Dissociative Ionization Mechanism and Appearance Energies in

Adipic acid, a model compound for oxygenated organic aerosol, has been studied at the VUV beamline of the Swiss Light Source. Internal energy selected...
1 downloads 13 Views 2MB Size
This is an open access article published under an ACS AuthorChoice License, which permits copying and redistribution of the article or any adaptations for non-commercial purposes.

Article pubs.acs.org/JPCA

Dissociative Ionization Mechanism and Appearance Energies in Adipic Acid Revealed by Imaging Photoelectron Photoion Coincidence, Selective Deuteration, and Calculations Maarten F. Heringa,†,‡ Jay G. Slowik,† André S. H. Prévôt,† Urs Baltensperger,† Patrick Hemberger,‡ and Andras Bodi*,‡ †

Laboratory of Atmospheric Chemistry, Paul Scherrer Institute, Villigen PSI, Switzerland Laboratory for Synchrotron Radiation and Femtochemistry, Paul Scherrer Institute, Villigen PSI, Switzerland



ABSTRACT: Adipic acid, a model compound for oxygenated organic aerosol, has been studied at the VUV beamline of the Swiss Light Source. Internal energy selected cations were prepared by threshold photoionization using vacuum ultraviolet synchrotron radiation and imaging photoelectron photoion coincidence spectroscopy (iPEPICO). The threshold photoelectron spectrum yields a vertical ionization energy (IE) of 10.5 eV, significantly above the calculated adiabatic IE of 8.6 eV. The cationic minimum is accessible after vertical ionization by H-transfer from one of the γ-carbons to a carbonyl oxygen and is sufficiently energetic to decay by water loss at the ionization onset. The slope of the breakdown curves, quantum chemical calculations, and selective deuteration of the carboxylic hydrogens establish the dissociative photoionization mechanism. After ionization, one γ-methylene hydrogen and the two carboxylic hydrogens are randomized prior to H2O loss. On the basis of the deuteration degree in the H2O + CO-loss product at higher energies, a direct water-loss channel without complete randomization also exists. The breakdown diagram and center of gravity of the H2O + CO-loss peak were modeled to obtain 0 K appearance energies of 10.77, 10.32, and 11.53 eV for H2O + CO loss, CH2COOH loss, and H2O + CH2COOH loss from adipic acid. These agree well with the CBS-QB3 calculated values of 10.68, 10.45, and 11.57 eV, respectively, which shows that threshold photoionization can yield energetics data as long as the dissociation is statistical, even when the parent ion cannot be observed. The results can be used as a starting point for a deeper understanding of the ionization and low-energy fragmentation of organic aerosol components. statistical approaches, such as principal component analysis,7 positive matrix factorization (PMF),8 and the multilinear engine (ME-2),9 or on marker fragments of specific sources.10 Spectral complexity can be mediated using soft and tunable VUV photoionization to suppress dissociative ionization processes. Although the use of synchrotron VUV sources makes the experiment immobile, studies at the Advanced Light Source have amply demonstrated the benefits of soft photoionization.11−13 Although even photoionization AMS spectra of ambient aerosols include contributions from thousands of compounds, investigation of the ionization and low-energy fragmentation of organic aerosol components is central to understanding the fragmentation pattern of different compound classes, and thus an important step toward improved source identification. Adipic acid (1,6-hexanedioic acid) is a C6 diacid chosen here as a surrogate for multifunctional organics found in SOA. Detailed studies of the energy-dependent fragmentation

1. INTRODUCTION Atmospheric organic aerosol (OA) is a major contributor to existing uncertainties regarding aerosol effects on climate and human health. This is in large part due to the complexity of secondary organic aerosol (SOA), which is formed by atmospheric reactions of directly emitted precursor gases and dominates the global OA.1,2 Although only 10−30% of SOA can typically be chemically speciated,3 multifunctional organics, especially polyacids, are thought to be a major component. In the past decade, online aerosol mass spectrometry (AMS) has become a well-established technique for time-resolved analysis of the concentration and composition of ambient aerosol4 and the source apportionment of its organic fraction.5,6 To ensure quantitative detection of most atmospherically relevant organic and inorganic species (excluding black carbon and mineral dust), the instrument utilizes vaporization at about 600 °C followed by electron ionization (EI) at 70 eV. This yields mass spectra that are significantly affected by both thermal decomposition and ionic fragmentation processes, making the molecular assignment of environmentally relevant mixtures intractable. Data analysis techniques therefore rely heavily on © XXXX American Chemical Society

Received: January 27, 2016 Revised: April 15, 2016

A

DOI: 10.1021/acs.jpca.6b00908 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

onset, as in diethyl ether,32−34 or unfavorable Franck−Condon factors at the adiabatic ionization energy meant that the first photoelectron band appeared only at higher energies, as in ethanol.35,36 Both phenomena can lead to IE error bars that underestimate the uncertainty significantly. In addition to unveiling and quantitatively modeling the unimolecular dissociative ionization processes of adipic acid, our secondary objective is to record its photoelectron spectrum for the first time.

