Synthesis, characterization and DFT studies of the cobalt(III) complex of a tetrapodal pentadentate N4S donor ligand

Article abstract of DOI:10.1039/b400710g

The synthesis of the pentadentate ligand 2,6-bis(3,3-dimethyl-2,4- dioxocyclohexanyl)-4-thiaheptane (N₄Samp) is described. The synthetic pathway involves the coupling of two 1,3-(dimethylenedioxy)-2-methyl-2- (methylene-p-toluenesulfonyl)propan

Full text of DOI:10.1039/b400710g

View Article Online / Journal Homepage / Table of Contents for this issue
Synthesis, characterization and DFT studies of the cobalt(III)
complex of a tetrapodal pentadentate N4S donor ligand
Clint A. Sharrad,a Germán E. Cavigliasso,b Robert Strangerb and Lawrence R. Gahan*a
a Chemistry Department, The University of Queensland, Brisbane, QLD 4072, Australia
b Department of Chemistry, Faculty of Science, Australian National University, ACT 0200,
Australia
Received 15th January 2004, Accepted 1st March 2004
First published as an Advance Article on the web 23rd March 2004
The synthesis of the pentadentate ligand 2,6-bis(3,3-dimethyl-2,4-dioxocyclohexanyl)-4-thiaheptane (N4Samp) is
described. The synthetic pathway involves the coupling of two 1,3-(dimethylenedioxy)-2-methyl-2-(methylene-p-
toluenesulfonyl)propane moieties with sodium sulfide and subsequent synthetic elaboration to prepare the final N4S
donor system. The cobalt() complex [Co(N4Samp)Cl]2ϩ has been prepared and subsequently crystallized as the
tetrachlorozincate salt. The X-ray structure analysis confirms the pentadentate nature of the ligand and shows the
thioether donor occupying one apex with four equivalent amine donors effectively occupying the equatorial plane
of the molecule. The sixth coordination site is occupied by a chloro ligand. The electronic absorption and 13C NMR
spectra have been studied. DFT calculations have been employed to explore structural and mechanistic comparisons
between [Co(N4Samp)Cl]2ϩ and an analogous pentaamine complex.
amine occupies the hinge site with four podal primary amine
Introduction
donors.3 Whereas pyN4 and PY5 arose from deliberate syn-
Pentadentate chelators have been of recent interest due to their
thetic strategies, ditame was isolated as a by-product from the
ability to give a stable coordination environment whilst allow-
synthesis of ethylidynetris(methanamine) (tame).3 For pyN4
ing the binding of an additional monodentate ligand resulting
and ditame, the Co() complexes have been isolated with a
in an octahedral structure about the metal cation.1–6 The
chlorine atom occupying the sixth site while the analogous
capacity of these ligands to invoke a single labile coordination
cobalt complex with PY5 stabilises the ϩ oxidation state.1,3,5
site allows the possibility of their complexes to be investigated
We now wish to report an addition to the small group of
as, for example, analogues of metalloenzymes involved in
pentadentate ligands of this type, 2,6-bis(3,3-dimethyl-2,4-
substrate and O2 activation.5,6 Other possible applications
dioxocyclohexanyl)-4-thiaheptane (N4Samp), structurally
include catalyst formation and the synthesis of coordination
analogous to ditame but with a N4S donor set where the
polymers.4 The syntheses of a small number of ligands with a
branched tetrapodal topology have been recently reported with
thioether occupies the hinge position trans to the vacant sixth
coordination site (Chart 1). The presence of a thioether in the
all examples consisting of only N donors (amine, pyridine)
donor set can potentially stabilize lower oxidation states and
(Chart 1).
lower spin states of the coordinating metal ion. Our interest
previously has been in the synthesis, electron transfer proper-
ties, 59Co NMR and visible spectroscopy of mixed donor
nitrogen–thioether cobalt() complexes.7–13 We have also
recently reported a potentially hexadentate N3OS2 ligand which
is able to give pentadentate chelation in a number of forms with
Co().7 N4Samp arises from a deliberate synthetic strategy
utilized previously to prepare a new set of hexadentate ligands
known as amplectors.7,8
Results and discussion
Syntheses
The synthetic strategy is described in Scheme 1. The reaction of
sodium sulfide nonahydrate with 1,3-(dimethylmethylenedioxy)-
2-methyl-2-(methylene-p-toluenesulfonyl)propane resulted in
2,6-bis(3,3-dimethyl-2,4-dioxocyclohexanyl)-4-thiaheptane,
which was converted into the tetraol under dilute acid con-
ditions. Reaction of the tetraol with toluenesulfonyl chloride
and subsequent reaction of the product with potassium
phthalimide in a high-boiling solvent resulted, after removal of
the phthalimide with hydrazine, in the potentially pentadentate
ligand 2,2,6,6-tetra(methyleneamine)-4-thiaheptane (N4Samp).
Subsequent reaction with cobalt() salt and oxygen resulted
in, after chromatographic purification and crystallization using
Chart 1
The ligands 2,6-bis(1Ј,3Ј-diamino-2Ј-methylprop-2Ј-yl)pyrid-
ine (pyN4) and 2,6-(bis(bis-2-pyridyl)methoxymethane)pyridine
(PY5) both contain a pyridine moiety at the hinge position
providing rigidity to the ligand structure.1,2,5,6 The podal donor
groups are all primary amines and pyridine nitrogens for pyN4
and PY5, respectively. In the case of 2,2Ј-dimethyl-2,2Ј-imino-
Ϫ
NaClO4, the Co() complex as a mixed ClϪ/ClO4 salt. The
2Ϫ
compound was subsequently crystallized as the ZnCl4 salt,
dimethylenebis(1,3-propanediamine) (ditame),
a
secondary
[Co(N4Samp)Cl](ZnCl4).
