads to a peak that increases linearly with hν. This feature can be assigned to the first excited state of the 9-C₁₄H₉ neutral radical. The energy gap between these values from our measurements is 1.2(1) eV, which is in agreement with the gap of 1.205(6) eV determined by Neumark and coworkers.⁴²

Fig. 3 False-colour plot of the frequency-resolved photoelectron spectra of 9-C₁₄H₉⁻.

In addition to the two direct detachment features, there is evidence of delayed autodetachment at 2.8 < hν < 3.5 eV. Instead of the expected increase of eKE with hν, the photoelectron spectra show signals at a lower eKE, indicating that some of the available eKE has been converted to kinetic energy of the nuclei.⁶⁰ Hence, in the 2.8 < hν < 3.5 eV energy range, a resonance appears to be excited and its nuclear wavepacket and/or internal conversion dynamics are in competition with autodetachment from the resonance.

For hν > 3.8 eV, there also appears to be a dramatic change in the branching ratio of formation of the neutral species in the ground or first excited state following electron detachment.

Specifically, the neutral excited state appears to be produced more favourably compared to the ground state. As discussed below, this arises from the presence of a resonance for which autodetachment to the first excited state may be more favourable.

Additional insight can be gained by considering the photoelectron angular distributions at all hν and eKE values. The angular distribution can be quantified using the anisotropy parameter, β₂, defined from a fit to the photoelectron angular distribution at a given eKE to^61,62

I(θ, eKE) ∝ 1 + ½[β₂P₂(cos θ)],

where P₂(cos θ) = 3 cos² θ − 1 is the second order Legendre polynomial. The two limiting values of β₂ are −1 and +2, which correspond to photoelectron emission predominantly perpendicular and parallel to the polarisation axis of the laser.

In Fig. 4, the frequency-resolved β₂ parameters are plotted as a false colour plot (the raw photoelectron images are in the ESI†). The photoelectron spectra show strong anisotropy in certain spectral regions. The direct detachment feature that forms the neutral ground state exhibits an overall positive β₂. At low hν, the direct detachment has β₂ = +1.3 ± 0.1. This is consistent with the observation of the Neumark group. The photoelectrons arising from delayed autodetachment (lower eKE) in the 2.8 < hν < 3.5 eV range are predominantly isotropic with β₂ ∼ 0. Similarly, detachment to form the first excited state of the neutral species appears to be isotropic.

Fig. 4 False-colour plot of the frequency-resolved β₂ spectra of 9-C₁₄H₉⁻. The signals below an intensity of 0.2 in Fig. 3 have been blacked out because the signal levels were too low to determine a reliable β₂ parameter.

Upon closer inspection, the anisotropy of the direct detachment feature that leaves the neutral radical in its ground state appears to change quite abruptly across the hν range studied. In Fig. 5, the β₂ parameter averaged over this direct detachment peak is shown as a function of hν (and of eKE of the photoelectron). If direct detachment was the only available channel, then one would expect to see slow variations with increasing hν. However, over the regions 2.8 < hν < 3.5 eV and hν > 3.8 eV, there are strong deviations from this expected behaviour, which supports our earlier suggestion that resonances are excited in these regions.

Fig. 5 Plot of the β₂ parameter of the direct detachment feature as a function of hν and eKE. The red points and dashed line indicate the experimentally determined values and the black solid line indicates the calculated β₂ parameter using the Dyson orbital shown. The experimental error shown on the first data point is similar for all the other data points.

Fig. 6 shows the relevant calculated molecular orbitals (MOs) of 9-C₁₄H₉⁻. The highest occupied MO (HOMO) corresponds to the n-orbital localised primarily at the C9 position (see Fig. 1), while the lowest unoccupied MO (LUMO) is a delocalised π* orbital. Our calculations show that there are several excited states and resonances of 9-C₁₄H₉⁻ that can be accessed in the hν range used in the present experiments. Fig. 6 also shows an overview of the most relevant excited states (those with significant oscillator strengths) and their dominant orbital contributions are given in the ESI.† Many of the lowest lying excited states involve ¹nπ* states and consequently have very low oscillator strengths; the lowest lying anion ¹nπ* resonance is included in the energy level diagram.

