Reactions within Fluorobenzene–Ammonia Heterocluster Ions

Jan 25, 2012 - Center for Computational Research, University at Buffalo, State University of .... collimated by two conical nickel skimmers, which del...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/JPCA

Reactions within Fluorobenzene−Ammonia Heterocluster Ions: Experiment and Theory Kristin Butterworth,† Chi-Tung Chiang,† Brian Cunningham,† Marek Freindorf,‡ Thomas R. Furlani,‡ Robert L. DeLeon,† and James F. Garvey*,† †

Department of Chemistry, University at Buffalo, State University of New York, Buffalo, New York 14260-3000, United States Center for Computational Research, University at Buffalo, State University of New York, Buffalo, New York 14260-3000, United States

‡

ABSTRACT: Reactions occurring within gas phase fluorobenenze− ammonia heterocluster cations (FC6H5−(NH3)n=1−4) have been studied through the use of a triple quadrupole mass spectrometer as well as employing density functional theory (DFT). Collision induced dissociation (CID) experiments were conducted in which mass selected cluster ions are accelerated into a cell containing argon gas and the resulting products then subsequently mass analyzed. Two dominate reaction channels are observed. The first is simple evaporative loss of neutral ammonia from the cluster ion. The second involves a substitution reaction occurring within the cluster ion to form the aniline cation, C6H5NH2+, where the reactivity was found to vary as a function of cluster size. DFT calculations have been performed to both help analyze the structure and the reactivity of these cluster ions. Pronounced differences in activation energies were found that provide an explanation for the observed variation of reactivity as a function of cluster size. An ad hoc model based upon the Arrhenius equation was developed to fit both the experimental collision energy dependence of the reaction and the observed lowering of the reaction barrier to aniline formation as a function of cluster size.

1. INTRODUCTION The ultimate goal of studying chemical reactions within gas phase clusters is to obtain an understanding of the factors that can govern reactions in solution but that are absent in gas phase processes.1−20 By observing the chemistry within these finite cluster systems, it is possible to learn how the behavior of the reactants changes from that of a gas phase bimolecular ion− molecule reaction to a typical ion−molecule reaction within solution. Chemistry can occur within these systems leading to unexpected products and unusual pathways. That is, solvation can help catalyze reactions, which otherwise would not occur. We wish to learn how the observed chemical reactivity contrasts with established gas phase chemistry and how it varies as a function of cluster size. Frequently, intracluster reactions are solvent catalyzed, in that a minimum number of solvent molecules are needed for a reaction to occur. In the most extreme cases, a single solvent molecule can serve as a catalyst that sufficiently lowers the activation barrier and allows a reaction to occur that is otherwise not observed. One example that has attracted considerable interest in recent years is an ion cluster that contains a single halogenated benzene solvated by varying numbers of ammonia molecules.21−29 In general, their reactivity is analogous to typical textbook organic reactions in polar liquid solvents in that the products observed are either the aniline cation and hydrogen halide or the anilinium cation and a halogen radical. Fluorobenzene−ammonia (FAn+) clusters are particularly © 2012 American Chemical Society

interesting due to their reactivity and have been studied by several groups including Brutschy et al.21,22 and Mikami et al.26 Both of these groups applied a similar REMPI methodology and although their data are similar, their interpretation of these spectra is very different. Brutschy et al. concluded that aniline formation did not occur from the FA1+ complex as shown in reaction 1. C6H5F+{NH3} → C6H5NH2+ + HF

(1)

Instead, in the case of single molecule catalysis, a second NH3 was required for the aniline formation to be observed and was said to proceed exclusively from the FA2+ complex, as shown in reaction 2.24 C6H5F+{NH3}2 → C6H5NH2+ + NH3 + HF

(2)

In contrast, Mikami et al. concluded that there were two different 1:1 complexes and that reaction to form aniline was from one of these complexes. In this article, we will employ both experiment and theory to probe the interaction of fluorobenzene radical cations with ammonia. By utilizing a triple quadrupole mass spectrometer, we can unambiguously mass-select any cluster ion, react it Received: August 18, 2011 Revised: January 25, 2012 Published: January 25, 2012 1877

dx.doi.org/10.1021/jp2079549 | J. Phys. Chem. A 2012, 116, 1877−1883

The Journal of Physical Chemistry A

Article

Figure 1. Schematic of the instrument used in this work.

anhydrous ammonia from Matheson, and fluorobenzene (99%) from Aldrich.

within a collision cell and then analyze the products to see if indeed aniline is formed. Larger complexes can also be studied and a comprehensive picture of aniline formation as a function of cluster size can be determined.