channels of dicarboxylic acids in mass spectrometers are technically challenging and quite rare, as illustrated by the most recent study of adipic acid, conducted in 1970.14 Most importantly, the parent ion peak is missing from the mass spectra of all α,ω-dicarboxylic acids with the possible exception of malonic acid, HOOC(CH2)COOH. Appearance energies measured by photoionization mass spectrometry15,16 and photoelectron photoion coincidence (PEPICO)17−19 can be used in thermochemical cycles to derive, e.g., enthalpies of formation or proton affinities. If, however, the sample molecule dissociatively photoionizes at the ionization threshold and no parent ion is observed in the mass spectrum, appearance energies can normally be derived only much less accurately on the basis of the kinetic energy imparted on the fragment ion.20−22 Two recent developments make the PEPICO study of adipic acid timely. First, iPEPICO experiments23 at the VUV beamline of the Swiss Light Source have shown that recording threshold photoionization mass spectra as a function of photon energy and plotting the fractional ion abundances in the breakdown diagram can directly unveil complex dissociative photoionization mechanisms involving isomerization steps.24,25 The photoion mass-selected threshold photoelectron spectra, msTPES, of pyrolysis products have also proved fruitful in identifying reactive intermediates, and thus, unimolecular thermal decomposition channels.26,27 Therefore, we expect to be able to disentangle the ionic fragmentation channels and to map them as a function of the parent ion internal energy. Second, up to about a decade ago, experimental results on molecules with more than a few heavy atoms could rarely be supported by accurate quantum chemical calculations. This has now changed, and even the adipic acid cation can be subjected to density functional theory (DFT) reaction path and, at the stationary points, reliable composite method calculations.28 As mentioned above, the last comprehensive overview of the mass spectrometry of dicarboxylic acids dates back to the 1970s.14,29−31 Holmes and Benoit discussed in great detail how steric constraints affect carboxyl−carboxyl and carboxyl− hydrocarbon chain interactions in the cation. They also employed selective deuteration to pinpoint the active carbon centers in the molecular ion. In the absence of γ-hydrogens, e.g., in malonic or oxalic acid, proton transfer from one carboxylic group to the other precedes carbon dioxide elimination and the CO2-loss peak is dominant in the electron ionization mass spectrum. When the hydrocarbon chain is at least two carbon atoms long, a H from the γ-carbon migrates to the carboxylic group, leading to water loss and CO loss in the first and second steps, respectively, with the corresponding H2O- and (H2O + CO)-loss peaks dominating the high-mass region of the mass spectrum. Having a γ-carbon, adipic acid falls into the latter case. Both of these decomposition mechanisms begin with an Htransfer isomerization step, i.e., a significant geometry change. If the isomerization takes place without an overall barrier, as the absence of the parent ion peaks in the mass spectra suggest, the nuclear wave function overlap integral, in other words, the Franck−Condon factor, for adiabatic ionization to the cationic minimum will be vanishingly small, and the first peak in the photoelectron spectrum will correspond to vertical ionization at a significantly higher photon energy. Adiabatic ionization energies (IE) are routinely reported on the basis of photoelectron spectra, and there have been examples in which either hot bands were mistakenly used to extrapolate to the ionization

2. EXPERIMENTAL AND COMPUTATIONAL METHODS Adipic acid was obtained from Sigma−Aldrich (purity ≥99.5%) and used as received. To gain more insight into the internal energy dependent fragmentation pathways, doubly deuterated adipic acid, DOOC(CH2)4COOD, was prepared by dissolving 250 mg of adipic acid in 10 mL of D2O (99.8%, Armar AG, Switzerland). The solution was continuously stirred at 40 °C using a rotary evaporator for 4 h, dried at 45 mbar, and stored in a desiccator until use. Vacuum ultraviolet (VUV) photoionization experiments were performed using the imaging photoelectron photoion coincidence (iPEPICO) end station at the VUV beamline of the Swiss Light Source. A detailed description of the beamline can be found elsewhere37 and is only briefly reviewed here. Synchrotron radiation from a bending magnet, dispersed by a grazing incidence monochromator using a 600 lines/mm laminar grating, is focused into a differentially pumped gas filter, operated with a mixture of Ar, Kr, and Ne to suppress higher-order radiation of the grating in the 7−14 eV energy range. The high harmonic free, monochromatic VUV radiation then enters the ionization chamber of the iPEPICO end station and ionizes the sample in a 4 × 2 mm2 interaction region. 2.1. iPEPICO Spectrometer. Adipic acid and adipic acid-d2 were introduced into the ionization chamber of the iPEPICO23 experiment effusively under high-vacuum conditions. The vapor pressure of adipic acid is almost 2 mbar at 170 °C,38 which means that it can easily be vaporized in a resistively heated oven. The sample was put in a 6 mm outer diameter tube, sealed using glass wool, put in a copper oven, and connected to the chamber by an Ultra-Torr fitting. Temperatures were kept at around 80 °C, well below the melting point to ensure thermal stability and prevent leaks.39 In the interaction region, the VUV beam crosses the effusive sample beam, and a constant, 120 V cm−1, electric field accelerates the resulting photoelectrons and photoions in opposite directions. The electrons are velocity map imaged onto a DLD40 Roentdek position-sensitive delay-line detector with a resolution of ca. 1 meV at threshold. As their time-of-flight (TOF) is negligible, electrons also act as start signal for the ion TOF analysis. The cations are accelerated downward in a linear TOF mass spectrometer and space focused on a microchannel plate detector. The relatively long and low first acceleration field leads to high ion residence times in the acceleration region and enables unimolecular dissociation rate constant measurements of the cations in the 103−107 s−1 range. Electron arrival times and positions, and ion arrival times are recorded using a triggerless multiple-start/multiple-stop data acquisition scheme.40 Threshold electrons with close to zero kinetic energy are imaged into a central spot on the detector. Kinetic energy electrons with negligible off-axis momentum make up the hot electron contamination that can be subtracted on the basis of a small ring around the center as proposed by Sztáray and Baer.41 B

DOI: 10.1021/acs.jpca.6b00908 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

distributions when the rate constant is below 107 s−1, and relative rates are obtained as branching ratios when two fast dissociation channels compete. In such cases, the transitional frequencies are scaled by a single factor in the transition state number of states calculation to fit the RRKM rate curves to the experimental results. The fitted rate curves are then extrapolated to the threshold, which corresponds to the 0 K appearance energy when referenced to the neutral.

In the threshold photoelectron spectrum (TPES), the threshold photoelectron signal is plotted as a function of the photon energy. The two-dimensional threshold photoionization mass spectra, i.e., the ion signal as a function of TOF and photon energy, can be reduced in two ways. First, fractional fragment ion abundances can be plotted in the breakdown diagram, which shows the open fragmentation channels and their branching ratios as a function of photon energy. Second, as mentioned above, the use of a relatively low and long extraction field leads to long ion residence times on the order of microseconds in the acceleration region and, in the case of slow dissociations, yields a quasi-exponential daughter ion peak shape, indicative of the dissociation rate.42 If the count rates are too low for peak profile analysis, the center of gravity of the daughter ion peak can be employed to the same effect,43 as will be done herein. We have scanned over the adipic acid TPES twice to rule out the possibility that changing experimental parameters distort the band. Furthermore, the photon flux is essentially constant within the TPES energy range. We note in passing that the breakdown diagram and the measured rates are independent of photon flux, collection efficiency and sample pressure. 2.2. Ab Initio Calculations. Constrained geometry optimizations and STQN44 calculations were carried out to locate stationary points, i.e., minima and transition states, on the potential energy surface of the adipic acid cation at the B3LYP/6-311++G(d,p) level of theory using Gaussian 09.45 The energies of the neutral adipic acid, the stationary points on the cationic potential energy surface as well as of the neutral fragments were then evaluated using the CBS-QB3 composite method.46 Overall, we expect CBS-QB3 results to be accurate to within 50 meV (4.8 kJ mol−1). The adiabatic ionization energies were also calculated using the G3, G3B3, and G4 composite methods.47 The ground and first five excited electronic states of the cation have been evaluated at the CBS-QB3 geometry by the EOM-IP-CCSD method with the cc-pVTZ basis set using Q-Chem 4.0.1.48 2.3. Statistical Model. Breakdown curves and experimental rate constants were modeled as described by Sztáray et al.49 The unimolecular dissociation rates are calculated according to the RRKM approach as k(E) =

