Metal-directed assembly of combinatorial libraries - Principles and establishment of equilibrated libraries with oligopyridine ligands

Article abstract of DOI:10.1039/b707601k

The cobalt(ii) complexes [Co(bpy)₃][PF₆]₂, [Co(Me₂bpy)₃][PF₆]₂ (Me ₂bpy = 4,4′-dimethyl-2,2′-bipyridine) and [Co(phen) ₃][PF₆]₂ give paramagnetically shifted ¹H NMR spectra which may be fully assigned; the complexes are labile and ligand exchange is complete within mixing time in CD₃CN solutions to give libraries of heteroleptic complexes which have been fully characterised by one- and two-dimensional ¹H NMR spectroscopy. A library comprising mer and fac isomers of [CoL₃]²⁺ (L = 2,2′-bipyridine-5-carbaldehyde) can be amplified by specific reaction of the fac stereoisomer with a triamine to give a new hexadentate ligand, although other ligand exchange processes compete. The Royal Society of Chemistry and the Centre National de la Recherche Scientifique.

Full text of DOI:10.1039/b707601k

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/njc | New Journal of Chemistry
Metal-directed assembly of combinatorial libraries—principles and
establishment of equilibrated libraries with oligopyridine ligands
Barbara Brisig, Edwin C. Constable* and Catherine E. Housecroft
Received (in Montpellier, France) 21st May 2007, Accepted 30th May 2007
First published as an Advance Article on the web 19th June 2007
DOI: 10.1039/b707601k
0
0
The cobalt(II) complexes [Co(bpy) ][PF ] , [Co(Me bpy) ][PF ] (Me bpy = 4,4 -dimethyl-2,2 -
2
3
6 2
2
3
6 2
1
bipyridine) and [Co(phen) ][PF ] give paramagnetically shifted H NMR spectra which may be
3
6 2
fully assigned; the complexes are labile and ligand exchange is complete within mixing time in
CD CN solutions to give libraries of heteroleptic complexes which have been fully characterised
by one- and two-dimensional H NMR spectroscopy. A library comprising mer and fac isomers
3
1
2
+
0
of [CoL
3
]
(L = 2,2 -bipyridine-5-carbaldehyde) can be amplified by specific reaction of the fac
stereoisomer with a triamine to give a new hexadentate ligand, although other ligand exchange
processes compete.
reactions is predicated upon the development and optimisa-
Introduction
tion of reactions which proceed rapidly and in high yield for as
wide a variety of substrates as possible. These constraints have
suggested an alternative approach in which traditional assem-
bly of components through covalent carbon–carbon or car-
bon–heteroatom bond formation is replaced by assembly of
suitably functionalised ligands at metal centres. This approach
has been elegantly demonstrated by Hamilton and co-workers,
Metal-directed self assembly of ligands around metal centres is
a powerful methodology in supramolecular chemistry that
allows the preparation of numerous topologically and topo-
graphically novel compounds in which the spatial arrangement
of organic components is determined both by the coordination
number and coordination geometry of the metal and the
number, type and distribution of coordination sites within
2+
2+ 26
who used the assembly of [Fe(terpy)
+ 27
2
]
or [Ru(terpy)
2
]
,
1
–10
the ligands.
Both labile and kinetically inert metal centres
and [Cu(phen)
2
] , motifs from functionalised terpy or phen
may be used for the construction of such molecular architec-
6
tures and the latter, usually d transition metal ions, have the
ligands to build libraries of receptors for dicarboxylic acids.
Conventional combinatorial libraries contain a statistically
defined number of components. Interest has developed in
advantage that a building block approach may be used in
which metal-containing units retain their structural integrity
through a synthetic sequence. This approach has been adopted
for the construction of photoactive metallodendrimers and
dynamic combinatorial libraries, in which the components
28–31
are in dynamic equilibrium.
Such libraries offer the pos-
sibility of responding to a challenge such as a template
molecule by redistribution of the components to produce
one predominant species. A prerequisite for such a dynamic
library is a reversible bond formation reaction as the key
synthetic step and imines have been a common motif in these
libraries. An obvious extension is to the use of metal–ligand
1
1–17
metallopolymers.
In contrast, labile metal centres allow
equilibrium to be reached rapidly and thermodynamic pro-
ducts to be isolated, although a disadvantage is that it is not
always possible to carry labile complexes through subsequent
reaction steps.
7
We have recently investigated the use of d cobalt(II) centres
32
bond formation as the reversible process, and the concept
for the assembly of tpy and bpy ligands into linear and
1
has been established in the optimisation of extractants for
zinc(II) and cadmium(II) ions, by presenting the metal ion as a
challenge to libraries of imines prepared from aminophenols
8–32
dendritic systems
and have demonstrated that oxidation
6
to kinetically inert d cobalt(III) allows the dynamic system to
be ‘‘frozen’’. This promises synthetic schemes in which the
advantages of labile metal centres (rapid reaction, thermody-
namic products) may be combined with subsequent reaction
steps associated with inert centres (stable complexes, no ligand
dissociation processes).
33
and 2-pyridinecarbaldehyde or optimising receptors for
34,35
lithium cations.
A number of examples of metal ion
involvement in the establishment of dynamic combinatorial
36–47
libraries have recently been described
48,49
referred to two excellent reviews.
and the reader is
Combinatorial chemistry has its basis in ‘‘organic’’ synthesis
and most is concerned with reactions involving the formation
or breaking of carbon–carbon or carbon–heteroatom
Most recently, cobalt(II)-centred assembly has been used for
the preparation of dynamic combinatorial libraries based
50
upon chiral bpy ligands or terpy ligands supporting a
2
4,25
bonds.
The construction of combinatorial libraries of
51
dynamic library of imines giving rise to a two-level system.
related compounds using these ‘‘organic’’ bond formation
Implicit in a dynamic combinatorial library of this type is the
assumption that all of the possible species are present in
statistical amounts; a modification of the concept to ‘‘virtual
combinatorial libraries’’ in which non-statistical distributions
of components are present has been made in the case of
Department of Chemistry, Spitalstrasse 51, CH 4056 Basel,
Switzerland. E-mail: Edwin.Constable@unibas.ch; Fax: +41 61 267
1005; Tel: +41 61 267 1001
This journal is ꢀc the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2007
New J. Chem., 2007, 31, 1437–1447 | 1437

View Article Online
3
0,48,49,52,53
0
molecular polygons.
We have discussed the ex-
vinyl]-2,2 -bipyridine as a yellow oil (0.534 g, 31%). d
H
0
/ppm
change of various substituted terpy ligands at cobalt(II) cen-
5
(250 MHz, CDCl
3
) 8.65 (1H, ddd, J 4.8, 1.8, 1.3 Hz, H6 ), 8.48
0,54,55
ters
and demonstrated how electron-transfer between
(1H, d, J 2 Hz, H6), 8.33 (1H, dt, J 7.9, 1.3 Hz, H3), 8.22 (1H,
0
cobalt(II) and cobalt(III) species may be monitored by ex-
change spectroscopy and used to unambiguously assign the
6
NMR spectra of cobalt(II) species. We now discuss the use of
d, J 8.2 Hz, H3 ), 7.78 (1H, td, J 7.9, 2.0 Hz, H4), 7.60 (1H, dd,
0
0
J 8.2, 2.3 Hz, H4 ), 7.23 (1H, m, H5 ), 6.92 (1H, d, J 13.8 Hz,
H8), 5.13 (1H, d, J 13.8 Hz, H7), 2.88 (6H, s, N(CH ) ); d /
5
3
2
C
1
H NMR spectroscopy in the establishment and quantification
3
ppm (100 MHz, CDCl ) 157.0, 150.9, 149.4, 145.6, 142.0,
of libraries of bpy and phen ligands with cobalt(II) centres.
137.2, 130.1, 125.7, 123.1, 121.3, 120.7, 93.8, 41.0.
