A Combined Experimental and Theoretical Electron Density Study of
Intra-and Intermolecular Interactions in Thiourea S,S-Dioxide
Chi-Rung Lee,*
[b]Ting-Hua Tang,
[a]Likey Chen,
[a]and Yu Wang*
[a]Abstract: The
thiourea
S,S-dioxide
molecule is recognized as a zwitterion
with a high dipole moment and an
unusually long C S bond. The molecule
has a most interesting set of
intermolec-ular
interactions
in
the
crystalline
state–a relatively strong O ¥¥¥H N
hy-drogen bond and very weak
intermolec-ular C ¥¥¥S and N ¥¥¥O interactions. The
molecule has C
ssymmetry, and each
oxygen atom is hydrogen-bonded to two
hydrogen atoms with O ¥¥¥H N
distan-ces of 2.837 and 2.826 ä and angles of
176.61 and 158.38
8. The electron density
distribution is obtained both from X-ray
diffraction data at 110 K and from a
periodic
density
functional
theory
(DFT) calculation. Bond
characteriza-tion is made in terms of the analysis of
topological properties. The covalent
characters of the C N, N H, C S, and
S O bonds are apparent, and the
agree-ment on the topological properties
be-tween experiment and theory is
ade-quate. The features of the Laplacian
distributions, bond paths, and atomic
domains are comparable. In a systematic
approach, DFT calculations are
per-formed based on a monomer, a dimer,
a heptamer, and a crystal to see the
effect on the electron density
distribu-tion due to the intermolecular
interac-tions. The dipole moment of the
mole-cule is enhanced in the solid state. The
typical values of
1
band H
bof the
hydrogen bonds and weak
intermolecu-lar C ¥¥¥S and N ¥¥¥O interactions are
given. All the interactions are verified
by the location of the bond critical point
and its associated topological properties.
The
isovalue
surface
of
Laplacian
charge density and the detailed atomic
graph around each atomic site reveal the
shape of the valence-shell charge
con-centration and provide a reasonable
interpretation of the bonding of each
atom.
Keywords: charge density ¥ density
functional calculations ¥ hydrogen
bonds ¥ topological analysis ¥ X-ray
diffraction
Introduction
Intermolecular interactions, particularly hydrogen bonds, play
a key role in molecular recognition in a wide range of
chemical and biological systems.
[1]These interactions are
almost ubiquitous in molecular crystals and biological
mole-cules. For the hydrogen-bond interaction, studies that use the
theory of atoms in molecules (AIM),
[2]based on both
theoretical
[3±14]and experimental
[15±25]electron densities, have
drawn considerable interest recently. Different types of
hydrogen bonding and weak interactions have been studied
to elucidate the nature of these interactions. The topological
properties of the electron density distribution of both a
molecule and a crystal are based on the gradient vector field
of the electron density
r1(r) and on the Laplacian
distribu-tion of the electron density
r
21(r). Several excellent reviews
have been published
[2, 15, 26±28]on this subject. In the light of the
AIM approach, experimental and theoretical research into
the charge density distributions of many different types of
chemical bond in organic, sulfur-containing compounds and
metal complexes
[29±32]has been studied in terms of local charge
concentration and local charge depletion of a Lewis acid ±
base concept.
[2]This general phenomenon has been well
studied in many other kinds of interaction, such as the
formation of hydrogen bonds,
[3±25]the directional
intermolec-ular interaction of Cl
2[33]and S
4N
4[34, 35]in the solid state, and
even the binding interaction in van der Waals dimers and a
trimer,
[36]as well as the adsorption of molecules on a
surface.
[37]The charge density, chemical bonding, and intermolecular
interactions of urea and thiourea have all been extensively
investigated.
[8, 25, 38±41]It was reported
[38]that the observed
planar geometries of guanidine and urea in the crystal form
[a] Prof. Y. Wang, Prof. T.-H. Tang, Dr. L. ChenDepartment of Chemistry National Taiwan University
Taipei 106, Taiwan (Republic of China) Fax: (886)-2-23636359
E-mail: yuwang@xtal.ch.ntu.edu.tw [b] Prof. C.-R. Lee
Department of Chemical Engineering Minghsin University of Science and Technology Hsin-chu 304, Taiwan (Republic of China) E-mail: crlee@mhit.edu.tw
Supporting information for this article is available on the WWW under http://www.chemeurj.org/ or from the author.
are probably due to hydrogen bonding. In a theoretical study,
hydrogen-bonded aggregates of urea and thiourea were
reported
[39]to form chains and ribbons. The structure and
charge density of the complex of thiourea with parabanic acid
were studied at room temperature,
[25]and two N H ¥¥¥S and
N H ¥¥¥O weak hydrogen-bonding interactions were located
in the cyclic parabanic acid ± thiourea dimer. A preliminary
study was carried out for crystalline urea a decade ago with
the ab initio approach.
[40]In a further study using the periodic
Hartree ± Fock (PHF) approach,
[8]with the geometry taken
from the accurate neutron diffraction,
[41]two in-plane and two
out-of-plane hydrogen bonds, as well as intramolecular
interactions, were described in terms of the AIM theory in
the crystal structure of urea. The topological properties
associated with the bond critical points (BCPs) were
inves-tigated in the gas phase and the solid state, for the purpose of
understanding changes in the properties due to changes in the
packing environment, in other words, the crystal packing
effect. It was found that, on going from molecule to crystal,
both the
j H
bj and 1
bvalues increase for the C N bond, while
these values decrease for all other bonds.
[8]This means that
only the C N bond is strengthened, where an increase of
p
character is induced by crystal packing. It is not often found
that a carbonyl oxygen atom is involved in four hydrogen
bonds, as in this case; the calculated sublimation energy was
91.21 kJ mol
1[8]in comparison with the experimental value of
88
2 kJmol
1.
[42]Recently, quantitative comparisons of the
experimental and theoretical charge densities of urea were
made.
[43]The intermolecular interaction energy in crystalline
urea was calculated both from diffraction data and with the
PHF approach by using a modified atom ± atom
approxima-tion
scheme;
the
calculated
sublimation
energy
was
90 kJ mol
1, which is very close to the value of 96 kJ mol
1obtained by experiment.
