IR and UV-NIR Absorption Spectroscopy of Matrix-Isolated C70+ and

Jul 5, 2016 - Bonnie Charpentier is 2018 ACS president-elect. Bonnie Charpentier, senior vice president for regulatory, quality, and safety at Cytokin...
7 downloads 42 Views 1MB Size
Article pubs.acs.org/JPCA

IR and UV-NIR Absorption Spectroscopy of Matrix-Isolated C+70 and C−70 Bastian Kern, Artur Böttcher, and Dmitry Strelnikov* Institute of Physical Chemistry II, KIT, Fritz-Haber-Weg 2, 76131 Karlsruhe, Germany S Supporting Information *

ABSTRACT: C+70 ions were mass-selectively deposited into a neon or an argon matrix at 5 K. Like in the case of C+60 deposition, soft landing into a rare gas matrix is associated with some charge-exchange processes such that C+70 as well as resulting C70 and C−70 can be probed simultaneously. In contrast with a very good coincidence of the experimental and DFT-calculated IR spectra of ± C±/2+ 60 , DFT predictions for C70 IR absorptions strongly deviate from our measurements. A possible explanation for this could be low-lying electronically excited states of C±70 in the vicinity of vibrational energy levels. The corresponding non-Born−Oppenheimer case is likely of significant interest to theory.





INTRODUCTION Fullerene ions have recently become of interest to the astronomy community1−5 and are also relevant in the field of organic photovoltaics.6,7 The first real breakthrough in cosmic fullerene research came in 2010 with the identification of neutral C60 and C70 in space via IR spectra.3 The second breakthrough is related to the long-standing puzzle of the diffuse interstellar bands (DIBs).8 In 2015 to 2016, 5 of the more than 500 (unassigned) DIBs known today were confirmed to be C+60 by laboratory gas-phase spectroscopy of C+60·He complexes.1,2,9 The very recent investigation of C+70·He complexes9 has provided accurate gas-phase values for parts of the C+70 absorption spectrum. C+70 should be also present in the diffuse interstellar medium because C70 is the second most abundant fullerene product after C60 in various physical processes, leading to the formation of fullerenes. For astronomy, the gas-phase absorption wavelengths are of great importance, allowing a direct comparison of laboratory and astronomical data. Matrix data can generally be more easily obtained than gas-phase measurements, provide information over a broad spectral region, give hints where to search for gas-phase absorptions, and also provide information about spectral regions inaccessible by gas-phase spectroscopy. Despite C70 being a commercially available fullerene, experimental infrared spectra of its ions in an inert environment have remained unknown up to now. IR spectra of fullerene ions are particularly important for astronomy because they give an opportunity to search for ionized fullerenes in objects different from the diffuse interstellar medium. For species with unknown IR absorptions or emissions, astronomers frequently use IR spectra simulated on the basis of (well-fitting) DFT calculations; however, this approach fails in the case of C±70, as we show in the current work. We present here IR data for C+70, C−70 and revise the partially existing information for UV-NIR absorptions in rare gas matrixes. Our measurements also uncover a new C+70 feature of potential relevance for the DIBs spectral region. We recommend to correspondingly extend the gas-phase spectroscopy measurement range. © 2016 American Chemical Society

EXPERIMENTAL METHODS A description of the experimental setup and measurement procedure has recently been published.10 In brief, C+70 ions were produced in an electron impact ionization source, mass-selected by a RF quadrupole filter, and codeposited with neon, argon and neon mixtures Ne+10%I2, Ne+1%CO2, and Ne+0.15%CCl4 onto a gold mirror substrate held at 5 K. Ion current was ∼100 nA during deposition. In addition to the CO2 and CCl4 electron scavengers, we have now also used I2. These molecules capture secondary electrons emitted from bombarded by the mass-selected cations metal surfaces along the incident ion beam path, forming counterions as a result. I2 as an electron scavenger simplifies the analysis of the infrared spectra. I2 monomers do not absorb IR light; however, I2 dimers and clusters do absorb at about 207 and 204 cm−1 (Figure 1) for symmetry reasons. Weak absorption of I−2 (I2)n at ∼116 cm−1 also does not complicate the analysis. The I−2 vibrational frequency was found to be close to the value of 110 ± 2 cm−1 previously measured by femtosecond photoelectron spectroscopy in the gas phase.11 The feasibility of I2 addition was initially tested with C+60 ions in Ne+10% I2 matrix. Despite the rather high amount of I2, no difference between C±/0 60 IR absorptions in pure neon and in neon diluted with molecular iodine was observed (see Supporting Information). The Ne+10% I2 mixture was empirically found to be suitable for the almost complete elimination of C−60 and its corresponding IR and NIR absorptions. The ratio of matrix gas (Ne 5.0, Ar 5.0, Air Liquide) to guest fullerene was about 5000:1 in all experiments. IR spectra were measured using a Bruker IFS66v/S FTIR spectrometer equipped with a CuGe-detector (6000−380 cm−1), a Bolometer (1000−100 cm−1), and an MCT-detector (6000− 500 cm−1). A Bolometer detector is capable also of measurements below 100 cm−1, but this region is dominated by distortions induced by residual mechanical vibrations associated with the Received: June 20, 2016 Revised: June 28, 2016 Published: July 5, 2016 5868

