Structural Aspects of Aromaticity: Tadeusz Marek Krygowski and Michał Ksawery Cyran Ski

Chem. Rev.
2001, 101, 1385−1419 1385
Structural Aspects of Aromaticity

Tadeusz Marek Krygowski* and Michał Ksawery Cyrański
Department of Chemistry, University of Warsaw, Pasteura 1, 02 093 Warsaw, Poland
Received August 3, 2000
Contents V. Acronyms 1416

VI. Acknowledgments 1416
I. Introduction: Importance of a Notion in the 1385
Chemical Vocabulary and an Outline of the VII. References 1416
Criteria of Aromaticity
II. Geometry-Based Indices of Aromaticity 1388
A. Sources and Reliability of the Molecular 1388 I. Introduction: Importance of a Notion in the
Geometry Chemical Vocabulary and an Outline of the
B. Outline of the Aromaticity Indices Based on 1389 Criteria of Aromaticity
Bond Length Alternation
C. Harmonic Oscillator Model of Aromaticity 1390 Aromaticity (aromatic) is considered as one of the
(HOMA) most important concepts in modern organic chemis-
D. Geometry-Based Energetic Indices of 1391 try.1 There is no doubt that aromatic/aromaticity are
Aromaticity among the terms most frequently used in chemistry,
1. Energetics of π-Electron Hydrocarbons 1391 compared also to the terms in biochemistry and
Evaluated from the Exponential medicine as shown in Table 1.2 The term aromaticity/
Relationship between Bond Length and aromatic has a long history dating back to its first
Bond Energy use in a chemical sense by Kekulé3,4 and Erlenmeyer5
2. Harmonic Oscillator Stabilization Energy: 1392 in the 1860s.
Estimation of Canonical Structure Weights Similarly, as in those days, in a modern sense the
and Stabilization Energy from Bond term is related either to some typical properties6-11
Lengths or to a specific structure. In the last few decades it
E. Discussion of the Pros and Cons of 1393 has been almost generally accepted that both terms,
Molecular Geometry-Based Indices aromaticity and aromatic character, are associated
III. Applications 1397 with the ground-state properties of cyclic π-electron
A. Topological Problems 1398 compounds which (i) are more stable than their chain
B. Benzenoid Hydrocarbons 1398 analogues, (ii) have bond lengths between those
C. Aza-Analogues of Benzenoid Hydrocarbons 1402 typical of the single and double ones, and (iii) have a
D. Other Heterocyclic Analogues of Benzenoid 1404 π-electron ring current that is induced when the
Hydrocarbons. system is exposed to external magnetic fields, leading
E. Nonbenzenoid Alternant Hydrocarbons and 1405 to increased values of the magnetic susceptibility and
Their Heteroanalogues specific values of 1H NMR chemical shifts. (iv) From
F. Nonalternant π-Electron Hydrocarbons 1406 the point of view of chemical reactivity, aromatic
G. Cyclopentadienyl and Cyclopentadiene 1407 compounds prefer substitution to addition, and hence,
Moieties Embedded in Various Chemical they exhibit a tendency to retain their π-electron
Environments structure.12 However, this criterion is not based on
H. Heterocyclic Analogues of Nonalternant 1408 the property of the ground state of a system, and in
Hydrocarbons addition, it cannot be presented in a quantitative way
I. Dependence of Aromaticity on Intermolecular 1410 for general use. That is why it is often rejected as
Interactions the base for aromaticity indices, although some
J. Aromaticity vs Deviation from Planarity 1411 attempts have been made at using reactivity to define
K. Experimental Arguments for the 1412 the aromatic character quantitatively.13
Hiberty−Shaik Hypothesis of the Role of Aromaticity can be defined only by convention,10
σ-Electron Structure in π-Electron Systems and since there is no unique and generally accepted
L. Substituent Effects on Aromaticity 1412 definition of the term,11 all the above-mentioned
1. Bond Localization Due to the Anneleation 1413 criteria may be used to develop indices of aromaticity.
of Small Rings to Benzene Aromaticity is essentially an “excess property”, i.e.,
2. Angular Group-Induced Bond Alternation 1414 a deviation from an additive scheme.1
(AGIBA) Aromatic properties are mostly due to π-electron
3. Charged Substituents in Benzenoid 1415 systems, which are much more polarizable than the
Systems σ-core. However, the π-electron structure (the so-
IV. Conclusions 1416 called π-electron delocalization observed as the equal-
10.1021/cr990326u CCC: $36.00 © 2001 American Chemical Society
Published on Web 03/08/2001
1386 Chemical Reviews, 2001, Vol. 101, No. 5 Krygowski and Cyrański
Table 1. Frequency of Chemical/Biochemical/Medical

Terms in Titles, Abstracts, or in Keywords of
Publications Retrieved from ISI (1981-1999)
term(s) ref
acid/acidity 639 166
AIDS 55 067
anomeric 2 273
aromatic/aromaticity 58 872
aqueous 114 340
base/basicity 125 694
benzene 28 327
cancer 270 647
carbohydrate/carbohydrates 44 750
catalyst/catalysis/catalytic 127 655
chiral/chirality 44 607
crystal/crystals 303 850
death 101 836
Tadeusz Marek Krygowski was born in Poznań, Poland (1937) and re- DNA 331 997
ceived his M.Sc. degree at the Adam Mickiewicz University (Poznań, 1961) electrophilic 6982
and his Ph.D. (1969) and D.Sc. degrees (habilitation; 1973) both from fullerene/fullerenes 8 526
the Department of Chemistry of Warsaw University. Since 1964 he has H-bond 504 236
been working at this university, from 1983 as a Professor of Chemistry. intramolecular 32 246
He has lectured at many universities in many countries, serving as an intermolecular 18 868
invited Professor in Canada (Guelph), France (Nantes), Austria (Linz/D), life 146 857
and Israel (BeerSheva). He has been president of the Polish Chemical molecular 453 162
Society (1994−1997) and is a national representative of the IUPAC Com- molecule/molecules 262 550
ission on Physical Organic Chemistry. His main research interests are nucleophilic 13 907
sedimentology (in geology, in the early 1960s), ion-pairing and generally polymer/polymers 161 497
medium effects in organic electrochemistry (in the 1960s and early 1970s), protein/proteins 804 002
solvent and substituent effects (1970s and 1980s), problems of aromaticity, solvent 71 519
and structural consequences of inter- and intramolecular interactions (until substituent 15 492
now). His hobby is national folk music and hiking in not-too-high mountains. substitution 61 397
water 404 451
virus 207 806
structuresmay be governed by the σ-electron de-

mand for the equalization of bond lengths.
The term delocalization may be understood in two
different ways: (a) as a result of the resonance
between various canonical structures,16 leading to the
equalization of bond lengths, and (b) as a result of
strong π-electron interaction (the so-called through-
resonance effect) between the substituents, leading
to a structure of strong quinoid character.17
In this review “delocalization” is used in the first
sense. The first three above-mentioned properties (i-
iii) are easily measurable and concern molecules in
Michał Ksawery Cyrański was born in Warsaw, Poland (1970). In 1994 their ground electronic state. Hence, they may be
received his M.Sc. degree and in 1999 his Ph.D. degree, both at the readily transformed into the numerical parameters
Chemistry Department of Warsaw University. In 1997 he was a holder of describing the aromatic character (aromaticity indi-
the Foundation for Polish Science grant; in 1998 he received the distinction ces) of molecules or their fragments. The last property
of the Kemula Reward (by the Polish Chemical Society) and in 1999 the (iv) depends on the energy of the transition state, and
Kołos Reward for the best Ph.D. work (by the Department of Chemistry hence, its value is difficult to estimate. Consequently,
of Warsaw University). In 1998−2000 he was a secretary of the Executive
Committee of the Polish Chemical Society (elected). His main scientific this criterion is usually applied mainly in a qualita-
interests are the structural chemistry of small organic compounds, tive way in the practice of synthetic organic chem-
crystallochemistry, molecular modeling (ab initio), structural aspects of istry.
aromatic character of π-electron systems, and definition of aromaticity. Criterion i leads to energy-based aromaticity indi-
So far, he has published over 30 original papers and 5 reviews and ces. The beginning of the energetic criteria is associ-
presented 5 lectures and over 30 poster and oral communications. His ated with the early discovered low reactivity of
hobby is classical music, especially by J. S. Bach, and singing in a
professional choir.
benzene (for historical details cf. ref 18); then a
numerical estimation of the thermodynamic stability,
ization of bond lengths) may be considered as a defined as the resonance energy (RE), came from the
byproduct of the σ-imposed geometric symmetry and work by Pauling and Sherman19 and Kistiakowsky
not necessarily as a driving force by itself (as it is in et al.20 Originally it was defined for the benzene
the case of benzene).14 There are also some argu- molecule as a difference in energy between that of
ments against the above hypothesis,15 and hence, this benzene and that of the hypothetical Kekulé struc-
topic should still be considered as disputable. In ture with bond lengths as in an olefine (a model
conclusion, it might be interpreted that the aromatic molecule called cyclohetratriene but with the double
propertiesseven if they are due to the π-electron bond like in ethylene). Those first attempts were
Structural Aspects of Aromaticity Chemical Reviews, 2001, Vol. 101, No. 5 1387
Table 2. Stabilization Energy Values for Benzene the (additive) reference system (reaction) and the
scheme stabilization level of the theory applied.1
of energy Another insight into ASE came from the so-called
level of theory reactiona [kcal/mol] graph-topological approaches. These methods are
MP2/RHF/SBK(d) 1 74.7b undoubtedly useful for large systems, but they do not
RHF/SBK(d) 1 61.4b take into account many subtle effects, and may be
MP2/6-31G//6-31G* 1 67.2c
used for estimation rather than detailed studies.24 An
HF/6-31G* 1 58.2c
MP4SDTQ/6-31G**/MP2(full)/6-31G** 2 23.9d excellent review on energetic criteria of aromaticity
MP4/6-31G+5D 2 24.3e is given in the monograph by Minkin.11
MP3/6-31G+5D 2 23.4e Criterion (ii) leads to the geometry-based aroma-
RMP2/6-311G** 2 28.0f ticity indices which are described in detail in the next
RMP2/6-311G* 2 28.7f
RMP2/6-31G* 2 28.9f section.
B3LYP/6-311+G** 2 22.2a1 Criterion (iii) is associated with a few indices of
B3LYP/6-311+G** 2i 34.1g aromaticity related to the magnetic properties of
6-31G** (SCF) 2 24.8f molecules. The earliest one was the exaltation of the
6-31G* (SCF) 2 24.7f
MP4SDTQ/6-31G**/MP2(full)/6-31G** 3 20.3d
magnetic susceptibility Λsthe quantity which can be
6-31G* (SCF) 3 23.4h estimated both experimentally25 or theoretically.26 As
in the case of energetic criteria, its reliability depends
a
Schemes of reactions: (1) C6H6 + 6CH4 f 3CH3CH3 + on the assumed reference structure, since by defini-
3CH2CH2; (2) C6H6 + 3 CH2CH2 f 3trans-CH2CHCHCH2; (3)
C6H6 + 3CH2CHCHCH2 f 3CH2CHCHCHCHCH2. b Matsu- tion
naga, N.; Cundari, T. R.; Schmidt, M. W.; Gordon, M. S. Theor.
Chim. Acta 1992, 83, 57. c Bock C. W.; George, P.; Trachtman, Λ ) χm - χa (1)
M. J. Phys. Chem. 1984, 88, 1467. d Schleyer, P. v. R.;
Glukhovtsev, M. N. Chem. Phys. Lett. 1992, 198, 547. e Had-
don, R. C.; Raghavachari, K. J. Am. Chem. Soc. 1985, 107, where χm and χa are the measured susceptibility
289. f Disch, R. L.; Schulman, J. M. Chem. Phys. Lett. 1988, value and that computed from the additive scheme,
152, 402. g Schleyer, P. v. R.; Jiao, H.; Hommes, N. J. R. v. E.;
Malkin, V. G.; Malkina, O. L. J. Am. Chem. Soc. 1997, 119,
respectively.27 The final results depend on the kind
12669. h Hess, B. A.; Schaad, L. J. J. Am. Chem. Soc. 1983, of Pascal’s constants employed28 and, by definition,
105, 7500. i In scheme 2 for cis CH2CHCHCH2. by the size of the ring in question.29 In the case of
theoretical treatments, reliability depends addition-
based on experimentally determined energies in ally on the level of theory and the extent of applied
which the enthalpy of formation19 or hydrogenation20 electron correlation.11 At times the anisotropy of the
of benzene was compared with the values of forma- magnetic susceptibility has been used as a descriptor
tion or hydrogenation of the hypothetical Kekulé of aromaticity,30 but this criterion is seldom applied
structure of benzene. In both cases, the value of RE now. Historically, a very important criterion and one
was almost exactly the same, 36 kcal/mol. Recent frequently used in practical work was the application
developments in computer techniques and theories of an easily measurable proton chemical shift in 1H
of quantum chemistry offered an exceptionally good NMR spectrastypically aromatic protons have sig-
opportunity for employing advanced quantum chem- nals around 7 ppm or more whereas olefinic systems
istry methods.21 The ideas of isodesmic22 and homo- are around 5 ppm.31 Very recently, a new and
desmotic23 reactions gave very effective methods for effective aromaticity index, NICS (abbreviation from
estimating various stabilization energies (SE), in- nucleus-independent chemical shift), has been intro-
cluding those of aromatic systems, called the aro- duced by Schleyer et al.32 It is defined as a negative
matic stabilization energy (ASE). This quantity de- value of the absolute shielding computed at a ring
pends dramatically on the kind of reaction, i.e., on center or some other interesting point of the system.
the “reference” structure. Table 2 presents the values Rings with negative NICS values qualify as aromatic,
of ASE calculated by use of various homodesmotic and the more negative NICS, the more aromatic are
and isodesmic reactions at various levels of theory.1 the rings. Consequently, antiaromatic systems have
It is immediately clear that any kind of stabiliza- NICS > 0. Table 3 presents the values of NICS and
tion energy (SE) depends strongly on the scheme of Λ for several typical π-electron systems.32
Table 3. Comparison of NICS and Exaltation of the Magnetic Susceptibility Values for Several Typical
Compounds
NICS [ppm] Λ [ppm]a NICS [ppm] Λ [ppm]a
compound HF/6-31+G* HF/6-31+G* compound HF/6-31+G* HF/6-31+G*
cyclobutadiene 27.6 18.0 benzene -9.7 -13.4
pyrrole -15.1 -12.1 naphthalene -9.9 -28.2
cyclopentadienyl anion -14.3 -17.2 phenanthrene (central ring) -6.5 -47.9
thiophene -13.6 -10.0 phenanthrene (outer ring) -10.2 -47.9
furan -12.3 -9.1 anthracene (central ring) -13.3 -49.8
phosphole -5.3 -3.3 anthracene (outer ring) -8.2 -49.8
cyclopentadiene -3.2 -2.4 tropylium ion -7.6 -20.5
alumole 6.5 11.2 azulene, 5-ring -19.7 -42.9
borole 17.5 12.8 azulene, 7-ring -7.0 -42.9
pentalene 18.1 30.9 heptalene 22.7 76.6
a Total value.
Table 4. Total Energies for Tetracene and Its Four theory for molecules of five benzenoid hydrocarbons
Benzenoid Structural Isomers Computed at the being the structural isomers of C18H12sthe differ-
HF/6-31G** Level. Only the σ-electron Structures Are
Shown ences are meaningful and illustrate the role of
topology well. The most stable chrysene is by about
14 kcal/mol more stable than the least stable tet-
racene! This is also well supported by the successes
in estimating various stabilization energies by use
of the theories based on graph theory and topological
approaches.24
The interplay between the topology of molecules
(or their fragments) and geometry can substantially
change their chemical, physicochemical, and bio-
chemical properties, including aromaticity. This kind
of problem is one of the main topics which will be
presented and discussed in this review. Geometry-
based indices of aromaticity are almost the only ones
which are readily accessible since the X-ray diffrac-
tion technique is an extremely powerful source of
information on molecular geometry. Moreover, reli-
able theoretically obtained molecular geometries may
be also used. Unlike energetic and magnetic criteria,
the geometric indices can be applied to both the
global and local aromatic character of a system.42
These kind of indices are well suited for studies of
the dependence of local or global aromaticity on the
environment of the molecule or its fragment.
An additional encouragement of using the geomet-
ric criterion stems from the well-known statement
by R. Hoffmann:43 “There is no more basic enterprise
One more criterion may be very useful: the Hückel in chemistry than the determination of the geo-
(4N + 2) rule,33 which states that cyclic (planar) metrical structure of a molecule. Such a determina-
π-electron systems with (4N + 2) π-electrons are tion, when it is well done, ends all speculation as to
more stable than those containing 4N π-electrons. the structure and provides us with the starting point
This is a rather qualitative criterion but has inspired for the understanding of every physical, chemical and
numerous attempts at synthesis of the systems which biological property of the molecule”.
could test its general validity.34 Those attempts were
particularly interesting for the π-electron systems
with a small number of π-electrons (N ) 0 or 1). II. Geometry-Based Indices of Aromaticity
Many of them were carried out in the 1960s: The Obviously the geometry-based aromaticity indices,
triphenyl derivative of cyclopropenyl cation was if they are to be applied in a reliable way, need a
synthesized and found to be stable,35 similar to reliable molecular geometry. However, even though
diphenylcyclopropenone36 in which the electronega- different experimental techniques, such as gas-phase
tive carbonyl group stabilizes the structure. Cyclo- electron diffraction (ED), microwave spectroscopy
butadiene is definitely unstable, but its complexes (MW), and X-ray and neutron diffraction, result in
with NiCl237 or AgNO338 are known. the molecular geometry defined in a slightly different
The Hückel rule was extended onto polycyclic way,44,45 these differences are not significant for the
nonalternant systems:39 Defining F and S as numbers purpose of describing aromaticity by use of the
of five- and seven-membered rings in a fused system, geometry-based indices. The quality of the experi-
the molecule is more stable when S ) F and less mental measurements or of the theoretical calcula-
stable for other cases. This rule was then demon- tions may be of key importance only in cases of very
strated and proved by use of the topological theories subtle analysis of small differences in aromaticity.
of aromaticity.40 Therefore, some notes in this field are given below,
Undoubtedly, aromatic character is related to the since some typical experimental errors or trends in
structure of the molecule (or its particular fragment) computational methods should be taken into account,
in question. However, the term “structure” may be particularly for studies of subtle changes in aromatic
understood in two ways. First, it may be understood character.
as a molecular geometry, when various geometrical
parameters are used to characterize the molecules. A. Sources and Reliability of the Molecular
Second, it may be understood in a topological sense, Geometry
i.e., taking into account only the bond connectivity
of atoms in the molecule (or its fragment). There is Dynamic development of computer techniques and
no doubt that the topology of the linking atoms exerts experimental techniques of the X-ray diffraction
a substantial influence on any molecular property. measurements have given the most powerful and
Table 4 presents as an example the total energies41 most frequently used method for determination of the
computed by ab initio at the HF/6-31G** level of molecular structure. Now more than 200 000 molec-
ular geometries are archived in the CSD (2000),46 and the first aromaticity index, AJ, based entirely on the
easy access to molecular geometry is also facilitated molecular geometry (bond lengths). This index is a
by use of routine X-ray diffraction measurements. normalized function of the variance of bond lengths;
Some consequences for estimated values of aroma- thus, it takes the mean value of the bond lengths as
ticity indices in dependence on the experimental a reference for estimating a decrease in aromaticity
conditions will be discussed later. Another important due to the bond length alternation (original formula-
and useful source of molecular geometry are the ab tion)55
initio techniques at a high level of theory.21 However,
in this case geometry depends dramatically on the
level of theory and quality of basis set. This short-
coming may be partly diminished if the data opti-
AJ ) 1 -
225
n ∑ 1- (
Rr
R ) 2
(2)
mized at the same level of theory are applied. The where n is the number of peripheral bonds with
above-mentioned techniques provide most of the individual bond lengths Rr and R is their mean value.
molecular geometries discussed in recent reports.8,45,47 The constant 225 results from the normalization
Much less useful are other techniques such as gas- conditions to obtain a zero value for the Kekulé
phase electron diffraction (ED) measurements or structure of benzene and 1 for any system with all
microwave (MW) determinations since they are ap- bonds of equal length.
plicable only to rather simple molecules (MW), often An obvious limitation of the model is the fact that
highly symmetric (ED).45 Additionally, due to the it can be applied only to carbocyclic systems. Some-
thermal condition of the experiment, the molecules times this approach has been used on species with
studied by electron diffraction are vibrationally ex- heteroatoms, and then only CC bonds were taken into
cited whereas the microwave measurements are very account.56,57 This limitation results in exclusion of the
time-consuming. contributions which are directly due to the interac-
tions between heteroatoms and carbon-heteroatoms
B. Outline of the Aromaticity Indices Based on leading to a potential misinterpretation of aromatic-
Bond Length Alternation ity of these systems. Julg’s index was applied to
Both archetypal aromatic systemssbenzene, the numerically describe the aromaticity of some π-elec-
paradigm of aromaticity in organic chemistry,48 and tron systems8,55,57 and served also to test the correct-
graphite49sare very stable among π-electron systems ness of other aromaticity indices such as HOMA58 as
and exhibit the same structural featuresjust one well as supporting other already accepted measures
bond length, the value of which (1.398350 and 1.422 of aromaticity such as aromatic stabilization energy,
Å,51 respectively) is between the length of the typical ASE, and magnetic susceptibility exaltation, Λ.56
single and double CC bonds for sp2-hybridized carbon To overcome the problem of heteroatoms, Bird59
atoms (e.g., in 1,3-butadiene they are 1.467 and 1.349 applied the same idea as Julg and François, replacing
Å,52 respectively). Consequently, bond length alterna- the bond lengths by Gordy’s60 bond orders
( )
tion has been taken as a measure of the decrease of
aromaticity. Annulenes following the Hückel (4N + V
I ) 100‚ 1 - (3)
2) rule have either no (benzene) or very little alterna- Vk
tion, whereas those which are 4N annulenes exhibit
strong alternation.11 Pauling and Sherman19 and where
later Kistiakowsky et al.20 considered the Kekulé
x ∑(N - Nh )2
structure of benzene as a reference for estimating the
resonance energy of benzene (see Figure 1). This 100
V)
h
N n
and
a
N) -b (4)
R2
R is the observed bond length, N is the individual

