Photoelectron Spectrum and Energetics of the meta-Xylylene Diradical

Communication Cite This: J. Am. Chem. Soc. 2017, 139, 14348-14351

pubs.acs.org/JACS

Photoelectron Spectrum and Energetics of the meta-Xylylene Diradical Mathias Steglich,† Victoria B. F. Custodis,‡ Adam J. Trevitt,¶ Gabriel daSilva,§ Andras Bodi,† and Patrick Hemberger*,† †

Paul Scherrer Institute, CH-5232 Villigen-PSI, Switzerland Department of Chemistry and Applied Biosciences, Institute for Chemical and Bioengineering, ETH Zurich, HCI D 130, Vladimir-Prelog-Weg 1, 8093 Zurich, Switzerland ¶ School of Chemistry, University of Wollongong, Wollongong, New South Wales 2522, Australia § Department of Chemical and Biomolecular Engineering, The University of Melbourne, Melbourne, Victoria 3010, Australia ‡

S Supporting Information *

heat of formation was obtained by pulsed-ion cyclotron doubleresonance spectroscopy.18 Later, it was reported at ΔHf,298K(mC8H8) = (338 ± 10) kJ mol−1 by combining the electron affinity of the diradical, (0.919 ± 0.008) eV,7 and acidity measurements of the 3-methylbenzyl radical as well as by collision-induced dissociation threshold energy measurements of 3-(chloromethyl)benzyl.19 Unlike the closed-shell isomers para- and ortho-xylylene, mC8H8 cannot be synthesized via unimolecular decomposition from the corresponding meta-xylyl radical C8H9,20 which isomerizes to ortho- or para-xylyl prior to hydrogen elimination in model flames and microreactors.21,22 Flash vacuum pyrolysis of 1,3-bis-iodomethyl-benzene (m-C8H8I2) provides efficient synthesis instead.17 Due to lower dissociation energy of the C−I bond compared to the C−H bond, loss of the second iodine atom is favored to yield the diradical. We used this method to produce m-C8H8 in an SiC reactor heated to (900 ± 100)K and probe it thereafter with vacuum ultraviolet (VUV) synchrotron light. Photoelectron spectra of diradical species are scarce, but much-needed to understand their electronic and magnetic properties.23−25 In addition to insights into the electronic and vibrational structure, isobaric species may be identified isomerselectively based on their characteristic spectral signature. This makes mass spectrometry in conjunction with photoelectron spectroscopy a selective and sensitive analytical technique.26 Before using photoelectron photoion coincidence (PEPICO)27 to address complex reactive systems,28 it is worthwhile to synthesize and characterize potential reactive intermediates purely.29 In this study, photoion mass-selected threshold photoelectron (ms-TPE) spectroscopy and photoelectron (PE) imaging was applied to obtain the adiabatic ionization energy (AIE) and characterize the ground and first excited state of the m-C8H8 cation. PEPICO is able to measure energetics data, such as enthalpies of formation or proton affinities with sub-kJ mol−1 accuracy in favorable cases.30 Its application to open-shell systems, however, is often limited by low signal levels and plagued by uncertainties in excess of 15 kJ mol−1.31,32 This is because, first, doublet

ABSTRACT: The meta-xylylene diradical m-C8H8 is a prototypical organic triplet that represents a building block for organic molecule-based magnets and also serves as a model compound for test and refinement of quantum chemical calculations. Flash vacuum pyrolysis of 1,3-bisiodomethyl-benzene (m-C8H8I2) produces m-C8H8 in gas phase; we used photoelectron spectroscopy to probe the first two electronic states of the radical cation, and resolve the vibrational fine structure of the ground state band. The determined adiabatic ionization energy of m-C8H8 is (7.27 ± 0.01) eV. Heat of formation of the diradical was established measuring C−I bond dissociation thresholds in the precursor cation and utilizing a thermochemical cycle to yield ΔHf,298K = (325 ± 8) kJ mol−1, ca. 10 kJ mol−1 below the previous value.