3. COMPUTATIONAL CALCULATIONS Theoretical calculations were carried out at the DFT level of theory using the B3LYP33,34 hybrid functional and the 6-311+G(d,p) basis set35 for the Q-Chem program36 on the server from the Center for Computational Research (CCR) at SUNY Buffalo. Optimized structures of clusters were calculated without any constraints and allowing C1 symmetry for all calculated cluster systems. This level of theory has been proven to be a precise computational approach in calculations involving hydrogen bonds.36 The reaction profiles were obtained by computing fully geometrically optimized calculations of the cluster, for a fixed value of the reaction coordinate. Then, the reaction coordinate was incrementally changed, and the calculation of the optimized geometry was repeated. We note that this technique of employing coordinate scans provides imprecise approximation to the actual transition states. We hope in the near future to perform such calculations. Calculations of the basis set superposition error (BSSE) for 1:1 heterocluster cluster ions performed using a counterpoise method indicated that the value of this error was smaller than 1 kcal/mol. The BSSE is anticipated to be somewhat larger for the larger clusters investigated in our study; however, we do not expect that even for the largest clusters investigated that the BSSE will exceed 2 kcal/mol. The computations provide a binding energy value for the cluster, which can be determined from the energy difference between the infinitely separated individual molecules (i.e., a fluorobenzene cation and neutral ammonias) and the final cluster using the following equation:

2. EXPERIMENTAL SECTION The experiments discussed here were performed using a Campargue-type30 continuous molecular beam source coupled with a triple quadrupole mass spectrometer (Extrel C-50). The operation of the triple quadrupole and the methods used during this experiment have been discussed in detail in previous work.31,32 A schematic of the instrument used can be seen in Figure 1. The instrument consists of three chambers, the source chamber, the buffer chamber, and the mass spectrometer chamber. The collision induced dissociation (CID) collision cell is located in the second quadrupole (Q2) of mass spectrometer. The fluorobenzene−ammonia mixtures were introduced to the system by passing gaseous ammonia through a bubbler containing the liquid fluorobenzene. A neutral cluster beam is generated in the source chamber by the adiabatic expansion of the mixed vapor at a stagnation pressure of 4 psig supersonically through a 250 μm nozzle. The neutral cluster beam is collimated by two conical nickel skimmers, which delimit the buffer chamber before entering the mass spectrometer chamber. These chambers are operated at 10−1−10−2 Torr, 10−4−10−5 Torr, and 10−5−10−7 Torr, respectively. Upon entering the mass spectrometer chamber, a small fraction of the beam is positively ionized by an electron impact ionization source. During operation, the electron ionization energy is held at 65.0 eV, and the emission current used is 3.0 mA. The ionized clusters are then focused and guided into the triple quadrupole mass spectrometer by collinear ion optics where survey scans, metastable decay, or CID scans can be used to study them. Unless otherwise stated, a collision energy of 11.5 eV, in the lab frame of reference, was used for all CID experiments and is defined as the potential difference between the ionization region and the (Q2) ion optics. For metastable scans not utilizing gas in the collision chamber, the collision energy was set to 7.5 eV. During collision energy studies, the laboratory frame collision energy was varied from 2 to 50 eV. Ions were detected by an off-axis channeltron device, and the output signal was then passed through a 1.5 kHz low pass filter. The final mass spectra were averaged and recorded by a LeCroy 9310A digital oscilloscope. The compounds used during the course of these experiments are as follows: argon was obtained from Irish Welding Supply,

Ebinding energy = (Ecluster − Emolecules)

(3)

The lowest energy clusters are the most stable and most likely structures to exist in the molecular beam. To allow comparison among clusters, energies given in this document are relative energies and are obtained by finding the difference in energy, ΔE, between the cluster in question and the most stable one.

