Structure, characterization, and metal-complexation properties of a new tetraazamacrocycle containing two phenolic pendant arms

Article abstract of DOI:10.1002/hlca.200490235

The new tetraazamacrocycle 2 (=2,2′-[[7-Methyl-3,7,11,17- tetraazabicyclo[11.3.1]heptadeca-1(17),13,15-triene-3,11-diyl]bis(methylene)] bis(4-bromophenol)) was synthesized and used as a ligand for different metal-ion complexes. The X-ray crystal structures of the complexes of the general formula [M(H-2)]⁺NO₃^-·MeOH (M = Ni ²⁺, Zn²⁺), in which only one of the two pendant phenolic OH groups of 2 is deprotonated, were determined. In both complexes, the coordination environment is of the [5 + 1] type, the four N-atoms of the macrocyclic framework defining a square-planar arrangement around the metal center, with similar Ni-N and Zn-N distances of 1.961(9) to 2.157(9) A and 2.021(9) to 2.284(8) A, respectively. In contrast, the M-O distances are markedly different, 2.060(6) and 2.449(8) A in the Ni^II complex, and 2.027(7) and 2.941(9) A in the Zn^II complex. The UV/VIS spectra of the Ni^II and Cu^II complexes with ligand 2, and the EPR spectra of the Cu^II system, suggest the same type of structure for the complexes in solution as in the solid state. Theoretical studies by means of density functional theory (DFT) confirmed the experimental structures of the Ni^II and Zn^II complexes, and led to a proposal of a similar structure for the corresponding Cu^II complex. The calculated EPR parameters for the latter and comparison with related data support this interpretation. The singly occupied molecular orbital (SOMO) in these systems is mainly made of a d orbital of Cu, with a strong antibonding (σ*) contribution of the axially bound phenolate residue.

Full text of DOI:10.1002/hlca.200490235

Helvetica Chimica Acta Vol. 87(2004)
2613
Structure, Characterization, and Metal-Complexation Properties of a
New Tetraazamacrocycle Containing Two Phenolic Pendant Arms
by Xiuling Cuia), Maria Jose¬ Calhordaa)b), Paulo J. Costaa)b), Rita Delgado*a)c), Michael G. B. Drewd),
and VÌtor Fe¬lixe)
a
¬
) Instituto de Tecnologia QuÌmica e Biologica, UNL, Apartado 127, P-2781-901 Oeiras
(phone: ‡351-214 469 737; fax: ‡351-214 411 277; e-mail: delgado@itqb.unl.pt)
b
√
) Departamento de QuÌmica e BioquÌmica, Faculdade de Ciencias, Universidade de Lisboa, P-1649-016 Lisboa
c
¬
) Instituto Superior Tecnico, Av. Rovisco Pais, P-1049-001 Lisboa
d) Department of Chemistry, University of Reading, Whiteknights, Reading RG6 6AD, UK
e) Departamento QuÌmica, CICECO, Universidade de Aveiro, P-3810-193 Aveiro
The new tetraazamacrocycle 2 (ˆ2,2'-[[7-Methyl-3,7,11,17-tetraazabicyclo[11.3.1]heptadeca-1(17),13,15-
triene-3,11-diyl]bis(methylene)]bis(4-bromophenol)) was synthesized and used as a ligand for different metal-
‡
ion complexes. The X-ray crystal structures of the complexes of the general formula [M(H-2)] NOÀ3 ¥ MeOH
(M ˆ Ni2‡, Zn2‡), in which only one of the two pendant phenolic OH groups of 2 is deprotonated, were
determined. In both complexes, the coordination environment is of the [5 ‡ 1] type, the four N-atoms of the
macrocyclic framework defining a square-planar arrangement around the metal center, with similar NiÀN and
ZnÀN distances of 1.961(9) to 2.157(9) ä and 2.021(9) to 2.284(8) ä, respectively. In contrast, the MÀO
distances are markedly different, 2.060(6) and 2.449(8) ä in the NiII complex, and 2.027(7) and 2.941(9) ä in the
ZnII complex. The UV/VIS spectra of the NiII and CuII complexes with ligand 2, and the EPR spectra of the CuII
system, suggest the same type of structure for the complexes in solution as in the solid state. Theoretical studies
by means of density functional theory (DFT) confirmed the experimental structures of the NiII and ZnII
complexes, and led to a proposal of a similar structure for the corresponding CuII complex. The calculated EPR
parameters for the latter and comparison with related data support this interpretation. The singly occupied
molecular orbital (SOMO) in these systems is mainly made of a d orbital of Cu, with a strong antibonding (s*)
contribution of the axially bound phenolate residue.
1. Introduction. Tetraazamacrocycles with two or more ionizable pendant arms
are able to encapsulate a wide range of divalent and trivalent metal ions of various sizes
[1]. The corresponding complexes usually give rise to high stability constants (Ks), and
these ligands have become useful in medicinal chemistry, e.g., in the formulation of
diagnostic agents or in the design of drugs with antitumor properties [2], as well as in
analytical applications [3].
We have recently studied the [14]pyN3 derivative 1 containing two 5-chloro-8-
hydroxyquinoline (CHQ) ligands as pendant arms [4]. The completely deprotonated
mononuclear complexes of 1 with several metal ions are formed only at pH > 8,
indicating the presence of one basic center on the metal complex. The acidity (in terms
II
of logKa) was found to be 8.56, 8.58, 9.17, and 9.52 for the corresponding Cu , ZnII, CdII,
and PbII complexes, which suggests that protonation is affected by both the size of the
metal ions and the stability (logKs) of the mononuclear complexes [4]. Interestingly, an
excess of metal (M) ion relative to the ligand (L) (M/L 2 :1) led to complete
deprotonation of the complex at much lower pH due to the formation of dinuclear
‡
complexes. This suggests that, in this case, H gets trapped by the phenolate O-atom of
¹ 2004 Verlag Helvetica Chimica Acta AG, Z¸rich

2614
Helvetica Chimica Acta Vol. 87(2004)
CHQ, the most-basic center of the ligand. However, it is puzzling to think that only one
of the two phenolates is protonated at high pH.
The aim of the present work was to address the above structural problem by slightly
modifying the macrocyclic ligand and to analyze the behavior of the resulting
complexes. To achieve this, the two CHQ ligands in 1 were replaced by shorter (4-
bromo-2-hydroxyphenyl)methyl moieties, which resulted in formation of macrocycle 2.
Various complexes of 2 were then investigated by UV/VIS, EPR, and X-ray analysis, as
well as by DFT calculations to rationalize the structural aspects and spectroscopic data.
2. Results and Discussion. 2.1. Synthesis. Hydroxylated benzyl substituents are
usually reacted with azacompounds in the form of halomethyl-substituted phenols [5].
However, a modified Mannich reaction, introduced by Lukyanenko et al. [6], was used
to synthesize 2. The parent macrocycle 3 (ˆ 7-methyl-3,7,11,17-tetraazabicyclo[11.3.1]-
heptadeca-1(17),13,15-triene; Me[14]pyN3) was reacted in the presence of both
paraformaldehyde and the corresponding substituted phenol in benzene, in analogy to
[7 9]. Although ortho-substitution in phenols is preferred over para-substitution, the
latter position was blocked with a Br-atom to avoid side reactions. The selection of the
solvent is crucial for the success of the reaction [8]. The previous preparation of the
3,11-bis(methoxymethyl) derivative of 3 and a one-pot reaction were tried, the former
leading to better yield [7]. The resulting compound 2 was subsequently used in the
preparation of the corresponding NiII, ZnII, and CuII complexes.

