Effects of metal co-ordination geometry on self-assembly: A dinuclear double belicate complex and a tetranuclear cage complex of a new bis-bidentate bridging ligand

Article abstract of DOI:10.1039/a909702c

Reaction of 3-(2-pyridyl)pyrazole with 3, 3'-bis(bromomethyl)biphenyl resulted in the new ligand L1 which contains two bidentate chelating pyrazolyl/pyridine fragments separated by a /w/n-biphenyl spacer; this ligand is designed to act only as a bridging ligand, as the two bidentate sites are too far apart to co-ordinate to the same metal ion. The dinuclear copper(n) complex [Cu2(L')2(OAc)2][BF4]2 is a double helicate in which each copper(n) centre is in a square pyramidal co-ordination geometry, arising from two bidentate pyrazolyl/pyridine groups (one from each ligand L1) and a monodentate acetate. The structure is stabilised by extensive inter-ligand Tc-stacking interactions. The complex [Ag2(Ll)2][ClO4]2 is also assumed to be a double helicate. In contrast, reaction with Co afforded the tetranuclear cage complex [Co4(L')6][BF4]8, in which each bridging ligand links two metal centres by spanning one edge of the Co4 tetrahedron. Each metal is therefore in a pseudo-octahedral tris-chelate geometry, with the three bidentate chelating arms each coming from a different ligand L1. Again there is substantial inter-ligand K stacking. Unlike other complexes with the same {M4L6} tetrahedral cage structure, the central cavity is not occupied by a counter ion, showing that although the templating eflect of a counter ion can be beneficial in the assembly of such cages it is clearly not essential. 'H NMR spectroscopy suggests that there is a mixture of species in solution arising from other metal : ligand combinations; B NMR spectroscopy shows that at -40 °C a [BF4]~ anion can become trapped in the cavity of the cage, giving a characteristic high-field resonance in addition to that for the free [BF4]~ anions. Reaction of L1 with Pd afforded a mixture of products arising from ligand decomposition, of which [Pd2(L')(pypz)2][BF4][OH] was structurally characterised. It has a near-planar {Pd2(u-pypz)2}2+ core [Hpypz = 3-(2-pyridyl)pyrazole, which has arisen from decomposition of L1] with an additional bridging ligand L1 co-ordinating in a 'basket-handle' mode, straddling the central core. The Royal Society of Chemistry 2000.

Full text of DOI:10.1039/a909702c

View Article Online / Journal Homepage / Table of Contents for this issue
Effects of metal co-ordination geometry on self-assembly:
a dinuclear double helicate complex and a tetranuclear cage
complex of a new bis-bidentate bridging ligand
Rowena L. Paul, Samantha M. Couchman, John C. Jeffery, Jon A. McCleverty,*
Zoe R. Reeves and Michael D. Ward*
School of Chemistry, University of Bristol, Cantock’s Close, Bristol, UK BS8 1TS.
E-mail: mike.ward@bristol.ac.uk
Received 9th December 1999, Accepted 3rd February 2000
Reaction of 3-(2-pyridyl)pyrazole with 3,3Ј-bis(bromomethyl)biphenyl resulted in the new ligand L1 which contains
two bidentate chelating pyrazolyl/pyridine fragments separated by a meta-biphenyl spacer; this ligand is designed to
act only as a bridging ligand, as the two bidentate sites are too far apart to co-ordinate to the same metal ion. The
dinuclear copper() complex [Cu2(L1)2(OAc)2][BF4]2 is a double helicate in which each copper() centre is in a square
pyramidal co-ordination geometry, arising from two bidentate pyrazolyl/pyridine groups (one from each ligand L1)
and a monodentate acetate. The structure is stabilised by extensive inter-ligand π-stacking interactions. The complex
[Ag2(L1)2][ClO4]2 is also assumed to be a double helicate. In contrast, reaction with CoII afforded the tetranuclear
cage complex [Co4(L1)6][BF4]8, in which each bridging ligand links two metal centres by spanning one edge of the
Co4 tetrahedron. Each metal is therefore in a pseudo-octahedral tris-chelate geometry, with the three bidentate
chelating arms each coming from a different ligand L1. Again there is substantial inter-ligand π stacking. Unlike
other complexes with the same {M4L6} tetrahedral cage structure, the central cavity is not occupied by a counter
ion, showing that although the templating effect of a counter ion can be beneficial in the assembly of such cages it is
clearly not essential. 1H NMR spectroscopy suggests that there is a mixture of species in solution arising from other
metal:ligand combinations; 11B NMR spectroscopy shows that at Ϫ40 ЊC a [BF4]Ϫ anion can become trapped in the
cavity of the cage, giving a characteristic high-field resonance in addition to that for the free [BF4]Ϫ anions. Reaction
of L1 with PdII afforded a mixture of products arising from ligand decomposition, of which [Pd2(L1)(pypz)2][BF4]-
[OH] was structurally characterised. It has a near-planar {Pd2(µ-pypz)2}2ϩ core [Hpypz = 3-(2-pyridyl)pyrazole,
which has arisen from decomposition of L1] with an additional bridging ligand L1 co-ordinating in a ‘basket-handle’
mode, straddling the central core.