σN ‡(E − E0) , hρ(E)

3. RESULTS AND DISCUSSION 3.1. Threshold Photoelectron Spectrum. The threshold photoelectron spectrum of adipic acid, shown in Figure 1, was

Figure 1. Threshold photoelectron spectrum of adipic acid based on the total threshold photoelectron yield as a function of photon energy. The EOM-IP-CCSD calculated vertical ionization energies are also shown.

recorded using 30 meV step size at an integration time of 4 min per data point. To correct for the short-time changes in sample number density and flux, a calibration point was taken at 12 eV for 30 s after every five data points, and the signal was found to decay linearly by 11% over the scan, which has been compensated for. The spectrum exhibits two bands between 10 and 12.5 eV. In fact, EOM-IP-CCSD calculations at the adipic acid minimum of C2h symmetry suggest that these bands are quite crowded with vertical ionization energies to the X̃ + 2Ag, Ã + 2Bu, B̃ + 2Bg, C̃ + 2Au, D̃ + 2Ag, Ẽ + 2Ag states of 10.67, 10.71, 11.97, 12.03, 12.40, and 12.50 eV, respectively. Therefore, at least some of the spectral structure observed in Figure 1 corresponds to ionization to different electronic states. Assuming that the first maximum in the TPES corresponds to vertical ionization to the ground X̃ + state, the vertical IE is determined to be 10.5 eV experimentally. When optimized in the ionic state, neutral rotamer geometries relax to 9.5−10.6 eV without H-transfer. According to our calculations, structure [1] is the global minimum (Figure 2), which was found by transferring one of the γ-C hydrogens to the carbonyl oxygen. The adiabatic ionization energy is calculated to be 8.64, 8.65, 8.61, 8.50, and 8.66 eV using the CBS-QB3, G3, G3B3, G4, and EOM-IP-CCSD/cc-pVTZ methods, respectively. Reaction path calculations indicate one or more curve crossings along the H-transfer coordinate, with an overall barrier smaller than 1.7 eV above [1], i.e., below a photon energy of 10.3 eV, suggesting that the global minimum is accessible at the ionization onset. Therefore, if Franck− Condon factors were favorable, the intact adipic acid parent ion could probably be observed close below 10 eV. There is a gigantic, almost 2 eV (193 kJ mol−1) gap between the vertical and adiabatic ionization energies in adipic acid, in which a thermodynamically and kinetically allowed fragmenta-

where σ is the

symmetry number of the reaction and h is Planck’s constant. The density of states of the reactant ion, ρ(E), and the number of states of the transition state, N‡(E − E0), are obtained on the basis of ab initio harmonic vibrational frequencies and rotational constants. The density of states of the neutral molecule is used to calculate the internal energy distribution of the sample at the experimental temperature, whereas the densities of states of the intermediate ions are used together with those of the leaving neutrals to partition the excess energy and obtain the internal energy distribution of the intermediate fragment ions. The dissociation barrier is the dissociation energy in the absence of a saddle point along the reaction coordinate and corresponds to the activation energy otherwise. In fast dissociations without a competing parallel channel, all ions with sufficient internal energy dissociate to form the fragment ion. Thus, the breakdown curve corresponds to the cumulative distribution function of the internal energy distribution of the dissociating ion, and the barrier is the only variable parameter of the model. It is fitted so that the measured breakdown diagram is best reproduced. Absolute experimental dissociation rates are measured in the TOF C

DOI: 10.1021/acs.jpca.6b00908 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

Figure 2. Fragmentation pathways and potential energy surface of the adipic acid ion at the CBS-QB3 level of theory. The energies are given in electronvolts and are relative to the neutral adipic acid. Dotted lines show minor reaction channels.

tion channel opens up. This explains the absence of the parent ion in the mass spectrum, and presumably, the dissociative photoionization of analogous dicarboxylic acids at the ionization onset, as well. It is not that the adipic acid cation is unbound or barely bound, as is the case in other, prototypically dissociatively photoionizing molecules, e.g., CCl4,50 but is the combination of two distinctly different effects. First, vertical ionization leads to highly energetic parent ions, and second, these ions can relax to the global minimum at the ionization onset, meaning that all of their internal energy can be used in fragmentation, as well. The 1.5 eV difference between the adiabatic ionization energy and the onset of the TPES signal also underscores the proper approach when ionization energies are determined on the basis of photoelectron spectra. In this case, a simple extrapolation to the ionization onset would yield 10.1 ± 0.1 eV as adiabatic IE, in error by 1.5 eV. If the vibrational structure of a band cannot be resolved, photoelectron spectra are only suitable to determine ionization energies when supported by ab initio calculations showing negligible geometry change upon ionization and a large Franck−Condon factor for the 0−0 transition. 3.2. Dissociative Photoionization. In the following sections, we will show that the low-lying fragmentation channels of the adipic acid cation can be identified on the basis of the breakdown diagram, isotopologue ratios in the threshold photoionization mass spectra of adipic acid-d2, and quantum chemical calculations. Illustrative threshold photoionization mass spectra are shown in Figure 3 at five selected photon energies. The ms-TPES belonging to mass channels m/z = 128, 100, 87, and 69, corresponding to the dissociative photoionization of adipic acid, was recorded in the 10−13 eV photon energy range with a 30 meV step size. The fractional ion abundances are plotted in the breakdown diagram in Figure 4. The adipic acid parent ion was not observed at m/z = 146. The first fragmentation channel is the loss of a water molecule, yielding the m/z = 128 peak, which dominates the threshold ionization mass spectrum from 10 to 10.5 eV. In the 10.5−11 eV photon energy range,

Figure 3. Illustrative threshold photoionization mass spectra of adipic acid recorded at five photon energies.

