Homo- and hetero-dinuclear anion complexes

Article abstract of DOI:10.1002/chem.200601682

The tripodal system 4, in which urea fragments are appended to the three terminal amine nitrogen atoms of a tris(2-aminoethyl)amine (tren) subunit, includes a Cu^II ion and two anions X^-, according to a cascade mechanism through three

Full text of DOI:10.1002/chem.200601682

FULL PAPER
DOI: 10.1002/chem.200601682
Homo- and Hetero-Dinuclear Anion Complexes
Michela Allevi, Marco Bonizzoni, and Luigi Fabbrizzi*[a]
Abstract: The tripodal system 4, in
which urea fragments are appended to
the three terminal amine nitrogen
the second anion Xꢀ interacts with the
trisurea cavity, but this occurs only if
the stronger H-bond acceptors, such as
N3 and H2PO4 , are used. Binding of
the second Xꢀ ion is favoured by the
preorganising effect exerted by the
metal and disfavoured by the steric and
electrostatic repulsions between the
anions. Under the appropriate condi-
tions, heterodinuclear complexes of for-
ꢀ
ꢀ
atoms of
a
tris(2-aminoethyl)amine
(tren) subunit, includes a CuII ion and
two anions Xꢀ, according to a cascade
mechanism through three well defined
stepwise equilibria in a DMSO solu-
mula [CuII(4)(Cl)
(H2PO4)] can be ob-
G
tained in solution, in which Clꢀ is
ꢀ
bound to the metal centre and H2PO4
interacts with the trisurea compart-
ment.
Keywords: anions · dinuclear com-
plexes · hydrogen bonds · molecular
recognition · polyamines
tion. The first anion Xꢀ (halide, N3 ,
ꢀ
ꢀ
ꢀ
NCSꢀ, NO2 , H2PO4 ) seeks the CuII
centre coordinated by the tren moiety;
Introduction
be supported by a given organic scaffold (e.g., 1,3,5-triethyl-
benzene),[12] thus providing a cavity suitable for establishing
both electrostatic and multipoint hydrogen-bonding interac-
tions with the anion. In this case, selectivity is based on the
intensity of the interaction and on matching the geometrical
features of the receptor (cavity size and position of the bind-
ing sites) and anion (size and shape, if polyatomic). Anion
binding fragments, such as amides,[13] ureas,[14] and pyr-
roles,[15] can also be neutral H-bond donors. In this case, the
energy of the receptor–anion interaction is lower than that
observed for positively charged receptors and, in most cases,
recognition studies have to be carried out in aprotic media
(CH2Cl2, MeCN, DMSO, in order of increasing polarity).
Artificial receptors for anions typically present a concave
space containing several interaction sites that have been
strategically positioned within the cavity to fulfil the binding
and geometrical requirements of the envisaged analyte.[1]
Receptor–anion interactions have to be reversible and can
be of various types, such as metal–ligand, electrostatic and
hydrogen bonding. A number of metal based anion recep-
tors contain a quadridentate-coordinating subunit, such as
tris(2-aminoethyl)amine (tren, 1)[2] and tris(2-pyridylmeth-
yl)amine (TPA, 5).[3] Tripodal ligands such as tren and TPA
tend to impose a trigonal-bipyramidal geometry on the
metal, leaving an apical position available for the coordina-
tion of the anion. As an example, bistren ligands, in which
two tripodal tetramine subunits are covalently linked by
spacers of varying length, have been shown to selectively in-
clude and recognise halides,[4] ambidentate polyatomic
ꢀ
anions (e.g., N3 )[5] and dicarboxylates (including l-gluta-
mate).[6,7] Moreover, positively charged fragments (e.g., am-
monium,[8] guanidinium,[9] pyridinium,[10] imidazolium)[11] can
[a] Dr. M. Allevi, Dr. M. Bonizzoni, Prof. L. Fabbrizzi
Dipartimento di Chimica Generale, Università di Pavia
via Taramelli 12, 27100 Pavia (Italy)
More recently, Anslyn et al. have designed systems in
which positively charged groups, either ammonium or guani-
dinium, were linked to a quadridentate tripodal subunit,
either tren or TPA, to give 3 and 6, respectively.[16] The cor-
responding CuII complexes are able to establish both metal–
Fax : (+39)0382-528-544
E-mail: luigi.fabbrizzi@unipv.it
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
Chem. Eur. J. 2007, 13, 3787 – 3795
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3787

ligand and electrostatic-hydrogen-bonding interactions with
one polyatomic anion (e.g., phosphate), giving rise to 1:1
complexes of various stabilities, in a neutral aqueous solu-
tion. The metal centre is important for both anion binding,
which affords a favourable enthalpy contribution, and its
role in preorganising the positively charged fragments to
generate a cavity suitable for anion inclusion (which ensures
an entropic advantage). The thermodynamic contributions
to the solution stability of the CuII complexes of 1 and 3
with phosphate have been investigated through isothermal
calorimetry studies.[17]
G
! !
anionic complex [CuII(4) X···Y] ( symbolises the coordi-
native bond and ··· the hydrogen-bonding interaction).
Results and Discussion
Anion interactions with the metal-free receptor 4: The tri-
podal receptor 4 provides three urea fragments suitable for
the interaction with anions. A single urea subunit can
donate two directional hydrogen bonds to a given anion
(either mono- or poly-atomic). The energy of the receptor–
substrate interaction is strictly related to the acidity of the
urea subunit (which can be modulated through substituents
of various electron-withdrawing properties) and to the in-
trinsic basicity of the anion. In this sense, the nitrophenyl
substituent is expected to enhance the H-bond-donor ten-
dencies of 4 and, conveniently, provide optical evidence of
the occurrence of the receptor–anion interaction. In fact,
the formation of the H-bond complex and the associated
We have considered a receptor containing two separate
compartments suitable for the interaction with a pair of
anions of either the same or different nature. In particular,
we designed the tripodal system 4, in which a nitrophenyl-
A
tren platform. It is expected that coordination of a CuII ion
by the tren subunit preorganises system 4 for the interaction
with a pair of anions: the first interaction site is available on
the axial position of the CuII
flow of negative charge from the anion to the NO2 group
ꢀ
second site to be provided by the three favourably arranged
nitrophenylurea fragments.
stabilises the excited state of the nitrophenyl chromophore,
inducing a red-shift of the intense charge-transfer band cen-
tred at ꢁ350 nm. Selected spectra recorded over the course
of the titration of a 1.99”10ꢀ5 m solution of 4 in DMSO with
[Bu4N]H2PO4 are reported in Figure 2. DMSO was chosen
as a medium because it ensures a fairly high solubility of 4
(ꢁ5”10ꢀ3 m).