D a l t o n T r a n s . , 2 0 0 4 , 1 1 6 6 – 1 1 7 2
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 4
1166

View Article Online
Table
1
Selected bond distances (Å) and bond angles (Њ) for
[Co(N4Samp)Cl](ZnCl4)
Co(1)–N(1)
Co(1)–N(3)
Co(1)–S(1)
1.973(5)
1.975(5)
2.1701(17)
Co(1)–N(2)
Co(1)–N(4)
Co(1)–Cl(1)
1.974(5)
1.982(5)
2.3247(17)
N(1)–Co(1)–N(2)
N(1)–Co(1)–N(3)
N(1)–Co(1)–N(4)
N(1)–Co(1)–S(1)
N(1)–Co(1)–Cl(1)
N(2)–Co(1)–N(3)
N(2)–Co(1)–N(4)
N(2)–Co(1)–S(1)
86.2(2)
98.7(2)
174.7(2)
84.76(16)
88.63(16)
175.1(2)
88.7(3)
N(2)–Co(1)–Cl(1)
N(3)–Co(1)–N(4)
N(3)–Co(1)–S(1)
N(3)–Co(1)–Cl(1)
N(4)–Co(1)–S(1)
N(4)–Co(1)–Cl(1)
S(1)–Co(1)–Cl(1)
92.57(17)
86.5(2)
83.98(16)
88.38(16)
97.04(17)
90.33(17)
169.02(7)
95.71(17)
substitution of a chloride ligand for a water molecule in the
N4Samp complex therefore shifts the 59Co resonance ∼200 ppm
downfield. A downfield 59Co shift of similar magnitude (370
ppm) is observed upon conversion of [Co(CN)5Cl]3ϩ to
[Co(CN)5(H2O)]2ϩ 14
.
Crystal structure
X-Ray analysis of crystals of the cobalt() complex of
N4Samp as the mixed chloride/perchloate salt indicated mixed
crystal forms with possibly both aqua and chloro forms present
([Co(N4Samp)(H2O)]Cl2(ClO4)ؒH2O and ([Co(N4Samp)Cl]Cl-
(ClO4)ؒ2H2O). Subsequent crystallization from acid solution in
the presence of zinc chloride resulted in the tetrachlorozincate
salt. The structure of [Co(N4Samp)Cl](ZnCl4) (Fig. 1) consists
of the molecular cation and a tetrachlorozincate anion. The
molecular cation has an N4SCl donor set where the amines are
coordinated to the Co() in a planar arrangement and the
chloride ion is coordinated to the cobalt centre at the position
trans to the thioether donor. Each of the six-membered chelate
rings adopt the unsymmetrical skew-boat conformation.
Scheme 1 Synthetic strategy for the formation of N4Samp: (a) Na2S
(0.75 eq.)/ethanol; (b) Hϩ, ethanol; (c) toluenesulfonyl chloride
(4.4 eq.)/pyridine; (d) potassium phthalimide (4.4 eq.)/diglyme/150 ЊC;
(e) hydrazine hydrate (excess)/ethanol/HCl.
NMR Spectroscopy
In D2O, the 13C NMR spectrum of the complex exhibits ten
resonances, suggesting the presence of two coordinated forms
of N4Samp rather than a single asymmetric complex. The
intensity of five signals due to one form of the complex relative
to the other five signals was found to be dependent on pH and/
or ClϪ concentration (both altered with HCl). Although the
two forms of the complex were detected immediately upon
dissolution the equilibrium between these forms was not
attained for several days after pH adjustment. The form
favoured at low pH/high [ClϪ] is suggested to be the chloro
species, [Co(N4Samp)Cl]2ϩ, whilst under neutral conditions
in the absence of added chloride, the aqua species [Co(N4-
Samp)(OD2)]3ϩ, becomes increasingly favoured. The 13C NMR
spectrum of [Co(N4Samp)Cl]2ϩ in d6-DMSO displays five
resonances indicative of a symmetric complex. Similar observ-
ations were obtained using 1H NMR spectroscopy.
The previously reported 13C spectrum of [Co(ditame)Cl]2ϩ in
D2O (pD ca. 5) does not show any ligand exchange process at
the sixth coordination site within the NMR time scale of these
experiments. This suggests that the thioether donor in the
Co() complex of N4Samp tends to labilise the ligand at the
trans position to this sulfur. However, this ligand lability is not
found for all complexes where a chloride ligand is located trans
to a thioether. We have recently reported two such complexes
(endo- and exo-[Co(Et(HO)N3S2amp)Cl]2ϩ) where the NMR
spectra in D2O gives no indication of any ligand exchange
processes occurring for both examples, within the NMR time
scale of these experiments.7
Fig. 1 ORTEP plot of the complex cation of [Co(N4Samp)Cl](ZnCl4),
giving the crystallographic atom numbering. Probability ellipsoids of
30% are shown.