[44]The purpose of the present study is to investigate the
intra-and intermolecular interactions, particularly the hydrogen
bonding, in a molecular crystal. Thiourea S,S-dioxide was
chosen due to its zwitterionic character and its C S bond,
which is the longest known (1.8592 ä), to best of our
knowledge. The extensive hydrogen-bond network
through-out the crystal may well be the result of the high dipole
moment of the molecule. The molecule is quite unique and,
together with the previous knowledge about urea and
thiourea,
[8, 25, 38±44]it is bound to give much insight into the
intermolecular interactions. A systematic study concerning
the intermolecular interactions of this molecule should be
very interesting. According to our previous studies, the crystal
structure
[45]of this molecule is in space group Pnma. Each
oxygen atom is involved in two hydrogen bonds similar to
those in urea. A preliminary electron density study
[46]was
made by single-crystal X-ray diffraction at 110 K and by a
molecular-orbital calculation based on an isolated molecule.
The agreement on the chemical bonding characterization
between the experimental and theoretical results was
ad-equate.
[46]The main interests of this work are, however, the
hydrogen bonds and other weak intermolecular interactions.
In order to understand the nature of such intermolecular
interactions in this molecule in the solid state, theoretical
charge densities are to be calculated systematically for the
monomer, dimer, heptamer, and crystal. The effect on the
chemical bonds due to molecular packing will be investigated
in terms of the AIM theory.
Results and Discussion
Structure and multipole model: The molecular structure of
thiourea S,S-dioxide with its atomic labeling and internal
coordinates is depicted in Figure 1. The molecular symmetry
is C
s, and two types of hydrogen bond exist in the crystal. In
Figure 1. The molecular structure of thiourea S,S-dioxide with atomic labeling and internal coordinates.
the first, molecules are linked to each other through hydrogen
bonds in pairs between the N H1 group and an O atom to
form a ™head-to-tail∫ type bond with an O ¥¥¥H1 N distance
of 2.837 ä and an angle of 176.61
8 (Figure 2a, b). The other
type of bond is between the N H2 group and a neighboring O
atom with an O ¥¥¥H2 N distance of 2.826 ä and an angle of
158.38
8 (Figure 2a, b). Each H atom is involved in one type of
hydrogen bond, but each O atom is involved in two hydrogen
bonds. These two types of hydrogen bond form a zig-zag
network and give rise to an infinite network of molecules
throughout the crystal (Figure 2 a, b). In addition, there are
weak C ¥¥¥S and N ¥¥¥O interactions in the crystal as shown in
Figure 2 c. The structural parameters are the same (within
standard deviations) as those given in the previous work,
[46]where the deformation density maps based on the multipole
model were also shown. In this work, multipole refinements
are performed with an additional quadrupole term for the H
atom and
k' for each atom. Parallel multipole refinements
based on the structural amplitudes (F values) derived from a
periodic density functional theory (PDFT) calculation are
also carried out. The residual maps after the multipole
refinements based on experimental structural amplitudes
and the theoretically derived ones are displayed in Figures 3 a
and b, respectively. The maps are essentially featureless. The
agreement indices and the multipole coefficients from various
refinements are qualitatively the same. Of course, the U
ijvalues were set to zero for the model derived from the PDFT
calculation. The list of parameters is given in the Supporting
FULL PAPER
C.-R. Lee, Y. Wang et al.
Figure 2. The intramolecular bonds (sticks) and intermolecular hydrogen bonds and weak interactions (dashed lines) in the crystal packing. a) Projection on the a,b plane showing the hydrogen bonds, b) display of the in-plane and out-of-plane hydrogen bonds, c) depiction of the C S and N O weak intermolecular interactions.
Figure 3. Residual maps after the multipole refinement based on a) ex-perimental data and b) theoretically derived structural factors. Solid contours denote positive values, dashed contours denote negative values. The contour interval is 0.1 e ä 3.
Information. The additional parameters in the present study
do improve the refinements. The refinement based on the
theoretically derived structure amplitudes is particularly
successful, with the lowest agreement indices shown in
Table 1. The refinements on the experimental data with
k'
constrained from, and free from, the values derived from the
PDFT refinement do not appear to show any difference. The
k'-restricted multipole model (KRMM)
[47]is used for the
subsequent density studies.
Intramolecular interactions: The comparison of the thiourea
S,S-dioxide molecule with the crystal geometry (CG) and the
optimized geometry (OG) in the gas phase was made, and the
Table 1. Agreement indices of various multipole refinements.R(F) Rw(F) R(F2) Rw(F2) GOF Variable
octapole[a][46] 0.0174 0.0217 0.0317 0.0468 1.47 97
KRMM 0.0169 0.0150 0.0278 0.0299 4.28 111
UMM[b] 0.0166 0.0149 0.0282 0.0297 4.25 115
PDFT/XD program 0.0097 0.0092 0.0119 0.0205 4.79 85 [a] Octapole with hexadecapole of the S atom. [b] UMM Unrestricted multipole model.
optimized geometry of the molecule was constrained to C
ssymmetry. Significant differences are found in these two
geometries: 1) Marked lengthening of the C S bond length is
found in the OG (1.9888 rather than 1.8592(6) ä); 2) a bigger
dihedral angle is found between the SO
2and (NH
2)
2C groups
in the OG (78.9 compared with 69.18); and 3) two extra
intramolecular hydrogen bonds (O ¥¥¥H2 N) are present in
the OG that are absent in the CG (Figure 4). However, the
Figure 4. Laplacian charge density distributions of the molecule with a) crystal geometry, and b) optimized geometry. Dashed contours denote positive values and solid contours denote negative values. Solid circles denote the BCPs. Heavy solid lines denote the bond paths. The locations of nuclei are labeled in a). Contours are in steps of 2m 10ne ä5(m 1 ± 3,
n 3 ± 1).
energy difference between the two geometries is relatively
small (30.59 kJ mol
1). Therefore the crystal geometry is used
for the subsequent analyses. In the light of the topological
properties analysis of the charge density distribution, the
1
b,
r
21
b
, and H
bvalues (Table 2) of all intramolecular bonds are
in good agreement between the two geometries. The only
discrepancy found is the values related to the S C bond, due
to the marked lengthening of the bond length in the OG. In
addition, the charge concentration of the S atom is different
between the CG and OG in the thiourea plane shown in
Figure 4; this is apparently owing to the extra hydrogen bond
in the OG. It is clearly shown that all the intramolecular bonds
are covalent in character. The C N and S O bonds have
relatively large
1
band
jH
bj values, which indicates a possible
double bond. The
r
21
b
value of the S O bond is positive, but
the H
bvalue is negative and the BCP is significantly closer to
the S atom; this indicates a highly polarized covalent bond,
although similar findings would also occur in a very short
polarized bond.