Helvetica Chimica Acta Vol. 87(2004)
2615
2.2. UV/VIS Spectroscopic Studies. The UV/VIS/near-IR spectral data of [NiII-2]0
and [CuII-2]0 are collected in Table 1. The electronic spectrum observed for the pale
yellow solution of [NiII-2]0 shows six well-defined bands of low intensity, characteristic
of a tetragonal (D4h) symmetry [10][11]. Following the considerations of Busch and co-
workers [11] for a large series of tetraazamacrocycles, we tentatively assigned these
3
bands to the two transitions 3B1g ! B2g, which is directly related to 10Dqxy, and 3B1g !
3Eag, which is equal to the difference between 10Dqxy and 35/(4Dt). Thereby, Dqz is
strongly influenced by the in-plane ligand field, and decreases as Dqxy increases. Values
of the equatorial and axial ligand field were calculated for Dqxy and Dqz as ˆ 1277 and
418 cmÀ1, respectively. The [CuII-2]0 complex exhibited UV/VIS bands at 614 and
430 nm. Upon doubling the amount Cu2‡, the longer-wavelength band was shifted to
644 nm, and the molar absorptivity of both bands increased (Table 1).
Table 1. UV/VIS/Near-IR Spectral Data of Selected NiII and CuII Complexes of the Ligand 2. Recorded at 208 in
MeOH; lmax in nm, e in mÀ1 cmÀ1
.
Complex
Stoichiometry
Color
lmax (e)
[NiII-2]0
Ni/2 1 : 1
Ni/2 2 : 1
Cu/2 1 : 1
Cu/2 2 : 1
Pale yellow
Pale yellow
Turquoise
Green
265 (391.6), 324 (263), 380 (sh, 95.2), 561 (10.3),
783 (11.6), 851 (sh, 7.1), 966 (7.1), 1180 (10)
430 (146), 614 (203)
[CuII-2]0
422 (242), 644 (269)
2.3. X-Ray Single-Crystal Structure Analyses. The X-ray crystal structures of the two
complexes [NiII(H-2)]NO3 ¥ MeOH1) and [ZnII(H-2)]NO3 ¥ MeOH are shown in Fig. 1.
Both compounds were found to be isomorphous. In both cases, the asymmetric unit
comprises one [M(H-2)] cation, one NOÀ3 counter ion, and one MeOH solvent
‡
molecule (crystallization solvent). The charge balances of the molecular formulae for
these two complexes require the protonation of only one donor atom of 2. Indeed, the
X-ray structures unequivocally showed that the pertinent H-atom is bound to the O-
atom of only one of the two pendant arms (see Exper. Part). A H-bond between this
phenolic OH group and the NOÀ3 anion was observed, with an intermolecular OÀH ¥¥¥ N
‡
distance (angle) of 1.91 ä (1408), respectively, for [NiII(H-2)] , and 1.81 ä (1758) for
the corresponding ZnII complex. No other significant intermolecular interactions were
found, suggesting that, except for the H-bonds mentioned, the crystal lattices are
assembled by weak packing interactions.
In both X-ray crystal structures, large anisotropic thermal displacements were
found for the three C-atoms bonded to N(7), suggesting that the macrocycle is
disordered. However, from the Fourier difference maps successively calculated for
‡
‡
[NiII(H-2)] and [ZnII(H-2)] , alternative positions for these atoms were not evident
and, consequently, all attempts to find a disorder model were unsuccessful. Therefore,
the refinement model used reveals only thermal disorder.
1
)
The abbreviation − H-2× means that only one of the two phenolic O-atoms of the pendant arms of the
ligand is deprotonated.

2616
Helvetica Chimica Acta Vol. 87(2004)
‡
‡
Fig. 1. X-Ray crystal structures of a) [NiII(H-2)] NO3À and b) [ZnII(H-2)] NOÀ3 . The ORTEP thermal
ellipsoids are drawn at the 20%-probability level. The OÀH ¥¥¥ N H-bonding interactions between phenolic OH
groups and nitrate anions are marked with dashed lines.
‡
The bond distances and angles of [MII(H-2)] NOÀ3 are listed in Table 2. The data
indicate that the coordination environment is of the [5 ‡ 1] type in both complexes
(M ˆ Ni2‡, Zn2‡). Furthermore, in both cases, the N-atoms of the macrocycle are
arranged in a square-planar geometry around the metal center, with the NiÀN and
ZnÀN distances ranging from 1.961(9) to 2.157(9), and from 2.021(9) to 2.284(8) ä,
respectively. In spite of the identical dispositions of the electron-donating N-atoms in
these two complexes, the distances between the axial O-atoms of the two phenolic
functions and the metal centers are substantially different (2.060(6) and 2.449(8) ä for
M ˆ Ni2‡ vs. 2.027(7) and 2.941(9) ä for M ˆ Zn2‡), which reflects the difference
between the OH (protonated) and OÀ (deprotonated) functions.
The coordination polyhedron of both complexes can be described as an octahedron
strongly distorted along the axial coordination axis. The axial angle OÀNiÀO of
‡
171.5(3)8 deviates by only ca. 88 from the ideal value (1808). In [ZnII(H-2)] NO3À, the
O(28)-atom is also directed towards Zn2‡, but at a much longer distance, suggesting
square pyramidal rather than octahedral geometry, with the O-atom of the second arm
occupying the apex. Furthermore, as would be expected for this type of coordination
polyhedron, Zn2‡ lies above the tetraaza basal plane (0.228(4) ä towards O(48)),
leading to an apical ZnÀO short distance of 2.027(7)8. The distance of this O-atom
from the mean least-squares plane defined by the four basal N-atoms is only
2.308(9) ä. In contrast, in the corresponding NiII complex, the metal protrudes by only
0.089(4) ä from this plane. In the ZnII complex, the metal forms an angle of 168.1(3)8
with the two phenolic O-atoms, which is similar to that found for the NiII complex. In
addition, the C(sp2)ÀOÀM angle is 115.6(6)8 for M ˆ Ni2‡, and 111.4(6)8 for M ˆ

