Di- versus Trinuclear Copper(II) Cryptate for the ... - ACS Publications

Article pubs.acs.org/IC

Di- versus Trinuclear Copper(II) Cryptate for the Uptake of Dicarboxylate Anions Catarina V. Esteves,† Pedro Mateus,† Vânia André,‡ Nuno A. G. Bandeira,‡,§,∥ Maria José Calhorda,∥ Liliana P. Ferreira,⊥,# and Rita Delgado*,† †

Instituto de Tecnologia Química e Biológica António Xavier, Universidade Nova de Lisboa, Av. da República, 2780-157 Oeiras, Portugal Centro de Química Estrutural, Instituto Superior Técnico, Universidade de Lisboa, Avenida Rovisco Pais, 1049-001 Lisboa, Portugal § Institut Català d’Investigació Química (ICIQ), Barcelona Institute of Science and Technology (BIST), 16, Av. Països Catalans, 43007 Tarragona, Spain ∥ Centro de Química e Bioquímica, DQB, Faculdade de Ciências, Universidade de Lisboa, 1749-016 Lisboa, Portugal ⊥ BioISI, Faculdade de Ciências, Universidade de Lisboa, Campo Grande, 1749-016 Lisboa, Portugal # Department of Physics, University of Coimbra, 3004-516 Coimbra, Portugal ‡

S Supporting Information *

ABSTRACT: Searching for receptors selective for the binding of dicarboxylate anions, the copper(II) complexes of the known ditopic octaazacryptand (t2pN8), derived from bistren [tren = tris(2aminoethyl)amine] linked by p-xylyl spacers, were re-examined, with the expectation of observing a selective binding of oxalate or malonate by bridging the two copper centers of the [Cu2(t2pN8)(H2O)2]4+ receptor. Solution studies involving the supramolecular species formed by the receptor and oxalate (oxa2−), malonate (mal2−), and succinate (suc2−) anions are reported. The determined association constants revealed the unexpected formation of a 3:1:1 Cu/t2pN8/ anion stoichiometry for the cascade species with oxa2− and mal2−, and the single crystal X-ray structural characterization confirmed the presence of tricopper(II) complexes, with an unusual binding mode for the dicarboxylate anions. Each of the two copper atoms binds four nitrogen donor atoms of the t2pN8 cryptand and one additional hydroxide group, which bridges to the third copper. The square planar environment of this one is complete with two oxygen atoms from the oxalate (or the malonate). The two copper centers bound to the tren heads are ∼6.5 Å apart, each one at about 3.5 Å from the third Cu center. These studies were complemented by SQUID magnetization measurements and DFT calculations. The magnetic susceptibility measurements of the oxalate cascade complex showed a strong magnetic coupling (J = − 210 cm−1) between the Cu centers at a short distance (3.5 Å), while the coupling between the two equivalent Cu atoms (∼6.5 Å) was only −70 cm−1. This result was well reproduced by DFT calculations.

■

crystal structures of bridged CN−3 and N3−6 showed a better fit of the N3− anion in the cavity, whereas for the inclusion of CN− the p-xylyl spacers needed to twist in order to decrease the distance between the two copper centers from 6.19 to 5.18 Å. This indicates that in spite of the rigid spacers, the cryptand has enough flexibility to adapt to the anion by adopting a different conformation with the consequent energy penalty and loss of selectivity.2 Bearing these data in mind, we expected to observe a selective binding of oxalate (oxa2−) or malonate (mal2−) by bridging the two copper centers of the [Cu2(t2pN8)(H2O)2]4+ receptor. Surprisingly, instead of simply bridging the two available sites of the dinuclear complex, both oxa2− and mal2− anions were involved in the formation of a trinuclear copper cryptate, while