4. RESULTS AND DISCUSSION 4.1. Survey Spectrum. The mass spectrum of fluorobenzene and ammonia, obtained through the method described above, is shown in Figure 2. A series of peaks denoted as (a,b), represent the cluster ion series (C6H5F)a(NH3)b+. A second series is denoted as (a,b)H corresponds to the protonated heterocluster ion series (C6H5F)a(NH3)b−1(NH4)+. Lastly, the 1878

dx.doi.org/10.1021/jp2079549 | J. Phys. Chem. A 2012, 116, 1877−1883

The Journal of Physical Chemistry A

Article

Figure 2. Survey mass spectrum of fluorobenzene and ammonia. (Peaks marked with an asterisk are from the fragmentation of fluorobenzene cation.)

final peak of interest in this spectrum is the aniline peak, N = C6H5NH2+ which is the product formed by the reaction between ammonia and fluorobenzene within a heterocluster ion. 4.2. Metastable Decay. Figure 3 shows metastable spectra for cluster ion, (C6H5F)(NH3)n=1−4+. The mass-selected cluster ion is accelerated at a collision energy of 7.5 eV through an

indicating that a substitution reaction has taken place within the reactant cluster ion. It can be seen from the figure that the aniline peak undergoes a drastic increase in intensity once a second ammonia is added signifying that single molecule catalysis might be occurring. The third channel is the formation of a protonated ammonia cluster, which can be generated from the deprotonation of fluororbenzene and ammonia clusters. 4.3. Collision Induced Dissociation. Figure 4 shows a set of CID spectra for the ion series (C6H5F)(NH3)n=1−4+. CID of

Figure 4. CID series of (C6H5F)(NH3)n+ where n = 1−4.

Figure 3. Metastable decay series of (C6H5F)(NH3)n+ where n = 1−4.

evacuated collision cell, and the third quadrupole is scanned to observe any decay products that were generated during transit through the cell. Three reaction channels are observed. The first more dominant channel is simple evaporative loss of 1 or 2 ammonias to generate the fluorobenzene cation. The second channel observed is the generation of the aniline cation,

the mass selected cluster ion (C6H5F)(NH3)+ exhibits only he evaporative loss of ammonia. For these conditions, no aniline product is observed. However, the CID of the mass selected cluster ion (C6H5F)(NH3)2+ exhibits not just evaporation of ammonia but the generation of the aniline cation, along with a protonated ammonia dimer. We also note the production of 1879

dx.doi.org/10.1021/jp2079549 | J. Phys. Chem. A 2012, 116, 1877−1883

The Journal of Physical Chemistry A

Article

(NH3)2+, which is produced via dissociative electron transfer (DET). DET is possible when the ionization potential (IP) of the charged moiety is higher than that of the neutral moiety, allowing the transfer of the positive charge. The IP values for the halogenated benzenes and various ammonia cluster sizes are shown in Table 1. DET becomes possible in these clusters

4.5. Fitting of Collision Energy Profile. Figure 6 shows the production of the aniline cation from both FA1+ and FA2+

Table 1. Ionization Potentials for the Fluorobenzene and (NH3)n=1−4

a

species

ionization potential (eV)

fluorobenzene NH3 (NH3)2 (NH3)3 (NH3)4

9.2022 (9.08a) 10.0738 (10.20a) 9.1922 (8.49a) 7.94a 7.34a

Denotes values calculated using Q-Chem. Figure 6. Plot of the production of the aniline cation from both FA1+ and FA2+ as a function of collision energy Data points are shown with filled symbols, while the theoretical fittings are shown with open symbols. Because of the large difference in intensity between the FA1+ peaks and the FA2+ aniline peaks, they are plotted on different axes. FA1+ is plotted against the right y axis, and FA2+ is plotted against the left y axis. Solid symbols indicate experiment, while the open symbols indicate theoretical fit.

