Functionalised azetidines as ligands: Pyridyl-complemented coordination

Article abstract of DOI:10.1039/b812295d

A relatively simple azetidine derivative with 2-aminoethyl and 2-pyridyl substituents functions as a tridentate ligand towards Cu(ii), giving species in which the CuN₃ unit is essentially planar and that, in the solid state, show a subtle balan

Full text of DOI:10.1039/b812295d

View Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
Functionalised azetidines as ligands: pyridyl-complemented coordination†
Young Hoon Lee,a Jack Harrowfield,b Yang Kim,*c Woo Taik Lim,d Yu Cheol Parka and Pierre Thue´rye
Received 18th July 2008, Accepted 16th September 2008
First published as an Advance Article on the web 13th November 2008
DOI: 10.1039/b812295d
A relatively simple azetidine derivative with 2-aminoethyl and 2-pyridyl substituents functions as a
tridentate ligand towards Cu(II), giving species in which the CuN3 unit is essentially planar and that, in
the solid state, show a subtle balance between mononuclear and binuclear forms, allowing both to be
simultaneously present in some lattices. Displacement of the ligand from Cu(II) by the use of cyanide
results in the formation of cyanogen, which appears to slowly react with the free amine to give an
oxamidine characterised in its doubly-deprotonated form as its binuclear Ni(II) complex.
Introduction
While the coordination chemistry of azetidine itself seems to
be quite limited,1–3 functionalised azetidines, commonly pro-
duced as intermediates in polyamine formation reaction,3 are
ligands which are strongly bound, at least by transition metal
ions.3–7 A significant feature of known structures of coordinated
Fig. 1 Synthesis of 1-aminoethyl-3-methyl-3-(2-pyridyl)azetidine, L.
N-aminoethyl azetidines is that the metal–azetidine N bond,
despite the tertiary nature of the donor atom,8 is even shorter than
means that spectroscopic measurements are of limited value in
(as is certainly the case for secondary azetidine-N4), or at most
defining the exact nature of its complexes and hence we have used
only very slightly extended3 compared to a metal–primary amine
single-crystal X-ray structure determinations as a basic method of
N bond. This indicates that an N-functionalised azetidine can
characterisation.
be considered as a strong donor and one which is immune to the
oxidation to imine species, which commonly occurs with N-donors
retaining a hydrogen substituent.9,10 Another N-donor similarly
resistant to oxidation is that of pyridine. (Although formally an
imine-N donor, delocalisation in the pyridine ring renders the
Results and discussion
Hydroxymethylation of 2-ethylpyridine13 leads to 2-methyl-2(2-
pyridyl)propane-1,3-diol, which after O-sulfonylation, provides
a useful intermediate for the synthesis of pyridyl–polyamines,
the present work being concerned with the reaction of this
intermediate with 1,2-diaminoethane (Fig. 1). Although we have
not made a precise kinetic study of the reaction of formation
of 1-aminoethyl-3-methyl-3-(2-pyridyl)azetidine, L, the reaction
conditions used, chosen to be similar to those used to successfully
obtain a related azetidine derivative,3,6 appear to result in a good
yield of the desired ligand, with no obvious contribution from the
ring-opening reaction of the azetidine in the presence of a large
excess of 1,2-diaminoethane at ~90 ◦C. While the bulk of the excess
diamine can be readily distilled from the final reaction mixture,
further purification of the viscous, high boiling product as the free
imine unit much less susceptible to nucleophilic addition than in
true imines.) Thus, the combination of pyridyl and azetidine units
within a multidentate ligand containing other amine centres
raises the prospect of forming strongly-coordinating polyamines
that are relatively resistant to oxidative degradation or that may
undergo selective oxidation reactions. We report herein some of
the basic coordination chemistry, established in terms of structural
characterisation of a variety of complexes (mainly of Cu(II)),
including three describable as intimate “molecular alloys”,11 for a
ligand of this type, 1-aminoethyl-3-methyl-3-(2-pyridyl)azetidine,
L (Fig. 1). This work forms part of our broader studies12 of
the utility of four-membered ring heterocycles in coordination
chemistry. As in other cases,6 the unsymmetrical nature of L and
the formation of complexes involving paramagnetic metal ions
amine is difficult and we have found (as in similar instances3,6,12
)
that formation of a metal complex is the most efficient way to rid
the ligand of minor contaminants.
aDepartment of Chemistry, Kyungpook National University, Daegu 702-701,
South Korea
As we will show elsewhere,14 the azetidine ring of L can be
preserved through reactions involving a variety of nucleophiles
and leading to a number of functionalised multidentate ligands,
so that L can in fact be considered as the basis of a quite
extensive coordination chemistry. In the structurally characterised
complexes described below, L acts as a “planar” tridentate ligand
and this apparent predilection of the three N-donor atoms to lie in
a plane with the metal reflects that seen in related systems where
all donor atoms involve saturated N.3,6 In all cases, L is chiral in its
coordinated form due to the conformation, d or l, of the chelate
bInstitut de Science et d’Inge´nierie Supramole´culaires, Universite´ Louis
Pasteur, 8, alle´e Gaspard Monge, 67083 Strasbourg, France
cDepartment of Chemistry and Advanced Materials, Kosin University, 149-1,
Dongsam-dong, Yeongdo-gu, Busan 606-701, South Korea
dDepartment of Applied Chemistry, Andong National University, Andong,
Kyungpook, South Korea
eCEA Saclay, DSM/IRAMIS/SCM (CNRS URA 331), Baˆtiment 125,
91191 Gif-sur-Yvette, France
† CCDC reference numbers 678919–678925. For crystallographic data in
CIF or other electronic format see DOI: 10.1039/b812295d
434 | Dalton Trans., 2009, 434–442
This journal is
The Royal Society of Chemistry 2009
©

View Online
ring unit involving the azetidine-N and primary-amino-N donor
atoms.
