Single and double ionization of corannulene and coronene

Article abstract of DOI:10.1002/1522-2675(20010613)84:6<1625::AID-HLCA1625>3.0.CO;2-0

Electron-transfer processes that involve single and doubly charged cations of corannulene (C₂₀H₁₀) and coronene (C₂₄H₁₂) are examined by three different mass-spectrometric techniques. Photoionization studies give first-ionization energies of IE(C₂₀H₁₀) = 7.83 ± 0.02 eV and IE(C₂₄H₁₂) = 7.21 ± 0.02 eV. Photoionizations of the neutrals to the doubly charged cations occur at thresholds of 20.1 ± 0.2 eV and 18.5 ± 0.2 eV for corannulene and coronene, respectively. Energy-resolved charge-stripping mass spectrometry yields kinetic energy deficits of Q_min(C₂₀H⁺₁₀ = 13.8 ± 0.3 eV and Q_min(C₂₄H⁺₁₂ = 12.8 ± 0.3 eV for the transitions from the mono- to the corresponding dications in keV collisions. Reactivity studies of the C₂₀H²⁺₁₀ and C₂₄H²⁺₁₂ dications in a selected-ion flow-tube mass spectrometer are used to determine the onsets for the occurrence of single-electron transfer from several neutral reagents to the dications, affording two different monocationic products. With decreasing IEs of the neutral reagents, electron transfer to doubly charged corannulene is first observed with hexafluorobenzene (IE = 9.91 eV), while neutrals with lower IEs are required in the case of the coronene dication, e.g., NO₂ (IE = 9.75 eV). Density-functional theory is used to support the interpretation of the experimental data. The best estimates of the ionization energies evaluated are IE(C₂₀H₁₀) = 7.83 ± 0.02 eV and IE(C₂₄H₁₂) = 7.21 ± 0.02 eV for the neutral molecules, and IE(C₂₀H⁺₁₀) = 12.3 ± 0.2 eV and IE(C²⁴H⁺₁₂) = 11.3 ± 0.2 eV for the monocations.

Full text of DOI:10.1002/1522-2675(20010613)84:6<1625::AID-HLCA1625>3.0.CO;2-0

Helvetica Chimica Acta ± Vol. 84 (2001)
1625
Single and Double Ionization of Corannulene and Coronene
by Detlef Schröder*a), Jessica Loosa), Helmut Schwarza), Roland Thissenb), Dorin V. Predac),
Lawrence T. Scottc), Doina Caraimand), Maxim V. Frachd), and Diethard K. Böhmed)
a) Institut für Chemie der Technischen Universität Berlin, Strasse des 17. Juni 135, D-10623 Berlin
b
Â
) Laboratoire de Chimie Physique, Bat. 350, Universite Paris-Sud, F-91405 Orsay
c) Merkert Chemistry Center, Boston College, Chestnut Hill MA 02467-3860, USA
d) Department of Chemistry, York University, 4700 Keele Street, Toronto, M3J 1P3, Canada
Dedicated to Edgar Heilbronner on the occasion of his 80th birthday
Electron-transfer processes that involve single and doubly charged cations of corannulene (C20H10) and
coronene (C24H12) are examined by three different mass-spectrometric techniques. Photoionization studies give
first-ionization energies of IE(C20H10) ˆ 7.83 Æ 0.02 eV and IE(C24H12) ˆ 7.21 Æ 0.02 eV. Photoionizations of the
neutrals to the doubly charged cations occur at thresholds of 20.1 Æ 0.2 eV and 18.5 Æ 0.2 eV for corannulene
and coronene, respectively. Energy-resolved charge-stripping mass spectrometry yields kinetic energy deficits of
‡
‡
Qmin(C20H10 ˆ 13.8 Æ 0.3 eV and Qmin(C24H12 ˆ 12.8 Æ 0.3 eV for the transitions from the mono- to the
2‡
2‡
12
corresponding dications in keV collisions. Reactivity studies of the C20H and C24H dications in a
10
selected-ion flow-tube mass spectrometer are used to determine the onsets for the occurrence of single-electron
transfer from several neutral reagents to the dications, affording two different monocationic products. With
decreasing IEs of the neutral reagents, electron transfer to doubly charged corannulene is first observed with
hexafluorobenzene (IE ˆ 9.91 eV), while neutrals with lower IEs are required in the case of the coronene
dication, e.g., NO2 (IE ˆ 9.75 eV). Density-functional theory is used to support the interpretation of the
experimental data. The best estimates of the ionization energies evaluated are IE(C20H10) ˆ 7.83 Æ 0.02 eV and
‡
‡
IE(C24H12) ˆ 7.21 Æ 0.02 eV for the neutral molecules, and IE(C20H10 ) ˆ 12.3 Æ 0.2 eV and IE(C24H12 ) ˆ 11.3 Æ
0.2 eV for the monocations.
Introduction. ± Polycyclic aromatic hydrocarbons (PAHs) are of environmental
relevance due to their toxicity combined with their ubiquity as combustion products of
organic material in fuel-rich flames [1]. Further, PAHs and their ions are believed to
contribute to interstellar absorption bands in the infrared [2a] (for criticism of this
hypothesis, however, see [2b]) [3]. For these and other reasons [4], the interest in
PAHs has a long tradition [5].