The course of a self-assembly reaction between a labile metal
ion and a multidentate bridging ligand depends principally on
the stereoelectronic properties of the metal ion, and the number
and disposition of the binding sites of the bridging ligands.1,2
In some cases the interaction between these is well understood
and a considerable degree of control can be exerted over the
self-assembly process by judicious choice of components, such
that the number, denticity and disposition of the binding sites
in the ligand, and the co-ordination number and geometric
preferences of the metal ions, can precisely be matched to
achieve a remarkable degree of specificity in the self-assembly
process.3 This is exemplified by the recent work of Raymond
and co-workers who have developed a symmetry-based argu-
ment to rationalise and predict the assemblies of some quite
complicated cage structures.4
In one such example we recently investigated the co-
ordination behaviour of the bridging ligand L2, with two biden-
tate pyrazolylpyridine chelates linked by a xylylene spacer.7
We isolated both an open-chain dinuclear complex [Ni2(L2)3]4ϩ,
and a tetranuclear tetrahedral cage [Co4(L2)6(BF4)]7ϩ in which a
[BF4]Ϫ anion is trapped in the central cavity. In both cases there
is the necessary 2:3 ratio of six-co-ordinate metal to tetra-
dentate ligand, but in the latter case the templating effect of the
[BF4]Ϫ anion entirely changes the course of the assembly pro-
cess. Reaction with CuI, which requires a pseudo-tetrahedral
geometry, predictably afforded the dinuclear double helicate
[Cu2(L2)2]2ϩ.9 These three examples neatly illustrate the import-
ance of both the nature of the metal ion and the presence
(or absence) of a counter ion templating effect on the course
of the self-assembly.
Sometimes, however, factors other than the nature of the
metal ion and the ligand need to be taken into account, a par-
ticularly significant one being incorporation of a third com-
ponent, the counter ion, into the structure. Recently there have
been several reports of reactions between particular metal/
ligand combinations where the assembly can proceed along two
different paths, depending on whether or not the counter ion
exerts a templating effect.5–8 Many of these have involved the
reaction of octahedral metal ions with bis-bidentate (four-co-
ordinate) bridging ligands. Such a combination is likely to lead
to a metal:ligand ratio of 2:3 in the absence of any other
strongly co-ordinating species, but more complicated structures
based on cyclic double helicates and tetrahedral cages can occur
in which an anion is trapped in the central cavity having fulfilled
the role of a template during the assembly process.5–8
We describe here the synthesis and co-ordination chemistry
of the new ligand L1, an analogue of L2 but with a more
extended spacer group incorporating two phenyl units instead
of one. In principle all of the structural types seen for the com-
plexes of L2 should still be possible, but a tetrahedral {M4(L1)6}
cage will have a much larger cavity, so that relatively small
anions such as [BF4]Ϫ will not be so effective as templates. We
were interested to see if the assembly of complexes of L1 pro-
ceeded along different lines because of this limitation, and the
results are presented in this paper.
Results and discussion
Synthesis and crystal structure of the ligand L1
The ligand L1 was prepared in 65% yield by reaction of 3,3Ј-
DOI: 10.1039/a909702c
J. Chem. Soc., Dalton Trans., 2000, 845–851
This journal is © The Royal Society of Chemistry 2000
845

View Article Online
Table 1 Selected bond distances (Å) and angles (degrees) for [Cu2(L1)2-
(OAc)2][BF4]2
Cu(1)–O(2)
Cu(1)–N(111)
Cu(1)–N(121)
1.934(4)
2.033(4)
2.042(3)
Cu(1)–N(221)
Cu(1)–N(211)
2.247(4)
2.056(4)
O(2)–Cu(1)–N(111)
N(111)–Cu(1)–N(121) 80.47(14) O(2)–Cu(1)–N(211)
169.57(14) O(2)–Cu(1)–N(121)
94.42(14)
90.5(2)
Fig. 1 Crystal structure of L1.
N(111)–Cu(1)–N(211) 93.7(2)
O(2)–Cu(1)–N(221) 97.1(2)
N(121)–Cu(1)–N(211) 172.2(2)
N(111)–Cu(1)–N(221) 93.20(14)
N(121)–Cu(1)–N(221) 108.37(14) N(211)–Cu(1)–N(221) 76.99(14)
Fig. 2 Crystal structure of the complex cation of [Cu2(L1)2(OAc)2]-
[BF4]2. One ligand strand is shown with hollow bonds for clarity.
[average 2.02 Å] in keeping with the requirements of the Jahn–
Teller effect. The two bridging ligands adopt significantly
different conformations. In ligand 2 [denoted by the first digit
of the numbering scheme, i.e. N(211) and so on, depicted with
hollow bonds in Fig. 2] the central biphenyl unit is involved
with π-stacking interactions with both pyridyl/pyrazolyl co-
ordinating units of ligand 1 (inter-planar separations in the
region of 3.5 Å) to give a ‘three-layer’ stack. In contrast the
central biphenyl unit of ligand 1 is more remote from the cen-
tral core of the complex and is not involved in any inter-
molecular stacking interactions. This is reflected in the torsion
angles between the rings within each biphenyl unit, which are
28.2Њ in ligand 2 (involved in π stacking) and 58.0Њ in ligand 1
(not involved in π stacking).
bis(bromomethyl)biphenyl with 3-(2-pyridyl)pyrazole and base
under phase-transfer conditions, and satisfactorily character-
ised by mass and 1H NMR spectroscopic data (see Experi-
mental section). This is the same method used to prepare
L2 from 1,2-bis(bromomethyl)benzene,7,9 and is a potentially
powerful general method for the preparation of bis-bidentate
bridging ligands with a wide variety of spacers.
X-Ray quality crystals were grown by diffusion of ether
vapour into a concentrated solution of L1 in CH2Cl2; the crystal
structure is in Fig. 1. The molecule lies across an inversion
centre which is located at the midpoint of the C(11)–C(11A)
bond, such that the two halves are crystallographically equiv-
alent. The pyridyl/pyrazolyl units adopt a transoid configur-
ation to minimise unfavourable electronic interactions between
the lone pairs of N(22) and N(32); the inter-ring torsion angle
is 16Њ. The two rings of the central biphenyl unit in contrast
are strictly coplanar. Bond distances and angles within the
molecule are unremarkable. It is clear from this structure
that, in contrast to the shorter ligand L2 which contains just one
phenyl spacer, it will be impossible for both bidentate arms
of L1 to co-ordinate to the same metal ion, as happened on
occasion with L2.7,9 The new ligand L1 can therefore only act
as a bridging ligand.