[2, 48]It was reported
[49]that the S O bond in
compounds containing a hypervalent sulfur atom in
> SO and
>SO
2fragments could involve possible participation of an
ionic
p bond, however more than 90% of the charge density is
located at the O atoms in such a
p bond. This is in agreement
with the results based on a different population analysis.
[50]In
one recent report,
[51]the S O bond was also expressed as a
typical polarized
s bond. Compared with the topological
analysis of the S O bond in Me
2SO
2and H
2SO
2[51]this bond is
longer in thiourea S,S-dioxide (1.500 compared with 1.456/
1.466 ä); the
1
band H
bvalues of the S O bond in this
compound are also slightly smaller in magnitude. With a
1
bvalue of 1.71 e ä
3(versus 1.97 in Me
2
SO
2) and a negative H
b( 2.04 Hartree ä
3), there is definitely an indication of a
partial double bond, and the polarity of the charge density is
toward the oxygen atom, a fact which is quite consistent with
earlier findings.
[49, 51]The C S bond is the longest single bond
found in the literature.
[46]The BCP is slightly closer to the S
atom, which indicates that the S atom is slightly more positive
than the C atom; this is indeed so, as observed from the
atomic charge (Table 4). The
1
band
jH
bj values of the C S
bond are relatively low. This indicates that the C S bond here
is a rather weak single bond. An empirical linear correlation
between the bond length and the values of
1
bwas reported
recently for C S bonds of several sulfur-containing
com-pounds.
[30]The
1
b
value and bond length of this molecule fit
well with this linear relationship. The C N bond here is
shorter than that of urea.
[8]The
1
b
values are comparable. The
large negative H
bvalue of
4.01 Hartree ä
3certainly
indicates that the C N bond is more than a single bond. It
was also reported that the
p character of the C N bond is
enhanced for urea from the isolated molecule to the solid
form.
[8]In order to realize the effect of intermolecular interactions
exerted on the chemical bonds in the solid, the systematic
topological properties analysis is applied to the theoretically
calculated electron densities of a monomer, a heptamer, and a
Table 2. The topological properties of the theoretically calculated charge densities at the bond critical point for the crystal geometry (CG) and the optimized geometry (OG).[a]Bond Bond length d1[b] 1
b r21b Hb [ä] [ä] [e ä 3] [e ä 5] [Hartree ä 3] S O 1.4997(6) 0.583 1.71 18.96 2.04 1.509 0.586 1.68 17.54 2.00 S C 1.8592(6) 0.884 1.12 6.69 0.87 1.988 0.975 0.88 2.76 0.50 C N 1.3096(7) 0.444 2.28 17.90 4.01 1.310 0.445 2.27 18.01 3.99 H1 N 1.030 0.259 2.12 41.18 3.22 1.013 0.250 2.20 43.36 3.39 H2 N 1.030 0.238 2.10 42.24 3.24 1.022 0.234 2.14 43.18 3.33 O ¥ ¥ ¥ H2 2.226 0.876 0.12 1.41 0.00
[a] First line for the CG, second line for the OG. [b] d1 is the distance from the BCP to the first atom.
FULL PAPER
C.-R. Lee, Y. Wang et al.
crystal; the results are tabulated in Table 3. The
correspond-ing values obtained by experiment and from the multipole
model imposed on theoretical structure factors are also listed
in the table for comparison. The trend of strengthening the
intramolecular bond in terms of
1
bfrom the monomer and the
heptamer to the crystal is detectable in the theoretical results.
Some discrepancies of the topological properties are found
between the experimental and theoretical values, particularly
on the sign of the
r
21
b
value of S O bonds. It is noticeable
that the BCP of the S O bond observed in the experiment is
not as close to the S atom as in the theoretical calculations,
therefore the
l
3value is reasonable to make the
r
21
bnegative
as expected.
[48]It was pointed out
[52]that topological
discrep-ancies between experimental and theoretical crystal charge
densities are mainly attributed to the nature of the radial
function in the experimental multipole model, which would
result in the difference of the
l
3value. It is worth mentioning
that the intramolecular bonds, except the N H bond, are all
enhanced by the intermolecular interactions according to the
topological analyses. The N H bonds are weakened in the
solid, which is expected when the corresponding
intermolec-ular hydrogen bond is strengthened. In general, the
agree-ment in
1
bbetween the experiment and theory improves when
intermolecular interactions are taken into account from a
monomer to a heptamer and then to a crystal. However, the
H
bvalues do not follow the same trend. This may result from
the approximate expression
[19, 53±55]in the experimental values,
which does not strictly fit for a covalent bond.
Atomic domains and atomic net charges: According to AIM
theory,
[2]a set of zero-flux surfaces partitions the molecule
into unique atom domains (
W). These zero-flux surfaces
projected onto a molecular plane, which contained the S, C,
and N atoms of a moiety, together with the total electron
density distribution, bond paths, and BCPs, are illustrated in
Figure 5 from a DFT calculation of a heptamer model and
from experimental results. The shape and the size of these
domains are nearly identical between the theory and
experi-ment.
The AIM net atomic charge (q
W) can be obtained by
numerical integration of the electron density distribution
within the atomic domain (W) and subtraction from the
atomic number (Z
n): q
W Z
n
W
1(r)dt. The net atomic
charges and the fragment charges are tabulated in Table 4.
The H, C, and S atoms are positively charged and the O and N
atoms are negatively charged. The agreement between
experiment and theory is reasonable. This charge distribution
gives a positive C(NH
2)
2moiety and a negative SO
2fragment,
which fits the zwitterionic description; however the negative
charge is mainly on the O atoms of the SO
2fragment. The
dipole moment of the molecule seems to be enhanced from
the monomer to the crystal (12.6 versus 14.6 Debye), which
was also found in the case of urea (2.02 versus 2.77 Debye)
[8]and nitro aniline compounds (13.3 versus 16.1 Debye).
[56, 57]The dipole moment derived from the Mulliken charges or the
monopole values obtained from the multipole refinement
gives roughly the same value. This confirms the earlier
finding.
