Thermodynamic characterization of the self-assembly process of a three component heterobimetallic bisporphyrin macrocycle

Article abstract of DOI:10.1021/jp9053742

The molecular self-assembly of macrocycle 4 is induced by the simultaneous coordination of two molecules of 4-pyridyldiphenylphosphine 3, a highly selective ditopic ligand, to Zn-bisporphyrin 1 and a square-planar Pd(II) complex 2·COD. We report a detaile

Full text of DOI:10.1021/jp9053742

J. Phys. Chem. B 2009, 113, 11479–11489
11479
Thermodynamic Characterization of the Self-Assembly Process of a Three Component
Heterobimetallic Bisporphyrin Macrocycle
‡
‡
,†,‡
´
Almudena Gonza´lez-Alvarez, Antoni Frontera, and Pablo Ballester*
Catalan Institution for Research and AdVanced Studies (ICREA), Pg. Llu´ıs Companys 23, 08010-Barcelona,
and Institute of Chemical Research of Catalonia (ICIQ), AV. Pa¨ısos Catalans 16, 43007-Tarragona, Spain
ReceiVed: June 8, 2009; ReVised Manuscript ReceiVed: June 30, 2009
The molecular self-assembly of macrocycle 4 is induced by the simultaneous coordination of two molecules
of 4-pyridyldiphenylphosphine 3, a highly selective ditopic ligand, to Zn-bisporphyrin 1 and a square-planar
Pd(II) complex 2 · COD. We report a detailed thermodynamic characterization of the assembly process based
on the quantification of each one of the two metal(Zn,Pd)-ligand(N,P) pairwise binding interactions implicated
in the supramolecular macrocycle and its effective molarity value (EM). The experimental values of the
pairwise metal-ligand interactions have been derived from UV-vis, NMR titrations, and isothermal titration
calorimetry experiments of reference model systems. In turn, an EM ) 1 × 10-2 M has been determined by
relating the experimental overall stability constant determined for the cyclic assembly with the equation to
evaluate the statistical (noncooperative) self-assembly equilibrium constant. We used numerical methods
(SPECFIT program) to predict the solution behavior (speciation) of two mixtures of the three molecules 1,
3, and 2 · COD in a 1:2:1 relative stoichiometry at two different overall concentrations. The method uses the
overall stability constant values and the stoichiometries of eleven species (complexes) implicated in the
multicomponent equilibrium self-assembly of 4. Estimated stability constants of some of the species were
statistically determined. The agreement observed between the theoretical simulations and the experimental
data validates the suitability of the theoretical treatment of self-assembly macrocyclization in a three component
strategy.
Introduction
Hunter7 and Ercolani.8 In particular, Ercolani has recently
presented an interesting and exhaustive overview of the
Simple monoporphyrins have found applications in catalysis,1
chemical sensing,2 and molecular optoelectronic devices.3
Consequently, the incorporation of porphyrin building blocks
into well-defined, self-assembled architectures represents a
logical strategy for the construction of elaborate and function-
alized multiporphyrin supramolecular assemblies4 and materials.5
In recent years, the structures of many discrete and kinetically
labile metal-mediated assemblies that contain porphyrin units
have been fully elucidated both in solution and in the solid state.6
Detailed understanding of the solution behavior of molecular
self-assembly systems becomes especially important when trying
to rationalize the results of their applications in different
functions like molecular recognition, catalysis, optoelectronic
molecular devices, and so forth. In general, accurate information
on the extent of the self-assembly process and the composition
of the mixture is required. In spite of these requirements, the
detailed thermodynamic characterization of the multiporphyrin
self-assembly process has only been undertaken in a limited
number of examples.7 This is most likely due to the inherent
complexity of these systems. In order to describe the solution
behavior of a mixture of molecules capable of reversible
association, it is necessary to know both their initial concentra-
tions as well as the stoichiometry and overall stability constants
of all the species (complexes) in which they are involved.
Theoretical treatments of self-assembled macrocyclizations
occurring under thermodynamic control were put forward by
principles underlying the thermodynamics of self-assembly in
solution using selected examples of metal-mediated porphyrin
assemblies.9 Both treatments are based in two fundamental
quantities, the effective molarity, EM,10 and the stability constant
for the intermolecular interaction, Kinter. Hence, the knowledge
of these two quantities, Kinter and EM, allows the modeling of
the solution behavior of a mixture of components operating
under thermodynamic equilibrium. Furthermore, the EM value
can also be used in assessing the presence of cooperativity in
self-assembly systems.11 It is worth noting that due to the
dependence of the molecular self-assembly process on concen-
tration (entropy), the quantitative formation of a cyclic aggregate
will only take place at a specific concentration and appropriate
monomer molar ratios if the condition EMKinter g 185n is
satisfied.12 Often times, the system will not produce a single or
major aggregate but an equilibrating mixture of varied cyclic
aggregates (i.e., trimer and tetramer) and open aggregates.
One of the less-explored strategies for the self-assembly of
multiporphyrin cyclic assemblies involves the combination of
three different but complementary ditopic molecular compo-
nents, a metalated bisporphyrin unit (PorM1sM1Por), a ditopic
ligand with two Lewis basic sites (LsL), and a metal center
(M2) containing two free coordination sites. This strategy grants
easy access to the preparation of tetrameric cyclic architectures
containing two different types of metallic centers as structural
units (Scheme 1). Different structures can be generated when
one considers that the ditopic ligand can have two basic sites
that can be identical (LsL) or not (LsL′).
* To whom correspondence should be addressed. E-mail: pballester@
iciq.es. Tel: +34 977 920 206. Fax: +34 977 920 221.
† ICREA.
‡ ICIQ.
10.1021/jp9053742 CCC: $40.75  2009 American Chemical Society
Published on Web 07/27/2009

´
11480 J. Phys. Chem. B, Vol. 113, No. 33, 2009
Gonza´lez-Alvarez et al.
SCHEME 1: Schematic Representations of
Three-Component Self-Assemblies Producing
Heterobimetallic Multiporphyrin Cyclic Structures
were prepared using the solution of the titrand as solvent. After
each addition, a new NMR spectrum was acquired. Deutero-
chloroform was previously deacidified by passing through a
short column of Aluminum Oxide 60 active basic.
UV-vis Spectroscopy and Titrations. UV-vis spectra were
measured on a UV-vis spectrophotometer Shimadzu UV-
2401PC. The UV-vis titrations were carried out through a
similar methodology as described above. In this case, the titrand
concentration was around 1 µM and to cover a wide range of
molar ratios it was necessary to prepare several titrant solutions
with different concentrations. The data obtained from the
UV-vis spectrophotometric titrations were analyzed by fitting
the whole series of spectra at 1 nm interval using the software
SPECFIT 3.0,18 which uses a global system with expanded factor
analysis and Marquardt least-squares minimization to obtain
globally optimized parameters.
Isothermal Titration Calorimetry (ITC). Stability constants
and enthalpies of complexation, ∆H, of the ligands 3 or 5 with
the metal Zn(II) or Pd(II) were determined using a microcalo-
rimeter Microcal VPITC. ITC titrations were performed by
adding microliter injections of the titrant solution to a titrand
solution 7-10-fold more diluted placed in the calorimetric cell.