The fact that L1 supports a double helical architecture sug-
gests that double helicates should also form using CuI and AgI,
which tend to adopt a pseudo-tetrahedral bis-chelate geometry
with this sort of ligand; for example, [Cu2(L2)2]2ϩ is a double
helicate.9 Attempts to isolate a copper() complex of L1 were
unsuccessful, giving only green copper() complexes after
work-up which are presumably similar to [Cu2(L1)2(OAc)2]2ϩ
.
We therefore prepared the complex with AgI, which has the
same structural preference for pseudo-tetrahedral geometry as
CuI but which is not redox-active. Reaction of L1 with AgClO4
in thf afforded a white precipitate of a material analysing as
[Ag(L1)][ClO4]. Its dimeric nature was confirmed by its electro-
spray mass spectrum, which showed a strong peak at m/z 1251
corresponding to {Ag2(L1)2(ClO4)}ϩ. We could not isolate
X-ray quality crystals despite several attempts, but the 2:2
metal:ligand stoichiometry suggests that, like the copper()
complex, the silver() complex is a double helicate.
Double helical complexes with CuII and AgI
Reaction of L1 with copper() acetate hydrate in MeOH (1:1
ratio) afforded a clear green solution from which a light green
solid precipitated on addition of aqueous NaBF4. Electrospray
mass spectrometric data indicated a 2:2 metal:ligand ratio,
with the peak at highest m/z being 1168.7 which corresponds to
{Cu2(L1)2(BF4)2(H2O)}ϩ. X-Ray quality crystals were grown by
slow evaporation of a solution of the complex in MeOH, and
the subsequent crystal structure determination (Fig. 2) showed
the complex to be the dimer [Cu2(L1)2(OAc)2][BF4]2 (see also
Table 1).
The complex is a fairly conventional double helicate, with
both bridging ligands spanning both metal ions. The two
copper() centres, which are crystallographically equivalent as
the complex lies across a C2 axis, are in a basically square
pyramidal five-co-ordinate geometry arising from two bidentate
pyridyl/pyrazolyl chelates and a monodentate acetate ligand.
The axial ligand N(221) is more remote from the metal centre
[Cu(1)–N(221), 2.247(4) Å] than are the four equatorial ligands
The tetranuclear cage [Co4(L1)6][BF4]8
We were next interested to see what type of complex would
form with a metal ion having a preference for regular octa-
hedral geometry. In the absence of any other ligands we expect
a metal:ligand ratio of 2:3, which can be achieved in several
different ways. Triple helicates are well known in which all three
bis-bidentate ligands bridge both metals.10 An alternative struc-
ture with the same stoichiometry is exemplified by [Ni2(L2)3]4ϩ 7
,
in which two ligands are terminal and only one is bridging, i.e.
[(L2)Ni(µ-L2)Ni(L2)]4ϩ; however this is unlikely with L1 which
cannot co-ordinate all donor sites to one metal ion because of
846
J. Chem. Soc., Dalton Trans., 2000, 845–851

View Article Online
approximately tetrahedral cage with one bridging ligand span-
ning each of the six edges of the tetrahedron. Each metal ion is
co-ordinated by a bidentate pyridyl/pyrazolyl unit from each of
three different ligands, and has a distorted pseudo-octahedral
geometry with the Co–N separations in the range 1.96–2.22 Å,
characteristic of high-spin CoII. Each metal centre is therefore
chiral with all four metal ions having the same chirality within
each tetranuclear unit. There are equal numbers of opposite
enantiomers in the unit cell related by an inversion centre. This
structure is topologically equivalent to those of other {M4L6}
tetrahedral cages based on bis-bidentate bridging ligands.4,6,7
The structure is stabilised by multiple aromatic π-stacking
interactions between overlapping adjacent ligand fragments,
which are emphasised in the space-filling view of the complex in
Fig. 4. The Co ؒ ؒ ؒ Co separations vary between 11.49 and 12.27
Å, with the average being 11.88 Å; this is significantly larger
than in [Co4(L2)6(BF4)][BF4]7 where the Co ؒ ؒ ؒ Co separations
are in the range 8.98–10.07 Å.7
The important difference between our earlier complex
[Co4(L2)6(BF4)][BF4]7 and this new example [Co4(L1)6][BF4]8 is
that in the latter there is no [BF4]Ϫ anion trapped in the central
cavity; all of the anions are ‘free’ in the lattice, although there
are close associations between some of the counter ions and the
periphery of the metal complex. This is illustrated in Fig. 3,
which depicts the Co4 tetrahedron with only two of the six
bridging ligands shown for clarity. One [BF4]Ϫ anion is perched
above the triangular face described by Co(1), Co(2) and Co(3),
in a shallow pocket between the three bridging ligands associ-
ated with that face. The result is several CH ؒ ؒ ؒ F contacts in
which the H ؒ ؒ ؒ F separations are around 2.5 Å and the C ؒ ؒ ؒ F
contacts are around 3.2 Å, indicative of weak CH ؒ ؒ ؒ F hydro-
gen bonding which is well known when [BF4]Ϫ anions are
used.11
Fig. 3 Crystal structure of part of the complex cation of [Co4(L1)6]-
[BF4]8ؒ6MeCN, showing (i) the tetrahedral arrangement of metal ions,
(ii) two of the six ligands which bridge the edges of the tetrahedron, and
(iii) one of the [BF4]Ϫ anions which is involved in weak CH ؒ ؒ ؒ F hydro-
gen bonding interactions with the periphery of the complex cation.