T

heir unique chemistry and electronic structure have thrust aromatic diradicals, such as meta-xylylene (m-C8H8), in the limelight.1−5 They are considered building blocks for future organic molecule-based magnetic materials, possibly replacing conventional metallic magnets in many applications.6 For example, m-C8H8 can act as a ferromagnetic coupling unit in a polymeric network3 due to its large triplet−singlet gap of 0.42 eV.7,8 As a consequence, small perturbations of the m-C8H8 structural motif will not change the spin nature of the ground state. The isoelectronic meta-benzoquinone has also been studied9−11 and a triplet−singlet gap of 0.39 eV was obtained.12 A tetraphenyl-substituted, stable derivative of m-C8H8 was first observed by Schlenk and Brauns.13 The confirmation of the triplet character of Schlenk’s hydrocarbon by electron spin resonance (ESR) spectroscopy dates to 1970.14 The parent mC8H8 is more difficult to investigate, because it has to be trapped applying matrix isolation or studied in the isolated phase using molecular beam techniques in vacuum conditions. In a 77 K npentane matrix, UV−vis absorption and fluorescence spectra of m-C8H8 have been reported, agreeing with a triplet ground state,15 eventually confirmed by ESR spectroscopy.16 Vibrations of the ground triplet and first singlet state have been characterized by argon matrix17 and anion photoelectron spectroscopy.7 A lower limit of 318 kJ mol−1 for the m-C8H8 © 2017 American Chemical Society

Received: June 30, 2017 Published: September 30, 2017 14348

DOI: 10.1021/jacs.7b06714 J. Am. Chem. Soc. 2017, 139, 14348−14351

Communication

Journal of the American Chemical Society

shaped bands labeled here G1 and G2. A weak angular anisotropy β2 is obtained, yielding ∫ G1β2 = 0.06 ± 0.06 and ∫ G2β2 = −0.12 ± 0.06, respectively, when integrated over both individual PE bands. The difference indicates G2 photoelectrons emerge from an orbital with slightly more s-character (HOMO1) compared to the G1 band (HOMO), i.e., two different electronic states contribute to the observed spectrum. To analyze both transitions more, the neutral and ionic ground state geometries and Hessians of m-C8H8, optimized at B3LYP/aug-cc-pVTZ level, were used to calculate Franck− Condon (FC) factors and to simulate the vibrational progression in the X̃ (3B2)→X̃ +(2B1) band system (Figure 2a).

radicals form closed-shell cations upon ionization, which rarely dissociate into two open-shell fragments at the thermochemical threshold. Second, thermochemical cycles, which bypass this complication by employing the ionization energy of the radical, rely on a well resolved origin transition in the PE spectrum, often unavailable.33 We surmount these difficulties in m-C8H8 by combining the accurate ionization energy with the measured dissociative ionization onset of the second C−I bond cleavage in m-C8H8I2. Using an ion cycle, the formation enthalpy of the diradical could thus be determined within 8 kJ mol−1. Photoelectron spectrum. In the ms-TPE spectrum of m-C8H8 (Figure 1a), two main features are apparent with maxima at 7.26

Figure 2. Franck−Condon simulations of electronic transitions compared to ms-TPE spectrum: (a) X̃ (3B2)→X̃ +(2B1) and (b) X̃ (3B2)→Ã +(2A2). At 600 K vibrational temperature, the rotational and instrumental broadening was taken into account by convoluting the stick spectrum with Gaussians (fwhm of 25 meV).

From this, the vibrational temperature in the experiment is about 600 K, a reasonable value considering the expansion conditions of the source35 and complexity of the radical. It is in agreement with other radical spectra from the same source.36,37 The progression is dominated by the origin band and excitations of the symmetric in-plane bending vibration of the methylene groups ν15(a1) calculated at 306 cm−1 in X̃ +(2B1). This value agrees well with the experimentally obtained (300 ± 20)cm−1. In X̃ (3B2), the frequency of this vibration can be approximated at (250 ± 60) cm−1, because several hot bands merge into a broad shoulder on the red wing of the origin band. However, ν15 was already observed7 more precisely in the neutral triplet state by anion PE spectroscopy at 290 cm−1. The ring breathing mode ν14(a1) overlaps with the 1520 band. It is probably at (560 ± 20) cm−1 in X̃ +(2B1), close to the calculated value of 542 cm−1 and to the value observed7 in X̃ (3B2) at 540 cm−1. The deviations of the calculated FC envelope and the experimental spectrum above 7.38 eV originate most likely from autoionization resonances, which are not considered in the simulation. It is interesting to note that the geometries are quite similar in the neutral triplet and cation doublet, as indicated by the strength of the 000 band and the comparable vibrational frequencies in both states. The same can be expected for m-C8H8 derivatives embedded in a magnetic network; upon charge transfer, electronic and magnetic properties of the material will change, but not the structural ones. Attempts to obtain optimized geometries and vibrational frequencies of the à +(2A2) state applying time-dependent DFT methods failed. Instead, we used equation-of-motion coupledcluster theory (EOM-IP-CCSD) in conjunction with the ccpVDZ basis set for the à + state and CCSD/cc-pVDZ theory for the X̃ (3B2) state to obtain the FC simulation shown in Figure 2b. The calculated adiabatic energy difference between X̃ + and à + of 0.28 eV compares well with the distance between the 000 band of the first transition (7.26 eV) and a shoulder at 7.55 eV in the