The 59Co NMR spectrum of the Co() complex of N4Samp
in water (pH 5.2) displays a broad resonance at δCo 7077 ppm
and a weaker, broader resonance at δCo 7269 ppm. It is most
likely the resonance at δCo 7077 ppm is due to [Co(N4Samp)-
Cl]2ϩ with a N4SCl chromophore. For pentacyanocobaltate()
complexes, the 59Co resonance when a chloride ligand occupies
the sixth site (δCo 1470 ppm) is shifted slightly downfield in
comparison to the case when an ammine group is bound at the
sixth position (δCo 1155 ppm).14 This suggests that the 59Co
NMR resonance of the N4SCl chromophore should be shifted
slightly downfield relative to N5S chromophores (δCo 6225–6250
ppm).8 The second weak resonance (δCo 7269 ppm) is likely
due to the presence of a small amount of [Co(N4Samp)-
(H2O)]3ϩ as observed in 13C and 1H NMR studies. The
The Co()–amine bond lengths for [Co(N4Samp)Cl]2ϩ
(1.973(5), 1.974(5), 1.975(5), 1.982(5) Å) (Table 1) fall within
the range of expected Co()–N bond lengths (1.94–2.01 Å).15
The Co()–thioether bond length for [Co(N4Samp)Cl]2ϩ
(2.1701(17) Å) falls below the range of Co()–S bond lengths
for similar ligand frameworks (2.194(5)–2.275(3) Å).16–18 The
Co–Cl bond for [Co(N4Samp)Cl]2ϩ (2.3247(17) Å) is slightly
longer than those reported for [Co(pyN4)Cl]2ϩ (2.2645(12) Å),1
[Co(ditame)Cl]2ϩ (2.305(4) Å)2 and a number of other com-
plexes where the chloride group is coordinated trans to an
amine or pyridine nitrogen.18–24 The elongation of the Co–Cl
bond length and the shorter Co–S bond length for [Co(N4-
Samp)Cl]2ϩ suggest that the thioether donor has a weakening
D a l t o n T r a n s . , 2 0 0 4 , 1 1 6 6 – 1 1 7 2
1167

View Article Online
Table 2 Comparison of calculated and experimental structural parameters (distances in Å, angles in Њ) for [Co(N4Samp)Cl]2ϩ and [Co(ditame)Cl]2ϩ
complexes
Complex
Parameter
LDA/TZP
PBE/ZORA/TZ2P
Experiment
[Co(N4Samp)Cl]2ϩ
Co–Cl
Co–S
2.22
2.14
2.25
2.18
2.33
2.17
Co–N
1.94–1.95
171
86–98
174
1.99–2.00
171
86–98
175
1.97–1.98
169
86–99
175
Cl–Co–S
N–Co–N (cis)
N–Co–N (trans)
[Co(ditame)Cl]2ϩ
Co–Cl
Co–N
Cl–Co–N (trans)
N–Co–N (cis)
N–Co–N (trans)
2.20
1.93–1.94
177
87–95
176
2.23
1.98–1.99
176
87–95
176
2.31
1.94–1.97
180
88–92
180
effect on the Co–Cl bond trans to itself. Similar elongation
of the Co–Cl bond length in the complexes endo- and exo-
[Co(Et(HO)N3S2amp)]2ϩ is not observed (2.281(2) and
2.2771(11) Å, respectively).7
(denoted LDA/TZP and PBE/ZORA/TZ2P). Both procedures
have been previously shown8 to satisfactorily reproduce the
experimental structural parameters of Co complexes similar in
size and ligand environment to those investigated in this work.
The calculated values for bond distances and angles involving
the central Co atom are compared in Table 2 with the corre-
sponding experimental results. All calculated parameters are in
good agreement with those observed in the crystal structures,
the largest (and only relatively significant) discrepancies corre-
sponding to the Co–Cl bond length. These discrepancies are,
however, likely to be caused by the fact that the computational
results correspond to isolated (“gas phase”) molecules, whereas
the experimental parameters refer to the crystalline form of
the complexes. Thus, the longer Co–Cl distances observed
experimentally probably reflect the interactions of the Cl site
with counterions in the solid-state environment.
UV-visible spectroscopy
In DMSO the UV-visible absorption spectrum displays one
transition at 17200 cmϪ1 (230 L molϪ1 cmϪ1), the remainder of
the spectrum being obscured by an intense charge transfer
band. At room temperature the Nafion film UV-visible spec-
1
trum of [Co(N4Samp)Cl]2ϩ shows the A1g
1T1g transition
(19860 cmϪ1), with the higher energy 1A1g
1T2g d–d transition
(28340 cmϪ1) clearly defined on a much more intense charge
transfer band at 11 K. Spin forbidden transitions were not
observed at 11 K. Assuming C = 6B, 10Dq was determined as
20810 cmϪ1 with B = 645 cmϪ1 11
.
The LDA/TZP and PBE/ZORA/TZ2P predictions for the
Co–N distances are, respectively, slightly shorter and longer
than the experimental observations, whereas the Co–S bond
length (in [Co(N4Samp)Cl]2ϩ) is better reproduced by the PBE/
ZORA/TZ2P approach, as found in our previous work on Co
complexes. Despite the aforementioned discrepancies between
computational and experimental values for the Co–Cl distance,
the calculations do predict a shortening of this bond in [Co-
Calculations
DFT calculations were used to explore structural and
mechanistic comparisons between [Co(N4Samp)Cl]2ϩ and the
analogous pentaamine complex [Co(ditame)Cl]2ϩ. The results
of these calculations are summarized in Tables 2 and 3, and
Figs. 2–4.
(ditame)Cl]2ϩ with respect to [Co(N4Samp)Cl]2ϩ
.
Molecular structure. The molecular structures of the [Co(N4-
Samp)Cl]2ϩ and [Co(ditame)Cl]2ϩ complexes have been fully
optimized using two different computational approaches
The experimental Cl–Co–S, Cl–Co–N, and N–Co–N angles
are closely reproduced by the calculations, with only minor
differences observed between the results obtained with the two
different approaches. In particular, the computational results
“correctly” predict a relatively significant deviation of the Cl–
Co–S angle (in [Co(N4Samp)Cl]2ϩ) from the “ideal” octahedral
value of 180Њ.
In general, both the LDA/TZP and PBE/ZORA/TZ2P pro-
cedures can be considered to be satisfactory approaches to the
computational investigation of the Co complexes studied in the
present work. However, given that the latter method should
(in principle) provide a more thorough description of physical
and chemical phenomena, all results presented in the remaining
sections are based only on PBE/ZORA/TZ2P calculations.
Electronic structure and bonding. The results of bond valency
calculations on the [Co(N4Samp)Cl]2ϩ and [Co(ditame)Cl]2ϩ
complexes, including Mulliken charge and Mayer covalency
for the Co atoms, and Mayer indexes for the Co–Cl, Co–S, and
Co–N bonds, are given in Table 3.
The calculated values for the charge and covalency of the Co
atoms are in accord with the differences in the ligand environ-
ment of the two complexes, as both results predict a higher
degree of covalent character in the overall bonding interactions
of the metal center in [Co(N4Samp)Cl]2ϩ. Examination of the
individual bond orders suggests that this is largely due to the
Fig. 2 Eigenvalue diagram showing some lowest-unoccupied and
highest-occupied molecular orbital levels for [Co(N4Samp)Cl]2ϩ (R =
N4S) and [Co(ditame)Cl]2ϩ (R = N4N) complexes at (a) fully optimized
geometry and (b) optimized geometry with experimental Co–Cl bond
distance.
greater covalency of the Co–S bond (in [Co(N4Samp)Cl]2ϩ
)
with respect to the (corresponding) Co–N bond (in [Co-
(ditame)Cl]2ϩ).