Figure 1. Model structures for the aromatic and non-
bond order, N h is the mean bond order, and n stands
aromatic forms of benzene for the number of bonds. Bird’s indices have been
defined for five- and six-membered rings, I5 and I6,
point has been a subject of dispute, since the systems for which the constants a and b of eq 3 are collected
such as benzene, which can be assumed as deformed in Table 5.59
to a localized structure by a simple “breathing” In the case of a fully delocalized heterocycle, V has
distortion, do so with little energy loss.53,54 In any a value of 1, whereas for a nondelocalized Kekulé
calculation of the aromatic stabilization energy (ASE), structure with alternating single and double bonds,
the reference structure consists of the idea of single the value depends on the type of ring system. Thus,
and double bonds (of different mutual delocalization). for a five-membered ring Vk ) 35 and for a six-
Following the idea that an increase in bond alter- membered ring Vk ) 33.3, and for systems consisting
nation is a good feature characterizing a decrease in of a five- and six-membered rings fused together, Vk
aromaticity, in 1967 Julg and François55 formulated ) 35.
Table 5. Values of Constants a and b Used in give HOMA ) 0 for the hypothetical Kekulé struc-
Calculation of Gordy’s Bond Orders tures of the aromatic systems (with the lengths for
type of bond a b CC bonds as in acyclic polyene, 1,3-butadiene52) and
C-C 6.80 1.71 1 for the system with all bonds equal to the optimal
C-N 6.48 2.0 value Ropt. The individual bond lengths are depicted
C-O 5.75 1.85 by Ri. The quantity Ropt is defined as a length of the
C-P 13.54 3.02 CC bond for which the energy (estimated by use of
C-S 11.9 2.59 the harmonic potential) of the compression to the
N-N 5.28 1.41
N-O 4.98 1.45 length of a double bond and expansion to the length
C-Se 15.24 3.09 of a single bond in 1,3-butadiene is minimal. The
C-Te 21.41 3.81 same procedure can be applied to bonds containing
C-P 13.54 3.02 heteroatoms, choosing the model single and double
N-S 10.53 2.50 bonds as a reference for the above-presented proce-
N-Se 13.31 2.86
O-S 17.05 5.58 dure.66,67
S-S 19.30 3.46 Within the framework of the HOMA model, some
more detailed information may be obtained. The
model makes it possible to separate the two terms
Bird’s indices I5 and I6 describe the percentage of describing different contributions to a decrease in
aromaticity and have been applied to a large number aromaticity:67,68 (i) due to the bond elongation (the
of hetero-π-electron systems.61 It is worth mentioning term called EN) and (ii) due to the bond length
that a similar idea based on Gordy’s bond order was alternation (the term called GEO).
published at the same time by Pozharskii,62 but then
it was not as widely applied as that of Bird. Another R
n∑ opt
aromaticity index based on variation of bond lengths HOMA ) 1 - (R - Ri)2 )
and bond orders was suggested by Nyulaszi et al.63 R
and applied to polyphosphaphospholes.64 1 - R(Ropt - Rav)2 + ∑(Rav - Ri)2 )
[ ]
n
Another approach in which the bond length alter-
1 - EN - GEO (7)
nation is taken into account is based on the sum of
squared differences of the sequential bond lengths. where Rav is an averaged bond length, Rav ) 1/n
The index called BAC (bond alternation coefficient)
∑i)1
n
Ri and R(Ropt - Rav)2 and R/n∑(Rav - Ri)2 repre-
reads65
sent EN and GEO terms, respectively.
In the cases where Rav is shorter than Ropt, the term
BAC )
x∑ n
(Rn - Rn+1)2 (5)
EN must be taken with a negative sign, since the
shorter the bond, the greater its energy.48,69 Hence,
the proper formula for HOMA is
where the summation runs over a defined cyclic route HOMA ) 1 - EN - GEO (8)
which is obvious in the case of monocyclic systems
but may be a more complex problem for polycyclic where
systems if the overall aromaticity of a molecule is in
question. This index cannot be used for systems EN ) f‚R(Ropt - Rav)2 (9)
containing heteroatoms. It also fails, at least partly,
{
in cases where the localization of double bonds is not
1 :Rav > Ropt
exactly alternant but such as that observed (for f)
example) in the quinoidal systems, where two of the -1 :Rav < Ropt
six differences of bond length are zero (for C2v
R
n∑ av
symmetry of the ring) or almost zero (in the lower GEO ) (R - Ri)2 (10)
symmetry cases).
Note that shorter Rav values than Ropt are sometimes
C. Harmonic Oscillator Model of Aromaticity encountered in molecular geometries obtained by
(HOMA) X-ray diffraction technique, where the measurements
were carried out at room temperature without cor-
Another approach based on bond lengths is the rections for thermal motion.
index HOMA.58,66 This index differs from all other The recently developed eq 8 cannot be used directly
geometry-based ones by assuming another reference for the heterocyclic systems since the averaging
bond length (bond order). Instead of the mean bond procedure is meaningless for heterogenic data (the
length (or bond order) taken as a reference, in the same limitation as in the case of Julg’s index55).
HOMA model a concept of the optimal bond length Therefore, in the case of hetero-π-electron systems,
is proposed a further modification of the HOMA index was
made.67 The bonds with heteroatoms are represented
R
n∑ opt
HOMA ) 1 - (R - Ri)2 (6) by the Pauling bond numbers n70
R(n) - R(1) ) -c ln(n) (11)

where n is the number of bonds taken into the
summation and R is an empirical constant chosen to then n is calculated by use of the model bond lengths
Table 6. Structural Parameters of the HOMA Index in the bond length: the shorter the bond, the more
Its Separated Form (8) for π-Electron Systems with energetic it is. Very recently it has been proposed that
Heteroatoms. nopt - a Value of the Bond Number for
Ropt bond lengths may be used in a model for calculating
the heats of formation of π-electron hydrocarbons.71
bond R(1) R(2) Ropt Ra c nopt The idea employed the Pauling concept of the bond
CC 1.467 1.349 1.388 257.7 0.1702 1.590 number, n70 (see section C)
CN 1.465 1.269 1.334 93.52 0.2828 1.589
CO 1.367 1.217 1.265 157.38 0.2164 1.602
CP 1.814 1.640 1.698 118.91 0.2510 1.587 R(n) - R(1) ) -c ln(n) (11)
CS 1.807 1.611 1.677 94.09 0.2828 1.584
NN 1.420 1.254 1.309 130.33 0.2395 1.590 where R(1) represents the length of a single bond,
NO 1.415 1.164 1.248 57.21 0.3621 1.586 R(n) the length of the bond with bond number n, and
a Applied in the case when no separated version of eq 6 is c is an empirical constant. Another empirical rela-
in use.66 tion72 combines bond energy with the bond number
for single and double bonds, R(1) and R(2), respec- E(n) ) E(1)‚np (14)
tively
where E(1) is the energy of a single CC bond, E(n) is
R(1) - R(n) the energy of a bond of bond number equal to n, and
n ) exp (12) p is an empirical constant.
c
Thus, a simple combination of eqs 11 and 14 leads
and consequently the bond numbers are converted to the final expression for calculation of the CC bond
into the “virtual” CC bonds via the formula energy of π-electron hydrocarbons
r(n) ) 1.467 - 0.1702 ln(n) (13) E(n) ) E(1) exp{R‚[R(1) - R(n)]} (15)
Finally, the r(n) values from eq 13 may be applied in
eq 8, leading to the HOMA index and its components where R is an empirical constant related to p and c.
EN and GEO. Table 6 collects all the constants used Applying the reference bond lengths and bond
for the carbo- and heterocyclic systems. energies for CC single and double bonds, one obtains
Even if the choice of the reference bond lengths an expression for the bond energy, E(R), calculated
R(1) and R(2) in Table 6 is somewhat arbitrary and directly from the bond length71
biased by some error (depending on the precision of
measurements and the choice of model molecule(s) E(n) ) 87.99 exp{2.255[1.533 - R(n)]} (16)
for a given bond length), a very narrow range of
variation of the nopt values, between 1.584 and 1.602 The reliability of eq 16 was tested by comparisons of
Å, makes the procedure encouraging. heats of formation from atoms, ∆Hf, calculated di-
rectly from experimental geometries of eight ben-
D. Geometry-Based Energetic Indices of zenoid hydrocarbons (assuming a constant energy of
Aromaticity CH bonds) with the respective data obtained from
calorimetric measurements. The mean error of esti-
One of the most fundamental and earliest criteria mation was less than 1%. Moreover, the above-
of the aromatic character is the high chemical stabil- presented model is in almost perfect agreement if the
ity of the systems regarded as aromatic. This feature bond lengths and energies based on charge density
was later quantified into measures of aromaticity studies of six benzenoid hydrocarbons with 34 types
named resonance energies19 or various types of of CC bonds are compared.73,74
stabilization energies.11 The above results provide a good basis for using
Undoubtedly, the energetic descriptors of molecules the model to estimate the energy of molecules and
are very valuable. In many cases they contribute in their fragments.71 The treatment might be particu-
a decisive way to the determination and interpreta- larly useful for large and complex molecular systems,
tion of the chemical and physicochemical properties for which it is very difficult to obtain reliable ener-
of molecules. However, quantum chemical methods gies. Moreover, it may be used for molecular frag-
are still less useful if one wishes to know the ments for which quantum chemical calculations are
energetic descriptors solely of some part of a mol- of limited value due to time-consuming computa-
ecule. In this matter, approximate models are much tions.75
more helpful than the methods originated from
Equation 16 provides an opportunity to calculate
quantum chemistry, although the latter are more
values of the energy content of a molecule or its
sophisticated and better grounded theoretically.
fragment, referred solely to CC bonds. In this way,
the ring energy content, REC, or the molecule energy
1. Energetics of π-Electron Hydrocarbons Evaluated from content, MEC, are defined.
the Exponential Relationship between Bond Length and
Bond Energy71 The REC values for the benzene ring in various
topological situations in benzenoid hydrocarbons vary
It is well-known from elementary textbooks on dramatically by about 80 kcal/mol (from 650 to 730
organic chemistry48,69 that bond energy depends on kcal/mol).71 The REC values correlated statistically
Table 7. Empirical Constants for HOSE Model

type of
bond Rso [Å] Rdo [Å] a (×104)Pa b (×104)Pa remarks
CC 1.467 1.349 44.39 -26.02 C-C and CdC from ED data for butadiene 1,3.52
CN 1.474 1.274 43.18 -25.73 C-N from MW data for methylamine;a CdN from ED data for
CH2)CH-CHdNHb
CO 1.428 1.209 52.35 -32.88 C-O taken from ED data for methanol;c CdO from ED data for acroleined
NN 1.420 1.254 78.82 -52.41 N-N and NdN taken from ED of MeCdN-NdNMee and
MeNdNMee, respectively.
NO 1.415 1.164 33.197 -19.92 N-O and NdO taken from MW measurements of MeONO.e
a Pulay, P.; Torok, F. J. Mol. Struct. 1975, 29, 236. b Penn, R. E. J. Mol. Spectrosc. 1978, 69, 373. c Kimma, K.; Kubo, M. J.
Chem. Phys. 1959, 30, 151. d Kuchitsu, K.; Fukuyama, T.; Morino, Y. J. Mol. Struct. 1969, 4, 41. e Vilkov, L. V., Mastryukov, V.
S., Sadova, N. I. Determination of the Geometrical Structure of Free Molecules; Mir Publishers: Moscow, 1983.
well (correlation coefficient R ) 0.911) with the Table 8. Empirical Constants to Bird’s Modification81
of the HOSE Model
Randić index,76 which describes the energetics of the
rings in benzenoid hydrocarbons computed by use of type of
topology-based theories. A slightly worse correlation bond Rso [Å] Rdo [Å] F a (×104)Pa b (×104)Pa
is observed for the scatter plot of REC vs HOMA CC 1.533 1.337 85.94 44.39 26.02
(correlation coefficient R ) 0.831). CN 1.474 1.274 62.19 43.18 25.73
CO 1.449 1.250 62.33 57.10 35.51
2. Harmonic Oscillator Stabilization Energy: Estimation of Table 9. Resonance Energies (in kcal/mol) from Bird’s
Canonical Structure Weights and Stabilization Energy Modification of the HOSE Model Compared with
from Bond Lengths Experimental Data
compound RE REexp compound RE REexp
Another attempt at using bond lengths to estimate
the stabilization energy is associated with the em- pyridine 44.1 43.3 imidazole 34.9 40.0
pirical model enabling estimation of the canonical pyrimidine 40.1 40.6 indole 73.3 73.8
pyrazine 38.8 40.9 isoindole 62.5 62.4
structure weights called HOSE (harmonic oscillator 1,3,5-triazine 44.9 44.9 benzimidazole 75.1 78.9
stabilization energy) for π-electron molecules or their quinoline 81.6 81.0 carbazole 113.5 111.7
fragments.77 HOSE is defined as a negative value of isoquinoline 82.6 81.0 acridine 101.1 108.3
the energy necessary to deform the geometry of the quinozaline 74.8 76.5 phenazine 100.3 110.3
real molecule into the geometry of one of its Kekulé pyrrole 31.8 34.8 phenanthridine 108.5 119.6
(or resonance) structures with localized single and
double bonds of typical lengths as illustrated for canonical structure is
benzene by Figure 2.
n1
HOSEi ) 301.15[ ∑ (R′r - Rso)2‚k′r +
r)1
n2
Figure 2. Real benzene and its Kekulé structure ∑ (R′′r - Rdo )2‚k′′r] (19)
r)1
The energy of deformation has been derived from
a simple harmonic oscillator potential giving the HOSE values in kJ/mol. Table 7 collects
parameters used in the model.77,78
n1 n2 The estimated stabilization energies from the
1
Edef ) - ‚ [ ∑ (R′r - Rso)2‚k′r + ∑ (R′′r - Rdo )2‚k′′r] HOSE model are in line with those obtained in other
2 r)1 r)1 ways. It was shown77 that HOSE correlates well with
(17) the Hess and Schaad resonance energies (correlation
coefficient R ) 0.991 for 22 alternant π-electron
where Rr′ and Rr′′ stand for the lengths of π bonds in hydrocarbons; R ) 0.937 for 12 nonalternant hydro-
the real molecule, whereas n1 and n2 are the numbers carbons).79,80 In the last case, the Hess and Schaad
of the corresponding formal single and double bonds RE values were given for nonalternant hydrocarbons
in the ith canonical structure, respectively. In the whereas geometries used to calculate the HOSE
process of deformation, the n1 bonds corresponding values originated from substituted species. Since the
to the single bonds in the ith canonical structure are substituent effect in nonalternant systems is quite
lengthened whereas the n2 bonds corresponding to considerable, these aspects will be discussed later.
the double bonds in the ith canonical structure are The idea of using HOSE to estimate stabilization
shortened to the reference bond lengths Rso and Rdo , energies of heteroaromatics was modified by Bird.81
respectively. The force constants (kr) follow the as- To grasp molecules with a broader range of bond
sumption length variation, he used different reference bond
lengths. By use of a scaling factor F, he also scaled
kr ) a + bRr (18) the resulting stabilization energies to reproduce RE
for benzene of 38.4 and 44.9 kcal/mol for 1,3,5-s-
Thus, the final formula for HOSEi for a given ith triazine. The reference data are given in Table 8,
Table 10. Geometry-Based Indices: HOMA, EN, GEO, I6, I5, and AJ in Comparison with NICS
system HOMA EN GEO I6 AJ NICS
benzene 0.979 0.021 0.000 100 1.00 -9.7
pentagons in C60 0.103 0.897 0.00 100 1.00
hexagons in C60 0.550 0.248 0.202 76.2 0.912
central ring of perylene 0.133 0.765 0.101 83.0 0.957 +6.8
outer rings of perylenea 0.806 0.039 0.154 79.2 0.931 -9.4
central ring of triphenylene 0.077 0.683 0.239 73.6 0.898 -3.0
outer rings of triphenylenea 0.936 0.017 0.046 89.0 0.979 -10.8
radialene -2.408 3.408, 0.000 100 1.00 +2.8
a Mean values.
Table 11. Comparison of IA with HOMA, EN, and GEO