Cu1¢¢ hydrogen bonds (symmetry code ¢¢ = 1 - x, 1 - y, 1 - z) cause
association. These dimers are present in a 1 : 1 molar ratio and
alternate in the “double sheets” by forming columns parallel to b.
Although in the given projection, the orientations of the monomer
components of adjacent, different dimers appear similar, if one
pair is considered to have d(up), l(down) conformations of
the 1,2-diaminoethane unit chelate rings, its neighbour (of the
other type) has l(up), d(down), and significant rotations of the
Cu(L)Cl units would be required to transform one dimer into
the other (Fig. 3(b)). To use the term “dimer” for either unit
is imprecise, since the seeming bis(complex) units are actually
linked into infinite chains by NH ◊ ◊ ◊ Cl hydrogen bonding and
complex units in the different chains appear to be involved in recip-
rocal (1,2-diaminoethane unit)CH2–p interactions. These chains
lie side by side in puckered sheets parallel to the ab plane and views
(Fig. 3(c)) along c of the two types of sheet further illustrate both
their similarities and their differences. In particular, the chains of
the non-chloro-bridged species do not appear to be cross-linked,
except by remote (aromatic)CH ◊ ◊ ◊ Cl contacts, while the chloride
ion “released” by chloro-bridge formation becomes involved
in symmetrical NH ◊ ◊ ◊ Cl interactions, which do cross-link the
chains of the other sheet (again possibly augmented by CH ◊ ◊ ◊ Cl
interactions). In terms of bonding changes in passing from one
sheet to another, it can be said that a Cu–Cl–Cu unit passes
to Cu–Cl ◊ ◊ ◊ HN as NH ◊ ◊ ◊ Cl ◊ ◊ ◊ HN becomes NH ◊ ◊ ◊ Cl–Cu,
so that it would be unsurprising if the difference in energy
were very small, thus explaining the simultaneous occurrence
of both. While the structure is not strictly identical to those
of so-called “molecular alloys”11 where a single crystal contains
macroscopic zones of distinct phases, it can be regarded as
defining the ultimate limit of intimate mixing of two species in
a crystal and the considerations above provide some indication
of reasons for the formation of such species in general. It is
remarkable that the complexes of Cu(II) chloride with a facially-
coordinating triamine (1,3,7-trimethyl-1,3,7-triazacyclononane)11
differ in a manner parallel to that of the present species and
perhaps this is in some way associated with the weakness of Cu(II)
interactions with a fifth donor atom. The presence of a primary
amino centre in L is clearly important in giving rise to H bonding
interactions, which may be competitive with Cu–Cl coordination
and, while 1,3,7-trimethyl-1,3,7-triazacyclononane contains no
NH centres, its two structurally characterised complexes do differ
in regard to incorporation of the H-bond donor, water, in the
lattice.
The crystal structure determination of the bis(benzenesulfonate)
1 confirms its anticipated molecular nature (Fig. 2(a)). A view of
the lattice down b (Fig. 2(b)) shows particularly clear evidence of
proximity of the aromatic entities presumably indicative of weak
interactions,15 although all C ◊ ◊ ◊ C contacts are relatively long,
16
˚
viz., >3.6 A, and a Hirshfeld surface analysis of intermolecular
contacts indicates that the most important aromatic–aromatic
approaches are essentially “edge-to-edge” (but with the rings
slightly tilted from co-planarity and C18(x, y, z) ◊ ◊ ◊ C21(x, 1 + y,
˚
z) 3.74 A, little different from that in some edge-to-face contacts),
and are accompanied by CH3 ◊ ◊ ◊ O(sulfony◦l) interactions (C ◊ ◊ ◊ O
˚
3.38 A). The (O)CCC(O) bond angle is 111 , indicating that there
is no structural aspect of this molecule favouring the formation of
a small (azetidine) ring upon its reaction with 1,2-diaminoethane,
though this of course, may not be true of the mono-substituted
reaction intermediate.
Whatever the reasons for the formation of lattices of mixed
species, its occurrence here is subject to seemingly subtle influ-
ences, again as observed for the triazacyclononane complexes.11
Thus, crystallisation from methanol rather than dichloromethane
suffices to produce a lattice containing only “monomer” units,
[Cu(L)Cl2]·H2O of the hydrated complex in 3. In fact, once again it
is possible to regard the lattice as containing dimer units involving
centrosymmetric monomer pairs linked by double NH ◊ ◊ ◊ Cl
hydrogen bonds as those in 2, though the conformations are
significantly different. A view of the lattice down a (Fig. 4(a))
shows that it has a layered structure, with the water molecules
serving to bridge the layers through Cl ◊ ◊ ◊ HOH ◊ ◊ ◊ Cl interactions.