As far as first- and second-ionization energies of PAHs are concerned, Tobita et al.
provided a comprehensive survey for more than 20 different compounds, ranging from
benzene to buckminsterfullerene [6]. However, the data available for the dications are
not all unambiguous, and differences may spread over several eV. For example, the
appearance energy measurements for the tetracene dication compiled by Tobita et al.
range from 18.6 Æ 0.2 eVobtained with photoionization to 20.1 Æ 0.5 eV determined by
charge stripping and 22.1 Æ 0.2 eV with electron ionization. Such a magnitude of
uncertainties in the evaluation of second-ionization energies appears rather large, thus
justifying a re-evaluation that aims at a more accurate determination of this
fundamental property for two representative PAHs.
Here, we report the results of three different experimental approaches aimed at
determining the second ionization energies of corannulene (C20H10) and coronene

1626
Helvetica Chimica Acta ± Vol. 84 (2001)
(C24H12) (Fig. 1). These two molecules are chosen as prototypes for bowl-shaped
(corannulene) and planar (coronene) PAHs. The experimental methods probe the
electron-transfer behavior of these hydrocarbons from three different aspects. i)
Photoionization (PI) methods involve ionization of the neutral molecules to either
‡
mono- or dications, i.e., PAH ! PAH and PAH ! PAH2‡. ii) Charge-stripping (CS)
mass spectrometry deals with the electron detachment from mass-selected mono- to
dications in high-energy collisions, i.e., PAH ! PAH2‡. iii) Reactivity studies in a
selected-ion flow tube (SIFT) apparatus can inter alia probe single-electron transfer
(SET) from several neutral molecules M to the dications at thermal energies, i.e., M ‡
‡
‡
PAH2‡ ! M ‡ PAH . Further, molecular-orbital calculations are used to estimate the
differences due to vertical and adiabatic electron transfer from neutrals to mono- and
dications, respectively.
Fig. 1. Structures of corannulene C20H10 and coronene C24H12
Experimental and Theoretical Details. ± The experiments with molecular dications [7] were performed
with three different types of mass spectrometers, which are very briefly described below; for more detailed
descriptions, the readers are referred to the original references quoted.
The photoionization experiments used the CERISES apparatus [8] for the determination of the first- and
second-ionization energies of corannulene and coronene. CERISES was installed on the beam line SA63 of the
synchrotron radiation source SuperACO at LURE (Orsay, France). This beam line, a normal incidence
monochromator, provides with monochromatic light in the range 7 ± 30 eV photon energy. Slits were at all time
opened to 1 mm, which gives a resolving power in the range of 500 (i.e., 20 meV at 10 eV). Photon-energy
accuracy was verified by recording the ionization potential of Ar, within 2 meV of its nominal value.
Measurements corresponding to the first-ionization threshold of the samples were performed after filtering of
higher-order light from the grating by means of a LiF window (removing all photons above 11.8 eV).
A small oven, filled with the samples, was adapted to the source in order, first, to obtain sufficient vapor
pressure of the neutral samples, and, second, to observe eventual temperature effects on the appearance curves.
The neutral species were ionized by monochromatic photons, and ions were extracted by a small extracting field
(1 V/cm), through octopolar ion guides, towards a quadrupole mass spectrometer (Nermag 350 mm), where
mass selection was performed. Detection was achieved by means of multi-channel plate operated in the counting
mode.
For charge stripping, a modified VG-ZAB-HF/AMD 604 four-sector mass spectrometer of BEBE
configuration (B and E stand for magnetic and electric sector, resp.) was used [9]. The ions of interest were
generated by electron ionization (EI, 70 eV) of the neutral PAHs and accelerated to a kinetic energy of 8 keV.
Because of the superior energy resolution of E(1), the energy-resolved CS experiments were conducted with
B(1)-only mass-selected ions [10]. To this end, the mono- and dication signals in CS experiments were recorded
at energy resolutions E/DE ꢀ 4999 by scanning E(1), and Qmin values were determined from the difference
between the high-energy onsets of the mono- and the dication peaks1). Calibration of the kinetic-energy scale
‡
‡
applied charge stripping of the molecular ions of both benzene, C6H6 ! C6H26‡ with Qmin(C6H6 ) ˆ 16.2 eV [6],
1
)
For a representative example of data analysis in CS experiments, see [11].

Helvetica Chimica Acta ± Vol. 84 (2001)
1627
‡
‡
and toluene, C7H8 ! C7H28‡ with Qmin(C7H8 ) ˆ 15.7 eV [12]. The values given further below are averages of at
least four different experiments, and the indicated errors comprise one standard deviation.
The selected-ion flow-tube mass spectrometer consist of an ion source, a first quadrupole mass filter (Q1)
for mass selection of the precursor ions, an ion flow-tube for carrying out reactivity studies, and a second
quadrupole mass filter (Q2) for product analysis [13]. In the ion source of the SIFT, corannulene and coronene
were subjected to electron ionization at 60 and 50 eV, resp. The dications formed were filtered with Q1 and
injected into the flow tube through a Venturi-type interface. He was used as the SIFT buffer gas at 0.35 Torr and
a flow tube temp. of 296 Æ 2 K. The selected ions entering the flow tube were allowed to thermalize by collisions
with He (ca. 4 ´ 105 collisions) prior to reaching the reaction region further downstream at the point of addition
of the reagent gas. Reactant and product ions were monitored still further downstream by sampling the reacting
mixture and analyzing the ions with Q2 as a function of the added reagent gas. Ions were counted with a
channeltron electron multiplier.