INTRODUCTION In search of dinuclear copper complexes for selective binding of dicarboxylate anions of short chain length, we decided to revisit the ditopic octaazacryptand (t2pN8) derived from bistren [tren = tris(2-aminoethyl)amine] linked by p-xylyl spacers, see Chart 1.1 This cryptand, and others in which tren heads are bound by related spacers (such as m-xylyl or furane), are able to coordinate two copper(II) cations, each one adopting more or less distorted square pyramidal or trigonal bipyramidal geometries with one coordinating position pointing to the three-dimensional cavity usually occupied by solvent molecules. The solvent positions can be easily replaced by an anion, given that the intermetallic distance and the shape of the formed pocket are appropriate for the formation of a cascade complex.2 Cascade dicopper(II) complexes with cyanide,3 cyanamide,4 hydroxide,5 azide,6 imidazole,7 and carbonate8 were found in the literature. In the case of the dicopper complex of t2pN8, © XXXX American Chemical Society

Received: April 15, 2016

A

DOI: 10.1021/acs.inorgchem.6b00945 Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

■

RESULTS AND DISCUSSION Single Crystal X-ray Diffraction Studies. The molecular structure of [Cu3(t2pN8)(OH)2(oxa)][NO3]2·2H2O·2MeOH is shown in Figure 1 along with the relevant atomic notation adopted. Selected bond lengths and angles are given in Table 1.

Chart 1. Representation of the Studied Cryptand t2pN8 and the Anions

Table 1. Selected Bond Distances (Å) and Angles (deg) in the Coordination Spheres of the [Cu3(t2pN8)(OH)2(oxa)][NO3]2·2H2O·2MeOH Complex bond lengths/Å Cu1−O1 Cu1−N1 Cu1−N2 Cu1−N3 Cu1−N4 Cu2−O1 Cu2−O2

2−

succinate (suc ) formed the expected bridge between the copper centers of the dinuclear complex through the reaction [Cu2(t2pN8)(H2O)2]4+ + suc2− ⇄ [Cu2(t2pN8) (suc)]2+ + 2H2O. The unusual behavior involving the first two dicarboxylate anions led us to study these systems in more detail. Additionally, trinuclear copper complexes, in themselves, can exhibit interesting magnetic properties and be candidates of possible synthetic models of the active sites of a number of metalloproteins, namely of blue oxidases (laccase, ceruloplasmin, particulate methane monooxygenase, and ascorbate oxidase) containing a triangular unit of copper atoms.9 In fact, these multicopper blue oxidases display a 3 + 1 arrangement of the copper atoms and catalyze the 4e−/4H+ reduction of O2 to H2O with concomitant one-electron oxidation of a variety of substrates such as ascorbate, polyphenols, and aromatic polyamines.10 In this work the solution studies involving the new supramolecular species containing oxa2−, mal2−, and suc2− are described. The association constants of the cascade species formed by the dicopper complex of t2pN8 with the three anions were determined, and solid state studies (structural characterization of the dicopper(II) complexes and SQUID magnetization measurements) were carried out. DFT calculations aimed at understanding the experimental findings were performed.

1.894(3) 2.041(4) 2.141(4) 2.195(5) 2.165(4) 1.898(3) 1.952(3) bond angles/deg

N1−Cu1−O1 N2−Cu1−N3 N2−Cu1−N4 N3−Cu1−N4 N1−Cu1−N4 N1−Cu1−N3 N1−Cu1−N2 O1−Cu1−N4 O1−Cu1−N3 O1−Cu1−N2 O1−Cu2−O2 O2−Cu2−O2′ O1−Cu2−O2′

176.01(14) 120.29(17) 123.12(17) 113.21(18) 84.42(14) 83.24(17) 83.96(16) 99.27(14) 96.61(16) 92.70(15) 175.68(13) 83.69(13) 92.16(13)

The molecule has a C2 symmetry axis that passes through the Cu2 atom and the Cg1 (center of gravity of the phenyl ring C19, C20, C21, C19′, C20′, and C21′) and bisects the oxalate anion.

Figure 1. Three views of the crystal structure of the [Cu3(t2pN8)(OH)2(oxa)]2+ complex cation determined from single crystal X-ray diffraction data. B


Article

Inorganic Chemistry

[Cu3(t2pN8)(OH)2(oxa)]2+, and the structural parameters are extremely close, though the three copper(II) atoms are all nonequivalent. Selected bond lengths and angles are given in Table 2.

