Dications of 3-phenyl-indenylidene dibenzo[ a. d ]cycloheptene: The role of charge in the antiaromaticity of cationic systems

Article abstract of DOI:10.1021/jo102220y

Dications of 9-(3-phenyl-1H-inden-1-ylidene)-5H-dibenzo[a,d]cycloheptene, 5²⁺, were prepared by oxidation with SbF₅ in SO ₂ClF, and their magnetic behavior was compared to dications of 9-(3-phenyl-1H-inden-1-ylidene)-9H-fl

Full text of DOI:10.1021/jo102220y

pubs.acs.org/joc
Dications of 3-Phenyl-indenylidene Dibenzo[a.d]cycloheptene: The Role
of Charge in the Antiaromaticity of Cationic Systems
Nancy S. Mills,* Francine E. Cheng, Joseph M. Baylan, Cornelia Tirla,†
Jennifer L. Hartmann, Kiran C. Patel, Bart J. Dahl,‡ and Sean P. McClintock
Department of Chemistry, Trinity University, San Antonio, Texas 78212-7200, United States.
†Current address: Department of Chemistry and Physics, University of North Carolina-Pembroke,
North Carolina. ‡Current address: Department of Chemistry, University of Wisconsin-Eau Claire,
Eau Claire, Wisconsin.
nmills@trinity.edu
Received November 6, 2010
Dications of 9-(3-phenyl-1H-inden-1-ylidene)-5H-dibenzo[a,d]cycloheptene, 52þ, were prepared
by oxidation with SbF5 in SO2ClF, and their magnetic behavior was compared to dications of
9-(3-phenyl-1H-inden-1-ylidene)-9H-fluorene, 22þ. The good correlation between the experi-
1
mental H NMR shifts for the dications that were oxidized cleanly and the chemical shifts
calculated by the GAIO method supported the use of the nucleus independent chemical shifts,
NICS, to evaluate the antiaromaticity of the indenyl systems of 22þ/52þ and their unsubstituted
parent compounds, 62þ and 72þ, as well as the antiaromaticity of the fluorenyl system of 22þ/72þ
and the aromaticity of the dibenzotropylium system of 52þ/62þ. Antiaromaticity was shown to be
directly related to the amount of charge in the antiaromatic systems, with the antiaromatic
systems more responsive to changes in the calculated NBO charge than the aromatic systems. The
antiaromaticity was also shown to be directly related to the amount of delocalization in the ring
system. The aromaticity of the dibenzotropylium system was much less responsive to changes in
the amount of charge in the tropylium system, because the aromatic system was much more
completely delocalized. Thus, antiaromatic species are more sensitive probes of delocalization
than aromatic ones.
Introduction
that contain an antiaromatic cationic fluorenyl system,7-18
for example, 12þ, and have recently extended those studies to
Substantial interest exists in the properties of antiaromatic
species1-6 but there are few examples of antiaromatic species
that can be prepared and characterized experimentally. We
have successfully prepared a number of hydrocarbon dications
(8) Dahl, B. J.; Mills, N. S. J. Am. Chem. Soc. 2008, 130, 10179–10186.
(9) Dahl, B. J.; Mills, N. S. Org. Lett. 2008, 10, 5605–5608.
(10) Herndon, W. C.; Mills, N. S. J. Org. Chem. 2005, 70, 8492–8496.
(11) Mills, N. S.; Tirla, C.; Benish, M. A.; Rakowitz, A. J.; Bebell, L. M.;
Hurd, C. M. M.; Bria, A. L. M. J. Org. Chem. 2005, 70, 10709–10716.
(12) Mills, N. S.; Benish, M. M.; Ybarra, C. J. Org. Chem. 2002, 67, 2003–
2012.
(13) Mills, N. S. J. Org. Chem. 2002, 67, 7029–7036.
(14) Levy, A.; Rakowitz, A.; Mills, N. S. J. Org. Chem. 2003, 68, 3990–
3998.
(15) Mills, N. S.; Levy, A.; Plummer, B. F. J. Org. Chem. 2004, 69, 6623–
6633.
(16) Mills, N. S.; Malinky, T.; Malandra, J. L.; Burns, E. E.; Crossno, P.
J. Org. Chem. 1999, 64, 511–517.
(17) Mills, N. S. J. Am. Chem. Soc. 1999, 121, 11690–11696.
(1) Tidwell, T. Chem. Rev. 2001, 101, 1333–1348.
(2) Breslow, R. In Magnetic Properties of Organic Materials; Lahti, P. M.,
Ed.; Marcel Dekker: New York, 1999; pp 27-40.
(3) Wiberg, K. B. Chem. Rev. 2001, 101, 1317–1331.
(4) Bally, T. Angew. Chem., Int. Ed. 2006, 45, 6616–6619.
(5) Herges, R. Nature 2007, 450, 36–37.
(6) Breslow, R.; Foss, F. W., Jr. J. Phys.: Cond. Matt. 2008, 20, 374104/
374101–374104/374107.
(7) Piekarski, A. M.; Mills, N. S.; Yousef, A. J. Am. Chem. Soc. 2008, 130,
14883–14890.
DOI: 10.1021/jo102220y
r
Published on Web 12/29/2010
J. Org. Chem. 2011, 76, 645–653 645
2010 American Chemical Society

Mills et al.
JOCArticle
dications containing an antiaromatic indenyl cation,19 22þ
,
fluorenyl/tropylium system on the antiaromaticity of the
indenyl system.
and to dianions containing an antiaromatic dibenzocyclo-
heptenyl system,7 32-. These systems have allowed us to
confirm that the aromaticity/antiaromaticity suggested by
nucleus independent chemical shifts is inversely related to the
square of the area of the ring system,20 have allowed an
assessment of the relationship between stability and mag-
netic measures of antiaromaticity,11 and have demonstrated
that the lack of bond length alternation seen as a character-
istic of aromatic compounds is also present in antiaromatic
species.7
Traditionally, the characterization of antiaromatic species
has been through theoretical calculation,3,21,22 which opens
the conclusions to concerns about how the calculation were
done, including issues of basis set choice. When theoretical
data can be shown to have a good correlation with experi-
mental data, the theoretical data is viewed as more reliable,
particularly by experimental chemists. We have been able to
successfully use this approach to “validate” measures of
aromaticity like nucleus independent chemical shifts, NICS,
which have no direct experimental analogue. NICS are the
The dicationic species that we have previously prepared
have all contained an antiaromatic ring system that is con-
nected to a substituent, 42þ, that can be a cyclic system as well
as an acylic substituent. The cyclic substituent can be an
aromatic ion, as in 12þ, or an antiaromatic ring system, as in
22þ. The antiaromaticity of the fluorenyl or indenyl ring
system is affected by the substituent, which is difficult to
rationalize because the fluorenyl ring is not planar with its
substituent Rþ. That is, the calculated geometry of 12þ shows
that the fluorenyl system is perpendicular to the dibenzo-
tropylium system,13 while for species such as 22þ, the calcu-
lated geometry shows that the dihedral between the indenyl
and fluorenyl systems is ∼50°, vide infra. If the ring systems
are not able to interact through resonance, what is the mode
of interaction?