Commercial coronene samples (Acros, Aldrich) were used without further purification. Corannulene was
prepared by a three-step synthesis reported in [14].
In the theoretical study, a compromise between accuracy and the computing-time expenditure has to be
found due to the large size of the systems under consideration. The B3LYP method implemented in Gaussian98
[15] was used with the corresponding 6-311G(d,p) basis sets. Full geometry optimizations as well as frequency
analyses were performed at this level of theory. To evaluate the performance of this approach, reference
computations for some smaller PAHs were also made (only benzene and naphthalene were included here).
Experimental Results. ± At first, let us describe the experimental findings about the
ionization energies2) of corannulene and coronene obtained by three different mass
spectrometric techniques.
Photoionization MS. The first-ionization energies of corannulene (C20H10) and
coronene (C24H12) can be derived in a straightforward manner from the appearance
curves of the corresponding monocation peaks as a function of photon energy. Both
curves show reasonably sharp onsets, and IE(C20H10) ˆ 7.83 Æ 0.02 eVand IE(C24H12) ˆ
7.26 Æ 0.05 eV were obtained accordingly from the first apparent rise of the PI signals
from the noise level. Both values are further consistent with previous studies: the
ionization energy of corannulene has been determined as IE(C20H10) ˆ 7.7 eV by
bracketing techniques [16] and that of coronene as IE(C24H12) ˆ 7.20 Æ 0.02 eV by
photoionization [6]. Considering the lower error limit of the previous measurement, we
use IE(C24H12) ˆ 7.21 Æ 0.02 eV as a weighted average. The determination of second-
ionization energies by photoionization is more difficult, because unlike monocations,
the dication signals do not show sharp onsets, but rather rise gradually. Therefore, the
values bear larger errors and should only be regarded as upper limits in a strict sense.
The best estimates for the dication thresholds are 20.1 Æ 0.2 eV for corannulene and
18.5 Æ 0.2 eV for coronene; again, the agreement with the previous value of 18.7 Æ
0.2 eV for coronene is reasonable [6]. Further, measurements were conducted at
different temperatures of the oven containing the PAH samples, because there might
exist an effect of the neutralꢀs internal energy on the apparent onsets of the dication
‡
2
)
Note the following definitions of the terms used. IEa(PAH) and IEa(PAH ) stand for the adiabatic
ionization energies of the neutral and monocationic PAHs; thus, these properties correspond to the
(adiabatic) first- and second-ionization energies of the neutral species. Accordingly, IEv(PAH) and
‡
IEv(PAH ) represent the vertical ionizations of the respective precursors to the next charge states.
2IEv(PAH) stands for the double ionization of a neutral to a dication as potentially occurring upon
2
‡
photoionization; the corresponding adiabatic property IEa(PAH) is equal to IEa(PAH) ‡ IEa(PAH ) as
well as to the sum of the (adiabatic) first and second IEs of the neutral. For all these properties, the
associated terms DIEv/a express the differences between vertical and adiabatic electron-transfer processes.

1628
Helvetica Chimica Acta ± Vol. 84 (2001)
signals. However, within the experimental uncertainty of Æ 0.2 eV, no dependencies of
the measured thresholds of the mono- and dications from the oven temperatures were
found for either corannulene or coronene.
As long as the structural differences between neutrals, mono-, and dications are not
enormous (see below), the onsets of the photoionization signals can further be assumed
to correspond closely to adiabatic ionization energies, thus, IEa(C20H10) ˆ 7.83 Æ
‡
0.02 eV and IEa(C24H12) ˆ 7.21 Æ 0.02 eV for the neutrals, and IEa(C20H10 ) ˆ 12.3 Æ
‡
0.2 eV and IEa(C24H12 ) ˆ 11.3 Æ 0.2 eV for the monocations, respectively.
Sector MS. Both, mass-selected corannulene and coronene monocations afford
good dication yields in keV collisions. Energy-resolved charge-stripping experiments
‡
‡
give averaged Qmin values of Qmin(C20H10 ) ˆ 13.8 Æ 0.3 eV and Qmin(C24H12 ) ˆ 12,8 Æ
0.3 eV. To a first approximation [6][7], the Qmin values correspond to the vertical
ionization energies IEv of the corresponding monocations, thus suggesting
‡
‡
IEv(C20H10 ) ˆ 13.8 Æ 0.3 eV and IEv(C24H12 ) ˆ 12.8 Æ 0.3 eV. Note that these values
are significantly above the adiabatic IEs indicating either quite unfavorable Franck-
Condon effects for the mono- to dication transitions and/or systematic errors (see
below).
SIFT MS. In the flow-tube studies, mass-selected corannulene and coronene
dications were reacted with various neutral molecules M. It is important to point out
that the experimental conditions safely ensure complete thermalization of the dication
precursors. In the present context, occurrence of single-electron transfer (SET) from
the neutrals to the PAH dications according to Reaction 1 is of particular interest.