Consequently, only two of the three copper(II) centers shown in Figure 1 are crystallographically independent. The cryptand adopts an ellipsoidal bowl conformation, with the phenyl ring Cg1 composing the bottom of the bowl, whereas the two sides are formed by the two phenyl rings Cg2 and Cg2′, which are almost perpendicularly oriented relative to Cg1 [83.85(13)]. The two bridgehead nitrogen atoms are at a distance of 10.585(8) Å, while the Cg2 and Cg2′ are 6.340(4) Å from each other. The two copper(II) centers, Cu1 and Cu1′, are located at 6.5173(12) Å from one another, each copper being pentacoordinate and bound to four nitrogen atoms of a tren subunit and to an oxygen atom of a hydroxide anion. The trigonal distortion calculated using the index structural parameter τ (τ = 0 for a perfect square pyramidal geometry and τ = 1 for an ideal trigonal bipyramidal geometry)11 assumes a value of 0.89, which is consistent with a trigonal bipyramidal stereochemistry. The distortion from a regular trigonal bipyramid is small, as the Neq−Cu−Neq angles do not deviate much from 120° [120.29(17), 123.12(17), and 113.21(18)°]. The copper(II) centers are located slightly above the trigonal plane [0.5490(16) Å] and directed toward the apical oxygen atoms. The Cu2 has coordination four and is bound to the two hydroxide anions and to two oxygen atoms of the oxalate anion, displaying a perfect square planar geometry. The Cu2 is at a distance of 3.5126(6) Å from Cu1 and Cu1′, which means that the three copper centers form the vertices of an isosceles triangle. The oxa2− anion is planar and coordinated to Cu2 in a bidentate fashion, forming a five-membered chelate ring. Two N−H···O hydrogen bonds between a secondary amine nitrogen atom and an oxygen atom (O2) of the oxalate anion contribute to the stability of the complex. The molecular structure of [Cu3(t2pN8)(OH)2(mal)][NO3]2· 6H2O, shown in Figure 2 along with the relevant atomic notation adopted, has features very similar to those observed in

Table 2. Selected Bond Distances (Å) and Angles (deg) in the Coordination Spheres of the [Cu3(t2pN8)(OH)2(mal)][NO3]2·6H2O Complex bond lengths/Å Cu1−O1 Cu1−N1 Cu1−N2 Cu1−N3 Cu1−N4 Cu2−O1 Cu2−O2 O1−Cu1−N1 O1−Cu1−N2 O1−Cu1−N3 O1−Cu1−N4 N1−Cu1−N2 N1−Cu1−N3 N1−Cu1−N4 N2−Cu1−N3 N2−Cu1−N4 N3−Cu1−N4 O1−Cu2−O2 O1−Cu2−O3 O1−Cu2−O4

1.909(6) Cu2−O3 2.042(8) Cu2−O4 2.185(7) Cu3−O2 2.180(7) Cu3−N5 2.152(7) Cu3−N6 1.901(6) Cu3−N7 1.938(6) Cu3−N8 bond angles/deg 177.4(3) 98.9(2) 95.8(2) 92.6(3) 83.6(3) 84.2(3) 85.1(3) 111.8(3) 124.7(3) 120.6(3) 89.9(2) 88.1(2) 176.8(3)

O2−Cu2−O3 O2−Cu2−O4 O3−Cu2−O4 O2−Cu3−N5 O2−Cu3−N6 O2−Cu3−N7 O2−Cu3−N8 N5−Cu3−N6 N5−Cu3−N7 N5−Cu3−N8 N6−Cu3−N7 N6−Cu3−N8 N7−Cu3−N8

1.966(6) 1.950(6) 1.912(6) 2.041(8) 2.237(8) 2.130(7) 2.161(8) 173.5(3) 90.4(2) 91.9(2) 175.3(3) 101.8(3) 90.7(2) 95.6(2) 82.8(3) 85.9(3) 83.6(3) 123.3(3) 111.2(3) 122.5(3)