Fig. 6 Summary of the relevant molecular orbitals (MOs) of 9-C₁₄H₉⁻ and the dominant electron configurations of the relevant electronic states of the 9-C₁₄H₉⁻ anion and the 9-C₁₄H₉ neutral species. On the right are the calculated excited state energy levels; the values in blue correspond to the adiabatic detachment energies from the anion to the neutral X and A states, with the experimental values in parentheses.

Based on our calculations, the most likely resonances that are excited around 2.8 < hν < 3.5 eV and hν > 3.8 eV are the three lowest ¹ππ* transitions. The lowest (1)¹ππ* transition is calculated to be at 3.14 eV. This calculated energy is consistent with the 2.8 < hν < 3.5 eV range over which the deviation of the β₂ parameter was noted in Fig. 5 and with the resonance signatures in the frequency-resolved photoelectron spectra in Fig. 3. For hν > 3.8 eV, deviations from the expected β₂ behaviour were also noted in Fig. 5, which is consistent with the next two ¹ππ* resonances: (2)¹ππ* was calculated at 3.77 eV and (3)¹ππ* at 4.12 eV photon energy. Interestingly, only the (1)¹ππ* resonance shows evidence of nuclear dynamics in competition with autodetachment (Fig. 3).

Fig. 6 additionally shows the orbital configurations associated with the neutral ground state (X²A₁) and its first excited state (A²B₁). The ground state is formed by the loss of the n-electron in the HOMO, while the first excited state is formed by the loss of a π-electron from the HOMO−1. All three ¹ππ* states predominantly involve the excitation of the HOMO−1 π-electron into a π* MO. Hence, these resonances are of Feshbach character with respect to the electron loss channel that forms the neutral ground state, but are of shape character with respect to the electron loss channel that leaves the neutral species in its first excited state. Because a Feshbach resonance involves a 2-electron transition, the autodetachment lifetimes are typically longer compared to shape resonances for which only a single electron transition is required.

The adiabatic energy of the neutral excited state lies very close to the onset of the (1)¹ππ* resonance, while the vertical energy of the neutral species (at the anion geometry) will be higher and was calculated to be at hν = 3.29 eV. Because excitation occurs in the vertical region and the electron loss cross section is small for low eKE electrons, the (1)¹ππ* resonance will predominantly decay by autodetachment to the ground state of the neutral species. But as this is a 2-electron transition, autodetachment may be expected to be relatively slow. Therefore, one might anticipate that nuclear dynamics could compete with autodetachment, which is consistent with the observations in Fig. 3. The (2)¹ππ* and (3)¹ππ* resonances are some way above the threshold region for the autodetachment channel that leaves the neutral species in its excited state. Therefore, loss of the electron can proceed through a one electron process (excited shape resonance) and the consequential increased autodetachment rate appears to outcompete the nuclear dynamics of the (2)¹ππ*/(3)¹ππ* resonances so that the photoelectron peak essentially takes the spectral form of a direct detachment process as seen in Fig. 3. We also previously noted the increase in yield of the neutral excited state when hν > 3.8 eV (see Fig. 3), and this is fully consistent with the fact that the (2)¹ππ* and (3)¹ππ* resonances are only of shape character with respect to the neutral excited state, but Feshbach with respect to the ground state. Hence autodetachment into the former channel is faster and therefore has a higher yield. Note that we cannot determine the branching ratios accurately because of the additional direct detachment channel that is always open for hν > ADE.