We describe here the formation of the complexes of the
[CuII(4)]2+ receptor with a variety of anions in DMSO. The
investigations were carried out by conducting spectrophoto-
metric titration experiments. We observed the changes 1) in
the d–d region of the spectra (to monitor the anion interac-
tion at the metal centre), and 2) in the UV region by look-
ing at the band of the nitrophenyl chromophore (to monitor
the hydrogen bonding interactions of the anion with the
urea fragments).[18] On this basis, we tried to answer the fol-
lowing questions: 1) How do the receptorsꢁ two binding sites
compete for the coordination of the first anion? 2) To what
extent do repulsive electrostatic and steric interactions
affect the coordination of the second anion? 3) How do the
different anions affinities toward metal–ligand and hydro-
gen-bonding interactions affect the stability of the 1:2 recep-
tor–anion complexes and how could the conditions be ad-
dressed to favour the formation of complexes with two
anions of different nature? Figure 1 illustrates the cascade
mechanism for the hypothesised formation of the heterodi-
Figure 2. Spectra taken over the course of the titration of a solution of 4
(1.99”10ꢀ5 m ) in DMSO against [Bu4N]H2PO4. Inset are titration pro-
files: 4 (!), at 400 nm, left vertical axis; 7 (*), at 400 nm, right vertical
axis.
Anion addition causes a significant red-shift of the nitro-
phenyl band, indicating the occurrence of a rather strong H-
bond interaction with the urea fragments. The titration pro-
file at a selected wavelength, shown in the inset of Figure 2
(!, left vertical axis), gives evidence of the formation of a
1:1 receptor–anion complex. Non-linear least-squares fitting
of titration data,[19] over the 300–450 nm spectral range, gave
a logK=6.05Æ0.03 for the association Equilibrium (1):
Figure 1. Cascade mechanism for the formation of the heterodianionic
complex [CuII(4)(X···Y)]. Spheres from the top: CuII ion, Xꢀ anion, Yꢀ
A
anion.
3788
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 3787 – 3795

Complexes for Molecular Recognition
FULL PAPER
ꢀ
ꢀ
4 þ H2PO4 Ð ½4 Á Á Á ðH2PO4ފ
ð1Þ
liminary equilibrium studies on the interaction of the
[CuII(2)]2+ complex with anions in a DMSO solution.
[CuII(2)]2+ was considered an appropriate system for model-
ling the interaction of anions at the metal centre in the two-
ꢀ
The fact that excess addition of H2PO4 did not induce
any further modification of the spectrum indicates the pres-
ence at equilibrium of the sole 1:1 complex, over the entire
titration range, and would indicate that in the H-bond com-
plex the anion interacts with all the urea fragments of 4. In
addition, models analysing the formation of 1:2 and 1:3 re-
ceptor–anion complexes were rejected by the least-squares
fitting of the titration data.
compartment [CuII(4)]2+ receptor. Typically, a 2–5”10ꢀ3
m
solution of [CuII(2)]
(ClO4)2 was titrated with a standard so-
A
lution of Xꢀ (in general, from a tetralkylammonium salt).
Figure 3 displays the spectra obtained over the course of the
ꢀ
The occurrence of a proton transfer from the H2PO4 ion
to a secondary amine group of 4 should be also considered.
However, the distinct red-shift of the nitrophenyl charge-
transfer band indicates anion interaction at the trisurea com-
partment under definite stoichiometry. The proton transfer
ꢀ
from the H2PO4 acid to a secondary amine, according to a
Brønsted equilibrium, is an energetically favoured process
that, in the present case, is in competition with and prevent-
ed by the establishment of rather strong H-bond interactions
of the anion with the three urea fragments.
Figure 3. Visible spectra taken over the course of the titration of a solu-
tion of [CuII(2)](ClO4)2 (3.74”10ꢀ3 m) in DMSO against NaN3. Inset: ti-
G
tration profiles at selected wavelengths.
ꢀ
For comparative purposes, the interaction of H2PO4 with
ꢀ
the urea derivative 7, corresponding to a single arm of the
tripodal receptor 4, was investigated under the same condi-
tions. The pertinent titration profile at 400 nm is shown in
the inset of Figure 2 (*, right vertical axis). Notice that the
curvature of the profile is much smoother than that ob-
served for 4, and an association constant logK=4.05Æ0.03
was calculated. This finding reinforces the hypothesis that
titration with N3 . On azide addition, a new, rather strong
band develops at 410 nm, while two less intense bands at
675 and 850 nm strengthen moderately. The two latter bands
have a d–d origin and correspond to the d 2!d 2ꢀ 2,dxy and
z
x
y
d 2!dxz,dyz transitions, respectively, in a d9 complex of trigo-
z
nal-bipyramidal geometry. In contrast, the absorption band
centred at 410 nm has a ligand-to-metal charge transfer
(LMCT) nature, arising from the transfer of an electron
ꢀ
the tripodal receptor 4 interacts with H2PO4 through more
than one arm. The advantage in favour of 4 by two orders
of magnitude, with respect to 7, results from the balance be-
tween the exoergonic contribution of H-bond interactions
and the endergonic term associated to the organization of 4,
upon creating a cavity.
from a p* level of the anion to the half-filled d 2 orbital of
z
the metal centre. Plots of the absorbance at 410 and 675 nm,
shown in the inset of Figure 3, clearly indicate the formation
of a 1:1 adduct, and multiwavelength least-squares treat-
ment of titration data gave a logK=4.57Æ0.04 for the Equi-
librium (2):
Similar titration experiments were carried out on the tri-
podal receptor 4 by using a variety of inorganic anions (Clꢀ,
½CuIIð2ފ2þ þ N3 Ð ½CuIIð2ÞðN3ފ
ꢀ
þ
ꢀ
ꢀ
ꢀ
Brꢀ, Iꢀ, N3 , NCSꢀ, NO3 , NO2 ), but no significant changes
were observed in the UV-visible spectra, even after a large
excess addition of the anion, ruling out the occurrence of
any anion–urea H-bond interaction. Lack of interaction has
to be ascribed to the relatively low basicity of the envisioned
anions, in the presence of the highly competing solvent
DMSO.