DOI: 10.1021/acs.jpca.6b06212 J. Phys. Chem. A 2016, 120, 5868−5873

Article

The Journal of Physical Chemistry A

Figure 1. Far-IR spectrum of C+70 deposited into Ne+10% I2 matrix. A red star indicates possible C+70 absorption. Broad interference features in the spectrum originate from the ice layer made of rest gas; these features are sample-specific and different in other samples.

Figure 2. Extracted C+70 absorption bands (red) were obtained by subtracting absorptions of neutral C70 from the spectrum of C+/0 70 in Ne+10%I2 matrix; for simplicity, spectral regions between absorptions were substituted by straight lines. Calculated (DFT BP86/def2SV(P)) IR spectra of doublet C70 in Cs and C1 symmetries are shown in blue and black. Overlapping experimental bands are fitted by Lorentzians (dark red). Tentatively assigned C+70 absorptions are shown in green. Intensities of theoretical C+70 in C1 have been scaled by factor two relative to C+70 in Cs.

coldhead, used to cool the matrix substrates. Relative to the previously described setup we have upgraded the FTIR spectrometer with a wedged diamond beam splitter to allow for a significantly broader spectral measurement range (diameter 57 mm, Diamond Materials). Diamond vacuum viewports for an infrared beam were also installed (free aperture 20 mm, Diamond Materials). Details of the upgrade will be published elsewhere. For all measurements the resolution was set to 0.25 cm−1. For Bolometer and CuGe measurements 600 and 2400 scans were averaged correspondingly. UV−vis measurements (200−1100 nm) were performed using a Princeton Instruments SCT-320 monochromator equipped with a PIXIS 256 OE CCD camera and halogen or deuterium light sources. The analysis of the spectra is as follows. Samples of C+70 deposited into pure neon contain various charge states of C70: C+70, C−70, and C70. Absorption lines, which considerably decrease after the addition of an electron scavenger, were attributed to C−70. The C+70 and C70 remaining absorptions can be distinguished from each other because we have measured a reference spectrum of neutral C70, sublimed from a Knudsen cell (see Supporting Information). The I2 effusion source was contaminated by trace amounts of ethanol and acetone, which are IR-active; the corresponding impurity absorption lines are shown in reduced contrast (Supporting Information, IR spectra). C+70 was also deposited into an Ar matrix. Obtained IR absorption frequencies of C±70 were compared with the absorptions in Ne. Gas-phase frequencies can be estimated by the extrapolation Ar → Ne → gas phase, assuming linear dependence on the polarizability of environment.5



COMPUTATIONAL DETAILS DFT quantum-chemical calculations were carried out using the Turbomole software package.12 Geometry optimization and vibrational analysis were done at the RI-DFT BP86/def2-SV(P) level of theory. First, the geometries of C±70 were optimized without symmetry restrictions in C1. The optimized geometry was analyzed to have some symmetry and was optimized again within the corresponding symmetry restriction. The symmetry of the lowest energy structure of C−70 was determined as C2v. The ground-state symmetry of C+70 is probably C5v. Unfortunately Turbomole12 fails to provide proper electronic occupation in this case. Therefore, we reduced symmetry in further calculations to Cs. The symmetry lowering from neutral C70 (D5h) is due to a static Jahn−Teller effect. Harmonic frequencies were calculated for doublet C+70 (Cs), C+70 (C1) as well as for doublet C−70 (C2v), C−70 (C1). Changing the functional from BP86 to B3LYP does not cause significant changes to predicted IR spectra. The same is true for an increase from def2-SV(P) to def2-TZVP basis sets. TDDFT calculations were performed to simulate electronic absorption spectra (RPA singlet excitations for singlet C70 and UHF RPA excitations for C+70 and C−70).12