because the IP of the ammonia trimer is significantly lower than the halobenzene. Although the IP of A2 is also lower than that of fluorobenzene, they are close enough to allow the production of aniline to compete with DET, explaining why the aniline and the ammonia dimer are both observed. In the CID of the mass selected cluster ion (C6H5F)(NH3)3+, the dominant process is the loss of all ammonias. However, aniline generation is still observed. The remaining peaks in the spectra are similar to that already assigned in the (C6H5F)(NH3)2+ CID spectra. Lastly, the CID spectra of (C6H5F)(NH3)4+ is very similar to that of the (C6H5F)(NH3)3+. In these two sizes of clusters, we also observed both series of (NH3)n+ and (NH3)nH+. 4.4. Collision Energy Study. The aniline peak is quite small in both the metastable and CID scans for the FA1+ (i.e., (C6H5F)(NH3)+) cluster ion, yet prominent for the FA2+ (i.e., (C6H5F)(NH3)2+) cluster ion. Figure 5 shows a side by side

as a function of collision energy. An ad hoc Arrhenius-type equation was used to fit the data for the formation of aniline from both FA1+ and FA2+. The model equation is given by eq 4 ⎛ ⎜ Nn F=⎜ ⎜ Elab ⎜ ⎝ mn

⎞ ⎟ ⎟e−Ea /(SEcm) ⎟ ⎟ ⎠

(4)

For this fit, F is the experimental ion intensity, the preexponential factor is a variable parameter (Nn) divided by the square root of the lab standard collision energy (Elab) divided by the mass of the respective cluster in amu (mn). The exponential factor is then the activation energy of the respective reaction (Ea) divided by another fitting parameter, S, and the center of mass energy (Ecm). The overall collision energy dependence of the data is very well fit by the model. Since, for the formation of aniline from fluorobenzene−ammonia clusters we were able to observe the product for both FA1+ and FA2+, we are also able to assess how the activation barrier can effect overall reactivity. Starting from a 15 kcal/mol DFT calculation of the activation energy for the formation of aniline from the FA1+ cluster, the calculated FA2+ cluster barrier is lowered to 5 kcal/mol. The relative barrier model fit of the experimental data determines a value of 9 ± 9 kcal/mol for the FA2+. This large error and the difference of 4 kcal/mol are likely due to the difficulty in accurately measuring such small signals. 4.6. Geometry Optimization of Fluorobenzene and Ammonia Cluster Ions. Figure 7 displays the optimized structures and their corresponding mappings of electron density of cluster ions. A red color indicates a negative electrostatic potential, which corresponds to sites with larger electronic density. A blue color indicates a positive electrostatic potential corresponding to sites with a smaller electron density. According to our calculations, the positive charge on the cluster is mainly localized on the hydrogen atoms on the ammonias, while a negative charge is mainly localized on the benzene and

Figure 5. CID spectra of fluorobenzene(F)−ammonia(A) cluster ions with 2.2 mTorr Ar in collision cell. FA1+ is shown with the collision energies at 11 eV and 25 eV. FA2+ is shown at 13 eV and 31 eV. The arrow indicates the aniline (N) product peak.

comparison of the CID spectra for both FA1+ and FA2+ at two different collision energies. For the FA1+, when the collision energy is increased from 11 to 25 eV, the intensity of the aniline peak doubles, and it is possible to see that the reaction does indeed occur. It can also be seen that when the collision energy is increased in similar fashion for the FA2+ cluster ion, the aniline peak intensity also increases, as expected. 1880

dx.doi.org/10.1021/jp2079549 | J. Phys. Chem. A 2012, 116, 1877−1883

The Journal of Physical Chemistry A

Article

Figure 7. Optimized structures and mappings of electron density of FAn+ clusters. The ΔE values given are only comparable within clusters of the same composition and size.

bond length usually indicates an increase in bond strength, which in this case implies that as the second ammonia is added, the cluster is gaining stability. With the addition of the third ammonia, the C4−N bond length does not decrease, indicating that there is not a significant strengthening of that bond. However, the C2−N bond is much shorter, and the hydrogen bound to C2 is being pushed out of plane. This difference in bond length aligns with the drastic increase in the energy of the C4 isomer relative to the C2 isomer. Also, the aromaticity of the ring is lost upon addition of the third ammonia to the C2 location. This is difficult to see at the angles at which the structures are shown, but a close examination of the structures, along with a change in the electron density diagrams, shows this to be the case. When the fourth ammonia is added and the first solvation shell is complete, the C4−N bond length actually increases, which indicates a lessening of the C−N bond strength. For the C2 bound isomer, the C−N bond stays about the same. During the optimization of FA4+, a third isomer emerged and was found to be more stable than the C4-bound isomer. In this F-bound isomer, the ammonia molecules associate with the fluorine through a hydrogen bond. Similar structures were attempted with the smaller clusters; however, all were found to have much higher relative energies than the structures shown here. Overall, the trend is that initially for the FA1+ cluster, the C1 structure is slightly higher in energy than the C4 structure, but