ð2Þ
ꢀ
It is suggested that N3 replaces a DMSO molecule that is
axially coordinated to the metal. The formation of the terna-
ry complex can be perceived visually, as the pale-blue solu-
tion of [CuII(2)(DMSO)]2+, on azide addition, takes a bright
R
blue-green colour.
Similar behaviour was observed on titration with tetraal-
kylammonium salts of halides and pseudohalides. On addi-
tion of Xꢀ, the solution of the [CuII(2)]2+ complex took a
colour varying from blue to green, essentially depending
upon the energy of the LMCT band. In particular, such a
band was distinctly observed at relatively high wavelengths
with Iꢀ (430 nm) and NCSꢀ (375 nm). For Clꢀ and Brꢀ, due
to their lower tendency to release electrons, the LMCT
Anion interactions at the [CuII(tren)]2+ cavity: Copper(II)
N
particular with those of the tribenzyl derivative 2, with Xꢀ =
2ꢀ [22]
Clꢀ,[20] NCSꢀ,[21] S2O3
.
On this basis, we carried out pre-
Chem. Eur. J. 2007, 13, 3787 – 3795
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3789

L. Fabbrizzi et al.
band was displaced toward lower wavelengths and perceived
as a shoulder of the band associated to the amine-to-cop-
per(II) charge-transfer transition. In all cases, the formation
of a 1:1 complex was ascertained. LogK values for the gen-
Interaction of anions of the same nature at the two compart-
ments of the [CuII(4)]2+ receptor: The tendencies of the tri-
podal system 4 to interact with CuII were investigated by ti-
trating a DMSO solution of 4 (2.075”10ꢀ3 m) with a stan-
eral equilibrium [CuII(2)]2+ +XꢀÐ
A
dard solution of CuII
(CF3SO3)2 in DMSO. Figure 5 displays
G
ed in Table 1.
the family of spectra obtained over the course of the titra-
tion.
Table 1. Equilibrium constants for the interaction of inorganic anions Xꢀ
with receptors [CuII(2)]2+, 4 and [CuII(4)]2+ in DMSO at 258C. The un-
certainty for the last figure is given in parentheses.
Anion
4
[CuII(2)]2+
U
G
logK
logK
logK1
logK2
ꢀ
N3
–
–
–
–
–
–
4.57(4)
4.16(2)
3.48(3)
3.74(1)
2.18(1)
4.79(4)
>5
3.86(3)
4.7(1)
4.7(1)
3.48(7)
2.13(2)
4.91(2)
>5
2.2(1)
–
–
–
–
–
ꢁ2
NCSꢀ
Clꢀ
Brꢀ
Iꢀ
NO2
H2PO4
ꢀ
ꢀ
6.05(3)
Titration experiments were carried out also with some
ꢀ
oxo
A
d bands were observed and the high-energy LMCT band ap-
peared as a shoulder. A 1:1 titration profile was obtained,
but it was too steep for an accurate determination of the
logK value. In particular, at the investigated concentration
(10ꢀ3 m), we can state only that the logK value is >5. How-
ever, very poor spectral modifications were observed upon
titration with [Bu4]NO3, for which a logK value of <2 was
estimated. A special case was that of the oxoanion NO2 for
which a well-defined low-energy LMCT band was observed
(see spectra in Figure 4) and a relatively high association
Figure 5. Visible spectra taken over the course of the titration of a solu-
tion of 4 (2.075”10ꢀ3 m) in DMSO against CuII
(CF3SO3)2. The two bands
A
at 750 and 850 nm indicate the formation of a trigonal-bipyramidal com-
plex. Inset are titration profiles of the band at 850 nm: experimental
values (*); values corrected for the absorbance of the [CuII
species (!).
A
ꢀ
The 1:1 stoichiometry of the [CuII(4)]2+ complex should
be demonstrated by the titration profiles of the two d–d
bands. As an example, the profile of the absorbance at
850 nm is shown in the inset of Figure 5 (circles). However,
this is not a definitive profile, as the measured absorbance
includes the contribution of the [CuII(DMSO)4]2+ complex,
A
present in solution after the addition of one equivalent of
CuII. The corrected titration profile (triangles in the inset of
Figure 5) definitely demonstrates the 1:1 stoichiometry of
the metal complex. Furthermore, the steep titration profile
and the sharp discontinuity at one equivalent indicate the
formation of a very stable coordination complex that has a
stability constant that can only be estimated to be >105 (at
10ꢀ3 m). Titration experiments with anions were carried out
on a DMSO solution containing equimolar amounts of
[CuII
(CF3SO3)2] and 4. Within the employed formation
G
range (2–5”10ꢀ3 m) the percent concentration of the
[CuII(4)]2+ receptor is >99%.
Figure 6a shows the visible portion of the spectra taken
over the course of the titration of a 2.075”10ꢀ3 m solution of
[CuII(4)]2+ in DMSO with [Bu4N]N3. The strong absorption
of the nitrophenyl chromophore prevents the observation of
Figure 4. Visible spectra taken over the course of the titration of a solu-
tion of [CuII(2)](ClO4)2 (3.74”10ꢀ3 m ) in DMSO against [Bu4N]NO2.
G
Inset: titration profiles at selected wavelengths.
ꢀ
the LMCT band expected upon interaction of N3 with the
ꢀ
[CuII(tren)]2+ subunit. However, the interaction of the N3
N
constant (logK=4.79Æ0.04) was determined. These findings
anion with the [CuII(tren)]2+ moiety is unambiguously dem-
G
ꢀ
may suggest the coordination of NO2 to the metal through
onstrated by the strengthening of the two d–d bands in the
600–900 nm region. The intensity of the two d–d bands
starts to increase noticeably upon the initial additions of the
the nitrogen atom (nitro mode), rather than through one of
the oxygen atoms (nitrito mode).