We have chosen to examine one aspect of this through
the preparation of indenyl cations with 3-phenyl substit-
uents that are connected to a dibenzotropylium system,
52þ. By changing the nature of the p-substituent on the
phenyl ring, we can moderate the electronic environment
in the indenyl system and examine the effects of that
interaction on the antiaromaticity of the indenyl system
as well as on the aromatic dibenzotropylium system
(described as the tropylium system in the rest of the
paper). We have previously reported the preparation
and characterization of derivatives of 22þ and can use
those results to understand the effect of changes in the
indenyl system on its antiaromaticity as well as on the
antiaromatic fluorenyl system. The comparison of 22þ
with 52þ will also help to understand the effect of the
˚
chemical shifts calculated for a ghost atom placed 1 A above
the planar ring system.23 The chemical shift is a tensor, and
the component of the chemical shift parallel to the π-system
has been shown to most reliably correlate with other mea-
sures of aromaticity, so we will draw our conclusions
from these NICS values, designated as NICS(1)zz.24 If a
good correlation exists between the experimental shifts
and the NMR chemical shifts calculated by the same
method as the NICS values, the method for calculation
of those NMR shifts, and by extension the NICS values,
are considered to be more reliable. In addition, a good
correlation between experimental and calculated shifts
also supports the quality of the calculated geometry for
these reactive species.
Our approach with this work will be to examine the effect
on the antiaromaticity of the indenyl systems of changing the
substituent Z on 22þ and 52þ, to see if the changes are
transmitted to the fluorenyl and tropylium systems, and if
so, to explore any differences in the antiaromaticity/aroma-
ticity of the latter two systems. We will use the NICS values,
rather than the 1H NMR shifts to assess the antiaromaticity
of the indenyl/fluorenyl systems, or the aromaticity of the
1
tropylium system. There is sufficient variability in the H
NMR shifts that they can only be used to evaluate aroma-
ticity in a quantitative sense for compounds that are carefully
(21) Schleyer, P. v. R.; Freeman, P. K.; Jiao, H.; Goldfuss, B. Angew.
Chem., Int. Ed. Engl. 1995, 34, 337–340.
(22) Fowler, P. W.; Zansai, R.; Cadioli, B.; Steiner, E. Chem. Phys. Lett.
1996, 251, 132–140.
(23) Schleyer, P. v. R.; Maerker, C.; Dransfeld, A.; Jiao, H.; Hommes,
N. J. v. E. J. Am. Chem. Soc. 1996, 118, 6317–6318.
(24) Fallah-Bagher-Shaidaei, H.; Wannere, C. S.; Corminboeuf, C.;
Puchta, R.; Schleyer, P. v. R. Org. Lett. 2006, 8, 863–866.
(18) Malandra, J. L.; Mills, N. S.; Kadlecek, D. E.; Lowery, J. A. J. Am.
Chem. Soc. 1994, 116, 11622–11624.
(19) Mills, N. S.; Llagostero, K. B.; Tirla, C.; Gordon, S.; Carpenetti, D.
J. Org. Chem. 2006, 71, 7940–7946.
(20) Mills, N. S.; Llagostera, K. B. J. Org. Chem. 2007, 72, 9163–9169.
646 J. Org. Chem. Vol. 76, No. 2, 2011

Mills et al.
JOCArticle
SCHEME 1. Preparation of p-Substituted 5
matched in terms of structure.25 We have chosen to evaluate
the effect of the substituent Z through changes in the distribu-
tion of charge of the indenyl, fluorenyl, and tropylium
systems. In theory, charge on a carbon should be able to be
probed experimentally but there are issues with this
approach.26 As well there are issues with the calculation of
charge.27 Recently, Baldridge, et al.,28 have observed a good
relationship between experimentally determined charges and
those calculated by natural population analysis, natural
bond order (NBO) calculations.27 We will use the NBO
calculated charges to observe changes in charge distribution
and will couple those with the NICS values.
There is one caveat with this study. We are making
antiaromatic species and there is no guarantee that the
species we seek will be stable enough for characterization.
We were able only to prepare 2a-c2þ with high enough
purity to have confidence in the assignment of their 1H NMR
chemical shifts. We report here the preparation and char-
acterization of 5b-d2þ by oxidation of the unsaturated
precursor and were unable to get sufficiently high quality
data for the oxidation of 5a and 5e2þ to make reliable
assignments. We argue, however, that if the calculated and
experimental shifts for 5b-d2þ show a linear relationship,
there is no reason to assume that the spectra of 5a and 5e
would not also show the same linear relationship. Thus we
can interpolate the data with calculated data and be in a
better position to understand trends in the effects of sub-
stituent Z.
zocycloheptenes is shown in Scheme 1. Reaction of 1-indanone
with an appropriately substituted phenyl Grignard reagent,
followed by dehydration gave 1-(p-substituted-phenyl)-
indene. Deprotonation of the indene with n-butyllithium fol-
lowed by reaction with trimethylsilyl chloride gave the TMS
indene derivative. Subsequent deprotonation with n-butyl-
lithium was followed by reaction with dibenzosuberenone,
5H-dibenzo[a,d]cyclohepten-5-one, to give the indenylidene
dibenzo[a.d]cycloheptenes via Peterson olefination. The tri-
fluoromethyl derivative, 2a, was synthesized from the
p-bromophenyl derivative by reaction with sodium trifluoro-
acetate with copper(I) iodide.