Within the layers, the dimers linked by H-bonds (Fig. 4(b))
appear to associate also through reciprocal azetidine–CH2–p
interactions.
Fig. 2 Upper: the molecular structure of the bis(benzenesulfonate) of
2-methyl-2-(2-pyridyl)propan-1,3-diol as found in the lattice of 1. Lower:
a view of the lattice of 1 down b, illustrating the layered structure of
the array. As in all figures, displacement ellipsoids are shown at the 50%
probability level and H atoms are omitted for clarity.
A convenient basis for an analysis of the lattice of what might
be described as an intimate “molecular alloy”, [{Cu(L)Cl}(m-
Cl){Cu(L)Cl}]Cl·2[Cu(L)Cl2], 2, is to take the view down b
(Fig. 3(a)). Thus, the first impression is that the lattice is composed
of dimeric units, which form double sheets of Cu atoms parallel to
the bc plane. In fact, there are two types of centrosymmetric dimer,
one in which a linear Cu2–Cl3–Cu2¢ bridge is the link (symmetry
code ¢ = 2 - x, 2 - y, 2 - z) and another in which two N1H ◊ ◊ ◊ Cl1¢¢–
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 434–442 | 435
©

View Online
Fig. 3 (a) A partial view of the lattice of the “molecular alloy” 2 down b, with double layers (seen edge-on) of the Cu(II) complex units in which every
second pair of complex units is bridged by chloride while the other pairs, bridged by H-bonds, have two chloride ligands nearly superimposed in this
view. (b) A perspective view of two adjacent dimer units within a “double layer”, showing the different relative orientations of monomer units. H-bonds
are indicated by dotted lines. (Symmetry codes: ¢ = 2 - x, 2 - y, 2 - z; ¢¢ = 1 - x, 1 - y, 1 - z.) (c) Partial views, down c, of the sheets of complexes lying in
planes parallel to ab, showing how H-bonds (dotted lines) crosslink chains of complex units only in the case of the chloro-bridged dimers.
When ammonium hexafluorophosphate is added to an aqueous
solution of [Cu(L)Cl2], the precipitate obtained, [Cu(L)Cl(PF6)],
4, has a lattice in which chains of hexafluorophosphate-bridged,
approximately square-planar [Cu(L)Cl]+ units lie parallel to c.
Although the Cu and P atoms are co-linear, the axis of the trans-
FPF unit involved in bridging is slightly inclined to c and there is
a slight asymmetry in the bridging bonds (Fig. 5). The hydrogen-
bonded dimer unit, seen in all the above structures, is found once
again in that for every columnar chain there is one adjacent,
which provides complex units suitably oriented for reciprocal
NH ◊ ◊ ◊ Cl interactions to occur (in addition to NH ◊ ◊ ◊ F). This
motif seems to be maintained despite numerous associated (methyl
and methylene)CH ◊ ◊ ◊ F and NH ◊ ◊ ◊ F contacts, which appear to
indicate that the hexafluorophosphate units are important not
only in creating the chains of complexes but also in cross-linking
“monomer” lattice) the product here was an intimate molecular
alloy, 5 (Fig. 6), isomorphous with the chloride as obtained
from dichloromethane. Similar changes in X–Cu and X ◊ ◊ ◊ HN
interaction energies in passing from X = Cl to X = Br may explain
why the same balance between species is retained.
When an aqueous solution containing both sodium tereph-
thalate (Na2tpa) and sodium perchlorate was added to a
solution of [Cu(L)Cl2] in methanol, a crystalline precipi-
tate, [{Cu(L)Cl}2(tpa)][{Cu(L)(CH3OH)}2(tpaH2)](ClO4)4, 6, of
a rather unexpected stoichiometry was obtained. The origin of
the protons added to the terephthalate is obscure but apparent H-
bonding interactions with chloride (see below) indicate that at least
one of the terephthalate units is fully protonated. The crystal struc-
ture of 6 provides a third example of a lattice in which two related
but significantly different complexes of L are present in a 1 : 1 ratio.
Though some uncertainty as to proton locations remains, these
two species can be described as the terephthalate-bridged dimers
[{Cu(L)Cl}2(tpa)] and [{Cu(L)(CH3OH)}2(tpaH2)](ClO4)4. They
are segregated in the lattice into sheets, which can be seen edge on
in the view down c (Fig. 7) and, within these sheets, the dimers
have significantly different orientations, though these are such
as to allow H-bonding between dimers in adjacent sheets, which
results in infinite “ladders” of alternating units. In contrast to the
systems described above where NH ◊ ◊ ◊ Cl interactions appear to
˚
these chains, with one of the shorter contacts (2.82 A) indeed
being F ◊ ◊ ◊ F. This is true also for adjacent chains, which are
not oriented so as to favour NH ◊ ◊ ◊ Cl interactions, though here
CH ◊ ◊ ◊ F contacts involving both aliphatic and aromatic hydrogen
may be augmented by CH2– and CH3–p interactions.
Further indication of the complexity of factors influencing the
appearance of crystals of mixed species is the fact that when
[Cu(L)Br2] was crystallised from methanol (the solvent, which
in the case of the chloride analogue, provided a homogeneous
436 | Dalton Trans., 2009, 434–442
This journal is
The Royal Society of Chemistry 2009
©

View Online
Fig.