Helvetica Chimica Acta Vol. 87(2004)
2617
Table 2. Selected Bond Lengths and Angles for the Crystallographically Derived Strucutres of Two Metal
Complexes of the Type [MII(H-2)]NO3 ¥ MeOH1). See also Fig. 1.
Bond length [ä]
Bond angle [8]
M ˆ Ni2‡
M ˆ Ni2‡
M ˆ Zn2‡
M ˆ Zn2‡
MÀN(3)
MÀN(7)
MÀN(11)
MÀN(17)
MÀO(48)
MÀO(28)
2.157(9)
2.091(9)
2.144(9)
1.961(9)
2.060(6)
2.449(8)
2.284(8)
2.061(10)
2.199(9)
2.021(9)
2.027(7)
2.941(9)
O(28)ÀMÀO(48)
N(7)ÀMÀN(3)
N(11)ÀMÀN(3)
N(17)ÀMÀN(3)
O(48)ÀMÀN(3)
O(48)ÀMÀN(11)
O(48)ÀMÀN(7)
N(7)ÀMÀN(11)
N(17)ÀMÀN(7)
N(17)ÀMÀN(11)
N(17)ÀMÀO(48)
C(27)ÀO(28)ÀM
171.5(3)
99.9(4)
160.9(3)
80.1(3)
91.3(3)
92.0(3)
94.6(3)
98.6(4)
174.0(3)
81.0(4)
91.4(3)
115.6(6)
168.1(3)
98.5(4)
154.3(3)
77.6(3)
98.4(3)
94.3(3)
95.3(3)
100.7(4)
163.0(4)
79.1(3)
101.7(3)
111.4(6)
Zn2‡, indicating that, in both cases, one lone-pair electron has a suitable orientation to
interact with the metal center.
‡
It is interesting to compare the structure of [NiII(H-2)] NO3À with that found for the
structurally related NiII complex of 4, which contains carboxymethyl substituents in the
3- and 11-positions [13]. The latter displays a distorted octahedral geometry, with four
N-atoms determining a square-planar arrangement around the metal center. The O-
atoms occupy the axial positions, with NiÀO distances of 2.095(5) and 2.090(5) ä,
‡
which is slightly longer than the shortest axial distance reported for [NiII(H-2)] NO3À ;
the OÀNiÀO axial angles, however, are similar in both complexes (171.5 vs. 175.28
(4)). In 4, the macrocycle is also monoprotonated. Unfortunately, the corresponding
Fourier difference map did not reveal the position of this H-atom; however, it is
probably located on one of the free O-atoms of the carboxylate groups.
Three more X-ray single-crystal structures of ZnII complexes of azamacrocycles of
the [14]pyN3 type with pendant arms were found in the Cambridge Structural Data Base
[14]. The first two are the mononuclear ZnII complexes of 5 [15] and 6 [16], which
display distorted square-pyramidal coordination environments, with the four basal N-
atoms defining the equatorial coordination plane. As reported for [ZnII(H-2)] NO3À,
‡
the Zn2‡ ion lies above the equatorial plane, pointing towards the N-atom of the apical
pyridylmethyl arm. The third compound, the mononuclear ZnII complex of 7 [17], also
exhibits a distorted square-pyramidal geometry, but the macrocycle adopts a folded
conformation. Here, atoms N(3), N(7), and N(11) of the macrocycle, together with one
I-atom, define the basal plane of the complex, while the N(7)-atom opposite to the
pyridine ring occupies the apical position.
‡
In complexes [MII(H-2)] NOÀ3 , the NiÀN and ZnÀN distances follow the usual
pattern found for tetraazamacrocycles containing a pyridine moiety, i.e., the MÀN(sp2)
distance is shorter than in MÀN(sp3)-type complexes. In addition, the ZnÀN(sp2) and
ZnÀN(sp3) distances (2.021(9) ä) and 2.181 ä) compare well with those found for the
related complexes of 5 7 [15 17]. The same is true when the NiÀN distances of the
corresponding Ni complexes are compared, the NiÀN(sp2) and NiÀN(sp3) average

2618
Helvetica Chimica Acta Vol. 87(2004)
distances being 1.961(9) and 2.131 ä, respectively, relative to 1.972 ä and 2.102 ä, for
the NiII complex of 4 [13].
‡
‡
In complexes [NiII(H-2)] NO3À and [ZnII(H-2)] NOÀ3 , the metal is encapsulated by
the macrocycle in a planar conformation, with two pendant arms oriented trans to each
other, but in a different overall geometry, as can be seen from Fig. 1. In the NiII complex
(Fig. 1,a), the N-Me group lies below the basal coordination plane, and on the same
side of the pendant arm positioned at the longest distance from the metal, which gives
rise to a (À, À , ‡ ) macrocycle conformation2). In the corresponding ZnII complex, the
N-Me group lies above that plane, leading to a geometric isomer of (À, ‡ , ‡ )
configuration [18]. The fitting of atomic coordinates of the macrocycle (excluding the
H-atoms) retrieved from the two complexes gave an rms (root mean-square) value of
0.30 ä, which could be reduced to 0.10 ä when only the non-H-atoms of the
macrocyclic framework were taken into account. Therefore, in these two complexes,
the macrocycles adopt the same conformation.
Another structural feature of [NiII(H-2)] NO3À and [ZnII(H-2)] NOÀ3 is that, in
both complexes, the pyridine ring is basically −sandwiched× between the two aromatic
rings of the pendant arms. The pyridine ring forms with the phenol ring a dihedral angle
of 37.6(4)8 (M ˆ Ni2‡) and 35.6(4)8 (M ˆ Zn2‡), and the corresponding dihedral angle
between the pyridine and the phenolate ring is reduced to 29.8(4)8 and 27.4(5)8,
respectively. In spite of the different distances between the O-atom donors of the side
arms and the metal center in each complex, they strictly determine the orientation of
the pendant arms.
‡
‡
A comparable situation was found in the related (pyridin-2-yl)methyl-appended
mononuclear CuII complex of 8 [19], which also has a [5 ‡ 1] coordination environ-
ment. Here, the macrocyle adopts a folded conformation, with the N-atom of the
macrocyclic framework trans to its pyridine ring occupying the apical position of a
distorted square pyramid. One of the pendant arms next to the pyridine ring of the
macrocyclic framework is in an equatorial coordination plane, giving rise to a CuÀN
distance of 2.043(7) ä, while the second is roughly parallel to the pyridine ring
(dihedral angle of 18.38). Moreover, the N-atom of the second pendant arm is directed
towards the Cu2‡ center, with a Cu ¥¥¥ N distance of 3.304 ä, which also suggests a weak
bonding interaction consistent with a [5 ‡ 1] coordination. The remaining third
(pyridin-2-yl)methyl pendant arm is not parallel to the macrocyclic pyridine ring and is
further away from the metal center.
2.4. DFT Structural Studies. Several features of the NiII and ZnII complexes are
intriguing, especially their distorted octahedral geometry and the presence of just one
deprotonated phenol moiety in the pendant arms. It is also interesting that the
macrocyclic framework displays different arrangements (configurations) in the ZnII
and NiII complexes, termed A and B, respectively (see Fig. 2, below). To shed light on
these phenomena, density-functional-theory (DFT) calculations [20] were carried out
with the ADF program [21] (see Exper. Part). Full-geometry optimizations were
performed for both Zn and Ni, in both the A and B arrangements (Fig. 2).
2
)
Plus (‡) and minus (À) mean that the N-substituent lies above (‡) and below (À) the N4 plane,
respectively.

Helvetica Chimica Acta Vol. 87(2004)
2619
‡
Fig. 2. DFT-Optimized Structures of the [NiII(H-2)] complex. The complex is drawn for two different ligand
configurations A (left) and B (right). The doubly instead of singly deprotonated system [NiII-2]0 is also shown
(middle)1).
The experimentally observed geometry (X-ray structure) was calculated to be the
most stable in each case, although the differences were very small. Configuration B of
the NiII complex was more stable by 0.2 kcal molÀ1 than configuration A, and
configuration A of the ZnII complex was more stable by 0.6 kcal molÀ1 than B. These
differences are too small to be traced to a specific feature. However, the agreement
between calculated and observed structures is extremely good. Indeed, calculated rms
values (for all non-H-atoms) between X-ray and calculated structures were 0.23 ä for
the ZnII and 0.20 ä for the NiII complexes, respectively. When the same rms calculation
was repeated, taking into account only the metal and its co-ordination sphere (seven
instead of 37atoms), the rms values drop even further to 0.21 ä (Zn) and 0.17ä (Ni),
indicating that the structures diverge more in the periphery of the molecules than in the
central parts. Such an agreement is not too surprising, given the known accuracy of
DFT calculations for these types of fairly rigid complexes, although the size of these
systems is relatively large and the Ni species offers the extra challenge of being
paramagnetic (two unpaired electrons). On the other hand, if we take these same seven
atoms (the metal and six donor atoms), the rms value between the experimental Ni
structure and the one calculated for the Zn arrangement is 0.24 ä, while the opposite
(experimental Zn structure vs. Zn in a DFT-optimized Ni environment) is 0.16 ä. This
means that, although the two arrangements differ, the coordination sphere is barely
affected, as further reflected by very similar calculated relative energies (DE) for both
forms (Table 3).
The MÀN and MÀO distances do not change significantly as one moves from Ni to
Zn (despite the two extra electrons in antibonding orbitals), except for one of the
MÀNcis) and the MÀO(H) bonds3), which are significantly longer in the ZnII complex.
However, these distances remain long when Ni is introduced into the Zn structure
‡
(Table 3, column [NiII(H-2)] , A), probably due to the influence of the Me substituent.
The most-interesting parameters in these complexes are certainly the axial MÀO
bonds. The MÀO(À) bond for the phenolate residue is significantly shorter than the
3
)
Ncis and Ntrans refer to N-atoms cis and trans to the pyridine N-atom, resp.