the fluorine substituent. This charge distribution indicates that the electron density is mainly localized on the π-electron system of the benzene aromatic ring and on the atomic orbitals of the fluorine substituent. The observed charge distribution of the investigated systems generates a strong dipole moment, as indicated by the calculations. Through optimization of various possible structures of the FA1+ cluster, two stable structures emerged, one with the ammonia located over the primary carbon (C1) and the other with the ammonia located over the para carbon (C4). To determine whether the ammonia could freely move about the ring, calculations were done in which the ammonia was moved around the ring or over the center. These calculations showed an energy barrier of 6−7 kcal/mol indicating that the two structures do not interconvert under the present experimental conditions. When this barrier is compared to the activation energies shown later, it can be seen that a reaction with an activation energy of 5 kcal/mol requires the extra energy provided by the argon collisions. When the second ammonia is added to the structure to form the FA2+ cluster, it was expected, based on the previous work with benzene and ammonia, that it would hydrogen bond to the first ammonia.37 The most stable structures did in fact follow this trend. When the second ammonia was placed anywhere else on the ring, the relative energy of the structure jumped at least 5 kcal/mol. The length of the C−N bond decreased in size from the FA1+ cluster to the FA2+ cluster (2.406 Å to 2.294 Å for the C4 bonded cluster). A decrease in 1881

dx.doi.org/10.1021/jp2079549 | J. Phys. Chem. A 2012, 116, 1877−1883

The Journal of Physical Chemistry A

Article

with the addition of each ammonia up to the FA4+ cluster, the C4 structure becomes progressively higher in relative energy compared to the C2 structure. Table 2 shows a comparison of C−N bond length and bond energy. The bond energy per ammonia was calculated using the following equation: Table 2. Comparison of Bond Length and Bond Energy per Ammonia between Clusters isomer 1 (C1 or C2-Bound NH3) no. NH3 1 2 3 4

C1−N bond length (Å)

bond energy per NH3 (kcal/mol)

2.50 2.25 1.54 1.53 isomer 2 (C4-bound NH3)

−17.0 −16.2 −18.0 −17.0

Figure 8. Calculated reaction profile for the formation of aniline from the FA1+ (green line) and FA2+ (blue line) clusters.

no. NH3

C4−N bond length (Å)

bond energy per NH3 (kcal/mol)

1 2 3 4

2.41 2.29 2.29 2.41

−18.4 −16.5 −15.3 −14.2

reaction. The additional ammonia molecules further withdraw electron density from fluorine, and the C−F bond consequently becomes weaker. Thus, the formation of a C−N bond dominates with additional ammonia molecules. The second possible reason is that the second ammonia stabilizes the transition state by forming a six-membered ring instead of the four-membered ring in the FA1+ cluster.

