Gas-phase electronic spectrum of the C₁₄ ring

Abstract

The electronic spectrum of a cyclic C₁₄ in the visible range has been detected in the gas phase by a mass selective resonant two-color two-photon ionization technique coupled to a laser ablation source. Absorption is localized in the 19 000 to 20 000 cm⁻¹ region and appears as a dozen of 3–7 cm⁻¹ narrow peaks belonging to one or two close-lying electronic states. Bands have structures which for the narrowest ones is likely to be the rotational profile contour. The spectrum is attributed to a cyclic form of C₁₄ based on time-dependent density-functional calculations and reactivity with H₂. The spectral pattern differs from that previously seen in the larger C_4n+2 member rings, C₁₈ and C₂₂, indicating some sort of a structural crossover.

---

1. Introduction

Pure carbon rings attract attention for several reasons. Apart from their established relation to fullerene formation^1,2 there are other intriguing aspects, the first of which is their expected importance in space chemistry. Rings become interesting in light of the discussions concerning polyaromatic hydrocarbons (PAH) as possible diffuse interstellar bands (DIB) carriers. This is because many PAH cations are inherently unstable to photolysis.^3–5 The photostability for two dozen PAH cations has been tested and it fell into four groups: stable, hydrogens loss only, also carbons loss, and destroyed.⁵ Another study concentrated on coronene cation, C₂₄H₁₂⁺, and indicates that depending on the photon flux, partial or even complete dehydrogenation may occur.⁴ Possibly produced in the circumstellar envelopes of carbon-rich red giants,⁶ PAHs spread out in interstellar space and become exposed to UV radiation, under which many of them may lose some, or all, of their hydrogens and fall apart. This results in a wealth of unsaturated skeletons that will be subject to ring-openings, ring-losses, and other reactions and will eventually end up in a variety of bare and partially hydrogenated chains, rings (monocyclic and polycyclic), PAHs (ionized, protonated, and partially dehydrogenated), and fullerenes. On the other hand, a good number of studies^7–13 on the fragmentation patterns of middle-sized (≈C_5–50) pure-carbon moieties (ionic and neutral) indicate losses of C₂ and C₃, and, though in much less copious amounts, C₅, C_10,14,18,22 (see ref. 14 and references therein). This may suggest that similar processes may occur with dehydrogenated PAHs in space.

Another aspect is that the 4n + 2 monocyclic rings seem to be suitable model systems for testing the handling of electron correlation in quantum chemical methods. These closed-shell molecules manifest a competition between electronic delocalization, second-order Jahn–Teller, and, at large sizes, Peierls instability,¹⁵ each of which favors different structures (Fig. 1). The real structure is a balance between these effects, which tends to vary with the size of the ring. Already the aromatic ring polyenes (C_4n+2H_4n+2) are known to have equal CC bond lengths at small sizes (C₆H₆), but exhibit alternation after 14 carbons,^16–18 caused by a distortive propensity of the π systems.¹⁹ Cyclic polyynes have even a greater tendency than polyenes to adopt bond-alternant forms,²⁰ which could be rationalized by the greater stability of acetylenes as compared to allenes.²¹ Therefore studying the effects occurring in 4n + 2 pure-carbon rings is of considerable interest to the chemistry and physics of extended π-systems and is important for understanding how electronic structures of finite systems approach those of infinite periodic ones.

Fig. 1 Cyclic isomers of C₁₄. 1 cumulenic, fully symmetric; 2 cumulenic, bond-angle alternating, halved symmetry, imposed by second-order Jahn–Teller effect; 3 intermediate structure between isomers 2 and 4; isomer 4 acetylenic, bond length alternating as a result of Peierls type instability.

Monocyclic C₁₄ has been detected in mobility experiments on both anions²² and cations.^11,23,24 It was also identified via anion UV photoelectron spectroscopy^25,26 where a relatively low electron affinity and a prominent energy gap between the ground and first excited electronic states of the corresponding neutral were detected. A direct absorption study of mass-selectively deposited and neutralized C₁₄ anions in the Ne matrix revealed an electronic absorption band near 347 nm, which was tentatively assigned to a D_14h structure.²⁷ A fluorescence survey of neutralized C₁₄ cations in a N₂ cryogenic matrix showed a strong and broad signal between 520 and 700 nm which was also assigned to the cyclic isomer.²⁸ The work presented here is the first observation of the gas-phase electronic spectrum of this molecule.