2620
Helvetica Chimica Acta Vol. 87(2004)
Table 3. Relative Energies (DE, in kcal molÀ1) and Selected Distances [ä] and Angles [8] of the DFT-Optimized
Structures of Different NiII and ZnII Complexes. For configurations A vs. B, see Fig. 2.
‡
‡
‡
Complex
Configuration
DE
[Ni(H-2)]
[NiII-2]0
[Zn(H-2)]
[ZnII-2]0
[ZnII(H-2)] NOÀ3
B
A
B
A
B
A
0.0
1.96
2.11
2.19
2.172.15
2.01
2.93
115.9
107.2
0.16
1.95
2.10
2.24
0.0
0.63
2.03
2.11
2.30
MÀN(py)
1.96
2.14
2.15
2.04
2.11
2.44
2.22
1.95
3.59
113.7
103.1
2.05
2.14
2.25
2.18
2.00
2.173.05
2.02
2.40
2.10
3
MÀNtrans
MÀNcis
MÀNcis
MÀO
)
2.19
2.21
2.23
2.01
3.15
116.6
105.0
2.16
2.13
117.7
115.8
1.972.13
3.42
114.1
104.1
MÀO(H)
CÀOÀM
116.7
118.2
113.2
111.8
CÀO(H)ÀM
MÀO(H) bond, as might be expected, and is shorter in the X-ray structure than in the
DFT-calculated one, which probably reflects the role of packing forces and H-bonds in
the crystal. The phenolic OH group is involved in one H-bond with the neighboring
counter ion (Fig. 1,a). When the geometry was optimized in the presence of the NO3À
ion, the distances did not change much (Table 3), but the Ni ÀO(H) bond was
shortened by 0.5 ä. The most-significant difference arises from the orientation of the
phenyl group involved in the interaction. The new position implies some rearrange-
ments in the peripheral atoms, although the coordination sphere of the metal is barely
changed. The calculated OÀH distance increases from 0.95 ä in the isolated ion to
1.05 ä in the presence of the counter ion. The OÀH ¥¥¥ O contact is 1.489 ä long, with
an O ¥¥¥ O distance of 2.514 ä and an OÀH ¥¥¥ O angle of 1638.
To find an answer to this point, the geometry of the doubly deprotonated species
[MII-2]0 was fully optimized, using the preferred arrangement in each case, for the Ni
(B) and Zn (A) derivatives. The optimized geometry of the [NiII-2]0 complex is also
shown in Fig. 2. It can be seen that the original MÀO(H) bond is shortened
considerably, the two new MÀO(À) bonds now adopting values between 2.10 and
2.15 ä, which is perfectly consistent with a −normal× NiÀO or ZnÀO bond. The rms
values for the DFT-optimized [MII-2]0 structures and the X-ray is 0.23 ä for M ˆ Ni,
and 0.40 ä for M ˆ Zn, when considering all 37non-H-atoms. These values essentially
reflect the approach of the additional phenolate O-atom to the metal, the rest of the
molecule being unperturbed.
2.5. Experimental and Theoretical EPR Studies. The electron-paramagnetic-
resonance (EPR) spectrum of the CuII complex of 2 was recorded in a H2O/DMSO
1:1 mixture. In the case of a 1:1 metal-to-ligand ratio, the spectrum indicated the
presence of two species (A and B), as easily observed by the splitting of the bands at
low field. Taking into account the similarity of this spectrum with that of the CuII
complex of 1 carried out under the same experimental conditions [4], the species A and
‡
B can be attributed to [MII(H-2)] and [MII-2]0 complexes. Each species exhibited
three well-resolved lines at low field, and no superhyperfine splitting. The strong,
unresolved band in the high-field region of the EPR spectrum overlaps the fourth line.
The hyperfine coupling constants and g values of glassy solutions of these complexes,
obtained by simulation of the spectrum [22], are compiled in Table 4, together with

Helvetica Chimica Acta Vol. 87(2004)
2621
Table 4. EPR Data of Different CuII Complexes. The tensors g and A are expressed in terms of 104 cmÀ1
.
Recorded at 102 130 K in DMSO/H2O. For ligands, see chemical formulae.
Complex
lmax [nm]
Species
gx
gy
gz
Ax
Ay
Az
Source
a
[Cu-2]0
614
A
B
2.049
2.050
2.039
2.034
2.032
2.034
2.0272.084
2.068
2.060
2.080
2.060
2.074
2.080
2.229
2.194
2.201
2.188
2.199
2.209
13.6
3.0
4.9
0.5
0.8
13.5
11.3
178.1
163.8
)
)
a
[Cu-3]2‡
[Cu-9]2‡
[Cu-7]2‡
[Cu-10]0
[Cu-4]0
560b)
560
600
614
630
14.7195.9 ] [17
3.4
17.2
8.5
192.9
170.3
167.1
[23]
[17]
[24]
b
)
14.3
2.221
14.9
21.3
165.4
[23]
a) This work. b) In DMF.
those of other CuII complexes based on the same macrocyclic framework, but with
different pendant arms [13][17][23][24]. The parameters of these species are typical of
CuII complexes of an axially elongated rhombic symmetry, and with a dx2Ày2 ground
state, consistent with elongated rhombic-octahedral or distorted square-based pyr-
amidal configurations [25].
From EPR spectroscopy and the ligand-field theory, it is known that the addition of
axial ligands to a square-planar complex decreases Az, but increases gz, with a
simultaneous red-shift in the electronic spectra [26]. In agreement with this, the EPR
spectrum of the square-planar CuII complex of the unsubstituted macrocycle 9, i.e.,
[CuII-9]2‡, for which an X-ray crystal structure is avaliable [18], shows lower gz and
higher Az values, in combination with a blue-shifted band of the maximum VIS
absorption [23]. This is in contrast to the corresponding values for [CuII-7]2‡ and [CuII-
10]0, which exhibit distorted square-pyramidal and octahedral geometries, respectively
[17][24]. The basal planes of the last two complexes are formed by the four N-atoms of
the macrocycle, and the fifth position of [CuII-7]2‡ is occupied by a Cl-atom, while the
six-coordination of [CuII-10]0 is completed via the two O-atoms of the appended
carboxylate groups, as shown by the X-ray structures.
The EPR parameters of both species A and B of the CuII complex of 2 are similar to
those of [Cu-7]2‡ and [Cu-10]0, indicating that both complexes are penta- or hexa-
coordinated. However, the gz values for species A is higher than expected for a square-
pyramidal geometry, and that of species B is lower than expected for an octahedral
environment.
DFT Calculations have been used in recent years to compute EPR parameters,
which can be compared to the experimental data. This is a complementary approach to
the classic ligand-field theory, which does neither directly include covalent bonds nor
allow for mixing of d orbitals, when the symmetry is lowered from the octahedral group
(Oh) [27]. Therefore, an attempt was made to predict the structure of [CuII-2]0, as no
crystal structure was available. Experimental evidence suggested a structure similar to
those of the corresonding NiII and ZnII complexes. Since this type of complex is
paramagnetic, and, because the difference between the Ni and Zn forms is very small,
the geometry was optimized starting from the Ni structure (configuration B). The
‡
optimized geometry of [CuII(H-2)] is shown in Fig. 3,a. Once again, the MÀO(À)
bond is much shorter than the MÀO(H) bond (2.14 vs. 3.20 ä, resp.), while the CuÀN