It is therefore apparent that the assembly of the [Co4(L1)6]8ϩ
cage is not dictated by an anion-based template effect. This is in
marked contrast to the behaviour of [Co4(L2)6(BF4)]7ϩ where
(i) the cage was only found with an encapsulated anion, which
remained trapped in the central cavity on the NMR timescale;
and (ii) in the absence of a templating effect, an alternative
open-chain structure [Ni2(L2)3]4ϩ was observed. The fact that a
[BF4]Ϫ anion is not observed in the cavity of [Co4(L1)6]8ϩ may be
ascribed partly to the large cavity size, which would make the
anion a much poorer fit and reduce the favourable electrostatic
interaction with the surrounding 8ϩ charge, and partly to the
gap in the centre of each triangular face of the complex which
could allow small guest molecules to drift in and out of the
cavity in solution. In addition, the fact that L1 cannot act as a
tetradentate chelate to one metal ion (unlike L2) prevents the
alternative open-chain structure from forming in which there
is one bridging and two terminal ligands, cf. [Ni2(L2)3]4ϩ. The
complex [Co4(L1)6][BF4]8 is unusual amongst the few examples
of this type.4,6,7 in that (i) it has a large central cavity, and (ii) the
cavity is empty; it does not even contain solvent molecules.4
Although the template effect is clearly operative in formation
of other {M4L6} tetrahedra,6,7 it does not appear to be all-
important.
Fig. 4 Space-filling picture of the complex cation [Co4(L1)6]8ϩ with
each ligand coloured differently, emphasising the inter-ligand
π-stacking interactions and easy access to the central cavity.
the large, relatively rigid spacer between them. A cyclic helicate
with a metal:ligand ratio of 8:12 has been observed.9 A fourth
possibility is a tetrahedral cage having six bridging ligands,
one along each edge of the tetrahedron,4,6,7 as exemplified by
[Co4(L2)6(BF4)][BF4]7.7 In this case the formation of the cage
appeared to be templated by inclusion of a [BF4]Ϫ anion in the
central cavity which was an ideal fit; the larger cavity which
would necessarily occur in a cage structure based on L1 may
diminish the templating action of the anion if the ‘good fit’
criterion is important.
In the area of self-assembly using labile metals ions there is
always the possibility of a mixture of species in solution even if
only one form is isolated in the crystalline state. Accordingly,
we examined the NMR behaviour of this complex in solution
to see to what extent the solution and the solid state structures
Reaction of L1 with cobalt() acetate hydrate in a 3:2 molar
ratio afforded a salmon-pink solution from which a pale orange
solid precipitated on addition of aqueous NaBF4. The electro-
spray mass spectrum shows numerous peaks, with the highest
m/z value of 1784.3 corresponding to {Co2(L1)3(BF4)3}ϩ, indi-
cating formation of at least a dinuclear complex. Diffusion of
ether vapour into a concentrated MeCN solution of the
complex afforded X-ray quality crystals; the structure of the
compound is in Figs. 3 and 4.
1
are comparable. The H NMR spectrum of [Co4(L1)6][BF4]8 in
CD3CN showed at least 42 resonances between δ Ϫ40 and
1
ϩ100. Although the H NMR spectra of high-spin cobalt()
complexes are paramagnetically shifted, it is nonetheless
possible to obtain useful information from them regarding the
number of magnetically inequivalent proton environments.12 If
[Co4(L1)6][BF4]8 were to adopt a regular tetrahedral geometry
in solution on the NMR timescale, we would expect only 12
magnetically inequivalent proton environments. Conversely, if
The complex has the formulation [Co4(L1)6][BF4]8ؒ6MeCN
(although the number of solvents quoted is an approximation
due to extensive disorder, see Experimental section), and is an
J. Chem. Soc., Dalton Trans., 2000, 845–851
847

View Article Online
Fig. 6 Crystal structure of the complex cation of [Pd2(L1)(pypz)2]-
[BF4][OH]ؒ4.5H2OؒMeCN.
Fig. 5 11B NMR spectrum of [Co4(L1)6][BF4]8 at Ϫ40 ЊC in CD3CN
showing the additional resonance corresponding to [BF4]Ϫ trapped
within the cavity of the tetrahedral cage.
electronic preference can be reconciled with the structural
limitations of this ligand.
Reaction of L1 with palladium() acetate (1:1 molar ratio)
in MeOH afforded an intense yellow solution, from which a
yellow solid was isolated on addition of aqueous NaBF4. The
FAB and electrospray mass spectra of this both had their most
abundant peak at m/z 574 with an isotope pattern characteristic
of one Pd atom; this corresponds to the fragment {Pd(L1)}ϩ.
In addition the FAB spectrum showed a weaker peak (8%
it retained in solution a low-symmetry structure like that
observed in the solid state then anything up to 144 magnetically
inequivalent proton environments would occur. Therefore we
have either a mixture of complexes in solution, or the one
cage complex adopting a low-symmetry conformation. The
appearance of this spectrum did not vary significantly with
temperature.
1
abundant) at m/z 680, corresponding to {Pd2(L1)}ϩ. The H
Further insight into the solution behaviour was provided by
11B NMR spectroscopy. For [Co4(L2)6(BF4)][BF4]7 the 11B and
19F resonances of the encapsulated [BF4]Ϫ anion were shifted
substantially upfield compared to those of the free [BF4]Ϫ anion
(11B, δ Ϫ1 and Ϫ105 for free and encapsulated [BF4]Ϫ; 19F,
δ Ϫ150 and Ϫ242 for free and encapsulated [BF4]Ϫ).7 For
[Co4(L1)6][BF4]8 in contrast, at room temperature in CD3CN a
single 11B resonance is seen at δ Ϫ5. On cooling the sample an
additional small signal appears at higher field which by Ϫ40 ЊC
is quite sharp and clear at δ Ϫ34 ppm (Fig. 5). This upfield shift
is consistent with a [BF4]Ϫ ion inside the [Co4(L1)6]8ϩ cavity, and
the temperature dependent behaviour indicates that a dynamic
process is occurring in which a [BF4]Ϫ ion can diffuse into and
out from the cavity, and that this exchange process can be
frozen out. The fast exchange at room temperature is presum-
ably facilitated by the gaps in the centre of each triangular face
of the complex cation which allow access to the central cavity.