2. Observations and discussion

C₁₄ was studied experimentally using the same approach as in the measurement of the ring species C₁₈ and C₂₂.²⁹ The molecules were produced using laser ablation of a rotating graphite rod. A 5–10 bar buffer gas pulse, coinciding with the laser, was expanded through the body of the source over the ablation spot and extracted the cloud of carbonaceous products into the vacuum chamber, cooling it in an adiabatic expansion. Any resulting ions were removed after the skimmer, and before entering the pulsed extraction zone of the time-of-flight (TOF) mass spectrometer, by an electric field perpendicular to the molecular beam. The neutral species were then ionized in a two-step REMPI process and ions were extracted into the TOF tube. The tunable excitation photons were delivered by an OPO system (≈5 mJ per 7 ns pulse, 0.2 cm⁻¹ linewidth), with the ionization step provided by a fluorine laser (7.9 eV, ∼1 mJ per 10 ns pulse).

Fig. 2 shows the electronic excitation spectrum of C₁₄ from 18 250 to 20 750 cm⁻¹. It consists of about 10 distinct peaks (Table 1), 3–7 cm⁻¹ broad, with substructure. At least some of them consist of components (Fig. 3). Contrary to the case of C₁₈,²⁹ this substructure can be described as a dip within a broad band rather than as being a convolution of Lorentzian peaks. Neither can this splitting be an artefact due to accidental coupling of a “bright state” with a “dark state”, as this would be isotope dependent (¹²C₁₄versus¹²C₁₃¹³C). Instead, the spectra of the two isotopomers appear similar with a small isotope shift (Fig. 3). Perhaps it is a rotational band profile complicated by Coriolis couplings or a set of overlapping sequence bands. The vibronic structure does not feature any obvious harmonic progressions. Assigning the first few bands as a progression in the lowest energy bending vibration would result in a flat and nonharmonic excited state potential surface as the vibronic spacings decrease toward the blue.

Fig. 2 Gas-phase electronic spectrum of C₁₄ detected by a resonant two-color two-photon ionization technique under supersonic molecular beam conditions. The broad background is due to sequence bands as its extent varies with source conditions. Band 1 appears to be the origin.

Fig. 3 Close-up of band 1 of mass-separated ¹²C₁₄ (m/z = 168 amu) and its isotopomer ¹²C₁₃¹³C (m/z = 169 amu) both showing non-trivial and profiles.

Table 1 Maxima of the prominent vibronic bands observed in the electronic spectrum of C₁₄, their positions relative to the observed origin, and the isotopic shifts of the corresponding absorptions for ¹³C¹²C₁₃ with respect to ¹²C₁₄

Label	λ air/nm ^a	Δν/cm^−1	(νi − ν)/cm^−1
a The positions of spectral features are the maxima of peaks or, for those double headed, as center of the dip. b This and further values are identical to those for the ^12C14 isotope as their shifts can not be defined with adequate precision because of their broadness.
1	521.23	0 (19180.2)	+0.8
2	519.58	61	+0.4
3	518.61	97	≈ +0.25
4	517.99	120	≈ +0.15
5	516.84	163	≈0^b
6	515.67	207
7	511.86	351
8	509.49	442
9	507.22	530
10	505.50	597

---

Although the natural isotope abundance of ¹³C is only 1.1%, there is a ca. 13.4% probability that a C₁₄ species will have one heavy atom. Unit resolution in the mass spectrometer allowed the separation of m/z 168 and 169, and the laser resolution (ca. 0.15 cm⁻¹) permitted the observation of the isotopic shifts of the first few vibronic lines in the spectrum of the C₁₄. Thus one sees band 1 in the spectrum of the heavy ¹²C₁₃¹³C isotopomer shifted ca. 0.8 cm⁻¹ to the blue compared to that in ¹²C₁₄ (Fig. 3), and the next few lines by ca. 0.4 cm⁻¹, ca. 0.25 cm⁻¹, and ca. 0.15 cm⁻¹. This decrease of the blue shift in the vibronic bands relative to that of the origin can be understood from a regular increase in the vibrational red shift. The latter seems reasonable, as the vibrational energy lowers upon increasing the mass of an oscillator. The blue isotopic shift of the origin of the electronic transition can in turn be rationalized as due to a more shallow excited state potential surface. In this case the isotopic substitution lowers the energy levels in the deeper and narrower ground potential more than in the shallower excited state so that there is an overall increase of the transition energy. Relaxation upon molecule excitation is common as an electron moves from a bonding to a non-bonding molecular orbital, causing the molecule to swell slightly in the upper state.

C₁₄ falls into the same 4n + 2 series as cyclic C₁₈ and C₂₂ do. Analogous to those cases, C₁₄ is an abundant graphite ablation product (as could be seen in mass spectra of neutral products ionized with 10.5 eV photons) which also demonstrates a low reactivity with respect to hydrogen.²⁹ And, according to density functional theory (DFT) calculations (B3LYP/6-31G(d)), the ground state prefers a cyclic form by about 3 eV over the linear one. Therefore the expected structure of C₁₄ is cyclic.