D a l t o n T r a n s . , 2 0 0 4 , 1 1 6 6 – 1 1 7 2
1168

View Article Online
Fig. 3 Plots of the two lowest-unoccupied and three highest-occupied molecular orbitals of [Co(N4Samp)Cl]2ϩ
.
Table 3 Mulliken charge and covalency index for Co atoms, and
Co–L (L = Cl, S, N) bond order indexes for [Co(N4Samp)Cl]2ϩ and
[Co(ditame)Cl]2ϩ complexes. Results are from [PBE/ZORA/TZ2P]
calculations
hedral environment of the [CoN4SCl] and [CoN4NCl] moieties.
The two lowest-unoccupied and the three highest-occupied
orbitals can be associated, respectively, with the eg and t2g levels
of a regular (transition metal) octahedral system. The gap
between the (highest-occupied) t2g and (lowest-unoccupied) eg
orbitals is predicted to be 18349 cmϪ1 for [Co(N4Samp)Cl]2ϩ
and 19148 cmϪ1 for [Co(ditame)Cl]2ϩ. Plots of these five
orbitals are given in Fig. 3 for [Co(N4Samp)Cl]2ϩ. (The corre-
sponding orbitals in [Co(ditame)Cl]2ϩ are of similar character
and composition.) The splitting due to the low molecular
symmetry is (relatively) small in general, but nonetheless
significantly larger for the t2g than the eg levels. These differences
can be explained by considering the nature of the orbital inter-
actions involved.
Property
[Co(N4Samp)Cl]2ϩ
[Co(ditame)Cl]2ϩ
Co charge
ϩ0.29
4.00
0.79
ϩ0.51
3.61
0.82
Co covalency
Co–Cl index
Co–S index
Co–N index
Co–N index
0.83
0.57
0.54
0.56
The five eg and t2g orbitals exhibit predominant Co d
character, but while both eg orbitals can be described as
σ-antibonding interactions, the three t2g orbitals can be divided
into two groups with distinctly different bonding properties.
2
The eg σ-antibonding interactions occur between Co dz and
“axial” (Cl, S or N) ligand p-type functions, and between Co
2
2
dx Ϫy and “equatorial” (N) ligand p-type functions. The two
highest-lying t2g levels correspond to π-antibonding inter-
actions involving primarily Co dxz or dyz and Cl p-type
functions, whereas the third t2g orbital possesses almost
exclusively Co dxy character with no significant contributions
from the ligands, and is therefore largely nonbonding in nature.
The splitting of the eg levels is only 50–100 cmϪ1 compared
to 3600 cmϪ1 for the t2g levels, but increases significantly
(to 700–800 cmϪ1) if the experimental rather than optimized
Co–Cl distances are used in the calculation. This result can be
related to the longer Co–Cl bond lengths observed in the crystal
structures, which should weaken the Co–Cl σ-antibonding
Fig. 4 Plot of a molecular orbital of [Co(N4Samp)Cl]2ϩ exhibiting S-p
(lone pair)–Co-d bonding character.
The Co–Cl indexes reflect the differences in the respective
bond lengths, but also suggest that the trans influence of the S
atom in [Co(N4Samp)Cl]2ϩ does not appear to be significantly
2
interaction involving the Co dz orbital, thus lowering its energy
2
2
with respect to the Co dx Ϫy orbital.
greater than that of the (“axial”) N atom in [Co(ditame)Cl]2ϩ
.
The two levels lying below the t2g orbitals can be described as
possessing primarily Co d and Cl p character, although in the
[Co(N4Samp)Cl]2ϩ complex some small S p contribution is also
observed. These orbitals mainly involve π-bonding interactions
between Co dxz or dyz and Cl p-type functions.
This result is confirmed by the relative magnitudes of the bond
dissociation energy, obtained from
[Co(N4Samp)Cl]2ϩ
[Co(ditame)Cl]2ϩ
[Co(N4Samp)]3ϩ ϩ ClϪ
[Co(ditame)]3ϩ ϩ ClϪ
(1)
(2)
The calculations on both complexes predict distortions of the
Co coordination environment from ideal octahedral geometry.
Three of the “equatorial” N–Co–N angles are predicted to be
slightly smaller than 90Њ, whereas the remaining angle has
calculated values of 98Њ in [Co(N4Samp)Cl]2ϩ and 95Њ in
[Co(ditame)Cl]2ϩ. In addition, the calculations predict greater
bending of the Cl–Co–S axis in [Co(N4Samp)Cl]2ϩ (at 171Њ)
than the Cl–Co–N axis in [Co(ditame)Cl]2ϩ (at 176–177Њ).
A possible reason for these distortions may be found in the
structural and electronic properties of the “axial” S–R2 and
NH–R2 fragments. Some of the bending of the Cl–Co–S and
Cl–Co–N axes can be related to the structural requirements and
steric effects of the C–H frameworks, as a significantly greater
distortion is observed for the Co–S and axial Co–N bonds than
The calculated values are 1282 and 1303 kJ molϪ1
respectively.