0
2
,2 -Bipyridine-5-carbaldehyde. 5-[2-(N,N-Dimethylamino)-
0
Experimental
vinyl]-2,2 -bipyridine (0.458 g, 2 mmol) was dissolved in 1 : 1
3
water-THF (130 cm ) and sodium periodate (1.72 g, 8 mmol)
NMR measurements
was added. The solution was stirred at room temperature for
1
H NMR spectra were measured on Bruker AM250 or AM500
4
h after which the insoluble products were filtered off and
spectrometers using 5 mM stock solutions of the complexes in
washed well with THF. The combined THF extracts were
3
CD CN. 500 MHz gradient COSY and NOESY spectra were
evaporated in vacuo, the residue dissolved in dichloromethane
3
100 cm ) and the solution washed with saturated sodium
recorded at 295 K using the Bruker programs cosygp and
noesytp, respectively. All spectra are referenced to the signal of
(
3
hydrogen carbonate solution (4 ꢁ 50 cm ). The organic phase
was dried over sodium sulfate and evaporated in vacuo to give
the product as a yellow–brown solid. Purification by column
CD HCN at d 1.94 ppm. The chemical shifts are somewhat
2
temperature dependent and all tabulated date refer to spectra
measured at 295 K. EXCSY spectra were acquired in the
phase-sensitive mode, using the noesytp programme.
2 3 2 2
chromatography (Al O , CH Cl ) yielded a white solid (0.315
g, 86%). (Found: C, 71.45, H 4.63, N 14.95. Calc. for
C H N O: C 71.73, H 4.38, N 15.21%); d /ppm (250 MHz,
CDCl ) 10.20 (1H, s, CHO), 9.16 (1H, dd, J 2.0, 0.8 Hz, H6),
3
1
1
8
2
H
Syntheses
The compounds bpy and phen were used as supplied; 2-
0
8
.76 (1H, ddd, J 4.7, 1.7, 1.1 Hz, H6 ), 8.65 (1H, d, J 8.0 Hz,
0
5
7
1-oxo-2-pyridinioethyl)pyridine iodide was prepared by
(
H3), 8.54 (1H, dt, J 8.0, 1.1 Hz, H3 ), 8.32 (1H, dd, J 8.0, 2.0
0
the literature method.
Hz, H4), 7.90 (1H, td, J 7.8, 1.7 Hz, H4 ), 7.42 (1H, ddd, J 7.4,
0
4
1
.7, 1.4 Hz, H5 ); dC/ppm (100 MHz, CDCl
55.2, 152.0, 149.9, 137.5, 137.3, 131.5, 125.2, 122.6, 121.7; m/z
3
) 190.9, 161.1,
0
5
-Methyl-2,2 -bipyridine.
2-(1-Oxo-2-pyridinioethyl)pyri-
3
dine iodide (8.16 g, 25 mmol), 2-methylpropanal (3.00 cm ,
+
+
ꢃ1
(
ESMS) 184.1 [M] , 155.1 [M ꢃ CHO] ; IR (n/cm ) 3379w,
3
6 mmol) and ammonium acetate (5.81 g, 75 mmol) were
3
dissolved in methanol (70 cm ) and the mixture refluxed for
.5 h. The solution was concentrated and the residue extracted
3
1
9
055w, 2916w, 2840w, 2715w, 1697m, 1589m, 1550m, 1450m,
434m, 1357m, 1303w, 1249w, 1203m, 1141w, 1087w, 1064w,
95w, 840m, 794s, 740s, 663w, 624w. Mp 91.8–93.0 1C.
3
3
with hexane (6 ꢁ 30 cm ). The combined organic layers were
3
washed with saturated sodium chloride solution (2 ꢁ 70 cm ),
Complex synthesis. The complexes [Co(bpy)
3
0
][PF
6
]
2
,
dried over sodium sulfate, and the solvent evaporated in vacuo.
The yellow liquid residue was dried in vacuo (3.10 g, 72%).
0
[
Co(Me
dine) and [Co(phen)
equivalent of Co(OAc)
2
bpy)
3
][PF
6
]
2
(Me
][PF
ꢂ 4H
2
bpy = 4,4 -dimethyl-2,2 -bipyri-
3
6
]
2
were prepared by reacting one
2
O with three equivalents of the
(
Found: C 75.61, H 6.43, N 16.14. Calc. for C11
H
10
N
2
ꢂ
2
1
/3H O: C 75.23, H 6.08, N 15.95%); m/z (ESMS) 250.8 [M
2
ligand in warm methanol followed by precipitation with
ammonium hexafluorophosphate.
+
+
+
Na + acetone] , 193.2 [M + Na] ; d /ppm (250 MHz,
CDCl ) 8.61 (1H, ddd, J 4.9, 1.9, 1.0 Hz, H6 ), 8.45 (1H, m,
3
H
0
0
H6), 8.31 (1H, dt, J 8.3, 1.0 Hz, H3 ), 8.24 (1H, d, J 7.5 Hz,
0
Results and discussion
H3), 7.72 (1H, td, J 8.0, 1.9 Hz, H4), 7.54 (1H, m, H4 ), 7.19
0
(
1H, ddd, J 7.5, 4.9, 1.3 Hz, H5 ), 2.29 (3H, s, CH
3
C
); d /ppm
Cobalt(II) centres for the construction of combinatorial libraries
(
100 MHz, CDCl
3
) 156.6, 153.9, 149.9, 149.4, 137.8, 137.2,
ꢃ1
Metal–ligand interactions present a powerful methodology for
the preparation of both static and dynamic combinatorial
libraries. The basic concept of preparing mixtures is alien to
the coordination chemist, whose traditional emphasis has been
the isolation of single pure substances. Although one may
regard first generations of metal-assembled libraries as nothing
more than equilibrium mixtures of complexes, subsequent
interactions at the periphery of the ligands may lead to useful
molecular recognition events and the potential to select mem-
bers of the library and perturb the distribution resulting in an
amplification of the library.
1
2
1
33.7, 123.7, 121.1, 120.9, 18.6; IR (n/cm ) 3047w, 3001w,
923w, 2869w, 1558m, 1458s, 1434m, 1380m, 1249w, 1218w,
126w, 1095w, 1064w, 1033m, 987m, 848m, 786s, 740s, 648w.
0
5
-[2-(N,N-Dimethylamino)vinyl]-2,2 -bipyridine. A solution
0
of 5-methyl-2,2 -bipyridine (1.32 g, 7.8 mmol) and tert-but-
3
oxybis(N,N-dimethylamino)methane (4.50 cm , 21.8 mmol) in
3
dry DMF (25 cm ) was degassed with argon for 15 min and
then heated to 140 1C under argon for 17 h. The reaction
mixture was cooled to room temperature and water (200 cm )
3
was added. The solution was extracted with dichloromethane
3
(
5 ꢁ 100 cm ), the combined organic layers dried over sodium
We have been particularly concerned with the construction
7
of libraries using cobalt(II) centres. The d cobalt(II) ion can
sulfate and the solvent evaporated in vacuo. Purification of the
crude product by column chromatography (aluminium oxide,
toluene: diethylamine 40 : 1) gave 5-[2-(N,N-dimethylamino)-
5
exhibit both high- and low-spin electron configurations (t eg
2
2
g
6
or t2g eg descriptions in pure O symmetry) and regardless of
1
h
1
438 | New J. Chem., 2007, 31, 1437–1447
This journal is ꢀc the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2007

View Article Online
the spin state, cobalt(II) complexes possess unpaired electrons
and mononuclear complexes will be paramagnetic. The elec-
1
tronic relaxation time is such that in favourable cases,
H
NMR spectra of the complexes can be readily measured and
interpreted as a result of the large chemical shift ranges
resulting from the perturbation of the local magnetic field by
that arising from the unpaired electrons and minimal broad-
ening arising from the coupling of electronic and nuclear
1
9,22
spins.
Ligand exchange reactions at cobalt(II) centres are
reasonably fast and even though Co–L bonds are thermody-
namically strong, they are also labile. This ensures that
systems will rapidly attain equilibrium and establish a library
1
of complexes. To date, few examples of the use of H NMR
spectroscopy to monitor exchange reactions in such systems
5
have been reported.