This is in contrast with the behaviour of [Co4(L2)6(BF4)]7ϩ for
which the anion remains trapped in the cavity (on the NMR
timescale) at all accessible temperatures. Integration of the two
peaks in Fig. 5 gives a ratio of ca. 1:10 for encapsulated vs. free
[BF4]Ϫ (we used a relaxation delay of 2 seconds when the
spectra were accumulated to ensure that the comparison of the
integrals for the two signals was as accurate as possible). If all
of the complex were present as the tetrahedral cage with no
other metal–ligand combinations present, and every cavity con-
tained an anion, then a ratio of 1:7 would be expected. Allow-
ing for the inevitable uncertainty in the integral values, this
suggests that most of the complex is present in solution at
Ϫ40 ЊC as the tetrahedral cage form [Co4(L1)6]8ϩ with an anion
trapped in the cavity, in interesting contrast to the solid-state
structure.
NMR spectrum of the solid showed a poorly resolved mixture
of overlapping signals in the aromatic region characteristic of a
mixture of products.
Recrystallisation of the crude solid by diffusion of diethyl
ether vapour into a concentrated MeCN solution afforded a
few crystals which were marginally suitable for an X-ray
diffraction study; the result is in Fig. 6. This compound has the
formulation [Pd2(L1)(pypz)2][BF4][OH]ؒ4.5H2OؒMeCN, and is
a dinuclear palladium() complex in which a near-planar
{Pd2(µ-pypz)2}2ϩ core is spanned by one L1 ligand which loops
over the top like a basket handle [Hpypz = 3-(2-pyridyl)-
pyrazole, which has arisen from decomposition of L1]. The
palladium() centres both have their favoured square-planar
co-ordination geometry, which is achieved by having the two
potentially bidentate chelating units of L1 acting as only mono-
dentate ligands. Thus Pd(1) is co-ordinated to the pyridyl donor
N(116) but not significantly to the associated pyrazolyl donor
N(111) which adopts a pseudo-axial position with a long, weak
interaction to the metal ion (2.86 Å). Similarly, Pd(2) co-
ordinates the pyrazolyl donor N(211) of the other potentially
chelating pocket of L1, but the adjacent pyridyl donor N(216)
only forms a very weak pseudo-axial interaction (2.89 Å).
These are in obvious contrast to the four Pd–N bonds in the
square plane which all lie between 1.97 and 2.07 Å. One phenyl
ring from the biphenyl unit of the intact L1 ligand [atoms
C(518)–C(523)] forms a π-stacking interaction with one of the
pypz ligands [atoms N(311)–C(321)] with an average inter-
planar separation of ca. 3.5 Å.
Given the poor quality of this structure (see Experimental
section) there is no point in discussing the minutiae of the struc-
tural parameters in any detail; however the overall structure is
clear and has some interesting features. Principally, it illustrates
a clear example of a self-assembly process being dominated by
the stereoelectronic requirement of the metal. Both the partial
decomposition of L1 to generate pypz units, and only mono-
dentate co-ordination of the potentially bidentate sites of L1,
are driven by the high thermodynamic stability of PdII in a
square-planar ligand field. Both of these types of behaviour
have been seen before in complexes of PdII where the ligand set
appeared to be poorly suited for providing a square-planar co-
A dinuclear complex of PdII arising from partial decomposition
of L1
Having examined the co-ordination behaviour of L1 with
metals having preferences for square pyramidal [CuII], tetra-
hedral [AgI] and octahedral [CoII] geometries we were interested
to use a metal ion having a strong preference for square-planar
co-ordination, viz. PdII; it is not easy to see how this stereo-
848
J. Chem. Soc., Dalton Trans., 2000, 845–851

View Article Online
ordination geometry.13,14 It is obvious from the NMR and mass
spectrometric data that this crystal structure represents but one
component of a mixture of decomposition products. A FAB
mass spectrum recorded on the material after recrystallisation
changed dramatically from that recorded on the freshly pre-
pared solid; the peaks at m/z 574 and 680, corresponding
to {Pd(L1)}ϩ and {Pd2(L1)}ϩ respectively, disappeared to be
replaced by new peaks at m/z 915, 878, 827 and 789. Of these
the peak at m/z 878 corresponds to the fragment {Pd2(L1)-
(pypz)}ϩ, consistent with the crystal structure, but no obvious
assignments can be found for the others.
[Ag2(L1)2][ClO4]2. To a solution of L1 (0.062 g, 0.12 mmol) in
thf (20 cm3) under N2 was added solid AgClO4 (0.035 g, 0.12
mmol). A white precipitate started to form immediately, and
the mixture was left to stir overnight. Filtration of the solid
and drying in vacuo afforded pure [Ag2(L1)2][ClO4]2 (0.060 g,
35%). ESMS: m/z 1251 (3, {Ag2(L1)2(ClO4)}ϩ) and 575 (100%,
{Ag2L2}2ϩ). Found: C, 52.9; H, 3.1; N, 12.3%; C30H24AgClN6O4
requires C, 53.4; H, 3.5; N, 12.5%.
[Co4(L1)6][BF4]8. A solution of L1 (0.070 g, 0.15 mmol) in
CH2Cl2 (10 cm3) was added dropwise to a solution of
Co(MeCO2)2ؒ4H2O (0.025 g, 0.10 mmol) in MeOH (10 cm3).