whereas a selection of the resulting RE compared for Quinoline, Isoquinoline, Indole, and Naphthalene
with the experimental data are given for some aza-
heterocycles in Table 9. In the case of five-membered compound IA HOMA EN GEO
rings, an energetic increment of 9.36 kcal/mol ac- quinoline 132 0.710 0.004 0.286
counting for the strain was included. This value was isoquinoline 134 0.667 -0.011 0.344
indole 146 0.909 0.006 0.086
estimated from a harmonic potential employing naphthalene 142 0.824 0.038 0.138
deformations of bond angles.
The results presented above are in quite satisfac- Table 12. HOMA, EN, and, GEO Values for
tory agreement with experimental data and encour- Naphthalene for Uncorrected (u) and Corrected (c)
age the use of HOSE in order to estimate stabiliza- Libration Geometries Measured by X-ray Diffraction
tion energies for systems with no calorimetric data. at 92, 143, and 239 K
The HOSE model was originally successfully condition of
used47,82 to estimate the weights of canonical struc- geometry ∆ HOMA
tures, Ci for the i-structure, by use of a formula estimations HOMA EN GEO (u-c)
92K (u) 0.8070 0.0705 0.1225 0.0155
-1 92K (c) 0.7915 0.0852 0.1233
[HOSEi] 143K (u) 0.8132 0.0623 0.1245 0.0239
Ci ) (20) 143K (c) 0.7893 0.0852 0.1255
N 239K (u) 0.8220 0.0406 0.1374 0.0460
∑[HOSEj]-1 239K (c) 0.7760 0.0852 0.1388
j)1
bonds of equal (or almost equal) length. Table 10
The weights of canonical structures estimated in this provides some typical examples.
way correlated well with those obtained by other The central rings in perylene and triphenylene are
methods. A good correlation between the HOSE definitely less aromatic than the outer rings: HOMA
weights estimated for 63 ground-state canonical and NICS are in line with this, but I6 indicates almost
structures of seven benzenoid hydrocarbons and the the same aromaticity for both kinds of rings. The ring
data calculated by use of the Randić model83 has been in radialene and the five-membered ring in fullerene
found (the correlation coefficient R ) 0.985). Simi- C60 are other excellent examples. In both cases, Julg’s
larly, a good but qualitative agreement was found index equals 1.00 and Bird’s I5 or I6 equals 100,
between HOSE-calculated canonical structure weights whereas NICS32 and HOMA show their weakly
for nitroaniline84 and those computed by use of aromatic (fullerene) and strongly antiaromatic (ra-
valence bond theory.85 dialene) character. In both cases, the fact that the
The application of the HOSE model permitted bond lengths are equal results in very high values of
estimation of aromaticity in the cases of very small AJ and I5 (I6), but this definitely does not mean that
changes in aromatic character of the molecule in the above systems are strongly aromatic. The reasons
question. This can be done by assuming equal or are obvious: the bond lengths in the radialene ring
almost equal weights86 of basic canonical structures and the pentagon in fullerene-C60 are 1.50387 and
as encountered in benzene or pyridine. 1.447 Å,88 respectively, and hence, the terms EN are
large and subtracted from 1 give low or even negative
E. Discussion of the Pros and Cons of Molecular values of HOMA.
Geometry-Based Indices An additional problem arises when the Bird index
is considered in its unified form IA.89 It is biased by
Most of the geometry-based aromaticity indices the assumption made in the unifying procedure. The
may be applied both to describe the local aromaticity Hückel delocalization energies (DE) of benzene (2.0β),
of the particular fragment(s) of molecules or to the the cyclopentadienyl anion (2.47β), naphthalene
global description of aromaticity of the whole mol- (3.68β), and indenyl anion (4.17β), i.e., for the nomi-
ecules. This is not the case for most of the energetic nal parents of the I6, I5, I6,6, and I5,6, were used in
or magnetic indices of aromaticity. the construction of an unified aromaticity index
All models, based only on the variance of bond
lengths or bond orders, suffer from the same disad- IA ) I6 ) 1.235I5 )1.341I6,6 ) 2.085I5,6 (21)
vantage: they overestimate the aromatic character
in cases where a molecule (or its fragment) has all The Hückel DE’s are well-known as an unreliable
Chart 1. HOMA, EN, and GEO Termsa for 27 Benzenoid Hydrocarbons Estimated from Experimental
Geometry182
a E, G, and H stand for HOMA, EN, and GEO, respectively.

representation of resonance energy.79,80,90 Conse- give more reliable lengths cause elongations and
quently, the aromaticity of many bicyclic hetero-π- consequently a higher EN term. Interestingly, the EN
electron systems described by IA is overestimated, as term for the corrected geometries is constantsthe
shown in Table 11, where a comparison with HOMA, extension of longer bonds is balanced by a shortening
EN, and GEO values is given for a few typical of the shorter bonds and the mean bond length does
examples. not change. This must correspond to variation in
In the case of HOMA, both terms (EN and GEO) bond alternation as detected by the GEO term. The
contribute to a decrease in aromaticity. An important above effects are not observed for uncorrected
question arises, are they independent? From a purely geometrysthe effect of libration on bond lengths is
statistical point of view, there is no reason the mean- too high and is reflected in shortening of bond
dependent function (EN) would have to be related to lengths, leading to a decrease in the EN value.
the variance-dependent function (GEO). Indeed, prac- Independent of the precision of the diffraction
tical correlations lead to the correlation coefficients, experiment, if a symmetrical molecule is placed in
R, for the regressions between these two terms equal the crystal lattice in a nonspecific position (i.e., it is
to 0.39 for 90 rings in TCNQ-EDA complexes and not in a position of symmetry required by the space
salts, 0.05 for 169 rings in 26 benzenoid hydro- group to which the crystal belongs), then the deter-
carbons, and 0.09 for 116 rings in p-disubstituted mined molecular geometry need not follow the sym-
benzene derivatives.68 From the point of view of the metry properties observed for the isolated molecule.
total energy, any geometrical deformation of a mol- This is due to an obvious experimental condition
ecule (or its fragment) requires extra energy. How- which holds that each bond length has its own, in
ever, the separation provided by the HOMA model principle independent, error of estimation. Therefore,
allows one to differentiate between two different for instance three outer rings in triphenylene (cf.
mechanisms of the decrease of aromaticity. Chart 1) may have different (geometry-based) indices
of aromaticity, despite C3 symmetry of the isolated
The geometry-based aromaticity indices employ
molecule.
bond lengths. The bond lengths determined by X-ray
or neutron diffraction may be biased by some sys- It should also be noted that there is no objection
tematic error due to data collection, method of to use molecular geometries computed by use of
refinement, static or dynamic disorder, or thermal reliable quantum chemical theories. In many cases
motion. This is well illustrated by the differences they are a source of important data, particularly for
found for benzene. Uncorrected for thermal motions, the systems which are unstable or where the effect
the CC bond length for benzene at 270 K is 1.378 of some (artificial) deformation is to be studied.
Å;91 with some correction for libration it is 1.392 Å, However, it should be emphasized here that only the
whereas the low temperature (15 K) neutron diffrac- equilibrium geometries should be applied, otherwise
tion geometry corrected for libration gives 1.3980 Å, some paradoxical situation may follow. For instance,
compared to 1.3972 Å for the uncorrected structure a constrained naphthalene geometry with all CC
at this temperature and 1.3940 Å for the uncorrected bonds of the length equal to Ropt would have HOMA
length at 123 K.50 These shortcomings of X-ray (and ) 1, but undoubtedly the energy of such a con-
neutron) diffraction data must always be taken into strained molecule is much higher than that of the
account when these kinds of measurements are the equilibrium structure. Nevertheless, in some con-
source of molecular geometry. Usually such correc- trolled cases nonequilibrium geometries were used
tions are much smaller, around 0.005 Å.45 Table 12 leading to interesting results and conclusions.
presents the geometry-based indices calculated for It should also be mentioned here that optimized
uncorrected and corrected geometries of naphtha- geometries obtained by ab initio calculations depend,
lene92 measured at three temperatures. usually in a systematic way,93 on basis sets and on
the level of theorystaking into account electron
As shown by the above data, the values of the correlation in many cases improves the situation.
indices do not vary much and 5-10% of the difference
in their values may be considered as significant
(depending on the precision of measurements). There III. Applications
are two tendencies observed: aromaticity character-
istics estimated from the uncorrected room temper- As previously underlined, the majority of geometry-
ature geometry are the highest (for HOMA and GEO based aromaticity indices may be calculated both for
terms) and the lowest for EN. When corrections for the whole molecule and/or locally for given frag-
thermal motions are applied, the aromaticity param- ment(s). This is clearly an advantage in comparison
eters for all three temperatures are much less with other indices, since in this way such indices may
dispersed. The parameters for the uncorrected and be compared with the other ones which are applicable
corrected geometries are closest to each other for the only for the whole molecules (e.g., exaltation of
lowest temperature measurements and most dis- diamagnetic susceptibility, ASE or RE values) or only
persed for room temperature. A significant increase for the individual rings (NICS, calculated ring cur-
of the EN term for the room temperature geometry rent values). Thus, the verification may be carried
for the corrected case (0.0852) as compared with the out from both sides in the same way as the compari-
value for the uncorrected geometry (0.0406) is quite sons of the aromatic character expressed by the
understandable: at room temperature, much strong- different indices or properties characteristic of aro-
er thermal motion leads to an artificial shortening maticity. Local aromaticity estimated for a given
of most of the bond lengths; hence, corrections which fragment of a larger molecule provides a possibility
to study the problems of the role of various environ- problems of this kind will be presented and discussed
mental perturbations which may affect aromaticity. in the following paragraphs.
This problem is particularly important in view of the
fact that the aromatic compounds (or molecules with B. Benzenoid Hydrocarbons
aromatic fragments) are active species in all aspects
For a long time benzenoid hydrocarbons have been
of life processes, both negative (poisons, environmen-
used to test various hypotheses dealing with aroma-
tal problems, wastes, etc.) and positive ones, enabling
ticity, stability, spectral properties, or reactivity.97,98
life processes (e.g., bases in DNA, RNA, porphyrines,
A great progress started with the work by Dewar
etc.), and also those externally affecting the quality
et al.99 Using the Pariser-Parr-Pople π-electron
of life (medicines, foods, etc.).94
method, they found that the bond energies of acyclic
Another important problem should be emphasized polyenes are additive. Hess and Schaad,79 applying
here. Since aromaticity is a phenomenon defined by the additivity rule and the simple HMO method,
convention, then we should be well aware of the fact presented a great number of REPE values (abbrevia-
that a geometry-based description of aromaticity may tion from Resonance energy per electron) for a
not always be in line with other descriptors,71,95 collection of benzenoid hydrocarbons100 (later ex-
although qualitatively it usually remains in agree- tended also onto nonalternants79 and even various
ment with them.41,42,95 hetero-π-electron systems101). It became clear that the
A general view may be presented in the following energy content varies dramatically in each of the
way. If benzene is assumed as the ideal model for an groups of systems studied.
aromatic compound, then the following way of intro- Application of HOMA and REC47,71 allowed re-
ducing various perturbations which convert this searchers to consider this variation by taking into
model into many other molecules can be considered. account the local aromaticity (i.e., when the indices
Let us start with the simplest case: one can replace are calculated for the individual rings). Chart 1
CH or CHCH fragments in a molecule of benzene by presents relevant data for the molecules for which
heteroatom(s). Thus, for instance, replacing CH by there are experimentally-accessible precise102 geom-
N and CHCH by NH leads to pyridine and pyrrole, etries.
respectively. These two molecules differ in aromatic The data presented above show a huge variation
character from benzene, and they also differ from in aromaticity (as expressed in HOMA values) for
each other. However, we can also think about greater individual rings depending dramatically on the to-
modifications leading to some more perturbed sys- pological position of the ring. The rings which have
tems, proceeding toward azoles and polyaza ana- the lowest aromatic character are those which, ac-
logues of benzene. Further modifications of this kind cording to the Clar characteristics,103 belong to the
may lead to macrocycles (annulenes, heteroanalogues class of “empty” rings, i.e., the ones containing fewer
of annulenes, etc.). Exocyclic substitution is still a π-electrons than other rings. The central rings in
new kind of perturbation modifying the properties triphenylene or perylene have HOMA values as low
of the π-electron system. Finally, we can consider a as 0.077 and 0.133, respectively. The REC values are
fusion of various systems of this kind, their interac- also very low: 668.9 and 648.1 kcal/mol, respectively.
tions with other systems (EDA complexes, sandwich These values may be compared with that for benzene
complexes, salts, complexes with metals, etc.). Re- itself for which REC ) 719.7 kcal/mol. An inde-
placing larger parts of benzene by other chemical pendent index, NICS,32 also shows very low aroma-
fragments may lead to almost all imaginable systems ticity for these two rings: -3.0 and +6.8,32 respec-
which may exhibit aromatic, nonaromatic, or anti- tively.
aromatic properties. Obviously in any of the above- HOMA with its components EN and GEO well
mentioned cases the aromatic character of new illustrate Clar’s classification103 of the rings in ben-
molecules is different from the original systems. Such zenoid hydrocarbons: empty rings (as the central one
different perturbation(s) may affect the chemical and in triphenylene) are less aromatic due to the elonga-
physicochemical properties in a nonequivalent way. tion of bonds, hence high values of EN and conse-
quently low values of HOMA; rings with localized
A. Topological Problems double bonds (as the central one in phenanthrene)
The topology of a molecule has a fundamental are less aromatic due to an increase of bond alterna-
influence on its physical, chemical, and biological Table 13. Correlation Coefficients for the Scatter
properties. The differences in the properties of the Plots between Aromaticity Indices NICS and HOMA
p-, m-, and o-disubstituted benzene derivatives are and the Ring Critical Points for σ and π Framework
well illustrated by different substituent constants σ (Gc, ∆Gc, and the Curvature of G, λ) for Benzene,
Naphthalene, Anthracene, Pyrene, Tetracene,
for these three positions.96 The energies of five Tiphenylene, Chrysene, Perylene, and Coronenea
structural isomers of C18H12 (Table 4) clearly show a
similar aspect, namely, a connectivity of atoms (i.e., index db ∇2Fπ λσ,3 or λπ,3 Fσ or Fπ
the topology of a molecule) that determines its HOMA/π -0.902 0.897 -0.919 0.919
chemical and physicochemical properties. The most HOMA/σ - -0.919 0.918
stable one, chrysene, is by almost 14 kcal/mol more NICS/π 0.870 -0.860 0.860 -0.857
NICS/σ - 0.876 -0.876
stable than the least stable one, tetracene! These
a For the π framework, a correlation with the distance of
kinds of variation in the properties due to topological
reasons no doubt must affect the aromatic character the stationary points from the nodal plane for the π-structure,
d, is given also. b A mean value of d is 0.8 Å.
of the systems or of their local fragments. Many
Table 14. Aromaticity Indices HOMA, EN, GEO, ASE [kcal/mol], and MEC/n [kcal/mol] Calculated as Global
Values (for whole molecules) for Linear and Angular Poliacenes (only σ-electron structures shown). Data in
Brackets Are Based on Experimental Geometries
Table 15. Correlation Coefficients of Regressions

between HOMA, EN, GEO, ASE/n (E), and MEC/n
Calculated for Whole Molecules (12 systems)
HOMA EN GEO E
EN -0.892
GEO -0.948 0.701
E 0.989 -0.847 -0.962
MEC/n 0.976 -0.939 -0.877 0.945
tion, hence higher values of GEO and consequently

again lower values of HOMA. Figure 3. Linear dependence of HOMA index on the
A comparison of HOMA with other geometry-based Cohen Benson ASE/n (E). Correlation coefficient R ) 0.989.
aromaticity indices has shown71 that the indices were (Reprinted with permission from ref 104b. Copyright 2000
Elsevier Science.)
not fully equivalent (even estimated for the same
geometry). Since the indices BAC and particularly BAC vs I6 correlates well, -0.94, due to almost the
I6 are “purely” bond alternation dependent, whereas same mechanistic background. Since both kinds of
HOMA depends also on the bond elongation, these deformations, bond alternation and bond elonga-
two kinds of indices clearly do not correlate with each tion, cost energy (HOMA contains components from
other (I6 and BAC vs HOMA give correlation coef- both), the correlation between REC and HOMA exists
ficients of only 0.56 and 0.67, respectively) whereas but is rather weak, whereas correlations of REC vs
Table 16. Mean Values of HOMA, EN, and GEO Terms for the Ring of Polysubstituted Benzene Derivatives
Compared with Topologically Equivalent Rings in Benzenoid Hydrocarbons
Table 16 (Continued)
I6 and BAC are practically nil (0.09 and -0.25,