6,58
Speciation in cobalt(II)-bpy (phen) systems
1
Fig. 1 250 MHz H NMR spectra at 298 K of 5 mM CD
solutions of the complexes (a) [Co(phen) ][PF , (b) [Co(Me bpy)
PF and (c) [Co(bpy) ][PF . Assignments are given using the
3
CN
Single ligand. In this section we develop the speciation
models for libraries constructed from symmetrical bidentate
ligands. The simplest case comprises the cobalt(II) ion and
three equivalents of a single ligand A. In this and subsequent
discussion, we assume that the thermodynamic stability of
3
6
]
2
2
3
]-
[
6
]
2
3
6 2
]
numbering scheme presented for the free ligands. Chemical shifts in
d/ppm
[
MA ] (M = cobalt(II), charges omitted for clarity) is such that
3
is presented in Fig. 2, and indicates both the number of species
and the contribution to the equilibrium mixture on statistical
the involvement of [MA] and [MA ] species may be disre-
2
2
+
59–61
2+
garded (log b [Co(phen) ] 19.9–20.1;
1
[Co(Me bpy) ]
2 3
3
3
grounds. At equilibrium, a 1 : 3 : 3 : 1 mixture of [MA ] :
3
6
7.55; [Co(bpy)3]
2
2+
59–66
15.82–17.59
). Furthermore, within
[MA B] : [MAB ] : [MB ] is expected. In the heteroleptic
2 2 3
the limits of integration of NMR data, the equilibrium con-
stants may be taken as approximately equal. In this case, the
species [MA B] the unique ligand B gives rise to one set of
2
signals as the two rings are in chemically and magnetically
identical environments whilst in ligand A the two rings of the
bpy are no longer equivalent as one is trans to A and the other
trans to B (Fig. 2). The result is that each pair of chemically
unique protons in the free A ligand (which give a single
3
library will consist of two (D and L) enantiomers of [MA ]. In
the absence of a chiral environment (solvent, counterions,
substituents etc.), NMR studies will detect a single species
corresponding to the enantiomeric pair D-[MA
3
] and L-[MA
3
].
All NMR spectra were recorded in CD CN solution and
3
resonance in [MA
3
]) will give two resonances with equal
6
Evans’ method for magnetic moment determinations re-
7
intensity in [MA B] as they are now chemically and magneti-
2
vealed the parent complexes [Co(bpy) ][PF ] , [Co(Me -
3
6 2
2
cally inequivalent. An identical argument applies to the B
ligands in [MAB2].
0
0
bpy) ][PF ] (Me bpy = 4,4 -dimethyl-2,2 -bipyridine) and
3
6 2
2
[
Co(phen) ][PF ] to be low spin in solution in this solvent.
3 6 2
The intensities of the resonances follow from the statistics
and the number of ligands present in each of the species and is
illustrated by considering ligand A (exactly similar arguments
apply to ligand B). In total, four resonances are expected for
1
Fig. 1 presents the 250 MHz H NMR spectra of 5 mM
solutions of [Co(bpy) ][PF , [Co(Me bpy) ][PF and [Co-
phen) ][PF . In each case, paramagnetically shifted and
3
6
]
2
2
3
6 2
]
(
3
6 2
]
broadened spectra are obtained which show four signals
appropriate to a single ligand environment for the pair of
enantiomers. Line-widths vary from 10–30 Hz for the majority
of the peaks, with the most shifted peak being significantly
broader with line-widths 50–120 Hz. From our previous
studies, we can assign the spectra to a low spin solution species
on the basis of a chemical shift in the region of d 100 ppm for
the bpy H or phen H protons. A comparison of the spectra
any given proton: one from [MA
from [MA B] (see discussion above). The intensities for any
proton from an A ligand in a statistical mixture of [MA ],
] is given by (number of A ligands
statistical weighting ꢁ proton weighting) which yields ratios
] 3 ꢁ 1 ꢁ 2 = 6) : ([MAB ] 1 ꢁ 3 ꢁ 2 = 6) : ([MA B] 2 ꢁ
ꢁ 1 = 6 and 2 ꢁ 3 ꢁ 1 = 6) : ([MB ] = 0); in other words,
3 2
], one from [MAB ] and two
2
3
[
2 2 3
MAB ], [MA B] and [MB
ꢁ
(
[MA
3
2
2
3
3
6
2,9
every proton from A and B should give four signals in a 1 : 1 :
allows the assignment of the d 14.6 ppm signal in [Co(bpy)
3
]-
1
: 1 ratio. Proton weightings indicate if the protons of the two
[
PF and the d 17.3 ppm peak in [Co(phen) ][PF to H .
6
]
2
3
6
]
2
4
Consideration of chemical shifts and equivalent protons in the
phen and bpy systems together with COSY studies leads to the
complete assignment that is presented in Table 1. The signals
1
Table 1 Assignment of the 295 K 500 MHz H NMR spectra of
CD CN solutions of [CoL ][PF ] complexes. The numbering scheme
3 3 6 2
refers to the bpy ligands, with the equivalent or unique protons of the
phen ligands denoted in parentheses. Chemical shifts d/ppm
1
1
are all paramagnetically broadened and no H– H coupling is
observed directly.
L
H
6
(H
2
)
H
5
(H
3
)
H
4
H
3
(H5,6
)
Methyl
0.40
Two ligands. If two ligands A and B are allowed to
equilibrate with a metal centre, then the library of complexes
has a greater diversity of species. The most convenient analysis
bpy
Me
phen
87.96
89.74
107.5
46.25
44.50
50.18
14.58
17.33
84.48
81.52
33.43
2
bpy
This journal is ꢀc the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2007
New J. Chem., 2007, 31, 1437–1447 | 1439

View Article Online
Fig. 2 Statistical equilibration of [MA
: 3 mixture of [MA ], [MB ], [MA B] and [MAB
B] and each B ligand in [MAB
2
3
] and [MB
3
] will give a 1 : 1 :
]. The two rings of
] are magneti-
3
3
3
2
2
each A ligand in [MA
cally inequivalent.
2
rings are equivalent (weighting two) or inequivalent (weighting
one).
The most convenient way of following the equilibration
involves mixing the preformed homoleptic complexes [MA3]
and [MB ] and this has been done with the three binary
3
combinations from the three complexes [Co(bpy) ][PF ] ,
3 6 2
[
Co(Me
2
bpy) ][PF
3
6
]
2
and [Co(phen)
3
][PF
6
]
2
. In each case,
equal amounts of 5 mM solutions in CD
3
CN were mixed to
give solutions 5 mM in cobalt and 7.5 mM in each of the
ligands. In all cases, equilibration was complete by the time
that the NMR spectra were recorded (within 5 min of mixing)
and four signals were observed for every chemically unique
proton in the ligands (Fig. 3).
The detailed analysis of the [Co(Me bpy) ][PF ] and [Co-
6 2
2
3
(
phen) ][PF ] mixture will be given, but this differs in no way
3 6 2
from the other two combinations. The first thing of note is that
every peak from the parent homoleptic species is now split to a
1
Fig. 3 (a) Sections of the 250 MHz H NMR spectrum at 298 K of a
1
: 1 : 1 : 1 set of four peaks. Within any cluster of four peaks,
one may be immediately assigned to the known homoleptic
Co(Me bpy) ][PF or [Co(phen) ][PF species. The re-
maining three peaks belong to the [Co(phen)
1 : 1 mixture of 5 mM CD CN solutions of the complexes
3
[Co(phen) ][PF ] and [Co(Me bpy) ][PF ] showing how four peaks
6 2
3
2
3
6 2
[
2
3
6
]
2
3
6
]
2
are obtained for a given ligand proton. Sections (b) and (c) show the
2
+
4 2
splitting of phen H and Me bpy methyl signals, respectively. Chemi-
2
2
(Me bpy)]
2
+
cal shifts in d/ppm
2 2
and [Co(phen)(Me bpy) ]
species, but cannot be further
assigned from these spectra. We were extremely surprised that
dispersion in the cobalt(II) shifted spectra was sufficient to
clearly allow the resonances of the inequivalent rings of the
major species in the heteroleptic complexes to be distin-
guished. This is even possible with the methyl signals of the
from A and B should give four signals in a 4 : 2 : 2 : 1 ratio.
Once again, we discuss the case of the [Co(Me bpy) ][PF
and [Co(phen) ][PF mixture in detail, but a similar analysis
is made for the other binary combinations. Spectra were
recorded with 1 : 2 and 2 : 1 mixtures of [Co(Me bpy) ][PF
and [Co(phen) ][PF . We now consider the phen H signals
2
3
6 2
]
3
6 2
]
5
,6
Me bpy ligand and the H resonances of the phen ligands,
2
2
3
6 2
]
which are rather remote from the paramagnetic centre. Re-
markably, all 32 resonances may be observed as equal intensity
signals in each spectrum. This experiment establishes both the
utility of the cobalt(II) complexes in quantifying the compo-
nents of the library and that statistical equilibrium distribu-
tions of compounds are obtained.