RESULTS AND DISCUSSION IR Spectrum. Absorptions of C+70 in the 1200−1600 cm−1 range corresponding to vibrations tangential to the cage surface 5869

C+70

DOI: 10.1021/acs.jpca.6b06212 J. Phys. Chem. A 2016, 120, 5868−5873

Article

The Journal of Physical Chemistry A

Table 1. Identified Absorption Lines of C+70 and C−70, Frequency in Ar and Ne, fwhm, and Integrated Molar Absorptivity, ANe, Measured in Ne Matrixesa νAr (cm−1)

νNe (cm−1)

fwhm(Ne) (cm−1)

ANe (km·mol−1)c

C+70 1540

1129.3

709.3

463

1541 1526 1509.5 1443 1289 1272 1129 745 733 711 618 598 510b 463.7 420b 321b 317b

9 19.5 20 21 19 10 2.2 5 4 2 9 17 5 3.5 2.5 3.5 3.5

19 41 57 39 42 12 5 6 2 1 7 12 4 7 5 3 2

19.7 20.2 8.7 6.8 5.9 8.5 41.4 17.6 23.9 159.1 149.4 8.5 0.7 1 1.4 35.9 28.9

98 55 80 70 105 95 99 55 75 1070 558 74 11 24 14 179 189

C−70

Figure 3. Density of electronic states of C±/0 70 in the vicinity of HOMO−LUMO. Energy levels of C+70 and C−70 were shifted by +2.95 and −2.95 eV, respectively. α and β designate α and β electrons. Unoccupied orbitals are shown in red. Blue arrows indicate the dominant contributions to the lowest in energy electronic transition calculated by the TDDFT. Symmetries of electronic states and orbitals involved in transition are also shown.

1507 1479 1409 1401 1389 1323.1 1240.8

are broad and consist of several overlapping bands (Figure 2). A strong absorption at 1025 cm−1, predicted by the calculations, is absent in the experimental data. Obviously, there is no acceptable agreement between the DFT calculated IR spectra and the experiment. TDDFT calculations predict lowlying electronically excited states at 1086 cm−1 (12A″ ← X2A″) and 1271 cm−1 (22A″ ← X2A″) for C+70 in Cs symmetry. These values are close to the vibrational energies. The density of C±/0 70 electronic orbitals and dominant orbital contributions to the lowest-in-energy electronic transitions predicted by TDDFT are shown in Figure 3. Apparently, one encounters a nonBorn−Oppenheimer case. Therefore, it is in fact questionable if one can expect agreement between DFT calculations and the experiment in this case. The predicted infrared lines of radial to the cage surface vibrational modes (below 750 cm−1) also display a relatively poor resemblance to the experiment. Our calculations agree with already reported IR spectra of C±70 obtained by DFT,13 but, as one can see now, this standard DFT approach fails. Such a behavior is particularly remarkable if one considers the almost perfect coincidence of experiment 10,14 and DFT calculations for C±/2+ and neutral C70. A detailed 60 + overview of C70 spectra in different matrixes can be found in the Supporting Information. The estimated experimental absolute absorption intensities are summarized in Table 1. C−70 IR Spectrum. IR features of C−70 are broad and complex, which may also be related to electronic/vibrational interactions (Figure 4). For C−60 one does not observe such a complex spectrum.10 In the case of C−70, apparently, one encounters some kind of a distortion. Similar to C+70, our TDDFT calculations predict low-lying electronically exited states for C−70 at 1177 cm−1 (12A2 ← X2B1) and 1553 cm−1 (22A2 ← X2B1). Therefore, it is likely that harmonic analysis and corresponding