2622
Helvetica Chimica Acta Vol. 87(2004)
bonds are comparable in size to those in the previous species. This complex, therefore,
adopts a geometry that can be considered a mixture between distorted octahedral and
square pyramidal. The metal does not lie in the plane of the N-atoms, protruding
toward the phenolate side. The unpaired electron in this d9 complex occupies an orbital
of mainly d2z type (SOMO), with a strong contribution from the phenolate arm, as can
be seen in Fig. 3,b.
‡
Fig. 3. a) DFT-Calculated structure of the [CuII(H-2)] complex and b) the corresponding singly occupied
molecular orbital (SOMO)
‡
The calculated spin density of [CuII(H-2)] is concentrated on Cu2‡ (42.1%), with
contributions from the N-atoms (ca. 6% each), the phenolate O-atom (9.2%), and C-
atoms from the adjacent phenyl ring (4.8, 7.2, and 5.6%).
To calibrate our DFT-calculated EPR data, the more-octahedral [CuII-10]0 complex
(see Table 4) was also analyzed based on the crystal-structure coordinates (for details,
see the Exper. Part). Here, the metal is bound to four N-atoms in the equatorial plane,
and to two axial O-atoms from the carboxylate groups (Fig. 4,a), and the experimental
geometry could be nicely reproduced.
Fig. 4. DFT-Calculated structure of a) the [CuII-10]0 complex and b) the corresponding singly occupied
molecular orbital (SOMO)

Helvetica Chimica Acta Vol. 87(2004)
2623
In [CuII-10]0, the spin density is also mainly concentrated on Cu (52.4%), with
contributions from the N-atoms (7.0% for the pyridine N-atom, and 7.4, 9.2, and 10.0%
for the others, resp.), and the carboxylate O-atoms (4.5 and 3.6%). The calculated
singly occupied molecular orbital (SOMO; Fig. 4,b) shows a strong contribution of
Cu(x2 À y2), with some mixing of the z2 orbital, and is s* antibonding toward the axial
O-atoms. The apparent discrepancy between the large contribution of N-atoms to the
spin density and their absence from the SOMO reflects the importance of internal
levels and spin-polarization effects. The latter are also responsible for the splitting of
the levels in a spin-unrestricted calculation when compared to the results of a spin-
‡
restricted one, as shown in Fig. 5 for the above two CuII complexes, i.e., [CuII(H-2)]
and [CuII-10]0.
‡
Fig. 5. Calculated energy diagram of the frontier orbitals of the complexes a) [CuI(H-2)] and b) [CuII-10]0
The electronic-transition energy estimated from the calculated molecular-orbital
(MO) levels, based on one-electron excitations ending at the SOMO, differs strongly
from the experimental results, as has been observed before for other CuII complexes
[27]. However, upon performing TD-DFT calculations with Gaussian98 [28], the most-
intense UV/VIS bands were found to be at 684 and 512 nm, as compared to
experimental values of 614 and 430 nm, respectively (see Table 1). The agreement is
even better when we consider that both bands are equally shifted to higher wave
numbers, which confirms the structural assignment (Fig. 3). The band at 512 nm
corresponds to a pure d ! d transition, while the one at 684 nm arises from a mixture of
d ! d contribution and interligand charge transfers.
‡
The calculated EPR parameters for [CuII(H-2)] and [CuII-10]0 are presented in
Table 5. Some general statements can be drawn: 1) the influence of the basis set on the
results is small (the B1 set is of lower quality than the default one, and B2 is better; in all
other cases, the standard basis set was used; see Exper. Part), and the effect of changing
the substituent R2 at N(7) from Me to H (see chemical formulae) is also negligible. 2)
With respect to the g values, there are significant discrepancies between the calculated
and experimental data, as has been described in other calculations [29]. 3) The patterns
of the EPR A values also exhibit a large deviation, Az being much larger than all the
others for both complexes. The calculated A values usually tend to deviate stronger

2624
Helvetica Chimica Acta Vol. 87(2004)
Table 5. DFT-Calculated and Experimental EPR Parameters of Two CuII Complexes. The tensors g and A are
expressed in terms of 104 cmÀ1
.
Complex
gx
gy
gz
Ax
Ay
Az
Source
a
‡
[CuII(H-2)]
2.049
2.050
2.013
2.010
2.034
1.998
1.998
2.068
2.060
2.0672.106
2.069
2.080
2.229
2.194
13.6
3.0
53.4
2.8
14.3
32.0
29.3
13.5
11.3
178.1
163.8
)
)
)
)
b
c
3.8
95.2
d
2.103
2.209
2.09729.2
2.099
54.8
8.5
96.6
167.1
[CuII-10]0
[24]
c
2.063
2.065
115.5
)
)
d
31.6
115.2
a) This work (A). b) This work (B). c) Basis set B1 (lower quality). d) Basis set B2 (higher quality).
from the experimental values compared to the g-tensor components [30]. These
deviations result from limitations in the ADF program, and it is, thus, not possible to
improve these results.
The solid-state and solution structures of [CuII-10]0 are likely to be very similar.
From the DFT results, its structure cannot be distinguished from the Me derivative
[CuII-4]0. Therefore, when dealing with the CuII complex of 2, the solution species of
configuration B should have a structure very similar to the one of the corresponding
NiII complex, which was used to model the Cu complex, without applying significant
modifications. The coordination geometry is somewhere intermediate between
octahedral and distorted square-pyramidal. In both cases, the SOMO consists mainly
of a d orbital, with a mixed contribution of x2 À y2 plus some mixing of z2 orbitals, but
there are also strong contributions of the axial ligands. Another point to be mentioned
is the contribution of the equatorial N-atoms to the spin density, although their
contribution to the SOMO is negligible. These observations illustrate the limitations of
the ligand-field approach in interpreting EPR spectra.
‡
3. Conclusions. The X-ray crystal structures of the complexes [NiII(H-2)] NO3À
‡
and [Zn(H-2)] NOÀ3 show that the macrocycle encapsulates the metal centers in a
fashion in which one phenol residue is deprotonated (phenolate), whereas the other is
not deprotonated but still points to the Ni2‡ or Zn2‡ ions, which is consistent with EPR
studies in solution. In the solid state, the −free× phenolic OH group is involved in a H-
bond with the NOÀ3 counter-ion. In aqueous solution, H2O molecules can easily replace
the nitrate anion. DFT Calculations have shown that the calculated and experimentally
determined structures generally agree well, the deprotonated complex being also
expected to be stable, with two comparable NiÀO bonds. Moreover, since potentio-
metric studies of various metal complexes of ligand 1 had shown earlier that the
residual phenolic OH group is not deprotonated even at high pH, its acidity must be
intrinsically tuned by its stereoelectronic disposition in the vicinity of the metal ion.
‡
Indeed, the pKa value of [MII(H-1)] complexes follows the trend Pb > Cd > Zn > Cu,
suggesting that the basicity of the coordinated phenolic OH group in complexes with
ligands 1 or 2 is the result of a delicate balance between the ligand-field strength, the
geometric constraints imposed both by the metal and the ligand, and potential H-bonds