‡
‡
PAH2‡ ‡ M ! PAH ‡ M
(1)
Table 1 summarizes the reactions studied in the order of increasing IEs of the
neutral reagents. With ethene and NH3 as neutral reagents (IEs ˆ 10.51 and 10.16 eV,
resp.), corannulene dication exclusively undergoes association to the corresponding
doubly charged adduct ions. Charge transfer according to Reaction 1 is first observed
for hexafluorobenzene, which has an IE of 9.91 eV [17]3). In fact, association of
corannulene dication with C6F6 to afford C20H10C6F26‡ cannot compete with the
electron-transfer process. Conversely, association occurs exclusively in the reaction of
the coronene dication with C6F6. In the case of NO (IE ˆ 9.264 eV [17]), electron
transfer is efficient for both PAH2‡ dications, though association can compete to some
extent. Although the IE of benzene is slightly lower than that of NO (IE ˆ 9.246 vs.
9.264 eV [17]), association prevails for the arene, probably due to the better
electrostatic interaction in the case of the polarizable benzene molecule compared to
NO and the larger number of rovibrational degrees of freedom of benzene. Among the
neutral reactants under study, distinct chemical reactivity in terms of the obvious
formation of new chemical bonds is only observed for strong oxidizing agents such as
NO2 (IE ˆ 9.75 eV [17]) and molecular chlorine (IE ˆ 11.48 eV [17]) with corannulene
dication.
The overall rate constants vary substantially and in particular do not correlate with
the neutralsꢀ IEs. For example, the rate constants for ammonia and benzene are of
3
)
For a compilation of ionization energies, also see: http://webbook.nist.gov/chemistry/

Helvetica Chimica Acta ± Vol. 84 (2001)
1629
2‡
Table 1. Branching Ratios and Observed Rate Constants kobs of the Reactions of Corannulene C20H
and
10
2‡
Coronene C24H Dications, Respectively, with C6D6, NO, NO2, C6F6, NH3, C2H4, and Cl2 in He Buffer Gas at
12
0.35 Torr and 298 K
IE [eV]a)
C20H
C24H
2‡
2‡
10
12
2‡
‡
‡
2‡
‡
‡
C6D6
NO
9.25
C20H ‡ C6D6 ! C20H10 ‡ C6D6
0.30
0.70
C24H ‡ C6D6 ! C24H12 ‡ C6D6
0.30
0.70
10
12
2‡
2‡
! C20H10C6D6
! C24H12C6D6
kobs ˆ 1.7 ´ 10 9 cm3 molec 1 s
kobs ˆ 1.6 ´ 10 cm3 molec 1 s
1
9
1
2‡
‡
2‡
‡
‡
9.26
9.75
C20H ‡ NO ! C20H10 ‡ NO
0.85
0.15
C24H ‡ NO ! C24H12 ‡ NO
0.70
0.30
10
12
! C20H10NO2‡
! C24H12NO2‡
kobs ˆ 2.3 ´ 10 cm3 molec
s
kobs ˆ 1.3 ´ 10 cm3 molec
s
10
1
1
10
1
2‡
‡
‡
2‡
‡
‡
NO2
C20H ‡ NO2 ! C20H10O ‡ NO
0.60
0.40
C24H ‡ NO2 ! C24H12O ‡ NO
0.60
0.25
0.10
0.05
10
12
! C20H10O2‡ ‡ NO
! C24H12O2‡ ‡ NO
‡
‡
! C24H12 ‡ NO2
! C24H12NO22‡
kobs ˆ 9.2 ´ 10 cm3 molec
s
kobs ˆ 6.3 ´ 10 cm3 molec 1 s
11
1
1
11
1
1
2‡
‡
‡
2‡
2‡
C6F6
NH3
C2H4
Cl2
9.91
10.16
10.51
11.48
C20H ‡ C6F6 ! C20H10 ‡ C6F6
1.00
1.00
1.00
1.00
C24H ‡ C6F6 ! C24H12C6F6
1.00
1.00
10
12
kobs ˆ 2.4 ´ 10 cm3 molec
s
kobs ˆ 1.1 ´ 10 cm3 molec
s
10
1
1
10
1
2‡
2‡
C20H ‡ NH3 ! C20H10NH32‡
C24H ‡ NH3 ! C24H12NH32‡
10
12
kobs ˆ 1.9 ´ 10 9 cm3 molec 1 s
kobs ˆ 1.4 ´ 10 cm3 molec 1 s
1
9
1
2‡
2‡
C20H ‡ C2H4 ! C20H10C2H4
10
kobs ˆ 1.8 ´ 10 cm3 molec 1 s
10
1
2‡
C20H ‡ Cl2 ! C20H10Cl2‡ ‡ Cl
10
kobs ˆ 6.2 ´ 10 cm3 molec
s
12
1
a) Taken from [17].
