Septulene: The heptagonal homologue of kekulene

Article abstract of DOI:10.1002/anie.201203266

A family likeness: The seven-sided homologue of kekulene, septulene (see structure), has been synthesized in seven steps by ring-closing metathesis. Its properties (spectroscopic and structural) are strikingly similar to those of kekulene despite its sevenfold symmetry and resulting non-alternant structure. The chair-type conformation in the crystal arises from packing forces. Septulene is probably a free pseudorotor in the gas phase. Copyright

Full text of DOI:10.1002/anie.201203266

Angewandte
Chemie
DOI: 10.1002/anie.201203266
Kekulene
Septulene: The Heptagonal Homologue of Kekulene**
Bharat Kumar, Ruth L. Viboh, Margel C. Bonifacio, William B. Thompson,
Jonathan C. Buttrick, Babe C. Westlake, Min-Soo Kim, Robert W. Zoellner, Sergey A. Varganov,
Philipp Mçrschel, Jaroslav Teteruk, Martin U. Schmidt, and Benjamin T. King*
Dedicated to Professor Robert G. Bergman on the occasion of his 70th birthday
Staab and Diederich defined cycloarenes as comprising
annelated benzene rings that form a macrocycle with
inward-pointing C H bonds.[1] Their synthesis and character-
ꢀ
ization of kekulene, the prototypical cycloarene,[1–4] answered
a long-standing and fundamental question about arenes and
the nature of aromaticity: do p electrons move throughout
the entire system, as hypothesized by Pauling,[5] or do they
remain localized in rings, as predicted by McWheeny[6] and
described phenomenologically by Clar.[7] The chemical shift
of the inner protons of 2 gave the answer. Paulingꢀs model
suggested that kekulene 2 should behave as concentric
annulenes (Scheme 1 bottom right), and that the inner
protons ought to be strongly shielded, as in [18]annulene.[8]
The localized model, as embodied by Clarꢀs aromatic sextets,
suggests that the ring currents arise from localized rings
(Figure 1 top right), and that the inner protons should be
deshielded, as in benzene. Experiments showed that the inner
protons are deshielded and resonate at d > 7 ppm. This
Scheme 1. Septulene 1 and kekulene 2. The unique radial double bond
in the Kekulꢀ structure of 1 is indicated.
deshielding demonstrates that electrons do not move freely
about the entire molecule, but are instead localized into
individual rings, just as in benzene.
We report the synthesis and properties of kekuleneꢀs
seven-sided cousin septulene (1), and find that its properties
reinforce the conclusions above and open some new ques-
tions.
The properties of 1 are strikingly similar to those of 2,
even though their Kekulꢁ structures (Scheme 1 bottom) are
fundamentally different. This difference arises from the odd
number of C atoms in both the inner and outer annuli of 1,
which necessitates a radial double bond in the Kekulꢁ
structure. It also follows that 1 is non-alternant. The
remarkable similarity between 1 and 2, despite their funda-
mentally different Kekulꢁ structures, dispels the notion that
a consideration of a few Kekulꢁ structures provides much
insight into the chemistry of condensed arenes.[9]
[*] B. Kumar, R. L. Viboh, M. C. Bonifacio, W. B. Thompson,
J. C. Buttrick, B. C. Westlake, M.-S. Kim, Prof. B. T. King
Department of Chemistry, University of Nevada, Reno
Reno, NV 89557 (USA)
The only previously known cycloarenes were 2 and its
hexaaza analogue.[10] Some computational[11] and synthetic
studies[12] towards 1 have appeared in dissertations, but
otherwise 1 appears to have escaped attention.