After each injection, the heat change was monitored. The
titration data were fitted using the binding model implement in
the Microcal ITC Data Analysis software or combining the
speciation simulation module of the SPECFIT software with
the equations for the calculation of the heat content (Q(i)) and
for the heat release (∆Q(i)) after any injection i as described in
the text.
In fact, using the three component strategy, one of us reported
convincing evidence for the quantitative self-assembly of a
heterobimetallic macrocycle at millimolar concentrations.13
Almost simultaneously, van Leeuwen et al. described the use
of the same methodology for the preparation of “supramolecular
bidentate ligands”.14 It is likely that the experimental difficulties
associated with the accurate determination of the overall stability
constant for a three component assembly (AmBnCl) has restricted
the discussion of the theoretical treatment of self-assembled
macrocyclization to examples involving aggregates formed by
just one or two different components with general formulas of
Am or AmBn. We report here the first application of the
theoretical treatment of self-assembled macrocyclization to the
experimental thermodynamic characterization of the self-as-
sembly process of the three-component heterobimetallic mac-
rocycle 4. We experimentally calculated the overall stability
constant of the cyclic tetrameric assembly 4 and derived its
effective molarity value (EM). The microscopic stability
constants of the two different intermolecular interactions that
are involved in the cyclic structure (Kinter) are extracted from
the study of simple reference models. Using the calculated
values for Kinter, we statistically determined the overall stability
constants of the plausible self-assembled complexes that may
exist at equilibrium. Next, we applied numerical methods to
investigate the effect of total concentration in the self-assembled
equilibrium speciation of the three monomers. We showed that
under strict stoichiometric control there is reasonably good
agreement between the simulated speciation and the experi-
mental data. This result serves to validate the application of
the theoretical treatment of self-assembled macrocyclization to
the three component strategy.
Results and Discussion
General Considerations. The promiscuity in the complex-
ation properties of a symmetric LsL type ligand, that is, 4,4′-
bipyridine, toward two metals ions, that is, M1 ) Zn(II) and
M2 ) Pt(II), implicated in the structure of the self-assembled
macrocycle (Scheme 1) hampered the accurate thermodynamic
characterization of the three component assembly.13 On the
contrary, the use of a dissymmetric ditopic LsL′ ligand, in
which the two individual basic centers are selective in coordi-
nating to one specific metal center of the assembly, simplifies
the analysis of the titration data avoiding the production of
competitive homometallic cyclic or acyclic architectures. van
Leeuwen et al. have reported that the pyridylphosphine ligand
3 selectively binds Zn-porphyrins through its pyridyl nitrogen,
leaving the phosphorus atom available for transition metal
coordination.19 For this reason, we decided to use pyridylphos-
phine ligand 3 in tackling the thermodynamic characterization
of the multicomponent equilibrium self-assembly of Zn-bispor-
phyrin 1 (previously used in our laboratory) and the Pd(II)
complex 2·COD (source of the external metal center) yielding
the supramolecular macrocyle 4.
Although the Pd(II) complex 2·COD includes a cis coordi-
nating cyclooctadiene (COD) ligand, it should not be considered
a cis-protected metal center. The COD can easily be replaced
by two phosphines yielding exclusively trans square-planar
complexes.20 However, it is also documented that square planar
complexes resulting from the substitution of COD with phos-
phine ligands can potentially exist as cis and trans isomers. In
fact, for chelating diphosphines it is well known that the bite
angle has a large influence on the cis-trans isomerization
equilibrium.21 In short, the use of 2·COD introduces a level of
flexibility in the coordination geometry adopted by the external
metal center in macrocyle 4 (cis or trans, see Figure 1).
Experimental Section
Synthesis. Bisporphyrin 1-H4 was obtained by reaction of
the monoamino porphyrin 5-(4-aminophenyl)-10,25,20-tris(4-
penthylphenyl)-21H,23H-porphine15 with 2,6-pyridinedicarboxy-
lic acid chloride in dichloromethane under anhydrous conditions.
Bisporhyrin 1-H4 was purified using column chromatography
and isolated in 60% yield. Metalation was achieved by reaction
of the free base form of porphyrin 1-H4 with Zn(II) acetate in
a mixture of chloroform and methanol. The Zn-metalated
porphyrin 1 was isolated in almost quantitative yield after
purification with an alumina column. The ditopic ligand
4-pyridyldiphenylphosphine 316 and the cycloocta-1,5-diene
methyl palladium(II) chloride 2·COD17 complex were synthe-
sized following procedures described in the literature.
NMR Measurements and Titrations. 1H NMR spectra were
recorded on Bruker Advance 400 and Bruker Advance 500 Ultra
shield NMR spectrometers. 1H NMR titrations were carried out
by running a spectrum of the titrand solution in a deuterated
solvent at a mM concentration and adding to it incremental
aliquots of a titrant solution that was 10 times more concentrated.
To avoid the dilution of the titrand solution, the titrant solutions

Three Component Heterobimetallic Bisporphyrin Macrocycle
J. Phys. Chem. B, Vol. 113, No. 33, 2009 11481
Figure 1. Chemical structures of the molecular components used in the self-assembly of the heterobimetallic macrocycle 4 studied in this work.
CAChe minimized structures of two possible cyclic assemblies with cis and trans coordination geometry in the external metal center (Pd). Nonpolar
hydrogen atoms are removed for clarity. The Zn and Pd metal centers are shown as CPK spheres.
SCHEME 2: Schematic Representation of the Species Involved in the Binding Equilibria of 3 with Zn-Bisporphyrin 1a
a The stepwise binding constants K11 and K11T12 are indicated as well as their relationship with Km (the microscopic binding constant), R
(cooperativity factor), and statistical correction factors.
Study of Reference Models. To assist in the study of the
three component self-assembly process of Zn-bisporphyrin 1
with the Pd(II) metal center 2·COD and 4-pyridyldiphenylphos-
phine ligands 3, we investigated in detail the following reference
models: (a) the binding interaction of Zn-bisporphyrin 1 with
4-pyridyldiphenylphosphine ligand 3, and (b) the coordination
properties of 4-pyridyldiphenylphosphine 3 and triphenylphos-
phine 5 to cycloocta-1,5-diene methyl Pd(II) chloride 2·COD.
The assembly process of the macrocyclic bimetallic architecture
4 based on the three molecular subunits (Zn-bisporphyrin 1,
4-pyridyldiphenylphosphine 3 and Pd(II) metal center 2·COD)
will be explained in detail in a forthcoming section.
Binding of 1 with 3. The interaction between 4-pyridyldiphe-
nylphosphine 3 and a simple Zn-monoporphyrin, 5,10,15,20-
tetraphenyl-21H,23H-porphine (TPP), leads to the exclusive
formation of a 1:1 complex. The 3·TPP complex has a stability
constant of K11 ) 6.1 × 103 M-1 in toluene.22 When we probed
the interaction of 3 and TPP using 1H NMR spectroscopy, we
observed that the addition of 0.2 equiv of 3 to a toluene solution
of TPP resulted in large upfield shifts of the pyridyl protons
(∆δR ) -5.19 ppm, ∆δꢀ ) -1.57 ppm) due to the shielding
effect of the porphyrin ring. The protons in the phenyl rings of
3 were observed as three different signals (δHo) 6.35 ppm, δHm
) 6.65 ppm, and δHp) 6.78 ppm) experiencing a moderate
upfield shift with respect to the corresponding proton signals
in free ligand (multiplet at δH) 7.29 ppm). The addition of
incremental amounts of 3 caused downfield shifts of all the
aforementioned proton signals. In agreement with the findings
reported by van Leuween et al.,22 these observations support
that ligand 3 interacts exclusively through axial coordination
of the nitrogen atom to the zinc porphyrin and that free and
bound 3 are involved in a chemical exchange that is fast on the
1H NMR time scale.