similar magnitudes while the IEs of the neutrals differ by almost 1 eV. It appears more
plausible to correlate the rate constants with the stabilities of the corresponding
electrostatic complexes. Thus, fast reactions dominated by adduct formation occur for
the dipolar ammonia as well as the polarizable benzene molecule, whereas the small,
less polarizable neutrals NO2, C2H4, and Cl2 react much more slowly. Exceptions to this
rule of thumb are the relatively fast rate for NO and the slow reaction of C6F6. The
former exception can be rationalized in terms of the effectively competing SET channel
that is unavailable to the other neutrals except benzene. The inefficiency of the addition
reaction of hexafluorobenzene with the two PAH dications can be attributed to the
weak binding properties of this perfluorinated molecule. Thus, the electron-with-
drawing F substituents disfavor coordination of cationic species to the aromatic ring
while the F-atoms themselves are also poor coordinators. This qualitative argument is
supported by the huge difference in the binding energies of gold cation to benzene and
‡
‡
its hexafluoro derivative, i.e., D0(Au C6H6) ˆ 3.0 eV [18] vs. D0(Au C6F6) ˆ 1.3 eV
[19].
Summarizing the SIFT data in terms of brackets for the occurrence of Reaction 1
yields thresholds of 10.1 Æ 0.2 eV for the corannulene dication and 9.5 Æ 0.3 eV for the
coronene dication, respectively. It is important to note that these thresholds are not
direct measures of the IEs of the monocations of the corresponding PAHs because
formation of two monocations in a dication/neutral reaction is associated with a
considerable Coulomb barrier [7]. For the same reason, however, these figures provide

1630
Helvetica Chimica Acta ± Vol. 84 (2001)
‡
solid lower limits for the monocation IEs, specifically, IE(C20H10 ) > 10.1 Æ 0.2 eV and
‡
IE(C24H12 ) > 9.5 Æ 0.3 eV.
Theoretical Results. ± Given the size of the systems under study, some compromise
in the level of theory applicable needs to be made. Thus, the B3LYP/6-311G(d,p)
approach chosen is certainly inappropriate to predict absolute IEs within chemical
accuracy (<0.1 eV). For example, a detailed examination by Pople and co-workers on
the performance of B3LYP for the computation of first IEs gave a mean error of Æ
0.18 eV for a set of 83 reference compounds [20]. Of particular interest in the present
context is that the IEs of all closed-shell hydrocarbons examined by Pople and co-
workers are underestimated by ca. 0.2 eV with B3LYP. While this effect is in part
associated with incompleteness of the basis sets, DiLabio et al. have attributed this
particular overestimation of the cation stabilities to unbalanced treatments of closed-
and open-shell hydrocarbons with B3LYP [21]. Hence, the major relevance of the
present theoretical studies is the prediction of the relative energetics of the various
states and the differences between vertical and adiabatic electron-transfer processes
(DIEv/a) in particular.
The B3LYP/6-311G(d,p) level of theory is employed to assess the energetics of the
neutral, the mono-, and dicationic species of interest (Table 2). For an evaluation of the
accuracy of the computational predictions, benzene and naphthalene are additionally
studied as reference molecules. Compared to experiment, the adiabatic first IEs of the
four hydrocarbons are slightly underestimated with B3LYP (mean deviation
0.25 eV). While the detailed evaluation of the second IEs is postponed to the
discussion, a similar trend is indicated by comparing the computed double-ionization
energies 2IEs of the neutrals to the dications with the photoionization thresholds given
in Table 2. Using the computed dication ground-states for comparison with the
photoionization thresholds, a mean deviation of 0.8 eV is derived. Roughly, these
values imply an underestimation of the IEs by 2 ± 3% with the B3LYP/6-311G(d,p)
approach for mono- and dications.
As far as the dication states are concerned, the B3LYP computations predict triplet
ground states for benzene, corannulene, and coronene, while the naphthalene dication
appears to have a singlet ground state (see also [22]). As a note of caution, we add,
however, that the performance of density-functional methods with respect to the
accurate energetic ordering of electronic states having different multiplicities is
somewhat limited [23]. Irrespective of the actual ground-state assignment, it should be
noted that, except for benzene, the dication state-splittings are more or less within the
accuracies of the experimental methods discussed here.
Most important in the present context, however, is the point that the differences
between vertical and adiabatic ionization energies are not especially large. Maximal
values occur for formation of benzene dication in the (excited) singlet state. The
differences DIEv/a are particularly small for the respective dication ground states, i.e.,
ca. 0.1 eV for first and second IEs and ranging from 0.1 to 0.5 eV for double ionization.
Therefore, differences in Franck-Condon factors [24] can certainly not be applied to
rationalize the spread of the experimental data discussed above.
Among the theoretical results, one interesting aspect deserves comment. For a
continuous change of electronic and geometric properties as a function of charge, it is

Helvetica Chimica Acta ± Vol. 84 (2001)
1631
Table 2. Adiabatic and Vertical Ionization Energies (in eV) of Benzene, Naphthalene, Corannulene, and
Coronene Computed at the B3LYP/6-311G(d,p) Level of Theory (dication ground states in bold)
a
b
c
Ion state
IEa
)
IEv
)
DIEv/a
IE1 ‡ IE2
)
DDIEd)
Exper.