E-mail: king@chem.unr.edu
Prof. S. A. Varganov
Department of Chemistry, University of Nevada, Reno
Reno, NV 89557 (USA)
Prof. R. W. Zoellner
Department of Chemistry, Humboldt State University
One Harpst Street, Arcata, California 95521-8299 (USA)
The scarcity of known cycloarenes is offset by the
abundance of schemes for their nomenclature. The current
schemes either assume a graphene lattice, which excludes 1,
or fail to convey the structure of the molecule. We agree with
the assertion by Staab and Diederich that, “The naming of (2)
as a polycyclic system according to the IUPAC rules on
nomenclature leads to an extraordinarily complicated name
which does not give any direct information about the
structure and symmetry of the molecule”.[1] We propose the
following solution: To handle the general case of cycloarenes,
we adopt the corannulene nomenclature developed by
Agranat et al.,[13] in which the name for kekulene (2) is
P. Mçrschel, J. Teteruk, Prof. Dr. M. U. Schmidt
Institut fꢁr Anorganische und Analytische Chemie
Goethe-Universitꢂt
Max-von-Laue-Strasse 7, 60438 Frankfurt am Main (Germany)
[**] This work was supported by the US-NSF (CHE-0957702). We thank
Dr. S. M. Spain and Dr. J. L. Dallas (NSF DBI-0722538) for
assistance with spectroscopy, and Profs. A. D. Schlꢁter, K. K.
Baldridge, and H. B. Bꢁrgi for useful suggestions.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201203266.
Angew. Chem. Int. Ed. 2012, 51, 12795 –12800
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
12795

.
Angewandte
Communications
corannulene [62,61,62,61,62,61,62,61,62,61,62,61], where the super-
script numbers designate the number of carbon atoms of the
small ring that are on the periphery of the molecule and are
not shared by any other small rings. In this case, all the small
rings are six-membered, hence the twelve sixes. The number-
ing of the small rings begins with the ring that contains the
carbon atom that is designated C1 when the molecule is
drawn in standard orientation.[14] Although not given by
Agranat et al., the trivial systematic name for 2 may be
abbreviated as “corannulene[(62,61)6]”, where the subscript
six clearly indicates the sixfold repeat unit. Thus, 1 becomes
“corannulene[(62,61)7]”, with the sevenfold repeat unit clearly
indicated.
Simple purification through silica gel gave a mixture of
open-chain polymers and cyclic oligomers. This unfractio-
nated polycondensation product[22] was treated with the
metathesis catalyst. By monitoring the disappearance of the
ꢀ
C D stretch in the IR spectrum, we optimized the RCM
reaction to push it to near completion.[23] Septulene (1)
crystallized from THF during workup and was purified by
HPLC.
Similar to 2, 1 is also insoluble in common organic
solvents. It dissolves slightly in 1,2-dichlorobenzene and 1,2,4-
trichlorobenzene (ca. 1 mg/20 mL at 1008C). The melting
point of 1 is > 5008C. This behavior is similar to that of 2.
Septulene (1) is stable, and no decomposition was observed
when stored at room temperature under air for several
months. Septulene (1) crystallizes from 1,2-dichlorobenzene
and 1,2,4-trichlorobenzene as yellow crystals having a light-
green fluorescence.
The key to our synthesis of 1 was to introduce seven
olefinic phenanthrene bridges onto the cyclohepta-m-phen-
ylene framework (Scheme 2) by ring-closing metathesis
The optical absorbance spectrum of 1 is
similar to that of 2,[4] with the same three
transitions (lmax 339 nm, loge 4.90; 365 (sh),
4.45; 396, 4.04; Figure 1) just slightly red-
shifted from those of 2 (lmax 326 nm, loge
4.93; 347 (sh), 4.74; 388, 4.22).[3] The fluores-
cence spectrum of 1 shows bands at 453, 466,
486, and 517 nm, which are remarkably
similar to those observed for 2 at 453, 468,
483, and 513 nm (ꢁ 5 nm).[4]
The NMR spectroscopic properties of
1 and 2 are also similar. The 1H NMR
spectrum of 1 in [D4]1,2-dichlorobenzene
(referenced to tetramethylsilane) showed
Scheme 2. a) [Pd{P(p-tol)3}3], K2CO3,THF/H2O, b) Grubbs 2nd generation catalyst, 1,2,4-
trichlorobenzene, toluene, 708C (4, n=2).
(RCM) of pendant olefins.[15–17] The cyclohepta-m-phenylene
was produced in low yield as a by-product in the polymeri-
zation of monomer 3, which was designed to have two
features that would simplify the analysis and handling:
1) Methyl groups help suppress the spontaneous vinylic
polymerization and cross-linking that plagued the simple
vinyl system, and 2) deuterium provided a spectroscopic
three singlets at 7.86, 8.36, and 10.19 ppm in the ratio 2:1:1
(Figure 2). The chemical shifts of 2 are 7.94, 8.37, and
ꢀ
handle for the RCM reaction: the disappearance of the C D
stretch in the infrared spectrum enabled the course of the
RCM to be monitored in the complicated mixture.