Helvetica Chimica Acta Vol. 87(2004)
2625
between this OH group and solvent molecules or counter-ions. All together, these
effects favor a [5 ‡ 1] coordination geometry over a perfect octahedral arrangement.
‡
An interpretation of the EPR spectrum of [CuII(H-2)] was attempted. Full-
geometry optimization of a model based on the structure of the corresponding NiII
complex led to a geometry similar to the one described above, with a coordination
between square-pyramidal and octahedral. The calculated EPR g- and A-tensor
components showed large deviations from those determined experimentally. The
calculation of the related complex [Cu-10]0, with a more-symmetric, tetragonally
distorted octahedral environment (as determined experimentally), allowed the
comparison of calculated and experimental results in another system. The differences
between the two sets of values are similar, and it is not possible to increase the quality
of these calculations. In both cases, the singly occupied molecular orbital (SOMO) is
mainly of the d-type, with a mixed contribution of x2 À y2 and some z2 orbitals, but with
strong antibonding contributions of the axial ligands.
Experimental Part
General. The parent macrocycle 3 (ˆ 7-methyl-3,7,11,17-tetraazabicyclo-[11.3.1]heptadeca-1(17),13,15-
triene) was synthesized by previously reported procedures [31]. All chemicals were of reagent grade and used as
supplied without further purification.
UV/VIS Spectra were recorded on a UNICAM UV-4 or (for UV/VIS/near- IR) on a Shimadzu UV-3100
spectrophotometer in MeOH or DMSO/H2O 1:1; nÄ in cmÀ1. The samples were prepared by the addition of
increasing amounts of metal ion (in the form of nitrate or chloride salts) to the ligands 1 10 at the appropriate
pH value. IR Spectra were recorded from KBr pellets on a UNICAM Mattson-7000 spectrometer. 1H- and
13C-NMR spectra were recorded on a Bruker AMX-300 spectrometer at 300/75 MHz; chemical shifts d are given
in ppm rel. to Me4Si by referencing to residual solvent signals. EPR Spectra were recorded on a Bruker ESP-380
spectrometer equipped with continuous-flow cryostats for liquid N2, operating at X-band. The EPR spectra of
CuII complexes (1 20 mm) were recorded in the range of 102 to 130 K and 7.5 to 32 K, in DMSO/H O mixtures
2
and in 1m tetrabutylammonium nitrate (TBAN) solutions. Computer-simulated EPR spectra were generated
with a program for a micro computer [22]. Microanalyses were carried out by the Instituto de Tecnologia
¬
QuÌmica e Biologica (ITQB) Microanalytical Service.
2,2'-{[7-Methyl-3,7,11,17-tetraazabicyclo[11.3.1]heptadeca-1(17),13,15-triene-3,11-diyl]bis(methylene)}-
bis(4-bromophenol) (2). Compound 3 (0.5 g, 2.0 mmol) [31] was treated with formaldehyde (0.165 g, 5.4 mmol)
in MeOH (10 ml). After standing overnight at r.t., the solvent was evaporated under reduced pressure, and to
the resulting, well-dried intermediate 11 (quant.) was added 4-bromophenol (0.69 g, 4.0 mmol) in benzene
(20 ml). The mixture was refluxed for 21 h, cooled down, and evaporated. The resulting residue was purified by
column chromatography (CC; SiO2; CHCl3/MeOH 20 :1) and recrystallization (CH2Cl2/hexane 2 :1): 0.93 g
(74.7%) of 2. Colorless crystals. M.p. 209 2108. IR (KBr): 3432, 2945, 2850, 2787, 1577, 1479, 1458, 1387, 1269,
1113, 974, 822, 758, 625. 1H-NMR (CDCl3): 1.67( q, 4 H, NCH2CH2CH2N); 2.00 (s, 3 H, MeN); 2.27(br., 4 H,
NCH2CH2); 2.62 (t, 4 H, NCH2CH2); 3.78 (s, 4 H, NCH2); 3.83 (s, 4 H, NCH2); 6.74 (d, J ˆ 8.7, 2 pyridyl H); 7.11
(s, 2 arom. H); 7.13 (d, 2 arom. H); 7.27 (d, 2 arom. H); 7.63 (t, J ˆ 7.6, 1 pyridyl H).13C-NMR (CDCl3; DEPT):
23.3; 43.3; 50.2; 53.7; 57.9; 58.1; 110.4; 118.0; 122.2; 124.4; 131.2; 131.3; 137.2; 156.5; 156.8. Anal. calc. for
C28H34Br2N4O2 ¥ 0.5 H2O: C 53.59, H 5.62, N 8.93; found: C 53.40, H 5.60, N 8.96.
‡
Preparation of the Complex [NiII(H-2)] NO3À ¥ H2O. An aq. soln. of Ni(NO3)2 ¥ 6 H2O (0.039 g, 0.136 mmol)
was added to a stirred soln. of 2 (42 mg, 67.9mmol) dissolved in a minimum (ca. 6 ml) of MeOH/CH2Cl2 1:2. The
mixture was stirred for 2 h and then concentrated to dryness. The residue was taken up in MeOH and filtered.
From the MeOH filtrate, dark-blue crystals were obtained by slow evaporation (2 d) of the solvent in contact
with air. Yield: 90%. Anal. calc. for C28H34Br2N5NiO5 ¥ H2O: C 44.40, H 4.79, N 9.25; found: C 44.37, H 4.68,
N 9.36.
‡
Preparation of the Complex [ZnII(H-2)] NO3À. Prepared as described above for the corresponding NiII
complex, but from Zn(NO3)2 (0.016 g, 55.8 mmol) and 2 (34 mg, 55.6 mmol). Yield: 90%. Yellow crystals.