Figure 1. Photoelectron spectroscopy of m-C8H8. (a) ms-TPE spectrum. (b) PE spectrum reconstructed from PE image, with graphical representations of HOMO (left) and HOMO-1 (right). (c) PE image and slice through center of reconstructed three-dimensional PE momentum distribution.

and 7.58 eV. The first one is accompanied by two minor peaks on the blue side and a shoulder on the red wing. It is assigned as X̃ (3B2)→X̃ +(2B1) transition. Orbital symmetries were obtained by density functional theory (DFT) at B3LYP/aug-cc-pVTZ level. The singly occupied HOMO and HOMO-1 of the triplet are depicted in Figure 1b. They are of a2 and b1 symmetry, respectively, resulting in a 3B2 ground state. Ionization from the HOMO generates a 2B1 ion, in line with Koopmans’ theorem and with a DFT calculation of the ion’s ground state. The band at 7.26 eV marks the adiabatic ionization energy of m-C8H8, which is (7.27 ± 0.01)eV, taking into account the Stark-shift.34 Good agreement was found with CBS-QB3 and G4 composite method calculations, yielding 7.32 and 7.30 eV, respectively. The second feature at 7.58 eV agrees neither with a fundamental, nor with an overtone vibrational band of the ion’s ground state. The first excited state is expected to be caused by ionization from the HOMO-1 of the triplet (b1 orbital), giving rise to the à +(2A2) state. This is confirmed by a time-dependent DFT calculation, predicting a vertical excitation energy of 0.52 eV, reasonably close to the measured value of 0.32 eV. The second excited state as produced from ionization of the HOMO-2 of the triplet is at 2.48 eV, which is too high to observe here. The assignment of the 7.58 eV feature to an electronic state different from the ground one is supported by Franck−Condon simulations (see hereafter) and the PE image (Figure 1c). The PE spectrum extracted from this image (Figure 1b) resembles the ms-TPE spectrum at lower spectral resolution. It exhibits two broad, approximately Gaussian 14349

DOI: 10.1021/jacs.7b06714 J. Am. Chem. Soc. 2017, 139, 14348−14351

Communication

Journal of the American Chemical Society

Information), i.e., the number of π electrons basically increases from 5.x to 6, leading to a stabilization due to aromaticity in the closed-shell cation and thus to a lower C−I bond energy in mC8H8I2+. Removal of the second iodine then creates a radical cation, which should be energetically less favorable than the first iodine removal. Indeed, the second removal costs 11.05 eV − 9.18 eV = 1.87 eV (180 kJ mol−1), as an additional resonance stabilization cannot be achieved. The enthalpy of formation of the precursor 1,3-bisiodomethylbenzene has not yet been determined experimentally. Here, we obtained it by calculating the reaction enthalpy of the isodesmic reaction (2) and using known heats of formation

second feature. Moreover, agreement of the FC simulation with the TPE spectrum is good enough to justify assignment of the shoulder to the origin band and of the maximum at 7.58 eV to the first excitation of the methylene bending vibration ν15(a1). Its energy is approximately (266 ± 40)cm−1, which is lower than in the neutral and cation ground states. Energetics. Enthalpies of formation are important thermochemical parameters for the calculation of reaction enthalpies and equilibrium constants at arbitrary temperatures and pressures. They are difficult to obtain for reactive molecules since standard approaches, such as bomb calorimetry, cannot be applied. Indirect methods, for example applying gas phase PE spectroscopy, have to be used.38 We determined the heat of formation of m-C8H8 by the positive ion cycle depicted in Figure 3. The heats of formation of m-C8H8 and m-C8H8I2 are