Monomer 3 was synthesized in five steps on 25 g scale (see
the Supporting Information). Deuterium was introduced at
the b-vinyl positions using the Wittig reagent derived from the
deuterated salt CH3CD2PPh3Br, which was easily prepared at
low cost.[18]
The polymerization of 3 under standard Suzuki condi-
tions[19–21] gave a mixture of polymers typically having about
100 repeat units and a polydispersity index (PDI) of 1.98, as
determined by gel-permeation chromatography versus poly-
styrene standards. MALDI-TOF analysis suggested that some
macrocycles were also present. By performing the polymer-
ization under dilute conditions, we were able to increase the
fraction of macrocycles formed. Although the cyclohexameric
precursor to kekulene (2) could be detected by mass
spectrometry, the cycloheptameric precursor (4) to septulene
was the dominant product. The reason for the preferential
formation of heptamer 4 over the hexamer is unclear.
Figure 1. UV/Vis absorption and fluorescence spectrum of 1.
Figure 2. 1H NMR spectrum of 1.
12796 www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 12795 –12800

Angewandte
Chemie
10.45 ppm, also in a 2:1:1 ratio. The chemical
shifts of 1 and 2 cannot be directly compared
because the spectrum of 1 was measured in
[D4]1,2-dichlorobenzene and that of 2 was mea-
sured in [D3]1,2,4-trichlorobenzene. To support
the assertion that the chemical shifts of 1 and 2
are quite similar, and to aid in the assignment, we
calculated the chemical shifts of 1 (the saddle-
shaped minimum, averaged over apparent D7h
symmetry) as 8.37 (outer vinyl), 8.5 (outer aryl),
and 10.31 (inner). For 2, we calculated similar
chemical shifts of 8.37, 8.79, and 10.05 ppm; thus
the calculated NMR spectra of 1 and 2 are
essentially the same.
In 1978, Staab and Diederich described the
acquisition of the 1H NMR spectrum of 2 as
“extraordinarily difficult.”[3] Their high standard
compelled us to record the 13C NMR spectrum of
1, an experiment that required 14 h on a high-
1
field NMR spectrometer (800 MHz for H, 200
MHz for 13C) equipped with a cryoprobe. [D]Bro-
moform was used to avoid interfering solvent
resonances. Five signals were evident at d = 131.7,
130.4, 129.8, 128.2, and 118.0 ppm, which are in
reasonable agreement with the calculated
(B3LYP/6-31G(d)) values of d = 134.0, 133.4,
131.3, 129.7, and 122.3.
Nucleus independent chemicals shifts (NICS-
1)[24] are widely used to quantify ring currents.
The NICS-1 values for the benzenoid rings of
1 and 2 (ꢀ11.3 and ꢀ11.5 ppm) match those of
the benzenoid rings of phenanthrene (ꢀ11.0) and
benzene (ꢀ10.5). The NICS-1 values for the
olefinic ring in 1 and 2 (ꢀ5.8 and ꢀ6.3 ppm)
indicate ring currents smaller than that of the
Figure 3. The conformation of 1 in the solid state viewed from the top and the side.
The least-squares molecular plane is shown in translucent gray, and thus atoms
behind this plane appear gray. Atomic displacement parameters are drawn at 50%
probability and H atoms are omitted.
olefinic ring of phenanthrene (ꢀ8.1 ppm). These NICS values
are in full accordance with Clarꢀs picture of aromatic sextets,
and suggest that the olefinic bonds in 1 and 2 are even more
pronounced than in phenanthrene.
X-ray diffraction analysis established the structure of
1 (Figure 3).[25] We were surprised that 1 is not saddle-shaped
in the crystal, as predicted by gas-phase density functional
theory (DFT) calculations (see below), but rather has a chair-
type conformation.[26] Nonetheless, it is not flat: the root-
mean-square deviation (RMS) of carbon atoms of 1 from the
least-squares plane passing through all 56 carbon atoms is
0.285 ꢂ (Figure 3). As in kekulene, the bond lengths are in
full agreement with Clarꢀs representation, and six unique
types of bonds can be identified. Indeed, the average bond
lengths in 1 and 2 are essentially identical (Figure 4).