Preparation of 52þ. NMR Shifts. Oxidation of 5b-5d with
SbF5 in SO2ClF gave 5b2þ-5d2þ. The 1H NMR spectrum of
5b2þ is shown in Figure 1; spectra of 5c2þ and 5d2þ can be
found in the Supporting Information. Assignments were
made via 2D COSY spectra and comparison with spectra
calculated using the GIAO method with basis set B3LYP/
6-311þg(d,p) on structures optimized at the B3LYP level. The
experimental spectrum of 5b2þ is not as pure as that of 5d2þ
.
There are additional peaks between 6.3 and 6.75 ppm, an
increased integration in the region between 6.9 and 7.8 and a
very large peak around 8, which may also contain protonated
water. However the resolution is much better than seen in the
spectrum of 5d2þ. The presence of the methyl substituent in
5d2þ would be expected to provide additional stabilization to
the dication. Oxidation of 5c resulted in the formation of the
dication, but the spectrum was much less clean. It was
possible to make assignments for 5c2þ based primarily on the
COSY spectrum and the calculated chemical shifts. The
experimental and calculated spectra are given in Table 1.
The calculated shifts clearly show the upfield shift antici-
pated for the antiaromatic indenyl system and the down-field
shift for the aromatic dibenzotropylium system. These same
regions are apparent in the experimental spectrum. A plot of
calculated vs experimental shifts for 5b2þ-5d2þ is shown in
Figure 2 and shows excellent agreement, with R = 0.987. The
quality of the agreement between experimental and calcu-
lated shifts gives support to NICS calculations using the
same method and basis sets, vide infra. Attempts to oxidize
Results and Discussion
Preparation of Substituted 3-Phenyl-indenylidene Dibenzo-
cycloheptenes and their Oxidation to Dications 52þ. The
synthetic strategy for preparation of the indenylidene diben-
(25) Mitchell, R. H. Chem. Rev. 2001, 101, 1301–1316.
(26) Wiberg, K. B.; Hammer, J. D.; Keith, T. A.; Zilm, K. Tetrahedron
Lett. 1997, 38, 323–326.
(27) Reed, A. E.; Weinstock, R. B.; Weinhold, F. J. Chem. Phys. 1985, 83,
735–746.
(28) Yerushalmi, R.; Scherz, A.; Baldridge, K. K. J. Am. Chem. Soc. 2004,
126, 5897–9505.
J. Org. Chem. Vol. 76, No. 2, 2011 647

Mills et al.
JOCArticle
FIGURE 1. 1H NMR spectrum of 5b2þ in SO2ClF, with d6- acetone/TMS as an external reference.
TABLE 1. Calculateda 1H NMR Shifts for 5a-e2þ and Experimentalb
Shifts for 5b-d2þ
5a2þ 5b2þ
5b2þ
5c2þ 5c2þ 5d2þ
5d2þ
5e2þ
proton
number calcd calcd
exp
calcd exp calcd
exp
calcd
Indenyl system
2
4
5
6
7
6.399 6.740
7.576 6.429
6.984 7.324
7.028 7.382
6.271 7.910
5.969 5.753 5.700 6.827
5.258 6.915 6.692 7.991
6.099 6.679 5.959 7.431
6.179 6.673 5.880 7.365
6.818 5.461 5.056 6.473
5.817 6.844
6.458 8.084
6.024 7.295
5.935 7.243
5.090 6.514
Dibenzotropylium system
1,10
2,9
3,8
4,7
5,6
10.017 9.692 ∼8.0
8.138 7.673 9.609 ∼8.0
9.553
7.089 8.477
7.373 8.833
7.512 8.892
8.629 8.797
8.880 9.063
8.925 9.134
7.359 7.621 7.095 8.761
7.656 8.072 7.494 9.041
7.737 7.888 7.356 9.119
9.289 9.518 ∼8.0
8.314 8.097 9.507 ∼8.0
9.259
Phenyl ring
8.380 8.376 ∼6.95 7.043 7.206 8.302
8.711 8.746 7.341 7.694 7.727 8.714
o
o
m
m
p
subst
7.029 8.050
7.583 9.094
6.607 7.271
6.825 7.436
8.030 8.113 ∼6.95 6.805 6.247 7.926
FIGURE 2. Comparison of calculated and experimental shifts for
5b2þ, 5c2þ, and 5d2þ. See Table 1 for details of the calculations and
experimental conditions.
8.190 8.264 ∼6.95 6.962 6.384 8.106
8.757
7.419
2.671
1.511 4.274
aShifts calculated with the GIAO method with the inclusion of solvent
using the PCM method using DMSO as the solvent with basis set
B3LYP/6-311þg(d,p) on geometries optimized at the B3LYP/6-31 g(d)
level. bChemical shifts determined in SO2ClF, referenced to an external
capillary with d6-acetone and TMS.
and these predictions have been given experimental support
1
in the H NMR characterization of [18]annulene and its
dianion.29 Nucleus independent chemical shifts, NICS, are
based on this phenomenon, with the calculation of the magnetic
shielding tensor for a dummy atom placed in the center of the
ring system.23 NICS for an aromatic system have a negative
sign; those for an antiaromatic system have a positive sign.
As mentioned earlier, the method has been refined by the
5a2þ and 5e2þ were unsuccessful. It is not surprising that the
highly electron-withdrawing trifluoromethyl substituent
would destabilize the carbocation, but we had anticipated
that the methoxy substituent on 5e2þ would stabilize the
carbocation. However, we have seen complexation of SbF5 with
a methoxy substituent in other systems16 and its complexation
would reduce the stabilizing effect of the methoxy group.
Nucleus Independent Chemical Shifts. The ring current
model of aromaticity predicts a magnetic environment in
the center of an aromatic ring that is opposite in sign to that
of the periphery, resulting in protons that are shifted down-
field when on the periphery of the aromatic species and would
be shifted upfield if present in the center of the aromatic ring.
These effects would be opposite for an antiaromatic species
˚
placement of the dummy atom 1 A above the plane of the
ring to avoid effects of electrons in σ-bonds in small ring
systems, and by the use of the component of the magnetic
shielding tensor that is perpendicular to the plane of the ring,
NICS(1)zz.24 We have recently shown that the summation of
the NICS values for each ring in a polycyclic system is an
accurate measure of the aromaticity/antiaromaticity of the
polycyclic system. We also determined that the magnitude of
(29) Oth, J. F. M.; Woo, E. P.; Sondheimer, F. J. Am. Chem. Soc. 1973,
95, 7337–7345.