The cryptand adopts the same ellipsoidal bowl conformation: the phenyl ring Cg1 composes the bottom and the two phenyl rings Cg2 and Cg3 are almost perpendicularly oriented relative to the latter [95.7(3) and 85.3(3)°, respectively], forming the

Figure 2. Three views of the crystal structure of the [Cu3(t2pN8)(OH)2(mal)]2+ complex cation determined from single crystal X-ray diffraction data. C


Article

Inorganic Chemistry

interfering anion).13 However, for the subsequent studies of the copper(II) complexes as receptors of anions, the KNO3 medium is the best choice. The effect of the large amount of nitrate anion in solution was taken into account by the corresponding values of the determined binding constant of NO3− with the protonated species of t2pN8 (see Table S2 of the Supporting Information). The stability constants of this macrobicycle with the Cu2+ metal ion, under the same experimental conditions, were also determined, and the results are collected in Table S4 of the Supporting Information. Mono- and dinuclear species are formed in solution; see the species distribution diagrams in Figure 3. Several mononuclear protonated complexes are formed from pH 4 to 10, and the dinuclear one, [Cu2(t2pN8)]4+, exists in the pH 4.5 to 9 range, being the dominant species at a 2:1 Cu/t2pN8 ratio, while the [Cu(t2pN8)]2+ mononuclear only forms for a 1:1 ratio at pH > 7.5, almost simultaneously with the hydroxocomplex [Cu(t2pN8)(OH)]+. The [Cu2(t2pN8)(OH)]3+ complex cation starts to be formed at a pH of about 5 and is the dominant species at pH > 8. This species results from the deprotonation of one water molecule coordinated directly to the copper center to complete the coordination sphere of the metal ion also bound to the four amine donors of each tren group of t2pN8. The deprotonation of a water molecule at such a low pH evidences a very strong binding and suggests a bridging mode coordination. Indeed, the association constant, corresponding to the equilibrium [Cu2(t2pN8)]4+ + OH− ⇄ [Cu2(t2pN8)(OH)]3+, calculated on the basis of the constants of Table S4, gives an impressive value of 7.00 in log units, as also occurred in corresponding complexes of cryptands with other spacers.14 Cascade Species Formed by the Copper(II) Complexes with Dicarboxylate Anions. At a 2:1 Cu/t2pN8 stoichiometry, the dinuclear complexes are the major species in solution at pH > 5.5, Figure 3b, which is an excellent indicator of their capability to behave as a ditopic receptor for the binding of anions, namely dicarboxylate anions. In fact, these anions are completely dissociated in the indicated pH range (see Table S3 of Supporting Information). However, as already mentioned, the OH− anion strongly competes for the uptake of other anions. The association constants of copper(II) complexes of t2pN8 with three dicarboxylate substrates were determined at the same experimental conditions, and the results are collected in Table 3. The stability constants of the copper(II) complexes of the