Typically, Zn-porphyrins display a red shift in both their
Soret and Q bands upon axial coordination to amine
ligands.23,24 The association constants for these complexes
are usually in the range of 103-105 M-1 and UV-vis
spectroscopy titrations are appropriate to calculate such values
with accuracy. In the case of the interaction between
bisporphyrin 1 and pyridylphosphine 3, it should be expected
to obtain species with higher stoichiometry than the simple
1:1 complex discussed above for the 3 · TPP complex. Thus,
when the concentration of 3 is low with respect to that of
the bisporphyrin, 1:1 complexes will be expected to form
preferentially. Further addition of the amine will lead to the
formation of complexes with 1:2 stoichiometry (bisporphyrin:
pyridylphosphine) as the major species (Scheme 2).
Our interest lies in determining the stepwise binding constants
for coordination of 3 with 1 (K11 and K11T12) and in evaluating
the extent of cooperativity in the binding process. Accordingly,
we performed UV-vis titrations of bisporphyrin 1 ([1] ) 4 ×
10-5 M) with pyridylphosphine 3 in toluene. We followed the
spectroscopic changes experienced in the Q bands of 1 upon
complexation with 3. The obtained titration data were fit, using
the SPECFIT18 software, to a binding model consisting of three
colored species, free Zn-bisporphyrin 1 and two complexes that
involve 1 in different binding stoichiometry with respect to
pyridylphosphine 3 (1:1 and 1:2). We determined the following

´
11482 J. Phys. Chem. B, Vol. 113, No. 33, 2009
Gonza´lez-Alvarez et al.
values for the stepwise binding constants: K11 ) 6.3 × 103 M-1
and K11T12 ) 1.6 × 103 M-1. Considering the relationship K11
) 2Km, we calculated the microscopic binding constant for the
Zn-pyridine interaction as Km(N-Zn) ) 3.15 × 103 M-1. This
value is in complete agreement with the one determined by van
Leeuwen et al. for the same interaction using TPP.22 We used
the relationship K11T12 ) KmR/2 to determine the cooperativity
factor. We obtained a value of R ) 1.01 indicating that the
porphyrin binding sites are independent.
We also probed the complexation process of bisporphyrin 1
with pyridylphosphine 3 by means of isothermal titration
calorimetry (ITC) experiments.25 ITC is a thermodynamic
technique that directly measures the heat released or absorbed
due to the interaction between any two molecules. The quantity
of heat absorbed or released is directly proportional to the
amount of binding. As the system reaches saturation, the heat
signal diminishes until only heat of dilution of the titrant is
observed. A binding isotherm can be obtained from a plot of
the normalized integration heats from each addition of titrant
against the ratio of titrant (ligand added) to titrand (binding
partner in the cell). The visual analysis of the obtained binding
curve usually gives a clear indication on the number of stepwise
binding events (number of sets of binding sites in the binding
partner) that are operative in the binding process and if all of
them are identical or not. Thus, a binding isotherm with a shape
of a single sigmoidal curve is indicative of the existence of “n”
binding events between the interacting substances that should
be almost identical in the case that n > 1. The value of the molar
ratio of the ligand to the binding partner at the inflection point
of the sigmoidal binding curve determines the number “n” of
stepwise binding events of the system. The “n” value is also
used to assign the stoichiometry of the complex or complexes
that are formed during the binding process. The determination
of n values <1 arises when a multitopic titrant (ligand) is used
producing complexes having stoichiometry values defined by
the ratio of titrant/titrand <1. Moreover, in the case of simple
binding models, it is also possible to determine values for the
equilibrium constant(s) of the binding events, Ka, by fitting the
obtained titration data to the corresponding theoretical binding
isotherms implemented in the data analysis software of the
instrument. Figure 2a depicts the heat (exothermic) evolved per
injection in terms of µcal/second plotted against time in minutes
for the calorimetric titration of 1 (1.07 × 10-3 M in toluene)
with 3 (1.88 × 10-2 M toluene solution). The normalized
integrated data (binding isotherm) is shown in Figure 2b in terms
of kcal/mol of injectant plotted against molar ratio. The
experimental binding isotherm is a sigmoidal curve with the
inflection point centered at a value of molar ratio 3/1 equal to
2. These observations indicate that the interaction of 3 with 1
takes place through two almost identical stepwise binding events.
The first interaction produces a 1:1 complex, while in a second
binding event, promoted by the increase in concentration of 3
in solution, yields a 1:2 complex. This finding is in complete
agreement with the binding model used to fit the UV-vis
titration data discussed above for the same systems. The
observation of a symmetrical sigmoidal ITC binding isotherm,
together with the value calculated for the cooperativity factor
in the UV-vis titration, prompted us to consider that both
binding sites are identical. Consequently, and for the sake of
simplicity, the ITC data were fitted to a simple 1:1 theoretical
binding isotherm. In doing so, the fitting procedure will directly
give the enthalpy change per mole of Zn-N bond and its
microscopic stability constants, that is, the average enthalpy
change and average microcoscopic binding constant for the two
Figure 2. ITC experiment of 1 (1.07 × 10-3 M) with 3 (1.88 × 10-2
M) in toluene. (a) Heat vs time plot; (b) normalized integration heat
vs 3/1 molar ratio; (c) 1:1 theoretical binding isotherm (red line) fit to
the experimental data assuming that the concentration of the binding
partner in the cell is 2-fold the concentration of 1. The values returned
in the fitting procedure are summarized on top of panel c.
Zn-N bonds formed in the process. For the fitting procedure,
we used the one set of sites model implemented in the Microcal
ITC Data Analysis software and we considered that the
concentration of the binding partner in the cell is 2-fold the
concentration of 1. We obtained a good fit of the ITC titration
data to the theoretical binding isotherm, which is now centered
at a molar ratio equal to 1 (Figure 2c). The fitting procedure
returned an average microcoscopic stability constant value for
the Zn-N interaction of Km(N-Zn) ) 2.80 × 103 M-1, which
agrees with the value obtained in the UV-vis titration, KmZn-N
) 3.15 × 103 M-1. We also calculated an average value of ∆Η
) -10.6 kcal/mol for the enthalpy of the Zn-N bonds formed.
This value also correlates nicely with the enthalpy calculated
for the Zn-N interaction, ∆Η ) -10.0 kcal/mol, in the
interaction of TPP with 3 using ITC measurements (see
Supporting Information).