First IEs
C6H6
C10H8
C20H10
C24H12
doublet
doublet
doublet
doublet
9.03
7.82
7.58
6.98
9.18
7.92
7.71
7.03
0.15
0.10
0.13
0.05
9.22e)
8.15e)
7.83 f)/7.7g)
7.20e)/7.26 f)
Second IEs
C6H6
singlet
triplet
singlet
triplet
singlet
triplet
singlet
triplet
15.36
15.28
13.22
13.66
12.01
11.66
10.90
10.79
16.10
15.47
13.31
13.79
12.21
11.76
10.98
10.86
0.74
0.19
0.09
0.13
0.20
0.10
0.08
0.07
16.2h )
14.6h )
13.8i )
12.8i )
C10H8
C20H10
C24H12
Double IEs
C6H6
singlet
triplet
singlet
triplet
singlet
triplet
singlet
triplet
24.39
24.31
21.04
21.48
19.59
19.24
17.88
17.77
25.55
24.50
21.41
21.61
20.14
19.71
18.13
17.83
1.16
0.19
0.37
0.13
0.55
0.47
0.25
0.06
25.28
24.65
21.23
21.71
19.92
19.47
18.01
17.89
0.27
0.15
0.18
0.10
0.22
0.24
0.12
0.06
25.3e )
C10H8
C20H10
C24H12
21.5e)
20.1f)
18.7e )/28.5f )
a) Computed adiabatic values including zero point energies. b) Computed vertical values assuming the adiabatic
zero-point energies. c) Sum of the vertical first- and second-ionization energies; only defined for the last four
d
rows. ) Difference DDIE ˆ 2IEv (IE1 ‡ IE2); this quantity is only defined for the last four rows. e) Photo-
ionization thresholds taken from [6]. f) Photoionization thresholds, this work. g) Ion/molecule bracketing, see
i
[16]. h) Charge stripping data taken from [6]. ) Charge-stripping data, this work.
‡
expected that stepwise (vertical) ionization PAH ! PAH ! PAH2‡ with intermediate
relaxation of the monocation is less energy-demanding than direct (vertical) double
‡
ionization PAH ! PAH2‡. Hence, DDIEv ˆ 2IEv(PAH) [IEv(PAH) ‡ IEv(PAH )] is
expected to be positive. While this trend is indeed observed for all singlet dications
under study, the transitions from the singlet neutrals to the triplet dications give the
opposite sign for benzene, naphthalene, and coronene. Qualitatively, this finding
implies that the structures of the triplet dication states show a better match with the
neutral ground states than with the monocation intermediates. The bent corannulene
molecule appears as an exception in this respect, as DDIEv is positive for both the
singlet and triplet dications.
Discussion. ± Before explicitly evaluating the absolute values of the second IEs, it is
worthwhile noting that all methods agree reasonably that coronene is significantly
easier to ionize than corannulene. Thus, both photoionization as well as charge
stripping yield differences of ca. 1.0 eV, a lower limit of 0.6 eV is implied by the SIFT

1632
Helvetica Chimica Acta ± Vol. 84 (2001)
study, and the calculations predict a difference of 0.87 eV between the adiabatic
ionizations of these PAHs monocations to their triplet ground-state dications.
Having made these introductory remarks, let us use the theoretical data to analyze
the mass spectrometric results in some more detail. As DIEv/a is comparable in
magnitude to the experimental error, we may safely assume that the onsets of the PI
signals of the mono- and dications correspond to the adiabatic IEs of corannulene and
coronene, respectively. Thus, the best estimates of ionization energies are IE(C20H10) ˆ
7.83 Æ 0.02 eV and IE(C24H12) ˆ 7.21 Æ 0.02 eV for the neutral molecules, and
‡
‡
IE(C20H10 ) ˆ 12.3 Æ 0.2 eV and IE(C24H12 ) ˆ 11.3 Æ 0.2 eV for the monocations. The
monocation IEs are in obvious disagreement with the charge stripping data, i.e.,
‡
‡
Qmin(C20H10 ) ˆ 13.8 Æ 0.3 eV and Qmin(C24H12 ) ˆ 12.8 Æ 0.3 eV; the corrections due to
DIEv/a are minor (0.1 eV) and cannot account for the differences of ca. 1.5 eV in both
cases. Note, however, that the relative difference agrees well. Thus, if we use coronene
‡
as the calibrant in charge stripping of corannulene, Qmin(C20H10 ) ˆ 12.2 Æ 0.3 eV is
‡
obtained, which agrees very well with IE(C20H10 ) ˆ 12.3 Æ 0.2 eV derived above.
Obviously, a yet unknown effect is operative in the calibration procedure of the charge-
stripping experiments in which the transitions of benzene and toluene monocations to
the corresponding dications were used as references. Among others [25], possible
reasons could be associated with the large differences in mass or size of the small PAHs
C6H6 and C7H8 used as references compared to C20H10 and C24H12, respectively. With
regard to the extensive CS studies of Beynon and co-workers in the 1980s, a mere mass
effect appears unlikely, however. For example, along the series Mg ± Ba, the Qmin values
of the monocations were found to agree with spectroscopic data [26]. In contrast, size
dependencies in the efficiency of energy transfer are well known in KeV collisions; the
conversion of translational into internal energy is particular relevant [27]. Thus,
rovibrational modes of the corannulene and coronene monocations could adsorb a
certain amount of collision energy prior to ionization to the corresponding dications,
thereby shifting the Qmin values to higher energies. In comparison, the energy
depositions in the benzene and toluene ions used as references in the CS experiments
are significantly smaller. Operation of such effects could well explain the fact that the
absolute Qmin values of corannulene and coronene are too large compared to the
ionization energies derived from the photoionization data while the relative data agree.