One aspect of the direct photodetachment from 9-C₁₄H₉⁻ that is particularly clear is its anisotropy. This has previously been discussed by the Neumark group in their interpretation of their SEVI experiment.⁴² The photoelectron angular distribution for directed detachment in the 0.0 < eKE < 0.3 eV range was measured to be β₂ = +1.3.⁴² This is in agreement with the β₂ parameters determined here in the same energy range (Fig. 5). For higher eKE values, we observe a gradual decrease in β₂ as well as rapid changes. In order to confirm that the rapid changes correspond to the deviation of a smoothly varying β₂ parameter due to direct detachment, we have performed similar calculations to the Neumark group but over a wider eKE range. The results from these calculations are included in Fig. 5 (solid line). The data are not scaled in any way and yield β₂ = +1.2 for low eKE electrons, in excellent agreement with the experiment.‡ As eKE increases, the β₂ parameter smoothly decreases as anticipated. This overall trend is reproduced in the experimental data. However, over the ranges where resonance dynamics were noted (2.8 < hν < 3.5 eV and hν > 3.8 eV), the measured photoelectron angular distributions are significantly more isotropic than that predicted for direct detachment, confirming that the frequency-resolved β₂ parameter has identified the resonances.

The quantitative agreement between the experiment and theory here is very encouraging. Our group has previously shown similar rapid changes in β₂ for para-benzoquinone radical anions,³³ which could also be correlated with the excitation and electron emission from resonances. However, in chemical derivatives of para-benzoquinone, the overall anisotropy tended to zero so that changes were difficult to discern.^20,21 For 9-C₁₄H₉⁻, the clear anisotropy arises from its high (C_2V) symmetry, enabling full use of the angular dimension. Further theoretical developments to accurately predict photoelectron distributions for photo-detachment and specifically autodetachment may assist in identifying the nature of the resonances in favourable cases such as 9-C₁₄H₉⁻.

Resonance dynamics in 9-C₁₄H₉⁻ compared to 1-C₁₈H₁₁⁻

The frequency- and angle-resolved photoelectron spectra clearly show that nuclear resonance dynamics can compete with autodetachment from the (1)¹ππ* resonance of 9-C₁₄H₉⁻. The dynamics of the resonance involves nuclear wavepacket motion on the (1)¹ππ* potential energy surface. Hence, autodetachment may occur along different coordinates of the (1)¹ππ* surface which leads to different Franck–Condon factors to the neutral ground state and consequently to photoelectrons with a lower eKE relative to that for direct detachment. Additionally, internal conversion can occur from the (1)¹ππ* resonance to lower-lying ¹nπ* resonances. Several such resonances were identified in our calculations. The population of the ¹nπ* resonances following a non-adiabatic transition from the (1)¹ππ* state is expected to result in very fast autodetachment as these are of shape character relative to the neutral ground state. Unfortunately, we cannot determine whether such internal conversion processes are occurring from the present data.

Internal conversion to repopulate the ground electronic state of the anion is barely observed: this would be evidenced by the presence of very low energies in the photoelectron spectra that are statistically emitted (thermionic emission) from the hot anion. There is the suggestion of thermionic emission in the photoelectron images (ESI†), but the yield is very low. Hence, unlike the case of the tetracenyl anion, C₁₈H₁₁⁻, for which intense thermionic emission was clearly seen, resonances in 9-C₁₄H₉⁻ are not efficient in recovering the ground state. While 1-C₁₈H₁₁⁻ has a higher density of resonances than 9-C₁₄H₉⁻, a key additional difference between the two systems is that 1-C₁₈H₁₁⁻ has several bound singlet excited electronic states. This is partly due to its higher ADE and partly due to the greater π-electron delocalisation in 1-C₁₈H₁₁⁻. We have previously shown that in the menadione radical anion, efficient anion ground state recovery was facilitated by a bound excited state, which acted as an intermediate in an internal conversion cascade.²⁰ Hence, we suggest that 9-C₁₄H₉⁻ does not efficiently reform the anion ground state in part because of the lack of anion bound states. We can therefore also suggest that 1-C₁₈H₁₁⁻ is the smallest astrophysically relevant polyacenyl anion for which efficient ground state recovery can occur. Note that 1-C₁₄H₉⁻, 2-C₁₄H₉⁻, 2-C₁₈H₁₁⁻ and 3-C₁₈H₁₁⁻ may prove to be more efficient in this regard, but only the lowest energy isomers (9-C₁₄H₉⁻ and 1-C₁₈H₁₁⁻) are expected to be prevalent in the interstellar medium; these can be formed by H-atom tunnelling from the other higher-lying isomers, which proceeds on timescales faster than the astronomically relevant ones.⁶³