724.8 693.5 563.4 533.6 488

1505.8 1479.1 1409.5 1400.7 1388.4 1383.0 1324.1 1239.5 1097.8 1031.3 916.9 722.1 694.3 562.5 532.6 518.7 488.0

a

Not all absorptions could be clearly identified in Ar matrixes because only mixed samples C0/± 70 in Ar were investigated. Only the strongest absorptions of C−70 are presented; the full list of the absorptions can be found in the Supporting Information. bTentative assignment. cRelative accuracy of ANe is ∼40%;

theoretical IR intensities do not provide an adequate description of the experimental data in this spectral range. A predicted strong vibrational mode at 1120 cm−1, interacting with the close-lying electronic level, may correspond to the experimental broad band at ∼1000 cm−1. Our DFT calculations also do not describe the experiment well in the region of radial vibrations (below 750 cm−1). The IR spectrum of C−70 is much richer and more intense in comparison with C+70. DFT also predicts stronger absorptions for C−70 than for C+70. In Figure 4, the intensity of the calculated C−70 spectrum has been divided by five relative to the DFT spectrum of C+70 (Cs) in Figure 2. The experimental frequencies and estimated absolute absorption intensities can be found in Table 1. C+70 Electronic Spectrum. C+70 electronic absorptions in the gas phase9 and in the Ne matrix15 have been reported. We extend that data here and compare our new matrix data with a recent gas-phase measurement.9 5870

DOI: 10.1021/acs.jpca.6b06212 J. Phys. Chem. A 2016, 120, 5868−5873

Article

The Journal of Physical Chemistry A

Figure 4. C−70 absorption spectrum in Ne matrix (blue) was obtained by subtracting absorption spectra of C70 and C+70 from the spectrum of −1 was omitted because C±/0 70 in Ne. The region from 1415 to 1450 cm of imperfect subtraction of neutral C70 absorption. Overlapping experimental bands are fitted by Lorentzians (dark red). Calculated (DFT BP86/def2-SV(P)) IR spectrum of doublet C−70 in C2v (black).

Figure 5. UV-NIR spectrum of three different samples: neutral C70 in Ne (black), C70 ions in Ne (blue), and C70 ions in Ne+1%CO2 (red). To better recognize C±70 ionic absorptions, spectra were scaled for equal heights of the narrow absorption peaks of neutral C70 in the 550−650 nm region. Red “+” signs indicate absorptions of C+70.

In addition to the previously observed series of vibronic features at ∼760 nm one sees also a broad and weaker absorption at 680 nm (Figures 5,6). This region was not considered by the previous matrix and gas-phase measurements. The relatively strong absorptions at 12 842 and 12 928 and a weak one at 12 682 cm−1, previously assigned to C+70,15 are not present in the Ne+1%CO2 and Ne+10%I2-matrixes, containing mostly C+70 and C70 (see Figure 6, blue vertical lines). Therefore, these absorption lines do not belong to C+70; however, all of these absorptions, are present in the Ne matrixes, which also contain considerable amount of C−70. Another strong absorption at 14 196 cm−1, which was also previously assigned to C+70,15 belongs to C2+ 70 and will be considered in detail in a future publication about C2+ 70 vibrational and electronic properties. TDDFT calculations for C0/± 70 predict very similar absorptions in the UV−vis range (see Supporting Information). Therefore, the remaining UV−vis absorption features of C+70 are probably masked by the C70. A comparison of the gas-phase C+70 absorption spectrum with the spectra measured in Ne+10%I2 and Ne+1%CO2 matrixes, mostly containing C+70, is presented in Figure 7. In contrast with C+60, the strongest NIR absorptions of C+70 in the matrixes have almost the same positions as in the gas phase. This suggests a weaker interaction of C+70 with solid matrix environment than in the case of C+60. There is an absorption of C+70 at 792.7 nm in the region not considered by the gas-phase measurement. Another absorption line at 743.6 nm is missing in the gas-phase spectrum.

Figure 6. Background continuum subtracted NIR absorption spectra of C±/0 70 in Ne+1%CO2, Ne+10%I2, Ne matrixes at 5 K. Intensities are scaled to be similar for C+70. Absorptions of C−70 are indicated by vertical blue lines. 5871

DOI: 10.1021/acs.jpca.6b06212 J. Phys. Chem. A 2016, 120, 5868−5873

Article

The Journal of Physical Chemistry A

UV−vis features to be quite similar to those of C70. Absorption bands, attributed to C−70 in amine solvents at ∼820 nm,17 cannot be observed in neon matrix. Another known strong C−70 NIR band15,18 at 7342 cm−1 in Ne is also present in our FTIR spectra.