Binding of 2 ·COD with 3 and 5. Complex 2·32 can be
isolated as a white solid from the reaction of 2·COD with 2
equiv of 3 in toluene and complexation studies of 2·32 with
Zn-TPP have been recently reported.22 When the 31P NMR
spectrum of the 2·32 complex was recorded, we only observed
a singlet resonating at δ ) 32.5 ppm, indicating that the
phosphines are in the trans orientation. The isomerization of
the initial cis 2·COD complex to the final trans 2·32 complex
has been previously described in related palladium complexes
through the intermediacy of chloro-bridged methyl Pd(II)
dimers.26 However, to the best of our knowledge, the coordina-
tion process leading to the 2·32 complex has not been studied
in detail and neither characterized thermodynamically.
1
We performed a H NMR titration of the Pd(II) complex
2·COD with the pyridylphosphine ligand 3. From the analysis
1
of the H NMR spectra (Figure 3) recorded along the titration

Three Component Heterobimetallic Bisporphyrin Macrocycle
J. Phys. Chem. B, Vol. 113, No. 33, 2009 11483
SCHEME 3: Proposed Binding Model for the Equilibria
Involved in the Coordination of 2 ·COD with 3a
1
Figure 3. Selected H NMR spectra acquired during the titration of
2·COD (0.01 M in toluene-d8) with 3. Equivalents of 3 added are (a)
0; (b) 0.5; (c) 1; (d) 2. “*” designates H2O.
a The CAChe minimized molecular structure for the putative 23 ·32
experiment we concluded that it is possible to distinguish three
different phases. In the first, the addition of less than 0.6 equiv
of 3 results in the appearance of a new set of proton signals.
Two new singlets resonating at δ ) 5.76 ppm and δ ) 2.25
ppm can easily be assigned to free COD, which is displaced by
the coordination of 3. We also observed two new singlets
appearing in the upfield region at δ ) 1.05 ppm and δ ) 1.23
ppm. These signals are assigned to the methyl protons of 2
initially coordinated to 6. These two singlets grow at the expense
of the methyl proton signal of 2·COD that resonates at δ )
1.28 ppm. Concurrently, we detected two new multiplets in the
downfield region at δ ) 9.31 and 7.50 ppm that were assigned
to the pyridyl protons of coordinated 3. The downfield shift
displayed by the pyridyl signals is indicative of the existence
of pyridyl N-Pd(II) coordination bonds in the structure of the
complex formed in this first phase of the titration. The other
proton signals of 3 are obscured by the strong solvent signal.
In the second phase of the titration, the addition of up to 1 equiv
of 3 produces the emergence of another new set of broad proton
signals at δ ) 7.59 and 8.40 ppm that we assigned to protons
of 3 in a different complex stoichiometry. Concomitantly, the
two signals assigned to the methyl protons of the coordinated
2 observed during the first phase diminish as a new broad signal
centered at δ ) 0.99 ppm grows. Finally, when more than 1
equiv of 3 is added, a second set of new signals for the protons
of 3 becomes distinguishable as three broad singlets resonating
at δ ) 7.42, 7.81, and 7.48 ppm, respectively. The signal
corresponding to the methyl protons of 2 in the new complex
emerges as a singlet centered at δ ) 0.23 ppm. All four proton
signals grow at the expense of the signals that appeared in the
second phase of the titration. On the basis of these observations,
the existence of nitrogen-palladium interactions in the initial
phase of the titration and the formation of dichloride-bridged
bimetallic complexes, we propose that three complexes of
different stoichiometry are generated from the titration of
2·COD with 3 (Scheme 3).
complex is also shown.
of this complex by the incremental addition of 3 produces an
intermediate dichloride-bridged bimetallic complex 22 ·32 that
finally collapses to the already characterized 2·32 complex. On
the basis of the available data, the assignment of a molecular
structure for the complex with the 3:2 stoichiometry is by no
means absolute. At most, the structure we propose for the 23 ·32
complex should be considered a likely scenario. The analysis
1
of the H NMR titration data is facilitated by the fact that the
species involved in the three consecutive equilibria are in slow
chemical exchange on the 1H NMR time scale. The high kinetic
stability observed for these complexes together with the fact
that they are formed quantitatively at their exact molar ratio
are indicative that they should also exhibit high thermodynamic
stabilities.
In order to gain insight into the thermodynamic stabilities of
the complexes and to verify our proposed binding scheme for
the interaction of 2·COD with 3, we also probed their binding
interactions using ITC experiments. Figure 4a depicts the
(exothermic) heat evolved per injection for the calorimetric
titration of 2·COD with 3 in terms of µcal/second plotted
against time in minutes, while Figure 4b corresponds to the
normalized integrated data in terms of kcal/mol of injectant
plotted against molar ratio ([2·COD] ) 9.2 × 10-4 M and [3]
) 13.3 × 10-3 M). The visual analysis of the integrated ITC
data reveals the existence of two sigmoidal binding isotherms
with inflection points centered at 3/2·COD molar ratios of ∼0.7
and 2. The two sigmoidal isotherms are separated by an abrupt
reduction in the heat produced by the injections close to a molar
ratio of 1. As mentioned above, the inflection point of a
sigmoidal ITC binding isotherm is related to the stoichiometry
of the complex that is produced. Taken together, these observa-
tions suggest the formation of three complexes during the
titration of 2·COD with 3 having stoichiometric molar ratios
of ∼0.7, 1, and 2. This result is in line with the conclusions
1
On the basis of the molar ratios that caused the transitions
between the three titration phases, we assign a stoichiometry
of 23 ·32 to the first complex formed in solution. The destruction
drawn from the H NMR titration experiments. Unfortunately,
the binding model that is required to fit the ITC data is not
included in the Microcal Data Analysis software.

´
11484 J. Phys. Chem. B, Vol. 113, No. 33, 2009
Gonza´lez-Alvarez et al.
TABLE 1: Calculated Values for the Apparent27 Overall
Stability Constants and Binding Enthalpies of the Complexes
Formed during the Coordination Process of 3 and 5 with
2·COD
complex (2m ·3l)
log K(2m ·3l) [units]
∆H° (kcal/mol)
23 ·33
22 ·32
2·32
27.5 [M-4
23.2 [M-3
14.5 [M-2
]
]
]
-10.0
-11.0
-19.3
complex (2m ·5l)
log K(2m ·5l) [units]
∆H° (kcal/mol)
22 ·52
2·52
22.0 [M-3
]
]
-12.0
-21.2
14.3 [M-2
SCHEME 4: Binding Model Used to Fit the ITC Data
Obtained in the Study of the Complexation of 2 ·COD
with 5
are clearly distinguishable in the heat versus time plot. The binding
isotherm (normalized integrated heat vs molar ratio) assigns the
inflection points of each one of the two sigmoidal curves at molar
ratios of 1 and 2, respectively. This result is in complete agreement
with the expected coordination behavior of triphenylphosphine 5
Figure 4. Experimental ITC measurements for the titration of
2 · COD with 3. (a) Heat vs time plot; (b) binding isotherm
(normalized integrated heat vs molar ratio) fit to a binding model
assuming that 2 is involved in the formation of three complexes
(3:2, 2:2 and 1:2) with 3.
We succeed in comparing the experimentally measured heat
with a calculated (Figure 4b) one by combining the data obtained
from the simulation module for complexation speciation of the
SPECFIT software with a spreadsheet in which the equations
for the calculation of the heat content (Q(i)) after any injection
i and for the correct expression for the heat release (∆Q(i)) for
the injection i are implemented (see Supporting Information).