Figure 4. Modeling of the breakdown curves of adipic acid. Markers show the experimental fractional ion abundances, whereas the solid lines show the modeled results. The fitted 0 K appearance energy are E0([8],C4H7O2) = 10.77 eV, E0([5],C5H8O2) = 10.32 eV, and E0([10],C4H5O) = 11.53 eV.

dissociative photoionization channels open up to m/z = 100 and 87. The former corresponds to sequential CO loss from m/ z = 128. CHCO loss from the water-loss intermediate is only possible above 13 eV; thus, the m/z = 87 channel must be CH2COOH loss from the parent adipic acid ion, i.e., a parallel process to the first, water-loss channel. The highest lying channel in the studied energy range, leading to m/z = 69, opens D

DOI: 10.1021/acs.jpca.6b00908 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A up at a photon energy of ca. 11.5 eV. The rising m/z = 69 fractional abundance coincides with a decrease in the m/z = 100 channel. Consequently, it is associated with this channel and is either a loss of 31 amu, e.g., H2COH, from m/z = 100, or a loss of 59 amu, CH2COOH, from its precursor, the m/z = 128 ion. In addition to the computed energetics (H2COH loss only opens up at 12.6 eV), the appearance of the breakdown diagram also speaks for the latter. As mentioned earlier, and discussed in detail elsewhere,51 the slope of the breakdown curve of a fragment ion corresponds to the internal energy distribution of its parent. Therefore, if the m/z = 69 were a dissociation product of m/z = 100, the slope of the m/z = 69 breakdown curve at 11.4 eV, corresponding to the internal energy distribution of m/z = 100 should only be moderately lower than that of m/z = 100 at 10.2 eV, corresponding to the internal energy distribution of m/z = 128. This is because only a small part of the excess energy is lost as kinetic energy and the internal energy of the leaving CO when making m/z = 100. Instead, the m/z = 100 signal reaches 50% slightly above 10.6 eV, i.e., 0.4 eV of its appearance, as opposed to only 20% for the m/z = 69 breakdown curve at 11.9 eV. This suggests strongly that the breakdown diagram is indicative of slowly changing branching ratios, and the m/z = 69 fragment ion is a dissociation product of m/z = 128 by CH2COOH loss, formed in parallel with m/z = 100. Besides the main fragments described above, there is a small peak at m/z = 110 (−2H2O) and several additional, mainly low-mass fragments appear at photon energies above 12 eV, corresponding to a set of the EI-MS peaks.14 The number of fragment peaks rising simultaneously in this energy range implies several reaction channels that open at very similar energies. As they quickly become intractable, and their individual abundance does not rise above 10% in the energy range up to 12.5 eV, we neglect them in the discussion below. As was shown by Holmes and St. Jean for shorter chain dicarboxylic acids,14 selective deuteration can be of aid in identifying the reaction pathways. For adipic acid, in addition to the mass spectrum of HOOC(CH2)4COOH, they also reported those of DOOC(CH2)4COOD and HOOCCD2(CH2) 2CD2COOH. Because peak intensities changed in unexpected ways, they could only conclude that the carboxyl−carboxyl interaction is negligible, as also evidenced by the missing CO2-loss peak, and that β-carbon hydrogens do not participate in water loss. We have prepared and studied the threshold ionization of adipic acid-1,6-d2, DOOC(CH2)4COOD. As expected, the breakdown diagram of the deuterated and light sample are almost indistinguishable when the mass shifts due to deuteration are taken into account. However, we can plot the ratio of differently deuterated mass spectral peaks as a function of the photon energy (Figure 5), which tells us the set of hydrogen atoms involved in a fragmentation and also if their randomization dynamics change as a function of energy. In adipic acid-d2, the ratio of the HOD- to D2O-loss product at m/ z = 129 and 128, respectively, is approximately constant at 1.94, similarly to what was observed in EI-MS.14 This means that in the energy range of the water-loss fragment as final product, the water hydrogens are randomly chosen from a set of three atoms, two deuterium and one hydrogen. In light of the previously established role of the γ-methylene groups, this means that one of the four γ-C hydrogen atoms migrates to a carbonyl oxygen in an irreversible step and is randomized with the two deuteriums at the carboxylic groups before the leaving

Figure 5. All-electron mass spectrum of adipic acid-d2 (M = 148 g/ mol) at 12.0 eV (top) and the isotopic ratios of the three main fragments in threshold photoionization (bottom). The horizontal lines represent the average isotopic ratio of 101/100 (2.68), 129/128 (1.94), and 89/88 (0.53).

water molecule is formed (Scheme 1). Considering the CH2COOH-loss product, we can see that one deuterium is removed 2/3 of the time by CH2COOD and no deuterium leaves 1/3 of the time by CH2COOH loss. This is in perfect agreement with the DDH randomization mechanism in water loss if the two D and the one, originally γ-C H atoms are mixed statistically, and the identity of the COOH hydrogen atom is determined by chance. The ratio of the 101 vs 100 peaks, expected to be 2 if DDH randomization were complete, as observed in the water-loss channel, is obscured by the slow fragmentation rates and the resulting daughter ion peak shape broadening in the low-energy region, up to 11 eV. However, at 2.7, it is still >2 in the fast dissociation regime at higher photon energies, similarly to the EI-MS.14 This ratio cannot be explained by complete randomization of a set of hydrogen atoms and indicates a secondary mechanism. A likely explanation is that at large excess energies, H-transfer from the carbon chain to the carboxyl group, perhaps directly to a hydroxyl oxygen, is instantaneous and water loss may follow without randomization; in such cases, a light hydrogen is lost deterministically. As discussed above, the fragment at m/z = 69 is formed dominantly from the sequential loss of CH2COOH from m/z = 128. As deuteration does not affect the peak position, both carboxylic hydrogens must leave. 3.3. Potential Energy Surface. The statistical model of the dissociative photoionization processes of adipic acid requires the vibrational frequencies of the reactants and transition states. These can be obtained by identifying the stationary points: transition states and intermediate minima on the adipic acid cation potential energy surface, which lead to the observed fragments. Bond length, bond angle, and dihedral angle scans as well as reaction path calculations were carried out in the light of the experimental results to identify (1) water loss, in which the two carboxylic hydrogens are randomized together with one γ-methylene hydrogen, (2) a sequential CO loss from the water-loss fragment, (3) CH2COOH loss parallel to (1) and yielding m/z = 87 in light adipic acid, during which one H or D leaves from the same set of randomized atoms as in water loss in adipic acid-d2, as well as (4) sequential CH2COOH loss following water loss and (5) the expectedly minor (vide inf ra) sequential water loss following CH2COOH loss from the parent ion. E