of benzyl iodide40 and benzene.41 Utilizing different theoretical approaches, the mean reaction enthalpy is ΔHR,298K = (0.3 ± 1) kJ mol−1 (see Supporting Information). After correcting to 0 K,42,43 the enthalpy of formation of the precursor is then derived as ΔHf,0K(m-C8H8I2) = 198.6 kJ mol−1. The corresponding value of the iodine radical is ΔHf,0K(I) = 107.2 kJ mol−1.44 Equation 1 can now be evaluated, resulting in ΔHf,0K(mC8H8) = (349 ± 8) kJ mol−1. The room temperature value is (325 ± 8) kJ mol−1, somewhat lower than what has been obtained by Hammad et al.,19 but still above the lower limit of Pollack et al. (see Table 1).18 To compare with the experimental

Figure 3. Experimental ionization and appearance energies. The thermochemical cycle connecting the heats of formation of the precursor and m-C8H8 is shown on the right.

Table 1. Heat of Formation of m-C8H8 Obtained from a Combination of Experiment and Theory in Comparison to Literature Values and Calculations Using Isodesmic Reactions (in kJ mol−1)

connected by the adiabatic ionization energy of the former and the appearance energy of the second sequential iodine loss in the dissociative photoionization (DPI) of the latter. To measure this threshold, the precursor was vaporized and photoionized without pyrolysis. Based on the breakdown diagram, which plots the fractional ion abundance in coincidence with threshold electrons as a function of the photon energy, the two C−I bonds break in sequential dissociation reactions, eventually yielding mC8H8+ along with two iodine atoms. A full description of the breakdown diagram fitting procedure is given in the Supporting Information. The experimentally measured fractional ion abundances and dissociation rate constants enable an accurate determination of dissociative photoionization onsets.39 In the absence of a reverse barrier, the appearance energy AE0K of the second iodine loss connects the minima of the m-C8H8I2 and mC8H8+ potential energy surfaces directly. The formation enthalpy of the diradical can then be determined via

ΔHf,298K this work Hammad et al.19 Pollack et al.18 G4 CBS-QB3 CBS-APNO average

eq 3 328.7 320.8 322.8 324.1

325 ± 8 335.1 ± 16/339.7 ± 13 ≥318 eq 4 330.6 321.2 323.7 325.2

eq 5 325.7 320.5 325.9 324.0

value, we also obtained the heat of formation of m-C8H8 by computations applying three different reactions (eq 3,4,5). The

ΔHf ,0K(m‐C8H8) = AE0K (m‐C8H8+) + ΔHf ,0K(m‐C8H8I 2) − 2ΔHf ,0K(I) − AIE(m‐C8H8)

(1)

The obtained appearance energies for the first and second iodine loss are AE0K(m-C8H8I+) = (9.18 ± 0.02) eV and AE0K(m-C8H+8 ) = (11.05 ± 0.08) eV, respectively. The average C−I bond dissociation energy in the precursor neutral can thus be calculated as one-half of the difference of the appearance energy of the m-C8H8+ cation and the ionization energy of the m-C8H8 diradical, yielding 1.89 eV (182 kJ mol−1). Interestingly, the first C−I bond breaking energy of the precursor cation is only 9.18 eV − 8.50 eV = 0.68 eV (66 kJ mol−1). The significant lower bond energy is explained by charge stabilization on the benzene ring. The positive charge in m-C8H8I2+ is about evenly distributed between the benzene ring and the iodine atoms, while in m-C8H8I+ it is almost exclusively located on the ring (see calculated electrostatic potentials in the Supporting

calculated values (Table 1) yield an average heat of formation of 324.4 kJ mol−1 at a standard deviation of 3.6 kJ mol−1, which is within 1 kJ mol−1 of our experimental one and well within the uncertainty of the measurement. To summarize, we investigated photoionization of the mC8H8 diradical using imaging PE and ms-TPE spectroscopy. The adiabatic ionization energy is (7.27 ± 0.01) eV. A symmetric inplane bending vibration of the methylene groups has comparable vibrational energy in the neutral and cation ground states. The multiplicity change upon charge transfer in a 14350