5
Partial view of adjacent chains of the hexafluorophos-
phate-bridged complex units present in complex 4 between which cen-
trosymmetric, H-bonded pairs can be discerned. Transformation of the
asymmetric unit -x, 1 - y, 2 - z.
Fig. 6 Two adjacent centrosymmetric pairs of complex units within the
“molecular alloy” 5. (Symmetry codes: ¢ = 1 - x, 1 - y, 1 - z; ¢¢ = -x, -y,
˚
-z.) The distance N ◊ ◊ ◊ Br in the H-bonded pair is 3.49(3) A.
Fig. 4 Upper: a partial view, down a, of the lattice of compound 3,
showing its layered structure. (The large, dark ellipsoids are those of
the water molecule oxygen atoms.) Lower: one of the centrosymmetric,
H-bonded pairs found within the layers of 3. (Symmetry code: ¢ = -x, 2 -
y, 1 - z.)
be important, in 6 the interactions of the ligand amino-group
protons appear to involve exclusively perchlorate oxygen, and
these interactions serve to link species both within and between
the sheets of the different dimers.
Removal of a ligand from Cu(II) using cyanide is expected to
produce cyanogen and this in turn is well known to undergo
reactions with amines.17 While the use of crude L to form the
Cu(II) complexes presently described provided no evidence of
the presence of cyanogen adducts, in the case of Ni(II), the
product 7 crystallographically characterised was clearly derived
from cyanogen, this perhaps reflecting the extended period of
storage of the crude ligand prior to its usage or a reaction catalysed
by Ni(II) in the solvent (CH3OH) used. (When the ligand was
isolated by treating the Cu(II) complex with Na2S, the addition
of Ni(II) gave none of the highly insoluble complex 7.) Thus, the
ligand L¢ found in the centrosymmetric binuclear Ni(II) complex is
presumed to have been formed by the reaction shown in Scheme 1.
The presence of but one chloride anion per nickel ion in the
complex indicates that the diamidine (oxamidine) ligand must
Scheme 1
be doubly deprotonated (as in [Ni2(L¢-2H)]Cl2·4H2O), a state
that would presumably be favoured by delocalisation of the
charge over the whole diamidine unit. Significantly, a chloride
˚
ion lies within 3.45 A of each unalkylated amidine-N of every
dimer and this is consistent with retention of a proton at these
sites, so that it appears that deprotonation must occur at the
coordinated, dialkylated amidine-N. (Alternative formulations
such as [Ni2(L¢)]Cl2(H3O2)2, with the hydrated hydroxide ion being
a well-known solid state species,18 are possible but seem unlikely
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 434–442 | 437
©

View Online
interactions. Thus, when the lattice is viewed down a, columns
of slipped stacks are seen in projection and, within these stacks,
close contacts are apparent between aromatic (pyridine) carbon
atoms and atoms of the amidine group. As well, pyridine units
of different columns lie parallel, producing here close contacts of
aromatic carbons only (Fig. 8).
Fig. 8 Upper: a partial view, down a, of the lattice of complex 7 showing
the stacked sheets of complex cations bridged by an H-bonding network
of chloride ions and water molecules. Lower: a perspective view of two
adjacent cations (displaced one unit along a) in the lattice of complex
7 showing the parallel array of the coordination planes and part of the
H-bonding network.
In all the complexes described above, the M–N bond lengths
for each of the three types (pyridine, azetidine and primary
amino) vary somewhat (Table 1). Mean values for the present
cases are consistent with those for related systems in the literature
˚
(Table 1) in indicating that M–N(azetidine) is ~0.03 A longer
than M–N(primary amine). This difference is much smaller
than that between an unconstrained tertiary amine donor and
a primary species19 and supports the notion that the donor
strength of azetidine-N is minimally affected by steric effects of its
substituents. On Cu(II), azetidine-N seems, on average, essentially
identical to pyridine-N, though in the one case available for Ni(II),
Ni–N(azetidine) is significantly shorter than Ni–N(pyridine). In
the case of this Ni complex, there may be an effect associated with
the restraints of the chelate ring formed involving the deprotonated
amidine unit and such a consideration raises the issue of whether
the pyridine and ethylamino arms of L serve to “anchor” the
azetidine-N on Cu(II), as has been suggested may be the case for
the tertiary donor of the ligand “tren”,20 for example. There appear
Fig. 7 Upper: a partial view of the lattice of complex 6 down c, showing
the appearance of layers alternating in thickness. Lower: the H-bonded
ladder of alternating forms of the different dimers found in 6. The contacts
Cl ◊ ◊ ◊ O(carboxylate) and (methanol)O ◊ ◊ ◊ O(carboxylate) are 2.70(2) and
˚
2.65(2) A, respectively.
given that the anhydrous triflate, [Ni2(L¢-2H)](CF3SO3)2, can be
crystallised and appears to contain the same form of the ligand.)
The chloride ions and water molecules form an infinite network
linking all the complex units together. The Ni centres are square
planar (four coordinate) and the relatively flat binuclear complex
entities form a lattice in which there is evidence for p-stacking
438 | Dalton Trans., 2009, 434–442
This journal is
The Royal Society of Chemistry 2009
©

View Online
Table 1 Metal–nitrogen bond lengths in complexes of azetidine derivatives
Liganda
M–N(primary)/A
M–N(azetidine)/A
M–N(pyridine)/A
Ref.