The similarity between 1 and 2 even extends to their
crystal packing. Both 1 and 2 form slipped stacks arranged in
ꢀ
Figure 4. Unique C C bonds lengths of 1 and 2.
angle—the inclination of molecules in adjacent stacks—is
54.58.[27] For 2, molecules form an angle 42.98 with the stacking
axis, adjacent stacks are slipped by 3.12 ꢂ, the interplanar
spacing is 3.35 ꢂ, and the herringbone angle is 868.
Gas-phase DFT calculations usually predict the correct
crystallographic conformation of distorted polyaromatic
hydrocarbons (PAHs), but they fail for 1. The solid-state
chair conformation of 1 could not be located in unconstrained
gas-phase calculations, but we could model the chair con-
formation by fixing the distance from the mean molecular
plane for seven atoms. We also considered the planar D7h,
saddle Cs, and saddle C2 conformations. The energies
calculated by several DFT methods (Table 1) are in agree-
a
herringbone pattern. The side-by-side comparison
(Figure 5) of the packing diagrams of 1 with those reported
for 2 compels us to discuss the packing of 1 in the same terms
used to describe the packing of 2.[2] The average molecular
plane forms an angle of 27.38 with the stacking axis, which is
parallel to the b axis. Stacked molecules are slipped by 5.86 ꢂ,
and the interplanar spacing is 3.19 ꢂ. The herringbone
Angew. Chem. Int. Ed. 2012, 51, 12795 –12800
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 12797

.
Angewandte
Communications
optimized under the same conditions. All the hypothetical
structures with molecules in Cs and C2 conformations have
less favorable packing arrangements. Their density is at least
4.6% (Cs conformation) or 5.5% (C2 conformation) lower
than the density of the optimized experimental structure. The
intermolecular energy is worse by at least 4.3 kcalmolꢀ1
,
which is on the same order as the difference in the intra-
molecular energies calculated by DFT. While we could not
rule out the existence of another polymorph that favors the
saddle conformer, the saddle geometry does not allow for
good, dense packing with low lattice energy.
The curvature of 1 invites a discussion of strain. Assuming
that 1, similar to 2,[32–34] does not enjoy any unusual aromatic
stabilization, we can define the strain energy of 1 by the
simple homodesmotic reaction 7(2)!6(1), which gives
a strain energy of only 9.8 kcalmolꢀ1 (B3LYP/6-31G(d)) for
the gas-phase C2 structure. This amounts to only about
0.2 kcalmolꢀ1 Catomꢀ1—about the same intrinsic strain as in
the gauche conformation of butane. This strain is incorpo-
rated in the RCM steps of the synthesis. Homodesmotic
reactions (B3LYP/6-31G(d)) based on the lowest energy
conformations[35] of the various intermediates and 2,2’-dipro-
penylbiphenyl show that the first three RCM steps reduce
strain by about 9 kcalmolꢀ1 and the last four steps increase
Figure 5. The crystal packing of 1 (bottom) and 2 (top).[28]
Table 1: Energies (kcalmolꢀ1) of various conformations of 1.[30]
B3LYP
B3LYP
6-311G
(2d,p)
B3LYP
B97D
M06-2X
6-31G(d)
6-311G
(2df,2pd)
6-31G(d)
6-31G(d)
strain by about 14 kcalmolꢀ1. The energy released in creation
of an aromatic ring (ca. 28 kcalmolꢀ1
)
in each RCM step
[15]
chair
D7h
Cs
5.4
4.5
0.1
0
5.6
4.9
nd[a]
0
5.5
4.8
nd
0
5.8
5.3
0.1
0
5.8
5.2
0.1
0
more than compensates for the strain built up in any step.
The undulating structure of 1 gives rise to vibrations
similar to the “Mexican wave” cheer in sport stadiums,[36]
where spectators stand and sit in response to their neighbor,
thereby generating a wave that rotates around a stadium. For
the chair conformation of 1, this motion is a pseudorotation[37]
that occurs on a flat potential energy surface. The Cs and C2
gas-phase minima (B3LYP/6-31G(d)) lie on the pseudorota-
tion coordinate and, because of the fourteen-fold symmetry of
the potential energy surface (sevenfold rotation and a twofold
reflection), are interconverted by a motion of less than 138
along the pseudorotation coordinate. These minima are
isoenergetic within 0.1 kcalmolꢀ1.[38] The proximity of these
minima on the pseudorotation coordinate and their energetic
similarity hints at a low energy barrier.