With regard to the arguments put forward in [6], it is to be stressed once more that the
above explanation does not refer to Franck-Condon effects in vertical ionization.
Having established the IEs, we can now analyze the ion/molecule reactions
examined with the SIFTapparatus in more detail. Single-electron transfer in a reaction
of a dication with a neutral species according to Reaction 1 is associated with
considerable barrier due to Coulomb repulsion of the two monocationic products
formed. In the experiment, charge separation is first observed to occur with
hexafluorobenzene (IE ˆ 9.91 eV) in the case of corannulene and with NO2 (IE ˆ
9.75 eV) in the case of coronene. The magnitude of the Coulombic barrier can be
estimated according to a model developed for single-electron transfer reactions of
fullerene dications [28]4). These barriers amount to 2.2 and 1.9 eV assuming that the
electron transfer occurs at contact upon an edge-on (rather than face-on) approach at
4
)
For more general discussions and leading references, see [29].

Helvetica Chimica Acta ± Vol. 84 (2001)
1633
intercharge separations equal to the diameters of the C-skeletons, which are computed
to be 6.48 and 7.49 Š for corannulene and coronene, respectively. Within this
‡
approximation, the ionization energies can be estimated to be IE(C20H10 ) ˆ 12.3 Æ
‡
0.2 eV and IE(C24H12 ) ˆ 11.4 Æ 0.2 eV with no uncertainty being assigned to the model
calculation. These values are in remarkable agreement with the values derived from
photoionization.
Conclusion. ± Three different experimental techniques in conjunction with ab initio
theory are used to study the mono- and dications of the prototype PAH molecules
corannulene and coronene. All experiments as well as theory agree upon the fact that
ionization of corannulene is more energy-demanding than that of coronene. Photo-
ionization thresholds give IE(C20H10) ˆ 7.83 Æ 0.02 eV, IE(C24H12) ˆ 7.21 Æ 0.02 eV,
‡
‡
IE(C20H10 ) ˆ 12.3 Æ 0.2 eV, and IE(C24H12 ) ˆ 11.3 Æ 0.2 eV as best estimates of the
adiabatic properties. While charge-stripping experiments agree upon the relative IEs of
the two PAHs, the measured Qmin values are substantially overestimated in comparison
to the photoionization thresholds. B3LYP Calculations reveal that neither differences
between vertical and adiabatic transitions nor formation of the dications in different
electronic states can explain this discrepancy. Instead, it is suggested that the CS
stripping are shifted to higher energies due to dissipation of some translational energy
into the rovibrational manifold of the larger PAH molecules. The proposed effect of
molecular size upon the calibration of energy-resolved CS experiments remains to be
analyzed further; the impulse-collision transfer (ICT) theory seems to provide a
starting point in this respect [27b]. Charge separations in ion/molecule reactions using a
SIFT apparatus are in good agreement with the PI studies, if the Coulomb repulsion of
the monocationic fragments formed is taken into account. As far as the existence of
multiply charged PAHs ions in interstellar media is concerned, the sizable Coulomb
barriers apparent in the reactivity studies in the SIFT apparatus further imply
considerable kinetic stabilizations of PAH2‡ dications with respect to SET in the
presence of the most abundant interstellar neutral molecules.
The Berlin group appreciates financial support by the Volkswagen-Stiftung, the Deutsche Forschungsge-
meinschaft, the European Commission (MCInet, RTN1-1999-254), the Fonds der Chemischen Industrie, and the
Gesellschaft von Freunden der Technischen Universität Berlin. The staff of LURE is thanked for operating the
Orsay Super-ACO ring and general facilities. The SIFT measurements and theoretical calculations at York
University were supported by the Natural Sciences and Engineering Research Council of Canada. D.S. thanks
Â
the Laboratoire Chimie Physique of the Universite Paris Sud, Orsay, for a visiting fellowship. L.T.S. gratefully
acknowledges financial support from the U.S. Department of Energy.
REFERENCES
[1] a) ꢁPolynuclear Aromatic Hydrocarbonsꢀ, Eds. R. I. Freudental, P. W. Jones, Raven Press, New York, 1976;
b) J. Ahrens, A. Keller, R. Kovacs, K.-H. Homann, Ber. Bunsen-Ges. Phys. Chem. 1998, 102, 1823.
[2] a) D. J. De Frees, M. D. Miller, D. Talbi, F. Pauzat, Y. Ellinger, Astrophys. J. 1993, 408, 530; b) S.
Schlemmer, D. J. Cook, J. A. Harrison, B. Wurfel, W. Chapman, R. J. Saykally, Science 1994, 265, 1686.
[3] D. K. Bohme, Chem. Rev. 1992, 92, 1487.
[4] H. Hopf, ꢁClassics in Hydrocarbon Chemistryꢀ, Wiley-VCH, Weinheim, 2000.