˚
˚
˚
Metal
Co(III)
Co(III)
Cu(II)d
Cu(II)e
Cu(II)f
Cu(II)g
Cu(II)h
Cu(II)i
A
B
2.01
1.98b
—
—
4
3, 6
1.96c
1.98c
L
1.999(3)
1.992(2)
2.008(3)
1.9952(15)
1.993(6)
2.004(6)
2.029(3)
2.015(3)
2.004(2)
—
2.038(3)
2.062(3)
2.038(2)
1.9975(14)
2.043(7)
2.054(7)
2.018(3)
2.003(3)
2.033(2)
1.9134(17)
2.022(3)
2.028(3)
2.049(2)
2.0498(15)
2.014(7)
2.007(6)
2.058(3)
2.050(3)
2.033(2)
1.9496(19)
Present work
Present work
Present work
Present work
Present work
Present work
Present work
Present work
Present work
Present work
L
L
L
L
L
Cu(II)j
L
Cu(II)k
Cu(II), mean
Ni(II)l
L
L
L¢–2H
a
.
b Secondary azetidine-N donors; mean value for a tris(ligand) complex. c Mean values for four
different complex species, also containing secondary-N donors for which Co–N = 1.97 A. d Chloro-bridged dimer in complex 2. e “Monomer” chloride
˚
i
in complex 2. f Complex 3. g Complex 4. h Bromo-bridged dimer in complex 5 (mean values for two inequivalent molecules). “Monomer” bromide in
complex 5 (mean values for two inequivalent molecules). j Complex 6, chloro species. k Complex 6, methanol species. l Complex 7.
to be no anomalies in either the Cu–N(pyridine) or Cu–N(amine)
bond lengths, however, so it is assumed that the Cu–N(azetidine)
distance is a “relaxed” value.
Experimental
Reagents and procedures
Although we have not as yet made any detailed study of the
nature of the complexes presently described in solution, the ESR
spectra of frozen DMSO solutions (example provided in ESI†) are
all similar and consistent with the presence of mononuclear Cu(II)
species in this relatively strongly coordinating solvent. The colours
of solutions of a given complex in solvents such as water, ethanol
and dichloromethane, however, are all quite different, indicating
that it may be possible to use particular solvents to preserve the
observed solid state structures.
2-Ethylpyridine, ethane-1,2-diamine and other basic chemicals
were purchased from Aldrich Chemical Co. and used as received.
Ion-exchange chromatography was conducted in glass columns
under gravity flow using SP-Sephadex C25 (Na+ form) ion-
exchange resin. Electronic absorption spectra were measured
on a SCINCO S-2100 diode-array spectrophotometer and IR
spectra on a Jasco 715 spectrometer. 1H NMR spectra (in CDCl3)
were recorded on a Bruker Advance Digital 400 instrument. All
solution measurements were made at 25
1
◦C. EPR spectra
were obtained with a JES-TE 300 spectrometer (9450 MHz) as
~1 ¥ 10-3 mol L-1 frozen (liquid nitrogen temperature) solutions in
DMSO. Elemental analyses were carried out with a Chemtronics
TEA-3000 instrument.
Conclusions
The ligand 1-aminoethyl-3-methyl-3-(2-pyridyl)azetidine acts as a
tridentate species favouring a co-planar array of its three N-donor
atoms with a bound metal. This indicates that it may be useful as a
species to partially block coordination sites on a square planar (or
square pyramidal) metal ion and we will report elsewhere on our
efforts to evaluate the utility of this in catalysis promoted by the
Pd(II) complex.21 Although an unusual ligand in the sense that its
three nitrogen centres are of formally different types, these appear
to be of similar donor strength and we have not encountered
circumstances where partial binding occurs. The protons on the
primary amino-N centre, however, do provide a unique site for
H-bonding interactions in the complexes. Such interactions appear
to be important in leading to the unusually high incidence of
lattices containing mixed complex species seen in the present work.
The primary amine centre is also unique in that it may become
involved in condensation reactions such as that with cyanogen
reported here and we will report subsequently on more extensive
investigations of such reactions for the production of a variety of
derivatives of the ligand.
Syntheses
2-Methyl-2-(pyridin-2-yl)propane-1,3-diol bis(benzenesulfonate),
1. An autoclave was charged with 2-ethylpyridine (15.0 g) and
aqueous formaldehyde (35%, 60.0 g), and heated at 140 ◦C for
72 h. After cooling to room temperature, the yellow solution was
evaporated under reduced pressure to give a viscous, yellow oil
that was used without further purification. The whole product
was dissolved in pyridine (200 mL) and cooled to 0 ◦C, then
benzenesulfonyl chloride (50.0 g) was added gradually from a
dropping funnel over 2 h. The resulting mixture was allowed to
warm to room temperature as it was stirred for another 12 h. It was
then poured into ice-water (600 mL), giving a brown precipitate.
The precipitate was collected by filtration and washed on the filter
with water, then methanol to remove the brown colouration, and
finally dried in a vacuum desiccator. Yield: 40 g. Recrystallisation
from boiling methanol provided colourless crystals suitable for
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 434–442 | 439
©

View Online
crystallography. Anal. calcd for C22H22O6S2: C 59.17, H 4.97.