The fitting process involves the following steps: (1) defin-
ing the stoichiometry of all possible complexes in the binding
model; (2) assigning estimated overall stability constants to each
species; (3) performing the equilibrium speciation, using
SPECFIT and the proposed theoretical model, at the exact
concentrations used to carry-out the ITC experiments; (4)
making initial guesses of ∆H°; (5) calculating of ∆Q(i) for each
injection using the spreadsheet and the simulated speciation
derived from SPECFIT and (6) comparing the calculated ∆Q(i)
values to the experimental measured heat. The procedure is
iterated starting either from step 2 or step 4 until no further
improvement in the fit can be obtained. The optimized values
calculated for the fitting variables (overall Kml and overall ∆H°)
that produce the theoretical isotherms shown in Figure 4b are
summarized in Table 1.
We also investigated the complexation of 2·COD with
triphenylphosphine 5 using ITC experiments. In this case, 5
should be involved in the formation of only two complexes with
2 with 2:2 and 1:2 stoichiometry, respectively (Scheme 4).17
Figure 5 shows the experimental ITC measurements for the
titration of 2·COD with 5. Two different sigmoidal binding events
Figure 5. Experimental ITC measurements for the titration of 2·COD
with 5. (a) Heat vs time plot; (b) binding isotherm (normalized
integrated heat vs molar ratio) fit to a binding model assuming that 2
is involved in the formation of two complexes (2:2 and 1:2) with 5.

Three Component Heterobimetallic Bisporphyrin Macrocycle
J. Phys. Chem. B, Vol. 113, No. 33, 2009 11485
SCHEME 5: Proposed Binding Model for the Complexation of Zn-Bisporphyrin 1 with the Palladium-Based Bispyridine
a
Ligand 2 ·32
a Schematic representations for the different stoichiometric states and structure of the complexes are shown.
with 2·COD and supports our hypothesis that the additional
calorimetrical process observed in the coordination of the same
metal center with 4-pyridyldiphosphine 3 should be ascribed to an
interaction involving the pyridyl nitrogen.
All titration experiments described in this section were
performed by adding incremental amounts of the Zn-bispor-
phyrin 1 (titrant) to a micromolar solution of 2·COD containing
2 equiv of pyridylphosphine 3. We have indicated that using
these experimental conditions, the diphosphine palladium
complex 2·32 is formed quantitatively in the titrand solution.
We hypothesized that the binding of the bidentate self-
assembled bispyridine ligand 2·32 with the complementary
ditopic Zn-bisporphyrin 1 could form three complexes with
different stoichiometry, (a) 1·(2·32), 1:1; (b) 1·(2·32)2, 1:2, and
(c) 12 ·(2·32) 2:1 (Scheme 5). The 1·(2·32) complex corresponds
to the macrocyclic self-assembly 4. Using eqs 1-3, it is possible
to estimate statistical overall stability constant values for these
complexes11
Figure 5b shows the theoretical binding isotherm obtained
from the fitting process of the ITC data of 2·COD with 5 using
the methodology described above. The optimized values cal-
culated for the fitting variables (overall Kml and overall ∆H°)
are also summarized in Table 1. From the data summarized in
Table 1, we can draw several conclusions. The overall stabilities
and binding enthalpies of the 2:2 and 1:2 complex formed upon
coordination of phosphines 3 and 5 with 2·COD are very
similar. Each interaction of the phosphorus atom with the Pd(II)
center in the 1:2 complexes produces a binding enthalpy of
approximately -10 kcal/mol. On the other hand, the binding
process of each phosphine, 3 or 5, to 2·COD affording the 1:2
complex follows a different binding model. Nevertheless, the
high values that we calculated for the apparent stability constants
of the 1:2 ternary complexes warrant their quantitative formation
in solution at micromolar concentration of 2·COD when 2 equiv
of 3 or 5 are added.
Study of the Molecular Self-Assembly Equilibrium Pro-
ducing the Formation of the Heterobimetallic Macrocyclic
Assembly 4. The study of the pairwise interactions of the two
molecular components of the macrocyclic assembly 4 that
contain a metal center, Zn-bisporphyrin 1 and Pd(II) unit
2·COD, with the pyridylphosphine ligand 3 have been detailed
above. These studies allowed us to experimentally quantify the
stability constants for the intermolecular interactions involved
in the self-assembly of the heterobimetallic macrocycle 4.
However, as already mentioned in the introduction, the complete
thermodynamic characterization of the macrocyclic assembly
4 also requires the determination of the value of the effective
molarity, EM, for the intramolecular interaction that closes the
cycle. Therefore, in this section we describe the results derived
from the study of the full self-assembly process of the three-
components (1, 2·COD, and 3) using NMR, UV-vis spec-
troscopy, and ITC titrations directed to the evaluation of an EM
value for the macrocylic assembly 4.
K(1·(2·32)) ) K11 ) 2Km(N-Zn)2 EM
(1)
(2)
(3)
2
K(1·(2·32)2) ) K12 ) 4Km(N-Zn)
2
K(12 ·(2·32)) ) K21 ) 4Km(N-Zn)
We performed three simulations for the speciation profiles
in the titration of the self-assembled bispyridine ligand ([2·32]
) 1 × 10-4 M) with Zn-bisporphyrin ([1] ) 0-1.6 × 10-4 M)
assuming EM values for the macrocycle 1·(2·32) ) 4 of 1,
10-2 and 10-3 M. Even for an EM value of 10-3 M, the
simulation predicts that the 1:2 complex, 1·(2·32)2, will be
formed in a maximum concentration of 4 × 10-6 M (<8% of
the 2·32 bispyridine ligand will be involved in this complex)
in the initial phases of the titration while the 2:1 complex,
12 ·(2·32), will reach a maximum concentration of 9.5 × 10-6
M (only 9.5% of the self-assembled bispyridine ligand 2·32 will
be involved in this complex) by the end of the titration (see
Supporting Information). In short, even for the smallest EM
value considered it is reasonable to assume that the titration of
the self-assembled bispyridine ligand 2·32 with 1 will afford

´
11486 J. Phys. Chem. B, Vol. 113, No. 33, 2009
Gonza´lez-Alvarez et al.
Figure 6. Series of UV-vis spectra obtained during the titration of a
toluene solution of 2·32 (1.04 × 10-4 M) with up to 1.6 equiv of 1.
Inset shows the calculated absorption spectrum (epsilon scale) of
macrocycle 4 (solid black line) and the experimentally determined
absorption for free 1 which was maintained fixed during the fit (blue-
dashed line).
Figure 7. ITC titration data of 2·32 with 1 in toluene solution. (a)
Raw data in terms of µcal/second plotted against time in minutes; (b)
normalized integration data in terms of kcal/mol of injectant plotted
against molar ratio. The normalized integration data (points) were fit
to one set of site theoretical binding model (line). Fitting function
parameters are summarized in the inset of the figure. Several integrated
data points were removed prior to fitting.
the nearly exclusive formation of the 1:1 macrocyclic complex
1·(2·32) ) 4, under the selected conditions.