,
Eigenvalue diagrams showing some lowest-unoccupied and
highest-occupied energy levels of the [Co(N4Samp)Cl]2ϩ and
[Co(ditame)Cl]2ϩ complexes are presented in Fig. 2. These
diagrams have been constructed for both the fully optimized
geometry and the partially optimized geometry where the
Co–Cl distance was fixed at the experimental value. Although
the actual molecular symmetry used in the calculations is C1, it
is possible to discuss the general properties of the molecular-
orbital schemes by concentrating on the approximate octa-
D a l t o n T r a n s . , 2 0 0 4 , 1 1 6 6 – 1 1 7 2
1169

View Article Online
for the Co–Cl bonds. Calculations on model [Co(NH3)4-
(SH2)Cl]2ϩ and [Co(NH3)5Cl]2ϩ systems (where the constraints
imposed by the C–H frameworks are removed) indicate that the
bending of the Cl–Co–S and Cl–Co–N axes is 3–4Њ less than in
the [Co(N4Samp)Cl]2ϩ and [Co(ditame)Cl]2ϩ complexes. In [Co-
(NH3)5Cl]2ϩ, the trans Cl–Co–N angle is only slightly bent (the
calculated value being 179Њ), but a relatively significant distor-
carried out at an approximate transition state geometry (corre-
sponding to Co–Cl distances between 4.50 and 5.00 Å). These
calculations suggest that the activation barriers for [Co(N4-
Samp)Cl]2ϩ and [Co(ditame)Cl]2ϩ may be similar but, due to the
approximations involved, no definitive conclusions can be
drawn.
tion of the Cl–Co–S angle remains in [Co(NH3)4(SH2)Cl]2ϩ
.
Conclusion
The larger bending of the Co–S bond in the [Co(N4-
Samp)Cl]2ϩ complex (and also in the model [Co(NH3)4-
(SH2)Cl]2ϩ system) is likely associated with the occurrence of
interactions between S-p (lone pair) and Co-d orbitals (Fig. 4),
which should favour tilting of the Co–S bond.
The distortions observed for the “equatorial” N–Co–N
angles may be necessary to accommodate the structural
requirements of the C–H framework in the pentadentate
ligands, but it is also possible that electronic factors are
involved. The spatial orientation adopted by the S–R2 and
NH–R2 fragments implies that some of the electron density
associated with the S “lone pair” in [Co(N4Samp)Cl]2ϩ or the
N–H bond in [Co(ditame)Cl]2ϩ lies over the plane defined by
the largest “equatorial” N–Co–N angle. Repulsive interactions
between the S or the N–H sites and the “equatorial” Co–N
bonds may be involved in causing the value of this N–Co–N
angle to increase beyond 90Њ.
The versatility of the amplector synthetic methodology, initially
developed for the formation of hexadentate ligands, has been
demonstrated by its capacity to deliberately form a penta-
dentate ligand with a N4S donor set. The ligand N4Samp is
capable of forming octahedral complexes where the sixth
coordination site is positioned trans to the thioether donor. The
lability of this sixth coordination site is shown by the structural
and solution properties of the complex [Co(N4Samp)Cl]2ϩ
.
DFT calculations were performed in order to explore some
structural characteristics and the chloride dissociation proper-
ties of [Co(N4Samp)Cl]2ϩ in comparison to those of the
analogous pentaamine [Co(ditame)Cl]2ϩ complex. The greater
tilting of the Co–S bond (in [Co(N4Samp)Cl]2ϩ) with respect to
the Co–N bond (in [Co(ditame)Cl]2ϩ) can be described as
resulting from the combined effect of steric and electronic
factors. The calculated Co–Cl bond dissociation properties
have not been found to be substantially different. Subsequent
work will investigate both the kinetics of dissociation and the
range of complexes with N4Samp formed by exchanging mono-
dentate and bridging ligands into the sixth coordination site.
Chloride dissociation. The dissociation of the chloro ligand as
a ClϪ ion can be a central step in the mechanism of substitution
reactions in species such as the [Co(ditame)Cl]2ϩ complex.3 As
described in the preceding section, the chloride dissociation
process – represented by eqns. (1) and (2) – in [Co(N4Samp)-
Cl]2ϩ and [Co(ditame)Cl]2ϩ complexes is predicted to be
highly endothermic if calculations correspond to “gas phase”
conditions.
Experimental
Measurements
1H, 13C[1H] and 13C DEPT NMR spectra were recorded at
301 K with a Bruker AC200F 200 MHz or a Bruker AV400
400 MHz spectrometer on internal lock. Chemical shifts for the
1H NMR spectra (CDCl3, d4-methanol and d6-DMSO) are
reported in parts per million (δ) as positive downfield of the
internal reference tetramethylsilane (TMS). In D2O, the 13C
chemical shifts are reported as positive downfield and negative
upfield of the internal reference 1,4-dioxane.7,8,11–13 13C NMR
spectra recorded in CDCl3 and d4-methanol were referenced to
the CDCl3 resonance at 77 ppm and the d4-methanol resonance
at 49 ppm, respectively. In d6-DMSO, the 13C chemical shifts are
reported in parts per million as positive downfield of the
internal reference TMS. For 13C assignments, quaternary and
aromatic carbons are denoted by Cq and Ar, respectively. 59Co
NMR spectra (0.1 M aqueous solutions) were recorded with a
Bruker AV400 400 MHz NMR spectrometer in H2O, without
lock, at 301 K (ν1/2 = resonance line width (Hz) at half-height).
Spectra were externally referenced to [Co(en)3]Cl3 in parts per
million at 7125 ppm. K3[Co(CN)6] (0.1 M) was used as a
secondary external reference at 0 ppm.
Solution UV-visible spectra were recorded with a Perkin-
Elmer Lambda 40 spectrometer. Nafion films (Aldrich Nafion
117 perflourinated membrane 0.0007 in. thick) of the metal
complexes as the mixed chloride perchlorate salt were prepared
by placing the film in dimethylformamide solutions of the
complex for 48 h. The films were removed from the solution and
dried on tissue paper. In order to observe weakly absorbing
bands several films were stacked on one another. UV-visible
spectra of these films were recorded with a Cary 17 spectro-
photometer at room temperature and at ∼14 K, the low temper-
ature spectra were obtained using a Leybold Heraeus ROK
10–300 closed cycle helium cryostat system. Where necessary,
peak positions were determined using Peakfit.25
The thermochemical results change considerably if an
approximate solvation treatment is incorporated into the
computational approach. The calculated values of the dissoci-
ation energy for eqns. (1) and (2) become 71 and 81 kJ molϪ1
respectively, in the presence of the solvent.