3790
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 3787 – 3795

Complexes for Molecular Recognition
FULL PAPER
terms of microconstants, as il-
lustrated in Scheme 1. In par-
ticular, the thermodynamic
constants K1 and K2 are related
to the microconstants k1 and
k2, and k3 and k4, respectively,
through the equilibria reported
in Scheme 1. In the present
case, in view of the much
greater propensity of the anion
Xꢀ for the coordinative inter-
action rather than for the hy-
drogen-bonding interaction, it
happens that: 1) k1 >>k2, from
which: K1 =k1, and 2) k4 >>k3,
from which K2 =k3.
Figure 6. a) Visible spectra taken over the course of the titration of a solution of [CuII(4)]2+ (2.075”10ꢀ3 m) in
DMSO against NaN3. b) Titration profiles at selected wavelengths: the d–d band at 675 nm gives evidence of
ꢀ
the coordination of the first N3 ion at the metal centre; the absorbance at 500 nm, tail of the band of the ni-
Interestingly, the metal-free
receptor 4 does not bind the
ꢀ
trophenyl chromophore, indicates the entry of the second N3 ion and its hydrogen-bonding interaction with
the three urea fragments.
ꢀ
N3 ion, (logK<2), probably
ꢀ
anion, indicating that the N3 first seeks the metal centre.
The titration profiles of the bands at 675 nm (see Figure 6 b,
triangles) and at 850 nm (not shown) correspond to the for-
mation of a 1:1 complex, with an association constant
logK1 =3.86Æ0.03. In contrast, the absorbance at 500 nm
begins to increase only after the addition of the first equiva-
lent of N3ꢀ (see the titration profile in Figure 6 b). This spec-
tral region pertains to the tail of the very intense absorption
band of the nitrophenyl chromophore, which is in most part
out of scale at this concentration level. Nevertheless, after
the addition of one equivalent of the anion, a red-shift of
the band was induced. It has been previously mentioned
!
Scheme 1. Interaction of the [LCu S]2+ receptor with two Xꢀ anions; S
that a red-shift of the nitrophenyl absorption indicates the
occurrence of a hydrogen-bonding interaction at a covalent-
ly linked urea fragment. Correspondingly, the increase of
!
represents a solvent molecule, ( ) indicates the coordinative bond, (···)
indicates the hydrogen-bonding interaction.
ꢀ
the absorption at 500 nm documents the interaction of N3
with the urea fragments, which are hypothesised to be spa-
tially disposed to donate up to six hydrogen bonds to the
anion. Multipoint interactions of the N3 ion with a polyam-
because the intrinsically small energy associated to the
urea–azide H-bond interaction does not compensate the
conformational energy spent by the receptor on organizing
a cavity for anion inclusion. However, the conformational
energy term does not have to be spent by the metal-contain-
ing receptor [CuII(4)]2+, in which the trisurea compartment
is preorganised (at least in part), thanks to the metal coordi-
nation of the tren subunit.
ꢀ
monium cage that donates six hydrogen bonds have been
described.[23] In the present case, logK2 =2.2Æ0.1 was calcu-
lated through non-linear least-squares fitting of spectral
data. Therefore, the interaction of the metal-containing re-
ꢀ
ceptor [CuII(4)]2+ with N3 can be described by the follow-
ing two distinct stepwise Equilibria (3) and (4).
A further example of stepwise anion recognition is pro-
ꢀ
ꢀ
ꢀ
½CuIIð4Þ DMSOŠ2þ þ N3 Ð ½CuIIð4Þ N3Šþ þ DMSO
vided by H2PO4 . As observed in the titration with N3 ,
anion addition induces spectral modifications first in the d–d
ð3Þ
bands of the [CuII(tren)]2+ subunit, then in the region of the
A
½CuIIð4Þ N3Šþ þ N3 Ð ½ðCuIIð4Þ N3Þ Á Á Á N3Š
ꢀ
nitrophenyl chromophore. Figure 7 displays the titration
profiles at wavelengths pertaining to the two spectral re-
gions.
ð4Þ
ꢀ
Although Equilibria (3) and (4) describe what happens in
the investigated solution, as monitored through the spectro-
photometric response, the equilibria involving the interac-
tion of a receptor (LCu), containing two non-equivalent
binding sites, with two anions Xꢀ should be discussed in
On addition of the first equivalent of H2PO4 , a steep in-
crease of the band at 820 nm indicates phosphate coordina-
tion at the CuII centre. The sharp discontinuity at one equiv-
alent prevents the determination of the constant of the
Equilibrium of type (3), which can be estimated to be >105.
Chem. Eur. J. 2007, 13, 3787 – 3795
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3791

L. Fabbrizzi et al.
cascade interaction of the [CuII(4)]2+ receptor with two dif-
ferent anions Xꢀ and Yꢀ would require that 1) the anion Xꢀ
exhibits a pronounced penchant for metal–ligand interac-
tions and a poor affinity for hydrogen bonding, and 2) the
opposite were true for the anion Yꢀ. Although there are a
number of anions that fulfil requirement (1), for example
Clꢀ, we have not found an anion that satisfies requirement
ꢀ
(2). In fact, H2PO4 , which establishes strong interactions at
the urea subunits, also shows a high affinity for the CuII
centre. To circumvent this problem, we prepared a diluted
solution 2.00”10ꢀ5
m m in
in [CuII(4)]2+ and 3.00”10ꢀ3
[Et3Bn]Cl. From the value of K1, one can calculate that the
!