Bond energy per ammonia = (Ecluster − Ecomponents)/#NH3

5. CONCLUSIONS We observe that the FA1+ (i.e., (C6H5F)(NH3)+) cluster ion does react to form aniline. However, the activation energy for this reaction (15 kcal/mol) is large, making this a minor channel at best. Our computational findings indicate that there are two low energy isomers, only one of which can react. It was shown through experimental and computational results that once a second ammonia is added to the cluster (i.e., FA2+ (i.e., (C6H5F)(NH3)2+), the reaction to form aniline proceeds much more readily. This second ammonia acts in essence as a catalyst, assisting the reaction by stabilizing the transition state and helping to remove the fluorine to form HF. While the FA1+ cluster can only form a four-membered ring, the FA2+ cluster is able to form a more stable six-membered ring. The new pathway, allowed by the addition of the second ammonia, has an activation energy 10 kcal/mol lower than the FA1+ pathway. These activation energies were shown to also be consistent with the experimental data through the collision energy data modeling. Because our instrument is able to study larger clusters than either Mikami or Brutschy, we were able to examine the bonding trends present throughout the entire first solvation shell, FA1−4+. The conventional survey scan, together with the calculated binding energies (Table 2), shows that the FA3+ cluster is the most stable cluster of the series. However, upon the addition of the fourth ammonia, the survey scan indicates that the cluster loses stability. This system has proven to resemble the benzene (B) and ammonia31 system in the way the ammonia molecules associate with the ring and with each other. Yet, there are two significant differences between the two systems. First, the most intense peak in the survey spectrum for the BA system is the BA4+ peak, as opposed to the FA system where the FA3+ peak is the most intense. Also, the BA system does not form aniline, making the evaporative loss of ammonia the dominant process. The addition of the fluorine acts to both direct the position of the Lewis acid−base bond with the ammonia and allow the substitution reaction to occur.

(5)

From these calculated bond energies, it can be seen that in isomer 1 (hydrogen bonded at C1 or C2), the FA3+ cluster has the largest stabilization energy per hydrogen bond. This stability corresponds well with the intensity of the FA3+ peak seen in the conventional survey scan (Figure 2). Also, the decrease in bond length and suggested progression from a van der Waals association to a covalent bond indicates that isomer 1 is the isomer that reacts to form aniline. In isomer 2 (hydrogen bonded at C4), each progressive cluster has slightly less stabilization energy per hydrogen bond than the one before it. This would suggest that the main dissociation pathway for this isomer is evaporative loss of the ammonia. Lastly, we note the drastic increase in the energy of the C4 isomer relative to the C2 isomer, in spite of the loss of aromaticity in the C2 isomer. 4.7. Reaction Profiles. We calculated the activation energy for the reaction through reaction profiles. The reaction profiles for the FA1−3+ clusters were obtained by optimizing the structure using a constrained distance between the atoms of the dissociating chemical bond (C1−F). The interatomic distance was then incrementally changed, and the calculation was repeated. The results are shown in Figure 8. From the calculations, the activation energies were found to be 15 kcal/mol for the FA1+ cluster and 5 kcal/mol for the FA2+ cluster. These activation energies agree with the experimental findings in that the reaction takes place much more readily in the FA2+ cluster. When the reaction takes place in the FA3+ cluster, the activation energy is similar to that of the FA2+ cluster (5 kcal/mol), in agreement with the experimental result that the intensity of the aniline peak does not change between the CID scans for FA2+ and FA3+. There are two possible explanations for the large difference between the energy barrier for the dissociation process within the FA1+ cluster as compared to the same reaction in the FA2+ and FA3+ clusters. The first is related to a change in the electronic density on the chemical bond dissociated during the 1882

dx.doi.org/10.1021/jp2079549 | J. Phys. Chem. A 2012, 116, 1877−1883

The Journal of Physical Chemistry A

■

Article

(33) Freindorf, M. Theory of Quantum Chemistry; Center for Computational Research: Buffalo, NY, 2008; p 82. (34) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (35) Hehre, W. Ab Initio Molecular Orbital Theory; John Wiley & Sons: New York, NY, 1986. (36) Kong, J.; White, C. A.; Krylov, A. I. J. Comput. Chem. 2000, 21, 1532. (37) Chiang, C.-T.; Freindorf, M.; Furlani, T.; DeLeon, R. L.; Richard, J. P.; Garvey, J. F. J. Phys. Chem. A 2007, 111, 6068. (38) Lias, S. G. 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, 2011; http://webbook.nist.gov.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was done in part at SUNY-Buffalo’s Center for Computational Research.