[47]Intermolecular interactions
Intermolecular hydrogen bonding: The existence of two types
of hydrogen bonds is confirmed by the topological properties
analysis of electron density (Table 5). According to the
topological properties associated with the BCP of the
Table 3. Intramolecular interactions : the topological properties associatedwith the BCPs.[a]
Bond/ d1 1b r21b Hb
Bond length [ä] [ä] [e ä 3] [e ä 5] [Hartree ä3]
S O 0.583 1.709 18.964 2.038 /1.4997(4) 0.583 1.749 19.469 1.958 0.581 1.779 20.866 2.012 0.609 1.772 6.825 1.677 0.608 1.852 5.353 1.772 S C 0.884 1.117 6.686 0.873 /1.8592(6) 0.922 1.173 7.790 0.878 0.939 1.186 6.984 0.817 0.933 1.167 4.967 0.867 0.985 1.195 2.977 0.838 C N 0.444 2.279 17.900 4.007 /1.3096(5) 0.452 2.322 22.396 4.103 0.442 2.419 18.488 4.125 0.450 2.444 22.419 3.135 0.479 2.302 17.770 2.906 H1 N 0.259 2.124 41.180 3.224 /1.030 0.237 2.102 41.805 3.246 0.230 2.170 41.395 3.217 0.233 1.916 30.605 2.482 0.203 1.776 29.992 2.382 H2 N 0.238 2.102 42.242 3.238 /1.030 0.224 2.077 41.942 3.227 0.230 2.170 41.395 3.227 0.241 1.946 30.513 2.503 0.179 1.532 25.074 2.165 S O[51] 1.97 22.94 2.38 /1.456 S O[51] 1.94 20.55 2.36 /1.466 S O[51] 0.878 1.357 0.502 /1.514 S C[30] 0.79 1.36 3.75 1.03 /1.712 0.71 1.42 8.97 1.60 S [30] 0.96 1.21 3.53 0.90 /1.824 0.96 1.22 6.78 0.83 S C[30] 0.83 1.35 6.91 1.13 /1.721 0.78 1.40 9.78 1.37 S C[30] 0.78 1.50 4.87 1.46 /1.657 0.64 1.45 0.18 1.71 C N[30] 0.71 2.45 11.15 2.43 /1.326 0.45 2.30 18.05 3.83 C N[30] 0.69 2.53 6.28 2.95 /1.318 0.45 2.32 18.43 3.83 C N[30] 0.60 2.51 24.80 2.82 /1.338 0.46 2.27 21.35 3.73 C N[8] 0.442 2.30 22.65 /1.345 0.451 2.36 27.71 ± C N[75] 0.52 2.31 20.65 /1.322 0.52 2.39 21.39 ±
[a] For thiourea S,S-dioxide: first line from DFT calculations of monomer, second line from DFT calculations of heptamer, third line from DFT calculations of crystal, fourth line from PDFT/XD program, fifth line from KRMM. For other compounds (with references): first line from calcula-tions, second line from experiment.
Figure 5. Total electron density, 1(r), bond paths, and atom domains partitioning by zero-flux surfaces in the molecular plane derived from a) the DFT calculation of a heptamer and b) experimental diffraction data. The nuclei at the plane are labeled in (a). The charge density contours are in steps of 2m 10ne ä 3(m 1 ± 3, n 3 ± 3).
H1 ¥¥¥O and H2 ¥¥¥O hydrogen bonds, these are relatively
strong hydrogen bonds with a
1
bvalue of
0.2 eä
3, which is
in accordance with values obtained elsewhere,
[19, 21, 55]for
example,
1
b 0.199 eä
3for the F ¥¥¥H hydrogen bond.
[55]There is no obvious difference in the
1
bor H
bvalues from a
dimer, a heptamer, and a crystal; nevertheless the
exper-imental values are significantly lower. However the kinetic
energy density, G
b(Table 5) fits well with the exponential
expression given in the literature.
[19]It also fits in the linear
relationship
[53]between G
b
and
l
3.The
1
bvalue of the O ¥¥¥H1
bond is slightly larger than the O ¥¥¥H2 interaction. The BSSE
(basis set superposition error) corrected binding interaction
energies are calculated based on the specifically chosen
dimers shown in Figures 6 and 7. This correction results in
energies of
40.83 and
54.53 kJ mol
1for N H1 ¥¥¥O and
N H2 ¥¥¥O, respectively. In general the nonbonded charge
concentrations are the preferred sites of protonation. There
are two Laplacian charge concentrations (vertex critical point
in the atomic graph) in the valence-shell charge concentration
(VSCC) of the O atom (see Supporting Information) and a
Laplacian local minimum (face critical point) of hydrogen
toward this charge concentration of the O atom. In principle,
the locations of the N H bond with its Laplacian local
minimum and the charge concentration of O should be in
alignment as far as possible. This is true in cases of both
N H2 ¥¥¥O and N H1 ¥¥¥O hydrogen bonds. A similar feature
was reported in crystalline urea.
[8]Furthermore the
BSSE-corrected binding energies of N H1 ¥¥¥O and N H2 ¥¥¥O
were calculated for a trimer including both N H1 ¥¥¥O and
N H2 ¥¥¥O type interactions. The values were
45.66 and
54.15 kJ mol
1, respectively. The binding energies increase a
bit for the N H1 ¥¥¥O bond from a dimer to a trimer.
Weak intermolecular binding interactions C ¥¥¥ S
' and O ¥¥¥ N':
In addition to the intermolecular hydrogen bonds, the
thiourea S,S-dioxide molecules in the crystal are linked to
one another through weak C ¥¥¥S
' and O ¥¥¥N' binding
interactions between neighboring moieties (Figure 2 c). These
binding interactions are again confirmed by the location of the
BCPs and the trace of the associated bond path to the relevant
nuclei, as displayed in Figure 8. The related topological
properties are listed in Table 5. It is obvious that the
1
bvalues
(approximately 0.02 ± 0.06 e ä
3) are much less than those of
hydrogen bonds (
0.2 eä
3). The bond paths of C ¥¥¥S
' and
O ¥ ¥ ¥ N
' are slightly longer than the respective geometrical
distances (bond lengths), so they are slightly bent as shown in
Figure 8. The nature of these weak intermolecular binding
interactions can be understood correctly by the VSCC
Laplacian critical points. The 3D isovalue surfaces of
Lap-lacian distributions derived by experiment and theory are
depicted in Figure 9, where a nonbonded charge
concentra-tion is located at the VSCC of the S atom and two charge
depletions are located above and below the C atom. While in
the crystal, the nonbonded charge concentration of the S atom
is directly inserted toward the charge depletion of the C atom
from the molecule below to form the C ¥¥¥S
' binding
interactions (Figure 9 a, b). In other words, the nonbonded
Laplacian charge concentration of the S atom serves as a
Lewis base or an electrophile and the Laplacian charge
Table 4. The atomic AIM charges (qW) and the molecular dipole moments (Debye).Atom Monomer Heptamer PDFT PDFT/XD program Experiment/KRMM
S 2.164 2.141 2.145 1.955 1.759 O 1.313 1.347 1.362 1.339 1.255 N 1.287 1.386 1.407 1.246 1.452 C 1.214 1.221 1.277 0.790 0.893 H1 0.410 0.502 0.522 0.611 0.671 H2 0.500 0.550 0.538 0.602 0.707 SO2 0.46 0.55 0.58 0.72 0.75 C(NH2)2 0.46 0.55 0.58 0.72 0.75 dipole moment 12.6 14.2 14.6 15.5 16.3
FULL PAPER
C.-R. Lee, Y. Wang et al.
depletion of the C atom is the Lewis acid or a nucleophile.