Found: C 59.1, H 4.9%. H NMR (CDCl3) d/ppm: 8.33, 8.32
(d, 1H), 7.78, 7.76 (d, 4H), 7.67, 7.65, 7.63 (dist. t, 3H), 7.54, 7.52,
7.50 (t, 4H), 7.21, 7.19 (d, 1h), 7.17, 7.16, 7.15 (t, 1H), 4.32 (s, 4H),
1.38 (s, 3H).
desired ligand. The 1H NMR spectrum (in CDCl3) of the material
immediately after isolation, however, indicated it to be largely
a single species, with the azetidine ring protons forming a well-
resolved AB doublet pair at d/ppm: 3.20, 3.18 and 3.12, 3.10, the
ethylene unit showing two clear triplets at d/ppm: 2.35, 2.33, 2.32
and2.24, 2.22, 2.21andthe methylgroup giving asingletat d/ppm:
1.33 with only baseline-level side peaks. 13C NMR (100 MHz;
CDCl3) d/ppm: 27.47, 38.76, 41.09, 59.82, 65.24, 119.96, 121.37,
136.97, 149.13, 166.45 plus numerous minor peaks with <10%
of the intensity of any of these major peaks. In work conducted
after completion of the other reactions described herein, the free
ligand was isolated by treating the Cu complex with Na2S, the
product here giving a 1H NMR spectrum identical with that
cited above. Attempts to purify both the ligand stripped from
the Cu(II) complex or that obtained directly from the reaction of
the bis(sulfonate) by chromatography on silica failed due to the
seemingly irreversible adsorption of the polyamine.
1
[{Cu(L)Cl}(l-Cl){Cu(L)Cl}]Cl·2[Cu(L)Cl2] (“molecular al-
loy”), 2, Cu(L)Cl2·H2O, 3 and [Cu(L)Cl(PF6)], 4. 2-Methyl-2-
(pyridin-2-yl)propane-1,3-diol bis(benzenesulfonate) (20.0 g) was
thoroughly mixed with ethane-1,2-diamine (30.0 g) then heated
at 90 ◦C for 72 h under N2. The bulk of the excess ethane-1,2-
diamine was distilled out under reduced pressure and a methanolic
solution of NaOH (3.6 g) was added. After stirring for 0.5 h, the
precipitate was filtered off and the solvent in the filtrate removed
under reduced pressure. The residual yellow oil was dissolved in
methanol (100 mL) and CuCl2·2H2O (7.6 g in 200 mL methanol)
was added to give a blue solution, which was then evaporated to
dryness under reduced pressure. The residue was dissolved in water
(1 L), the solution filtered and applied to a column of Na+-form SP
Sephadex. Elution with 0.3 mol L-1 NaCl separated a major blue-
violet component. Its eluate was taken to dryness under reduced
pressure and the residue extracted with ethanol to leave as much
as possible of the NaCl behind. The extract was taken to dryness
and the blue solid extracted with hot CH2Cl2 (3 ¥ 50 mL), filtered,
then set aside to give a green solution and a blue precipitate. The
precipitate was collected, washed with a little cold CH2Cl2 and
then redissolved in boiling CH2Cl2. Blue crystals suitable for a
structure determination and thereby identified as [{Cu(L)Cl}(m-
Cl){Cu(L)Cl}]Cl·2[Cu(L)Cl2]), 2, deposited as the solution cooled
and slowly evaporated. The crystals were collected, washed with
cold hexane and then dried under vacuum. Yield: 7.4 g. Anal. calcd
for C11H17Cl2CuN3: C 40.56, H5.26, N12.90. Found:C 40.4, H 5.3,
[{Cu(L)Br}(l-Br){Cu(L)Br}]Br·2[Cu(L)Br2] “molecular alloy”
5. CuBr2 (0.28 g) was added to a solution of crude ligand (L)
(0.24 g) in methanol (10 mL) giving a green solution, which
deposited crystals suitable for structure determination upon slow
evaporation at room temperature. Yield: 0.48 g. Anal. calcd for
C11H17Br2CuN3: C 31.86, H 4.13, N 10.13. Found: C 32.0, H 4.0,
N 10.1%. Visible spectrum (in water): lmax = 622.0 nm, logemax
=
2.06.
[{Cu(L)Cl}2(tpa)][{Cu(L)(CH3OH)}2(tpaH2)](ClO4)4 “molecu-
lar alloy”, 6. Cu(L)Cl2·H2O (0.5 g) was dissolved in methanol
(5 mL) and mixed with a solution of sodium terephthalate (Na2tpa,
0.15 g) and LiClO4 (0.10 g) in water (5 mL). The reaction mixture
was stirred for 0.5 h and left to stand at room temperature for
one week to give blue crystals suitable for crystallography. Yield:
0.50 g. Anal. calcd for C61H80Cl4Cu3N12O17: C 46.20, H 5.08,
N 10.60. Found: C 46.1, H 5.2, N 10.7%. Visible spectrum (in
water): lmax = 626.0 nm, logemax = 2.26.