3
6
]
2
4
in detail to clarify the analysis (Fig. 4). In the 1 : 1 mixture,
four peaks of equal intensity are observed at d 17.29, 17.18,
17.06 and 16.96 ppm. The peak at d 17.29 ppm corresponds to
the single peak observed in pure [Co(phen)
assigned to this species. In the 1 : 2 [Co(Me
(phen) ][PF mixture, the four peaks are observed in a 4 : 2 :
2 : 1 ratio of intensities. It follows immediately that the d 17.29
ppm is from [Co(phen) ][PF , the 17.18 and 17.06 peaks to
the non-equivalent rings of the phen in [Co(phen)
][PF
3
]
6 2
and may be
bpy)
2
][PF –[Co-
]
6 2
3
A full assignment of the spectrum follows from spectra
recorded with variable ratios of the reactants. Consider a
3
6 2
]
2
: 1 mixture of [MA ] and [MB ]; a statistical mixture will
3
3
6 2
]
3
contain the species [MA ], [MA B], [MAB ] and [MB ] in the
2
2
-
3
2
3
ratio 8 : 12 : 6 : 1. For any given proton in ligand A, we expect
four signals, corresponding to [MA ], the two environments in
MA B] and [MAB ]. The intensities for a proton from ligand
A in the mixture of [MA ], [MAB ], [MA B] and [MB ] is
given by (number of A ligands ꢁ statistical weighting ꢁ
proton weighting) which yields ratios ([MA
(Me
(Me
2
bpy)][PF
bpy) ][PF
bpy) ][PF
6
]
2
and the d 16.96 ppm peak to [Co(phen)-
. Final confirmation comes from the 2 : 1
–[Co(phen) ][PF mixture, in which the
3
2
2
6 2
]
[
2
2
[Co(Me
2
3
6
]
2
3
6 2
]
four peaks are observed in a 1 : 2 : 2 : 4 ratio of intensities.
Additional information regarding assignments and the estab-
lishment of statistical distributions of complexes in the li-
3
2
2
3
3
] 3 ꢁ 8 ꢁ 2 =
B] 2 ꢁ 12 ꢁ 1 = 24 and
] = 0); in other words, every proton
1
4
8) : ([MAB
2
] 1 ꢁ 6 ꢁ 2 = 12) : ([MA
2
braries comes from the H NMR spectra of solutions
2
ꢁ 12 ꢁ 1 = 24) : ([MB
3
containing the parent complexes in a 3 : 1 ratio. In this case,
1
440 | New J. Chem., 2007, 31, 1437–1447
This journal is ꢀc the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2007

View Article Online
1
1
Fig. 6 Part of the 500 MHz H– H COSY spectrum of a 1 : 1 mixture
of [Co(bpy) ][PF and [Co(phen) ][PF showing the correlation of
the phen H and bpy H protons with bpy and phen H signals The
CN at 298 K and is 5 mM in cobalt(II). Chemical
Fig. 4 The phen H
mixtures of [Co(Me
K, 5 mM in cobalt(II) 250 MHz). The abbreviations P3, P2M and PM2
4
signals observed in (a) 1 : 1 (b) 1 : 2 and (c) 2 : 1
bpy) ][PF and [Co(phen) ][PF (CD CN, 298
3
6
]
2
3
6
]
2
3
3
6
]
2
3
6 2
]
2
3
5
4
2+
solution is in CD
3
refer to the signals from the species [Co(phen)
3
]
, [Co(phen)
. There are two environments for
the phen in the P2M species. Chemical shifts in d/ppm.
2 2
(Me -
2+
2+
shifts in d/ppm
2 2
bpy)] and [Co(phen)(Me bpy) ]
be further assigned using COSY spectroscopy of the 1 : 1
3
mixtures. It is not possible to completely assign these to rings
1
the expectation values for a 3 : 1 mixture of [MA
are 27 : 27 : 9 : 1 ratio of [MA ], [MA B], [MAB
3
] and [MB
3
]
trans to A or B through NOSY spectroscopy, as exchange
3
3
2
2
] and [MB ].
3
1
Applying the weighting scheme for the number of protons in
each compound, we predict that each proton of each ligand
will give rise to four signals in a 9 : 3 : 3 : 1 ratio. This is
illustrated for the methyl protons of a 3 : 1 mixture of
cross-peaks will also be obtained. The signals for the H
protons of bpy and Me bpy and the H2,9 protons of phen are
6
2
too broad and show no detectable cross-peaks in the COSY
spectrum. The COSY technique is illustrated in Fig. 6, which
[
Co(Me bpy) ][PF ] –[Co(phen) ][PF ] in Fig. 5; within ex-
2
shows the correlation of the phen H
3 5
and bpy H protons with
3
6 2
3
6 2
perimental error, the expected 9 : 3 : 3 : 1 ratio of peaks is
obtained.
bpy and phen H signals. The final assignments for the binary
4
combinations are presented in Table 2.
It follows from the statistical distribution of complexes, that
the dimensionless equilibrium constant for the equilibration
reaction
2
In the MA B complexes, each proton of the A ligand gives
rise to two signals and two sub-spectra are obtained which can
CoA + CoB = CoA B + CoB A
2
3
3
2
given by
½
CoA2Bꢄ½CoB2Aꢄ
K ¼
½
CoA3ꢄ½CoB3ꢄ
2
will be (3) /(1) = 9.
2
Three ligands. The analysis for an equilibrium mixture of
three ligands A, B and C is presented in Fig. 7, and indicates
both the number of species and the contribution to the
equilibrium mixture on statistical grounds. A total of ten
chemical species are present; of these ten species, only the
1
heteroleptic complex [MABC] will give rise to signals in the H
NMR spectrum which have not already been observed in the
binary mixtures. At statistical equilibrium, a 1 : 1 : 1 : 3 : 3 : 3 :
Fig. 5 The resonances of the methyl protons in a 3 : 1 mixture of
3
: 3 : 3 : 6 mixture of species is expected, of which the minor
[
2 3 6 2 3 6 2 3
Co(Me bpy) ][PF ] and [Co(phen) ][PF ] . The solution is in CD CN
components with relative concentration 1 are the homoleptic
compounds, the species of intensity 3 are the various binary
complexes and the dominant component is the heteroleptic
compound [MABC]. In the heteroleptic compound, each ring
of each ligand is magnetically independent. Once again, the
analysis will be given in detail considering a pair of protons
at 298 K and is 5 mM in cobalt(II) and was recorded on a 250 MHz
instrument; within experimental error, the expected 9 : 3 : 3 : 1 ratio of
peaks is obtained. The abbreviations M3, M2B and MB2 refer to the
2+
2+
species [Co(Me
2
bpy)
, respectively. There are two environments for Me
M2B species. Chemical shifts d/ppm.
3
]
, [Co(Me
2
bpy)
2
(bpy)]
2
and [Co(Me bpy)-
2+
(
bpy)
2
]
2
bpy in the
This journal is ꢀc the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2007
New J. Chem., 2007, 31, 1437–1447 | 1441

View Article Online
1
Table 2 500 MHz H NMR spectra (295 K) of CD
3
CN solutions of complexes (B = bpy, M = Me
2
bpy or P = phen). Chemical shifts d/ppm
B
M
P
H
5
H
4
H
3
H
6
H
5
H
3
Methyl
81.52
H
2
H
3
H
4
H
5
BBB
MMM
PPP
88.0
46.25
14.58
84.48
89.74
44.50
0.40
107.20
113.7
50.18
50.64
—
50.46
50.23
17.33
16.33
33.43
31.73
a
a
BBP
84.5
83.9
80.9
45.57
45.76
45.24
14.58
15.69
15.66
86.58
86.16
88.13
a
a
a
PPB
110.7
1
17.33
16.33
32.64
32.58
a
10.0
MBB
MMB
PMM
MPP
88.6/88.0
8.6/88.0
90.7
47.32
45.31
46.41
15.32
14.53
15.26
85.51
84.26
85.28
87.10
44.31
80.71
0.50
8
a
a
a
a
a
a
a
89.5/88.8
43.39
45.38
44.79
42.89
43.35
81.73
80.51
83.78
82.06
84.22
1.16
ꢃ0.25
1.06
8
9.5/88.8
86.4
4.9
82.5
114.8
111.5
50.89
17.00
32.06
a
8
ꢃ0.12
a
a
a
a
0.49
51.40
49.51
17.10
17.22
33.01
32.52
1
10.0
a
Assignment to individual subsystem is ambiguous.