respectively).
Another representative sample of benzenoids stud-
ied by use of HF/6-31G** calculations has shown41
that for the optimized geometry of 10 molecules (18
individual, symmetrically independent rings), the
HOMA index correlates well with NICS leading to a
regression line with a correlation coefficient R )
-0.896. This is a quite reasonable correlation if we
take into account the completely different physical
basis for their definition. For the same sample the
HOMA values exhibited quite good correlations with
topological charge density parameters75 such as Fc
(charge density at the unique ring critical points),
∆∇2Fc, and the curvature λ of F. Table 13 presents Figure 4. Analogy in the topology of the ring in hexa-
substituted benzene derivatives and of the central ring in
correlation coefficients between these parameters for triphenylene, perylene, and coronene. (Only the σ-electron
σ and π electron structures. structures are shown.)
Since the topological charge density parameters are
excellent descriptors of the electronic structure, it number of CC bonds), together with HOMA, EN, and
may be concluded that HOMA and NICS alike are GEO terms for whole molecules.
associated with this characteristic of π and σ electron When the whole-molecule aromaticity indices are
systems, at least as far as benzenoid hydrocarbons concerned, HOMA, EN, GEO, MEC/n, and Cohen-
are concerned. This result supports a view that there Benson E, it is clear that the aromaticity of these
is an interplay between the π and σ electron struc- systems varies monotonically with the increase of the
ture. The charge density analysis shows that both number of rings. This tendency is observed in both
structures contribute to the overall effect observed series, however with different sensitivity, which is
as aromaticity via indices as HOMA or NICS. in line with an observation of the changes of the
The molecular geometry of the whole molecule of position of the p-band in UV spectra in arenes. The
seven linear and five angular polyacenes was used aromaticity of the linear polyacenes is smaller than
to calculate the MEC values, which are compared that of the corresponding angular systems, and
with the aromatic stabilization energy (ASE) calcu- moreover, it decreases more rapidly with an increase
lated from the Benson and Cohen group additivity in the size of the system. This tendency is clearly
values (GAV) within the thermochemical proce- shown by E and HOMA and is roughly in line with a
dure.104 Table 14 presents normalized ASE values decrease in the chemical stability of polyacenes.
(divided per one π-electron) depicted as E, MEC/n (n Evidently the above observation is followed by the
mutual linear relationships between the above- One important conclusion may be drawn from the
mentioned indices, as shown by the statistical data above results. If we consider the links between the
of Table 15. ring in question and the closest atoms in the molec-
The best correlation was found between HOMA ular environments, we can distinguish two situations.
and ASE/n (E). This dependence is shown in Figure In the case of benzenoid hydrocarbons, all carbon
3. atoms are bound either to other carbon atoms or to
hydrogen atoms, both characterized by very similar
The benzene ring has often been used as a probe electronegativity (Pauling’s electronegativity values
of aromaticity in topologically analogous situations: for C and H are 2.5 and 2.1, respectively;106 Sander-
benzenoid hydrocarbons vs polysubstituted benzene son’s values are 2.746 and 2.592, respectively107). In
derivatives. Polysubstituted benzene derivatives (PBD) the case of the substituted benzene derivatives, the
may lead in the ring to a topologically equivalent situation is quite different. The substituents are of
situation as in benzenoid hydrocarbons: the ring in different kinds, and electronegativities of the atom(s)
naphthalene is topologically equivalent to o-disub- linked to the ring may vary significantly. Thus,
stituted benzene derivatives, whereas the central considerable inductive or resonance interactions may
ring in triphenylene (and, e.g., in coronene too) is operate in, or through, the ring. The data in Table
equivalent to the hexasubstituted benzene deriva- 16 lead to an important conclusion: perturbations
tives. Figure 4 illustrates this point: due to substituent effects acting on the ring in
A question arises, how does the ring respond in question cause a much smaller decrease in aroma-
these two situations in terms of aromaticity? The ticity than the topological ones. This finding may well
problem was analyzed in detail for X-ray measured be a reason for successful applications of theoretical
molecular geometries of polysubstituted benzenes methods based on topology and graph theory24 in
and benzenoid hydrocarbons.105,182 According to the describing some properties of the π-electron systems.
positions of substitution in benzene, 12 classes were It is interesting to note that perturbation of the
defined and used for both groups of systems, as π-electron system in the benzene ring due to proton-
shown in the schemes in Table 16. ation changes the aromaticity of the sp2 fragment
The range of variation of the mean HOMA values only to a very small extent. MP2/4-31G* geometry
for polysubstituted benzene derivatives is much optimization108 gives HOMA ) 0.9255, which may be
narrower (0.866-0.994) than that for the rings in compared with HOMA ) 0.9958 for benzene opti-
benzenoid hydrocarbons (0.219-0.862). The only mized at the same level of theory. For the whole ring
clear but rough conclusion is that the ortho substitu- (including the saturated carbon atom), HOMA )
tion leads to a greater decrease of aromaticity than 0.493. In that case the decrease in aromaticity stems
any other type: the mean HOMA value for eight from two C(sp2)-C(sp3) bonds. These results seem to
cases where ortho-substitution is present is 0.949 as support a conclusion that breaking the π-electron
compared with 0.993 for the four other cases where delocalization in aromatic systems by an sp3 carbon
ortho-substitution is absent. When the aromaticities atom does not lead to any dramatic change in the
of the rings in the benzenoid hydrocarbons are aromaticity of the remaining part of the molecule.
compared with substituted analogues, in all cases
they are lower than those in the substituted systems. C. Aza-Analogues of Benzenoid Hydrocarbons
In the latter case, alternation of bond length (the Aza-analogues of the benzenoid hydrocarbons are
GEO term) is always decisive concerning the decrease well-known to be similar in aromaticity to the parent
of aromaticity, whereas in the case of benzenoid compounds. Application of the geometry-based indi-
hydrocarbons this is true only of the analogues of di- ces of aromaticity makes it possible to analyze the
and trisubstituted benzene. In other cases an in- problem of changes of aromaticity due to the replace-
crease in the mean elongation of the CC bonds (an ment of the carbon atom(s) by aza-nitrogen atom(s),
increase in the EN term) becomes decisive, leading both in a local and global sense.109 There are at least
to a dramatically low aromaticity of the central rings two aspects of the problem: (a) the direct effect of
in groups VII, IX, X, and even XI. The following replacement and (b) the consequences in further
general conclusions may be drawn. (i) The terminal fragments of the π-electron system. Chart 2 presents
rings (group I) are more aromatic than in all other the HOMA, EN, and GEO values for aza-analogues
cases. The decrease in their aromaticity is mostly due of the benzenoid hydrocarbons.
to the GEO term (bond alternation). The same is true Following are some observations based on the data
of the rings of group II. (ii) Angularly fused rings in Chart 2, typified by the HOMA values for naph-
(central rings in phenanthrene, group III, and pyrene- thalene and its aza-derivatives. (i) All aza-analogues
type, group V) have lower aromatic character due to have higher HOMA than naphthalene. (ii) In all
an increase in the GEO term (bond alternation). (iii) cases, the decisive factor determining aromaticity is
In the case of more than three rings fused to the ring the GEO term, which is always much larger than EN.
in question (except group VI), the decrease in aro- (iii) The EN term is in all cases larger for naphtha-
maticity results from bond elongation (the EN term) lene than for any aza-analogue.
and these rings are usually of low aromaticity. (iv) Taking into account all aza-analogues of benzenoid
Other groups, except III, are represented by small hydrocarbons recently studied,109 it is apparent that
samples (groups IV, VI, VII), hence any conclusions in most cases replacement of the carbon atom by the
are limited, but the decrease in their aromatic nitrogen atom increases aromaticity, as compared
character is due to both terms. with the parent hydrocarbon. Moreover, low-aromatic
Chart 2. Local Values of HOMA, EN, and GEO Termsa for Aza-Analogues of Benzenoid Hydrocarbons109
a E, G, and H Stand for EN, GEO, and HOMA, respectively. (Reprinted with permission from ref 109. Copyright 1996 Elsevier Science.)
rings in the molecule are strongly affected by the A good illustration of the above statements is given
substitution of nitrogen atoms into the ring, and the by aza-analogues of phenanthrene. The HOMA value
effect is still significant if the nitrogen atom is closest for the central ring in phenanthrene is 0.400. How-
(i.e., directly linked) to the given ring. ever, if the carbon atoms in position 9 and then in
Table 17. Aromaticity Parameters: HOMA (H), EN (E), Table 18. Aromaticity Parameters: HOMA, EN, GEO,
GEO (G), and I6 for Aza-Analogues of Benzene67 and I6 for Hetero-Analogues of Benzene
D. Other Heterocyclic Analogues of Benzenoid

Hydrocarbons.
There are many possibilities of replacing the CH
group in benzenoid hydrocarbons by heteroatoms,
and a few good examples are presented in Table 18,
which contains a few analogues of benzene.
Compared with pyridine, which shows high aro-
maticity, the changes observed in the aromatic
character of phosphorus, sulfur, and oxygen ana-
logues are in line with expectation.11
The pyrylium ring in a series of 2,6-diphenyl-4-(4-
X-substituted-phenyl)pyrylium perchlorates (Figure
positions 9 and 10 are replaced by the nitrogen 5) was one of the systems studied in more detail.112
atoms, then HOMA values rise up to 0.618 and 0.791,
respectively. When the carbon atoms in positions 4
and 5 are replaced, a substantial increase in the
aromaticity of the central ring is observed. HOMA
rises from 0.400 to 0.625. The increase in aromaticity
is also observed in a CT complex with TCNQ with
HOMA ) 0.705 or in the protonated species where
HOMA ) 0.666. A similar result is shown for chry-
sene: HOMA for the central ring is 0.503, whereas
in the aza-analogue with nitrogen atoms in positions Figure 5. 2,6-Diphenyl-4-(4-X-substituted-phenyl)pyr-
4 and 10 it rises dramatically up to 0.651. Also, in ylium perchlorates.
perylene the central ring becomes much more aro-
matic when the carbon atoms in positions 1 and 7 The aromatic character of the pyrylium ring was
are replaced by nitrogen; HOMA increases from 0.133 studied in dependence of the substituent. The het-
to 0.466. erocyclic ring is strongly electron attracting; hence,
However, another question arises, namely, whether the donating substituents should affect its aroma-
the replacement of CH by N always leads to an ticity.
increase in the aromaticity of the system. Recent For four substituted species with X ) NMe2, H,
studies on acridine derivatives showed110 a slightly COOH, and NO2, the HOMA values for the pyrylium
increased aromaticity of the central ring (HOMA ) ring depended significantly on the σ+ substituent
0.810 as compared with 0.763 in anthracene) in constants. In a similar way, the variation of the CC
acridine and a slightly decreased aromatic character bond lengths (for the link between the phenyl and
of carbocylic fragments. pyrylium rings) and of the dihedral angles between
Apart from the above problems, there are many the rings were observed. The pyrylium ring is a
aza-analogues of benzene itself. Table 17 presents strongly electron-accepting moiety; hence, the induc-
aromaticity characteristics for them.67 tion of the quinoid structure caused by the electron-
The aromaticity of aza-analogues of benzene is high donating substituents (NMe2) results in a substantial
and not differentiated. This conclusion is qualita- decrease of the aromatic character of the pyrylium
tively in line with resonance energies computed for ring (HOMA values ranging between 0.237 for NMe2
reactions with proton transfer.11,111 and 0.605 for NO286). The difference in the HOMA
As observed for aza-analogues of the benzenoid values (∆ ) 0.368) for NMe2- and NO2-substituted
hydrocarbons, when NN bonds are present in the derivatives shows that the aromatic character of the
ring, its aromaticity is usually lowered. In contrast, ring is very sensitive to the substituent effect. For
the “meta”- or “para”-type replacements lead to comparison, an analogous difference for the pyri-
higher aromaticity. s-Pyrazine is even more aromatic dinium ring in 1-X-4-(4-phenyl)-2,6-diphenylpyri-
than benzene itself, but the difference does not seem dinium perchlorates is equal only to 0.15.113 In both
to be significant. This observation is not too distant cases the substituent effect is transmitted through
from chemical experience and intuition. the benzene ring. In the case of p-substituted nitro-
Chart 3. Aromaticity of Oxa-Analogues of Benzenoid Hydrocarbons
benzene, where the substituent effect acts directly Table 19. HOMA, EN, and GEO Terms of Annulenes
and Some of Their Derivatives
on the ring, the difference is also very small, ∆ )
0.11. It is apparent that HOMA values for the system HOMA EN GEO comments
pyrylium ring are much more dependent on the same annulene [4] -2.570 1.317 2.253 ref 11
changes of substituent than the highly aromatic annulene [6] 0.996 0.004 0.000 X-raya
pyridinium or benzene rings. 0.974 0.026 0.000 ND50
annulene [8] -0.295 0.103 1.192 EDb
Replacement of the CH group in benzenoid hydro- -0.383 0.028 1.355 X-rayc
carbons by nitrogen has been discussed in a previous annulene [10] 0.930 0.004 0.066 D5hd,e
section. An analogous study of replacement by oxygen 0.783 0.217 0.000 D10hd,e
has shown that the rings which contain the oxygen annulene [14] 0.942 -0.004 0.064 X-rayf
atom definitely become less aromatic (or even anti- Annulene [18] 0.899 0.010 0.091 X-rayg
aromatic) whereas the vicinal rings gain higher a Cox, E. G.;Cruickshank, D. W. J.; Smith, J. A. S. Nature
aromaticity.114 Chart 3 provides a good illustration 1955, 175, 766. b Traettenberg, M. Acta Chem. Scand. 1966,
of these trends. 20, 1724. c Cobbledick, R. E.; Einstein, F. W. B. Acta Crystal-
logr. 1977, B33, 2339. d At MP2/6-31G//HF/6-31G: Haddon,
R. C.; Raghavachari, J. Am. Chem. Soc. 1982, 104, 3516. e At
E. Nonbenzenoid Alternant Hydrocarbons and MP2/6-31G//HF/6-31G: Haddon, R. C.; Raghavachari, J. Am.
Their Heteroanalogues Chem. Soc. 1985, 107, 289. f Vogel, E.; Engels, H.-W.; Huber,
W.; Lex, J.; Mullen, K. J. Am. Chem. Soc. 1982, 104, 3729.
While benzenoid hydrocarbons are the subject of g Bergman, J.; Hirshfeld, F. L.; Rabinovich, D.; Schmidt, G.
long-standing studies (many of their derivatives are M. J. Acta Crystallogr. 1965, 19, 227.
applied in all fields of human activities), nonben-
zenoid hydrocarbons are a subject of rather moderate
interest. Undoubtedly, some of these systems have
substantial applications and create a significant
theoretical interest. Hence, many of them have been
subject to thorough structural studies. Their molec-
ular geometries may serve as a good source of Figure 6. Syn (1) and anti (2) conformers of bridged
[14]annulene. (Reprinted with permission from ref 116.
information to study their aromaticity. Copyright 1986 Wiley-VCH Verlag GmbH.)
The first group of alternant nonbenzenoid hydro-
carbons are annulenes. Except for benzene and All (4N + 2) annulenes clearly exhibit aromatic
cyclobutadiene, all others are nonplanar. Table 19 character, whereas 4N annulenes are definitely anti-
presents the aromaticity indices for some annulenes aromatic, with negative HOMA.
and their derivatives calculated either from the A slightly different class of molecules is bridged
experimental or ab initio calculated geometries. annulenes, which are often close to the benzenoid
Table 20. HOMA, EN, and GEO Values for Bridged in the case of fulvene and electron-accepting ones in
[14]Annulene Derivatives Estimated from Perimeter the case of heptafulvene strongly stabilize the rings,
Bond Lengths
leading to a substantial increase in their aromatic
compound HOMA EN GEO character, which is in line with Figure 7.
anti-1,6:7,12-bismethano[14]annulene116 0.241 0.034 0.215
syn-1,6:7,12-bismethano[14]annulene115 0.749 0.036 0.215
syn-1,6;8,13-bisoxido[14]annulenea 0.992 0.000 0.008
1,6:8,13-propanediylidene[14]annulenea 0.900 0.052 0.048
syn-1,6:8,13-biscarbonyl[14] annuleneb 0.939 0.017 0.044
7-methoxycarbonyl-anti-1,6:8,13- 0.021 0.066 0.913
dimethano[14]annulenec
a Vogel, E. Pure Appl. Chem. 1971, 355. b Destro, R.; Simo-
netta, M. Acta Cryst. 1977, B33, 3219. c Gramaccioli, C. M.; Figure 7. Illustration of the Hückel rule for the fulvene
Mimun, A. S.; Mugnoli, A.; Simonetta, M. J. Am. Chem. Soc. and heptafulvene derivatives. (Reprinted with permission
1973, 95, 3149. from ref 65. Copyright 1995 Slovak Academy of Sciences.)
hydrocarbons. A very interesting situation is ob- The changes in aromaticity are unlike those for the
served for bridged [14]annulenes with a phenan- substituted benzenes, where they are much smaller.105
threne perimeter. Recent studies on exocyclically substituted fulvene
Table 20 presents HOMA, EN, and GEO values for derivatives have demonstrated a very good system-
these two kind of conformers of bridged [14]annulene. atic dependence between aromaticity and the nature
The syn115 and anti116 conformers of bridged an- of substituents, as shown in Figure 8.
nulene[14] derivatives differ much in aromaticity.
Application of HOMA and its components to the
perimeter bond lengths allowed us to look at it in a
more subtle way. Low aromaticity for the anti
conformers of the bridged [14]annulenes is docu-
mented by the low values of HOMA (<0.24) with a
dominant contribution to the decrease in aromaticity
from the geometric factor (large GEO terms). The
situation is quite different for the syn conformers for Figure 8. Dependence between aromaticity index HOMA
which HOMA values for the perimeter bond lengths and the substituent constants σ+. The correlation coefficient
R ) 0.970.
are always high, thus pointing to their aromatic
character. The above values may be compared with
Pentalene is undoubtedly antiaromatic, with a
the value obtained for [14]annulene with HOMA )
large negative value of REPE which equals -0.018,
0.942 and phenanthrene (HOMA ) 0.750.)
large and positive value of NICS (+18.0 at HF/6-
31+G* for the geometry at DFT B3LYP/6-311+G**
F. Nonalternant π-Electron Hydrocarbons level of theory), and HOMA ) -0.381.119 Following
There are four classical nonalternant π-electron the Hückel rule, its dianion is considered to be
hydrocarbons which have long been the subject of aromatic, and hence, the metal complexes are well-
many studies of their aromatic character: fulvene, known.120 Schleyer’s NICS value for the pentalene
heptafulvene, azulene, and pentalene.117 Fulvene, dianion is -12.1, indicating its strong aromatic
heptafulvene, and pentalene have been considered for character.32,121 Similarly, the HOMA value calculated
a long time as nonaromatic (or even antiaromatic), from the optimized geometry at the DFT B3LYP/6-
both for their chemical instability and the negative 311+G** level of theory is 0.468, in a qualitative
REPE values.79 More recent studies strongly support agreement with NICS.
this hypothesis. The X-ray determined geometry of Azulene is a valence isomer of naphthalene, and
heptafulvene118 allowed us to estimate the aromatic- the fusion of five- and seven-membered rings allows
ity: the HOMA value for the ring is equal to 0.257. them to interact in order to fulfill the Hückel rule.
The case of fulvene is more complicatedsonly geom- Old estimations of its aromaticity via REPE ) 0.02379
etries for the substituted species are known. The indicates a moderate aromatic character, in line with
mean HOMA value calculated for 11 exocyclically HOMA values for the whole molecule122 (HOMA )
substituted fulvenes is 0.001.65 0.665). Individual rings may be treated by NICS and
The situation changes dramatically if fulvene and HOMA: for the five-membered ring the values are
heptafulvene are substituted exocyclically by electron- -19.7 and 0.286, respectively, and for the seven-
donating (e.g., NMe2) and electron-accepting (e.g., membered ring -7.0 and 0.514, based on the lowest
CHO) groups, respectively. In both cases, aromaticity energy isomer at B3LYP/6-31G*.
strongly increases to the values for HOMA of 0.669 Tropolone is another famous nonalternant system
and 0.769, respectively. In both cases an increase in whose aromaticity has long been disputed. Recent
the REC values is also observed: for fulvene rings studies of its anion and two conformers of protonated
from 576.7 to 591.6 kcal/mol and for heptafulvene tropolone123 have given us a good opportunity to
rings from 842.2 to 850.3 kcal/mol. In both cases the analyze how the aromaticity is affected by the pro-
rings very clearly tend to follow the Hückel (4N + 2) tonation. The HOMA for tropolone is equal to 0.642.
rule and may serve as good examples of the substitu- Its anion is definitely less aromatic, HOMA ) 0.176,
ent effect on aromaticity. Electron-donating groups whereas cis and trans conformers of protonated
tropolone are highly aromatic, HOMA ) 0.924 and Table 21. Statistical Parameters for HOMA and C-X
0.910, respectively. This is a nice illustration of the Interatomic Distances127
Hückel rule: a positively charged hydroxy group in RC-X averaged
the cationic form withdraws electrons from the ring, sub- inter- interatomic inter- variance
leading closer to the (4N + 2) number of electrons, sample quartile distance quartile of C-X
element size HOMAav range (averaged) range in rings
whereas a negatively charged phenolic oxygen acts
in the opposite direction. Li 18 0.602 0.424 2.394 0.085 0.051
Be 3 0.927 0.111 1.887 0.031 0.000
Finally, we should mention some more complex B 13 0.303 0.098 2.943 0.043 0.916
fused systems with nonalternant rings. A good ex- N 14 0.466 0.349 2.724 0.067 0.889
ample is dicyclohepta[cd,gh]pentalene derivative O 15 0.331 0.746 2.682 0.062 0.891
(methyl 10-ethyldicyclohepta[cd,gh]pentalene-5-car- Na 2 0.729 0.398 2.954 0.153 0.015
boxylate)124 for which the hydrocarbon skeletone is Mg 6 0.703 0.344 2.412 0.148 0.002
Al 5 0.429 0.695 2.648 0.911 0.280
shown in Figure 9:
Si 104 0.540 0.384 3.143 0.023 0.835
P 132 0.614 0.156 3.123 0.045 0.881
S 43 0.636 0.255 3.087 0.015 0.895
Cl 12 0.795 0.091 3.034 0.012 0.876
K 1 0.851 0.000 3.035 0.000 0.003
Ca 9 0.816 0.038 2.649 0.014 0.001
Ti 68 0.788 0.192 2.352 0.024 0.004
V 64 0.822 0.178 2.292 0.033 0.003
Figure 9. Dicyclohepta[cd,gh]pentalene. Cr 145 0.778 0.220 2.235 0.055 0.009
Mn 214 0.752 0.151 2.158 0.020 0.006
Fe 1222 0.723 0.237 2.076 0.061 0.005
The molecule is aromatic (global HOMA value is
Co 398 0.709 0.268 2.083 0.038 0.008
0.734), resulting from the rule39 stating that when Ni 105 0.689 0.238 2.124 0.025 0.002
the numbers of five-membered and seven-membered Cu 2 0.931 0.055 2.213 0.004 0.001
rings are equal, then the system should maintain the Zn 5 0.758 0.293 2.447 0.340 0.075
aromatic character. Interestingly, the aromatic char- Ga 3 0.599 0.000 3.484 0.300 0.636
acter is different in these two kinds of rings: the Ge 7 0.741 0.157 2.754 0.751 0.266
As 4 0.592 0.107 2.843 0.760 0.413
seven-membered rings are more aromatic and have Se 6 0.630 0.094 3.224 0.004 0.891
HOMA equal to 0.818, whereas the five membered Br 8 0.675 0.300 3.183 0.030 0.873
rings have only 0.500. This difference between the Y 2 0.728 0.045 2.654 0.002 0.000
aromatic character of seven- and five-membered rings Zr 15 0.888 0.127 2.528 0.055 0.000
becomes even more dramatic if in positions 7 and 14 Nb 35 0.868 0.101 2.423 0.046 0.002
Mo 481 0.781 0.175 2.350 0.027 0.005
carbonyl groups are attached.125 Then the five- Tc 2 0.497 0.175 2.299 0.018 0.001
membered rings become antiaromatic with HOMA Ru 332 0.690 0.231 2.241 0.067 0.014
equal to -0.258, whereas aromaticity of the seven- Rh 231 0.584 0.261 2.237 0.048 0.002
membered rings is almost unchanged: HOMA ) Pd 15 0.642 0.289 2.444 0.039 0.063
0.822. In an unsubstituted molecule,126 five-mem- In 10 0.777 0.004 2.863 0.396 0.008
Sn 23 0.597 0.299 2.864 0.766 0.243
bered rings are also less aromatic with a HOMA
Sb 1 0.865 0.000 2.682 0.000 0.092
value 0.663 whereas a seven-membered ring has Ba 12 0.760 0.064 2.971 0.067 0.001
HOMA ) 0.823, with global HOMA ) 0.791. It seems La 2 0.896 0.018 2.813 0.022 0.003
that five-membered rings are sensitive to substitu- Hf 5 0.769 0.183 2.498 0.022 0.000
tion, and the perturbation does not go further to Ta 13 0.673 0.307 2.457 0.063 0.002
seven-membered rings. W 122 0.764 0.181 2.345 0.030 0.002
Re 76 0.600 0.352 2.319 0.038 0.017
Os 8 0.185 0.323 3.063 0.775 0.571
G. Cyclopentadienyl and Cyclopentadiene Ir 44 0.539 0.192 2.225 0.058 0.001
Moieties Embedded in Various Chemical Pt 7 0.656 0.140 2.635 0.951 0.259
Environments Au 1 -0.025 0.000 2.961 0.000 0.347
Hg 10 0.300 0.553 3.393 0.011 0.920
Cyclopentadiene is known to be a relatively strong Tl 4 0.570 0.077 2.925 0.336 0.003
Brønsted acid (pKa ) 1648), and as such it is able to
form salts. It also forms covalent compounds with characteristics of the dispersion since the distribu-
most nonmetal elements as well as complexes with tions were frequently not normal).128
metals. Cyclopentadienyl rings form different compounds
Until the mid-1990s, the CSD base46 archived 4079 depending on the element linked to them. In some
derivatives of cyclopentadienyl and cyclopentadiene cases, a more or less deformed cyclopentadiene-like
bound to almost all elements from Li to Tl.127 Table ring is involved in the complex systems of interac-
21 presents the values of HOMA index estimated as tions with atoms (ions) other than carbon elements.
a mean value for the subsamples built up of the The typically nonmetallic elements are bound via a
compounds of cyclopentadiene or cyclopentadienyl covalent link C-X. Typical ionic interactions are
moieties with particular elements. In most cases the encountered for compounds with s-block elements of
data observed are greatly dispersed; hence, statistical the Periodic Table. Coordination compounds are
characteristics are given as averages and the inter- formed mainly with transition metals. These kind of
quartile ranges of HOMA values as well as the mean compounds are the most complex, since they are often
C-X interatomic distances are given with their built up of several ligands which are located in
interquartile ranges (variances could not be used for various places in the spectrochemical series and may
Table 22. HOMA, EN, GEO, and Bird’s I5 Indices for