DOI: 10.1021/jacs.7b06714 J. Am. Chem. Soc. 2017, 139, 14348−14351

Communication

Journal of the American Chemical Society

(11) Fu, Q.; Yang, J.; Wang, X.-B. J. Phys. Chem. A 2011, 115, 3201− 3207. (12) Chen, B.; Hrovat, D. A.; Deng, S. H. M.; Zhang, J.; Wang, X.-B.; et al. J. Am. Chem. Soc. 2014, 136, 3589−3596. (13) Schlenk, W.; Brauns, M. Ber. Dtsch. Chem. Ges. 1915, 48, 661− 669. (14) Kothe, G.; Denkel, K.-H.; Sümmermann, W. Angew. Chem., Int. Ed. Engl. 1970, 9, 906−907. (15) Migirdicyan, E.; Baudet, J. J. Am. Chem. Soc. 1975, 97, 7400− 7404. (16) Wright, B. B.; Platz, M. S. J. Am. Chem. Soc. 1983, 105, 628−630. (17) Neuhaus, P.; Grote, D.; Sander, W. J. Am. Chem. Soc. 2008, 130, 2993−3000. (18) Pollack, S. K.; Raine, B. C.; Hehre, W. J. J. Am. Chem. Soc. 1981, 103, 6308−6313. (19) Hammad, L. A.; Wenthold, P. G. J. Am. Chem. Soc. 2000, 122, 11203−11211. (20) Hemberger, P.; Trevitt, A. J.; Ross, E.; da Silva, G. J. Phys. Chem. Lett. 2013, 4, 2546−2550. (21) Hemberger, P.; Trevitt, A. J.; Gerber, T.; Ross, E.; da Silva, G. J. Phys. Chem. A 2014, 118, 3593−3604. (22) Bierkandt, T.; Hemberger, P.; Osswald, P.; Kohler, M.; Kasper, T. Proc. Combust. Inst. 2017, 36, 1223−1232. (23) Chrostowska, A.; Xu, S.; Lamm, A. N.; Mazière, A.; Weber, C. D.; et al. J. Am. Chem. Soc. 2012, 134, 10279−10285. (24) Osborn, D. L.; Vogelhuber, K. M.; Wren, S. W.; Miller, E. M.; Lu, Y.-J.; et al. J. Am. Chem. Soc. 2014, 136, 10361−10372. (25) Oliveira, A. M.; Lu, Y.-J.; Lehman, J. H.; Changala, P. B.; Baraban, J. H.; et al. J. Am. Chem. Soc. 2015, 137, 12939−12945. (26) Bodi, A.; Hemberger, P.; Osborn, D. L.; Sztáray, B. J. Phys. Chem. Lett. 2013, 4, 2948−2952. (27) Baer, T.; Tuckett, R. P. Phys. Chem. Chem. Phys. 2017, 19, 9698− 9723. (28) Hemberger, P.; Custodis, V.; Bodi, A.; Gerber, T.; van Bokhoven, J. Nat. Commun. 2017, 8, 15946-1−15946-8. (29) Reusch, E.; Holzmeier, F.; Constantinidis, P.; Hemberger, P.; Fischer, I. Angew. Chem., Int. Ed. 2017, 56, 8000−8003. (30) Bodi, A.; Csontos, J.; Kállay, M.; Borkar, S.; Sztáray, B. Chem. Sci. 2014, 5, 3057−3063. (31) Holzmeier, F.; Wagner, I.; Fischer, I.; Bodi, A.; Hemberger, P. J. Phys. Chem. A 2016, 120, 4702−4710. (32) Holzmeier, F.; Fischer, I.; Kiendl, B.; Kruger, A.; Bodi, A.; et al. Phys. Chem. Chem. Phys. 2016, 18, 9240−9247. (33) Holzmeier, F.; Herbert, M.-P.; Fischer, I.; Steglich, M.; et al. J. Anal. Appl. Pyrolysis 2017, 124, 454−460. (34) Chupka, W. A. J. Chem. Phys. 1993, 98, 4520−4530. (35) Guan, Q.; Urness, K. N.; Ormond, T. K.; David, D. E.; Ellison, G. B.; et al. Int. Rev. Phys. Chem. 2014, 33, 447−487. (36) de Miranda, B. K. C.; Alcaraz, C.; Elhanine, M.; Noller, B.; Hemberger, P.; et al. J. Phys. Chem. A 2010, 114, 4818−4830. (37) Savee, J. D.; Zador, J.; Hemberger, P.; Sztaray, B.; Bodi, A.; et al. Mol. Phys. 2015, 113, 2217−2227. (38) Blush, J. A.; Clauberg, H.; Kohn, D. W.; Minsek, D. W.; Zhang, X.; et al. Acc. Chem. Res. 1992, 25, 385−392. (39) Sztáray, B.; Bodi, A.; Baer, T. J. Mass Spectrom. 2010, 45, 1233− 1245. (40) Walsh, R.; Golden, D. M.; Benson, S. W. J. Am. Chem. Soc. 1966, 88, 650−656. (41) Roux, M. V.; Temprado, M.; Chickos, J. S.; Nagano, Y. J. Phys. Chem. Ref. Data 2008, 37, 1855−1996. (42) Chase, M. W. J. Phys. Chem. Ref. Data, Monogr. 1998, 9, 1−1963. (43) Harvey, J.; Bodi, A.; Tuckett, R. P.; Sztáray, B. Phys. Chem. Chem. Phys. 2012, 14, 3935−3948. (44) Ruscic, B. Active Thermochemical Tables 2013, values based on ver. 1.119 of the Thermochemical Network, available at ATcT.anl.gov.