We expect that the not well-resolved weaker absorption series of C+70 with maximum at 680 nm is also similar to the gas-phase positions. From the total number of deposited C+70 ions and relying on previous C+60 measurements,5 we estimated oscillator strengths of C+70 with an accuracy of ∼50% by integrating over the individual wavelength ranges; see Table 2. To compare these

781.2 739−774 645−710 C−70 773.7 778.2 788.9 a

f Ne (nm)

fgasa

1.7 ± 0.8 × 10−4 1.5 ± 0.7 × 10−3 8 ± 4 × 10−4

1.6 ± 0.7 × 10−4 2.0 ± 0.8 × 10−3



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpca.6b06212. IR absorption spectra of C0/± 60 in Ne+10%I2 and Ne matrixes; IR absorption spectra of C0/± 70 in Ar and Ne matrixes; IR absorption spectra of C0/± 70 in Ne, Ne+10%I2 matrixes vs DFT C±70 IR spectra; IR absorption spectra of 0/± in Ne, Ne+10%I2, Ne+1%CO2, Ne+0.1%CCl4 C70 matrixes; Full list of the C−70 identified IR absorption lines in Ne matrix; IR absorption spectra of neutral C70 in Ne matrix vs DFT; NIR spectra of C0/± 70 in Ne, Ne+1% CO2, Ne+10%I2 matrixes; and electronic absorption spectra of ionized and neutral C70 as predicted by TDDFT calculations. (PDF)

Table 2. Absorption Oscillator Strengths f of C+70 in Ne+10% I2 Matrix, Gas Phase, and C−70 in Ne Matrix C+70

CONCLUSIONS

We codeposited mass-selected C+70 ions and neon with/without electron scavengers on cryogenic substrates and obtained sufficient amounts of matrix-isolated ions to enable IR and UV−vis absorption spectroscopy. In contrast with C±/2+ and C70, IR 60 spectra of C±70 cannot be well-fitted by the DFT calculations, most probably due to electronic−vibrational coupling in these species. Higher level non-Born−Oppenheimer calculations would clearly be of interest to better understand the experimental results, which could also serve as benchmarks for the development of such theoretical methods. Because C0/± 70 IR absorptions have very similar positions in Ar and Ne matrixes, we do not expect considerable shifts of gas-phase vibrational frequencies relative to the matrix data. Consistent with C0/± 60 , the electronic spectra of C70, C+70, and C−70 in the UV are similar and dominated by plasmonlike absorptions. Previously published matrix isolation data for C+70 in the visible range were extended and revised. Comparison of the C+70 electronic transitions in the matrix and in the gas phase revealed almost identical absorption positions. We suggest to extend a measurement range of the gas-phase spectroscopy to the 645−710 nm region with another series of C+70 absorptions.

Figure 7. NIR spectra C+70 ions in Ne+1%CO2 (red) and Ne+10%I2 (magenta) versus C+70 gas-phase absorptions9 (blue vertical lines). Absorption intensities were scaled for better comparison.

λNe (nm)



2.1 ± 1 × 10−4 1.4 ± 0.7 × 10−4 6 ± 3 × 10−5



Derived from the absorption cross sections and fwhm9

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: +49-72160847373; +49-72160843310.

values with the gas-phase absorption cross sections, we used relationships from ref 16. C−70 Electronic Spectrum. As was already pointed out in the previous section, absorptions at 12 842, 12 928, and 12 682 cm−1 are present in the spectrum only if the sample contains considerable amount of C−70, as in the case of pure Ne matrix (see Figure 6 and Supporting Information). Therefore, we reassign these absorptions to C−70. Assuming an overall neutral charge of the Ne matrix, containing C0/± 70 , we estimated also oscillator strengths of these absorptions; see Table 2. We do not observe electronic UV−vis absorptions of C−70 differing from those of C70. Again, taking into an account the TDDFT predictions (see Supporting Information), we would expect most of the C−70

Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We thank Prof. M. M. Kappes for useful comments and fruitful discussions, Prof. R. Ahlrichs for valuable comments regarding theory, Mr. Waltz and his mechanical workshop for hardware construction, and K. Stree and H. Halberstadt for the help in electronic issues. This work was supported by the Deutsche Forschungsgemeinschaft (KA 972/10-1). We also acknowledge support by KIT and Land Baden-Württemberg. 5872