However, the REMPI spectrum (Fig. 2) of C₁₄ differs in appearance from those of C₁₈ and C₂₂.²⁹ Located solely in the visible region around 520 nm, it contains a dozen lines that could be attributed to just one or two close lying electronic transitions. The larger family members showed absorptions features in the whole studied region (up to 36 900 cm⁻¹) starting from the origin (16 862 cm⁻¹ for C₁₈) and becoming increasingly intense in the UV.²⁹ Such a change in the spectral pattern from C₂₂ and C₁₈ to C₁₄ is striking and probably indicates a structural change upon decreasing the size of the doubly aromatic ring.

The small number of allowed electronic transitions suggests that the carrier has high symmetry: thus, for instance, a D_14h ring 1 (Fig. 1) has fewer allowed transitions than a ring with D_7h (2, 4) or C_7h (3). However calculations do not place this structure as an energy minimum for any of the C_4n+2 monocyclic compounds.^{15,20,21,30,31} Instead, they indicate it as a transition state for the automerization of the less symmetrical cumulenic isomer 2.

Literature suggests that there is another structural “crossover” expected in these systems. Upon increasing the size of the ring the aromatic delocalization is overruled¹⁵ by the distortive propensity of the π systems to electron-localization known as Peierls instability³² (or “Kekuléan distortion”¹⁹). As a result smaller rings would adopt cumulenic structures equivalent to 2, whereas “larger” rings should acquire the acetylenic structures 4 (or 3) at some point. The theoretical approaches account for these many-body effects variously and this turning point is predicted at different ring sizes. The results of available calculations are summarized in Table 2. The Hartree–Fock (HF) method, known to render a less homogeneous structure in order to increase the impact of exchange effects,^15,21,33,34 predicts a bond length alternating geometry for rings larger than C₁₀. Conversely, DFT generally “overshoots” the effect of correlation.^15,21 Thus the local density approximation (LDA) adheres to the aromatic structure 2 at least until C₈₂.^15,31 For a system as large as C_14,18,22 the more elaborate ab initio methods can not be applied. The hybrid DFT and quantum Monte Carlo (QMC) with their ability to account for electron correlation and favorable scaling with the number of particles, could be sensibly employed. Table 2 demonstrates that the calculations available generally agree with the presence of a crossover and place it somewhere within the C₁₀–C₂₂ interval. Yet it is not clear how to correlate the observed evolution of the spectral pattern upon decreasing the size of the monocyclic rings with the expected structural change. Perhaps one can draw a conclusion when the electronic spectrum of the next family member, C₁₀, becomes available. The latter has so far escaped spectroscopic detection for no clear reasons (tentatively attributed to its high IP). The presented electronic spectrum of C₁₄ shows that a detailed theoretical treatment for both ground and excited electronic states is needed to solve this challenge.

Table 2 Calculated ring size at which a structural change from cumulenic (1, 2) to polyynic (3, 4) regime occurs. Crossover type refers to transitions between structures specified in Fig. 1

	Method	Crossover size	Crossover type	Reference
a DFT energies corrected to the experimental relative energies for acetylene and cumulene model compounds.
1	HF/cc-pVDZ	C10–C14	2 → 4	36
2	LDA	>C82	2 → 3 (or 4)	31
3	Corrected DFT^a	C14	2 → 3	21
4	hybrid DFT/cc-pVDZ	C18	2 → 3	30
5	B3LYP/6-31+g(d,p)	C18–C22	2 → 4	34
6	QMC	C10–C14	2 → 4	15

---

An interesting aspect concerning the detection of the gas phase spectrum of the C₁₄ cyclic ring is that the peak positions can be compared to DIB observations. This was done for the apparent origin band of the system at 521.23 nm, as well as for the other prominent peaks. In the available DIB compilations no evident features are given there. This was also the case for cyclic C₁₈³⁵ and a similar conclusion can be drawn: as the oscillator strength of the transition is rather low, the species would have to be present in concentrations ≳10¹² cm⁻² to lead to discernible features with present DIB detection limits.

Acknowledgements

This work has been supported by the Swiss National Science Foundation (200020-107386) and the European Union project “Molecular Universe” (MRTN-CT-2004-512303). We wish to thank Prof. Michael Morse for helpful discussions and Prof. Pavel Rosmus for advice concerning the calculations.