The resultant salmon-pink solution was left to stir at room
temperature for 1 h whereupon dropwise addition of aqueous
NaBF4 precipitated a pale orange solid which was filtered
off and dried in vacuo. Yield 0.063 g, 20%. ES-MS: m/z
1784.3 (5, {[Co2L3](BF4)3}ϩ), 1316.2 (8, {Co2L2][BF4]3}ϩ),
1160.3 (34, {[Co2L2][BF4]F}ϩ), 1014.7 (56, {CoL2F}ϩ), 546.2
(100, {CoLF}ϩ), 527.5 (5, {CoL}ϩ), 469.3 (62, {LH}ϩ) and
263.6 (55%, {CoL}2ϩ). Found: C, 55.5; H, 3.7; N, 12.4%;
C180H144B8Co4F32N36ؒ6H2O requires C, 56.1; H, 4.1; N,
13.1%. X-Ray quality crystals were grown by slow diffusion of
diethyl ether vapour into a MeCN solution of [Co4L6][BF4]8
at 0 ЊC.
Conclusion
Reaction of the new ligand L1 with various transition-metal
ions afforded a variety of structures in which L1 always acts
as a bridging ligand. Whereas dinuclear double helicates are
observed for the complexes with CuII and AgI, the complex with
CoII is a tetrahedral [Co4(L1)6]8ϩ cage with a large central cavity
into which the [BF4]Ϫ counter ions can diffuse as shown by
variable-temperature 11B NMR spectroscopy. Significantly, this
tetrahedral [Co4(L1)6]8ϩ cage formed even without a strong
anion-templating effect, in marked contrast to [Co4(L2)6-
(BF4)]7ϩ, although it is clearly less robust in consequence as
shown by its fragmentation under mass spectrometric con-
ditions. Reaction of L1 with PdII resulted in decomposition of
the ligand to generate pzpy ligand fragments which allowed PdII
to adopt its preferred square-planar co-ordination geometry.
[Pd2(L1)(pzpy)2][BF4][OH]. A solution of L1 (0.10 g, 0.22
mmol) in CH2Cl2 (10 cm3) was added dropwise to a solution
of Pd(MeCO2)2 (0.050 g, 0.22 mmol) in MeOH (20 cm3). An
intense yellow solution developed which was left to stir at room
temperature for 1 h. An excess of aqueous NaBF4 was then
added dropwise and over a 10 minute period the solution
became cloudy. After reducing the solvent volume by half
a yellow precipitate resulted which was filtered off and dried
in vacuo. Yield 0.18 g. FABMS: m/z 680 (8, {Pd2(L1)}ϩ), 574
(100, {Pd(L1)}ϩ) and 430 (30%, {Pd(L1) Ϫ pypz}ϩ). Recrystal-
lisation of this material by diffusion of ether vapour into a
concentrated MeCN solution of the complex afforded a few
crystals of [Pd2(L1)(pypz)2][BF4][OH]ؒ4.5H2OؒMeCN; the 1H
NMR spectrum of the mother liquor suggested that a mixture
of many other components was present.
Experimental
General details
Instrumentation used for routine spectroscopic studies has been
described previously.2 The starting materials 3-(2-pyridyl)-
pyrazole15 and 3,3Ј-bis(bromomethyl)biphenyl16 were prepared
according to the literature methods.
Syntheses
L1. A mixture of 3,3Ј-bis(bromomethyl)biphenyl (0.91 g, 2.68
mmol), 3-(2-pyridyl)pyrazole (0.86 g, 5.90 mmol), aqueous
NaOH (10 M, 7 cm3) and thf (50 cm3) was heated to reflux with
stirring for 24 h. After cooling the yellow solution was dried
over MgSO4, filtered, and the solvent removed in vacuo to
afford an off-white solid which was shown to be impure by
TLC. The crude solid was dissolved in dichloromethane and
applied to a silica column using ethyl acetate–methanol (99:1)
as eluent to yield the desired product as the third fraction.
White crystals were obtained upon solvent evaporation. Yield:
0.81 g, 65%. EI mass spectrum: m/z 468 (100, Mϩ) and 323
(64%, Mϩ Ϫ Hpypz). 1H NMR (400 MHz, CDCl3): δ 8.65 (1 H,
d, J 4.4, pyridyl H6), 7.97 (1 H, d, J 8.0, pyridyl H3), 7.74 (1 H,
pseudo-t, pyridyl H4), 7.50–7.38 (4 H, m), 7.23–7.21 (2 H, m),
6.97 (1 H, d, J 2.2 Hz, pyrazolyl) and 5.44 (2 H, s, CH2). Found:
C, 76.4; H, 5.2; N, 17.6%; C5H4N requires C, 76.9; H, 5.2; N,
17.9%. X-Ray quality crystals were grown by slow diffusion of
diethyl ether vapour into a CH2Cl2 solution of L1.
X-Ray crystallography
Suitable crystals were quickly transferred from the mother
liquor to a stream of cold N2 on a Siemens SMART diffract-
ometer fitted with a CCD-type area detector. In all cases a
full sphere of data was collected at low temperature using
graphite-monochromatised Mo-Kα radiation.
A detailed
experimental description of the methods used for data collec-
tion and integration using the SMART system has been pub-
lished.17 Empirical absorption corrections were applied using
SADABS,18 and structure solution and refinement was per-
formed with the SHELX suite of programs.19 Table 2 con-
tains a summary of the crystal parameters, data collection and
refinement details.
Crystals of both L1 and [Cu2(L1)2(OAc)2][BF4]2 were stable
and diffracted well; these structural determinations presented
no problems. In L1 the molecule lies astride an inversion
centre such that only half of it is crystallographically independ-
ent. In [Cu2(L1)2(OAc)2][BF4]2 the molecule lies astride a C2
axis such that again only half of it is crystallographically
independent.
[Cu2(L1)2(OAc)2][BF4]2. A solution of L1 (0.050 g, 0.11 mmol)
in CH2Cl2 (10 cm3) was added dropwise to a solution of
Cu(MeCO2)2ؒH2O (0.032 g, 0.071 mmol) in MeOH (10 cm3).