DOI: 10.1021/acs.jpca.6b00908 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

Scheme 1. (a) Low-Lying Dissociative Photoionization Channels of Adipic Acid and (b) H-Atom Randomization after Irreversible Transfer of One γ-C Hydrogen to a Carboxylic Groupa

[2] can be formed by CH2−COOH rotation in either [1]′ or [1]″, after which HDO or D2O are lost, respectively. The marked H atoms have been exchanged for D in d2-adipic acid. a

parent losing CH2COOH; i.e., H2O + CH2COOH loss is more favorable energetically than CH2COOH + H2O loss. Second, at least some H atom randomization could be expected in [9]‡ by CH3 internal rotation, leading to an m/z = 70 fragment ion in adipic acid-d2, which is absent in the mass spectra. Therefore, the contribution of the CH2COOH + H2O-loss channel to forming m/z = 69 is expected to be small, in agreement with the breakdown diagram analysis. The water-loss product [3] can pass over transition state [11]‡ at an energy of only 10.96 eV and form the intermediate [12] before CH2COOH loss yields the m/z = 69 product [10] at 11.57 eV, without an overall reverse barrier to the process. No matter whether a light hydrogen atom or a deuterium is found in the COOH group of [3], it will leave in this process, meaning that the product must be of m/z = 69 also in adipic acid-d2. Holmes and St. Jean14 were perplexed that this peak shifts to m/z = 71 if the α-hydrogens are exchanged for deuterium, which is in fact easily explained by the mechanism proposed here: the four α-hydrogens are inert in the dissociation, and two of them stay in the daughter ion. 3.4. Unimolecular Dissociation Model. The dissociative photoionization mechanism of adipic acid can now be put to the test by constructing a statistical model and determining experimental thresholds to the processes proposed in the previous sections. RRKM dissociation rate curves were calculated,49 and 10 additional (nonphysical) low-frequency vibrational modes were added to the H 2 O-loss and CH2COOH-loss transition states to increase the flexibility of the statistical model by increasing the number of states so that the observed rate curves could be reproduced.49,52,53 The breakdown diagram for the first step, the dissociation of the parent ion, remains hidden, because it takes place below the Franck−Condon envelope of the cation ground state. As the statistical redistribution of the excess energy affects the breakdown diagram, we have included it in the model with the computed onset energy of E0 = 10.07 eV. The sequential CO loss, yielding m/z = 100, [5], is the main channel opening up at 10.4 eV and complete at around 11.0 eV. The daughter

The structure of the cation minimum, [1] in Figure 2, readily implies an H-randomization mechanism. The bridging H can be transferred between the oxygens over a negligible barrier, and even linear conformers are accessible at 1 eV above [1], i.e., below the ionization onset. In this conformational space, the two carboxyl H atoms and the transferred H lose their original locational identity and are randomized (Scheme 1b). Within this conformational space, in a somewhat higher-lying conformer, [2], a leaving water molecule is all but formed. From here, H2O can leave along a purely attractive potential curve, i.e., without a reverse barrier and form the water-loss product [3] at a CBS-QB3-calculated energy of 10.07 eV, at about the onset of the photoionization signal in the TPES. Carbon monoxide can leave from [3] by C−C bond rupture. The product at 10.30 eV is the five-membered ring structure [5], but ring formation involves a transition state with a reverse barrier at 10.68 eV and structure [4]‡. The tight ring-formation transition state explains why CO loss is a slow process at threshold. In competition with water loss, adipic acid cations may also lose CH2COOH. C−C bond breaking, concurrent with H atom back-transfer to the methylene chain, could yield the fivemembered ring product [14], analogous to [5], over the transition state [13]‡ at 10.78 eV. The bond breaking coordinate in a different rotamer, with the H on the leaving carboxylic group pointing outward, involves no H atom backtransfer in the transition state [6]‡ at a lower energy of 10.45 eV and is more likely to compete with water loss. Afterward, the CH2COOH-loss daughter ion [8] is formed after passing an intermediate minimum [7] at 9.80 eV. This mechanism leads to one of the equivalent HDD atoms leaving in adipic acid-d2, which explains the observed constant 0.5 ratio between m/z = 89 and 88 in Figure 5. Because water is a good leaving group, H-transfer could be assumed to take place in [8] or, for that matter, in [14], and [10] could be formed by releasing water over the transition state [9]‡ at 11.75 eV. However, two arguments speak against this path. First, there is a lower-lying channel, which involves the water-loss daughter ion from the F

DOI: 10.1021/acs.jpca.6b00908 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

4. CONCLUSIONS Similar to most dicarboxylic acids, adipic acid does not form the intact cation but dissociates by water loss at the ionization onset. On the basis of the calculated adiabatic ionization energy of 8.6 eV and the measured vertical ionization energy of 10.5 eV, we argue that the reason is the thermodynamically and kinetically allowed water-loss channel opening up in the 1.5 eV energy window above the adiabatic ionization energy, in which vanishingly small Franck−Condon factors prohibit ionization. Parent ions formed in the Franck−Condon envelope can access the global minimum [1] and are energetic enough to dissociate afterward. Three hydrogen atoms, namely, the two carboxylic hydrogens and one of the four γ-methylene hydrogens become equivalent following an irreversible H-transfer step, which explains the effect of deuteration and the observed mass ratios in the H2O- and CH2COOH-loss peaks, [3] and [8], respectively. Two leading themes have been proposed in the literature for the fragmentation mechanism of dicarboxylic acid cations: neighboring COOH groups can interact, which leads to CO2 loss, or the COOH group interacts with the γ-C to transfer a hydrogen, characterized by H2O loss in the initial step. Our results suggest that there is an intermediate mechanism, in which, although the two carboxylic groups interact with each other, this is mediated by one of the γmethylene hydrogens, and which determines the fragmentation mechanism of the adipic acid cation. However, the randomization of the γ-methylene hydrogen and the carboxylic deuterium atoms is not complete at higher energies, which suggests that a direct water-loss channel may also open up. The calculated potential energy surface, as constructed on the basis of the overall appearance of the breakdown diagram and deuteration effects, was validated by a statistical model fitted to the experimental data. The agreement with the experimental results and the calculated onsets is generally good, with threshold energies to H2O + CO loss, CH2COOH loss, and H2O + CH2COOH loss obtained as 10.77, 10.32, and 11.53 eV experimentally and calculated to be 10.68, 10.45, and 11.57 eV, respectively. The results not only show that recording the breakdown diagram can help unveil the dissociative photoionization processes, particularly when supported by ab initio calculations, but also illustrate that the data analysis may yield reliable onset energies in statistical dissociations even when the parent ion cannot be observed. This opens up the use of PEPICO to study the energetics of numerous dissociatively photoionizing oxygenated organic compounds.