Taking into account the results of the simulations described
above, we performed a UV-vis titration of a toluene solution
of 2·32 (1.04 × 10-4 M) with a solution of bisporphyrin 1 (10-3
M) until 1.6 equiv of 1 are added. We followed the spectroscopic
changes experienced in the Q-band region of 1, because the
Soret band of the Zn-bisporphyrin 1 is too intense in the
concentration range used. As shown in Figure 6, the incremental
additions of the Zn-bisporphyrin 1 produced not only the
expected increase in the absorbance of the Q bands but also a
small hypsochromic shift (blue shift) of their maxima. It is well
known that Zn-porphyrins typically experience a red shift of
the Q bands upon axial coordination with amines. Therefore,
the spectroscopic behavior observed during the titration indicates
that although initially Zn-bisporphyrin 1 is involved in the
formation of interactions with the bispyridine ligand 2·32, as
the concentration of 1 increases, the percentage of free 1 in
solution also increases with respect to complexed 1, resulting
in the observed blue shift of the Q bands. This explanation is
also in complete agreement with the simulated speciation profiles
(see Supporting Information).
We analyzed the experimental titration data using the
SPECFIT software and a 1:1 binding model that considers only
two colored species, free Zn-bisporphyrin 1 and the macrocyclic
complex 4, that is, the 1:1 complex formed by the ditopic
interaction of 1 with the self-assembled palladium bispyridine
moiety 2·32. We calculated a stability constant value for the
macrocyclic assembly 4 of K11 ) K(1·(2·32)) ) 1 ( 0.2 × 105
M-1. Using this value in eq 1 (Km(N-Zn) ) 3.15 × 103 M-1), we
determined an EM value for the cyclic assembly 4 of 1 × 10-2
M, which is in complete agreement with the values obtained
for related macrocyles.28 The fact that the determined EM value
is higher than our lowest estimate (EM ) 1 × 10-3 M) used in
the simulation profiles justifies the use of the simple 1:1 binding
model to fit the experimental titration data.
bisporphyrin solution ([1] ) 9.75 × 10-4 M) in the syringe.
The integrated ITC data show a sigmoidal binding isotherm with
an inflection point centered at a molar ratio equal to 1 (Figure
7). This finding is in complete agreement with the formation of
a simple 1:1 complex. The titration data were fit to a theoretical
binding curve using the one set of sites binding model available
in the Microcal Data Analysis software. The values determined
for the fitting parameters (n, K, ∆H) are summarized in the inset
of Figure 7b. The calculated stability constant for the macro-
cyclic assembly 4 using this technique is K11 ) K4 ) 1.0 × 105
M-1, a value in good agreement with the one we derived
previously from the UV-vis titration. It is worth mentioning
that the enthalpy measured for the ditopic binding of Zn-
bisporphyrin 1 and the self-assembled bispyridine 2·32 (∆H )
-16.2 kcal/mol) is approximately 4 kcal/mol smaller than the
enthalpy value expected for the formation of two N-Zn
interactions. We have previously determined a value of ∆H )
-10.6 kcal/mol for an optimum N-Zn interaction in the ITC
titration of 1 with 3. Most likely, this enthalpy difference is a
direct consequence of the cyclic strain and distortion from the
optimum geometry required for the axial coordination of an
amine ligand with a Zn-porphyrin that occurs in macrocycle 4.
Both effects are imposed by the ditopic coordination of the self-
assembled bipyridine 2·32 to 1.
Calculated Speciation for the Self-Assembly of Macrocycle
4. The calculated value for the effective molarity, EM, of
macrocycle 4 (EM ) 1 × 10-2 M) can be incorporated in eq 4
2
together with the values Km(N-Zn)2 ) 9.9 × 106 M-2 and Km(P-Pd)
) 3.2 × 1014 M-2 (see model systems) to estimate the statistical
overall stability constant of the heterobimetallic macrocyclic
assembly 4 from its three molecular components 1, 3, and
Next, we studied the assembly of macrocycle 4 by means of
an ITC experiment using a toluene solution of the self-assembled
bispyridine ligand ([2·32] ) 1 × 10-4 M) in the cell and a Zn-
2·COD as K4 ) 6.3 × 1019 M-3
.

Three Component Heterobimetallic Bisporphyrin Macrocycle
K4 ) 2Km(Zn-N)2Km(Pd-P)2 EM (4)
J. Phys. Chem. B, Vol. 113, No. 33, 2009 11487
equiv of 3, we calculated that 80% of 1 is involved in the
formation of macrocyle 4 ([4] ) 4.0 × 10-3 M). The remaining
species are present in concentrations below 2 × 10-4 M. In
other words, the binding model predicts that macrocycle 4
assembles almost quantitatively at 5 mM concentration under
strict stoichiometric control and should be the only detectable
species using 1H NMR spectroscopy. To validate this hypothesis,
With this estimate in hand, we became interested in determin-
ing numerically the effect that the concentration (entropy) has
on the relative stabilization of the cyclic assembly 4 in favor of
other species that can be formed during its assembly process.
For this reason, we simulated the speciation of a solution
containing an equimolar mixture of 1 and 2·COD in the
presence of 2 equiv of 3 at two different concentrations, (a) 5
mM, which corresponds to a typical concentration for the
1
we collected a H NMR spectrum of a toluene-d8 solution
containing an equimolar mixture of 1 and 2·COD ([1] )
[2·COD] ) 5 × 10-3 M), with 2 equiv of the pyridylphosphine
3 (Figure 8). The 1H NMR spectrum of the mixture shows that
the observed number of proton signals, their chemical shifts
values and multiplicities are fully consistent with the almost
exclusive presence in solution of macrocycle 4. This result again
agrees with the prediction derived from the binding model
developed to describe the solution behavior of the three
component mixture. Most of the signals in the 1H NMR
spectrum were assigned to the corresponding protons in the
macrocyclic structure 4 with the help of 2D NMR experiments
(COSY and ROESY) (see Supporting Information and Figure
8 for proton assignments). The ditopic coordination of the 2·32
bipyridyl ligand, through its two nitrogen atoms, to the two Zn-
bisporphyrin component 1 is clearly evidenced by the high
thermodynamic stability of the assembly and the upfield shift
experienced by the pyridyl protons in the macrocylic structure.
The signals of the protons in the R and ꢀ positions of the
pyridine ring (H12 and H11 in Figure 8) of 2·32 resonate at δ )
5.62 and 3.33 ppm, respectively, when embedded in macrocyclic
assembly 4. This intense upfield shift is characteristic for protons
of pyridyl ligands axially coordinated to Zn-porphyrins and is
due to their close proximity to the porphyrin ring current
anisotropy. Concomitantly, and as consequence of the ditopic
coordination with 2·32, all the proton signals of the bisporphyrin
units in 1 are shifted slightly downfield. Thus, the amide protons
of 1 (H3), which resonate at δ ) 9.46 ppm in the free
bisporphyrin, appear at δ ) 10.04 ppm in assembly 4. On the
other hand, the complete displacement of the COD ligand from
2·COD was confirmed by the exclusive observation of proton
signals corresponding to free cyclooctadiene (δ ) 5.58 and 2.25
ppm). Both the methyl group and the phenyl protons of 2·32
experience upfield shifts due to coordination. This observation
suggests that these protons are also influenced by the anisotropy
of the ring current of the porphyrin. In particular, the intense
upfield shift (∆δ(H10) ) -1.63 ppm) experienced by the methyl
group of 2·32 when coordinated with 1 is better explained by
accommodating this substituent in an endo position (see
chemical structure in Figure 8). However, no additional evidence
is available to support this assignment at the present time. We
also assigned a trans geometry to the coordination of palladium
center 2 within macrocyclic structure 4. This assignment derives
1
acquisition of a H NMR experiment, and (b) 0.1 mM, the
concentration used during the UV-vis titration of 2·32 with 1.