[Ni2(L¢–2H)]Cl2·4H2O, 7, L¢ = the product of addition of two
moles of L to cyanogen. NiCl2·6H2O (0.97 g) was added to a
methanolic solution of crude ligand L (0.5 g in 20 mL) and
stirred for 1 h to give a dark orange solution, which was then
allowed to slowly evaporate at room temperature to give orange
crystals suitable for crystallographic analysis. Visible spectrum (in
water): = 451.0 nm, logemax = 2.26; lmax = 546 nm, logemax = 2.31.
Yield: 0.73 g. Anal. calcd for C24H40Cl2N8Ni2O4: C 41.60, H 5.82,
N 16.17. Found: C 40.9, H 5.7, N 15.8% (indicative of a slightly
higher degree of hydration (~4.5H2O) of the bulk relative to the
single crystal taken for the structure determination).
N 12.8%. Visible spectrum (in water): lmax = 633.5 nm, logemax
=
2.07.
The green filtrate was evaporated to dryness and the residue
dissolved in the least amount of methanol. Slow evaporation at
room temperature gave green crystals of Cu(L)Cl2·H2O, 3, suitable
for a structure determination. They were filtered, washed with cold
hexane and then dried under vacuum. Yield: 1.8 g. Anal. calcd for
C11H19Cl2CuN3O: C 38.44, H 5.57, N 12.22. Found: C 38.6, H 5.5,
N 12.1%. Visible spectrum (in ethanol): lmax = 725.0 nm, logemax
=
2.17.
The filtrate from this second precipitate clearly contained fur-
ther complex and, on evaporation of the methanol, redissolution
of the purplish residue in the minimum volume of ethanol and
addition of NH4PF6, purple crystals of [Cu(L)Cl(PF6)], 4, suitable
for a structure determination were obtained as the solvent slowly
evaporated. Yield: 1.1 g. Anal. calcd for C11H17ClCuF6N3P: C
30.36, H 3.94, N 9.65. Found: C 30.9, H 3.9, N 9.7%. Visible
spectrum (in water): lmax = 545.0 nm, logemax = 2.03.
Isolation of the free ligand, L. Cu(L)Cl2·H2O, (2.0 g) was
dissolved in water (50 mL) and NaCN (1.0 g) added with stirring.
The colour of the solution rapidly changed to yellow and it was
then extracted with CH2Cl2 (3 ¥ 20 mL). The combined extracts
were washed with water (2 ¥ 10 mL), dried with anhydrous
sodium sulfate and evaporated under reduced pressure to give a
yellowish-brown oil. Yield: 1.0 g. The ligand was used to prepare
complexes without further purification, the reactions showing (see
below and Discussion) that cyanogen was present along with the
Crystallography
The data for compounds 2 and 4 were collected at 100(2) K
on a Nonius Kappa-CCD area detector diffractometer22 using
˚
graphite-monochromated MoKa radiation (l = 0.71073 A). The
crystals were introduced in glass capillaries with a protecting
“Paratone-N” oil (Hampton Research) coating. The unit cell
parameters were determined from ten frames, then refined on all
data. The data (j- and w-scans) were processed with HKL2000.23
The structures were solved by direct methods (2) or Patterson
map interpretation (4) with SHELXS-97 and subsequent Fourier-
difference synthesis and refined by full-matrix least-squares on F2
440 | Dalton Trans., 2009, 434–442
This journal is
The Royal Society of Chemistry 2009
©

View Online
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 434–442 | 441
©

View Online
with SHELXL-97.24 Absorption effects were corrected empirically
with the program SCALEPACK.23 All non-hydrogen atoms were
refined with anisotropic displacement parameters. The hydrogen
atoms bound to nitrogen atoms were found on Fourier-difference
maps and all the others were introduced at calculated positions;
all were treated as riding atoms with a displacement parameter
equal to 1.2 (NH, CH, CH2) or 1.5 (CH3) times that of the parent
atom.
References
1 J. Reedijk in Comprehensive-Coordination Chemistry, ed. G. Wilkinson,
R. D. Gillard and J. A. McCleverty, Pergamon Press, Oxford, 1987,
vol. 2, ch. 13.2, pp. 73–98 (see p. 81 in particular).
2 R. D. Adams and G. Chen, Organometallics, 1993, 12, 2070–2077.
3 J. M. Harrowfield, G. H. Jang, Y. Kim, P. Thue´ry and J. Vicens, J. Chem.
Soc., Dalton Trans., 2002, 1241–1243, and references therein.
4 R. J. Geue, M. G. McCarthy, A. M. Sargeson, P. Jørgensen, R. G.
Hazell and F. K. Larsen, Inorg. Chem., 1985, 24, 2559–2565.
5 L. Keller, M. Vargas Sanchez, D. Prim, F. Couty, G. Evano and J.
Marrot, J. Organomet. Chem., 2005, 690, 2306–2311.
6 J. Harrowfield, J. Y. Kim, Y. Kim, Y. H. Lee and P. Thue´ry, Polyhedron,
2005, 24, 1569–1576.
7 Gas phase binding of alkali metal ions to azetidine-2-carboxylic acid is
discussed in: P. B. Armentrout and R. M. Moision, J. Phys. Chem. A,
2006, 110, 3933–3946.
8 D. Meyerstein, Coord. Chem. Rev., 1999, 185–186, 141–147.
9 (a) A. M. Sargeson, Pure Appl. Chem., 1986, 11, 1511–1522; (b) P.
Bernhard, D. J. Bull, H.-B. Burgi, P. Osvath, A. Raselli and A. M.