two sides of the bowl. Two of the copper(II) centers, Cu1 and Cu3, are pentacoordinate and bound to four nitrogen atoms of a tren subunit and to an oxygen atom of a hydroxide anion, assuming trigonal bipyramidal stereochemistries (τ values of 0.87 and 0.84 for Cu1 and Cu3, respectively). These copper(II) centers are located at 6.528(2) from one another, a distance only marginally shorter than that observed in the [Cu3(t2pN8)(OH)2(oxa)]2+ complex. The third copper center, Cu2, was also found coordinated by two hydroxide anions and two oxygen atoms, this time from a malonate anion. One of the oxygen atoms of malonate is located at 0.283(10) Å from the leastsquares plane defined by Cu2 and the three other oxygen atoms, yielding a slightly distorted square planar geometry around the metal ion. In this case, the triangle formed by the three copper(II) centers is not exactly isosceles as the Cu2···Cu3 and Cu2···Cu1 distances are slightly different (3.5967(18) and of 3.5077(16) Å, respectively). The malonate anion is coordinated to the Cu2 in a bidentate fashion, forming a six-membered chelate ring and exhibiting a skew boat conformation.12 Besides the two N−H···O hydrogen bonds involving two secondary amine nitrogen atoms and the two coordinated oxygen atoms of the anion, as observed in the [Cu3(t2pN8)(OH)2(oxa)]2+ complex cation, two water molecules mediate hydrogen bonding interactions between the two other secondary amine nitrogen atoms and the two noncoordinated oxygen atoms of the anion. Solution Studies. Acid−Base Behavior of the Cryptand t2pN8 and Stability Constants of the Copper(II) Complexes. Six protonation constants from the eight basic centers of t2pN8 could be determined by potentiometry in aqueous solution at 298.2 K and an ionic strength of 0.10 M in KNO3. The results are collected in Table S1 of the Supporting Information, together with the values determined before in KCl and KTsO.13 The values have similar magnitude in pairs, as they correspond to protonation of amine centers at alternating positions in the macrobicyclic backbone, far from each other, and therefore the differences in values for each pair are mainly due to statistical factors. The protonation constants of the tertiary amines are very low due to the presence of positive charges at short distance and could not be obtained. In KNO3, the values for the last three constants are larger than in other media, and the last two constants are inverted. This is due to the association of the NO3− anion to the more charged species of the cryptand, as also observed for Cl− and TsO− (the latter being the less

Figure 3. Species distribution diagrams calculated for the complexes of t2pN8 (L) with a Cu2+ metal ion at 1:1 (a) and 2:1 (b) Cu2+/L stoichiometries, respectively. CCu = Ct2pN8 = 1.0 × 10−3 M in a, and CCu = 2Ct2pN8 = 2.0 × 10−3 M in b. L denotes t2pN8. D


Article

Inorganic Chemistry

However, unexpectedly, trinuclear complexes were formed in solution with oxa2− and mal2− anions, as also found in the corresponding crystals (see above), but not with suc2−, as can be observed in Figure S1 of the Supporting Information. And more remarkable is the fact that the trinuclear copper(II) complexes are formed at 2:1:1 Cu/t2pN8/anion stoichiometry in aqueous solutions (see Figure 4a and b), and they are even the main species at a 2:1:1 Cu/t2pN8/anion ratio for mal2− above pH 5.5, as can be observed in the distribution diagrams of the Figure 4b. Species distribution diagrams were also presented in Figure 4c and d at a 3:1:1 Cu/t2pN8/anion ratio to show that in this case the trinuclear copper(II) complexes are the main species for both anions above pH 5. These features were confirmed by ESI mass spectrometry (see Figures S2−S5 in the Supporting Information). It is necessary to emphasize that dinuclear species are also formed for the systems containing oxa2− and mal2− and coexisting with the trinuclear ones, as clearly evidenced in Figure 4, especially in the 2:1:1 ratio (Figure 4 a and b). However, for the system with suc2−, only dinuclear species were found (see Figure S1 of the Supporting Information). Indeed, when the model including the trinuclear species was also tried for the latter case, this species was found in such a small amount (less than 10% of the total copper amount) that its stability constant could not be accurately determined. SQUID Magnetometry. Magnetization measurements of the oxalate complex [Cu3(t2pN8)(OH)2(oxa)]2+ were performed as a function of temperature from 2 to 300 K under six different magnetic fields (0.2, 0.5, 1.0, 2.0, 3.5, and 5.5 T). In Figure 5,

Table 3. Overall (βCumHhLlAa) and Stepwise (KCumHhLlAa) Association Constants of the Copper(II) Complexes of t2pN8 with the Anions at 298.2 ± 0.1 K in 0.10 M KNO3 Aqueous Solution equilibrium reactiona