After stirring the mixture at room temperature for 1 h to give
a clear solution, the product was precipitated by addition of
aqueous NaBF4 to give a light green solid which was filtered off
and dried in vacuo. Yield: 0.044 g, 30%. ES-MS: m/z 1168.7
(18, {[Cu2L2][BF4]2ؒH2O}ϩ), 1150.9 (10, {[Cu2L2][BF4]2}ϩ),
550.4 (65, {[Cu2L2]ؒH2O}2ϩ), 531.3 (100, {Cu2L2}2ϩ) and 265.7
(68%, [CuL]2ϩ). Found: C, 56.2; H, 3.5; N, 12.6%; C32H27-
BCuF4N6O2 requires C, 56.7; H, 4.0; N, 12.4%. X-Ray quality
crystals were grown by slow evaporation of a saturated
methanolic solution of the complex.
The complex [Co4(L1)6][BF4]8ؒ6MeCN formed nice-looking
crystals which however lost solvent very fast. After many
attempts a suitable crystal was mounted without too much
decomposition; nevertheless diffraction was weak and only data
with 2θ ≤ 40Њ were used in the final refinement as there was no
significant diffracted intensity at higher angles. Owing to the
weakness of the data and the large number of parameters to
refine, extensive use of geometric restraints was made to keep
J. Chem. Soc., Dalton Trans., 2000, 845–851
849

View Article Online
Table 2 Crystallographic data for the four crystal structures
[Cu2(L1)2(MeCO2)2]-
[BF4]4
[Co4(L1)6][BF4]8ؒ6MeCN
[Pd2(L1)(pypz)2][BF4][OH]ؒ
4.5H2OؒMeCN
L1
Formula
M
C30H24N6
468.55
C64H54B2Cu2F8N12O4
1355.9
C192H162B8Co4F32N42
3987.84
C48H36BF4N13O5.5Pd2
1182.5
T/K
173
173
123
173
System, space group
a/Å
b/Å
c/Å
Monoclinic, P21/c
6.0948(8)
10.967(2)
17.606(3)
94.07(2)
1173.9(3)
2
Monoclinic, C2/c
14.702(3)
20.284(4)
20.669(6)
94.561(11)
6144(3)
Monoclinic, P21/n
19.221(8)
37.42(2)
31.929(9)
90.71(2)
22963(16)
4
0.364
Monoclinic, C2/c
35.916(11)
14.059(4)
26.247(7)
120.314(7)
11441(6)
β/Њ
U/Å3
Z
4
8
µ/mmϪ1
0.081
0.776
0.695
Reflections collected:
total, independent, Rint
Final R1, wR2a
7270, 2679, 0.0242
19097, 6969, 0.0801
73509, 21435, 0.1179
18410, 5338, 0.1766
0.0372, 0.0968
0.0654, 0.1773
0.1576, 0.4853
0.1696, 0.4440
a Structure was refined on Fo2 using all data; the value of R1 is given for comparison with older refinements based on Fo with a typical threshold of
F ≥ 4σ(F ).
the refinement stable. In particular, all aromatic rings (pyridyl,
Acknowledgements
pyrazolyl and phenyl) were constrained to be regular hexagons
We thank the EPSRC for financial support and Dr Martin
or pentagons as appropriate; the bonded B–F distances in the
Murray for assistance with the NMR spectroscopy.
[BF4]Ϫ anions were restrained to be similar to one another, as
were all of the non-bonded F ؒ ؒ ؒ F distances.
Of the [BF4]Ϫ anions, six refined well with unit site occu-
pancy, but three more appeared which refined satisfactorily
References
1 D. Philp and J. F. Stoddart, Angew. Chem., Int. Ed. Engl., 1996, 35,
1155; J.-M. Lehn, Supramolecular Chemistry, VCH, Weinheim,
1995; C. Piguet, G. Bernardinelli and G. Hopfgartner, Chem. Rev.,
1997, 97, 2005.
2 Previous paper in this series: R. L. Paul, A. J. Amoroso, P. L. Jones,
S. M. Couchman, Z. R. Reeves, L. H. Rees, J. C. Jeffery, J. A.
McCleverty and M. D. Ward, J. Chem. Soc., Dalton Trans., 1999,
1563.
3 A. M. Garcia, D. M. Bassani, J.-M. Lehn, G. Baum and D. Fenske,
Chem. Eur. J., 1999, 5, 1234; G. S. Hanan, D. Volmer, U. S.
Schubert, J.-M. Lehn, G. Baum and D. Fenske, Angew. Chem., Int.
Ed. Engl., 1997, 36, 1842; C. Piguet, G. Hopfgartner, A. F. Williams
and J.-C. G. Bünzli, J. Chem. Soc., Chem. Commun., 1995, 2575;
C. Piguet, E. Rivera-Minten, G. Bernardinelli, J.-C. G. Bünzli and
G. Hopfgartner, J. Chem. Soc., Dalton Trans., 1997, 421; M. Fujita,
Chem. Soc. Rev., 1998, 27, 417; N. Takeda, K. Umemoto, K.
Yamaguchi and M. Fujita, Nature (London), 1999, 398, 794;
B. Olenyuk, J. A. Whiteford, A. Fechtenköttner and P. J. Stang,
Nature (London), 1999, 398, 796; P. J. Stang and B. Olenyuk,
Acc. Chem. Res., 1997, 30, 502.
4 T. Beissel, R. E. Powers and K. N. Raymond, Angew. Chem., Int. Ed.
Engl., 1996, 35, 1084.
5 B. Hasenknopf, J.-M. Lehn, G. Baum, B. O. Kneisel and D. Fenske,
Angew. Chem., Int. Ed. Engl., 1996, 35, 1838; R. Vilar, D. M. P.
Mingos, A. J. P. White and D. J. Williams, Angew. Chem., 1998, 37,
1258.