648 J. Org. Chem. Vol. 76, No. 2, 2011

Mills et al.
JOCArticle
TABLE 2. Nucleus Independent Chemical Shiftsa, NICS, for 5a-e2þ and 62þ
indenyl system
dibenzotropylium system
5-ring
6-ring
sum
NICS/area2
7-ring
6-ring
sum
NICS/area2
total
5a2þ
5b2þ
5c2þ
5d2þ
5e2þ
62þ
44.64
41.49
38.45
36.44
31.54
90.33
4.26
1.13
1.24
-2.48
-5.12
41.34
48.90
42.62
39.69
33.97
26.42
131.68
0.675
0.576
0.532
0.456
0.356
1.760
-20.72
-20.62
-17.07
-18.88
-17.71
-16.23
-30.53
-30.58
-24.75
-27.51
-26.10
-30.73
-81.78
-81.78
-66.58
-73.90
-69.92
-77.69
-0.261
-0.279
-0.237
-0.249
-0.222
-0.247
-32.87
-39.17
-26.89
-39.93
-43.49
53.99
aCalculated with the GIAO method and basis set B3LYP/6-311þg(d,p) on geometries optimized at the B3LYP/6-31 g(d) level.
dibenzotropylium system with a fluorenyl cation, that is,
comparing the antiaromaticity of the indenyl system of 52þ
with that of 22þ, and on the effect of those changes on the
tropylium and fluorenyl systems, respectively. We had pre-
viously prepared dications 2a-c2þ by oxidation with SbF5.19
To assess the effect on antiaromaticity/aromaticity of replac-
ing the dibenzotropylium system with a fluorenyl system,
we rely on comparison of NICS. However, the NICS(1)zz
values for 2a-e2þ previously reported were done at a differ-
ent basis set and used the average of the three magnetic
tensors rather than the tensor perpendicular to the plane of
the ring system, so the values were recalculated and are
reported in Table 3, along with that of the unsubstituted
indenylidene fluorene dication, 72þ. The comparison of
FIGURE 3. Optimized geometry for 5b2þ. (a) View showing all
atoms. (b) Orientation showing the perpendicular relationship of
the indenyl and tropylium ring systems.
1
experimental vs H NMR shifts for 2a-c2þ is given in the
Supporting Information. The agreement is very good, R =
0.967, and improves to 0.984 when data for 5b-d2þ is
included. The excellent agreement when both systems are
examined supports the use of calculations with GIAO at the
B3LYP/6-311þg(d,p) level for NICS.
the NICS value is inversely related to the square of the ring area,
which must be considered in the use of NICS to evaluate
aromaticity/antiaromaticity in ring systems of different sizes.20
The calculated NICS(1)zz values for 5a-e2þ are given in
Table 2, along with the NICS(1)zz values of the unsubstituted
indenylidene dibenzotropylium dication, 62þ, vide infra. The
antiaromaticity of the indenyl systems are evident through
the positive sign of NICS and the aromaticity of the diben-
zotropylium systems is supported by their negative NICS.
They were calculated using the GIAO method with basis set
B3LYP/6-311þg(d,p) on geometries optimized at the
B3LYP/6-31 g(d) level using Gaussian03. As noted prev-
iously, this approach gives good agreement between calcu-
As was true for 52þ, the antiaromaticity of the indenyl
system decreased with phenyl substitution and as Z became
more electron-rich. The antiaromaticity of the fluorenyl
system behaved in an analogous way. We asked what role
the substituted phenyl ring could be playing in affecting the
antiaromaticity of the indenyl and fluorenyl ring systems and
it seemed reasonable that the effect of Z might be on the
stabilization of the positive charge in the indenyl system; that
as Z became more electron-donating, it would stabilize the
positive charge by delocalizing it into the phenyl ring. We
chose to calculate the charges by the natural population
analysis method, giving natural bond order (NBO) charges
which have been shown to represent experimental charges
reliably.28 The table of the calculated NBO charges for each
atom of 22þ, 52þ, 62þ and 72þ can be found in the Supporting
Information. Table 4 contains the sum of the charges for
1
lated and experimental H NMR shifts. It is clear that the
more electron-deficient the 3-phenyl substituent, the more
antiaromatic the indenyl system becomes and that the greatest
antiaromaticity in the indenyl system is seen for 62þ
,
which lacks the stabilization of the phenyl ring. The opti-
mized geometry of 5a-e2þ reveals that the indenyl and
dibenzotropylium ring systems are perpendicular to each
other, as is shown in Figure 3 for 5b2þ
.
each ring system of 22þ, 52þ, 62þ and 72þ
.
Examining the unsubstituted systems, 62þ and 72þ first,
the charges are unevenly partitioned between the ring sys-
tems, with slightly more positive charge on the fluorenyl and
tropylium systems. This may simply reflect the greater ability
of the larger system to support a positive charge. As would be
expected, the presence of the phenyl substituent in the
3-position causes additional delocalization of the positive
charge throughout the entire indenyl system, with greater
movement of charge into that system as Z becomes more
electron-donating and therefore more capable of stabilizing
the charge. While the magnitude of the sum of the charges on
the phenyl substituent as a function of substituent are not
equal, for example, the charge on the anisyl ring of 2e2þ is
Comparison with Related Dications: NICS. As was de-
scribed previously, we were interested in the effect on the
antiaromaticity of the indenyl system of replacing the aromatic
J. Org. Chem. Vol. 76, No. 2, 2011 649

Mills et al.
JOCArticle
TABLE 3. Nucleus Independent Chemical Shifts,a NICS(1)zz, for 2a-e2þ and 72þ
indenyl system
fluorenyl system
5-ring
6-ring
sum
NICS/area2
5-ring
6-ring
sum
NICS/area2
total
2a2þ
2b2þ
2c2þ
2d2þ
2e2þ
72þ
43.89
40.38
38.15
33.22
28.22
87.09
7.03
4.28
2.40
-2.47
-5.10
47.46
50.91
44.66
40.55
30.75
23.11
134.54
0.677
0.596
0.539
0.410
0.307
1.793
68.61
66.99
66.91
65.70
62.45
87.38
25.93
24.30
23.97
22.57
19.25
52.95
120.48
115.60
114.78
110.83
100.94
186.17
0.631
0.605
0.601
0.579
0.527
1.018
171.39
160.25
155.41
141.57
124.05
320.71
aCalculated with the GIAO method and basis set B3LYP/6-311þg(d,p) on geometries optimized at the B3LYP/6-31 g(d) level.