[CuII(4) Cl]+ complex is present in solution at 99% (the
Figure 7. Titration profiles at selected wavelengths taken over the course
of the titration of a solution of [CuII(4)]2+ (2.00”10ꢀ3 m) in DMSO
against [Bu4N]H2PO4: the absorbance at 650 nm gives evidence of the co-
formation of this complex cannot be monitored spectropho-
tometrically at this concentration level, in view of the in-
trinsically poor intensity of the d–d bands) In contrast, chlo-
ride addition did not significantly modify the intense band
of the nitrophenyl chromophore centred at 355 nm. The so-
ꢀ
ordination of the first H2PO4 ion at the CuII centre (logK>5); the ab-
sorbance at 426 nm indicates the establishing of H-bond interactions of
ꢀ
the urea compartment of [CuII(4)]2+ with the second H2PO4 ion (logK=
2.0Æ0.1).
!
lution containing the [CuII(4) Cl]+ complex was then titrat-
ed with [Bu4N]H2PO4, inducing a distinct red-shift of the ni-
trophenyl absorption band, indicating anion interaction at
the urea compartment.
The titration profile at 400 nm is shown in Figure 8 (!).
Multiwavelength non-linear fitting of titration data indicated
Upon addition of the second equivalent of phosphate, the
absorbance at 425 nm begins to increase, providing evidence
of the interaction of H2PO4 with the urea compartment.
ꢀ
The smooth profile corresponds to a logK2 value of 2.0Æ0.1,
for an equilibrium of type (4). Noticeably, this value is
nearly four orders of magnitude lower than that observed
for the metal free receptor 4. Such a huge difference should
reflect
a significant repulsive effect between the two
ꢀ
H2PO4 , of both steric and electrostatic nature. In this
regard, a concomitant change in the d–d spectral region
takes place, shown in Figure 6 by the decrease in the absorb-
ance at 820 nm. This may indicate a distortion of the geome-
try of the [CuII(tren)]2+ subunit, as part of a general confor-
A
mational rearrangement of the receptor for accommodating
ꢀ
the second H2PO4 anion. It is concluded that the unfavour-
able energy terms associated with electrostatic and steric re-
pulsions and to the conformational rearrangement of the re-
ceptor cancel the favourable term associated with the metal-
driven preorganisation of the urea compartment.
!
Figure 8. Titration profiles: at 400 nm for a solution of [CuII(4) Cl]+
(2.00”10ꢀ5 m) against [Bu4N]H2PO4 (!); at 400 nm for a solution of 4
(2.00”10ꢀ5 m) against [Bu4N]H2PO4 (*).
Upon titration of a 2.00”10ꢀ3 m solution of [CuII(4)]2+
with Clꢀ, Brꢀ and Iꢀ, respectively, spectral changes were ob-
served only in the d–d region, whereas no shift of the nitro-
phenyl band was observed. This excludes any interaction at
the urea compartment and reflects the intrinsically weak H-
bond acceptor tendencies of halide ions. LogK1 values that
have been determined through least-squares processing of
spectral data in the d–d region are reported in Table 1. Most
logK1 values for anions are similar to those determined for
the model system [CuII(2)]2+, whereas those for anions Clꢀ
the formation of a 1:1 adduct, with a logK2 =4.83Æ0.01. The
constant corresponds to the following Equilibrium (5):
½CuIIð4Þ ClŠþ þ H2PO4 Ð ½ðCuIIð4Þ ClÞ Á Á Á H2PO4Š
ꢀ
ð5Þ
Surprisingly, the K2 value is ꢁ700-fold higher than that as-
sociated with the formation of the corresponding homodi-
!
ꢀ
and N3 are more than one order of magnitude higher. This
anionic complex [(CuII(4) H2PO4)···H2PO4]. This pro-
may suggest the occurrence of an additional interaction of
nounced advantage is probably related to the smaller size of
the chloride anion, which is expected to exert less pro-
nounced steric and electrostatic repulsions toward the in-
ꢀ
the metal bound anion with the N H groups of the proxi-
mate urea fragments.
ꢀ
coming H2PO4 . However, this value is still 17-fold lower
ꢀ
Interaction of anions of different nature at the two compart-
than that observed for the interaction of H2PO4 at the tris-
ments of the [CuII(4)]2+ receptor: The occurrence of the
A
3792
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 3787 – 3795

Complexes for Molecular Recognition
FULL PAPER
!
ꢀ
cates that, in the [(CuII(4) Cl)···H2PO4] complex, interan-
centre the H2PO4 anion, which translocates to the trisurea
compartment, according to Equilibrium (6):
ion repulsions are still working and cancel, in part, the
energy gain due to the metal-induced organization of the re-
ceptor. The distinctly higher affinity of the metal free recep-
½CuIIð4Þ H2PO4Šþ þ Clꢀ Ð ½ðCuIIð4Þ ClÞ Á Á Á H2PO4Š
ð6Þ
ꢀ
tor 4 toward H2PO4 with respect to that containing the
CuIICl+ ion pair can be seen in Figure 8, in which the two ti-
tration profiles taken under the same conditions are com-
pared.
ꢀ
The progress of the H2PO4 sliding from one compart-
ment to another is illustrated by the titration profile in
Figure 9 (right).
The anion exchange process described by Equilibrium (6)
emphasises the lability and reversibility of the non-covalent
interactions responsible for the anion binding properties of
the [CuII(4)]2+ receptor, either metal–ligand or hydrogen
bonding, and provides a further example of supramolecular
dynamics.
The experiments described above may contribute to the
impression that in the [CuII(4)]2+ system, the first anion
binds irreversibly to the metal centre, giving rise to a rigid
and unalterable receptor that provides an H-bond-donating
cavity. Indeed, the system investigated here should not
differ too much from previously described anion receptors,
in which a metal ion plays a mere architectural role in preor-
ganising the organic receptorꢁs framework. However, the
anion interacting with the metal centre in the [CuII(4)]2+
system is not irreversibly bound. This is not unexpected in
view of the intrinsic lability of the CuII metal ion. Conse-
quently, an Xꢀ anion bound to CuII in the [CuII(4)]2+ system
should be replaced by an anion Yꢀ, under appropriate equi-
librium conditions, according to a fast-exchange process.