Five-Membered Heterocyclics67
Figure 10. Molecular structure of tris(cyclopentadienyl)-

nitrosomolybdenium. (Reprinted with permission from ref
127. Copyright 1996 Elsevier Science.)
differ considerably in their power of interacting with

the central atom, affecting its interactions with the
cyclopentadienyl moieties. Consequently, dramatic
differences in aromaticity of these rings may be
observed. A good example is the structure of tris-
(cyclopentadienyl)nitrosomolybdenium129 shown in
Figure 10. Two rings interact with the Mo central
atom in almost the same way. The variances in the
C-Mo distances for those two rings are small: 0.025
and 0.016 Å2. Both rings are still aromatic, with
HOMA ) 0.913 and 0.717. The third ring is bound
to the Mo atom in a quite different waysone carbon
atom is sp3 type and much closer to the Mo atom than
any others, and hence the variance of C-Mo inter-
atomic distances for this ring is large (0.58 Å2), which
is responsible for a significant decrease of aromatic-
ity, HOMA ) 0.016. The ring simply resembles a and benzothiophene, calculating optimized geom-
cyclopentadiene ring. etries and a collection of aromaticity indices. Chart
From the data in Table 21 it is easy to draw 4 presents these data enriched by HOMA values
conclusions based on a similar method of interpreta- computed from the optimized geometry.
tion. In the right-hand column are collected averaged
The topology of the heteroatoms in benzofuran and
variances of C-X distances calculated for each ring
benzothiophene exerts different effects on aromatic-
separately and the variances averaged for the whole
ity, as seen in a different way by HOMA and NICS:
subsample. Low values of these quantities mean that
in both cases, 1-endo-substitution decreases the aro-
all carbon atoms of the five-membered ring are a very
maticity of the benzene ring in terms of HOMA, 0.969
similar distance from the X atom. This is encountered
and 0.937 compared with benzene (0.987), but it
mostly in the case of sandwich-type complexes and
increases the aromaticity in terms of NICS, -11.6
in ionic derivatives of cyclopentadienyl anion in the
and -10.7 compared with -9.7 for benzene. In the
crystalline states.
case of HOMA, a similar situation is observed for the
five-membered rings: 0.064 and 0.547 compared with
H. Heterocyclic Analogues of Nonalternant 0.200 and 0.745 for the oxygen- and sulfur-containing
Hydrocarbons rings. In the case of 2-endo-substitution, both NICS
Monocyclic systems, heteroanalogues of the cyclo- and HOMA exhibit a decrease in the aromaticity of
pentadienyl anion, are most often encountered in the benzene ring, but again there is no full agreement
organic synthesis. Application of HOMA, EN, and in the description of five-membered rings.
GEO to this class of compounds allows one to The aromatic character of furofuran and thienothio-
estimate their aromaticity from experimental geom- phenes strongly depends on the topology of endo-
etries. Table 22 presents a collection of data. substituents. The same observation is found for
A classical problem of the changes in aromaticity pyrropyrrole and a mixed systemsfuropyrrole for
in the series, pyrrole, thiophene, and furan,56,130-132 which the calculations were carried out at the same
is also shown by the present data and is in line with level of theory.119 Chart 5 presents these data.
chemical expectation. In all cases, the [1,4]-derivatives are more aromatic
Recently Schleyer et al.121 studied the aromaticity than [1,6]-di-endo-substituted pentalene. Interest-
of bicyclics, furofuran, thienothiophene, benzofuran, ingly, only pyrrole fused to furan with [1,4]-endo-
Chart 4. Aromaticity Indices HOMA and NICS of Furofuran, Thienothiophenes, Benzofuran, and
Benzothiophene
Chart 5. Aromaticity Indices of Pyropyrroles and Furopyrroles
substitution exhibits an increase of the aromaticity and 0.792 for 1H and 2H tautomers, respectively,
compared with the individual ring. In all other cases which is in line with the energetic expectation.
the fusion leads to a decrease in aromaticity. These However, a really dramatic difference is observed for
effects are supported by both NICS and HOMA and the benzene fragment in these two tautomeric
are in line with the topological charge stabilization forms: for the 1H tautomer HOMA for this ring is
rule.133 0.897 compared with HOMA ) 0.810 for the 2H
Another problem is when the changes in aroma- tautomer. The decrease in aromaticity of the benzene
ticity of heterobicyclics are considered as a result of ring is associated here with the shortening of both
exo-substitution. X-ray and ab initio studies134 of CN and CC bonds from the five-membered rings.
eight derivatives of furo[3,2-b]pyrrole and five de- A similar picture was observed in the case of
rivatives of furo-[2,3-b]pyrrole lead to the conclusion bispyrazolo[3,4-b;4,3-e]pyridine derivatives.138 The
that the aromaticity of both π-electron systems pyrazole rings fused to pyridine have much lower
exhibits a strong dependence on the substituent aromaticity (HOMA around 0.65) than unfused spe-
effect, much stronger than that observed in benzene cies, for which HOMA is about 0.92. These data based
derivatives.105 on experimental geometries are extended by the
Very recently experimental geometries of 1,3,4- theoretically computed geometries onto a wider range
thiadiazole derivatives have been studied with regard of substituted systems, as shown in Table 23. The
to their aromatic character.135 It was shown that the pyrazole ring becomes less aromatic when exocyclic
aromaticity of the ring in these systems depends bonds to substituents (or in the fused ring) and are
significantly on the nature of the substituent. The shortened as shown in the schemes in Figure 12.
electron-donating NH2 group increases the 1,3,4- As can be seen in Table 23, a decisive component
thiadiazole ring aromaticity by about 0.06-0.07 unit of the decrease in aromaticity in those cases is an
of HOMA as compared with the systems substituted increase in bond alternation measured by an increase
by less donating groups such as NHCOMe or S.
Moreover, the ring in thiadiazoles is significantly less
aromatic than in triazole.135
In the case of nonalternant heterocyclic systems
fused to benzene, an interesting case is that in
indazole two existing tautomers (Figure 11) differ by
2.9 kcal/mol, thus indicating 1H tautomer to be more
stable.136,137 HOMA shows the values equal to 0.808 Figure 11. Tautomers of indazole.
Table 23. Aromaticity Indices for the Pyrazole Table 24. HOMA, EN, GEO, and NICS Values for
Fragment of Model Molecules Based on RHF/6-311G** Porphyrins, Their Metal Complexes, and the Dianiona
Geometry
NICS
HOMAb HOMAc ENb GEOb (ppm)c
porphyrine 0.652 0.666 0.107 0.240 -16.5d
ring (1,3) 0.666 0.702 0.132 0.202 -15.2
ring (2,4) 0.452 0.381 0.175 0.373 -4.5
internal cross 0.880 0.932 0.075 0.045 -21.7
porphyrine-metal 0.656 0.671 0.094 0.249
ring (1,3) 0.566 0.556 0.158 0.276 -10.0
ring (2,4) 0.524 0.556 0.153 0.323 -10.0
internal cross 0.872 0.906 0.071 0.057 -21.0
porphyrine dianion 0.448 -14.9d
ring (1,2,3,4) 0.287 -4.0
internal cross 0.908 -17.5
porphyrine-MII 0.638 0.102 0.259
porphyrine-MIII 0.682 0.084 0.233
a MII and MIII stand for divalent and trivalent metals,
respectively b Mean experimental geometries. c Calculated

geometries. d Ring center value, see Figure 13.
Figure 13. Enumeration of the rings in porphyrine.