The observation of resonances in 9-C₁₄H₉⁻ and their dynamics also has implications for the SEVI work reported by the Neumark group. Specifically, the SEVI experiment assumes that direct detachment into the continuum determines the vibrational structure of the neutral species. However, as the direct detachment cross section is low near the threshold, even optically weak transitions may become important, and we note that our calculations suggest that the lowest ¹nπ* resonance is close to the neutral state energy. If a resonance was accessed at the SEVI detachment energy, then this would lead to changes in the Franck–Condon factors of the neutral species, as was shown by Schiedt and Weinkauf in para-benzoquinone radical anions.⁶⁴ This should be apparent from changes in the peak intensities as the photon energy was changed and perhaps in the β₂ parameters associated with specific photoelectron peaks. In the data presented for the neutral ground state of 9-C₁₄H₉⁻, there is no evidence of the participation of resonance excitation, however, the same is certainly not true for the first excited state of the neutral species. Fig. 3 clearly shows that the (1)¹ππ* resonance produces photoelectrons essentially over the entire eKE < hν − ADE range. Indeed, it is noted in their work that the β₂ parameters are likely distorted because of high-eKE electrons that are arising from direct detachment into the neutral ground state.⁴² From our data, these “distorting” electrons are actually also at low-eKE because of the resonance dynamics of the (1)¹ππ* state.

Conclusions

A new methodology for studying the resonances of closed shell PAH anions has been presented based on the CID of their respective carboxylic acids formed by electrospray ionisation. For different anthracene carboxylic acid isomers, only the most stable isomer of the anthracenyl anion (9-C₁₄H₉⁻) is formed. Frequency- and angle-resolved photoelectron imaging of 9-C₁₄H₉⁻ was used to identify and characterise the dynamics of a number of resonances. The lowest lying π* resonance shows rapid nuclear dynamics that are in competition with autodetachment. However, ground state recovery was observed to be inefficient. Two higher-lying π* resonances were observed to rapidly decay by autodetachment. All observations could be accounted for using supporting electronic structure calculations. In particular, our data and calculations show how photoelectron anisotropy can be used to identify the presence of resonances.

Closed shell PAH anions have been implicated to exist in the interstellar medium, where, in the absence of a sufficiently large dipole moment, electron attachment through resonances may be a key formation mechanism. Our results show that the predominant decay mechanism for resonances in 9-C₁₄H₉⁻ is autodetachment, which is in contrast to the next polyacene increment, the tetracenyl anion, for which ground state recovery has been reported previously.¹⁹ Therefore, the latter is the smallest polyacenyl anion for which this process can occur and therefore is also the smallest likely polyacenyl anion to be found in the interstellar medium. With the development of this new methodology, we will now be able to explore a wide range of large PAHs to determine the role of resonances in their anion formation.

Acknowledgements

We thank Berwyck Poad for many insightful discussions concerning CID and bringing it and its versatility to our attention. This work was supported by the European Research Council (Starting Grant 306536).