1
that are chemically and magnetically equivalent in free ligand
A. Ligand A is present in six of the ten species; in [MA3],
8(a)–(c) present parts of the H NMR spectra of the ternary
mixture and clearly show the occurrence of eight peaks for
every chemically unique proton in the ligand. The assignment
presented in Table 3 follows from a COSY spectrum of the
ternary mixture.
[
MAB ] and [MAC ] the two rings of A remain in identical
2 2
environments and each species will give a single A resonance,
whereas in [MA B], [MA C] and [MABC] the rings are
2
2
inequivalent and each A ligand will give rise to two reso-
nances. Thus each chemically unique proton in the free ligand
A will give a total of nine resonances.
The 1 : 1 : 1 : 3 : 3 : 3 : 3 : 3 : 3 : 6 statistical distribution of
complexes, according to the equation
7
CoA + 7CoB + 7CoC = 3CoA B + 3CoB A +
3 3 3 2 2
The intensities of the resonances again follow from the
statistics and the number of ligands present in each of the
species and is illustrated considering ligand A. As shown
above, nine resonances are expected for any proton from A:
one each from [MA ], [MAC ] and [MAB ], and two each
3
CoA C + 3Co AC + 3CoB C + 3CoBC + 3CoABC
2 2 2 2
leads to an equilibrium constant given by
3
3
3
3
3
3
3
½CoA2Bꢄ ½CoAB2ꢄ ½CoA2Cꢄ ½CoAC2ꢄ ½CoB2Cꢄ ½CoBC2ꢄ ½CoABCꢄ
3
2
2
K ¼
7
7
7
from [MA B], [MA C] and [MABC]. The intensities in a
2
½CoA3ꢄ ½CoB3ꢄ ½CoC3ꢄ
2
statistical mixture of [MA ], [MAB ], [MAC ], [MA B],
3
2
2
2
6
3
10
and numerically equal to (27) ꢁ (6) /1 = 8.37 ꢁ 10 .
[
MA C] and [MABC] will be (3 ꢁ 1 ꢁ 2 = 6) : (1 ꢁ 3 ꢁ 2
2
=
6) : (1 ꢁ 3 ꢁ 2 = 6) : (2 ꢁ 3 ꢁ 1 = 6) + (2 ꢁ 3 ꢁ 1 = 6) :
(
2 ꢁ 3 ꢁ 1 = 6) + (2 ꢁ 3 ꢁ 1 = 6) : (1 ꢁ 6 ꢁ 1 = 3). Every
proton from A (and by a similar analysis every proton from B
and C) should give nine signals of equal intensity.
The only new signals which have not been observed in the
binary mixtures are the two assigned to the non-equivalent
+
2
2
rings of each ligand in [Co(bpy)(phen)(Me bpy)] . Fig.
1
Fig. 8 Sections of the 500 MHz H NMR spectrum at 295 K of a 1 :
1
: 1 mixture of 5 mM CD
[Co(phen) ][PF , [Co(bpy) ][PF
how nine peaks are obtained for a given ligand proton; (a) the phen H
signals (b) the phen H , bpy H and Me bpy H signals and (c) the
methyl signals of Me bpy, respectively. Chemical shifts in d/ppm.
3
CN solutions of the complexes
Fig. 7 Statistical equilibration of [MA
: 1 : 1 : 3 : 3 : 3 : 3 : 3 : 3 : 6 mixture of [MA
MAB ], [MAC ], [MA C], [MB C], [MBC ] and [MABC]. In [MABC]
each ring of each ligand is magnetically inequivalent.
3
], [MB
3
] and [MC
3
] will give a
3
6
]
2
3
6
]
2
and [Co(Me bpy) ][PF showing
2
3
6 2
]
1
3
], [MB
3
], [MC
3
], [MA B],
2
2
[
2
2
2
2
2
3
5
2
5
2
1
442 | New J. Chem., 2007, 31, 1437–1447
This journal is ꢀc the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2007

View Article Online
Table 3 Assignment of the ternary complex [Co(bpy)(Me
2
bpy)-
CN solutions of the
and [Co(Me
2+
(
phen)]
complexes [Co(phen)
bpy) ][PF as determined from the 500 MHz H and H– H COSY
in a 1 : 1 : 1 mixture of 5 mM CD
3
3
][PF [Co(bpy) ][PF
6
]
2
,
3
6
]
2
2
-
1
1
1
3
6 2
]
NMR spectra. The numbering scheme refers to the bpy ligands, with
the equivalent or unique protons of the phen denoted in parentheses.
Each chemically unique proton within the ligands gives rise to two
signals in the ternary complex. These are grouped in subspectra, but
cannot be correlated between ligands. Chemical shifts d/ppm
Fig. 9 The mer and fac stereoisomers of a complex containing three
0
2
,2 -bipyridine-5-carbaldehyde ligands. The colouring (black, grey or
white) indicates the three chemically and magnetically inequivalent
ligands in the mer diastereomer.
In the case of the asymmetrical ligand bpyal, the speciation
involves the formation of mer and fac stereoisomers of the tris
complex (Fig. 9). For a complex of the type [ML
3
] with
chelating bidentate ligands, a statistical distribution of 1 : 3
8
fac : mer is expected.
7
L
H
6
(H
2
)
H
5
(H
3
)
H
4
H
3
(H5,6
)
Methyl
In the fac stereoisomer, each chemically unique proton will
give a single signal as all three ligands are chemically and
magnetically equivalent, whereas in the mer stereoisomer each
chemically unique proton of each ligand is in a magnetically
unique environment. Using our usual analysis, with the ex-
pectation ratio of 1 : 3 fac : mer, the intensities in a statistical
Bpy
87.96
89.74
46.66
4
15.32
15.59
87.58
85.91
82.79
4.79
43.84
3.68
a
a
a
a
a
Me
2
bpy
0.38
0.64
a
4
82.30a
32.00
31.69
a
Phe
115.2
113.1
51.77
49.74
17.07
16.25
a
a
a
Assignment to individual sub-spectrum ambiguous.
mixture of fac and mer [MA ] will be (1 ꢁ 3 = 3) : (3 ꢁ 1 = 3)
3
+
(3 ꢁ 1 = 3) + (3 ꢁ 1 = 3). Every proton will give four
signals of equal intensity. Fig. 10 presents regions of the 500
1
MHz H NMR spectrum of the mixture of mer and fac
Asymmetric ligands—mer and fac isomers
Although the establishment of the libraries with the cobalt(II)
complexes is elegant, it serves no significant purpose if the
information cannot be read-out or the distribution of species
within the library cannot be altered by external constraints.
Accordingly, we have extended the studies to systems contain-
ing functional groups which can react in subsequent steps.
Initially, we considered the use of the monofunctionalised bpy
complexes. It is immediately evident that all signals are present
in sets of four of equal intensity.
1
The H NMR spectrum of the mer/fac mixture consists of a
number of sets of four signals (Fig. 10(a)–(c)). Eight broad
signals (d 105.5, 99.3, 96.4, 91.4, 90.9, 86.0, 84.5 and 79.6 ppm)
are assigned to H and H (Fig. 10(a)); no COSY or NOESY
0
6 6
peaks are observed and further assignment was not possible. A
0
ligand, 2,2 -bipyridine-5-carbaldehyde, bpyal.
6
set of eight peaks in the range d 74–92 ppm (Fig. 10(a)) are
8–75
The ligand bpyal
76
in which the known compound 5-methyl-2,2 -bipyridine was
was prepared in a two-step procedure
assigned to H
51.67, 46.39 and 46.29 ppm (Fig. 10(b)) are assigned to H
remaining set of 12 signals in the range d 19–7 ppm (Fig. 10(c))
are due to H , H and the aldehyde CHO protons. A more
detailed assignment follows from the H– H COSY spectrum.