Â
[5] ꢁPolycyclic Aromatic Hydrocarbons and Astrophysicsꢀ, Eds. A. Leger, L. dꢀ Hendecourt, N. Boccara,
NATO ASI Series 191, Reidel, Dordrecht, 1987.
[6] S. Tobita, S. Leach, H. W. Jochims, E. Rühl, E. Illenberger, H. Baumgärtel, Can. J. Phys. 1994, 72, 1060.

1634
Helvetica Chimica Acta ± Vol. 84 (2001)
Â
[7] a) K. Vekey, Mass Spectrom. Rev. 1995, 14, 195; b) D. Schröder, H. Schwarz, J. Phys. Chem. A 1999, 103,
7385.
[8] O. Dutuit, C. Alcaraz, D. Gerlich, P. M. Guyon, J. W. Hepburn, C. Metayer-Zeitoun, J. B. Ozenne, T. Weng,
Chem. Phys. 1996, 209, 177.
[9] C. A. Schalley, D. Schröder, H. Schwarz, Int. J. Mass Spectrom. Ion Processes 1996, 153, 173.
[10] D. Schröder, H. Schwarz, Int. J. Mass Spectrom. Ion Processes 1995, 146/147, 183.
[11] S. McCullough-Catalano, C. B. Lebrilla, J. Am. Chem. Soc. 1993, 115, 1441.
[12] K. Lammertsma, P. v. R. Schleyer, H. Schwarz, Angew. Chem. 1989, 101, 1313; Angew. Chem., Int. Ed. 1989,
28, 1321.
[13] G. I. Mackay, G. D. Vlachos, D. K. Bohme, H. I. Schiff, Int. J. Mass Spectrom. Ion Phys. 1980, 36, 259.
[14] L. T. Scott, P.-C. Cheng, M. M. Hashemi, M. S. Bratcher, D. T. Meyer, H. B. Warren, J. Am. Chem. Soc.
1997, 119, 10963.
[15] Gaussian 98, Revision A.7, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R.
Cheeseman, V. G. Zakrzewski, J. A. Montgomery Jr., R. E. Stratmann, J. C. Burant, S. Dapprich, J. M.
Millam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi, B.
Mennucci, C. Pomelli, C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala, Q. Cui, K.
Morokuma, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz, A. G.
Baboul, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R. L. Martin, D. J.
Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, C. Gonzalez, M. Challacombe, P. M. W. Gill,
B. Johnson, W. Chen, M. W. Wong, J. L. Andres, C. Gonzalez, M. Head-Gordon, E. S. Replogle, J. A. Pople,
Gaussian, Inc., Pittsburgh PA, 1998.
[16] H. Becker, G. Javahery, S. Petrie, P.-C. Cheng, H. Schwarz, L. T. Scott, D. K. Bohme, J. Am. Chem. Soc.
1993, 115, 11636.
[17] S. G. Lias, J. E. Bartmess, J. F. Liebman, J. L. Holmes, R. D. Levin, W. G. Mallard, J. Phys. Chem. Ref. Data,
Suppl. 1 1988, 17.
[18] T. K. Dargel, R. H. Hertwig, W. Koch, Mol. Phys. 1999, 96, 583.
Ï Â
[19] D. Schröder, H. Schwarz, J. Hrusak, P. Pyykkö, Inorg. Chem. 1998, 37, 624.
[20] L. A. Curtiss, P. C. Redfern, K. Raghavachari, J. A. Pople, J. Chem. Phys. 1998, 109, 42.
[21] G. A. DiLabio, D. A. Pratt, J. S. Wright, Chem. Phys. Lett. 1999, 311, 215.
[22] L. Ghergel, J. D. Brand, M. Baumgarten, K. Müllen, J. Am. Chem. Soc. 1999, 121, 8104.
[23] W. Koch, M. C. Holthausen, ꢁA Chemistꢀs Guide to Density Functional Theoryꢀ, Wiley-VCH, Weinheim,
2000, p. 168.
[24] C. Wesdemiotis, R. Feng, E. R. Williams, F. W. McLafferty, Org. Mass Spectrom. 1986, 21, 689.
[25] a) M. Guilhaus, R. G. Kingston, A. G. Brenton, J. H. Beynon, Int. J. Mass Spectrom. Ion Processes 1985, 63,
101; b) N. Jeffreys, I. W. Griffiths, D. E. Parry, F. M. Harris, Int. J. Mass Spectrom. Ion Processes 1997, 164,
133.
[26] M. Rabrenovic, T. Ast, J. H. Beynon, Int. J. Mass Spectrom. Ion Processes 1984, 61, 31, and refs. cit. therein.
[27] a) E. Uggerud, P. J. Derrick, J. Phys. Chem. 1991, 95, 1430; b) C. D. Bradley, P. J. Derrick, Org. Mass
Spectrom. 1993, 28, 390.
[28] S. Petrie, G. Javahery, J. Wang, D. K. Bohme, J. Phys. Chem. 1992, 96, 6121.
[29] a) Z. Herman, Int. Rev. Phys. Chem. 1996, 15, 299; b) S. D. Price, J. Chem. Soc., Faraday Trans. 1997, 93,
2451.
Received March 12, 2001