2626
Helvetica Chimica Acta Vol. 87(2004)
‡
‡
Crystallography. The crystallographic data for [NiII(H-2)] NO3À ¥ MeOH and [ZnII(H-2)] NO3À ¥ MeOH
are collected in Table 6. The data for these two compounds were collected on an MAR research-image-plate
system equipped with graphite-monochromated MoKa radiation (l ˆ 0.71073 ä). A total of 95 frames were
measured at 28 intervals using a counting time adequate to the crystal under study. Data analyses were
performed with the XDS program [32]. Empirical absorption corrections were applied to the intensities by
means of the DIFABS program [33].
Table 6. Room-Temperature Crystallographic Data and Pertinent Refinement Details for the Structures Solved in
This Work
‡
‡
[NiII(H-2)] NO3À ¥ CH3OH
[ZnII(H-2)] NO3À ¥ CH3OH
Formula
Mr
C29H37Br2N5NiO6
770.17
C29H37Br2N5O6Zn
776.83
Crystal system
Space group
Unit cell:
a [ä]
b [ä]
c [ä]
Monoclinic
P21/c
Monoclinic
P21/c
10.774(14)
22.75(3)
13.339(15)
100.06(1)
3219(7)
10.870(17)
22.83(3)
12.983(15)
97.99(1)
3191(7)
4
b [8]
V [ä3]
Z
4
Dc [mg mÀ3
]
1.589
3.134
1.617
3.323
m [mmÀ1
]
Reflections collected
Unique reflections, [Rint
R1, wR2 (I > 2s(I))
R1, wR2 (all data)
1980719659
5883, [0.1593]
0.0959, 0.2028
0.2051, 0.2521
]
6176, [0.0628]
0.0924, 0.2780
0.1314, 0.2951
‡
The merge of the intensity data of the [NiII(H-2)] complex in the Laue symmetry group 2/m led to a higher
Rint value of 0.1593, suggesting that the crystal used in the data collection had a lower degree of crystallinity. In
fact, several crystals of the NiII complex were investigated, and the image-plate-picture patterns showed that all
of them had fairly poor and weak diffraction patterns. In spite of this experimental drawback, the structure was
solved unambiguously, and the quality of the final data was comparable to that for the corresponding ZnII
complex.
‡
‡
The structures of [NiII(H-2)] and [ZnII(H-2)] were solved by direct methods and refined by full-matrix
least-squares analysis against F2, using SHELXS and SHELXL from the SHELX97package [34]. All non-H-
atoms were refined with anisotropic thermal parameters. The H-atoms bonded to C-atoms were inserted in
geometric positions, and those attached to O-atoms were localized according to Fourier difference maps. The
OH H-atoms were introduced in the refinement, assuming OÀH distances and HÀO ¥¥¥ H angles constrained to
0.82 ä and 104.58, resp. All atomic H-atom positions were refined, giving rise to an isotropic thermal parameter
1.2 times those of the atoms to which they were bonded. The structures were refined convergence was achieved,
leading to R values listed in Table 6. The largest peaks and holes in the final difference Fourier maps were 0.684
and À0.355 eäÀ3 for the NiII complex, and 0.960 and À0.521 eäÀ3 for the Zn analogue, which is within the
values expected.
DFT Calculations. DFT Calculations [20] were performed with the Amsterdam-density-functional program
package (ADF) [21]. The local spin density (LSD) exchange correlation potential was used with the local
density approximation of the correlation energy [35]. Full-geometry optimizations, without any symmetry
constraints, were performed by means of the generalized gradient approximation [36], using Perdew Wang
exchange and correlation corrections [37]. Spin-unrestricted calculations were performed for all the CuII and
NiII paramagnetic complexes studied. The inner shells of Ni2‡, Cu2‡, and Zn2‡ ([1 2]s, 2p); C, N, and O (1s); Br
([1 3]s, [1 3]p) were frozen. An uncontracted triple-z STO basis set, with one polarization function, was used
to describe the valence shells of all elements. For the EPR calculations, all electron basis sets, consisting of
uncontracted triple-z STO functions, augmented by one polarization function, were used for all elements (unless

Helvetica Chimica Acta Vol. 87(2004)
2627
otherwise mentioned). The basis set B1 consisted of a double-z STO basis with one polarization function; for the
B2 basis set, all electron basis sets with uncontracted triple-z STO functions, and augmented by two polarization
functions, were considered. The ZORA method was used to account for relativistic effects and spinÀorbit
coupling [38]. The EPR A values were obtained from an unrestricted calculation, and the g values from a spin-
restricted calculation with spinÀorbit correction.
The starting geometries were taken from the crystal structures of the complexes with the ligand 2 (described
in this work) and the published structure of [CuII-10]0 [24]. Graphical representations of molecular orbitals were
drawn with MOLEKEL [39]. TD-DFT [40] Calculations were performed with the Gaussian98 package [28] on
‡
the ADF-optimized structure of [CuII(H-2)] . The unrestricted B3LYP [41] formalism was adopted (10 states
were requested). The standard LANL2DZ [42] basis set along with the associated ECP was used for Cu and Br,
a standard 6-31G(d) basis set [43] was used for C, O and N, and the 3-21G basis set [44] was taken for H.
√
R. D. and V. F. acknowledge financial support from FundaÁaƒo para a Ciencia e Tecnologia (FCT) and
POCTI, with co-participation of the EU fund FEDER (Project No. POCTI/1999/QUIM/35396). X. C. and P. J.
C. acknowledge FCT for the grants SFRH/BPD/1502/ 2000 and SFRH/BD/10535/2002, resp.
REFERENCES
[1] K. P. Wainwright, Coord. Chem. Rev. 1997, 166, 35; T. A. Kaden, Adv. Supramol. Chem. 1993, 3, 65.
[2] S. Liu, D. S. Edwards, Bioconjugate Chem. 2001, 12, 7; A. Heppeler, S. Froidevaux, A. N. Eberle, H. R.
Maecke, Curr. Med. Chem. 2000, 7, 971; C. J. Anderson, M. J. Welch, Chem. Rev. 1999, 99, 2219.
[3] J. S. Bradshaw, R. M. Izatt, Acc. Chem. Res. 1997, 30, 338; S. C. Lee, R. M. Izatt, X. X. Zhang, E. G. Nelson,
J. D. Lamb, P. B. Savage, J. S. Bradshaw, Inorg. Chim. Acta 2001, 317, 174.
[4] X. Cui, M. F. Cabral, J. Costa, R. Delgado, Inorg. Chim. Acta 2003, 356, 133.
[5] U. Auerbach, U. Eckert, K. Wieghardt, B. Nuber, J. Weiss, Inorg. Chem. 1990, 29, 938.
[6] N. G. Lukyanenko, V. N. Pastushok, A. V. Bordunov, Synthesis 1991, 241; A. V. Bordunov, N. G.
Lukyanenko, V. N. Pastushok, K. E. Krakowiak, J. S. Bradshaw, N. K. Dalley, X. Kou, J. Org. Chem.
1995, 60, 4912.
[7] X. X. Zhang, A. V. Bordunov, J. S. Bradshaw, N. K. Dalley, X. Kou, R. M. Izatt, J. Am. Chem. Soc. 1995,
117, 11507.
[8] K.-W. Chi, H.-C. Wei, T. Kottke, R. J. Lagow, J. Org. Chem. 1996, 61, 5684.
[9] A. V. Bordunov, J. S. Bradshaw, X. X. Zhang, N. K. Dalley, X. Kou, R. M. Izatt, Inorg. Chem. 1996, 35,
7229.
[10] A. B. P. Lever, −Inorganic Electronic Spectroscopy×, 2nd edn., Elsevier, Amsterdam, 1984; R. W. Renfrew,
R. S. Jamison, D. C. Weatherburn, Inorg. Chem. 1979, 18, 1584; R. Smierciak, J. Passariello, E. L. Blinn,
Inorg. Chem. 1977, 16, 2646.
[11] L. Y. Martin, C. R. Sperati, D. H. Busch, J. Am. Chem. Soc. 1977, 99, 2968.
[12] L. Spek, −PLATON, a Multipurpose Crystallographic Tool×, Utrecht University, Utrecht, The Nederlands,
1999.
¬
[13] J. Costa, R. Delgado, M. G. B. Drew, V. Felix, R. T. Henriques, J. C. Waerenborgh, J. Chem. Soc., Dalton
Trans. 1999, 3253.
[14] F. H. Allen, Acta Cryst., Sect. B 2002, 58, 380; I. J. Bruno, J. C. Cole, P. R. Edginton, M. Kessler, C. F.
Macrae, P. McCabe, J. Pearson, R. Taylor, Acta Cryst., Sect. B 2002, 58, 389.
[15] S. J. Grant, P. Moore, H. A. A. A. Omar, N. W. Alcock, J. Chem. Soc., Dalton Trans. 1994, 485.
[16] N. W. Alcock, K. P. Balakrishnan, A. Berry, P. Moore, C. J. Reader, J. Chem. Soc., Dalton Trans. 1988, 1089.
¬
[17] J. Costa, R. Delgado, M. G. B. Drew, V. Felix, A. Saint-Maurice, J. Chem. Soc., Dalton Trans. 2000, 1907.
¬
[18] V. Felix, M. J. Calhorda, J. Costa, R. Delgado, C. Brito, M. T. Duarte, T. Arcos, M. G. B. Drew, J. Chem.
Soc., Dalton Trans. 1996, 4543.
¬
[19] J. Costa, R. Delgado, M. G. B. Drew, V. Felix, J. Chem. Soc., Dalton Trans. 1999, 4331.
[20] R. G. Parr, W. Yang, −Density Functional Theory of Atoms and Molecules×, Oxford University Press, New
York, 1989.
¬
[21] E. J. Baerends, A. Berces, C. Bo, P. M. Boerrigter, L. Cavallo, L. Deng, R. M. Dickson, D. E. Ellis, L. Fan,
T. H. Fischer, C. Fonseca Guerra, S. J. A. van Gisbergen, J. A. Groeneveld, O. V. Gritsenko, F. E. Harris, P.
van den Hoek, H. Jacobsen, G. van Kessel, F. Kootstra, E. van Lenthe, V. P. Osinga, P. H. T. Philipsen, D.
Post, C. C. Pye, W. Ravenek, P. Ros, P. R. T. Schipper, G. Schreckenbach, J. G. Snijders, M. Sola, D.
Swerhone, G. te Velde, P. Vernooijs, L. Versluis, O. Visser, E. van Wezenbeek, G. Wiesenekker, S. K. Wolff,