Although the resonances of the H and H protons are too
3
and H
3
0
, whilst four peaks between d 51.80,
0
0
5
. A
0
converted to the N,N -dimethylaminovinyl derivative by reac-
tion with Bredereck’s reagent and then oxidised to bpyal with
sodium periodate (Scheme 1) in a reaction sequence directly
analogous to the one we have reported for the synthesis of
4
0
4
1
1
0
6
6
0
77
2
,2 -bipyridine-4-carbaldehyde.
broad to show any detectable COSY cross peaks, the remain-
ing J-coupled signals do, even though the line-width is so great
that no coupling is observed in the 1D spectra. Commencing
with the twelve peaks in the range 19–7 ppm, the four peaks at
d 13.71, 10.97, 9.66 and 7.24 ppm show no cross peaks and are
assigned to the aldehyde protons. Of the remaining eight
signals, four show cross peaks to the H
unambiguously assigned to H (Fig. 11(a)). Similarly, the
and H signals may be unambiguously correlated with
the H and H signals (Fig. 11(b)). The full assignment of the
5
0
signals and are
0
4
H
3
0
3
0
4
4
spectrum is given in Table 4.
Although the statistical distribution of complexes is strong
evidence for a dynamic equilibrium between mer and fac
stereoisomers, we have established the dynamic equilibrium
through the use of EXCSY spectroscopy in a similar manner
to that demonstrated for electron exchange between cobalt(II)
0
Scheme 1 Synthesis of 2,2 -bipyridine-5-carbaldehyde.
This journal is ꢀc the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2007
New J. Chem., 2007, 31, 1437–1447 | 1443

View Article Online
1
Fig. 10 Regions of the 500 MHz H NMR spectrum of a CD
3
CN
solution of the statistical mixture of diastereoisomeric mer and fac
2
+
complexes [Co(I)
the H and H and the eight sharper signals from H
presents the four signals from H and (c) shows the twelve signals
arising from H , H and the aldehyde, with the latter signals denoted.
3
]
; (a) shows the eight broad signals arising from
0
and H , (b)
0
3
6
6
3
5
0
1
1
4
4
Fig. 11 500 MHz H– H COSY spectra (295 K, CD
the H and H signals with H , H and H in the mixture of mer and
fac stereoisomers of [Co(I)
3
CN) correlating
Chemical shifts in d/ppm.
0
0
0
3
4
4
5
3
2
+
3
]
. Chemical shifts in d/ppm.
5
6
and cobalt(II) terpy complexes.
The standard NOESY
(
noesytgp) spectrum of the mixture of mer and fac isomers
exhibited EXCSY cross peaks corresponding to chemical
exchange but no detectable NOESY peaks. This is illustrated
in Fig. 12 which shows the region of the H
3
protons and
demonstrates that every H proton shows an exchange peak
3
1
Table 4 Assignment of the 295 K 500 MHz H NMR spectrum of a
with every other H environment; in other words, the ligand
3
2+
CD CN solution of the mixture of mer and fac [Co(bpyal) ]
3
3
exchange is occurring both between the mer and fac isomers
and between the three sites of the mer isomer.
complexes. The spin systems A to C for the (H
3
/H
4
) and the
(
H
3
0
/H ) are not correlated with one another and the aldehyde,
0
and H proton groups are not correlated with the other groups as
6
4
0
/H
5
0
H
6
they are either too broad or show no NOE cross peaks. The groups of
peaks are assigned from COSY spectra but cannot be unambiguously
attributed to particular rings in particular complexes. Chemical shifts
d/ppm
Amplification—selective reaction and interconversion of mer
and fac isomers
We now have a system which fulfills the prerequisites for a
dynamic combinatorial library—two reactive aldehyde com-
A
B
C
D
0
plexes in dynamic equilibrium with one another. Both 2,2 -
0
H
3
H
4
H
3
H
4
H
5
82.31
16.12
90.20
18.32
51.80
80.20
15.08
87.86
16.99
51.66
77.53
12.41
85.64
14.54
46.29
74.80
11.25
82.68
13.10
46.39
0
bipyridine-5-carbaldehydes and 2,2 -bipyridine-5,5 -carbalde-
5
hydes are known to react with tris(2-aminoethyl)amine (tren)
0
0
0
0
to give open-chain (cap) or encapsulating ligands, respec-
7
3,79,80
tively
and we planned to use this reaction to amplify
the library of complexes.
H6,6
0
105.5, 99.3, 96.43, 91.4, 90.9, 86.0, 84.5, 79.6
13.71, 10.97, 9.66, 7.24
Modelling and the known crystallographic data for the
tris(imines) confirmed the expectation that only the minor
CHO
1
444 | New J. Chem., 2007, 31, 1437–1447
This journal is ꢀc the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2007

View Article Online
1
1
Fig. 12 Part of the H– H NOESY spectrum of a CD
3
CN solution of
complexes showing the
environments. Chemical shifts in d/ppm.
2+
a mixture of mer and fac [Co(bpyal)
3
]
1
3 3 6 2
Fig. 13 250 MHz H NMR spectra (in CD CN) of [Co(bpy) ][PF ]
exchange of all four H
3
before and after the addition of different equivalents of tren. Chemical
shifts in d/ppm.
component fac stereoisomer can react with tren and the
consequence will be conversion of mer isomer to fac to return
to equilibrium (Scheme 2).
of bpyal are likely to be less stable than those of bpy as they
have an electron-withdrawing substituent and secondly, when
the stability constants are expressed in conventional concen-
The addition of one equivalent of tris(2-aminoethyl)amine
2
+
to the equilibrium mixture of mer and fac [Co(bpyal)
3
]
ꢃ3
3
tration units, b
mol L (or mol L and mol L if acetonitrile solvent is
explicitly included), respectively and they are dimensionally
3
and K (= b ) have units of mol L and
1
1
complexes resulted in the immediate loss of all H NMR
signals of these compounds and the development of the
spectrum of a new paramagnetically shifted species.
ꢃ1
3
ꢃ3
2
ꢃ2
incommensurate. Accordingly, we investigated the worst case
2+
2
+
Before discussing the formation of [Co(cap)] we make an
by titrating tren into a CD
3
CN solution of [Co(bpy)
3
]
(Fig.
observation and describe a separate series of experiments.
1
3). After the addition of half an equivalent of tren to the
2
+
First, we note that the solutions of [Co(bpyal)3] contained
no free bpyal, which has a characteristic resonance in CD CN
solution at d 10.2 ppm assigned to the aldehyde proton. At
1
solution, the H NMR spectrum shows resonances assigned to
3
2+
Co(bpy)3] , free bpy and a tren complex. After the addi-
[
tion of 1.5 equivalents of tren, the solution contains no
2+
2
+
first sight, complexes related to [Co(bpy)
5
3
]
(with log b
) are expected to be stable with respect to
3
[Co(bpy)3] . These results indicate that the desired capping
9–66
1
5.82–17.59
reaction with tren is likely to compete with the displacement of
the bpyal by tren.
1
The H NMR spectra of the mixtures with the complex to
8
1–90
displacement by tren (log K
1
12.4–14.7
). However inspec-
tion reveals that this may not strictly true. Firstly, complexes
template ratios 1 : 0.5, 1 : 1 and 1 : 1.6 are shown in Fig. 14.
After the addition of half an equivalent of tren, new para-
magnetically shifted signals in addition to the starting mixture
1
Fig. 14 250 MHz H NMR spectra of the paramagnetically shifted
region of equilibrated solutions of [Co(bpy-CHO)
3 6 2 3
][PF ] in CD CN
Scheme 2 Amplification of the library of mer and fac stereoisomers
by reaction with tren to give [Co(cap)]
2+
2+
.
of [Co(bpyal)
3
]
before and after addition of tren. Chemical shifts in d/ppm.