Similarly the N atom possesses two nonbonded charge
concentrations, one on each side of the plane. One of these
charge concentrations of the N atom can align itself toward
the Laplacian charge depletion of the O atom of a
neighbour-ing molecule to form the O ¥¥¥N
' binding interaction. These
weak C ¥¥¥S and N ¥¥¥O interactions together with hydrogen
bonds demonstrate the 3D directional ™key ± lock∫
architec-ture in the crystal. Based on the chosen dimer model, the
BSSE-corrected binding interaction energies,
DE, are 9.48
and
5.80 kJ mol
1for C ¥¥¥S
' and O ¥¥¥N', respectively.
Conclusion
An exact comparison has been accomplished between the
experiment and theory of electron density based on the
multipole model. The covalent bonding characters of the C S,
S O, and C N bonds in thiourea S,S-dioxide have been
illustrated by topological analysis. The partial double bond
character of the S O and C N bonds is recognized, with ionic
p character of the S O bond. A highly polar single C S bond
between two oppositely charged fragments has been
estab-lished. The large dipole moment results from a positive
C(NH
2)
2moiety and a negative SO
2fragment, a firm
illustration of the zwitterionic character. Two categories of
intermolecular interactions are identified: the relatively
strong hydrogen bonds, with
1
b0.2 eä
3and binding
energies (
DE) of 50 kJmol
1, and the relatively weak C ¥¥¥S
'
and O ¥¥¥N
' interactions, with 1
b0.02 ± 0.05 eä
3and
DE 6 ± 9 kJmol
1. The natures of these intermolecular
interactions have been demonstrated as 3D directional
interactions. The effect of such intermolecular interactions
on the chemical bond is detected through the systematic
studies. The chemical reactivity of this molecule can be
understood according to the fragment charges of the SO
2and
C(NH
2)
2groups as well as the knowledge of VSCC
distribu-tion.
Computational Methods
Multipole refinement: Thiourea S,S-dioxide was prepared by oxidation of thiourea with hydrogen peroxide at 08C and crystallized in an aqueous solution.[45]The intensity data were collected on a CAD4 diffractometer at
110 K. Details were described in our previous work.[46]A multipole model
refinement was reinvestigated by using the XD program.[58]The multipolar
model is expressed as a series expansion of spherical harmonic terms (ylmp)
multiplied by a Slater-type radial function Rl(r).[17, 59]Spherical harmonic
expansion terms up to hexadecapoles were included for S atoms, up to octapoles for O, N, and C atoms, and up to quadruples for H atoms. The core and valence electron scattering factors for each atom are taken from International Tables for X-rayCrystallography(1974, Vol. IV). The core electron configurations are assumed to have Ne core for S and He core for O, N, and C atoms. During the refinement, H atoms are moving along N H vectors to make an N H distance of 1.03 ä.[60]The n
lvalues (l 1 ± 4) of S
are 4, 4, 4, 4, those of N, O, and C are 2, 2, 3, and those of H are 1, 2. The k'-restricted multipole model (KRMM)[47]was carried out at the final stage.
The radial k' coefficients were fixed at values derived from multipole refinement of theoretical structure factors obtained from PDFT calcula-tions (k' coefficient for S: 0.890; O: 2.315; N: 0.939; C: 0.725; and H: 1.2). Table 5. Intermolecular interactions: the topological properties associated with the BCPs.[a]
Bond/bond length [ä]/ d1 1b r21b Gb Hb DE[b]
bond path [ä] [ä] [e ä3] [e ä5] [Hartree ä 3] [Hartree ä3] [kJ mol 1]
O ¥¥¥H1 1.176 0.222 2.486 0.176 0.002 40.83 /1.810(2) 1.179 0.225 2.479 0.176 0.003 /1.811 1.179 0.235 2.376 0.171 0.005 1.192 0.216 2.232 0.167 0.010 1.202 0.178 2.883 0.180 0.022 O ¥¥¥H2 1.199 0.199 2.171 0.152 0.000 54.53 /1.843(2) 1.197 0.196 2.210 0.155 0.001 /1.849 1.189 0.204 2.267 0.157 0.002 1.221 0.166 2.649 0.164 0.022 1.237 0.146 2.323 0.141 0.022 S ¥ ¥ ¥ C' 1.820 0.054 0.649 0.031 0.010 9.48 /3.3128(6) 1.813 0.055 0.666 0.036 0.010 /3.323 1.805 0.058 0.659 0.036 0.010 1.776 0.064 0.718 0.042 0.009 1.769 0.062 0.766 0.044 0.010 O ¥ ¥ ¥ N' 1.794 0.018 0.270 0.015 0.004 5.80 /3.619(2) 1.787 0.019 0.278 0.016 0.004 /3.630 1.791 0.019 0.279 0.016 0.004 1.851 0.017 0.247 0.012 0.005 1.836 0.019 0.261 0.013 0.005 O ¥ ¥ ¥ H[21] 0.273 3.40 0.216 0.013 /1.72 0.295 3.10 0.242 0.020 O ¥ ¥ ¥ H[21] 0.230 2.88 0.202 0.000 /1.81 0.248 2.66 0.202 0.000 O ¥ ¥ ¥ H[21] 0.174 2.49 0.155 0.013 /1.93 0.188 2.30 0.149 0.007
[a] First line from DFT calculations of chosen dimers, second line from DFT calculation of heptamer, third line from PDFT calculations of crystal, fourth line from PDFT/XD program, fifth line from KRMM. The intermolecular binding interaction energies are calculated based on the corresponding dimer models in the BSSE correction.