ion peak shape is asymmetrically broadened toward higher times of flight (see mass spectrum at hν = 10.9 eV in Figure 3), which indicates a metastable dissociation that takes place while the photoion is accelerated in the mass spectrometer. The breakdown diagram is determined by three factors: (1) The internal energy distribution of the sample is shifted to the ion manifold upon threshold photoionization and is also available for fragmentation reactions. (2) Some of the excess energy distribution in the first, water-loss step is released as kinetic energy or is converted to the internal energy of the leaving water and is lost to the intermediate water-loss cation, which decreases the slope of the breakdown curves of sequentially formed fragment ions and shifts them to higher photon energy. Finally (3), slow dissociation rate constants close to the threshold mean that not all ions dissociate that have enough energy to do so. Some extra energy, called the kinetic shift, is needed to speed up the dissociation sufficiently, which also shifts the breakdown curves to higher energy. Among these, (1) can be calculated using the Boltzmanndistribution, (2) can be established on the basis of statistical theory, and information on (3) is contained in the m/z = 100 peak shapes, i.e., in the peak center of gravity as a function of photon energy (Figure 6). The last unknown, thus, remains the

Figure 6. Center of gravity (CoG) analysis of fragment ion m/z 100 (C5H8O2). The statistical dissociation model (blue) is fitted to reproduce the experimentally determined CoG (red) to account for the kinetic shift.

threshold to sequential CO loss, which was calculated to be 10.68 eV using the CBS-QB3 method and the statistical model yields E0 = 10.77 eV, in good agreement with the ab initio calculation, and confirming the dissociation mechanism. There are two main parallel dissociation channels, CH2COOH loss from the adipic acid ion yielding [8] and competing with water loss, and CH2COOH loss from the water-loss intermediate yielding [10] and competing with CO loss. The branching ratios in the breakdown diagram correspond to the rate constant ratios, which can be used to construct a dissociation rate curve for the higher-lying channel and extrapolate it to the threshold energy. For these channels, the statistical model yields 10.32 and 11.53 eV; cf. the CBSQB3 computational results of 10.45 and 11.57 eV, respectively. The agreement is also quite good here, especially considering that fitting parallel channels is generally fraught with more uncertainty.19 The statistical model, thus, validates the fragmentation mechanism proposed and confirms the calculated barriers to the three processes.



AUTHOR INFORMATION

Corresponding Author

*E-mail address: [email protected]. Tel: +41 56 310 4471. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This research received funding from the European Community’s Seventh Framework Program (FP7/2007-2013) under grant agreement no. 290605 (COFUND: PSI-FELLOW), the Swiss Federal Office for Energy (BFE Contract Numbers 101969/152433 & SI/501269-01), and the Swiss National Science Foundation (Starting grant BSSGI0_155846). The authors thank Jan Blees for his help producing the deuterated adipic acid sample. G

DOI: 10.1021/acs.jpca.6b00908 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A



(18) Bodi, A.; Stevens, W. R.; Baer, T. Understanding the complex dissociation dynamics of energy selected dichloroethylene ions: neutral isomerization energies and heats of formation by imaging photoelectron-photoion coincidence. J. Phys. Chem. A 2011, 115 (5), 726− 734. (19) Harvey, J.; Tuckett, R. P.; Bodi, A. A halomethane thermochemical network from iPEPICO experiments and quantum chemical calculations. J. Phys. Chem. A 2012, 116 (39), 9696−9705. (20) Bodi, A.; Hemberger, P.; Gerber, T.; Sztaray, B. A new double imaging velocity focusing coincidence experiment: i2PEPICO. Rev. Sci. Instrum. 2012, 83 (8), 083105. (21) Chim, R. Y. L.; Kennedy, R. A.; Tuckett, R. P.; Zhou, W. D.; Jarvis, G. K.; Collins, D. J.; Hatherly, P. A. Fragmentation of energyselected SF5CF3+ probed by threshold photoelectron photoion coincidence spectroscopy: Bond dissociation energy of SF5-CF3 and its atmospheric implications. J. Phys. Chem. A 2001, 105 (37), 8403− 8412. (22) Tang, X. F.; Zhou, X. G.; Wu, M. M.; Gao, Z.; Liu, S. L.; Liu, F. Y.; Shan, X. B.; Sheng, L. S. Dissociation limit and dissociation dynamic of CF4+: Application of threshold photoelectron-photoion coincidence velocity imaging. J. Chem. Phys. 2013, 138 (9), 094306. (23) Bodi, A.; Johnson, M.; Gerber, T.; Gengeliczki, Z.; Sztaray, B.; Baer, T. Imaging photoelectron photoion coincidence spectroscopy with velocity focusing electron optics. Rev. Sci. Instrum. 2009, 80 (3), 034101. (24) Hemberger, P.; Bodi, A.; Gerber, T.; Würtemberger, M.; Radius, U. Radius, U., Unimolecular reaction mechanism of an imidazolin-2ylidene: an iPEPICO study on the complex dissociation of an Arduengo-type carbene. Chem. - Eur. J. 2013, 19 (22), 7090−7099. (25) Hemberger, P.; Bodi, A.; Berthel, J. H. J.; Radius, U. Intramolecular C-N bond activation and ring-expansion reactions of N-heterocyclic carbenes. Chem. - Eur. J. 2015, 21 (4), 1434−1438. (26) Liang, S. Y.; Hemberger, P.; Neisius, N. M.; Bodi, A.; Grutzmacher, H.; Levalois-Grutzmacher, J.; Gaan, S. Elucidating the thermal decomposition of dimethyl methylphosphonate by vacuum ultraviolet (VUV) photoionization: pathways to the PO radical, a key species in flame-retardant mechanisms. Chem. - Eur. J. 2015, 21 (3), 1073−1080. (27) Hemberger, P.; da Silva, G.; Trevitt, A. J.; Gerber, T.; Bodi, A. Are the three hydroxyphenyl radical isomers created equal? - The role of the phenoxy radical. Phys. Chem. Chem. Phys. 2015, 17 (44), 30076−30083. (28) Peterson, K. A.; Feller, D.; Dixon, D. A. Chemical accuracy in ab initio thermochemistry and spectroscopy: current strategies and future challenges. Theor. Chem. Acc. 2012, 131 (1), 1079. (29) Benoit, F.; Holmes, J. L. Mass-spectra of carboxylic-acids 5. Carboxyl-carboxyl interaction in fragmentation of some cycloalkene1,2-dicarboxylic acids. Org. Mass Spectrom. 1972, 6 (5), 549−557. (30) Benoit, F.; Holmes, J. L. Mass-Spectra of Carboxylic-Acids 0.4. Carboxyl-Carboxyl Interaction in Some Cycloalkane 1,2-DicarboxylicAcids and Its Relationship with Molecular Geometry. Org. Mass Spectrom. 1972, 6 (5), 541−548. (31) Benoit, F.; Holmes, J. L.; Isaacs, N. S. Mass Spectra of Carboxylic Acids.I. Fragmentation Mechanisms in Maleic and Fumaric Acids and Related Compounds. Org. Mass Spectrom. 1969, 2 (6), 591− 601. (32) Botter, R.; Pechine, J. M.; Rosenstock, H. M. Photoionization of dimethyl ether and diethyl-ether. Int. J. Mass Spectrom. Ion Phys. 1977, 25 (1), 7−25. (33) Cocksey, B. J.; Eland, J. H. D.; Danby, C. J. Effect of Alkyl Substitution on Ionisation Potential. J. Chem. Soc. B 1971, 5, 790−792. (34) Voronova, K.; Mozaffari Easter, C. M.; Covert, K. J.; Bodi, A.; Hemberger, P.; Sztaray, B. Dissociative photoionization of diethyl ether. J. Phys. Chem. A 2015, 119 (43), 10654−10663. (35) Lias, S. G. Ionization Energy Evaluation. In NIST Chemistry WebBook; NIST Standard Reference Database Number 69; Linstrom, P. J., Mallard, W.G., Eds.; National Institute of Standards and Technology: Gaithersburg, MD, http://webbook.nist.gov (retrieved 17 November 2015).