This journal is ꢀc the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2007
New J. Chem., 2007, 31, 1437–1447 | 1445

View Article Online
of mer and fac isomers of the bpyal complex ar observed. The
chemical shift of the most downfield peaks (d 4 170 ppm) is
free bpyal) was 1 : 3. Over hours to days, the spectrum of bpyal
is replaced by a new species identified as free cap. From this we
may construct the scheme presented in Scheme 3 and we see
that the apparent amplification of the mer and fac library in
6
3
typical of high-spin cobalt(II) solution species. It was there-
fore assumed that whatever new species had formed, the spin
state of the central cobalt(II) had changed from low-spin to
2
+
the reaction of [Co(bpyal)3] with tren is actually a destruc-
tive process in which the mer compound is lost in a competing
reaction. Specifically, we see that the capping reaction occurs
1
high-spin. After the addition of one equivalent of tren, the H
NMR spectrum showed that all signals of the starting bpyal
complex had disappeared and a single new paramagnetic
species was present. This showed a pattern of NMR signals
typical of a high spin cobalt(II) complex with signals at d 171,
2
+
selectively with the fac isomer to give [Co(cap)] . However,
2
+
the displacement of bpyal from mer-[Co(bpyal)
3
]
by tren is
2
+
faster than the interconversion of mer-[Co(bpyal)
2
3
]
to
+
1
58, 69.2, 67.2, 53.3, 26.7, 9.35, 8.83 ppm together with peaks
fac-[Co(bpyal)
3
]
.
assigned to the aliphatic protons. The interpretation is that the
2
+
tren reacted with the fac isomer of [Co(bpy-CHO)
3
]
and
Conclusions
formed the cap complex, driving the dynamic library to form a
single complex through the conversion of mer bpyal complex
to the fac isomer, which then subsequently reacts with addi-
tional tren. COSY spectra allowed the assignment of the
1
We have shown that H NMR spectroscopy is a powerful
technique for the analysis of libraries of paramagnetic cobal-
t(II) complexes. The establishment of a dynamic library with
the ligand bpyal has been demonstrated by the use of EXCSY
spectroscopy. The amplification of the library occurs by reac-
paramagnetically shifted peaks as follows: H6/6
ppm, H6/6 d 171/158 ppm, H d 69.2 ppm, H d 67.2 ppm,
d 53.3 ppm, H d 9.35 ppm, H d 8.83 ppm, NQCH
d 26.7 ppm.
As anticipated, competition between imine formation and
0
d 171/158
0
0
3
3
2
+
2+
H
5
0
4
0
4
tion of the fac-[Co(bpyal)
+
mer-[Co(bpyal) ]
3
3
]
to give [Co(cap)]
whilst the
2
reacts to give bpyal and cobalt(II)–tren
complexes.
1
H
ligand displacement does complicate the reaction. The
NMR spectrum in the region d 1–11 ppm is complicated,
but contains peaks in the ‘‘normal’’ aromatic region in addi-
tion to poorly resolved signals assigned to tren and tren
complexes. Comparison with known standards revealed that
these peaks corresponded to free bpyal and possibly free cap.
This is compatible with the observations made with bpy. Even
after the addition of 0.5 equivalents of tren, there is a
significant concentration of free bpyal (d 10.2 ppm).
Acknowledgements
We thank the University of Basel and the Swiss National
Science Foundation for financial support. We particularly
¨
wish to thank Professor Andreas Zuberbuhler for numerous
helpful discussions in the course of this work and Drs K.
Kulicke, V. Jullien and A. Mahmood for measurement of
1
H NMR spectra.
We have not investigated the speciation in the reaction of
2
+
[
Co(bpyal)
3
]
with tren at all concentrations in detail, but we
References
may make a number of comments. In the presence of excess
tren, the only cobalt-containing species are tren complexes.
1
Comprehensive Supramolecular Chemistry, ed. J. L. Atwood, J. E.
D. Davies, D. D. MacNicol and F. Vogtle, Pergamon Press,
Oxford, 1996, vol. 9.
2
+
¨
The solutions containing [Co(bpyal) ]
3
and tren at a 1 : 1
2
ratio contain [Co(cap)] and free bpyal. The solutions con-
+
2 G. F. Swiegers and T. J. Malefetse, Chem. Rev., 2000, 100, 3483.
3 G. F. Swiegers and T. J. Malefetse, Chem.–Eur. J., 2001, 7, 3637.
4 G. F. Swiegers and T. J. Malefetse, Coord. Chem. Rev., 2002, 225,
2
+
taining [Co(bpyal)3]
2+
contain no [Co(cap)] , [Co(bpyal)3] or free bpyal. Integra-
and tren at a 1 : 1.6 or greater ratio
2+
9
1.
1
tion of the H NMR spectrum of a 1 : 1 [Co(bpyal)
2+
3
]
to tren
5
E. C. Constable, Metals and Ligand Reactivity, VCH, Weinheim,
1995.
solution revealed that the ratio of the d 26.7 ppm peak to that
2
+
6
7
8
9
E. C. Constable, Chem. Ind. (London), 1994, 56.
O. Costisor and W. Linert, Rev. Inorg. Chem., 2003, 23, 289.
M. Fujita, Chem. Soc. Rev., 1998, 27, 417.
C. J. Jones, Chem. Soc. Rev., 1998, 27, 289.
of the d 10.2 ppm peak (i.e. the [Co(cap)]
complex to the
1
0 V. G. Machado, A. S. Mangrich and J.-M. Lehn, J. Braz. Chem.
Soc., 2003, 14, 777.
1 G. R. Newkome, E. F. He and C. N. Moorefield, Chem. Rev., 1999,
1
9
9, 1689.
1
1
2 V. Balzani and A. Juris, Coord. Chem. Rev., 2001, 211, 97.
3 A. M. Juris, A. M. Venturi, P. Ceroni, V. Balzani, S. Campagna
and S. Serroni, Collect. Czech. Chem. Commun., 2001, 66, 1.
4 S. Kelch and M. Rehahn, Chem. Commun., 1999, 1123.
5 S. Kelch and M. Rehahn, Macromolecules, 1999, 32, 5818.
6 E. C. Constable, in Comprehensive Coordination Chemistry II
Chemistry, ed. J. A. McCleverty and T. J. Meyer, Elsevier, Oxford,
1
1
1
2
003.
1
1
7 E. C. Constable, Chem. Commun., 1997, 1073.
8 E. C. Constable and A. M. W. Cargill Thompson, J. Chem. Soc.,
Dalton Trans., 1995, 1615.
1
9 E. C. Constable, P. Harverson and C. E. Housecroft, J. Chem.
Soc., Dalton Trans., 1999, 3693.
Scheme 3 Amplification of the library of mer and fac stereoisomers
2
+
2+
by reaction with tren to give [Co(cap)] , bpyal and
of [Co(bpyal)
3
]
cap through reaction of the stereoisomers in different pathways.
20 E. C. Constable, P. Harverson and M. Oberholzer, Chem. Com-
mun., 1996, 1821.
1
446 | New J. Chem., 2007, 31, 1437–1447
This journal is ꢀc the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2007

View Article Online
21 E. C. Constable, C. E. Housecroft, V. Jullien, M. Neuburger and S.
Schaffner, Inorg. Chem. Commun., 2006, 9, 504.
Ziolkowski, University Publishing House, Wydawnictwa Uniwer-
sytetu Wroclawskiego, 2000, vol. 7, p. 159.
2
2 E. C. Constable, C. E. Housecroft, M. Neuburger, S. Schaffner and
L. J. Scherer, Dalton Trans., 2004, 2635.
3 E. C. Constable, Coord. Chem. Rev., submitted.
4 Combinatorial Chemistry. A Practical Approach, ed. H. Fenniri,
OUP, Oxford, 2000.
56 H. S. Chow, E. C. Constable, C. E. Housecroft, K. J. Kulicke and
Y. Q. Tao, Dalton Trans., 2005, 236.
57 F. Krohnke, Synthesis, 1976, 1.
58 S. G. Telfer, X. J. Yang and A. F. Williams, Dalton Trans., 2004,
699.
2
2
2
2
2
5 Combinatorial Chemistry. A Practical Approach, ed. W. Bannwarth
and E. Felder, Wiley-VCH, Weinheim, 2000.
6 M. S. Goodman, V. Jubian and A. D. Hamilton, Tetrahedron
Lett., 1995, 36, 2551.