We used an extended binding model that considers the formation
of all the species (complexes) summarized in Table 2 (see
Supporting Information for the schematic binding model).29 The
overall stability constant assigned to each species is either
derived directly from the study of the adequate model system
described in previous sections or estimated statistically using
the same methodology applied in the calculation of the overall
stability constant of 4 from its three molecular components.
The simulated composition of the mixture at the two different
concentrations turns out to be slightly different (see Supporting
Information for the graphical representations). Thus, the addition
of 2 equiv of pyridylphosphine 3 to a 0.1 mM equimolar mixture
of 1 and 2·COD induces the self-assembly of macrocycle 4 in
a concentration of [4] ) 7.2 × 10-5 M (73% of Zn-bisporphyrin
1 is involved in macrocyclic assembly 4). Other species present
in solution to a considerable extent include free bisporphyrin 1
([1]free ) 2.2 × 10-5 M) and the self-assembled bipyridyl ligand
(the 1:2 complex of Pd(II) center 2 with pyridylphosphine 3;
[2·32] ) 1.5 × 10-5 M). The remaining species in the mixture
are present at concentrations below 9 × 10-6 M. We compared
the speciation data derived from the extended binding model
with the one obtained from the UV-vis titration of the self-
assembled bipyridine ligand 2·32 with 1 discussed in the
preceding paragraphs. In the latter case, we analyzed the titration
data using a simple 1:1 binding model. The corresponding
speciation of the UV-vis titration indicated that the concentra-
tion of the 1:1 complex, the macrocyclic heterobimetallic
assembly, was [1·(2·32)] ) [4] ) 7.6 × 10-5 M when 1 equiv
of 1 is added to a 0.1 mM solution of preformed 2·32 bipyridyl
ligand (76% of 1 is involved in the macrocyclic assembly 4).
The existence of an acceptable agreement between the percent-
age of the macrocyclic assembly 4 (73 vs 76%) calculated using
the two binding models gives some indication of the quality
and validity of the theoretical treatment of self-assembly
macrocyclization to the thermodynamic characterization of the
self-assembly process of 4.
When the extended binding model was used to simulate the
speciation for a 5 mM solution of 1 and 2·COD containing 2
TABLE 2: Calculated and Estimated Values for the Overall Stability Constants of the Species Involved in the Multicomponent
Self-Assembly Equilibrium of Macrocycle 4 from Its Molecular Components, 1, 3 and 2 ·COD
species
methodology used for the determination of the K value
K value
log K
23 ·32
22 ·32
2·32
1·3
calculated experimentally model system
calculated experimentally model system
calculated experimentally model system
calculated experimentally model system
calculated experimentally model system
3.16 × 1027 M-4
1.6 × 1023 M-3
3.2 × 1014 M-2
7 × 103 M-1
27.5
23.2
14.5
3.8
1·32
1.2 × 107 M-2
6.3 × 1019 M-3
4.0 × 1018 M-3
4.0 × 1036 M-6
1.3 × 1022 M-4
5.0 × 1040 M-7
1.3 × 1022 M-4
7.0
2
1 (2·32)·(Macrocycle 4)
estimated statistically K ) 2Km(N-Zn) Km(P-Pd)2 EM
19.8
18.6
36.6
22.1
40.7
22.1
2
1·(2·32) (Acyclic, monotopic interaction)
estimated statistically K ) 4Km(N-Zn)Km(P-Pd)
2
4
1 (2·32)2
estimated statistically K ) 4Km(N-Zn) Km(P-Pd)
2
2
(1)2 (2·32)
(1)2 (2·32)2
1·(2·32)·3
estimated statistically K ) 4Km(N-Pd) Km(P-Pd)
3
4
estimated statistically K ) 16Km(N-Zn) Km(P-Pd)
2
2
estimated statistically K ) 4Km(N-Zn) Km(P-Pd)

´
Gonza´lez-Alvarez et al.
11488 J. Phys. Chem. B, Vol. 113, No. 33, 2009
Figure 8. 1H NMR spectrum of the self-assembled heterobimetallic bisporphyrin macrocycle 4 ([1] ) [2·COD] ) 5 × 10-3 M, [3] ) 1 × 10-2
M) in toluene-d8, 500 MHz. “*” designates impurity.
from the observation of a single phosphorus signal in the 31P
NMR spectrum of the cyclic structure (δ(31P, 4) ) 31.88 (s)
ppm, δ(31P, 2·32) ) 32.46 ppm, and δ(31P, free 4) ) -3.87
ppm). Unfortunately, additional ROESY experiments did not
reveal any sign of intermolecular contacts between the different
molecular components of macrocycle 4.
Finally, it is worth mentioning that the quantitative self-
assembly of 4 at 5 × 10-3 M concentration under strict
stoichiometric control was also observed in deuterated chloro-
form (see Supporting Information). In chloroform solution, the
proton signals of cyclic assembly 4 are sharper and better
resolved than in toluene. This is likely because the dynamic
processes in which 4 is involved are much faster in chloroform
solution.
the detailed understanding of the solution behavior derived from
this thermodynamic study provides a useful knowledge base
for the design and construction of other self-assembled systems,
as well as for rationalizing their applications in different
functions.
Acknowledgment. We thank DGI, Ministerio de Ciencia y
Tecnolog´ıa (CTQ2008-00222/BQU, Consolider Ingenio 2010
Grant CSD2006-0003), ICIQ Foundation, and Generalitat de
Catalunya DURSI (Grant 2005SGR00108) for generous finan-
cial support. A.F. thanks DURSI for a predoctoral fellowship
and A.G-A thanks Gobierno del Principado de Asturias for a
postdoctoral contract. We also thank Professor Shannon M.
Biros from the Chemistry Department of the Grand Valley State
University, Allendale, Michigan for her assistance in writing
this manuscript.
Conclusions
In conclusion, we have performed a complete thermodynamic
characterization of the self-assembly of the tetrameric macro-
cycle 4 by means of UV-vis, ITC and NMR titrations. We
have developed an extended binding model that describes the
solution behavior of the multicomponent (1, 2·COD, and 3)
equilibrium self-assembly of 4. Under strict stoichiometric
control ([1])[2] ) 2 × [3]), we have used the binding model
to predict the speciation of the multicomponent equilibrium at
two different overall concentrations of the three components.
The simulated speciation profile has been tested experimentally
Supporting Information Available: NMR and UV-vis
spectra mentioned in the text, procedure for the calculation of
the theoretical ITC binding isotherms and schematic representa-
tion of the multicomponent (11) equilibrium considered in the
prediction of the speciation of the assembly process of macro-
cycle 4 are included. This material is available free of charge
via the Internet at http://pubs.acs.org.