REFERENCES

(1) Hallquist, M.; Wenger, J. C.; Baltensperger, U.; Rudich, Y.; Simpson, D.; Claeys, M.; Dommen, J.; Donahue, N. M.; George, C.; Goldstein, A. H.; et al. The formation, properties and impact of secondary organic aerosol: current and emerging issues. Atmos. Chem. Phys. 2009, 9 (14), 5155−5236. (2) Jimenez, J. L.; Canagaratna, M. R.; Donahue, N. M.; Prevot, A. S.; Zhang, Q.; Kroll, J. H.; DeCarlo, P. F.; Allan, J. D.; Coe, H.; Ng, N. L.; et al. Evolution of organic aerosols in the atmosphere. Science 2009, 326 (5959), 1525−1529. (3) Hoffmann, T.; Huang, R. J.; Kalberer, M. Atmospheric Analytical Chemistry. Anal. Chem. 2011, 83 (12), 4649−4664. (4) Allan, J. D.; Jimenez, J. L.; Williams, P. I.; Alfarra, M. R.; Bower, K. N.; Jayne, J. T.; Coe, H.; Worsnop, D. R. Quantitative sampling using an Aerodyne aerosol mass spectrometer - 1. Techniques of data interpretation and error analysis. J. Geophys. Res-Atmos. 2003, 108 (D3), 4090. (5) Crippa, M.; DeCarlo, P. F.; Slowik, J. G.; Mohr, C.; Heringa, M. F.; Chirico, R.; Poulain, L.; Freutel, F.; Sciare, J.; Cozic, J.; et al. Wintertime aerosol chemical composition and source apportionment of the organic fraction in the metropolitan area of Paris. Atmos. Chem. Phys. 2013, 13 (2), 961−981. (6) DeCarlo, P. F.; Ulbrich, I. M.; Crounse, J.; de Foy, B.; Dunlea, E. J.; Aiken, A. C.; Knapp, D.; Weinheimer, A. J.; Campos, T.; Wennberg, P. O.; et al. Investigation of the sources and processing of organic aerosol over the Central Mexican Plateau from aircraft measurements during MILAGRO. Atmos. Chem. Phys. 2010, 10 (12), 5257−5280. (7) Heringa, M. F.; DeCarlo, P. F.; Chirico, R.; Tritscher, T.; Clairotte, M.; Mohr, C.; Crippa, M.; Slowik, J. G.; Pfaffenberger, L.; Dommen, J.; et al. A new method to discriminate secondary organic aerosols from different sources using high-resolution aerosol mass spectra. Atmos. Chem. Phys. 2012, 12 (4), 2189−2203. (8) Lanz, V. A.; Alfarra, M. R.; Baltensperger, U.; Buchmann, B.; Hueglin, C.; Prevot, A. S. H. Source apportionment of submicron organic aerosols at an urban site by factor analytical modelling of aerosol mass spectra. Atmos. Chem. Phys. 2007, 7 (6), 1503−1522. (9) Canonaco, F.; Crippa, M.; Slowik, J. G.; Baltensperger, U.; Prevot, A. S. H. SoFi, an IGOR-based interface for the efficient use of the generalized multilinear engine (ME-2) for the source apportionment: ME-2 application to aerosol mass spectrometer data. Atmos. Meas. Tech. 2013, 6 (12), 3649−3661. (10) Alfarra, M. R.; Prevot, A. S. H.; Szidat, S.; Sandradewi, J.; Weimer, S.; Lanz, V. A.; Schreiber, D.; Mohr, M.; Baltensperger, U. Identification of the mass spectral signature of organic aerosols from wood burning emissions. Environ. Sci. Technol. 2007, 41 (16), 5770− 5777. (11) Gloaguen, E.; Mysak, E. R.; Leone, S. R.; Ahmed, M.; Wilson, K. R. Investigating the chemical composition of mixed organic-inorganic particles by ″soft″ vacuum ultraviolet photoionization: The reaction of ozone with anthracene on sodium chloride particles. Int. J. Mass Spectrom. 2006, 258 (1−3), 74−85. (12) Mysak, E. R.; Wilson, K. R.; Jimenez-Cruz, M.; Ahmed, M.; Baer, T. Synchrotron radiation based aerosol time-of-flight mass spectrometry for organic constituents. Anal. Chem. 2005, 77 (18), 5953−5960. (13) Skeen, S. A.; Michelsen, H. A.; Wilson, K. R.; Popolan, D. M.; Violi, A.; Hansen, N. Near-threshold photoionization mass spectra of combustion-generated high-molecular-weight soot precursors. J. Aerosol Sci. 2013, 58, 86−102. (14) Holmes, J. L.; St. Jean, T. The mass spectra of carboxylic acidsII: Fragmentation mechanisms in the homologous series HOOC(CH2)nCOOH. Org. Mass Spectrom. 1970, 3 (12), 1505−1518. (15) Ruscic, B.; Berkowitz, J. The heats of formation of some C2H5O+ isomers, relevant bond-energies in ethanol and PA(CH3CH). J. Chem. Phys. 1994, 101 (12), 10936−10946. (16) Harvey, Z. A.; Traeger, J. C. The propionyl cation heat of formation revisited. J. Mass Spectrom. 2004, 39 (7), 802−807. (17) Bodi, A.; Csontos, J.; Kallay, M.; Borkar, S.; Sztaray, B. On the protonation of water. Chem. Sci. 2014, 5 (8), 3057−3063. H