Thez values of H, N, O, and S are 3.1762, 3.8407, 4.4724 and 3.9496, respectively. The experimental Laplacian distributions are depicted with the XD program,[58]and the
contour maps of the charge density distri-butions, zero-flux surfaces, and bond paths are drawn with the PROP program.[61]
The total energy density at the bond critical point Hb values evaluated from
the experimental electron density by the multipole model was derived according to the approximate expression,[54]where the
kinetic energy density, Gb, is directly
related to the electron density, introduced by the semiclassical Thomas ± Fermi equa-tion, whereD1(r) 1(r) 1ci(r); in
prac-tice,1ci(r) is the deformation density, that
is,1ci(rc) 1mul(rc) 1iam(rc). The Hbvalue
is the sum of Gb and Vb, which can be
estimated roughly by this generalized ap-proach.
Theoretical calculations: All DFT[62]
cal-culations, including those for the single molecule and oligomers of thiourea S,S-dioxide, were carried out with the GAUS-SIAN 98 program.[63]The correlation
cor-rection by Lee, Yang, and Parr (LYP)[64]
Figure 6. a) Model of the chosen dimer with O ¥¥¥H1 N hydrogen bonds. Laplacian charge density distributions in the plane containing O ¥¥¥H1 N hydrogen bonds are shown as calculated from b) the DFT calculation of the chosen dimer, c) the PDFT calculation, d) the multipole model of theoretically derived structure factors and e) the multipole model of experimental data. Contours are stepped as in Figure 4.
Figure 7. a) Model of the O ¥¥¥H2 N hydrogen bond. Laplacian charge density distributions of the O ¥¥¥H2 N hydrogen bond are shown in b) ± e) as defined as in Figure 6. Contours are stepped as in Figure 4.
Figure 8. Weak intermolecular interactions indicated by bond paths (black lines), critical points (black dots), and charge density distribution (con-tours) in a heptamer model: a) C ¥¥¥S' interaction, b) O ¥¥¥N' interaction. Contours in1(r) are stepped as in Figure 5.
FULL PAPER
C.-R. Lee, Y. Wang et al.
together with Becke×s nonlocal, gradient approach to the exchange functional in its three-parameter hybrid density form (B3LYP)[65] was
used. The standard split-valence 6-31G(d,p) basis set[62, 66]was employed for
all the calculations. Fully PDFT calculation of the crystal was incorporated in the CRYSTAL algorithm by Saunders et al.[67]by using the CRYSTAL 98
program.[67]According to the Bloch function for an insulator, the K-point
sampling of 6, 6, and 6 was chosen isotropically along the reciprocal axis a*, b*, and c* respectively. The structural amplitudes (Ftheo) were thus derived
from this calculation. The multipole refinement[59]was then applied to these
theoretically calculated structural amplitudes, and thek' thus obtained was adapted in the experimental multipole model.
Topological properties analysis: The total electron density obtained from the experiment was calculated according to the multipole refinement model.[59] The total electron density for the theoretical model was
calculated on the basis of the aforementioned DFT and PDFT calculations. The topological properties, maps of the charge density distributions, and Laplacian distributions were performed with the programs AIMPAC[68]and
AIM98PC[69]for the single molecule and oligomers, with the program
TOPOND[70]used for the crystal.
The magnitude of the electron density at the BCP,1b, correlates with the
bond distance and the bond order and, therefore, with the bond strength.[71, 72] The sign of the Laplacian distribution of1 at the BCP,
r21
b, could be used to distinguish bonding features between a shared
interaction (covalent bond) and a closed-shell interaction (ionic bond). It was suggested[27, 73]that the negative total energy density value, H
b, at the
BCP can be interpreted as the sufficient condition of a covalent bond and could be used as a qualitative measure for covalence. Correlation between the bonding and strain energies of hydrocarbon molecules was reported with1band Hbvalues.[74]The network of bond paths defines the shape of
the molecule. A gradient path for which the electron density decreases most rapidly is developed in all directions normal to the bond. The set of such gradient paths starting at each BCP defines a zero-flux surface separating two bonded atoms. The network of these surfaces (one per
bond) will partition the molecule into unique atomic domains (basins) for which the hypervirial theorem is satisfied. Numerical integration of the electron density within such a region yields the net charge of the given atom[75]called the AIM charge. The Laplacian distribution,r21(r), is also a
very useful tool for understanding the chemical reactivity and revealing a simple 3D directional interaction in the molecular crystal. Similarly to the aforementioned1(r) distribution, the topological properties of r21(r) can
be summarized by its critical points. The atomic graph thus defined denotes the connectivity of the local valence shell charge concentration (VSCC). Such a polyhedron obeys Eular×s formula: V EF 2, where V, E, F stand for the number of vertices, edges, and faces, respectively. A correlation between these critical points ofr21(r) in the valence shell
and the location of the active site has been established.[2, 14, 76]
Acknowledgement
The authors thank the National Science Council of the Republic of China for the financial support and are grateful to the National Center of High-Performance Computing for access to their computing facilities and software packages.
[1] a) C. F. Mellot, A. M. Davidson, J. Eckert, A. K. Cheethan, J. Phys. Chem. B 1998, 102, 2530 ± 2535; b) S. E. Barber, K. E. S. Dean, A. J. Kirby, Can. J. Chem. 1999, 77, 792 ± 801; c) C. Cox, T. Lectka, J. Am. Chem. Soc. 1998, 120, 10 660 ± 11 668; d) K.-J. Lin, Y. Wang, J. Phys. Chem. 1994, 98, 11 685 ± 11 693; e) A. Ivancich, T. A. Mattioli, Photo-synth. Res. 1998, 55, 207 ± 215; f) A. Aurora, G. D. Rose, Protein Sci., 1998, 7, 21 ± 38; g) G. R. Desiraju, Crystal Engineering ± The Design of Organic Solids, Elsevier, New York, 1989.
[2] R. F. W. Bader, Atoms in Molecules : a Quantum Theory, Clarendon Press, Oxford, 1990.
[3] M. T. Caroll, C. Chang, R. F. W. Bader, Mol. Phys. 1988, 63, 387 ± 405. [4] M. T. Caroll, R. F. W. Bader, Mol. Phys. 1988, 65, 695 ± 722. [5] P. L. A. Popelier, R. F. W. Bader, Chem. Phys. Lett. 1992, 189, 542 ±
548.