The “Mexican wave” vibrational modes are remarkably
low in energy (2 cmꢀ1 for the C2 minimum and 4 cmꢀ1 for the
Cs minimum). These frequencies are comparable to the
uncertainty of the calculations. We therefore attempted to
locate a transition state for the pseudorotation. Geometrical
interpolation between the superimposed C2 and Cs confor-
mations provided a set of structures that lie on the pseudor-
otation trajectory. DFT calculations at points along this
trajectory reveal that the surface is flat (the calculated barrier
was 0.05 kcalmolꢀ1). The energy barrier to psuedorotation is
less than the uncertainty of the calculations. Similar to
cyclopentane,[39] 1 is a free pseudorotor. Not surprisingly,
a closely related pseudorotation is manifest in [7]circulene.[40]
In conclusion, septulene was prepared by a sevenfold
RCM reaction. Its conformational behavior is unusual: Two
saddle-shaped conformers were identified by gas-phase
calculations, and a chair-type conformer was observed in the
crystal structure. The two saddle-shaped conformers inter-
C2
[a] nd=not determined.
ment, and show that the saddle structures are lowest in energy
and the planar D7h and chair-shaped structures are about 5–
6 kcalmolꢀ1 higher in energy. So why does septulene adopt
a chair conformation in the solid state instead of its lower-
energy saddle conformation?
The answer to this question is evident from lattice-energy
calculations using force fields.[29] If one replaces the chair
conformation in the experimental crystal structure by a saddle
conformation and optimizes the lattice with fixed molecular
geometry, the lattice is drastically distorted. The resulting
packing has a lower density and an unfavorable lattice energy
(see the Supporting Information). Similar to DFT methods,
the force-field methods used favor a saddle conformation in
the gas phase. However, a full optimization from the saddle
conformation in the crystal packing leads back to the
experimental structure with molecules in a chair conforma-
tion. These calculations demonstrate that the preference in
the crystal for the chair conformation is a packing effect.
To investigate whether a saddle-type molecule can
crystallize with a different molecular arrangement we per-
formed a global crystal structure prediction with molecules in
Cs and C2 conformations.[31] The calculations were carried out
with fixed molecular geometries in various space groups with
free lattice parameters, starting from more than 300000
random crystal packing arrangements (details are given in the
Supporting Information). The energetically best resulting
structures were compared with the experimental structure
12798 www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 12795 –12800

Angewandte
Chemie
[13] I. Agranat, B. A. Hess, L. J. Schaad, Pure Appl. Chem. 1980, 52,
convert through a vanishingly small energy barrier, thus
making 1 a free pseudorotor. The chair conformation that is
observed in the crystal structure is not a gas-phase minimum,
but exists as a result of packing forces.
The electronic and structural properties of 1 are essen-
tially the same as those of 2. This similarity follows from their
Clar structures. The similarity does not—and indeed cannot—
follow from their Kekulꢁ structures, which fundamentally
differ, by virtue of their inner and outer annuli having an odd
or even number of atoms. Septulene (1) unequivocally
demonstrates that Clar structures, not Kekulꢁ structures,
should be used to represent bonding in polycyclic aromatic
molecules.
The conclusion that Clar structures are the correct way to
represent bonding in PAHs has not permeated popular
chemical thinking. Most textbook authors avoid circles in
aromatic rings altogether, an expedient work-around that
deprives the student of this simple and useful model. Indeed,
Carey and Sundberg draw kekulene 2 as concentric annu-
lenes,[41] even though the main lesson from kekulene was that
the concentric annulene model was wrong. In other textbooks,
circles are drawn willy-nilly, simply to indicate that a ring is
“aromatic” in some ill-defined sense, thus compounding the
suffering studentꢀs confusion. We recommend that chemists
treat circles in rings with the same respect and care that they
give to doubly barbed, curly arrows.
1399 – 1407.
[14] IUPAC Commission on Nomenclature of Organic Chemistry, A
Guide to IUPAC Nomenclature of Organic Compounds (Rec-
ommendations 1993), Blackwell Scientific Publications, Oxford,
1993.