TABLE 4. Summation of calculateda NBO charges by ring systemb for
One possibility is that the delocalization of charge, rather
than simply the amount of charge, is very important in
the manifestation of aromaticity and antiaromaticity, con-
ceivably that the greater the delocalization, the greater the
aromaticity/antiaromaticity. Delocalization can be quantified
by considering the lack of bond length alternation. We have
previously observed with the dications and dianions of
tetrabenzo[5.7]fulvalenes that aromatic systems are much
less responsive to changes in the amount of bond length
alternation than are antiaromatic systems. That is, when the
optimized geometry for an aromatic system was modified so
that the overall area was unchanged but the bond length
alternation was reduced, the effect on aromaticity was very
small, while the effect of similar changes for an antiaromatic
system was much greater.7 Thus antiaromatic systems are much
more responsive to the effectiveness of the delocalization in
the system, as evaluated through bond length alternation.
Evaluation of Bond Length Alternation using the GEO
Term of the Harmonic Oscillator Model of Aromaticity. Bond
length alternation is the deviation of each bond length from
the average bond length. This measure is considered as one
term of the evaluation of structural effects on aromaticity/
antiaromaticity described as the Harmonic Oscillator Model
of Aromaticity, HOMA.30-32 The HOMA approach con-
tains two terms; the GEO term considers the effect of the
degree of bond length alternation while EN term examines
the deviation of the average bond length from the optimal
one, Ropt, on the energy of an aromatic compound. The
degree of bond length alternation is obtained by the summa-
tion of the square of the difference of each bond, Ri, from the
average bond length, Rav, divided by the number of bonds, as
shown below. For benzene both the GEO and EN terms
would be 0, thus the HOMA value would be 1. The coeffi-
cient R was chosen so that the HOMA value for a localized
polyene would be 0; n is the number of bonds. The larger the
value of the geometric term, GEO, the greater is the degree of
bond length alternation. The calculated HOMA GEO terms
for 22þ, 52þ, 62þ and 72þ are shown in Table 5. The calculated
bond lengths from which the GEO terms were determined
can be found in the Supporting Information.
22þ, 52þ, 62þ and 72þ
ΣNBO
charge
phenyl
ind
Fl
phenyl
ind
trop
2a2þ
2b2þ
2c2þ
2d2þ
2e2þ
72þ
0.429
0.468
0.504
0.541
0.651
0.630 0.941 5a2þ
0.601 0.931 5b2þ
0.568 0.928 5c2þ
0.544 0.921 5d2þ
0.462 0.887 5e2þ
0.999 1.001 62þ
0.399
0.434
0.466
0.500
0.598
0.588 1.018
0.554 1.012
0.524 1.010
0.495 1.006
0.404 0.998
0.954 1.046
aCharges calculated with the NPA method in Gaussian 09 with
B3LYP/6-311 g(d,p) on geometries optimized with B3LYP/6-31 g(d).
bCharges from hydrogens were summed into the charges of the carbons
to which they were attached.
0.651 whereas for 5e2þ it is 0.598, the substituents are causing
a proportional response, as is shown by a linear plot of the
summation of charges by substituent (see Supporting Informa-
tion). One would expect that as charge is removed from the
antiaromatic systems, the magnitude of the antiaromaticity
would decrease; a decrease in the amount of charge would
also result in a decrease of the aromaticity of the tropylium
system. These changes should be apparent in changes in the
NICS values for each ring system, ΣNICS. Because the
magnitude of the NICS values are related to the size of the
ring, they must be divided by the square of the area, ΣNICS/
area2.20 The magnitude of delocalization of the charge in
each ring system is also affected by the area. We compared
ΣNICS/area2 to ΣNBO charges/area. The plots of ΣNICS/
area2 to the ΣNBO charges/area2 were very similar in terms
of the quality of the correlation because there was relatively
little change in the ring areas for the different systems (see
Supporting Information). The plots of ΣNICS/area2 vs ΣNBO
charges/area are given in Figure 4. Figure 4a shows the linear
relationship for both indenyl systems of 22þ, 52þ, 62þ and
72þ; Figures 4b shows the linear relationships for the fluo-
renyl and tropylium substituents. The positive slopes for the
antiaromatic systems show that as the amount of positive
charge increases, the amount of antiaromaticity increases. That
is, the ability of the electron-donating substituent to stabilize
the cationic system is related to a decrease in antiaromaticity.
Figure 4 suggests at least two questions. The slopes for the
indenyl systems are very similar, and are around 25 ppm/
charge, but the slope of the fluorenyl system is more steep,
about 60 ppm/charge. The steeper slope suggests that anti-
aromaticity of the fluorenyl system is more sensitive to charge
than the antiaromaticity of the indenyl systems. Antiaroma-
ticity must be affected by more than simply the charge
present in the system. The more obvious question is the flat
slope of the tropylium system. This system, which is the only
aromatic system in the species examined, shows a marked
lack of responsiveness of aromaticity to changes in charge.
HOMA ¼ 1 - ½GEO þ ENꢀ
X
2
2
¼ 1 - ½R=n
ðRaν - RiÞ þ RðRopt - RaνÞ ꢀ
(30) Kruszewski, J.; Krygowski, T. M. Tetrahedron Lett. 1972, 3839–
3842.
(31) Krygowski, T. M.; Cyranski, M. K. Tetrahedron 1999, 55, 11143–
11148.
650 J. Org. Chem. Vol. 76, No. 2, 2011

Mills et al.
JOCArticle
FIGURE 4. Relationship between the magnetic measure of aromaticity/antiaromaticity, corrected for the area of the ring system, ΣNICS/sq
area, with the charge of that ring system, ΣNBO charges/area, for 22þ, 52þ, 62þ and 72þ. (a) Indenyl systems. (b) Fluorenyl (22þ/72þ) and
tropylium (52þ/62þ systems.