[6] T.-H. Tang, D. M. Whitefield, S. P. Douglas, J. J. Krepinsky, I. G. Csizmadia, Can. J. Chem. 1992, 70, 2434 ± 2448.
[7] D. M. Whitefield, T.-H. Tang, J. Am. Chem. Soc. 1993, 115, 9648 ± 9654.
[8] C. Gatti, V. R. Saunders, C. Roetti, J. Chem. Phys. 1994, 101, 10 686 ± 10 696.
[9] U. Koch, P. L. A. Popelier, J. Phys. Chem. 1995, 99, 9747 ± 9754. [10] T.-H. Tang, Y.-P. Cui, Can. J. Chem. 1996, 74, 1162 ± 1170. [11] J. A. Platts, S. T. Howard, J. Chem. Phys. 1996, 105, 4668 ± 4674. [12] J. A. Platts, S. T. Howard, B. R. F. Bracke, J. Am. Chem. Soc. 1996, 118,
2726 ± 2733.
[13] P. L. A. Popelier, J. Phys. Chem. 1998, A102, 1873 ± 1878.
[14] C. Gatti, R. Bianchi, R. Destro, F. Merati, J. Mol. Struct.(THEO-CHEM) 1992, 255, 409 ± 433.
[15] T. S. Koritsanszky, P. Coppens, Chem. Rev. 2001, 101, 1583 ± 1627. [16] P. Coppens, X-rayCharge Densities and Chemical Bonding, Oxford
University Press, Oxford, 1997.
[17] P. Coppens, Acta Crystallogr. Sect. A 1998, 54, 779 ± 788.
[18] R. Flaig, T. Koritsa¬nszky, D. Zobel, P. Luger, J. Am. Chem. Soc. 1998, 120, 2227 ± 2238.
[19] E. Espinosa, E. Molins, C. Lecomte, Chem. Phys. Lett. 1998, 285, 170 ± 173.
[20] J. Overgaad, B. Schi˘tt, F. K. Larsen, A. J. Schultz, J. C. MacDonald, B. B. Iversen, Angew. Chem. 1999, 111, 1321 ± 1324; Angew. Chem. Int. Ed. 1999, 38, 1239 ± 1242.
[21] P. Coppens, Y. A. Abramov, M. Carducci, B. Korjov, I. Novozhilova, C. Alhambra, M. R. Pressprich, J. Am. Chem. Soc. 1999, 121, 2585 ± 2593.
[22] R. Flaig, T. Koritsa¬nszky, J. Janczak, H.-G. Krane, W. Morgenroth, P. Luger, Angew. Chem. 1999, 111, 1494 ± 1497; Angew. Chem. Int. Ed. 1999, 38, 1397 ± 1400.
[23] F. Benabicha, V. Pichon-Pesme, C. Jelsch, C. Lecomte, A. Khmou, Acta Crystallogr. Sect. B 2000, 56, 155 ± 165.
Figure 9. The isovalue surfaces (r21(r) 11.0 eä 5) of Laplacian
dis-tributions concerning the C ¥¥¥S' interaction between two moieties from a) theory based on a heptamer model and b) experiment.
[24] W. D. Arnold, L. K. Sanders, M. T. McMahon, A. V. Volkov, G. Wu, P. Coppens, S. R. Wilson, N. Godbout, R. Oldfield, J. Am. Chem. Soc. 2000, 122, 4708 ± 4717.
[25] H.-P. Weber, B. M. Craven, Acta Crystallogr. Sect. B 1987, 43, 202 ± 209.
[26] R. F. W. Bader, Chem. Rev. 1991, 91, 893 ± 928.
[27] E. Kraka, D. Cremer in The Concept of the Chemical Bond, Theoretical Models of Chemical Bonding, Part 2 (Ed.: Z. B. Makic), Springer-Verlag, New York, 1990.
[28] R. F. W. Bader, J. Phys. Chem. 1998, A102, 7314 ± 7323.
[29] C.-R. Lee, T.-H. Tang, C.-C. Wang, L. Chen, Y. Wang, unpublished results.
[30] C.-R. Lee, L. Y. Tan, Y. Wang, J. Phys. Chem. of Solids, 2001, 62, 1613 ± 1628.
[31] C.-C. Wang, T.-H. Tang, Y. Wang, J. Phys. Chem. 2000, A104, 9566 ± 9572.
[32] C.-R. Lee, C.-C. Wang, K.-C. Chen, G.-H. Lee, Y. Wang, J. Phys. Chem. 1999, A103, 156 ± 165.
[33] V. G. Tsirelson, P.-F. Zou, T.-H. Tang, R. F. W. Bader, Acta Crystallogr. Sect. A 1995, 51, 143 ± 153.
[34] T.-H. Tang, R. F. W. Bader, P. J. MacDougall, Inorg. Chem. 1985, 24, 2047 ± 2053.
[35] W. Scherer, M. Spiegler, B. Pedersen, M. Tafiposky, W. Hieringer, B. Reinhard, A. J. Downs, G. S. McGrady, Chem. Commun. 2000, 635. [36] R. G. A. Bone, R. F. W. Bader, J. Phys. Chem. 1996, 100, 10 892 ±
10 911.
[37] a) Y. Aray, R. F. W. Bader, Surf. Sci. 1996, 351, 233 ± 249; b) Y. Aray, J. Rodriguez, D. Vega, J. Phys. Chem. 2000, B104, 5225 ± 5231. [38] A. Gobbi, G. Frenking, J. Am. Chem. Soc. 1993, 115, 2362 ± 2372. [39] A. Masunov, J. J. Dannenberg, J. Phys. Chem. 1999, A103, 178 ± 184;
A. Masunov, J. J. Dannenberg, J. Phys. Chem. 2000, B104, 806 ± 810. [40] R. Dovesi, M. Causa¬, R. Orlando, C. Roetti, V. R. Saunders, J. Chem.
Phys. 1990, 92, 7402 ± 7411.
[41] S. Swaminathan, B. M. Craven, R. K. McMullan, Acta. Crystallogr. Sect. B 1984, 40, 300 ± 306.
[42] K. Suzuki, S. Onishi, T. Koide, S. Seki, Bull. Chem. Soc. Jpn. 1956, 29, 127 ± 131.
[43] a) V. Zavodnik, A. Stash, V. Tsirelson, R. de Vries, D. Feil, Acta Crystallogr. Sect. B 1999, 55, 45 ± 54; b) K. Y. Suponitsky, V. G. Tsirelson, D. Feil, Acta Crystallogr. Sect. A 1999, 55, 821 ± 827; c) R. Y. de Vries, D. Feil, V. G. Tsirelson, Acta Crystallogr. Sect. B 2000, 56, 118 ± 123.