References and Notes
(1) Doyle, M. P. Angew. Chem., Int. Ed. 2009, 48, 850–852.
(2) D’Urso, A.; Mammana, A.; Balaz, M.; Holmes, A. E.; Berova, N.;
Lauceri, R.; Purrello, R. J. Am. Chem. Soc. 2009, 131, 2046–2047.
(3) (a) Otsuki, J.; Narutaki, K.; Bakke, J. M. Chem. Lett. 2004, 33,
356–357. (b) Otsuki, J.; Suka, A.; Yamazaki, K.; Abe, H.; Araki, Y.; Ito,
O. Chem. Commun. 2004, 1290–1291.
1
using UV-vis and H NMR spectroscopy and has shown to
correlate nicely. We conclude that our results serve to validate
the application of the theoretical treatment of self-assembled
macrocyclization to the three component strategy. We hope that

Three Component Heterobimetallic Bisporphyrin Macrocycle
J. Phys. Chem. B, Vol. 113, No. 33, 2009 11489
(4) (a) Nakamura, Y.; Aratani, N.; Osuka, A. Chem. Soc. ReV. 2007,
36, 831–845. (b) Satake, A.; Kobuke, Y. Tetrahedron 2005, 61, 13–41. (c)
Choi, M. S.; Yamazaki, T.; Yamazaki, I.; Aida, T. Angew. Chem., Int. Ed.
2004, 43, 150–158.
(5) Drain, C. M.; Goldberg, I.; Sylvain, I.; Falber, A. Top. Curr. Chem.
2005, 245, 55–88.
(6) (a) Iengo, E.; Zangrando, E.; Alessio, E. Eur. J. Inorg. Chem. 2003,
237, 1–2384. (b) Iengo, E.; Zangrando, E.; Alessio, E. Acc. Chem. Res.
2006, 39, 841–851.
(20) Ladipo, F. T.; Anderson, G. K. Organometallics 1994, 13, 303–
306.
(21) Harvey, J. N.; Heslop, K. M.; Orpen, A. G.; Pringle, P. G. Chem.
Commun. 2003, 278–279.
(22) Slagt, V. F.; Roder, M.; Kamer, P. C. J.; van Leeuwen, P. W. N. M.;
Reek, J. N. H. J. Am. Chem. Soc. 2004, 126, 4056–4057.
(23) Ballester, P.; Costa, A.; Deya, P. M.; Frontera, A.; Gomila, R. M.;
Oliva, A. I.; Sanders, J. K. M.; Hunter, C. A. J. Org. Chem. 2005, 70,
6616–6622.
(7) (a) Chi, X. L.; Guerin, A. J.; Haycock, R. A.; Hunter, C. A.; Sarson,
L. D. J. Chem. Soc., Chem. Commun. 1995, 2563–2565. (b) Ercolani, G.
Chem. Commun. 2001, 1416–1417. (c) Ballester, P.; Oliva, A. I.; Costa,
A.; Deya, P. M.; Frontera, A.; Gomila, R. M.; Hunter, C. A. J. Am. Chem.
Soc. 2006, 128, 5560–5569.
(8) (a) Ercolani, G. J. Phys. Chem. B 1998, 102, 5699–5703. (b)
Ercolani, G.; Ioele, M.; Monti, D. New J. Chem. 2001, 25, 783–789.
(9) Ercolani, G. Struct. Bonding (Berlin) 2006, 121, 167–215.
(10) (a) Kirby, A. J. AdV. Phys. Org. Chem. 1980, 17, 183–278. (b)
Mandolini, L. AdV. Phys. Org. Chem. 1986, 22, 1–111.
(11) Ercolani, G. J. Am. Chem. Soc. 2003, 125, 16097–16103.
(12) Ercolani, G. J. Phys. Chem. B 2003, 107, 5052–5057.
(13) Capo, M.; Ballester, P. Tetrahedron Lett. 2004, 45, 1055–1058.
(14) Slagt, V. F.; van Leeuwen, P. W. N. M.; Reek, J. N. H. Chem.
Commun. 2003, 2474–2475.
(24) Ballester, P.; Costa, A.; Castilla, A. M.; Deya, P. M.; Frontera, A.;
Gomila, R. M.; Hunter, C. A. Chem.sEur. J. 2005, 11, 2196–2206.
(25) Examples of ITC applicability to supramolecular chemistry and
host-guest recognition: (a) ten Cate, M. G. J.; Huskens, J.; Crego-Calama,
M.; Reinhoudt, D. N. Chem.sEur. J. 2004, 10, 3632–3639. (b) Kerckhoffs,
J.; ten Cate, M. G. J.; Mateos-Timoneda, M. A.; van Leeuwen, F. W. B.;
Snellink-Ruel, B.; Spek, A. L.; Kooijman, H.; Crego-Calama, M.; Reinhoudt,
D. N. J. Am. Chem. Soc. 2005, 127, 12697–12708. (c) Jadhav, V. D.;
Schmidtchen, F. P. Org. Lett. 2005, 7, 3311–3314. (d) Tobey, S. L.; Anslyn,
E. V. J. Am. Chem. Soc. 2003, 125, 10963–10970. (e) Ballester, P.; Costa,
A.; Deya`, P. M.; Vega, M.; Morey, J.; Deslongchamps, G. Tetrahedron
Lett. 1999, 40, 171–174.
(26) Ladipo, F. T.; Anderson, G. K. Organometallics 1994, 13, 303–
306.
(27) For simplicity, we haven’t considered the thermodynamic equilib-
rium of formation/dissociation of the initial palladium complex 2·COD in
our binding model. Consequently, we refer to the reported values for the
overall stability constants as apparent.
(15) Gardner, M.; Guerin, A. J.; Hunter, C. A.; Michelsen, U.; Rotger,
C. New J. Chem. 1999, 23, 309–316.
(16) Weiner, M. A.; Schwartz, P. Inorg. Chem. 1975, 14, 1714–1716.
(17) Rulke, R. E.; Ernsting, J. M.; Spek, A. L.; Elsevier, C. J.; van
Leeuwen, P. W. N. M.; Vrieze, K. Inorg. Chem. 1993, 32, 5769–5778.
(18) (a) SPECFIT, ver. 3.0; Spectrum Software Associates. (b) Gampp,
H.; Maeder, M.; Meyer, C. J.; Zuberbu¨hler, A. D. Talanta 1986, 33, 943–
951. (c) Gampp, H.; Maeder, M.; Meyer, C. J.; Zuberbu¨hler, A. D. Talanta
1985, 32, 95–101.
(28) Ballester, P.; Costa, A.; Deya, P. M.; Frontera, A.; Gomila, R. M.;
Oliva, A. I.; Sanders, J. K. M.; Hunter, C. A. J. Org. Chem. 2005, 70,
6616–6622.
(29) For a related method to predict the speciation in multicomponent
equilibrium self-assembly, see Fyles, T. M.; Tong, C. C. New J. Chem.
2007, 31, 296–304.
(19) Slagt, V. F.; Reek, J. N. H.; Kamer, P. C. J.; van Leeuwen,
P. W. N. M. Angew. Chem., Int. Ed. 2001, 40, 4271–4274.
JP9053742