TABLE 5. Bond Length Alternation from Calculated Geometry of 22þ
52þ, 62þ and 72þa
,
The GEO term shows the degree of bond alternation with
a smaller value indicating less bond length alternation,
and therefore greater delocalization. Considering first the
indenyl system of 52þ, there is an increase in bond length
alternation of the entire system as the substituent Z becomes
more electron donating. The bond length alternation
becomes greater in the 5-membered ring as the substituent
becomes more electron-donating, but less in the 6-membered
ring. Thus, the 6-membered ring becomes more benzene-like,
more completely delocalized as the 5-membered ring
becomes less completely delocalized. This suggests that the
pattern of delocalization changes from that shown for 52þ to
that indicated as 5*2þ. In effect, as the phenyl ring becomes
more able to stabilize the carbocation, it begins to behave as
an allyl cation.
52þ/ 62þ
5a2þ
5b2þ
5c2þ
5d2þ
5e2þ
62þ
indenyl
5-ring
6-ring
tropylium
5-ring
6-ring
0.3577
0.4220
0.1041
0.2008
0.1861
0.2296
0.3678
0.4426
0.0968
0.2010
0.1863
0.2289
0.3698
0.4505
0.0893
0.2013
0.1866
0.2289
0.3751
0.4604
0.0863
0.2011
0.1870
0.2274
0.3893
0.4970
0.0672
0.2008
0.1873
0.2249
0.2733
0.2146
0.1997
0.2087
0.2220
0.2315
22þ/72þ
2a2þ
2b2þ
2c2þ
2d2þ
2e2þ
72þ
indenyl
5-ring
6-ring
fluorenyl
5-ring
6-ring
0.3501
0.4035
0.0504
0.1809
0.0958
0.0842
0.3556
0.4141
0.0466
0.1809
0.0957
0.0829
0.3538
0.4145
0.0424
0.1809
0.0955
0.0824
0.3555
0.4184
0.0396
0.1809
0.0952
0.0804
0.3504
0.4016
0.0311
0.1809
0.0943
0.0768
0.3204
0.3066
0.2036
0.1821
0.0929
0.0971
aCalculated using the geometric term from the HOMA method; R =
257.7; geometries optimized at the B3LYP/6-31G level.
requirements than the tropylium system because the internal
bond angles of the five-membered ring of the fluorenyl
system are smaller. That is, the benzene rings of the fluorenyl
system are effectively tied back more than in the tropylium
system. The fluorenyl and indenyl systems can move closer to
planarity than was true for the tropylium and indenyl
systems of 52þ/62þ. That allows the fluorenyl system to
interact with the indenyl system to a greater extent than
was true for the interaction of the tropylium system with the
indenyl system of 52þ/62þ. A partial consequence of this
might account for the greater variability of the dihedral
angles for the 5-membered ring of 22þ compared to 52þ
(see Supporting Information). Support for the change in the
pattern of delocalization from 52þ to 5*2þ comes from a
consideration of the NICS values. The 6-membered rings of
both 22þ and 52þ become more aromatic and more negative
as Z becomes more electron donating while the 5-membered
rings become less antiaromatic.
There is a similar pattern in the 6-membered ring of the
indenyl system of 22þ but more variability in the 5-membered
ring, and therefore in the bond length alternation of the
entire indenyl system. We are inclined to attribute this
variability to differences in geometry. As was demonstrated
in Figure 3, the dihedral angle between the ring systems of
52þ is 90°, which is a consequence of the greater spacial
requirements of the tropylium system because of the larger
bond angles in the seven-membered ring. In comparison, the
dihedral angle of 22þ is approximately 50° (see Supporting
Information). The fluorenyl ring system has smaller spacial
While the indenyl systems show a change in the pattern of
delocalization as a consequence of the effect of Z, both the
fluorenyl system of 22þ and tropylium system of 52þ show
much less bond length alternation in all ring systems, there-
fore much greater delocalization. This is reflected in the
greater antiaromaticity of both 5- and 6-membered rings of
(32) Krygowski, T. M.; Cyranski, M. K. Chem. Rev. 2001, 101, 1385–
1419.
J. Org. Chem. Vol. 76, No. 2, 2011 651

Mills et al.
JOCArticle
22þ, Table 3, and greater aromaticity of both rings of 52þ
,
presumably because they were already substantially delo-
calized.
Table 2. The tropylium system shows much less change in
NICS as a function of charge, Figure 4b, because it is already
appreciably delocalized.
Thus, subtle changes in electronic effects are much more
visible in antiaromatic systems, as measured through bond
length alternation, changes in the amount of charge, and
measures of antiaromaticity. We suggest that using antiaro-
matic species to study delocalization is both possible to do
experimentally and is a very powerful tool.
When the unsubstituted species 62þ and 72þ are compared
to the dications with phenyl substituents, there is greater
delocalization in the indenyl systems because there is no
phenyl substituent to shift the pattern of delocalization to
one resembling 5*2þ. However, both the tropylium system of
62þ and the fluorenyl system of 72þ show very slightly more
Experimental Section
bond length alternation in comparison to 52þ/22þ
.
The olefin precursors to 5b-e2þ were synthesized by Peterson
olefination of the appropriate substituted 3-phenyl indene19
with dibenzosuberenone, as described below for the synthesis
of 5b. 5a was prepared from 3-(p-bromophenyl)-1-(1H-
indenylidene)-5H-dibenzo[a.d]cycloheptene, which was synthe-
sized by Peterson olefination of 1-(4-bromophenyl)-indene with
dibenzosuberenone, followed by reaction with sodium trifluor-
oacetate in the presence of copper(I) iodide, as described below.
1H and 13C NMR spectra for 5a-e can be found in the
Supporting Information.
Summary
Dications 5b-d2þ were prepared by oxidation with SbF5
in SO2ClF and characterized by 1H NMR spectroscopy.
There was an excellent relationship between the experimen-
tal and calculated chemical shifts for 5b-d2þ as well as for
previously prepared 2a-c2þ, which gave validity to the
NICS values calculated using the same method. The behav-
ior of 22þ/72þ, which contained indenyl and fluorenyl
cationic ring systems and of 52þ/62þ, which contained
indenyl and tropylium cationic ring systems was examined by a
magnetic measure of aromaticity/antiaromaticity, ΣNICS-
(1)zz. By that measure, the antiaromaticity of the indenyl ring
system decreased with the introduction of a phenyl substi-
tuent in the 3-position. In addition, as substituents on that
phenyl ring became more electron rich, the antiaromaticity
of the indenyl ring system decreased. This suggested that the
amount of charge in the system might be related to the degree
of antiaromaticity. The sum of the calculated NBO charge
showed a good correlation with the amount of antiaromati-
city in the indenyl system. A similar examination of the
fluorenyl systems of 22þ/72þ showed a linear relationship
between ΣNICS(1)zz and ΣNBO charges, but in the tropy-
lium systems of 52þ/62þ, the degree of aromaticity was
basically unresponsive to changes in the degree of charge.