2628
Helvetica Chimica Acta Vol. 87(2004)
T. K. Woo, T. Ziegler, ADF-2002 program (see www.scf.com); E. J. Baerends, D. Ellis, P. Ros, Chem. Phys.
1973, 2, 41; E. J. Baerends, P. Ros, Int. J. Quantum Chem. 1978, S12, 169; P. M. Boerrigter, G. te Velde, E. J.
Baerends, Int. J. Quantum Chem. 1988, 33, 87; G. te Velde, E. J. Baerends, J. Comput. Phys. 1992, 99, 84.
[22] F. Neese, Diploma Thesis, University of Konstanz, 1993.
[23] J. Costa, R. Delgado, M. C. Figueira, R. T. Henriques, M. Teixeira, J. Chem. Soc., Dalton Trans. 1997, 65.
¬
[24] J. Costa, R. Delgado, M. G. B. Drew, V. Felix, J. Chem. Soc., Dalton Trans. 1998, 1063.
[25] B. J. Hathaway, Coord. Chem. Rev. 1983, 52, 87; H. Yokoi, M. Sai, T. Isobe, S. Ohsawa, Bull. Chem. Soc. Jpn.
1972, 45, 2189.
[26] P. W. Lau, W. C. Lin, J. Inorg. Nucl. Chem. 1975, 37, 2389; M. J. Maroney, N. J. Rose, Inorg. Chem. 1984, 23,
2252.
[27] R. Deeth, J. Chem. Soc., Dalton Trans. 2001, 664.
[28] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski,
J. A. Montgomery Jr., R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Millam, A. D. Daniels, K. N. Kudin,
M. C. Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C. Adamo, S.
Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala, Q. Cui, K. Morokuma, N. Rega, P. Salvador, J. J.
Dannenberg, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz, A. G.
Baboul, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R. L. Martin, D. J.
Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson,
W. Chen, M. W. Wong, J. L. Andres, C. Gonzalez, M. Head-Gordon, E. S. Replogle, J. A. Pople, Gaussian
98, Revision A.11.3, Gaussian, Inc., Pittsburgh PA, 2002.
[29] M. Stein, E. van Lenthe, E. J. Baerends, W. Lubitz, J. Phys. Chem., A 2001, 105, 416.
[30] P. J. Carl, S. L. Isley, S. C. Larsen, J. Phys. Chem., A 2001, 105, 4563; S. C. Larsen, J. Phys. Chem., A 2001,
105, 8333.
[31] K. P. Balakrishnan, H. A. A. Omar, P. Moore, N. W. Alcock, G. A. Pike, J. Chem. Soc., Dalton Trans. 1990,
2965; J. Costa, R. Delgado, Inorg. Chem. 1993, 32, 5257.
[32] W. Kabsch, J. Appl. Cryst. 1988, 21, 916.
[33] G. M. Sheldrick, Acta Crystallogr., Sect. A 1990, 46, 467; G. M. Sheldrick, SHELX-97, University of
Gˆttingen, 1997.
[34] S. H. Vosko, L. Wilk, M. Nusair, Can. J. Phys. 1980, 58, 1200.
[35] N. Walker, D. Stuart, Acta Crystallogr., Sect. A 1983, 39, 158.
[36] L. Versluis, T. Ziegler, J. Chem. Phys. 1988, 88, 322; L. Fan, T. Ziegler, J. Chem. Phys. 1991, 95, 7401.
[37] J. P. Perdew, J. A. Chevary, S. H. Vosko, K. A. Jackson, M. R. Pederson, D. J. Singh, C. Fiolhais, Phys. Rev.,
B 1992, 46, 6671.
[38] E. van Lenthe, A. Ehlers, E.-J. Baerends, J. Chem. Phys. 1999, 110, 8943.
[39] S. Portmann, H. P. L¸thi, Chimia 2000, 54, 766.
[40] R. E. Stratmann, G. E. Scuseria, M. J. Frisch, J. Chem. Phys. 1998, 109, 8218; R. Bauernschmitt, R.
Ahlrichs, Chem. Phys. Lett. 1996, 256, 454; M. E. Casida, C. Jamorski, K. C. Casida, D. R. Salahub, J. Chem.
Phys. 1998, 108, 4439.
[41] C. Lee, W. Yang, R. G. Parr, Phys. Rev., B 1988, 37, 785; A. D. Becke, J. Chem. Phys. 1993, 98, 5648.
[42] T. H. Dunning Jr., P. J. Hay, in −Modern Theoretical Chemistry×, Ed. H. F. Shaefer III, Plenum Press, New
York, 1977, Vol. 3, p. 1 27; P. J. Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 270; P. J. Hay, W. R. Wadt, J.
Chem. Phys. 1985, 82, 299.
[43] R. Ditchfield, W. J. Hehre, J. A. Pople, J. Chem. Phys. 1971, 54, 724; W. J. Hehre, J. A. Ditchfield, J. A.
Pople, J. Chem. Phys. 1972, 56, 2257; P. C. Hariharan, J. A. Pople, Theor. Chim. Acta 1973, 28, 213; P. C.
Hariharan, J. A. Pople, Mol. Phys. 1974, 27, 209; M. S. Gordon, Chem. Phys. Lett. 1980, 76, 163.
[44] J. S. Binkley, J. A. Pople, W. J. Hehre, J. Am. Chem. Soc. 1980, 102, 939.
Received February 17, 2004