,
Sargeson and coworkers have studied the reactivity of the
[Co(ditame)Cl]2ϩ complex in basic solution.3 Under these con-
ditions, the substitution reactions are likely to proceed via a
conjugate base mechanism. We have also carried out calcu-
lations for the chloride dissociation process which involves the
conjugate bases of the [Co(N4Samp)Cl]2ϩ and [Co(ditame)Cl]2ϩ
complexes.
The calculated results for the dissociation energy in the cases
where deprotonation occurs at an “equatorial” (eN) site are
48 kJ molϪ1 for both
[Co(eNHϪ eN3SR)Cl]ϩ
[Co(eNHϪ eN3SR)]2ϩ ϩ ClϪ (3)
and
[Co(eNHϪ eN3 aNR)Cl]ϩ
[Co(eNHϪ eN3 aNR)]2ϩ ϩ ClϪ (4)
For the [Co(ditame)Cl]2ϩ complex, deprotonation can also
occur at the “axial” (aN) site. In this case, the dissociation
energy for
[Co(aNϪ eN4R)Cl]ϩ
[Co(aNϪ eN4R)]2ϩ ϩ ClϪ (5)
is predicted to be 34 kJ molϪ1
.
In addition to exploring the thermodynamics of chloride
dissociation, we have also attempted to calculate the activation
barrier for this process. However, performing a full transition
state optimization, including solvation effects, is a complicated
and costly computational procedure, and we have only been
able to obtain results from a constrained transition state search,
in which a number of single-point calculations have been
Calculation details
All density functional calculations reported in this article were
carried out with the ADF (2002.03) program.26–28 Functionals
D a l t o n T r a n s . , 2 0 0 4 , 1 1 6 6 – 1 1 7 2
1170

View Article Online
based on the Volko–Wilk–Nusair (VWN)29 form of the Local
Density Approximation (LDA),30 and on the gradient-corrected
expressions proposed by Perdew, Burke and Ernzerhof (PBE)31
were utilized. Basis sets of triple-zeta quality and one (TZP) or
two (TZ2P) polarization functions, incorporating frozen cores
(Co 2p, C 1s, N 1s, S 2p, Cl 2p), were employed.26–28 Relativistic
corrections were included through the ZORA approach.32–34
The COSMO model35 was used for the treatment of solvation
effects. Atomic charges and valency indexes36 were obtained
with a program37 designed for their calculation from the ADF
output file. Graphics of molecular orbitals were generated with
the MOLEKEL program.38 Calculations on all complexes
investigated utilized C1 molecular symmetry.
cm3) to precipitate a brown oil. The mixture was allowed to
stand for 24 h and filtered to yield a pale brown solid. The solid
was dissolved in CHCl3 (600 cm3), dried over Na2SO4, filtered
and the solvent removed under reduced pressure to give a pale
brown oil (35.8 g). The product was used without further
purification.
2,2,6,6-Tetra(methyleneamine)-4-thiaheptane
(N4Samp).
2,2,6,6-Tetra(methylenephthalimido)-4-thiaheptane (19.1 g)
was suspended in ethanol (250 cm3) and heated at reflux.
Hydrazine monohydrate (51 cm3, 98%) was added to the
refluxing solution. Over a period of 5 min the solution became
clear then a dense white precipitate formed. The reflux was
maintained for 2 h. The solution was cooled in an ice bath and
concentrated HCl (40 cm3) added dropwise. The mixture was
heated at reflux for a further 40 min, cooled and the solvent
removed under reduced pressure. The residue was dissolved in
water (200 cm3) and the solution filtered. The filtrate was made
strongly alkaline with KOH and the product extracted with
CHCl3 (3 × 100 cm3). The CHCl3 extracts were combined, dried
over Na2SO4, filtered and the solvent removed under reduced
pressure to yield a brown oil (4.1 g). The product was used
without further purification.
Materials
1,3-(Dimethylenedioxy)-2-methyl-2-(methylene-p-toluene-
sulfonyl)propane was prepared as described previously.39
Sodium sulfide nonahydrate was purchased from Aldrich and
used without further purification.
Synthesis of N4Samp
2,2,6-Bis(3,3-dimethyl-2,4-dioxocyclohexanyl)-4-thiaheptane.
Sodium sulfide nonahydrate (57 g, 0.24 mol) was stirred in
ethanol (500 cm3) for 5 min. 1,3-(Dimethylenedioxy)-2-methyl-
2-(methylene-p-toluenesulfonyl)propane (100 g, 0.32 mol) was
added and the solution heated to reflux for 6 h. Upon cooling
the solution was filtered and the solvent removed from the
filtrate under reduced pressure. The residue was dissolved in
CHCl3 (300 cm3) and the solution washed with water (3 × 100
cm3). The organic layer was separated, dried over Na2SO4,
filtered and the solvent removed under reduced pressure to yield
a yellow oil (42.8 g, 56%). 13C NMR (CDCl3): δC 19.1 (–CH3);
20.8, 26.4 (CH3–CqO); 34.2 (Cq); 40.5 (–CH2–S); 67.9 (–CH2–
CAUTION: Although the perchlorate salts described in this
work do not appear to be sensitive to shock or heat these
materials, like all perchlorates, should be treated with caution.
[Co(N4Samp)Cl](ZnCl4). Cobaltous nitrate hexahydrate
(19.5 g) in methanol (500 cm3) was added dropwise to the
stirred mixture of ligand (19.5 g) dissolved in methanol (400
cm3). A stream of air was bubbled through the ligand solution
for the duration of the addition and continued for a further 4 h.