[15] M. C. Bonifacio, C. R. Robertson, J. Y. Jung, B. T. King, J. Org.
Chem. 2005, 70, 8522 – 8526.
[16] A. Iuliano, P. Piccioli, D. Febbri, Org. Lett. 2004, 6, 3711 – 3714.
[17] S. K. Collins, A. Grandbois, M. P. Vachon, J. Cꢃtꢁ, Angew. Chem.
2006, 118, 2989 – 2992; Angew. Chem. Int. Ed. 2006, 45, 2923 –
2926.
[18] M. Schlosser, Chem. Ber. 1964, 3219 – 3233.
[19] A. D. Schlꢄter, J. Polym. Sci. Part A 2001, 39, 1533 – 1556.
[20] V. Hensel, K. Lꢄtzow, J. Jacob, K. Gessler, W. Saenger, A. D.
Schlꢄter, Angew. Chem. 1997, 109, 2768 – 2770; Angew. Chem.
Int. Ed. Engl. 1997, 36, 2654 – 2656.
[21] R. Kandre, A. D. Schlꢄter, Macromol. Rapid Commun. 2008, 29,
1661 – 1665.
[22] MALDI-TOF MS of the Suzuki polycondensation mixture
suggested the formation of macrocycles, from the cycloheptamer
to the cyclododecamer, with abundances decreasing with size.
[23] We note that the exhaustive RCM of the polymers should afford
helicene polymers.
[24] Z. Chen, C. S. Wannere, C. Corminboeuf, R. Puchta, P. von R.
Schleyer, Chem. Rev. 2005, 105, 3842 – 3888.
[25] X-ray data for septulene were collected on a Bruker APEX
CCD area detector with graphite-monochromated MoKa X-ray
source. The data were processed using SADABS (BRUKER,
2002) and structure was solved by direct methods (SHELXTL-
97). The structure was refined on F2 using full-matrix least-
squares (SHELXTL-97). Crystal data for septulene: C56H28;
Mr = 700.78, orthorhombic, a = 16.056(16), b = 30.14(3), c =
6.675(7), V= 3230(6) ꢂ3, T= 100 K, space group Pnma
(no. 62), Z = 4, 1 = 1.445 gcmꢀ3, 64919 reflections collected,
2897 unique (Rint = 0.153) which were used in all calculations.
Data completeness 100%, GOF = 0.902. The final R1 and wR2
(F2) were 0.0762 (I > 2s(I)) and 0.1225 (all data). CCDC 878569
contains the supplementary crystallographic data for this paper.
These data can be obtained free of charge from The Cambridge
Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_
request/cif.
[26] We use Mislowꢀs definition of conformation, “a particular
geometry of the molecule”, not the myopic IUPAC definition,
“the spatial arrangement of the atoms affording distinction
between stereoisomers which can be interconverted by rotations
about formally single bonds”: K. M. Mislow, Introduction to
Stereochemistry, Dover Publications, Mineola, NY, 2002;
IUPAC. Compendium of Chemical Terminology, 2nd ed. (the
“Gold Book”; Eds.: A. D. McNaught, A. Wilkinson), Blackwell
Scientific Publications, Oxford, 1997; XML on-line corrected
version: http://goldbook.iupac.org (2006-) created by M. Nic, J.
Jirat, B. Kosata; updates compiled by A. Jenkins. ISBN 0-
9678550-9-8. DOI: 10.1351/goldbook.
[27] For both 1 and 2, crystallographic symmetry ensures that the
herringbone angles are simply twice the inclination from the
stacking axis.
[28] The top figure (copyright 1983 Wiley) is used with permission.
[29] DREIDING force field (S. L. Mayo, B. D. Olafson, W. A.
Goddard III, J. Phys. Chem. 1990, 94, 8897 – 8909) within the
program suite Materials Studio (Materials Studio, Release 4.4,
Accelrys Software Inc., San Diego, 2008).
[30] These relative energies were obtained with a restricted version
of DFT as implemented in GAMESS suite of programs: M. W.
Schmidt, K. K. Baldridge, J. A. Boatz, S. T. Elbert, M. S.
Gordon, J. H. Jensen, S. Koseki, N. Matsunaga, K. A. Nguyen,
S. J. Su, T. L. Windus, M. Dupuis, J. A. Montgomery, J. Comput.
To close, we note the irony that molecules named in honor
of August Kekulꢁ have diminished the utility of a bonding
model also named in his honor.