Indeed, this has been demonstrated by adding one equiva-
lent of [Bu4N]H2PO4 to a DMSO solution of [CuII(4)]2+
(2.00”10ꢀ3 m). The high binding constant (K1 >105) ensures
Conclusion
This work has shown that in a receptor containing two un-
equivalent recognition sites, the anion first seeks to bind at
the metal centre, then to form H-bond interactions at the
urea compartment. The latter interaction suffers from 1) the
competition from the solvent and 2) the interanion repulsive
effects of both electrostatic and steric nature. In particular,
the trisurea compartment is able to host only anions display-
ing pronounced H-bond acceptor tendencies. On this basis,
the formation of a heterodinuclear anion complex can be
designed, firstly by saturating the metal containing subunit
with a poor H-bond acceptor (e.g., Clꢀ), through a concen-
tration effect, then allowing the trisurea compartment to in-
!
the formation of 100% of the [CuII(4) H2PO4]+ complex,
in which all the dihydrogenphosphate ion is bound to the
metal centre. To this solution, aliquots of a standard concen-
trated solution of [Et3Bn]Cl in DMSO (0.620m) were added.
Upon chloride addition, a distinct shift of the nitrobenzyl
charge transfer absorption band was observed, as shown in
Figure 9 (left).
ꢀ
teract with a powerful H-bond acceptor (e.g., H2PO4 ).
Moreover, due to the labile and reversible nature of metal–
ligand and hydrogen-bonding interactions, hetroditopic sys-
tems such as those investigated here offer the opportunity
to carry out fast anion interexchange processes, such as the
ꢀ
sliding of H2PO4 from the metal centre cavity to the tris-
urea compartment, to leave room for Clꢀ (present in excess
concentration).
Anion-coordination chemistry is the younger sister of the
traditional and well-established discipline of transition-metal
coordination chemistry.[24] A number of examples of anions
inside receptors that imitate the role of metals inside multi-
dentate ligands have been reported recently. These include
chelate, macrocyclic and cage complexes, helicates,[25] rotax-
anes[26] and catenates.[27] In this work, we have explored the
design of homodinuclear and heterodinuclear anion com-
plexes. Dinuclear metal complexes display interesting prop-
Figure 9. Spectra (left) taken over the course of the titration of a solu-
tion of [CuII(4)]2+ (2.00”10ꢀ3 m) and [Bu4N]H2PO4 in DMSO against
[Et3Bn]Cl. Titration profile (right) at 435 nm.
Only the very final portion of the tail of the nitrophenyl
band could be distinctly monitored, due to the relatively
high concentration of the chromophore, which caused opti-
cal saturation. However, a definite red-shift of the band was
observed, which indicates the occurrence of an H-bond in-
teraction at the trisurea compartment. As the Clꢀ anion is
not able to interact with the trisurea cavity of the [CuII(4)]2+
receptor, even at high concentration, the spectral change
[28]
[29]
erties in magnetism,
dicinal chemistry.
catalysis,
It is difficult to forecast equally domi-
biochemistry[30] and me-
[31]
nant roles for anion analogues, which nevertheless, deserve
investigation. Here, we have tried to outline the principles
governing the formation of stable homo- and hetero-dinu-
clear anion complexes in solution.
ꢀ
has to be ascribed to the binding of the H2PO4 ion. There-
fore, due to a mass effect, Clꢀ displaces from the metal
Chem. Eur. J. 2007, 13, 3787 – 3795
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3793

L. Fabbrizzi et al.
(50 mL) was added dropwise with constant stirring. After ꢁ40 min a
yellow-white precipitate began to develop. After 2.5 h stirring was dis-
continued and the solid isolated by filtration. The mother liquor was
evaporated to half its volume and some more solid was collected. The
Experimental Section
General procedures: The spectra for the UV/Vis titration experiments
were recorded by using either a Hewlett–Packard 8452 A diode-array or
a scanning Varian Cary 100 spectrophotometer. The temperature of the
cell holder was maintained at 25.08C by a thermostat. Typically, aliquots
of a fresh tetrabutylammonium salt standard solution of the respective
anion were added and the UV/Vis spectra of the samples were recorded.
For chloride, the [Et3Bn]Cl salt was used, for azide, NaN3 was used The
HYPERQUAD program was used to fit all of the spectrophotometric ti-
tration curves.[19]
combined solids were dried in
a desiccator overnight (yield 61%).
1H NMR (400 MHz, [D6]DMSO): d=9.45 (s, 1H; urea NH), 9.05 (s, 1H;
urea NH), 8.20 (d, 2H), 7.60 (d, 2H), 7.30 (m, 4H), 5.60 (s, 1H), 4.10 (m,
2H), 3.90 ppm (m, 2H); negative-ion ESI-MS: m/z: 328.3 [MꢀH]ꢀ,
364.2–366.2 [M+Cl]ꢀ.
1-(4-Formylphenyl)-3-(4-nitrophenyl)urea (11): Complex 10 ( 2 g,
6.08 mmol) was dissolved in THF (350 mL). To this, HClO4 (19 mL,
70%) was added and the solution was left stirring overnight at ꢁ308C,
during which time the solution turned orange. Solid NaOH was added to
neutralise the excess acid: the colour changed abruptly to brownish-red
and a solid precipitated (inorganic salts). The solution was dried three
times over fresh Na2SO4, then evaporated to a reddish solid, still contain-
ing inorganic salts. Assuming nearly quantitative yield, a purity of around
50% was estimated. Further purification was not undertaken and the
product was used directly for the formation of the Schiff base mentioned
above; a trial run showed that in this step, only the desired product pre-
cipitates out of the reaction mix. ESI-MS: m/z: 286.2 [M+H]+, 308.3
[M+Na]+.
Synthesis of ligands and complexes: Reagents were bought from Sigma-
Aldrich. Reagents and solvents were used without further purification.
N1-benzyl-N2,N2-bis(2-(benzylamino)ethyl)ethane-1,2-diamine (2) was
synthesised by following a reported procedure.[32] Deionised water was
used throughout. NMR spectra were recorded by using a Bruker Avance
at 400 MHz, mass spectra were acquired by using a Thermo-Finnigan
ion-trap LCQ Advantage MAX instrument equipped with an ESI source.
Catalytic hydrogenations were carried out in a Parr stainless steel-pres-
sure vessel with mechanical stirring. The synthetic route to 4 is outlined
in Scheme 2.