Notes and references

1. C. Barckholtz, T. A. Barckholtz and C. M. Hadad, J. Am. Chem. Soc., 1999, 121, 491–500.
2. H. F. Calcote, Combust. Flame, 1981, 42, 215–242.
3. H. Richter, O. A. Mazyar, R. Sumathi, W. H. Green, J. B. Howard and J. W. Bozzelli, J. Phys. Chem. A, 2001, 105, 1561–1573.
4. R. G. Harvey and N. E. Geacintov, Acc. Chem. Res., 1988, 21, 66–73.
5. J. L. Durant, W. F. Busby, A. L. Lafleur, B. W. Penman and C. L. Crespi, Mutat. Res., Genet. Toxicol., 1996, 371, 123–157.
6. T. M. Figueira-Duarte and K. Müllen, Chem. Rev., 2011, 111, 7260–7314.
7. X. Guo, M. Baumgarten and K. Müllen, Prog. Polym. Sci., 2013, 38, 1832–1908.
8. J. L. Puget and A. Léger, Annu. Rev. Astron. Astrophys., 1989, 27, 161–198.
9. L. J. Allamandola, A. G. G. M. Tielens and J. R. Barker, Astrophys. J., Suppl. Ser., 1989, 71, 733–775.
10. A. G. G. M. Tielens, Annu. Rev. Astron. Astrophys., 2008, 46, 289–337.
11. H. A. Galué and J. Oomens, Astrophys. J., 2012, 746, 83.
12. E. K. Campbell, M. Holz, D. Gerlich and J. P. Maier, Nature, 2015, 523, 322–323.
13. S. Brünken, H. Gupta, C. A. Gottlieb, M. C. McCarthy and P. Thaddeus, Astrophys. J., Lett., 2007, 664, L43.
14. M. C. McCarthy, C. A. Gottlieb, H. Gupta and P. Thaddeus, Astrophys. J., Lett., 2006, 652, L141.
15. J. Cernicharo, M. Guélin, M. Agúndez, K. Kawaguchi, M. McCarthy and P. Thaddeus, Astron. Astrophys., 2007, 467, L37–L40.
16. P. J. Sarre, Mon. Not. R. Astron. Soc., 2000, 313, L14–L16.
17. K. D. Jordan and F. Wang, Annu. Rev. Phys. Chem., 2003, 54, 367–396.
18. J. Simons, J. Phys. Chem. A, 2008, 112, 6401–6511.
19. J. N. Bull, C. W. West and J. R. R. Verlet, Phys. Chem. Chem. Phys., 2015, 17, 32464–32471.
20. J. N. Bull, C. W. West and J. R. R. Verlet, Chem. Sci., 2015, 6, 1578–1589.
21. J. N. Bull, C. W. West and J. R. R. Verlet, Phys. Chem. Chem. Phys., 2015, 17, 16125–16135.
22. M. Bixon and J. Jortner, J. Chem. Phys., 1968, 48, 715–726.
23. K. D. Jordan and P. D. Burrow, Chem. Rev., 1987, 87, 557–588.
24. H. S. Taylor, G. V. Nazaroff and A. Golebiewski, J. Chem. Phys., 1966, 45, 2872–2888.
25. R. V. Khatymov, R. F. Tuktarov and M. V. Muftakhov, JETP Lett., 2011, 93, 437–441.
26. M. Larsson, W. D. Geppert and G. Nyman, Rep. Prog. Phys., 2012, 75, 66901.
27. A. G. G. M. Tielens, Rev. Mod. Phys., 2013, 85, 1021–1081.
28. G. Israël, C. Szopa, F. Raulin, M. Cabane, H. B. Niemann, S. K. Atreya, S. J. Bauer, J.-F. Brun, E. Chassefière, P. Coll, E. Condé, D. Coscia, A. Hauchecorne, P. Millian, M.-J. Nguyen, T. Owen, W. Riedler, R. E. Samuelson, J.-M. Siguier, M. Steller, R. Sternberg and C. Vidal-Madjar, Nature, 2005, 438, 796–799.
29. A. J. Coates, F. J. Crary, G. R. Lewis, D. T. Young, J. H. Waite and E. C. Sittler, Geophys. Res. Lett., 2007, 34, L22103.
30. C. Walsh, N. Harada, E. Herbst and T. J. Millar, Astrophys. J., 2009, 700, 752.
31. R. C. Fortenberry, J. Phys. Chem. A, 2015, 119, 9941–9953.
32. J. N. Bull, C. W. West and J. R. R. Verlet, Chem. Sci., 2016, 7, 5352–5361.