Sargeson, Inorg. Chem., 1997, 36, 2804–2815, and references therein;
(c) An example of such oxidation in a ligand-closely related to the
present is given in: J. M. Harrowfield, Y. Kim, Y. H. Lee and P. Thue´ry,
Eur. J. Inorg. Chem., 2003, 2913–2916.
10 F. R. Keene, Coord. Chem. Rev., 1999, 187, 121–149.
11 J. W. Steed, A. E. Goeta, J. Lipkowski, D. Swierczynski, V. Panteleon
and S. Handa, Chem. Commun., 2007, 813–815.
12 P. V. Bernhardt, J. H. Cho, J. M. Harrowfield, J. Y. Kim, Y. Kim,
Sujandi, P. Thue´ry and D. C. Yoon, Polyhedron, 2006, 25, 1811–1822.
13 (a) W. Koenigs and G. Happe, Ber. Dtsch. Chem. Ges., 1902, 35, 1343–
1349; (b) A. Grohmann and F. Knoch, Inorg. Chem., 1996, 35, 7932–
7934.
The data for compounds 1, 3, 5, 6 and 7 were collected at 293(2)
K using an ADSC Quantum 210 collector at Beamline 4A MXW
of the Pohang Light Source. The unit cell determinations and data
˚
collections were done using 0.77 A radiation with a detector-to-
crystal distance of 6.0 cm. Preliminary cell constants and an ori-
entation matrix were determined from 36 sets of frames collected
at scan intervals of 5◦ with an exposure time of 1 s per frame. The
data were processed with HKL200023 and reflections were indexed
using the automated indexing routine of DENZO. A total of 7893,
6286, 41 724, 10 123 and 5406 reflections for compounds 1, 3, 5, 6
and 7 were obtained by collecting 72 sets of frames with 5◦ scans
and an exposure time of 1 s per frame. These highly redundant
data sets were corrected for Lorentz and polarisation effects and a
(negligible) correction for crystal decay applied. The space groups
were assigned using the program XPREP (version 6.13, Bruker-
AXS). The structures were solved by direct methods and refined
on F2 by full-matrix least-squares procedures. All non-hydrogen
atoms were refined using anisotropic displacement parameters.
Hydrogen atoms were included in the structure factor calculations
at idealised positions using a riding model but the positions were
not refined. The structure of 1 was refined as corresponding to a
racemic twin, consistent with a Flack parameter of 0.48 and the
fact that no missing symmetry element was detected by the use
of ADDSYM(PLATON). In the case of complex 5, a somewhat
inferior data set gave a high residual, with the highest residual
electron density peaks being located near to the bromine atoms and
being possibly indicative of unresolved disorder and/or imperfect
absorption corrections.
14 Y. H. Lee, J. M. Harrowfield, J. S. Kim, M. H. Lee, W. T. Lim, Y. C.
Park and P. Thue´ry, Dalton Trans., 2009, DOI: 10.1039/b812298a.
15 W. B. Jennings, B. M. Farrell and J. F. Malone, Acc. Chem. Res., 2001,
34, 885–894.
16 J. J. McKinnon, D. Jayatilaka and M. A. Spackman, Chem. Commun.,
2007, 3814–3816.
17 H. M. Woodburn and B. G. Pautler, J. Org. Chem., 1954, 19, 863–867,
and references therein.
18 K. Abu-Dari, K. N. Raymond and D. P Freyberg, J. Am. Chem. Soc.,
1979, 101, 3688–3689.
19 (a) P. S. Donnelly, J. M. Harrowfield, B. W. Skelton and A. H. White,
Inorg. Chem., 2000, 39, 5817–5830;(b) P. S. Donnelly, J. M. Harrowfield,
B. W. Skelton and A. H. White, Inorg. Chem., 2001, 40, 5645–5652.
20 G. Bombieri, F. Benetello, A. Del Pra, M. L. Tobe and D. A. House,
Inorg. Chim. Acta, 1981, 50, 89–94.
Crystal data and structure refinement parameters are given in
Table 2. The molecular plots were drawn with SHELXTL.25
21 D.-H. Lee, Y. H. Lee, D. I. Kim, Y. Kim, W. T. Lim, J. M. Harrowfield,
P. Thue´ry, M.-J. Jin, Y. C. Park and I.-M. Lee, Tetrahedron, 2008, 64,
7178–7182.
22 Kappa-CCD Software, Nonius B.V., Delft, The Netherlands, 1998.
23 Z. Otwinowski and W. Minor, Methods Enzymol., 1997, 276, 307.
24 G. M. Sheldrick, SHELXS-97, Program for solution of crystal struc-
tures, University of Go¨ttingen, Germany, 1997; G. M. Sheldrick,
SHELXL-97, Program for refinement of crystal structures, University
of Go¨ttingen, Germany, 1997.
Acknowledgements
This work was supported by Kosin University and the Brain Korea
21 program. We gratefully acknowledge support for the use of the
beamline 4A MXW and computing facilities at the Pohang Light
Source, Korea.
25 G. M. Sheldrick, SHELXTL, Version 5.1, Bruker AXS Inc., Madison,
WI, USA, 1999.
442 | Dalton Trans., 2009, 434–442
This journal is
The Royal Society of Chemistry 2009
©