However, the responsiveness of the antiaromaticity of the
fluorenyl system to changes in charge was much greater than
the responsiveness of the indenyl system; the amount of
charge alone was not primarily responsible for changes in
antiaromaticity. To determine if the lack of bond length
alternation, a measure of delocalization, might play a role,
bond length alternation was calculated using the GEO term
of the HOMA approach. The 5-membered ring of the
indenyl system of 52þ showed an increase in bond alternation
while the 6-membered ring became more benzene-like and
3-(Phenyl)-1-(1H-indenylidene)-5H-dibenzo[a.d]cycloheptene,
5b. To 1-phenyl-indene (1.11 g, 5.77 mmol) in 40 mL of dry THF
-78 °C was added 5.4 mL n-butyllithium (7.9 mmol), giving a
dark red solution. After 1.5 h, trimethylsilyl chloride (1.1 mL,
8.1 mmol) was added to the reaction mixture and the reaction
allowed to stir for two hours, with the temperature warming
to -15 °C. An additional 1.5 equiv of n-butyllithium were added
and the reaction mixture stirred for 2 h. Dibenzosuberenone
1.19 g, 5.77 mmol) was added and stirred overnight at room
temperature. The solution was quenched with water and the
aqueous layer was extracted with 2 ꢁ 50 mL of diethyl ether. The
combined organic layers were washed with with 2 ꢁ 25 mL
saturated sodium bicarbonate. The solvent was removed under
vacuum to yield a brown oil. Flash column chromatography
with hexane was performed to yield the pure product, yellow
crystals, 0.973 g, 44.2% yield. MP 172-174 °C.
1H NMR (400 MHz, CDCl3) δ 7.53-7.62 (m, 4H), 7.30-7.53
(m, 10H), 7.17 (tt, J=7.52, 1.18 Hz, 1H), 7.02 (d, J=11.82 Hz,
1H), 6.96 (d, J=11.80 Hz, 1H), 6.90 (t, J=7.58 Hz, 1H), 6.60
(d, J = 1.97 Hz, 1H), 6.59 (d, J = 4.66 Hz, 1H); 13C NMR
(100 MHz, CDCl3) δ 145.0, 143.7, 143.2, 141.5, 139.3, 138.5,
137.4, 136.6, 135.9, 134.6, 134.1, 131.4, 131.2, 128.9, 128.8,
128.8, 128.7, 128.6, 128.3, 128.2, 128.1, 127.9, 127.4, 127.4,
126.4, 125.3, 124.1, 120.3. Analysis calculated for C30H20: C,
94.70; H, 5.30. Found: C, 94.40; H, 5.81.
3-(p-Fluorophenyl)-1-(1H-indenylidene)-5H-dibenzo[a.d]cyclo-
heptene, 5c. Same procedure as for 5b, yellow crystals, 72%
yield. MP 173-174 °C.
1H NMR (400 MHz, CDCl3) δ 7.53-7.23 (m, 11H), 7.10 (td,
J=7.83, 1.23 Hz, 1H), 7.01 (t, J=8.91 Hz, 2H), 6.93 (d, J =16.40
Hz, 1H), 6.88 (d, J = 16.40 Hz, 1H), 6.83 (td, J = 7.31, 1.19 Hz,
1H), 6.52 (d, J = 7.83 Hz, 1H), 6.47 (s, 1H); 13C NMR (100
MHz, CDCl3) δ 164.4, 161.1, 144.0, 143.0, 139.2, 138.3, 137.3,
136.5, 134.6, 134.1, 131.9, 131.3, 131.2, 129.6, 129.5, 128.8,
128.8, 128.7, 128.4, 128.2, 128.1, 127.5, 126.2, 125.4, 124.2,
120.1, 115.9, 115.6. Analysis calculated for C30H19F: C, 90.43;
H, 4.81; F, 4.77. Found: C, 90.06; H, 4.70.
showed a change in the pattern of delocalization to 5*2þ
.
This change in the pattern of delocalization was apparent in
the behavior of the 6-membered ring of 22þ but was less clear
for the 5-membered ring, presumably because of different
geometric constraints in the system. Both the fluorenyl
system of 22þ and the tropylium system of 52þ showed much
greater delocalization, as shown by smaller bond length
alternation. Finally, for systems without the phenyl substi-
tuent, 62þ and 72þ, the indenyl ring showed much greater
delocalization than systems with the phenyl substituent. For
the indenyl and fluorenyl systems, greater delocalization
resulted in greater antiaromaticity, as evaluated through
NICS. Aromatic systems, such as the tropylium system of
52þ/62þ, were much less responsive to changes in delocaliza-
tion caused by changes in the degree of charge in the system,
3-(p-Methylphenyl)-1-(1H-indenylidene)-5H-dibenzo[a.d]cyclo-
heptene, 5d. Same procedure as for 5b, with the following
exceptions: purification was achieved by column separation
and recrystallization in pentane. Yellow crystals, 48.4% yield.
MP 189-192 °C.
1H NMR (400 MHz, CDCl3) δ 7.60-7.52 (m, 2H) 7.52-7.31
(m, 9H), 7.22 (t, J=8.47 Hz, 1H), 7.20 (t, J=6.85 Hz, 1H), 7.17
(t, J=7.48 Hz, 1H), 7.01 (d, J=11.8 Hz, 1H), 6.96 (d, J=11.8 Hz,
1H), 6.89 (t, J=7.62 Hz, 1H), 6.58 (d, J=8.02 Hz, 1H), 6.57 (s, 2,
652 J. Org. Chem. Vol. 76, No. 2, 2011

Mills et al.
JOCArticle
1H), 2.38 (s, 3H); 13C NMR (100 MHz, CDCl3) δ 145.0, 143.1,
139.3, 138.5, 138.0, 137.5, 136.6, 134.6, 134.1, 133.0, 131.3,
131.2, 129.4, 128.9, 128.8, 128.8, 128.7, 128.3, 128.1, 128.0, 127.8,
127.4, 127.4, 125.9, 125.2, 124.1, 120.3, 21.4. Analysis calculated
for C29H20: C, 94.38; H, 5.47. Found: C, 94.01; H, 5.60.