magnetic network consisting of m-C8H8 units is expected to alter the electronic and magnetic properties of the material, but not the structural ones, as indicated by the high Franck− Condon factor of the 000 transition in the PE spectrum. The first excited ion state can be populated only 0.3 eV above the origin transition, producing electrons with weak angular anisotropy. Utilizing a cation cycle, we could obtain the heat of formation of the diradical and found agreement with computational results using isodesmic and isogyric reactions. Reliable thermochemical data as well as information on the electronic states of organic diradicals are still scarce in the literature, but are essential, e.g., to evaluate theoretical methods or to describe magnetic properties. Furthermore, adiabatic ionization energies are required in general whenever spectroscopic techniques rely on ionization detection schemes, such as resonance-enhanced multiphoton ionization spectroscopy.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.7b06714. Experimental information and details on appearance energy modeling and formation enthalpy (PDF) Outputs of quantum chemical calculations, including energies and coordinates of calculated structures (ZIP)

■

AUTHOR INFORMATION

Corresponding Author

*[email protected] ORCID

Adam J. Trevitt: 0000-0003-2525-3162 Gabriel daSilva: 0000-0003-4284-4474 Andras Bodi: 0000-0003-2742-1051 Patrick Hemberger: 0000-0002-1251-4549 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This project was funded by the Swiss Federal Office of Energy (SFOE, Contract Number SI/501269-01) and by the Laboratory for Thermal Processes and Combustion (LTV) located at PSI. The authors thank Patrick Ascher for technical support.

■

REFERENCES

(1) Abe, M. Chem. Rev. 2013, 113, 7011−7088. (2) Maneru, D. R.; Pal, A. K.; de P. R. Moreira, I.; Datta, S. N.; Illas, F. J. Chem. Theory Comput. 2014, 10, 335−345. (3) Pal, A. K.; Hansda, S.; Datta, S. N. J. Phys. Chem. A 2015, 119, 2176−2185. (4) Saito, T.; Kitagawa, Y.; Takano, Y. J. Phys. Chem. A 2016, 120, 8750−8760. (5) Tsuji, Y.; Hoffmann, R.; Strange, M.; Solomon, G. S. Proc. Natl. Acad. Sci. U. S. A. 2016, 113, E413−E419. (6) Miller, J. S. Chem. Soc. Rev. 2011, 40, 3266−3296. (7) Wenthold, P. G.; Kim, J. B.; Lineberger, W. C. J. Am. Chem. Soc. 1997, 119, 1354−1359. (8) Lineberger, W. C.; Borden, T. Phys. Chem. Chem. Phys. 2011, 13, 11792−11813. (9) Fort, R. C.; Getty, S. J.; Hrovat, D. A.; Lahti, P. M.; Borden, W. T. J. Am. Chem. Soc. 1992, 114, 7549−7552. (10) Fattahi, A.; Kass, S. R.; Liebman, J. F.; Matos, M. A. R.; Miranda, M. S.; et al. J. Am. Chem. Soc. 2005, 127, 6116−6122. 14351

DOI: 10.1021/jacs.7b06714 J. Am. Chem. Soc. 2017, 139, 14348−14351