DOI: 10.1021/acs.jpca.6b00908 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A (36) Bodi, A.; Brannock, M. D.; Sztaray, B.; Baer, T. Tunneling in H loss from energy selected ethanol ions. Phys. Chem. Chem. Phys. 2012, 14 (46), 16047−16054. (37) Johnson, M.; Bodi, A.; Schulz, L.; Gerber, T. Vacuum ultraviolet beamline at the Swiss Light Source for chemical dynamics studies. Nucl. Instrum. Methods Phys. Res., Sect. A 2009, 610 (2), 597−603. (38) Stull, D. R. Vapor pressure of pure substances - organic compounds. Ind. Eng. Chem. 1947, 39 (4), 517−540. (39) Wright, S. F.; Alexander, K. A.; Dollimore, D. The initial thermal characterization of hair color rinse ingredients - adipic acid. Thermochim. Acta 2001, 367, 29−35. (40) Bodi, A.; Sztaray, B.; Baer, T.; Johnson, M.; Gerber, T. Data acquisition schemes for continuous two-particle time-of-flight coincidence experiments. Rev. Sci. Instrum. 2007, 78 (8), 084102. (41) Sztaray, B.; Baer, T. Suppression of hot electrons in threshold photoelectron photoion coincidence spectroscopy using velocity focusing optics. Rev. Sci. Instrum. 2003, 74 (8), 3763−3768. (42) Baer, T.; Sztaray, B.; Kercher, J. P.; Lago, A. F.; Bodi, A.; Skull, C.; Palathinkal, D. Threshold photoelectron photoion coincidence studies of parallel and sequential dissociation reactions. Phys. Chem. Chem. Phys. 2005, 7 (7), 1507−1513. (43) West, B.; Joblin, C.; Blanchet, V.; Bodi, A.; Sztaray, B.; Mayer, P. M. On the dissociation of the naphthalene radical cation: new iPEPICO and tandem mass spectrometry results. J. Phys. Chem. A 2012, 116 (45), 10999−11007. (44) Peng, C. Y.; Ayala, P. Y.; Schlegel, H. B.; Frisch, M. J. Using redundant internal coordinates to optimize equilibrium geometries and transition states. J. Comput. Chem. 1996, 17 (1), 49−56. (45) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; et al. Gaussian 09; Gaussian, Inc.: Wallingford, CT, USA, 2009. (46) Montgomery, J. A.; Frisch, M. J.; Ochterski, J. W.; Petersson, G. A. A complete basis set model chemistry. VII. Use of the minimum population localization method. J. Chem. Phys. 2000, 112 (15), 6532− 6542. (47) Curtiss, L. A.; Redfern, P. C.; Raghavachari, K. Gaussian-4 theory. J. Chem. Phys. 2007, 126 (8), 084108. (48) Shao, Y. H.; Gan, Z. T.; Epifanovsky, E.; Gilbert, A. T. B.; Wormit, M.; Kussmann, J.; Lange, A. W.; Behn, A.; Deng, J.; Feng, X. T.; et al. Advances in molecular quantum chemistry contained in the Q-Chem 4 program package. Mol. Phys. 2015, 113 (2), 184−215. (49) Sztaray, B.; Bodi, A.; Baer, T. Modeling unimolecular reactions in photoelectron photoion coincidence experiments. J. Mass Spectrom. 2010, 45 (11), 1233−45. (50) Harvey, J.; Tuckett, R. P.; Bodi, A. Shining new light on the multifaceted dissociative photoionisation dynamics of CCl4. Phys. Chem. Chem. Phys. 2014, 16 (38), 20492−20499. (51) Baer, T.; Guerrero, A.; Davalos, J. Z.; Bodi, A. Dissociation of energy selected Sn(CH3)4+, Sn(CH3)3Cl+, and Sn(CH3)3Br+ ions: evidence for isolated excited state dynamics. Phys. Chem. Chem. Phys. 2011, 13 (39), 17791−17801. (52) Fischer, K. H.; Schneider, M.; Fischer, I.; Pfaffinger, B.; Braunschweig, H.; Sztaray, B.; Bodi, A. Bonding in a borylene complex investigated by photoionization and dissociative photoionization. Chem. - Eur. J. 2012, 18 (15), 4533−4540. (53) Hemberger, P.; Bodi, A.; Schon, C.; Steinbauer, M.; Fischer, K. H.; Kaiser, C.; Fischer, I. A pass too far: dissociation of internal energy selected paracyclophane cations, theory and experiment. Phys. Chem. Chem. Phys. 2012, 14 (34), 11920−11929.

I

DOI: 10.1021/acs.jpca.6b00908 J. Phys. Chem. A XXXX, XXX, XXX−XXX