3-(p-Methoxyphenyl)-1-(1H-indenylidene)-5H-dibenzo[a.d]cyclo-
heptene, 5e. Same procedure as for 5b yellow crystals, 37% yield.
MP 204-205.5 °C.
a vortex stirrer until homogeneous, and the solution was cooled
to -78 °C. The neutral precursor (∼3 mmol) was added in small
portions, followed by vortex mixing and cooling to -78 °C.
Samples for NMR analysis were kept at -78 °C until needed and
transferred by chilled pipet into a chilled NMR tube. A capillary
tube with acetone-d6 was then inserted into the NMR tube to
serve as an external standard and deuterium lock.
1H NMR (400 MHz, CDCl3) δ 7.60-7.30 (m, 9H), 7.42-7.31
(m, 2H), 7.17 (tt, J=7.52, 1.26 Hz, 1H), 7.02 (d, J=11.8 Hz, 1H),
6.96 (d, J=11.8 Hz, 1H), 6.94 (d, J=7.56 Hz, 2H), 6.89 (t, J =
7.62 Hz, 1H); 13C NMR (400 MHz, CDCl3) δ159.6, 144.6, 143.5
143.0, 139.4, 138.5, 137.5, 136.6, 134.6, 134.2, 131.3, 131.2,
129.1, 128.9, 128.8, 128.8, 128.7, 128.3, 128.1, 127.4, 126.1,
125.3, 125.2, 123.9, 120.3, 114.4, 55.4. Analysis calculated for
C31H22O: C, 90.70; H, 5.40; O, 3.90. Found: C, 90.34; H, 5.38.
3-(p-Trifluoromethylphenyl)-1-(1H-indenylidene)-5H-dibenzo-
[a.d]cycloheptene, 5a. A solution of 3-(p-bromophenyl)-1-(1H-
indenylidene)-5H-dibenzo[a.d]cycloheptene (0.4594 g, 1.000 mmol)
in 40 mL of dry N-methyl-2-pyrrolidone was purged with argon
for 30 min. Copper(I) iodide (0.762 g, 4.000 mmol), which had
been purified by Soxhet extraction with CH2Cl2, and NaO2-
CCF3 (1.08 g, 8.00 mmol) were added and the solution refluxed
under Ar overnight. After cooling to room temperature, 250 mL
ether was added and the reaction mixture filtered through Celite.
The solution was extracted with twice with water and twice with
brine. The solution was dried, solvent removed under vacuum,
and the product isolated by chromatography on silica with
gradient eleution of hexanes to 5% CH2Cl2/hexanes to give
yellow crystals. Yield: 0.250 g, 55.8%. MP 187-191 °C.
H1 NMR (CDCl3, TMS internal standard, 300 MHz):
δ7.67-7.64 (4H, m), δ7.61-7.39 (9H, m) 7.19 (t, J=7.62 Hz,
1H), 7.02 (2H, d, J=4.69 Hz) 6.95 (1H, t, J=7.62 Hz), 6.66 (1H,
s), 6.66 (1H, d, J = 8.20 Hz). 13C NMR (100 MHz, CDCl3)
δ145.1, 142.5, 139.0, 138.2, 137.1, 136.3, 134.5, 134.0, 131.3,
131.2, 128.9, 128.8, 128.7, 128.6, 128.4, 128.3, 128.1, 127.6,
127.6, 125.7, 125.7, 125.6, 124.3, 120.0. Analysis calculated for
C33H19F3: C, 83.02; H, 4.27; F, 12.71. Found: C, 82.90; H, 4.28.
Preparation of Dications by Chemical Oxidation. SbF5 (∼0.7
mL, ∼9 mmol) was added to a graduated centrifuge tube in a
drybox and the tube was capped with a septum and placed in an
ice bath. SO2ClF33 (1.3 mL) at -78 °C was transferred by
cannula into the centrifuge tube. The contents were mixed on
Computational Methods
Geometries were optimized at B3LYP/6-31G(d) density func-
tional theory levels with the Gaussian 98 and 03 program
packages, (see Supporting Information). The chemical shifts
were calculated at B3LYP/6-311þG(d,p) using the GIAO
approach with the Gaussian 98 or 03 program packages on
the optimized geometries. The best agreement between experi-
mental and calculated shifts was found when the effect of solvent
was included in the calcuation through the polarization con-
tinuum method, PCM. The nucleus-independent chemical shifts
(NICS(1)zz)23,24 were obtained from the chemical shift tensor
˚
perpendicular to the ring for a dummy atom placed 1 A above
the center of each ring. Natural bond order, NBO, charges were
calculated using the natural population analysis27 in Gaussian
09 with basis set B3LYP/6-311þg(d,p) on geometries optimized
as described above. Areas were calculated from the Cartesian
coordinates for each ring system oriented in the xy plane. See
Supporting Information for specific details of the calculation.
Acknowledgment. We thank the Welch Foundation, grant
W-794, and the National Science Foundation, grant CHE-
0242227 and CHE-0948445 for support of this work.
Supporting Information Available: Spectra of 5a,c,e2þ, plots
of experimental vs calculated shifts for 2a-c2þ and for 2a-c2þ
/
5b,d2þ, table of the calculated NBO charges for each atom of
22þ, 52þ, 62þ and 72þ, plot of summation of charges by sub-
stituent, plots of ΣNICS(1)zz/area2 vs NBO charge/area, by
indenyl systems and by fluorenyl/tropylium systems, calculated
bond lengths for 22þ, 52þ, 62þ and 72þ, relationship between
ΣNICS(1)zz/area2 and the GEO term for 22þ and 52þ, H and
1
13C NMR spectra for 5a-e, calculated ring areas including
details of their calculation, complete reference for Gaussian 98,
Gaussian 03, and Gaussian 09, Cartesian coordinates, total
energies for 5a-e2þ, 62þ and 72þ. This material is available free
of charge via the Internet at http://pubs.acs.org.
(33) Reddy, V. P.; Bellew, D. R.; Prakash, G. K. S. J. Fluorine Chem.
1992, 56, 195–197.
J. Org. Chem. Vol. 76, No. 2, 2011 653