33. C. W. West, J. N. Bull, E. Antonkov and J. R. R. Verlet, J. Phys. Chem. A, 2014, 118, 11346–11354.
34. D. A. Horke, Q. Li, L. Blancafort and J. R. R. Verlet, Nat. Chem., 2013, 5, 711–717.
35. C. W. West, A. S. Hudson, S. L. Cobb and J. R. R. Verlet, J. Chem. Phys., 2013, 139, 71104.
36. C. W. West, J. N. Bull, A. S. Hudson, S. L. Cobb and J. R. R. Verlet, J. Phys. Chem. B, 2015, 119, 3982–3987.
37. C. W. West, J. N. Bull and J. R. R. Verlet, J. Phys. Chem. Lett., 2016, 7, 4635–4640.
38. R. R. Squires, Acc. Chem. Res., 1992, 25, 461–467.
39. Z. Tian and S. R. Kass, Chem. Rev., 2013, 113, 6986–7010.
40. C. H. DePuy, V. M. Bierbaum, L. A. Flippin, J. J. Grabowski, G. K. King, R. J. Schmitt and S. A. Sullivan, J. Am. Chem. Soc., 1980, 102, 5012–5015.
41. I. H. Krouse and P. G. Wenthold, J. Am. Soc. Mass Spectrom., 2005, 16, 697–707.
42. M. L. Weichman, J. A. DeVine, D. S. Levine, J. B. Kim and D. M. Neumark, Proc. Natl. Acad. Sci. U. S. A., 2016, 113, 1698–1705.
43. S. W. Froelicher, B. S. Freiser and R. R. Squires, J. Am. Chem. Soc., 1986, 108, 2853–2862.
44. D. A. Horke, G. M. Roberts, J. Lecointre and J. R. R. Verlet, Rev. Sci. Instrum., 2012, 83, 63101.
45. G. Roberts, Doctoral, Durham University, 2010.
46. J. Lecointre, G. M. Roberts, D. A. Horke and J. R. R. Verlet, J. Phys. Chem. A, 2010, 114, 11216–11224.
47. W. C. Wiley and I. H. McLaren, Rev. Sci. Instrum., 1955, 26, 1150–1157.
48. A. T. J. B. Eppink and D. H. Parker, Rev. Sci. Instrum., 1997, 68, 3477–3484.
49. G. M. Roberts, J. L. Nixon, J. Lecointre, E. Wrede and J. R. R. Verlet, Rev. Sci. Instrum., 2009, 80, 53104.
50. M. J. Frisch, G. W. Trucks, J. R. Cheeseman, G. Scalmani, M. Caricato, H. P. Hratchian, X. Li, V. Barone, J. Bloino, G. Zheng, T. Vreven, J. A. Montgomery, G. A. Petersson, G. E. Scuseria, H. B. Schlegel, H. Nakatsuji, A. F. Izmaylov, R. L. Martin, J. L. Sonnenberg, J. E. Peralta, J. J. Heyd, E. Brothers, F. Ogliaro, M. Bearpark, M. A. Robb, B. Mennucci, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, A. Rendell, R. Gomperts, V. G. Zakrzewski, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao and H. Nakai, Gaussian 09.
51. J. Schiedt and R. Weinkauf, Chem. Phys. Lett., 1997, 266, 201–205.
52. S. Hirata and M. Head-Gordon, Chem. Phys. Lett., 1999, 314, 291–299.
53. A. I. Krylov, Annu. Rev. Phys. Chem., 2008, 59, 433–462.
54. Y. Shao, Z. Gan, E. Epifanovsky, A. T. B. Gilbert, M. Wormit, J. Kussmann, A. W. Lange, A. Behn, J. Deng, X. Feng, D. Ghosh, M. Goldey, P. R. Horn, L. D. Jacobson, I. Kaliman, R. Z. Khaliullin, T. Kuś, A. Landau, J. Liu, E. I. Proynov, Y. M. Rhee, R. M. Richard, M. A. Rohrdanz, R. P. Steele, E. J. Sundstrom, H. L. Woodcock III, P. M. Zimmerman, D. Zuev, B. Albrecht, E. Alguire, B. Austin, G. J. O. Beran, Y. A. Bernard, E. Berquist, K. Brandhorst, K. B. Bravaya, S. T. Brown, D. Casanova, C.