“Internal cross” shown in bold (see text). (Reprinted with
permission from ref 42. Copyright 1998 Wiley-VCH Verlag
GmbH.)
instead are an integral part of the aromatic system.
The unprotonated five-membered rings have low
aromatic character, implying that the C2H2 groups
function as exocyclic “bridges”. A more appropriate
Figure 12. Scheme of the decrease of aromaticity for description of the free base porphyrin should prob-
pyrazole ring. (Reprinted with permission from ref 138. ably be as a 22 π-electron system rather than the
Copyright 1998 Elsevier Science.) [18]annulene model as shown at Figure 14.
Lower aromaticity of both kinds of pyrrole moieties
of the term GEO. A very good agreement between
is associated with an increase of the double-bond
the HOMA estimation of aromaticity and the values
character of CC(H)C bridges, leading to more local-
of Schleyer’s NICS is observed.
ized structures.42 This kind of effect has been dis-
Porphyrins are a good example in which the local
cussed in detail in the case of porphyrins42 and
aromatic properties in pyrrole moieties is associated
pyrazoles.138
with a global property measured in the center of the
In the case of metal complexes, all four pyrrole
macrocycle.42 An extensive sample of 456 porphyrines
rings are incorporated into the aromatic system,
including complexes with metals (424) of which the
while in the case of dianions, the π-electronic struc-
molecular geometry was released from CSD46 was
ture is more strongly-localized on the central part,
treated by the HOMA model. On the other hand, the
which can genuinely be considered as an 18 π-elec-
ab initio optimization of the unsubstituted por-
tron species plus four C2H2 bridging groups.
phyrine, its Mg2+ complex, and dianion at the B3LYP/
6-31G* level of theory allowed us to look at these
systems in an independent way including application
I. Dependence of Aromaticity on Intermolecular
of the magnetic index NICS.32 In spite of the quite
Interactions
different sources of data, the conclusions drawn from Since aromaticity undoubtedly depends on the
experimental and theoretical data were very similar. electron structure of the system in question, it should
Table 24 presents the resulting data. also depend on the intermolecular interactions. In-
In the free base porphyrin, the most aromatic part deed, it was shown140 that the geometry of the ring
is the “internal cross” (Figure 13). The pyrrole rings in hydrated nitrosophenolate anions in salts with
with NH groups are clearly more aromatic (HOMA sodium and magnesium cations varies significantly.
) 0.666, NICS ) -15.2) than the other five- When the HOMA, EN, and GEO indices are calcu-
membered rings (HOMA ) 0.452, NICS ) -4.5). lated from the molecular geometry of the ring, in
Hence, the NH groups are not the inert bridging these two cases the changes in the aromatic character
groups as suggested by the [18]annulene model139 but are significant, as shown in Figure 15.
Figure 14. Two models of the π-electron system of porphyrin: (left) as a bridged [18]annulene derivative;139 (right) as
macrocyclic 22 π-electron system.42 (Reprinted with permission from ref 42. Copyright 1998 Wiley-VCH Verlag GmbH.)
points to a known fact that in quinoidal structures

four bonds are lengthened and only two are short-
ened.
Another group of systems showing a significant
influence of the intermolecular interactions on aro-
maticity are EDA complexes. These compounds crys-
tallize well and have often been subject to thorough
structural studies. A good illustration is given by
tetracyanoquinodimethane (TCNQ) and its salts.
HOMA for TCNQ is 0.377, whereas for its Na salt is
0.724. Evidently the quinoidal structure of TCNQ is
not aromatic and acceptance of π-electrons leads to
an increase in aromaticity. A rough dependence
between the charge transferred to the TCNQ moi-
ety141,142 and the HOMA value of the ring was found.8
J. Aromaticity vs Deviation from Planarity

Benzene itself in the crystalline state studied at
Figure 15. Interatomic distances in H-bonds (Å, in 15 K is not planar.50 Its slight chair conformation is
parentheses), their energies (kcal/mol) of interactions, and
estimated aromaticity descriptors for the p-nitrosopheno- caused by weak intermolecular interactions in the
late anion in two salts: sodium and magnesium, respec- crystal lattice. According to Kitaygorodsky,143 the
tively. (Reprinted with permission from ref 47. Copyright energy of interactions between two adjacent chemical
1998 Elsevier Science.) species is in the range of 1-3 kcal/mol. This is
sufficient to bend the ring in the crystalline benzene.
The difference between the sodium and magnesium Furthermore, this means that the out of plane
salts consists of the nature of the hydration sphere deformation of the aromatic ring does not need much
around the anion. In the case of the sodium salt (I), energy.144 Indeed, it was shown that close contacts
there is a trihydrate and the water molecules interact in the crystal lattice, exemplified by the studies on
via the hydrogen bonding mainly with the nitroso sodium p-nitrobenzoate trihydrate, cause significant
group, whereas in the case of the magnesium salt (II), deformations in the ring, which are proportional to
which exists as a hexahydrate, both basic terminals the strength of these interactions.145 In view of those
are almost equally hydrated. The energies presented observations, a question arises: how do distortions
in the scheme visualize that the oxygen atom in the from planarity affect the aromaticity of the ring(s)?
nitroso group of I is much more hydrated (E ) 3.54 Analysis of the experimental molecular geometry
kcal/mol) than the oxygen atom in the nitroso group (X-ray) of 1,4-dichloro-2,6[5]metacyclophane146 shows
of II (E ) 2.53 kcal/mol), particularly if compared its high aromaticity (HOMA ) 0.985), despite sub-
with the hydration of the oxygen atom in the phe- stantial distortions from the planar conformation, as
nolate group of I (E ) 0.98 kcal/mol) and II (E ) 4.52 shown in Figure 16. This is in line with the 1H NMR
kcal/mol). In consequence, the negative charge in I signals of the ring protons but against the increased
is mostly located at the nitroso oxygen, whereas in reactivity of this compound.147
the case of II, it is shared with the phenolate oxygen. A similar situation is observed for [5]-p-cyclo-
Thus, in I shorter CO and CN bond lengths are phane,119,148 where the HOMA is also high, 0.953,
observed and hence a stronger contribution of the despite a substantial folding of the ring. The same
quinoidal structure than in II, leading to a larger results were obtained from magnetic studies of the
decrease in the aromatic character in I as compared folded benzenes.149
with II. Per-substituted naphthalene derivatives are also
Analysis of the HOMA components shows that the convenient model systems to exemplify the changes
main contribution to the observed decrease in aro- in aromaticity due to the folding of the aromatic
maticity stems from the GEO term (around 70% of moiety. The distortions lead the moiety to a chiral
the total decrease) but the other part, due to EN, conformation as shown in Figure 17.
Figure 16. Scheme of [5]metacyclophane indicating fold-

ing of the ring: R1 ) 26.8°, R2 ) 12°.
Figure 18. The σ- and π-separated energies156 for benzene

and its Kekulé structures.
Chart 6. Local Values of HOMA, EN, and GEO

Terms for Chrysene and Its Moiety in EDA
Figure 17. Projection of per-bromo-substituted naphtha- Complexes with TCNQ
lene along C4a-C8a bond.
Table 25. HOMA Values and Nonplanarity Parameters
for a Few Persubstituted Derivatives of Naphthalene
(octa-X-naphthalene)
maximal deviation of the
carbon atom from the best plane
X HOMA of the naphthalene moiety (in Å)
H 0.810 0.003
methyl 0.723 0.234
S-Ph 0.715 0.293
F 0.743 0.007
Cl 0.730 0.213
Br 0.431 0.427
Table 25 presents a magnitude of the deformation

from planarity (as the value of the maximal deviation
of carbon atoms from the least-squares plane for
carbon atoms) and HOMA values. The presented
results show that small folding of the naphtha- then their distortive tendency might be decreased
lene moiety does not substantially affect the aroma- and the system should exhibit less bond alternation.
ticity. Consequently, a higher aromaticity of the system152
Recent synthesis of pyrenophanes150 allowed us to is expected. Such a situation is encountered in the
study the dependence of the global aromaticity of the case of chrysene and its EDA complexes with TCNQ
pyrene moiety on the degree of its bending and also and fluoranil.152,157 Chart 6 presents HOMA values
how individual rings react to bending.151 for chrysene and for its moiety in complexes with
TCNQ. A similar picture is found for phenazine and
its EDA complexes.109
K. Experimental Arguments for the Hiberty−Shaik It is immediately apparent that the aromaticity of
Hypothesis of the Role of σ-Electron Structure in the central ring increases markedly for chrysene
π-Electron Systems involed in the EDA complexes. Moreover, the same
is also observed for the terminal rings. Consequently,
Analysis of the changes in aromaticity of chry- higher aromaticity is observed for the whole mol-
sene152 and phenazine109,153,154 gave interesting argu- ecule. Looking at the EN and GEO terms for chrysene
ments for the Hiberty and Shaik hypothesis,14 which and its moiety in the EDA complex with TCNQ, it
claims that14,155 “the benzene picture that emerges appears that the EN term for the central ring is
from (these) all-electron studies, the π-component practically the same in both cases (0.119 and 0.121)
enjoys resonance stabilization, but is also distortive whereas the GEO term decreases from 0.378 to 0.257.
and prefers a D3h symmetry, and it is the resistance TCNQ is a well-known electron-accepting system;158
of the σ-frame to the distortion that restores the D6h thus, in the EDA complexes it strips electrons from
structure”. A good illustration of these ideas is given the chrysene moiety, enabling the σ-electrons to
by a scheme of the dependence of the σ- and π-ener- partially equalize the length of all CC bonds.
gies on the resonance coordinate for benzene156
presented in Figure 18.
In this case, the σ-structure dominates in the L. Substituent Effects on Aromaticity
balance between trends of σ- and π-electron struc-
tures. Following this hypothesis, if the amount of This section is devoted to some special kind of
π-electrons in a weakly aromatic system is decreased, interactions, due to substituents but characteristic
Chart 7. HOMA, EN, and GEO Terms Estimated from the Experimental Geometry for Naphthalene,
Phenanthrene, and Triphenylene
of the π-electron systems and not related to typical lene) or by two benzenes in an angular manner
substituent effects, usually associated with Ham- (phenanthrene) leads to an increase in bond alterna-
mett-type plots or similar problems. The latter kind tion. However, in the case of smaller rings, the effect
of substituent effects on aromaticity are presented is usually much more conspicuous.
and discussed throughout the whole review. While looking at benzene rings which are annel-
eated to small rings, an important fact should be
1. Bond Localization Due to the Anneleation of Small taken into account. The bond lengths calculated as
Rings to Benzene the shortest distance between two adjacent atoms
may not be a good picture of the bonding. In such
The bond length alternation in benzene anneleated cases the bonds are bent and an adequate description
to small rings has long been observed and originally is a bond path.75 Figure 19 presents the problem
attributed to the so-called Mills-Nixon effect.159
However, recent studies oppose using the Mills-
Nixon effect for these kinds of phenomena.160-163
Stanger suggested the name strain-induced bond
localization (SIBL)161 and mentioned three approaches
which can be helpful in understanding the observed
phenomena: (a) a classical π-electron effect which
relies on the π-electron count in the small anneleated
ring;164,165 (b) localization due to the rehybridization
caused by the strained atomssthe bonds with the
wider bond angles would have a lower π-character
and following the Bent-Walsh rule should form
shorter bonds to the adjacent atom;166 finally, (c) a
combination of σ- and π-electron effects.167,168 Un-
doubtedly increasing the bond localization means Figure 19. X-X electron densities for 1,2,3,4,5,6,-hexa-
that the system becomes less aromatic, and Chart 7 hydrotricyclobuta[a,c,e]benzene.183 (Reprinted with permis-
presents a few typical examples. It is apparent that sion from ref 183. Copyright 1994 Wiley-VCH Verlag
anneleating benzene by another benzene (naphtha- GmbH.)
clearly showing that the maximum charge density From an energetic point of view, the observed effect
is located outside of the shortest distance between is very small. The differences in total energy for cis-
the atoms nuclei. Thus, for these kinds of molecules, and trans-p-dinitrosobenzene and p-dimethoxyben-
the resulting geometry-based aromaticity indices zene computed at the B3LYP/6-311G** level177 are
should be considered with caution. -0.14 and -0.17 kcal/mol, respectively. In both cases
the cis conformers show a marked AGIBA effect with
2. Angular Group-Induced Bond Alternation (AGIBA) a difference in ipso-ortho bond lengths of 0.0089 and
0.0102 Å, respectively, whereas the trans conformers
Very recently it has been shown that open substit-
exhibit only 0.0022 and 0.0036 Å, respectively. The
uents can also induce bond alternation in aromatic
rings, provided the attached groups are angular as total energy decomposition scheme of Bader,75 ap-
for instance in the nitroso or methoxy group.169-173 plied to determine separately the energies of sub-
The effect has also been observed experimentally for stituents and rings, led to an increase of this differ-
ring systems with equalized bond lengths such as ence for p-dinitrosobenzene up to 0.76 kcal/mol, when
boraxine174 and borazine,175 which despite the bond only the ring is taken into account. Figure 21
length equalization are not considered to be aro- presents the HOMA, EN, and GEO values for all four
matic.176 cases calculated from the optimized geometry (HF/
6-311G**).178 Additionally, the weights of two canoni-
The observed phenomenon is schematically pre-
cal structures are given (calculated by use of the
sented in Figure 20: the double-bonded groups
HOSE model77,78). For a benzene ring (without
AGIBA effect), the contributions are equal to each
other (50%). In the case with angularly-substituted
benzenes (AGIBA substituents), one of the structures
dominates as shown in the figures.
An interesting point in AGIBA-affected systems is
Figure 20. Scheme of the AGIBA effect. (Reprinted with that when the angular group is bent toward the ring,
permission from ref 173. Copyright 1997 Elsevier Science.) the effect is enhanced.169,170,173 Since the enhancement
of the AGIBA effect is nearly additive,173 these effects
-XdY cause a lengthening of the cis bond and a are nicely observed for symmetrically trisubstituted
shortening of the trans one, whereas the -X-Y benzenes.171 Table 26 shows how canonical structure
groups act in the opposite way, they elongate the weights and aromaticity depend on the bending of
trans bonds and shorten the cis ones. For the type of the substituent toward the ring. Even if the changes
substituent, the effect is named as angular group- in HOMA values are not dramatic, the ratio of the
induced bond alternation (AGIBA). Kekulé structure contributions indicates substantial
Figure 21. Aromaticity and canonical structure weights of the phenyl ring for two conformers of p-dimethoxy- and
p-dinitrosobenzene. (Reprinted with permission from ref 47. Copyright 1998 Elsevier Science.)
Table 26. Aromaticity Indices and the Kekulé Structure Weights (Figure 22) of 1,3,5-Tridiazobenzene with
Varying CNN Angle Based on HF/6-31G*
CNN angle(χ) 128° 124° 120° 116° 112°
HOMA H ) 0.994 H ) 0.971 H ) 0.929 H ) 0.866 H ) 0.779
E ) 0.000 E ) 0.000 E ) -0.001 E ) 0.000 E ) 0.001
G ) 0.006 G ) 0.029 G ) 0.072 G ) 0.134 G ) 0.220
canonical stucture K1 57.63% 66.53% 74.72% 81.95% 88.21%
canonical stucture K2 42.37% 33.47% 25.28% 18.05% 11.79%
a The Kekulé structure of the ring K1 is shown in Figure 22. Structure K2 depicts the opposite one.
Chart 8. Aromaticity Indices for Anthracene and Its 1-, 2-, and 9-CH2+ Derivatives
changes in electron structures. The changes in aro- thracene derivatives as well as the values of HOMA,
maticity are solely due to the increase in bond length EN, and GEO.
alternation (GEO term). 1-CH2+ and 2-CH2+ anthracenes differ from each
other dramatically. The substituted ring in the
1-CH2+ derivative has a less aromatic character
(HOMA ) 0.348) than in the parent molecule (0.519),
but the ring in the 2-CH2+ derivative is much more
dearomatized (HOMA ) 0.179). Two other rings in
the case of 2- CH2+ derivatives are much less aro-
matic (HOMA ) 0.668 and 0.652) than in the case of
the 1-CH2+ derivative (HOMA ) 0.767 and 0.715).
Figure 22. Scheme of one of the Kekulé structures (K1) The CC bond, via which the quinoidal structure may
of 1,3,5-triazobenzene. propagate over the whole molecule, is equal to 1.401
Å in the case of the 1-CH2+ derivative as compared
3. Charged Substituents in Benzenoid Systems with the value of 1.375 Å in the case of the 2-CH2+
derivative (both bonds marked in bold lines in Chart
Systematic studies on CH2+-substituted benzenoid 8).
hydrocarbons179 carried out by use of ab initio HF/ These local characteristics of aromaticity are also
6-31G*-optimized geometries revealed a substantial well seen in the HOMA values for the whole mol-
influence of the position of substitution on the ecule. 1-Derivatives and 9-derivatives have HOMA
aromatic character of the substituted ring and other values nearly the same as an unsubstituted system,
rings of the system in question. If the CH2+ group is only the 2-derivative is strongly dearomatized. This
attached to the ring in the benzenoid hydrocarbon
is due to a possibility of a long-range quinoidal
in such a way that the induction of the quinoidal
structure (Figure 23) which cannot be realized in
structure onto the further fragments of the mol-
other derivatives.
ecule proceeds via a short CC bond, then the in-
duced quinoidal structure has strong π-bond local- The least effect of CH2+ on the adjacent rings is
ization. If this is done via a long CC bond, then the observed in the case of 9-derivatives. The substituted
localization is much weaker. Consequently, in the ring is of low aromaticity (HOMA ) 0.174), but the
first case the decrease in aromaticity is greater other two are markedly more aromatic, even more
whereas in the second case weaker. The case of so than in the parent compound (HOMA ) 0.892).
anthracene is a good illustration. Chart 8 presents This observation is in line with the well-known fact
anthracene and 1-, 2-, and 9-CH2+-substituted an- that 9-hydroxyanthracene easily tautomerizes to its
Figure 23. Selection of quinoidal canonical structures showing the long-range transmission of the CH2+ substituent effect.
ranging from strongly antiaromatic to typically aro-