Schiff base condensation between 11 and tren: Crude 11 (3.5 g, contain-
ing ꢁ6 mmol of 11) was suspended in MeOH (1 L) and thoroughly soni-
cated accompanied by heating, then the insoluble matter was filtered off.
To this solution, tris(2-aminoethyl)amine (tren) (300 mL, 2 mmol) was
added and the mixture was left stirring overnight. During this time a
solid separated, which was isolated and stored in a desiccator (yield 55%
based on tren). 1H NMR (400 MHz, [D6]DMSO): d=9.55 (s, 1H; urea
NH), 9.20 (s, 1H; urea NH), 8.20 (d, 2H), 8.10 (s, 1H; imine CH), 7.65
(d, 2H), 7.48 (m, 4H), 3.60 (t, 2H), 2.75 ppm (t, 2H); ESI-MS: m/z:
948.2 [M+H]+.
Reduction of the Schiff base to give 4: The Schiff base (969 mg,
1.02 mmol) was suspended in MeOH (180 mL). The suspension was stir-
red at 608C while NaBH4 (1.16 g, 31 mmol) was added in small portions.
The mixture was refluxed for another 2.5 h; then left stirring overnight at
RT. The solvent was removed completely to give an orange solid that
was taken up in dilute aqueous NaOH (200 mL) to give a suspension
that was washed with CH2Cl2 (2”40 mL). The organic phases were dis-
carded. The bright-yellow solid in the water phase was filtered off and
dried to constant weight in a vacuum desiccator (yield 73%). 1H NMR
(400 MHz, [D6]DMSO): d=9.60 (s, 1H; urea NH), 9.20 (s, 1H; urea
NH), 8.15 (d, 2H), 7.70 (d, 2H), 7.30 (d, 2H), 7.15 (d, 2H), 3.60 (brm,
4H; benzylic and aliphatic CH2), 2.75 ppm (t, 2H); ESI-MS: m/z: 954.3
[M+H]+.
Scheme 2. Synthetic route to the receptor 4. i) H2, 7 atm, PtO2 5 mol%,
CH
A
in THF, 2 h, RT; iv) MeOH, RT; v) NaBH4, MeOH, 608C.
2-(4-Nitrophenyl)-[1,3]-dioxolane
(8):
4-Nitro-benzaldehyde
(4 g,
26.5 mmol) was dissolved in benzene (160 mL) with magnetic stirring
(toluene was also tried, but yielded inferior results). To this solution, eth-
ylene glycol (7.3 mL, 132 mmol) and p-toluenesulfonic acid (60 mg,
0.59 mmol, catalytic) were added and the solution refluxed for 4.5 h;
after which TLC (SiO2, AcOEt:hexanes=3:7) showed the reaction to be
complete. Once cooled to RT, the solution was washed with water (3”
20 mL), then brine (3”20 mL). The combined aqueous phases from these
extractions were washed with CH2Cl2 (3”20 mL). The combined organic
phases from all the extraction steps were then dried with Na2SO4
(45 min) and the solvent was removed in vacuo to give the desired pro-
tected aldehyde in 90% yield. 1H NMR (400 MHz, CDCl3): d=8.25 (d,
2H), 7.65 (d, 2H), 5.95 (s, 1H), 4.15 ppm (m, 4H).
Acknowledgements
The financial support of the Italian Ministry of University and Research
(PRIN-Dispositivi Supramolecolari; FIRB-Project RBNE019H9K) is
gratefully acknowledged.
2-(4-Aminophenyl)-[1,3]-dioxolane (9): The pressure reactor was purged
with nitrogen, then charged with 8 (9.75 g, 50 mmol), CH(OEt)3 (14.8 g,
A
100 mmol) and dry THF (125 mL), to obtain a solution in which PtO2
(1 g, 4 mmols, catalytic) was suspended. The reactor was then sealed,
placed in a water/ice bath, evacuated and charged with H2 to 7 atm
(100 psi). Mechanical stirring was maintained throughout the reaction.
Hydrogen consumption was compensated by repressurisation every
15 min. After ꢁ1.5 h hydrogen absorption ceased and the pressure re-
mained constant. The reactor was then opened and the catalyst removed
by filtration. After drying over Na2SO4, the solution was concentrated in
vacuo (water bath ꢁ658C) to obtain a yellow viscous liquid that solidifies
on cooling (yield 63%). The product was pure enough for the subsequent
reaction steps. 1H NMR (400 MHz, CDCl3): d=7.30 (d, 2H), 6.70 (d,
2H), 5.70 (s, 1H), 4.20 (m, 2H), 4.00 (m, 2H), 3.60 ppm (brs, 2H; NH2).
[1] J.-M. Lehn, Supramolecular Chemistry, Concepts and Perspectives,
Wiley-VCH, Weinheim, 1995; Supramolecular Chemistry ofAnions
(Eds.: A. Bianchi, K. Bowman-James, E. García-EspaÇa), Wiley-
VCH, New York, 1997; F. P. Schmidtchen, M. Berger, Chem. Rev.
1997, 97, 1609–1646; P. D. Beer, Acc. Chem. Res. 1998, 31, 71–80;
P. D. Beer, P. A. Gale, Angew. Chem. 2001, 113, 502–532; Angew.
Chem. Int. Ed. 2001, 40, 486–516; Coord. Chem. Rev. 2003, 240, 1–
226 (whole volume); R. Martínez-Mµnez, F. Sancenón, Chem. Rev.
2003, 103, 4419–4476.
[2] G. De Santis, L. Fabbrizzi, M. Licchelli, A. Poggi, A. Taglietti,
Angew. Chem. 1996, 108, 224–227; Angew. Chem. Int. Ed. Engl.
1996, 35, 202; M. Di Casa, L. Fabbrizzi, M. Licchelli, A. Poggi, A.
Russo, A Taglietti, Chem. Commun. 2001, 825–826.