■

REFERENCES

(1) Muntean, F.; Armentrout, P. B. J. Chem. Phys. 2001, 115, 1213. (2) Armentrout, P. B. J. Anal. At. Spectrom. 2004, 19, 571. (3) Armentrout, P. B. Annu. Rev. Phys. Chem. 2001, 52, 423. (4) Armentrout, P. B. J. Mass Spectrom. 1999, 34, 74. (5) Dedonder-Lardeux, C.; Jouvet, C.; Martrenchard-Barra, S.; Solgadi, D.; Talbot, F.; Vervloet, M.; Dimicoli, I.; Richard-Viard, M. Chem. Phys. 1996, 212, 371. (6) Brutschy, B. Chem. Rev. 2000, 100, 3891. (7) Brutschy, B. Chem. Rev. 1992, 92, 1567. (8) Thoelmann, D.; Gruetzmacher, H. F. J. Am. Chem. Soc. 1991, 113, 3281. (9) Martrenchard-Barra, S.; Dedonder-Lardeux, C.; Jouvet, C.; Rockland, U.; Solgadi, D. J. Phys. Chem. 1995, 99, 13716. (10) Buchhold, K.; Reimann, B.; Djafari, S.; Barth, H. D.; Brutschy, B.; Tarakeshwar, P.; Kim, K. S. J. Chem. Phys. 2000, 112, 1844. (11) Riehn, C.; Buchhold, K.; Reimann, B.; Djafari, S.; Barth, H. D.; Brutschy, B.; Tarakeshwar, P.; Kim, K. S. J. Chem. Phys. 2000, 112, 1170. (12) Djafari, S.; Barth, H. D.; Buchhold, K. J. Chem. Phys. 1997, 107, 10573. (13) Castleman, A. W.; Bowen, K. H. J. Phys. Chem. 1996, 100, 12911. (14) Dermota, T. E.; Zhong, Q.; Castleman, A. W. Chem. Rev. 2004, 104, 1861. (15) Justes, D. R.; Mitrić, R.; Moore, N. A.; Bonačić-Koutecký, V.; Castleman, A. W. J. Am. Chem. Soc. 2003, 125, 6289. (16) Wisniewski, E. S.; Hershberger, M. A.; Castleman, A. W. J. Chem. Phys. 2002, 116, 5738. (17) Folmer, D. E.; Wisniewski, E. S.; Stairs, J. R.; Castleman, A. W. J. Phys. Chem. A 2000, 104, 10545. (18) MacTaylor, R. S.; Castleman, A. W. J. Atmos. Chem. 2000, 36, 23. (19) Tu, Y.-P.; Holmes, J. L. J. Am. Chem. Soc. 2000, 122, 5597. (20) Zagorevskii, D. V.; Holmes, J. L.; Stone, J. A. Eur. J. Mass Spectrom. 1996, 2, 341. (21) Brutschy, B. J. Phys. Chem. 1990, 94, 8637. (22) Brutschy, B.; Eggert, J.; Janes, C.; Baumgaertel, H. J. Phys. Chem. 1991, 95, 5041. (23) Grover, J. R.; Cheng, B. M.; Herron, W. J.; Coolbaugh, M. T.; Peifer, W. R.; Garvey, J. F. J. Phys. Chem. 1994, 98, 7479. (24) Maeyama, T.; Mikami, N. J. Am. Chem. Soc. 1988, 110, 7238. (25) Maeyama, T.; Mikami, N. J. Phys. Chem. 1990, 94, 6973. (26) Maeyama, T.; Mikami, N. J. Phys. Chem. 1991, 95, 7197. (27) Tholmann, D.; Grutzmacher, H.-F. Chem. Phys. Lett. 1989, 163, 225. (28) Tholmann, D.; Grutzmacher, H.-F. Org. Mass Spectrom. 1989, 24, 439. (29) Van der Hart, W.; Luijten, W.; Thuijl, T. Org. Mass Spectrom. 1980, 15, 463. (30) Campargue, R. J. Phys. Chem. 1984, 88, 4466. (31) Chiang, C.-T.; Freindorf, M.; Furlani, T.; DeLeon, R. L.; Richard, J. P.; Garvey, J. F. J. Chem. Phys. Lett 2011, 509, 102. (32) Shores, K. S.; Charlebois, J. P.; Chiang, C.-T.; DeLeon, R. L.; Freindorf, M.; Furlani, T. R.; Garvey, J. F. J. Phys. Chem. A 2009, 113, 2268. 1883

dx.doi.org/10.1021/jp2079549 | J. Phys. Chem. A 2012, 116, 1877−1883