The solvent was removed under reduced pressure, the residue
dissolved in water and the solution filtered. The filtrate was
diluted to 2 L and loaded on Dowex cation exchange resin
(50W × 2 (200–400 mesh) Hϩ form). The column was washed
with water and 1 M HCl to elute minor products. The major
pink/purple band was eluted with 2 M HCl. The solvent was
removed from pink/purple eluent under reduced pressure to
give an impure pink residue. The residue was dissolved in a
minimum volume of aqueous NaClO4 and pink needle like
crystals formed overnight (0.3 g, 0.7%). Analysis of these
crystals indicated a mixed chloride/perchlorate salt of the com-
plex had formed. Analysis. Calc. for C10H26N4SCoCl2(ClO4)ؒ
2H2O: C, 24.03; H, 6.05; N, 11.21. Found: C, 24.16; H, 5.74;
N, 11.10%. The crystals were dissolved in water, acidified
with HCl, and ZnCl2 was added to give an immediate purple
precipitate. The precipitate was filtered and dissolved in a
minimum of water with heating. Purple crystals were grown
from the aqueous solution by vapour diffusion with ethanol.
Analysis. Calc. for [C10H26N4SClCo]ZnCl4: C, 22.41; H, 4.89;
N, 10.46. Found: C, 22.22; H, 4.95; N, 10.21%. UV-visible spec-
trum [λmax /nm (εmax/L molϪ1 cmϪ1) in DMSO]: 580 (230). 13C
NMR (d6-DMSO): δC 24.1 (–CH3); 33.6 (–CH2–S); 40.2 (Cq);
44.3, 45.0 (–CH2–N). 1H NMR (d6-DMSO): δH 0.93 (–CH3, s);
2.75 (–CH2–S, dd); 4.7, 6.0 (–NH2, dd). 59Co NMR (H2O, pH
5.2): δCo 7077 (ν1/2 = 5350 Hz). ESI-MS: Calc. for [Co(N4-
Samp)35Cl]2ϩ Ϫ Hϩ, m/z 327. Found, m/z 327 (91%). Calc. for
[Co(N4Samp)]3ϩ Ϫ 2Hϩ, m/z 291. Found, m/z 291 (100%).
1
O); 97.7 (Cq–O). H NMR (CDCl3): δH 0.89 (–CH3, s); 1.39,
1.41 (CH3–CqO, s); 2.78 (–CH2–S, s); 3.63 (–CH2–O, dd).
2,2,6,6-Tetra(hydroxymethyl)-4-thiaheptane. 2,6-Bis(3,3-di-
methyl-2,4-dioxocyclohexanyl)-4-thiaheptane (42.8 g) was dis-
solved in ethanol (400 cm3) and heated at reflux. Concentrated
HCl (20 cm3) was added and the reflux continued for 10 min.
Upon cooling, the solvent was removed under reduced pressure
to yield a brown oil (36.4 g, quantitative). 13C NMR (d4-
methanol): δC 18.8 (–CH3); 40.3 (–CH2–S); 42.3 (Cq); 66.9
1
(–CH2–O). H NMR (d4-methanol): δH 0.92 (–CH3, s); 2.60
(–CH2–S, s); 3.47 (–CH2–O, dd).
2,2,6.6-Tetra(methylene-p-toluenesulfonyl)-4-thiaheptane.
2,2,6,6-Tetra(hydroxymethyl)-4-thiaheptane (36.4 g) was dis-
solved in dry pyridine (200 cm3) and the solution cooled in an
ice-bath. To the stirred solution, p-toluenesulfonyl chloride (128
g) dissolved in dry pyridine (400 cm3) was added dropwise over
2 h. The reaction mixture was allowed to warm to room tem-
perature and stirring maintained for 48 h. The mixture was
poured into a solution of concentrated HCl (275 cm3), water
(350 cm3) and methanol (700 cm3) to precipitate an off-white
solid which was extracted with CHCl3 (3 × 300 cm3). The
extracts were combined and washed with water (2 × 300 cm3).
The CHCl3 solution was separated, dried over Na2SO4, filtered
and the solvent removed under reduced pressure to yield a
golden oil (101.3 g, 78%). 13C NMR (CDCl3): δC 18.0 (–CH3);
21.6 (–CH3 (tosylate)); 38.7 (–CH2–S); 39.7 (Cq); 71.2 (–CH2–
O); 127.8, 130.0, 132.0, 145.2 (Ar (tosylate)). 1H NMR
(CDCl3): δH 0.85 (–CH3, s); 2.36 (–CH2–S, s); 2.45 (–CH3 (tosyl-
ate), s); 3.77 (–CH2–O, dd); 7.49 (Ar–H (tosylate), dd).
Crystallography
Data collection and processing. For diffractometry the
crystal was mounted onto glass fibres with Supa Glue. Lattice
parameters were determined by least squares fits to the setting
parameters of 25 independent reflections, measured and refined
with an Enraf-Nonius CAD4 diffractometer using graphite-
monochromated Mo-Kα radiation (λ = 0.71073 Å). Formula:
C10H26Cl5CoN4SZn, M = 535.96, monoclinic, space group P21/
c, T = 293(2) K, a = 7.861(2), b = 15.432(2), c = 16.856(2) Å, β =
2,2,6,6-Tetra(methylenephthalimido)-4-thiaheptane. 2,2,6.6-
Tetra(methylene-p-toluenesulfonyl)-4-thiaheptane (39.6 g) and
potassium phthalimide (37.8 g) were suspended in diethylene
glycol dimethyl ether (150 cm3) and the mixture heated at
150 ЊC for 18 h. The cooled solution was poured into water (600
D a l t o n T r a n s . , 2 0 0 4 , 1 1 6 6 – 1 1 7 2
1171

View Article Online
98.64(2)Њ, V = 2021.6(6) Å3, Z = 4, µ = 27.70 cmϪ1, 3550 unique
data, Rint = 0.0488, R(Fo) = 0.0510, Rw = 0.0869.
15 P. Hendry and A. Ludi, Adv. Inorg. Chem., 1990, 35, 117.
16 P. Osvath, A. M. Sargeson, A. McAuley, R. E. Mendelez,
S. Subramanian, M. J. Zaworotko and L. Broge, Inorg. Chem., 1999,
38, 3634.
Structure analysis and refinement. The structure was solved
by heavy-atom methods (direct methods) and refined using
full-matrix least squares on F 2. Programs used were SHELXS-
86 40 for solution, SHELXL-97 41 for refinement and ORTEP-3
for Windows42 for plotting. The geometry of the molecule is
shown in Fig. 1 together with the atomic numbering scheme.