1-(4-(1,3-Dioxolan-2-yl)phenyl)-3-(4-nitrophenyl)urea (10): Compound 9
(2 g, 8.21 mmol) was dissolved in THF (50 mL). To this solution, 4-nitro-
phenyisocyanate (1.36 g, 8.29 mmol, 1.01 equiv) dissolved in THF
3794
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 3787 – 3795

Complexes for Molecular Recognition
FULL PAPER
[3] J. W. Canary, J. Xu, J. M. Castagnetto, D. Rentzeperis, L. A. Marky,
J. Am. Chem. Soc. 1995, 117, 11545–11547; Z. Dai, X. Xu, J. W.
Canary, Chem. Commun. 2002, 13, 1414–14145. S. Zahn, J. W.
Canary, J. Am. Chem. Soc. 2002, 124, 9204–9211.
[4] V. Amendola, E. Bastianello, L. Fabbrizzi, C. Mangano, P. Pallavici-
ni, A. Perotti, A. Manotti-Lanfredi, F. Ugozzoli, Angew. Chem.
2000, 112, 3039–3042; Angew. Chem. Int. Ed. 2000, 39, 2917–2920.
[5] L. Fabbrizzi, A. Leone, A. Taglietti, Angew. Chem. 2001, 113, 3156–
3159; Angew. Chem. Int. Ed. 2001, 40, 3066–3069.
[17] S. L. Tobey, E. V. Anslyn, J. Am. Chem. Soc. 2003, 125, 14807–
14815.
[18] M. Boiocchi, L. Del Boca, D. Esteban-Gómez, L. Fabbrizzi, M. Lic-
chelli, E. Monzani, J. Am. Chem. Soc. 2004, 126, 16507–16514.
[19] P. Gans, A. Sabatini, A. Vacca, Talanta 1996, 43, 1739–1753.
[20] a) M. Schatz, M. Becker, O. Walter, G. Liehr, S. Schindler, Inorg.
Chim. Acta 2001, 324, 173–179; b) K. Komiyama, H. Furutachi, S.
Nagatomo, A. Hashimoto, H. Hayashi, S. Fujinami, M. Suzuki, T.
Kitagawa, Bull. Chem. Soc. Jpn. 2004, 77, 59–72.
[6] M. Boiocchi, M. Bonizzoni, L. Fabbrizzi, G. Piovani, A. Taglietti,
Angew. Chem. 2004, 116, 3935–3940; Angew. Chem. Int. Ed. 2004,
43, 3847–3852.
[7] M. Bonizzoni, L. Fabbrizzi, G. Piovani, A. Taglietti, Tetrahedron
2004, 60, 11159–11162.
[8] J. M. Llinares, K. Bowman-James in Encyclopedia ofSupramolecular
Chemistry, (Eds.: J. L. Atwood, J. W. Steed), Marcel Dekker, New
York, 2004, pp. 1170–1175.
[9] E. V. Anslyn, S. L. Tobey in Encyclopedia ofSupramolecular
Chemistry, (Eds.: J. L. Atwood, J. W. Steed), Marcel Dekker, New
York, 2004, pp. 615–627.
[21] Y.-S. Xie, H. Jiang, X-T. Liu, Z.-Y. Zhou, Q.-L. Liu, X.-L. Xu, Col-
lect. Czech. Chem. Commun. 2002, 67, 1647–1657.
[22] A. J. Fischmann, A. C. Warden, J. Black, L. Spiccia, Inorg. Chem.
2004, 43, 6568–6578.
[23] B. Dietrich, J. Guilhem, J.-M. Lehn, C. Pascard, E. Sonveaux, Helv.
Chim. Acta 1984, 67, 91–104.
[24] K. Bowman-James, Acc. Chem. Res. 2005, 38, 671–678.
[25] J. Keegan, P. E. Kruger, M. Nieuwenhuyzen, J. OꢁBrien, N. Martin,
Chem. Commun. 2001, 2192–2193.
[26] M. R. Sambrook, P. D. Beer, J. A. Wisner, R. L. Paul, A. R. Cowley,
F. Szemes, M. G. B. Drew, J. Am. Chem. Soc. 2005, 127, 2292–2302.
[27] K-Y. Ng, A. R. Cowley, P. D. Beer, Chem. Commun. 2006, 3676–
3678.
[10] V. Amendola, M. Boiocchi, L. Fabbrizzi, A. Palchetti, Chem. Eur. J.
2005, 11, 5648–5660.
[11] J. Yoon, S. K. Kim, N. J. Singh, K. S. Kim, Chem. Soc. Rev. 2006, 35,
355–360.
[28] J. A. McCleverty, M. D. Ward, Acc. Chem. Res. 1998, 31, 842–851.
[29] V. S. I. Sprakel, M. C. Feiters, W. Meyer-Klaucke, M. Klopstra, J.
Brinksma, B. L. Feringa, K. D. Karlin, R. J. M. Nolte, Dalton Trans.
2005, 3522–3534.
[30] L. J. Murray, R. Garcia-Serres, S. Naik, B. H. Huynh, S. J. Lippard, J.
Am. Chem. Soc. 2006, 128, 7458–7459.
[31] H. T. Chifotides, K. R. Dunbar, Acc. Chem. Res. 2005, 38, 146–156.
[32] A. A. Naiini, W. M. P. B. Menge, J. G. Verkade, Inorg. Chem. 1991,
30, 5009–5011.
[12] J. W. Steed, Chem. Commun. 2006, 2637–2649.
[13] P. A. Gale in Encyclopedia ofSupramolecular Chemistry , (Eds.: J. L.
Atwood, J. W. Steed), Marcel Dekker, New York, 2004, pp. 31–41.
[14] V. Amendola, D. Esteban-Gomez, L. Fabbrizzi, M. Licchelli, Acc.
Chem. Res. 2006, 39, 343–353.
[15] P. A. Gale, J. L. Sessler, S. Camiolo in Encyclopedia ofSupramolec-
ular Chemistry, (Eds.: J. L. Atwood, J. W. Steed), Marcel Dekker,
New York, 2004, pp. 1176–1185.
[16] S. L. Tobey, B. D. Jones, E. V. Anslyn, J. Am. Chem. Soc. 2003, 125,
4026–4027.
Received: October 13, 2006
Published online: January 24, 2007
Chem. Eur. J. 2007, 13, 3787 – 3795
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3795