7 Goodman, A. D. Hamilton and J. Weiss, J. Am. Chem. Soc., 1995,
59 G. Anderegg, Helv. Chim. Acta, 1963, 46, 2397.
60 H. Irving and D. Mellor, J. Chem. Soc., 1962, 5222.
61 H. Irving and D. Mellor, J. Chem. Soc., 1962, 5237.
62 H. Schwarz, C. Creutz and N. Sutin, Inorg. Chem., 1985, 24, 433.
63 P. Langer, S. Fallab and H. Erlenmeyer, Helv. Chim. Acta, 1955,
38, 92.
117, 8447.
2
2
8 J.-M. Lehn, Chem.–Eur. J., 1999, 5, 2455.
9 P. A. Brady and J. K. M. Sanders, J. Chem. Soc., Perkin Trans. 1,
64 R. Davies and K. Dunning, J. Chem. Soc., 1965, 4168.
65 P. Daniele, G. Ostacoli and V. Zelano, Ann. Chim. (Rome), 1975,
65, 617.
1997, 3237.
3
3
0 I. Huc and J.-M. Lehn, Proc. Natl. Acad. Sci. USA, 1997, 94, 2106.
1 L. J. Prins, P. Timmerman and D. N. Reinhoudt, Pure Appl.
Chem., 1998, 70, 1459.
66 I. Ozherelev, T. Spolitak and D. Palade, Zh. Neorg. Khim., 1992,
37, 341.
67 D. F. Evans, J. Chem. Soc., 1959, 2003.
3
3
2 M. Albrecht, J. Incl. Phenom. Macrocycl. Chem., 2000, 36, 127.
3 D. M. Epstein, S. Choudhary, M. R. Churchill, K. M. Keil, A. V.
Eliseev and J. R. Morrow, Inorg. Chem., 2001, 40, 1591.
4 I. Saur, R. Scopelliti and K. Severin, Chem.–Eur. J., 2006, 12, 1058.
5 R. L. E. Furlan, Y. F. Ng, S. Otto and J. K. M. Sanders, J. Am.
Chem. Soc., 2001, 123, 8876.
68 V. Amendola, L. Fabbrizzi, C. Mangano, A. M. Lanfredi, P.
Pallavicini, A. Perotti and F. Ugozzoli, J. Chem. Soc., Dalton
Trans., 2000, 1155.
69 J. Polin, E. Schmohel and V. Balzani, Synthesis, 1998, 321.
70 S. Leroy, T. Soujanya and F. Fages, Tetrahedron Lett., 2001, 42,
1665.
3
3
3
3
3
6 T. Haino, H. Mitsuhashi, Y. Ishizu and Y. Fukazawa, Tetrahedron
Lett., 2006, 47, 7915.
7 N. Giuseppone, J. L. Schmitt and J.-M. Lehn, J. Am. Chem. Soc.,
71 S. Tanji, T. Shibata, I. Sato and K. Soai, J. Chem. Soc., Perkin
Trans. 1, 2001, 217.
72 S. Belvedere and R. Breslow, Bioorg. Chem., 2001, 29, 321.
73 V. Amendola, L. Fabbrizzi, C. Mangano, H. Miller, P. Pallavicini,
A. Perotti and A. Taglietti, Angew. Chem., Int. Ed., 2002, 41,
2553.
2
006, 128, 16748.
8 C. Gennari, C. Monti and U. Piarulli, Pure Appl. Chem., 2006, 78,
03.
3
3
4
4
9 J. G. de Vries and L. Lefort, Chem.–Eur. J., 2006, 12, 4722.
0 A. Buryak and K. Severin, J. Comb. Chem., 2006, 8, 540.
1 C. Monti, C. Gennari, U. Piarulli, J. G. de Vries, A. H. M. de Vries
and L. Lefort, Chem.–Eur. J., 2005, 11, 6701.
74 A. Gonzalez-Cabello, P. Vazquez, T. Torres and D. M. Guldi, J.
Org. Chem., 2003, 68, 8635.
75 R. Prabaharan, R. E. Oakes, N. C. Fletcher and M. Nieuwenhuy-
zen, Tetrahedron: Asymmetry, 2004, 15, 2527.
76 T. L. J. Huang and D. G. Brewer, Can. J. Chem., 1981, 59, 1689.
77 E. Figgemeier, L. Merz, B. Hermann, Y. C. Zimmermann, C. E.
¨
Housecroft, H.-J. Guntherodt and E. C. Constable, J. Phys. Chem.
B, 2003, 107, 1157.
78 A. von Zelewsky, Stereochemistry of Coordination Compounds,
Wiley, Chichester, 1996.
79 J. de Mendoza, E. Mesa, J.-C. Rodriguez-Ubis, P. Vazquez, F.
4
4
4
4
4
2 F. Dumitru, E. Petit, A. van der Lee and M. Barboiu, Eur. J. Inorg.
Chem., 2005, 4255.
3 P. Dastidar, S. Okabe, K. Nakano, K. Iida, M. Miyata, N. Tohnai
and M. Shibayama, Chem. Mater., 2005, 17, 741.
4 C. Monti, C. Gennari and U. Piarulli, Tetrahedron Lett., 2004, 45,
6859.
5 M. Horn, J. Ihringer, P. T. Glink and J. F. Stoddart, Chem.–Eur.
J., 2003, 9, 4046.
6 E. L. Doyle, L. Riera and D. S. Wright, Eur. J. Inorg. Chem., 2003,
¨
Vogtle, P. M. Windscheif, K. Risannen, J.-M. Lehn, D. Lilien-
baum and R. Ziessel, Angew. Chem., Int. Ed. Engl., 1991, 30,
1331.
3279.
4
4
7 K. Severin, Chem. Commun., 2006, 3859.
80 N. Weibel, L. J. Charbonniere and R. F. Ziessel, J. Org. Chem.,
2002, 67, 7876.
81 G. Anderegg and V. Gramlich, Helv. Chim. Acta, 1994, 77, 685.
82 D. A. Buckingham and C. R. Clark, Inorg. Chem., 1994, 33, 6171.
83 H. A. Gazzaz, N. M. E-Guindi and A. A. El-Awady, J. Chem.
Soc., Dalton Trans., 1993, 2313.
8 P. J. Corbett, J. Leclaire, L. Vial, K. R. West, J.-L. Wietor, J. K.
M. Sanders and S. Otto, Chem. Rev., 2006, 106, 3652.
9 J. R. Nitschke, Acc. Chem. Res., 2007, 40, 103.
4
5
0 E. C. Constable, C. E. Housecroft, T. Kulke, C. Lazzarini, E. R.
Schofield and Y. Zimmermann, J. Chem. Soc., Dalton Trans., 2001,
2
864.
84 D. J. Aymes, A. Suifil, E. Collange and M. R. Paris, Bull. Soc.
Chim. Fr., 1990, 145.
85 S. A. Bedell and A. E. Martell, J. Am. Chem. Soc., 1985, 107,
7909.
86 R. J. Motekaitis, A. E. Martell and D. A. Nelson, Inorg. Chem.,
1984, 23, 275.
5
5
1 V. Goral, M. I. Nelen, A. V. Eliseev and J. M. Lehn, Proc. Natl.
Acad. Sci. USA, 2001, 98, 1347.
2 M. Hochgurtel, H. Kroth, D. Piecha, M. W. Hofmann, C.
Nicolau, S. Krause, O. Schaaf, G. Sonnenmoser and A. V. Eliseev,
Proc. Natl. Acad. Sci. USA, 2002, 99, 3382.
5
5
5
3 T. Nishinaga, A. Tanatani, K. C. Oh and J. S. Moore, J. Am.
Chem. Soc., 2002, 124, 5934.
4 E. C. Constable, T. Kulke, M. Neuburger and M. Zehnder, New J.
Chem., 1997, 21, 1091.
5 E. C. Constable, in Education in Advanced Chemistry, Perspectives
in Coordination Chemistry, ed. A. M. Trzeciak, P. Sobota and J. J.
87 R. J. Motekaitis, A. E. Martell, J.-M. Lehn and E. I. Watanabe,
Inorg. Chem., 1982, 21, 4253.
88 R. M. Jellish and L. C. Thompson, J. Coord. Chem., 1975, 4,
199.
89 G. McLendon and A. E. Martell, J. Coord. Chem., 1975, 4, 235.
90 J. E. Prue and G. Schwarzenbach, Helv. Chim. Acta, 1950, 33, 963.
This journal is ꢀc the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2007
New J. Chem., 2007, 31, 1437–1447 | 1447