References

1. G. von Helden, N. G. Gotts and M. T. Bowers, Nature, 1993, 363, 60–63.
2. S. W. McElvany, M. M. Ross, N. S. Goroff and F. Diederich, Science, 1993, 259, 1594–1596.
3. P. Boissel, G. Lefevre and P. Thiébot, in AIP Conf. Proc. 312: Molecules and Grains in Space, ed. I. Nenner, AIP, New York, 1994, pp. 667–674.
4. S. P. Ekern, A. G. Marshall, J. Szczepanski and M. Vala, Astrophys. J., 1997, 488, L39–L41.
5. S. P. Ekern, A. G. Marshall, J. Szczepanski and M. Vala, J. Phys. Chem. A, 1998, 102, 3498–3504.
6. M. Frenklach and E. D. Feigelson, Astrophys. J., 1989, 341, 372–384.
7. M. B. Sowa, P. A. Hintz and S. L. Anderson, J. Chem. Phys., 1991, 95, 4719–4720.
8. M. B. Sowaresat, P. A. Hintz and S. L. Anderson, J. Phys. Chem., 1995, 99, 10736–10741.
9. P. P. Radi, T. L. Bunn, P. R. Kemper, M. E. Molchan and M. T. Bowers, J. Chem. Phys., 1988, 88, 2809–2814.
10. G. von Helden, N. G. Gotts and M. T. Bowers, J. Am. Chem. Soc., 1993, 115, 4363–4364.
11. J. Hunter, J. Fye and M. F. Jarrold, J. Phys. Chem., 1993, 97, 3460–3462.
12. R. L. Whetten, C. Yeretzian and P. M. Stjohn, Int. J. Mass Spectrom., 1994, 138, 63–76.
13. G. Martinet, S. Diaz-Tendero, M. Chabot, K. Wohrer, S. Della Negra, F. Mezdari, H. Hamrita, P. Desesquelles, A. Le Padellec, D. Gardes, L. Lavergne, G. Lalu, X. Grave, J. F. Clavelin, P. A. Hervieux, M. Alcami and F. Martin, Phys. Rev. Lett., 2004, 93, 063401.
14. A. E. Boguslavskiy and J. P. Maier, J. Chem. Phys., 2006, 125, 094308.
15. T. Torelli and L. Mitas, Phys. Rev. Lett., 2000, 85, 1702–1705.
16. C. H. Choi and M. Kertesz, J. Chem. Phys., 1998, 108, 6681–6688.
17. M. Kertesz, C. H. Choi and S. J. Yang, Chem. Rev., 2005, 105, 3448–3481.
18. C. S. Wannere, K. W. Sattelmeyer, H. F. Schaefer and P. V. R. Schleyer, Angew. Chem., Int. Ed., 2004, 43, 4200–4206.
19. P. C. Hiberty, G. Ohanessian, S. S. Shaik and J. P. Flament, Pure Appl. Chem., 1993, 65, 35–45.
20. C. X. Liang and H. F. Schaefer, J. Chem. Phys., 1990, 93, 8844–8849.
21. D. A. Plattner and K. N. Houk, J. Am. Chem. Soc., 1995, 117, 4405–4406.
22. G. von Helden, P. R. Kemper, N. G. Gotts and M. T. Bowers, Science, 1993, 259, 1300–1302.
23. G. von Helden, M. T. Hsu, P. R. Kemper and M. T. Bowers, J. Chem. Phys., 1991, 95, 3835–3837.
24. G. von Helden, N. G. Gotts and M. T. Bowers, J. Am. Chem. Soc., 1993, 115, 4363–4364.
25. S. Yang, K. J. Taylor, M. J. Craycraft, J. Conceicao, C. L. Pettiette, O. Cheshnovsky and R. E. Smalley, Chem. Phys. Lett., 1988, 144, 431–436.
26. Y. Achiba, M. Kohno, M. Ohara, S. Suzuki and H. Shiromaru, J. Electron Spectrosc. Relat. Phenom., 2005, 142, 231–240.
27. M. Grutter, M. Wyss, E. Riaplov, J. P. Maier, S. D. Peyerimhoff and M. Hanrath, J. Chem. Phys., 1999, 111, 7397–7401.
28. G. A. Rechtsteiner, C. Felix, A. K. Ott, O. Hampe, R. P. Van Duyne, M. F. Jarrold and K. Raghavachari, J. Phys. Chem. A, 2001, 105, 3029–3033.
29. A. E. Boguslavskiy, H. Ding and J. P. Maier, J. Chem. Phys., 2005, 123, 034305.
30. M. Saito and Y. Okamoto, Phys. Rev. B, 1999, 60, 8939–8942.
31. E. J. Bylaska, R. Kawai and J. H. Weare, J. Chem. Phys., 2000, 113, 6096–6106.
32. R. E. Peierls, Quantum Theory of Solids, Oxford University Press, 1955.
33. J. D. Watts and R. J. Bartlett, Chem. Phys. Lett., 1992, 190, 19–24.
34. S. Sen, P. Seal and S. Chakrabarti, Phys. Rev. B, 2006, 73, 245401.
35. J. P. Maier, A. E. Boguslavskiy, H. B. Ding, G. A. H. Walker and D. A. Bohlender, Astrophys. J., 2006, 640, 369–372.
36. J. M. L. Martin, J. El-Yazal and J. P. Francois, Chem. Phys. Lett., 1995, 242, 570–579.

---