6 D. L. Caulder, R. E. Powers, T. N. Parac and K. N. Raymond,
Angew. Chem., Int. Ed., 1998, 37, 1840; R. W. Saalfrank, R. Burak,
A. Breit, D. Stalke, R. Herbst-Irmer, J. Daub, M. Porsch, E. Bill,
M. Müther and A. X. Trautwein, Angew. Chem., Int. Ed. Engl.,
1994, 33, 1621; M. Scherer, D. L. Caulder, D. W. Johnson and
K. N. Raymond, Angew. Chem., Int. Ed., 1999, 38, 1588.
7 J. S. Fleming, K. L. V. Mann, C.-A. Carraz, E. Psillakis, J. C. Jeffery,
J. A. McCleverty and M. D. Ward, Angew. Chem., Int. Ed., 1998, 37,
1279.
with site occupancies of 0.67 each, to give the required 8 anions.
Three well behaved MeCN molecules could be identified per
complex molecule; in addition, a large collection of electron-
density peaks clearly corresponded to further badly disordered
solvent molecules. These were all refined as carbon atoms with
site occupancies of either 100 or 50% as appropriate to give a
total electron density consistent with the presence of 6 MeCN
molecules, although this is necessarily an approximation. There
is a large apparent void in the lattice of volume ca. 900 Å3 in
which we could only locate a seemingly random collection of
low electron-density peaks (≤1 e ÅϪ3); these could not be made
to look like anything sensible and presumably arise from an
extensively disordered collection of solvent molecules. Use of
the ‘SWAT’ command during refinement to try and account for
the scattering of these was not successful. The atoms of all of
the solvent molecules that were located, and the atoms of the
disordered [BF4]Ϫ anions, were refined with isotropic thermal
parameters. All other atoms were refined anisotropically but
with isotropic restraints. Considering the problems associated
with this structural determination, and the size and complexity
of the problem, the final R1 value of 15.7% is reasonable and
compares well with other refinements on similarly elaborate
complexes. A table of detailed bond distances and angles is
however not included.
Crystals of [Pd2(L1)(pypz)2][BF4][OH]ؒ4.5H2OؒMeCN suf-
fered from very similar problems, arising from poor crystal-
linity, loss of solvent and extensive disorder of solvent
molecules. Again only data with 2θ ≤ 40Њ were used in the final
refinement. The complex dication, one [BF4]Ϫ anion and one
MeCN molecule could be located easily but there was no obvi-
ous sign of the additional anion necessary to achieve charge
balance. A group of closely spaced electron-density peaks was
modelled as a cluster of hydrogen-bonded water molecules, and
it is reasonable to assume that one of these could be deproton-
ated to give a hydroxide ion. It was necessary to apply restraints
to the geometric and thermal parameters (particularly of the
anion and solvent molecules) to ensure stable refinement. A
table of detailed bond distances and angles is not included for
this reason.
8 P. L. Jones, K. J. Byrom, J. C. Jeffery, J. A. McCleverty and M. D.
Ward, Chem. Commun., 1997, 1361.
9 J. S. Fleming, K. L. V. Mann, S. M. Couchman, J. C. Jeffery, J. A.
McCleverty and M. D. Ward, J. Chem. Soc., Dalton Trans., 1998,
2047.
10 P. Baret, D. Gaude, G. Gellon and J.-L. Pierre, New J. Chem., 1997,
21, 1255; M. Albrecht, M. Schneider and R. Frohlich, New J. Chem.,
1998, 22, 753; M. Albrecht and C. Riether, Chem. Ber., 1996, 129,
829.
11 K. L. V. Mann, J. C. Jeffery, J. A. McCleverty and M. D. Ward,
J. Chem. Soc., Dalton Trans., 1998, 3029 and references
therein.
CCDC reference number 186/1843.
See http://www.rsc.org/suppdata/dt/a9/a909702c/ for crystal-
lographic files in .cif format.
12 E. C. Constable, R. Martínez-Máñez, A. M. W. Cargill Thompson
and J. V. Walker, J. Chem. Soc., Dalton Trans., 1994, 1585; E. C.
850
J. Chem. Soc., Dalton Trans., 2000, 845–851

View Article Online
Constable, M. A. M. Daniels, M. G. B. Drew, D. A. Tocher, J. V.
Walker and P. D. Wood, J. Chem. Soc., Dalton Trans., 1993, 1947.
13 M. D. Ward, J. S. Fleming, E. Psillakis, J. C. Jeffery and M. D. Ward,
Acta Crystallogr., Sect. C, 1998, 54, 609.
14 E. C. Constable, S. M. Elder, J. Healy, M. D. Ward and D. A.
Tocher, J. Am. Chem. Soc., 1990, 112, 4590.
15 A. J. Amoroso, A. M. W. Cargill Thompson, J. C. Jeffery, P. L.
Jones, J. A. McCleverty and M. D. Ward, J. Chem. Soc., Chem.
Commun., 1994, 2751; H. Brunner and T. Scheck, Chem. Ber., 1992,
125, 701.
17 P. L. Jones, A. J. Amoroso, J. C. Jeffery, J. A. McCleverty,
E. Psillakis, L. H. Rees and M. D. Ward, Inorg. Chem., 1997, 36, 10;
SMART and SAINT, Area-Detector Control and Integration
Software, Siemens Analytical X-Ray Instruments Inc., Madison,
WI, 1995.
18 SADABS, A program for absorption correction with the Siemens
SMART area-detector system, G. M. Sheldrick, University of
Göttingen, 1996.
19 SHELXTL 5.03 program system, Siemens Analytical X-Ray
Instruments, Madison, WI, 1995.
16 M. Asakawa, P. R. Ashton, S. E. Boyd, C. L. Brown, S. Menzer,
D. Pasini, J. F. Stoddart, M. S. Tolley, A. J. P. White, D. J. Williams
and P. G. Wyatt, Chem. Eur. J., 1997, 3, 463.
Paper a909702c
J. Chem. Soc., Dalton Trans., 2000, 845–851
851