DOI: 10.1021/acs.jpca.6b06212 J. Phys. Chem. A 2016, 120, 5868−5873

Article

The Journal of Physical Chemistry A



REFERENCES

(1) Campbell, E. K.; Holz, M.; Gerlich, D.; Maier, J. P. Laboratory Confirmation of C+60 as the Carrier of Two Diffuse Interstellar Bands. Nature 2015, 523, 322−323. (2) Walker, G. A. H.; Bohlender, D. A.; Maier, J. P.; Campbell, E. K. Identification of More Interstellar C+60 Bands. Astrophys. J., Lett. 2015, 812, L8. (3) Cami, J.; Bernard-Salas, J.; Peeters, E.; Malek, S. Detection of C60 and C70 in a Young Planetary Nebula. Science 2010, 329, 1180−1182. (4) Omont, A. Interstellar fullerene compounds and diffuse interstellar bands. Astron. Astrophys. 2016, 590, A52. (5) Strelnikov, D.; Kern, B.; Kappes, M. M. On observing C+60 and C2+ 60 in laboratory and space. Astron. Astrophys. 2015, 584, A55. (6) Dennler, G.; Scharber, M. C.; Brabec, C. J. Polymer-Fullerene Bulk-Heterojunction Solar Cells. Adv. Mater. 2009, 21, 1323−1338. (7) Liu, Y.; Zhao, J.; Li, Z.; Mu, C.; Ma, W.; Hu, H.; Jiang, K.; Lin, H.; Ade, H.; Yan, H. Aggregation and morphology control enables multiple cases of high-efficiency polymer solar cells. Nat. Commun. 2014, 5, 5293. (8) Snow, T. P.; McCall, B. J. Diffuse Atomic and Molecular Clouds. Annu. Rev. Astron. Astrophys. 2006, 44, 367−414. (9) Campbell, E. K.; Holz, M.; Maier, J. P.; Gerlich, D.; Walker, G. A. H.; Bohlender, D. Gas Phase Absorption Spectroscopy of C+60 and C+70 in a Cryogenic Ion Trap: Comparison with Astronomical Measurements. Astrophys. J. 2016, 822, 17. (10) Kern, B.; Strelnikov, D.; Weis, P.; Böttcher, A.; Kappes, M. M. IR Absorptions of C+60 and C−60 in Neon Matrixes. J. Phys. Chem. A 2013, 117, 8251−8255. (11) Zanni, M. T.; Taylor, T. R.; Greenblatt, B. J.; Soep, B.; Neumark, D. M. Characterization of the I−2 anion ground state using conventional and femtosecond photoelectron spectroscopy. J. Chem. Phys. 1997, 107, 7613−7619. (12) TURBOMOLE V6.4 2012, a development of University of Karlsruhe and Forschungszentrum Karlsruhe GmbH, 1989−2007, TURBOMOLE GmbH, since 2007; available from http://www. turbomole.com. (13) Bauschlicher, C. W. The infrared spectra of nonplanar polycyclic aromatic hydrocarbons with five- or seven-membered rings. Chem. Phys. 2015, 448, 43−52. (14) Kern, B.; Strelnikov, D.; Weis, P.; Böttcher, A.; Kappes, M. M. 3+ IR, NIR, and UV Absorption Spectroscopy of C2+ 60 and C60 in Neon Matrixes. J. Phys. Chem. Lett. 2014, 5 (3), 457−460. (15) Fulara, J.; Jakobi, M.; Maier, J. P. Electronic spectra of the C70 molecule and C+70, C−70 ions in neon matrices. Chem. Phys. Lett. 1993, 206, 203−209. (16) Hilborn, R. C. Einstein coefficients, cross sections, f values, dipole moments, and all that. ArXiv Physics e-prints, 2002. (17) Cataldo, F.; Iglesias-Groth, S.; Manchado, A. On the Radical Anion Spectra of Fullerenes C60 and C70. Fullerenes, Nanotubes, Carbon Nanostruct. 2013, 21, 537−548. (18) Hase, H.; Miyatake, Y. Electronic spectra of C70 anions produced in γ-irradiated organic glasses at 77 K. Chem. Phys. Lett. 1993, 215, 141−143.

5873

DOI: 10.1021/acs.jpca.6b06212 J. Phys. Chem. A 2016, 120, 5868−5873