Selected bond lengths and bond angles are given in Table 1.
CCDC reference number 228914.
17 G. J. Grant, S. S. Shoup, C. E. Hadden and D. G. VanDerveer,
Inorg. Chim. Acta., 1998, 274, 192.
18 G. Wei, C. C. Allen, T. W. Hambley, G. A. Lawrance and M. Maeder,
Inorg. Chim. Acta, 1997, 261, 197.
19 G. A. Lawrance, M. Martinez, B. W. Skelton and A. H. White,
J. Chem. Soc., Dalton Trans., 1992, 1649.
20 J. A. Stanko and I. C. Paul, Inorg. Chem., 1967, 6, 487.
21 G. Bombieri, E. Forsellini, A. D. Pra and M. L. Tobe, Inorg. Chim.
Acta, 1980, 40, 71.
See http://www.rsc.org/suppdata/dt/b4/b400710g/ for crystal-
lographic data in CIF or other electronic format.
22 G. Bombieri, E. Forsellini, A. D. Pra and M. L. Tobe, Inorg. Chim.
Acta, 1981, 51, 177.
23 G. Bombieri, E. Forsellini, A. D. Pra, M. L. Tobe and
R. Henderson, Gazz. Chim. Ital., 1979, 109, 207.
24 J. L. Flückiger, C. W. Schläpfer and C. Couldwell, Inorg. Chem.,
1980, 19, 2493.
References
1 A. Grohmann and F. Knoch, Inorg. Chem., 1996, 35, 7932.
2 J. P. López, F. W. Heinemann, R. Prakash, B. A. Hess, O. Horner,
C. Jeandet, J.-L. Oddou, J.-M. Latour and A. Grohmann, Chem.
Eur. J., 2002, 8, 5709.
3 B. Fabius, R. J. Geue, R. G. Hazell, W. G. Jackson, F. K. Larsen,
C. J. Qin and A. M. Sargeson, J. Chem. Soc., Dalton Trans., 1999,
3961.
4 K. Hegetschweiler, O. Maas, A. Zimmer, R. J. Geue, A. M.
Sargeson, J. Harmer, A. Schweiger, I. Buder, G. Schwitzgebel,
V. Reiland and W. Frank, Eur. J. Inorg. Chem., 2002, 41, 4633.
5 R. J. M. K. Gebbink, R. T. Jonas, C. R. Goldsmith and T. D. P.
Stack, Inorg. Chem., 2002, 41, 4633.
25 PeakFit, Version 4, AISN Software Inc., 1991-1995.
26 ADF2002.03, SCM, Theoretical Chemistry, Vrije Universiteit,
Amsterdam, The Netherlands (http://www.scm.com).
27 C. Fonseca Guerra, J. G. Snijders, G. te Velde and E. J. Baerends,
Theor. Chem. Acc., 1998, 99, 391.
28 G. te Velde, F. M. Bickelhaupt, S. J. A. van Gisbergen, C. Fonseca
Guerra, E. J. Baerends, J. G. Snijders and T. Ziegler, J. Comput.
Chem., 2001, 22, 931.
29 S. H. Vosko, L. Wilk and M. Nusair, Can. J. Phys., 1980, 58, 1200.
30 W. Kohn and L. J. Sham, Phys. Rev., 1965, 140, A1133.
31 J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77,
3865.
6 C. R. Goldsmith, R. T. Jonas, A. P. Cole and T. D. P. Stack, Inorg.
Chem., 2002, 41, 4642.
7 C. A. Sharrad, S. R. Lüthi and L. R Gahan, Dalton Trans., 2003,
3693.
32 E. van Lenthe, E. J. Baerends and J. G. Snijders, J. Chem. Phys.,
1993, 99, 4597.
33 E. van Lenthe, E. J. Baerends and J. G. Snijders, J. Chem. Phys.,
1994, 101, 9783.
8 C. A. Sharrad, G. E. Cavigliasso, R. Stranger and L. R. Gahan,
Dalton Trans., 2004, 767.
9 T. M. Donlevy, L. R. Gahan, T. W. Hambley, K. L. McMahon and
R. Stranger, Aust. J. Chem., 1993, 46, 1799.
34 E. van Lenthe, A. E. Ehlers and E. J. Baerends, J. Chem. Phys., 1999,
110, 8943.
35 C. C. Pye and T. Ziegler, Theor. Chem. Acc., 1999, 101, 396.
36 I. Mayer, Chem. Phys. Lett., 1983, 97, 270.
10 L. R. Gahan, T. W. Hambley, A. M. Sargeson and M. R. Snow,
Inorg. Chem., 1982, 21, 2699.
11 L. R. Gahan, T. M. Donlevy and T. W. Hambley, Inorg. Chem.,
1990, 29, 1451.
37 A. J. Bridgeman and C. J. Empson, MAYER, University of Hull,
England, 2002.
38 S. Portmann and H. P. Lüthi, Chimia, 2000, 54, 766.
39 V. W. Gash, J. Org. Chem., 1972, 37, 2197.
12 T. M. Donlevy, L. R. Gahan, T. W. Hambley and R. Stranger, Inorg.
Chem., 1992, 31, 4376.
13 J. I. Bruce, L. R. Gahan, T. W. Hambley and R. Stranger, Inorg.
Chem., 1993, 32, 5997.
14 T. Fujihara and S. Kaizaki, J. Chem. Soc., Dalton Trans., 1993, 1275.
40 A. Altomare, G. Cascorano, C. Giacovazzo and A. Gualardi,
J. Appl. Crystallogr., 1993, 26, 343.
41 G. M. Sheldrick, SHELXL-97, A program for crystal structure
determination, University of Göttingen, 1997.
42 L. J. Farrugia, J. Appl. Crystallogr., 1997, 30, 565.
D a l t o n T r a n s . , 2 0 0 4 , 1 1 6 6 – 1 1 7 2
1172