matic systems.95
V. Acronyms
HOMA harmonic oscillator model of aromaticity
AGIBA angular group-induced bond alternation
Figure 24. Tautomerism of 9-hydroxyanthracene.
NMR nuclear magnetic resonance
SE stabilization energy
keto form,97 leaving two other rings highly aromatic, ASE aromatic stabilization energy
Figure 24. NICS nucleus-independent chemical shift
Depending on the position of substitution, the CC CSD Cambridge Structural Database
bond between CH2+ and the benzenoid hydrocarbon ED (gas-phase) electron diffraction
varies strongly in length. The changes are related to MW microwave (measurement of molecular
the total atomic charge at the CH2+ group, Q(CH2+), geometry)
AJ Julg’s index of aromaticity
estimated by Mulliken population analysis,180 which BAC bond alternation coefficient
in turn depends on the Hammett-Streitwieser posi- I 5 , I6 Bird’s aromaticity indices of five- and six-
tion constants.181 membered rings (with heteroatoms)
The analysis of HOMA, EN, and GEO terms made EN term measuring the decrease of aromaticity
for individual rings and the whole moieties of CH2+- estimated by HOMA due to bond elongation
substituted naphthalene, anthracene, phenanthrene, GEO term measuring the decrease of aromaticity
and pyrene molecules allowed us to draw the follow- estimated by HOMA due to bond alternation
ing conclusions. If a single substituent able to form REC ring energy content
MEC molecule energy content
a double bond is attached to the benzenoid hydro-
HOSE harmonic oscillator stabilization energy
carbon in a position which permits the formation of DE (Hückel’s) delocalization energy
the quinoidal structure along a larger part of the TCNQ tetracyanoquinodimethane
π-electron system, then it acts as a dearomatizing EDA electron donor-acceptor (complexes)
factor for this fragment and consequently may do this HMO Hückel molecular orbital (theory)
for the whole system. In all cases, a decrease in REPE resonance energy per electron (π)
aromaticity is associated with an increase in the GAV (Cohen and Benson’s) group additivity values
geometric term. The magnitude of the π-electron PBD polysubstituted benzene derivatives
localization via the formation of the quinoidal struc- DFT density functional theory
tures depends on the accepting power of the sub- SIBL strain-induced bond localization
stituent, which leads to the changes in bond lengths,
which causes this induction. The shorter this bond, VI. Acknowledgments
the greater localization is observed. Very inspiring discussions with Professor Paul von
Rague Schleyer (Erlangen, Germany, and Athens,
IV. Conclusions United States), Dr. Siân Howard (Cardiff, U.K.) and
Professor Günter Häfelinger (Tübingen, Germany)
The molecular geometry of cyclic π-electron sys- are gratefully acknowledged. We also gratefully
tems is an important and readily accessible source acknowledge Gesellschaft Deutscher Chemiker
of information from which corresponding aromaticity (GDCh) for supplying the photo of Friedrich August
indices may be easily obtained. Unlike other indices, Kekulé von Stradonitz on the cover and for the
the geometry-based aromaticity indices are applicable permission of using it for the purpose of this article.
to both local and the global π-electron systems. They
can even be applied to noncyclic systems, and their
values then have an interpretation as measures of VII. References
π-electron delocalization. (1) Krygowski, T. M.; Cyrański, M. K.; Czarnocki, Z.; Häfelinger,
G.; Katritzky, A. R. Tetrahedron 2000, 56, 1783 and references
Molecules with too much strained rings (containing therein.
three- or four-member rings) should not be treated (2) Institute for Scientific Information, Philadelphia (1981-1999).
by geometry-based aromaticity indices, since their Retrieved in February 2000.
(3) Kekulé, A. Bull. Soc. Chim. Fr. 1865, 3, 98.
bonds are not straight links between the atoms but (4) The Kekulé Riddle. A Challenge for Chemists and Psychologists;
rather banana-shaped “bond paths” of which the Wotiz, J. H., Ed.; Clearwater, FL, Vienna, 1993.
lengths are not easily accessible and which differ in (5) Erlenmeyer, E. Ann. Chem. Pharm. 1866, 137, 327.
(6) Elvidge, J. A.; Jackman, L. M. J. Chem. Soc. 1961, 859.
length from the direct links (cf. Figure 19). Sondheimer, F. Pure Appl. Chem. 1964, 859. Dewar, M. J. S.
Among geometry-based indices, the HOMA index Tetrahedron Suppl. 1966, 8, 75.
(7) Schleyer, P. v. R.; Jiao H. Pure Appl. Chem. 1996, 68, 209.
works most effectively; not only does it measure the (8) Krygowski, T. M.; Cyrański, M. Aromatic Character of Carbo-
decrease in aromaticity due to an increase in bond cyclic π-Electron Systems Deduced from Molecular Geometry.
Advances in Mocular Structure Research; Hargittai, I. M., Ed.;
alternation (the GEO term), but it also estimates the JAI Press: London, 1997; Vol. 3, p 227.
decrease due to bond elongation (the EN term). (9) Aromaticity, Pseudoaromaticity and Antiaromaticity, Proceed-
ings of an International Symposium held in Jerusalem 1970,
In most cases, HOMA correlates well with mag- Bergmann, E. D., Pullman, B., Eds.; Israel Academy of Sciences
netic indices, such as NICS and the magnetic sus- and Humanities: Jerusalem, 1971.
ceptibility exaltation, and this is well observed in a (10) Bergmann, E. D. Aromaticity, Pseudoaromaticity and Antiaro-
maticity, Proceedings of an International Symposium held in
large sample of five-membered hetero-π-electron Jerusalem 1970, Bergmann, E. D., Pullman, B., Eds.; Israel
systems, which consists more than 100 systems Academy of Sciences and Humanities: Jerusalem, 1971; p 386.
(11) Minkin, V. I.; Glukhovtsev, M. N.; Simkin, B. Ya. Aromaticity (43) Hoffmann, R., Foreword. Determination of the Geometrical
and Antiaromaticity. Electronic and Structural Aspects; J. Wiley: Structure of Free Molecules; Vilkov, L. V., Mastryukov, V. S.,
New York, 1994. Sadova, N. L., Eds.; Mir Publishers: Moscow, 1983.
(12) Armit, J. W.; Robinson, R. J. Chem. Soc. 1925, 127, 1604. (44) Diffraction studies on non-crystalline substances; Hargittai, I.,
(13) (a) Kruszewski, J.; Krygowski, T. M. Tetrahedron Lett. 1970, Orville-Thomas, W. J., Eds.; Akademiai Kiado: Budapest, 1981.
319. (b) Krygowski, T. M. Tetrahedron Lett. 1970, 1311. (c) (45) International Union of Crystallography. Accurate Molecular
Dixon, W. T. J. Chem. Soc. 1970, 612. Structures, Their Determination and Importance; Domenicano,
(14) (a) Longuet-Higgins, H. C.; Salem, L. Proc. R. Soc. 1959, A251, A., Hargittai, I., Eds.; Oxford University Press: Oxford, 1992.
172. (b) Berry, R. S. J. Chem. Phys. 1961, 35, 2253. (c) Shaik, S. (46) Allen, F. H.; Davies, J. E.; Galloy, J. J.; Johnson, O.; Kennard,
S.; Hiberty, P. C. J. Am. Chem. Soc. 1985, 107, 3089. (d) Hiberty, O.; McRae, E. M.; Mitchell, G. F.; Smith, J. M.; Watson, D. G.
P. C.; Shaik, S. S.; Ohanessian, G.; Lefour, J. M. J. Org. Chem. J. Chem. Inf. Comput. Sci. 1991, 31, 187.
1986, 51, 3908. (e) Hiberty, P. C. Top. Curr. Chem. 1990, 153, (47) Krygowski, T. M.; Cyrański, M. K. Molecular Geometry as a
27. Source of Chemical Information for π-Electron Compounds.
(15) (a) Glendening, E. D.; Faust, R.; Streitwieser, A.; Vollhardt, K. Theoretical Organic Chemistry; Párkányi, C., Ed.; Elselvier:
P. C.; Weinhold, F. J. Am. Chem. Soc. 1993, 115, 10952. (b) Amsterdam, 1998; Vol. 5, p 153.
Ichikawa, H.; Kagawa, H. J. Phys. Chem. 1995, 99, 2307. (48) March, J. Advanced Organic Chemistry, Reaction, Mechanism
(16) (a) Pross, A. Physical and Theoretical Principles of Organic and Structure, 4th ed.; J. Wiley: New York, 1993.
Reactivity; J. Wiley: New York, 1996. (b) Coulson, C. A. Valence; (49) (a) Minot, C. J. Phys. Chem. 1987, 91, 6380. (b) Berresheim, A.
Oxford University Press: Oxford, 1961. J.; Muller, M.; Müllen, K. Chem. Rev. 1999, 99, 1747. (c)
(17) (a) Charton, M. Prog. Phys. Org. Chem. 1981, 13, 119. (b) Schleyer, P. v. R., Private communication, 1998.
Krygowski, T. M.; Kalinowski, M. K.; Turowska-Tyrk, I.; Hiberty, (50) Jeffrey, G. A.; Ruble, J. R.; McMullan, R. K.; Pople, J. A. Proc.
P. C.; Milart, P.; Silvestro, A.; Topsom, R. D.; Daehne, S. Struct. R. Soc. London 1987, A414, 47.
Chem. 1991, 2, 71. (51) Trucano, P.; Chen, R. Nature 1975, 258, 136.
(18) Mueller, E. Neuere Anschauungen der Organischen Chemie; (52) Kveseth, K.; Seip, R.; Kohl, D. A. Acta Chem. Scand. 1980, A34,
Springer-Verlag: Göttingen, 1957. 31.
(19) Pauling, L.; Sherman, J. J. Chem. Phys. 1933, 1, 606. (53) Fleischer, U.; Kutzelnigg, W.; Lazzeretti, P.; Muhlenkamp, V.
(20) Kistiakowski, B.; Ruhoffm J. R.; Smith, H. A.; Vaughan, W. E. J. Am. Chem. Soc. 1994, 116, 5298.
J. Am. Chem. Soc. 1936, 58, 146. (54) Alkorta, I.; Elguero, J. New J. Chem. 1999, 23, 951.
(21) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.; (55) Julg, A.; François, Ph. Theor. Chim. Acta 1967, 7, 249.
Johnson, B. G.; Robb, M. A.; Cheesman, J. R.; Keith, T.; Peterson, (56) Schleyer, P. v. R.; Freeman, P. K.; Jiao, H.; Goldfuss, B. Angew.
G. A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Chem. Int. Ed. 1995, 34, 337.
Zakrzewski, V. G.; Ortiz, J. V.; Foresman, J. B.; Cioslowski, J.; (57) Julg, A., ref 9 p 383.
Stefanov, B. B.; Nanayakkara, A.; Challacombe, M.; Peng, C. (58) Kruszewski, J.; Krygowski, T. M. Tetrahedron Lett. 1972, 3839.
Y.; Ayala, P. Y.; Chen, W.; Wong, M. W.; Andres, J. L.; Replogle, (59) Bird, C. W. Tetrahedron 1985, 41, 1409.
E. S.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Binkley, J. S.; (60) Gordy, W. J. J. Chem. Phys. 1947, 14, 305.
Defrees, D. J.; Baker, J.; Stewart, J. P.; Head-Gordon, M.; (61) (a) Bird, C. W. Tetrahedron 1986, 42, 89. (b) Bird, C. W.
Gonzales, C.; Pople, J. A. Gaussian 94; Gaussian Inc.: Pitts- Tetrahedron 1987, 43, 4725. (c) Bird, C. W. Tetrahedron 1990,
burgh, PA, 1995. 46, 5697. (d) Bird, C. W. Tetrahedron 1992, 48, 1675. (e) Bird,
(22) (a) Hehre, W. J.; Ditchfield, D.; Radom L.; Pople, J. A. J. Am. C. W. Tetrahedron 1992, 48, 7862. (f) Bird C. W. Tetrahedron
Chem. Soc. 1970, 92, 4796. (b) George, P.; Trachtman, M.; Bock, 1993, 49, 8441.
C. W.; Brett A. M. J. Chem. Soc., Perkin Trans. 2 1976, 1222. (62) (a) Pozharski, A. F. Khim. Geterotsikl. Soed. 1985, 867. (b)
(23) (a) Hehre, W. J.; Taylor R. T.; Pople, J. A.; Schleyer, P. v. R. J. Pozharski, A. F. Theoretical Principles of Heterocyclic Chemistry;
Am. Chem. Soc. 1974, 96, 7162. (b) Radom, L. J. Chem. Soc., “Khimia” Publishing: Moscow, 1985.
Chem. Commun. 1974, 403. (c) George, P.; Trachtman, M.; Brett,
(63) Nyulaszi, L.; Varnai, P.; Veszpremi T. J. Mol. Struct. (Theochem)
A. M.; Bock, C. W. J. Chem. Soc., Perkin Trans. 2 1977, 1036.
1995, 358, 55.
(24) (a) Trinajstić, N. Chemical Graph Theory; CRC Press: Boca
(64) Dransfeld, A.; Nyulaszi, L.; Schleyer, P. v. R. Inorg. Chem. 1998,
Raton, FL, 1983. (b) Graovać, A.; Gutman, I.; Trinajstić, N.
37, 4413.
Lecture Notes in Chemistry; Springer-Verlag: Berlin, 1977. (c)
Gutman, I.; Polansky, O. E. Mathematical Concepts in Organic (65) Krygowski, T. M.; Ciesielski, A.; Cyrański, M. Chem. Pap. 1995,
Chemistry; Springer-Verlag: Berlin, 1986. (d) Cyvin, S. J.; 49, 128.
Brunvoll, J.; Cyvin, B. N. Theory of Coronoid Hydrocarbons; (66) Krygowski, T. M. J. Chem. Inf. Comput. Sci. 1993, 33, 70.
Springer-Verlag: Berlin, 1991. (67) Krygowski, T. M.; Cyrański, M. Tetrahedron 1996, 52, 10255.
(25) (a) Flygare, W. H. Chem. Rev. 1974, 74, 653. (b) Bothner-By, A. (68) Krygowski, T. M.; Cyrański, M. Tetrahedron 1996, 52, 1713.
A.; Pople, J. A. Rev. Phys. Chem. 1965, 16, 43. (69) (a) Pine, S.; Hendrickson, J. B.; Cram, D. J.; Hammond, G. S.
(26) (a) London, F. J. Phys. Radium 1937, 8, 397. (b) Kutzelnigg, W. Organic Chemistry, 4th ed.; McGraw-Hill: New York, 1980. (b)
Isr. J. Chem. 1980, 19, 193. (c) Woliński, K.; Hinton, J. F.; Pulay, Allinger, N. L.; Cava, M. P.; de Jong, D. C.; Johnson, C. E. R.;
P. J. Am. Chem. Soc. 1990, 112, 8251. Lebel, N. A.; Stevens, C. L. Organic Chemistry, 2nd ed.; W. de
(27) Dauben, H. J.; Wilson, J. D.; Laity, J. L. J. Am. Chem. Soc. 1968, Gruyter: Berlin, 1980. (c) Vollhardt, K. P. C.; Schore, N. E.
90, 811. Organic Chemistry, 2nd ed.; W. H. Freeman: New York, 1987.
(28) (a) Selwood, P. W. Magnetochemistry; Interscience: New York, (d) Streitwieser, A.; Heathcock, C. H.; Kosower, E. M. Introduc-
1956. (b) Haberdietzl, H. Angew. Chem., Int. Ed. 1966, 5, 288. tion to Organic Chemistry, 4th ed.; Macmillan: New York, 1992.
(29) Maouche, B.; Gayoso, J.; Ouamerali, O. Rev. Roum. Chim. 1984, (e) Fox, M. A.; Whitesell, J. K. Organic Chemistry, 2nd ed.; Jones
29, 613. and Bartlett Publ.: Boston, 1997. (f) Beyer, H.; Walter, W.
(30) Salem, L., The Molecular Orbital Theory of Conjugated Systems; Lehrbuch der Organischen Chemie, 22 ed.; S. Hirzel: Stuttgart
Benjamin:, New York, 1966. 1991. (g) Carey, F. A.; Sandberg, R. J. Advanced Organic
(31) Reference 11 p 72. Chemistry, 3rd ed.; Plenum Press: New York, 1993.
(32) Schleyer, P. v. R.; Maerker, C.; Dransfeld, A.; Jiao, H.; Hommes, (70) Pauling, L. J. Am. Chem. Soc. 1947, 69, 542.
N. J. R. v. E. J. Am. Chem. Soc. 1996, 118, 6317. (71) Krygowski, T. M.; Ciesielski, A.; Bird, C. W.; Kotschy, A. J.
(33) (a) Hückel, E. Z. Phys. 1931, 76, 204. (b) Hückel, E. Z. Chem. Inf. Comput. Sci. 1995, 35, 203.
Elektrochem. 1937, 43, 752. (72) Johnston, H. S.; Parr, Ch. J. Am. Chem. Soc. 1963, 85, 2544.
(34) Berson, J. A., Chemical Creativity. Ideas from the Work of (73) Grimme, S. J. Am. Chem. Soc. 1996, 118, 1529.
Woodward, Hückel, Meerwein, and Others; Wiley-VCH: Wein- (74) Exner, K.; Schleyer, P. v. R. Private communication, 1999.
heim, 1999; Chapter 3. (75) Bader, R. F. W. Atoms in Molecules: a quantum theory; Oxford
(35) Breslow, R.; Yuan C. J. Am. Chem. Soc. 1958, 80, 5991. University Press: Oxford, 1990.
(36) (a) Breslow, R.; Haynie, R.; Mirra, J. J. Am. Chem. Soc. 1959, (76) Randić, M. Chem. Phys. Lett. 1976, 38, 68.
81, 247. (b) Volpin, M. E.; Koreshkov, Y. D.; Kursanov, D. N. (77) Krygowski, T. M.; Anulewicz, R.; Kruszewski, J. Acta Crystallogr.
Izvest. Akad. Nauk SSSR 1959, 560. 1983, B39, 732.
(37) Avram, M.; Dinu, D.; Matescu, G.; Nenitescu, C. D. Ber. 1960, (78) Krygowski, T. M.; Anulewicz, R.; Wisiorowski, M. Pol. J. Chem.
93, 1789. 1995, 69, 1579.
(38) Fritz, H. P.; McOmie, J. F.; Sheppard, W. Tetrahedron Lett. 1960, (79) Hess, B. A.; Schaad, L. J. Am. Chem. Soc. 1971, 93, 305.
26, 35. (80) Hess B. A.; Schaad, L. J. Org. Chem. 1971, 36, 3418.
(39) Kruszewski, J.; Krygowski, T. M. Can. J. Chem. 1975, 53, 946. (81) Bird, C. W. Tetrahedron 1997, 53, 13111.
(40) Gutman, I.; Trinajstic, N. Can. J. Chem. 1976, 54, 1789. (82) (a) Krygowski, T. M. Structure and Reactivity; Liebman, J. F.;
(41) Howard, S. T.; Krygowski, T. M. Can. J. Chem. 1997, 75, 1174. Greenberg, A., Eds.; VCH: New York, 1988; Chapter 6, p 231.
(42) (a) Cyrański, M. K.; Krygowski, T. M.; Wisiorowski, M.; Hommes, (b) Krygowski, T. M. Progres in Physical Organic Chemistry;
N. J. R. v. E.; Schleyer, P. v. R. Angew. Chem. Int. Ed. 1998, 37, Taft, R. W., Ed.; J. Wiley: New York, 1991; p 239.
177. (b) Cyrański, M. K.; Krygowski, T. M.; Wisiorowski, M.; (83) Randić, M. Tetrahedron 1977, 33, 1905.
Hommes, N. J. R. v. E.; Schleyer, P. v. R. Angew. Chem. 1998, (84) Krygowski, T. M.; Maurin, J. J.Chem. Soc., Perkin Trans. 2 1989,
110, 187. 695.
(85) Hiberty, P. C.; Ohanessian, G. J. Am. Chem. Soc. 1984, 106, (121) Subramanian, G.; Schleyer, P. v. R.; Jiao, H. Angew. Chem., Int.
6963. Ed. 1996, 35, 2638.
(86) Turowska-Tyrk, I.; Krygowski, T. M.; Milart, P. J. Mol. Struct. (122) Nendel, M.; Houk, K. N.; Tolbert, L. M.; Vogel, E.; Jiao, H.;
1991, 263, 235. Schleyer, P. v. R. J. Phys Chem. 1998, A102, 7191.
(87) Cyrański, M. K.; Krygowski, T. M. Tetrahedron 1999, 55, 6205. (123) Mo, O.; Yanez, M.; Essefar, M.; Herreros, M.; Notario, R.;
(88) Bürgi, H. B.; Blanc., E.; Schwartzenbach, D.; Liu, S.; Lu. Y.-J.; Abboud, J. L. M. J. Org. Chem. 1997, 62, 3200.
Kappes, M. M.; Ibers, J. A. Angew. Chem., Int. Ed. 1992, 31, (124) Kabuto, C.; Fujimori, K.; Yasunami, M.; Takase, K., Morita, N.;
640. Asao, T. Acta Crystallogr. 1983, C39, 1245.
(89) Bird, C. W. Tetrahedron 1992, 48, 335. (125) Vogel, E.; Schmalstieg, L.; Henk, P.; Wilmes, O.; Lex, J.; Gleiter,
(90) Streitwieser, A. Molecular Orbital Theory for Organic Chemists; R.; Langer, M. Angew. Chem., Int. Ed. 1991, 30, 681.
J. Wiley: New York, 1961. (126) Vogel, E.; Wieland, H.; Schmalstieg, L.; Lex, J. Angew. Chem.,
(91) Cox, E. G. Rev. Mod. Phys. 1955, 30, 159. Int. Ed. Engl. 1984, 23, 717.
(92) Brock, C. P.; Dunitz, J. D. Acta Crystallogr. 1982, B38, 2218. (127) Krygowski, T. M.; Ciesielski, A.; Cyrański, M. J. Mol. Struct.
(93) Häfelinger, G.; Regelmann, K.; Krygowski, T. M.; Woźniak, K. 1996, 374, 277.
J. Comput. Chem. 1989, 10, 329. (128) Czermiński, J.; Iwasiewicz, A.; Paszek, Z.; Sikorski, A. Statistical
(94) Pozharskii, A. F.; Soldatenkov, A. T.; Katritzky, A. R. Hetero- Methods in Applied Chemistry; PWN-Polish Scientific Publish-
cycles in Life and Society; J. Wiley: New York, 1997. ers: Warsaw, Poland; Elsevier: Amsterdam, 1990.
(95) (a) Cyrański, M. K.; Krygowski, T. M.; Katritzky, A. R.; Schleyer, (129) Calderon, J. L.; Cotton, F. A.; Legzdins, P. J. Am. Chem. Soc.
P. v. R. Manuscript in preparation. (b) Cyrański, M. K.; Kry- 1969, 91, 2528.
gowski, T. M.; Katritzky, A. R.; Schleyer, P. v. R. Chem. Listy (130) Cook, M. J.; Katritzky, A. R.; Linda, P.; Tack, R. D. J. Chem.
2000, 9, 788. Soc., Perkin Trans. 2 1972, 1295.
(96) (a) Exner, O. The Hammett Equation in Advances in Linear Free (131) Katritzky, A. R.; Karelson, M.; Slid, S.; Krygowski, T. M.; Jug,
Energy Relationships; Chapman, N. B., Shorter, J., Eds.; Plenum K. J. Org. Chem. 1998, 63, 5228.
Press: London. 1972, p 1. (b) Charton, M. The Upsilon Steric (132) Krygowski, T. M.; Kruszewski, J. Bull. Acad. Sci. Pol. 1974, 22,
Parameter-Definition and Determination. Top. Curr. Chem. 871.
1987, 114, 58. (133) Gimarc, B. M. J. Am. Chem. Soc. 1983, 105, 1979.
(97) Clar, E. Polycyclic Hydrocarbons; Academic Press: London, 1964; (134) Cyrański, M. K.; Krygowski, T. M.; Krutosı́ková, A.; Sleziak, R.
Vols. 1 and 2. Manuscript submitted.
(98) Harvey, R. G. Polycyclic Aromatic Hydrocarbons; Wiley-VCH: (135) (a) Mrozek, A.; Karolak-Wojciechowska, J.; Amiel, P.; Barbe, J.
New York, 1997. J. Mol. Struct. 2000, 524, 151. (b) Mrozek, A.; Karolak-
(99) Dewar, M. J. S.; de Llano, C. J. Am. Chem. Soc. 1969, 91, 789. Wojciechowska, J.; Amiel, P.; Barbe, J. J. Mol. Struct. 2000, 524,
(100) Hess, B. A.; Schaad, L. J. Am. Chem. Soc. 1971, 93, 2413. 159.
(101) (a) Hess, B. A.; Schaad, L.; Holyoke, C. W. Tetrahedron 1972, (136) Catalan, J.; del Valle, J. C.; Claramunt, R. M.; Boyer, G.; Laynez,
28, 3657. (b) Hess, B. A.; Schaad, L.; Holyoke, C. W. Tetrahedron J.; Gomez, J.; Jimenez, P.; Tomas, F.; Elguero, J. J. Phys. Chem.
1972, 28, 5299. (c) Hess, B. A. Schaad, L. J. Am. Chem. Soc. 1994, 98, 10 606.
1973, 95, 3907. (137) MP2/6-31G** optimized geometry: Catalan, J.; de Paz, J. L. G.;
Elguero, J. J. Chem. Soc., Perkin Trans. 2 1996, 57.
(102) Definition of precise geometry follows CSD requirements, see
(138) Krygowski, T. M.; Anulewicz, R.; Cyrański, M. K.; Puchała, A.;
ref 46.
Rasała, D. Tetrahedron 1998, 54, 12295.
(103) Clar, E. The Aromatic Sextet; J. Wiley & Sons: Chichester, 1972.
(139) e.g. Vogel, E. J. Heterocycl. Chem. 1996, 33, 1461.
(104) (a) Cohen, N.; Benson, S. W. Chem. Rev. 1993, 93, 2419. (b) (140) Krygowski, T. M.; Kalinowski, M. K.; Turowska-Tyrk, I.; Hiberty,
Cyrański, M. K.; Stȩpień, B. T.; Krygowski, T. M. Tetrahedron P. C.; Milart, P.; Silvestro, A.; Topsom, R. D.; Daehne, S. Struct.
2000, 56, 9663. Chem. 1991, 2, 71.
(105) Cyrański, M.; Krygowski, T. M. J. Chem. Inf. Comput. Sci. 1996, (141) Flandrois, S.; Chasseau, D. Acta Crystallogr. 1977, B33, 2744.
36, 1142. (142) Woźniak, K.; Krygowski, T. M. J. Mol. Struct. 1989, 193, 81.
(106) Pauling, L. The Nature of the Chemical Bond and Structure of (143) Kitaygorodsky, A. I. Molecular Crystals and Molecules; Academic
Molecules in Crystals, 3rd ed.; Cornell University Press: Ithaca, Press: New York, 1973.
1960. (144) Lipkowitz, K. B.; Peterson, M. A. J. Comput. Chem. 1993, 14,
(107) Sanderson, J. J. Am. Chem. Soc. 1983, 105, 2259. 121.
(108) Howard, S. T.; Woźniak, K. Chem. Phys. Lett. 1993, 212, 1. (145) Krygowski, T. M.; Turowska-Tyrk, I. Chem. Phys. Lett. 1987,
(109) Cyrański, M.; Krygowski, T. M. Tetrahedron 1996, 52, 13795. 138, 90.
(110) Karolak-Wojciechowska, J. Acta Crystallogr. 2000, submitted for (146) (a) Jenneskens, L. W.; Klamer, J. C.; de Boer, H. J. R.; de Wolf,
publication. W. H.; Bickelhaupt, F.; Stam, C. H. Angew. Chem. 1984, 96, 236.
(111) Hayon, E.; Simic M. J. Am. Chem. Soc. 1970, 92, 7487. (b) Jenneskens, L. W.; Klamer J. C.; de Boer, H. J. R.; de Wolf,
(112) Krygowski, T. M.; Anulewicz, R.; Pniewska, B.; Milart, P. J. Phys. W. H.; Bickelhaupt, F.; Stam, C. H. Angew. Chem., Int. Ed. 1984,
Org. Chem. 1991, 4, 121. 23, 238.
(113) Anulewicz, R.; Ba̧k, T.; Cyrański, M.; Krygowski, T. M.; Pawlak, (147) Bickelhaupt, F. Pure Appl. Chem. 1990, 62, 373.
D.; Pniewska, B.; Rasała, D.; Gawinecki R. Acta Chem. Scand. (148) Rice, J. E.; Lee, T. J.; Remington, R. B.; Allen, W. D.; Clabo, D.
1995, 49, 515. A.; Schaefer, H. F. J. Am. Chem. Soc. 1987, 109, 2902.
(114) Kataoka, M.; Sato T. J. Heterocycl. Chem. 1999, 36, 1323. (149) Dijkstra, F.; van Lenthe, J. H. Int. J. Quant. Chem. 1999, 74,
(115) Vogel, E.; Puttmann, W.; Duchatsch, W.; Schieb, T.; Schmickler, 213.
H. Angew. Chem., Int. Ed. 1986, 25, 720. (150) Bodwell, G. J.; Houghton, T. J.; Koury, H. E.; Yarlagadda, B.
(116) Vogel, E.; Schieb, T.; Schulz, W. H.; Schmidt, K.; Schmickler, Synlett 1995, 751.
H.; Lex, J. Angew. Chem., Int. Ed. 1986, 25, 723. (151) Manion, M.; Bodwell, G. J.; Cyrański, M. K.; Krygowski, T. M.
(117) Ginzburg, D. (Ed.), Non-Benzenoid Aromatic Compounds; In- Manuscript in preparation.
terscience: New York, 1959. (152) Krygowski, T. M.; Ciesielski, A.; SÄ wirska, B.; Leszczyński, P.
(118) Thomas, R.; Coppens, P. Acta Crystallogr. 1972, B28, 1800. Pol. J. Chem. 1994, 68, 2097.
(119) Cyrański, M. K. Analysis of the Aromatic Character of π-Electron (153) Woźniak, K.; Kariuki, B.; Jones, W. Acta Crystallogr. 1991, C47,
Systems by Separating Geometric and Energetic Contributions. 1113.
Ph.D. Thesis (in Polish), University of Warsaw, Warsaw, 1998. (154) Goldberg, I.; Shmueli, U. Acta Crystallogr. 1973, B29, 440.
(120) (a) Howard, J. A. K.; Knox, S. A. R.; Riera, V.; Stone, F. G. A.; (155) Hiberty, P. C.; Danovich, D.; Shurki, A.; Shaik, S. J. Am. Chem.
Woodward, P. Chem. Commun. 1974, 452. (b) Howard, J. A. K.; Soc. 1995, 117, 7760.
Stansfield, R. F. D.; Woodward, P. J. Chem. Soc., Dalton Trans. (156) Jug, K.; Köster A. J. Am. Chem. Soc. 1990, 112, 6772.
1979, 1812. (c) Howard, J. A. K.; Woodward, P. J. Chem. Soc., (157) Munnoch, P. J.; Wright, J. D. J. Chem. Soc., Perkin Trans. 2
Dalton Trans. 1978, 412. (d) Stezowski, J. J.; Hoier, H.; Wilhelm, 1974, 1397.
D.; Clark, T.; Schleyer, P. v. R. Chem. Commun. 1985, 1263. (e) (158) Brigleb, G. Elektronen-Donator-Acceptor-Komplexe; Springer-
Bunel, E. E.; Valle, L.; Jones, N. L.; Carroll, P. J.; Barra, C.; Verlag: Berlin, 1961.
Gonzalez, M.; Munoz, N.; Aizman, A.; Manriquez, J. M. J. Am. (159) (a) Mills, W. H.; Nixon, I. G. J. Chem. Soc. 1930, 2510. (b) See
Chem. Soc. 1988, 110, 6596. (f) Manriquez, J. M.; Ward, M. D.; also Maksić, Z. B.; Eckert-Maksić, M.; Mó, O.; Yáñez, M. The
Reiff, W. M.; Calabrese, J. C.; Jones, N. L.; Carroll, P. J.; Bunel, Mills-Nixon Effect: Fallacies, Facts and Chemical Relevance.
E. E.; Miller, J. S. J. Am. Chem. Soc. 1995, 117, 6182. (g) Theoretical and Computational Chemistry; Politzer, P., Maksić,
Alvarez-Larena, A.; Brianso, J. L.; Piniella, J. F.; Farran, J.; Z. B., Eds.; Elsevier: Amsterdam, 1999; Vol. 6, p 47.
Manriquez, J. M.; Chavez, I.; Oelckers, B., Molins, E.; Mirav- (160) Stanger, A. J. Am. Chem. Soc. 1991, 113, 8277.
itlles, C. Acta Crystallogr. 1996, C52, 2754. (h) Miravitlles, C.; (161) Stanger, A. J. Am. Chem. Soc. 1998, 120, 12034.
Molins, E.; Maniukiewicz, W.; Mas, M.; Manriquez, J. M.; (162) Siegel, J. S. Angew. Chem., Int. Ed. 1994, 33, 1721.
Chavez, I.; Oelckers, B.; Alvarez-Larena, A.; Brianso, J. L. Acta (163) Frank, N. L.; Siegel, J. S. Advances in Theoretically Interesting
Crystallogr. 1996, C52, 3047. (i) Abbasali, Q. A.; Cloke, F. G. Molecules; JAI Press Inc., Greenwich, CT, 1995; Vol. 3 p 209.
N.; Hitchcock, P. B.; Joseph, S. C. P. Chem. Commun. 1997, (164) Diercks, R.; Vollhardt, K. P. C. J. Am. Chem. Soc. 1986, 108,
1541. 3150.
(165) Faust, R.; Glendening, E. D.; Streitwieser, A.; Vollhardt, K. P. Schleyer, P. v. R.; Jiao, H.; Hommes, N. J. R. v. E.; Malkin, V.
C. J. Am. Chem. Soc. 1992, 114, 8263. G.; Malkina, O. L. J. Am. Chem. Soc. 1997, 119, 12669.
(166) Bent, H. A. Chem. Rev. 1961, 61, 275. (177) Howard S. T.; Krygowski, T. M.; Ciesielski, A.; Wisiorowski, M.
(167) Shurki, A.; Shaik, S. S. Angew. Chem., Int. Ed. 1997, 36, 2205. Tetrahedron 1998, 54, 3553.
(168) Shaik, S. S.; Shurki, A.; Danovich, D.; Hiberty P. C. J. Mol. (178) Krygowski, T. M.; Cyrański, M.; Wisiorowski, M. Pol. J. Chem.
Struct. 1997, 398, 155. 1996, 70, 1351.
(169) Krygowski, T. M.; Anulewicz, R.; Jarmuła, A.; B_k, T.; Rasała,
(179) (a) Krygowski, T. M.; Cyrański, M. K.; Nakata, K.; Fujio, M.;
D.; Howard, S. T. Tetrahedron 1994, 50, 13155.
Tsuno, Y. Tetrahedron 1997, 53, 11383. (b) Krygowski, T. M.;
(170) Krygowski, T. M.; Anulewicz, R.; Hiberty, P. C. J. Org. Chem.
Cyrański, M. K.; Nakata, K.; Fujio, M.; Tsuno, Y. Tetrahedron
1996, 61, 8533.
1998, 54, 3303.
(171) Howard, S. T.; Krygowski, T. M.; Glówka, M. L. Tetrahedron
1996, 52, 11379. (180) Mulliken, R. S. J. Chem. Phys. 1955, 23, 1833.
(172) Krygowski, T. M.; Howard, S. T.; Martynowski, D.; Główka, M. (181) Krygowski, T. M. Tetrahedron 1972, 28, 4981.
L. J. Phys. Org. Chem. 1997, 10, 125. (182) Krygowski, T. M.; Cyrański, M.; Ciesielski, A.; SÄ wirska B.;
(173) Krygowski, T. M.; Wisiorowski, M.; Howard, S. T.; Stolarczyk, Leszczyński P. J. Chem. Inf. Comput. Sci. 1996, 36, 1135.
L. Tetrahedron 1997, 53, 13027. (183) (a) Boese, R.; Blaser, D.; Billups, W. E.; Haley, M. M.; Maulitz,
(174) Boese R.; Polk, M.; Blaser, D. Angew. Chem. Int. 1987, 26, 245. A. H.; Mohler, D. L.; Vollhardt, K. P. C. Angew. Chem. 1994,
(175) Maulitz, A. H.; Stellberg, P.; Boese, R. J. Mol. Struct. (THEO- 106, 321. (b) Boese, R.; Blaser, D.; Billups, W. E.; Haley, M. M.;
CHEM) 1995, 338, 131. Maulitz, A. H.; Mohler, D. L.; Vollhardt, K. P. C. Angew. Chem.,
(176) (a) Boyd, R. J.; Choi, S. C.; Hale, C. C. Chem. Phys. Lett. 1984, Int. Ed. 1994, 33, 313.
112, 136. (b) Haddon, R. C. Pure Appl. Chem. 1982, 54, 1129.
(c) Jemmis, E. D.; Kiran, B. Inorg. Chem. 1998, 37, 2110. (d) CR990326U