[44] H. G. M. De Wit, J. C. Van Miltenburg, C. G. De Kruif, J. Thermodyn. 1983, 15, 651 ± 654.
[45] I.-C. Chen, Y. Wang, Acta Crystallogr. Sect. C 1984, 40, 1937 ± 1938. [46] Y. Wang, N.-L. Chang, C.-T. Pai, Inorg. Chem. 1990, 29, 3256 ± 3259. [47] A. Volkov, C. Gatti, Y. Abramov, P. Coppens, Acta Crystallogr. Sect. A
2000, 56, 252 ± 258.
[48] T.-S. Hwang, Y. Wang, J. Phys. Chem. 1998, A102, 3726 ± 3731. [49] J. Cioslowski, P. R. Surja¬n, J. Mol. Struct.(THEOCHEM) 1992, 255,
9 ± 33.
[50] A. E. Reed, P. v. R. Schleyer, J. Am. Chem. Soc. 1990, 112, 1434 ± 1445. [51] J. A. Dobado, H. Martinez-Garcia, J. M. Molina, M. R. Sundberg, J. Am. Chem. Soc. 1998, 120, 8461 ± 8471; J. A. Dobado, H. Martinez-Garcia, J. M. Molina, M. R. Sundberg, J. Am. Chem. Soc. 1999, 121, 3156 ± 3164.
[52] A. Volkov, Y. Abramov, P. Coppens, C. Gatti, Acta Crystallogr. Sect. A 2000, 56, 332 ± 339.
[53] E. Espinosa, C. Lecomte, E. Molins, Chem. Phys. Lett. 1999, 300, 745 ± 748.
[54] Y. A. Abramov, Acta Crystallogr. Sect. A 1997, 53, 264 ± 272.
[55] E. Espinosa, I. Alkorta, I. Rozas, J. Elguero, E. Molins, Chem. Phys. Lett. 2001, 336, 457 ± 461.
[56] S. T. Howard, M. B. Hursthouse, C. W. Lehmann, P. R. Mallinson, C. S. Frampton, J. Chem. Phys. 1992, 97, 5616 ± 5630.
[57] Y. A. Abramov, A. V. Volkov, P. Coppens, Chem. Phys. Lett. 1999, 311, 81 ± 86.
[58] T. Koritsa¬nszky, S. Howard, T. Richter, Z. Su, P. R. Mallinson, N. K. Hansen, XD (Computer Program Package for Multipole Refinement and Analysis of Electron Densities from X-ray Diffraction Data), Free University of Berlin, Germany, 1996.
[59] N. H. Hansen, P. Coppens, Acta Crystallogr. Sect. A 1978, 34, 909 ± 921. [60] N H bond length 1.03 ä; see: F. H. Allen, Acta Crystallogr. Sect. B
1986, 42, 515 ± 522.
[61] M, Souhassou, R. H. Blessing, J. Appl. Crystallogr. 1999, 32, 210 ± 217. [62] a) P. Hohenberg, W. Kohn, Phys. Rev. 1964, 136(B), B864 ± 871; b) W. Kohn, S. J. Sham, Phys. Rev. 1965, 140(A), A1133 ± 1138; c) R. G. Parr, W. Yang, Density-Functional Theory of Atoms and Molecules, Oxford University Press, Oxford, 1989.
[63] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgomery, R. E. Strat-mann, J. C. Burant, S. Dapprich, J. M. Millam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala, Q. Cui, K. Morokuma, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, C. Gonzalez, M. Challacombe, P. M. W. Gill, B. G. Johnson, W. Chen, M. W. Wong, J. L. Andres, M. Head-Gordon, E. S. Replogle, J. A. Pople, Gaussian 98 (Revision A.7), Gaussian Inc., Pittsburgh, PA, 1998.
[64] C. Lee, W. Yang, R. G. Parr, Phys. Rev. B. 1988, 37(2), 785 ± 789. [65] A. D. Becke, J. Chem. Phys. 1993, 98, 5648 ± 5652.
[66] M. M. Francl, W. J. Pietro, W. J. Hehre, J. S. Binkley, M. S. Gordon, D. J. DeFrees, J. A. Pople, J. Chem. Phys. 1982, 77, 3654 ± 3665. [67] V. R. Saunders, R. Dovesi, C. Roetti, M. Causa¬, N. M. Harrison, R.
Orlando, C. M. Zicovich-Wilson, CRYSTAL 98 User×s Manual, Uni-versity of Torino, Italy, 1998.
[68] R. F. W. Bader, F. W. Biegler-Kˆnig, J. R. Cheeseman, J. A. Duke, T. A. Keith, P. Krug, K. E. Laidig, B. Langdon, AIMPAC (a set of programs for the theory of atoms in molecules), McMaster University, Hamilton, ON, Canada, 1994.
[69] D.-C. Fang, T.-H. Tang, AIM98PC (modified PC version of AIMPAC), 1998.
[70] C. Gatti, TOPOND98 User×s Manual 1998, CNR-CSRSRC, Milano, Italy.
[71] R. F. W. Bader, T.-H. Tang, Y. Tal, F. W. Biegler-Kˆnig, J. Am. Chem. Soc. 1982, 104, 946 ± 952.
[72] a) R. F. W. Bader, T. Slee, D. Cremer, E. Kraka, J. Am. Chem. Soc. 1983, 105, 5061 ± 5068; b) S. M. Bachrach, J. Comput. Chem. 1989, 10, 392 ± 406; c) S. M. Bachrach, J. Mol. Struct.(THEOCHEM) 1992, 255, 207 ± 219; d) S. M. Bachrach, J. Mol. Struct.(THEOCHEM) 1995, 337, 201 ± 207.
[73] D. Cremer, E. Kraka, Croatica Chem. Acta 1984, 57, 1259 ± 1281. [74] S. Grimme, J. Am. Chem. Soc. 1996, 118, 1529 ± 1534.
[75] F. W. Biegler-Kˆnig, R. F. W. Bader, T.-H. Tang, J. Comp. Chem. 1982, 3, 317 ± 328.
[76] R. F. W. Bader, P. L. A. Popelier, C. Chang, J. Mol. Struct.(THEO-CHEM) 1992, 255, 145 ± 17.
Received: October 21, 2002 Revised version: March 6, 2003 [F 4519]