-M. Chang, Y. Chen, S. H. Chien, K. D. Closser, D. L. Crittenden, M. Diedenhofen, R. A. DiStasio Jr, H. Do, A. D. Dutoi, R. G. Edgar, S. Fatehi, L. Fusti-Molnar, A. Ghysels, A. Golubeva-Zadorozhnaya, J. Gomes, M. W. D. Hanson-Heine, P. H. P. Harbach, A. W. Hauser, E. G. Hohenstein, Z. C. Holden, T.-C. Jagau, H. Ji, B. Kaduk, K. Khistyaev, J. Kim, J. Kim, R. A. King, D. Kosenkov, T. Kowalczyk, C. M. Krauter, K. U. Lao, A. D. Laurent, K. V. Lawler, S. V. Levchenko, C. Y. Lin, F. Liu, E. Livshits, R. C. Lochan, A. Luenser, P. Manohar, S. F. Manzer, S.-P. Mao, N. Mardirossian, A. V. Marenich, S. A. Maurer, N. J. Mayhall, E. Neuscamman, C. M. Oana, R. Olivares-Amaya, D. P. O'Neill, J. A. Parkhill, T. M. Perrine, R. Peverati, A. Prociuk, D. R. Rehn, E. Rosta, N. J. Russ, S. M. Sharada, S. Sharma, D. W. Small, A. Sodt, T. Stein, D. Stück, Y.-C. Su, A. J. W. Thom, T. Tsuchimochi, V. Vanovschi, L. Vogt, O. Vydrov, T. Wang, M. A. Watson, J. Wenzel, A. White, C. F. Williams, J. Yang, S. Yeganeh, S. R. Yost, Z.-Q. You, I. Y. Zhang, X. Zhang, Y. Zhao, B. R. Brooks, G. K. L. Chan, D. M. Chipman, C. J. Cramer, W. A. Goddard III, M. S. Gordon, W. J. Hehre, A. Klamt, H. F. Schaefer III, M. W. Schmidt, C. D. Sherrill, D. G. Truhlar, A. Warshel, X. Xu, A. Aspuru-Guzik, R. Baer, A. T. Bell, N. A. Besley, J.-D. Chai, A. Dreuw, B. D. Dunietz, T. R. Furlani, S. R. Gwaltney, C.-P. Hsu, Y. Jung, J. Kong, D. S. Lambrecht, W. Liang, C. Ochsenfeld, V. A. Rassolov, L. V. Slipchenko, J. E. Subotnik, T. V. Voorhis, J. M. Herbert, A. I. Krylov, P. M. W. Gill and M. Head-Gordon, Mol. Phys., 2015, 113, 184–215.
55. C. M. Oana and A. I. Krylov, J. Chem. Phys., 2007, 127, 234106.
56. Y. Liu and C. Ning, J. Chem. Phys., 2015, 143, 144310.
57. C. M. Oana and A. I. Krylov, J. Chem. Phys., 2009, 131, 124114.
58. A. I. Krylov and S. Gozem, ezDyson.
59. J. L. Goldfarb and E. M. Suuberg, J. Chem. Eng. Data, 2008, 53, 670–676.
60. C. S. Anstöter, J. N. Bull and J. R. R. Verlet, Int. Rev. Phys. Chem., 2016, 35, 509–538.
61. R. N. Zare, Mol. Photochem., 1972, 4, 1.
62. K. L. Reid, Annu. Rev. Phys. Chem., 2003, 54, 397–424.
63. F. Trixler, Curr. Org. Chem., 2013, 17, 1758–1770.
64. J. Schiedt and R. Weinkauf, J. Chem. Phys., 1999, 110, 304–314.

---

Footnotes

† Electronic supplementary information (ESI) available: Raw photoelectron spectra and images; fragment photoelectron spectra from 9- and 1-C₁₄H₉–CO₂⁻ precursors; and excited state molecular orbital contributions. See DOI: 10.1039/c6sc05405f

‡ We believe that this specific β₂ parameter calculated by Neumark and coworkers may be erroneous because of the older version of QChem and/or ezDyson used in their calculations. In fact, visual inspection of the Dyson orbital shows that, while a large component of the orbital is of n-character which would lead to β₂ = +2, there is also significant π-character which should reduce this substantially, which is consistent with the predicted and observed β₂ parameter presented here.

---