oxa2−

mal2−

suc2−

log βCumHhLlAa

b

Cu2+ + 4H+ + L + A2− ⇄ [CuH4LA]4+ 2Cu2+ + H+ + L + A2− ⇄ [Cu2HLA]3+ 2Cu2+ + L + A2− ⇄ [Cu2LA]2+ 2Cu2+ + L + A2− ⇄ [Cu2LH−1A]+ + H+ 2Cu2+ + L + A2− ⇄ [Cu2LH−2A] + 2H+ 3Cu2+ + L + A2− ⇄ [Cu3LH−2A]2+ + 2H+ [CuH3L]5+ + HA− ⇄ [CuH4LA]4+ [Cu2L]4+ + HA− ⇄ [Cu2HLA]3+ [Cu2L]4+ + A2− ⇄ [Cu2LA]2+ [Cu2LOH]3+ + A2− ⇄ [Cu2LOHA]+ [Cu2L(OH)2]2+ + A2− ⇄ [Cu2L(OH)2A] [Cu2L(OH)2]2+ + [CuA] ⇄ [Cu3L(OH)2A]2+

46.29(6)

45.08(9) 43.38(5) 31.03(7) 29.01(9) 26.87(9) 25.64(5) 21.11(9) 18.99(9) 10.98(9) 8.37(4) 22.90(8) 22.03(5) log KCumHhLlAab 7.21

4.55

6.48 5.36 6.80 13.37

4.34 3.24

2.87 3.24 3.11 4.19

12.87

a

L denotes the ligand t2pN8, and A the anion (oxa2−, mal2−, and suc2−). Values in parentheses are standard deviations in the last significant figures. b

studied dicarboxylates were obtained before15 under the same experimental conditions and used in the calculations (Table S3 in the Supporting Information). The binding constant of the [Cu2L]4+ receptor with oxa2− is very high (6.48 in log units) and significantly decreases for mal2− and then for suc2−.

Figure 4. Species distribution diagrams calculated for the species formed between [CuxHyt2pN8] and each dicarboxylate. At CCu = 2Ct2pN8 = 2CA = 2.0 × 10−3 M in a and b. At CCu = 3Ct2pN8 = 3CA = 3.0 × 10−3 M in c and d. CNO3− = 0.10 M. E


Article

Inorganic Chemistry

Figure 5. Temperature dependence of the inverse magnetic susceptibility (χ−1 m , solid symbols) and of the χmT product (open symbols) obtained for 0.2 T (a) and 2.0 T (b) applied magnetic fields. The solid lines represent the fitting of the isotropic spin Hamiltonian (eq 1), using program PHI.16 The inset in a shows the χmT curve at the lowest temperatures.

curve, the average fitting parameter values were determined to be the following: J(Cu2−Cu1) = J(Cu2−Cu1′) = −210(10) cm−1, J′(Cu1−Cu1′) = −70(7) cm−1, g2(Cu2) = 1.97(6), and g1(Cu1) = g1′(Cu1′) = 2.18(1). The value obtained for the exchange parameter J shows that the antiferromagnetic coupling between Cu2−Cu1 and Cu2−Cu1′ centers is dominant. As deduced from DFT calculations (see below), this coupling is explained by the magnetic orbitals overlap between the Cu2−Cu1 and Cu2−Cu1′ ions. In spite of the 10% uncertainty in exchange parameter J′, its negative sign seems to be unequivocal, indicating the existence of magnetic frustration among the three Cu ions of the oxalate complex. DFT Calculations. Calculations were performed in order to rationalize the formation of the trimetallic cascade complex with the oxalate anion and to study the spin coupling. Since no dinuclear copper(II) derivative was obtained, the energy balance for the reaction between [Cu2(t2pN8)]4+ and oxa2− or [Cu(oxa) (OH)2)]2− was determined, as shown in eqs 2 and 3. The formation of the trinuclear species is more exergonic than the binding of oxa2−. An energy decomposition energy analysis was then performed on both the intended ([Cu2(t2pN8) (oxa)]2+) and the obtained structure ([Cu3(t2pN8)(OH)2(oxa)]2+), partitioning them in two fragments, one of them always being the [Cu2(t2pN8)]4+ cation and the other oxa2− or [Cu(oxa) (OH)2]2−. The bonding energy, ΔEbonding, can be defined as the difference between the energy of the optimized molecule and the sum of the energies of the fragments in their optimized geometry. In the energy decomposition energy analysis, it is taken as the sum of the interaction energy ΔEint (the energy difference between the final molecule and the energy sum of its frozen fragments) and the fragment preparation energies ΔEprep (the energy difference between the geometry of the frozen fragments in the final molecule and the optimized fragments, in their resting state). The difference in interaction energies (ΔEint(3) = −577 kJ mol−1 and ΔEint(2) = −450 kJ mol−1) between both association reactions defined as ΔEint(3) − ΔEint(2) = ΔΔEint = −127 kJ mol−1 is also negative as indeed ΔΔEbonding= −119 kJ mol−1. The reason for this favoring of process 3 might be explained by the higher electron richness of the [Cu(oxa) (OH)2]2− anion with respect to a competing oxalate anion. The overall fragment to fragment electron population change in 3 yields 0.28 e− (both spins) flowing from [Cu(oxa) (OH)2]2− to [Cu2(t2pN8)]4+, whereas in 2 this value is only 0.11e− from the oxalate to Cu dimer. Another thermodynamic analysis addressing the addition of Cu(κ2-dianion) to the [Cu2(t2pN8)(OH)2]2+ cryptate or substitution