Structural Aspects of Aromaticity: Tadeusz Marek Krygowski and Michał Ksawery Cyran Ski

Uploaded by

Copyright:

Available Formats

Structural Aspects of Aromaticity: Tadeusz Marek Krygowski and Michał Ksawery Cyran Ski

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Structural Aspects of Aromaticity: Tadeusz Marek Krygowski and Michał Ksawery Cyran Ski

Uploaded by

Copyright:

Available Formats

Chem. Rev.

2001, 101, 1385−1419 1385

Structural Aspects of Aromaticity

Received August 3, 2000

Contents V. Acronyms 1416

Table 1. Frequency of Chemical/Biochemical/Medical

structuresmay be governed by the σ-electron de-

R is the observed bond length, N is the individual

R(n) - R(1) ) -c ln(n) (11)

Table 7. Empirical Constants for HOSE Model

Table 11. Comparison of IA with HOMA, EN, and GEO

a E, G, and H stand for HOMA, EN, and GEO, respectively.

Table 15. Correlation Coefficients of Regressions

tion, hence higher values of GEO and consequently

I6 and BAC are practically nil (0.09 and -0.25,

D. Other Heterocyclic Analogues of Benzenoid

Chart 3. Aromaticity of Oxa-Analogues of Benzenoid Hydrocarbons

Table 22. HOMA, EN, GEO, and Bird’s I5 Indices for

Figure 10. Molecular structure of tris(cyclopentadienyl)-

differ considerably in their power of interacting with

Chart 5. Aromaticity Indices of Pyropyrroles and Furopyrroles

respectively b Mean experimental geometries. c Calculated

Figure 13. Enumeration of the rings in porphyrine.

points to a known fact that in quinoidal structures

J. Aromaticity vs Deviation from Planarity

Figure 16. Scheme of [5]metacyclophane indicating fold-

Figure 18. The σ- and π-separated energies156 for benzene

Chart 6. Local Values of HOMA, EN, and GEO

Table 25 presents a magnitude of the deformation

ranging from strongly antiaromatic to typically aro-

You might also like