Received: April 27, 2012
Revised: September 9, 2012
Published online: October 12, 2012
Keywords: aromaticity · fused-ring systems · kekulenes ·
.
macrocycles · metathesis
[1] H. A. Staab, F. Diedrich, Chem. Ber. 1983, 116, 3487 – 3503.
[2] H. A. Staab, F. Diederich, C. Krieger, D. Schweitzer, Chem. Ber.
1983, 116, 3504 – 3512.
[3] F. Diederich, H. A. Staab, Angew. Chem. 1978, 90, 383 – 385;
Angew. Chem. Int. Ed. Engl. 1978, 17, 372 – 374.
[4] C. Krieger, F. Diederich, D. Schweitzer, H. A. Staab, Angew.
Chem. 1979, 91, 733 – 735; Angew. Chem. Int. Ed. Engl. 1979, 18,
699 – 701.
[5] L. Pauling, J. Chem. Phys. 1936, 4, 673 – 677.
[6] R. McWeeny, Proc. Phys. Soc. London Sect. A 1951, 64, 921 –
930.
[7] E. Clar, The Aromatic Sextet, Wiley, New York, 1972.
[8] J.-M. Gilles, J. F. M. Oth, F. Sondheimer, E. P. Woo, J. Chem. Soc.
B 1971, 2177 – 2186.
[9] This conclusion is not new. Herndon discusses this point and
demonstrates the utility of considering all Kekulꢁ structures in:
W. C. Herndon, J. Am. Chem. Soc. 1973, 95, 2404 – 2406.
[10] A. Tatibouꢁut, R. Hancock, M. Demeunynck, J. Lhomme,
Angew. Chem. 1997, 109, 1238 – 1239; Angew. Chem. Int. Ed.
Engl. 1997, 36, 1190 – 1191.
[11] Z. E. Katsamanis, PhD thesis, Stony-Brook University (NY),
2005.
[12] L. Zhou, PhD thesis, Stony-Brook University (NY), 2006.
Angew. Chem. Int. Ed. 2012, 51, 12795 –12800
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 12799

.
Angewandte
Communications
Chem. 1993, 14, 1347 – 1363; “Advances in electronic structure
theory: GAMESS a decade later”: M. S. Gordon, M. W. Schmidt
in Theory and Applications of Computational Chemistry: The
First Forty Years (Eds.: C. E. Dykstra, G. Frenking, K. S. Kim,
G. E. Scuseria), Elsevier, Amsterdam, 2005, pp. 1167 – 1189.
[31] Program CRYSCA (M. U. Schmidt, U. Englert, J. Chem. Soc.
Dalton Trans. 1996, 2077 – 2082), using the DREIDING force
field for intermolecular interactions.
[32] J. Aihara, J. Am. Chem. Soc. 1992, 114, 865 – 868.
[33] H. Jiao, P. von R. Schleyer, Angew. Chem. 1996, 108, 2548 – 2551;
Angew. Chem. Int. Ed. Engl. 1996, 35, 2383 – 2386.
[34] C. Jerzy, OꢀC. Peter, F. D. Eugene, J. Am. Chem. Soc. 1991, 113,
1086 – 1089.
[35] Omega 2.3.2, Openeye Scientific Software, Santa Fe, New
Mexico (USA), 2008.
[36] I. Farkas, D. Helbing, T. Vicsek, Nature 2002, 419, 131 – 132.
[37] H. L. Strauss, Annu. Rev. Phys. Chem. 1983, 34, 301 – 328.
[38] The relative energies (0.02 kcalmolꢀ1) were obtained with
restricted version of DFTand 6-31G(d) basis set as implemented
in Gaussian suite of programs.
[39] W. Cui, F. Li, N. L. Allinger, J. Am. Chem. Soc. 1993, 115, 2943 –
2951.
[40] M. Shen, I. S. Ignatyev, Y. Xie, H. F. Schaefer III, J. Phys. Chem.
1993, 97, 3212 – 3216.
[41] F. A. Carey, R. J. Sundberg, Advanced Organic Chemistry, 4th
ed., Kluwer Academic/Plenum Publishers, New York, 2000,
p. 523.
12800 www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 12795 –12800