the molar magnetic susceptibility χm (= M/H) obtained from the magnetization data measured under 0.2 and 2.0 T is plotted in the form of χ−1 m and χmT vs temperature. In both curves, a χmT value of 0.6 cm3 K mol−1 was obtained at 300 K, which is well below 1.25 cm3 K mol−1 expected for three magnetically noninteracting copper(II) ions with S = 1/2 (g = 2). For low temperatures, on the other hand, the χmT product decreases to 0.48 cm3 K mol−1 at 20 K, which is above the expected value (0.38 cm3 K mol−1) for a model of one isolated Cu(II) with S = 1/2 and an antiparallel coupling of the other two Cu(II) ions (S = 0). Taking into account the crystallographic data, namely the distances between the copper(II) ions and the fact that the local symmetry and environment of two of them are equivalent, the spin coupling was modeled assuming a simple isosceles triangle exchange between the three S = 1/2 spins (as sketched in Figure 6), with two different exchange pathways (J21 = J21′ = J and J11′ = J′) and g factors (g2 and g1 = g1′).

Figure 6. Schematic representation of the three Cu ions in the metallic center of the oxalate complex, indicating the exchange pathways between them.

The χm(T), χ−1 m (T), and χmT(T) curves were fitted simultaneously considering this model and using the isotropic Spin Hamiltonian below (eq 1) and program PHI. The best fitting curves in the case of data collected under 0.2 and 2.0 T are represented as solid lines in Figure 5a and b, respectively. Ĥ = −2[J(S2̂ ·S1̂ ) + J(S2̂ ·S1̂ ′) + J (́ S1̂ ·S1̂ ′)] + μB (g1S1̂ + g1 ′S1̂ ′ + g2S2̂ ) ·B⃗

(1)

As shown in the inset of Figure 5a, a small anomaly, only visible in the χMT curves and less pronounced for higher applied magnetic fields, appears around 12 K. In all χMT(T) curves, a steep decrease is found for temperatures below 6 K. As a single fit of eq 1 did not reproduce reasonably the behavior of all six different χMT(T) curves over the temperature range above the anomaly, separate fits were performed for curves obtained under different applied fields allowing small variations of the fitting parameters. After obtaining the best fit for each F


Article

Inorganic Chemistry of two OH− anions by each of the three dianions was also carried out, in order to account for the appearance of the unexpected tricopper structures and to clarify the coordination of the succinate anion. The results are summarized in Table 4 (see also Figures S6 and S7 in the Supporting Information). Table 4. Energy Change (ΔE, kJ/mol) for the Addition of Cu(κ2-dianion) to the [Cu2(t2pN8)(OH)2]2+ Cryptate (1) or for Substitution of two OH− Anions by each of the three Dianions (2) [Cu 2(t 2pN8)(OH)2 ]2 + (aq) + X (aq) → [Cu 2(t 2pN8)(OH)2 X]2 + (aq) 2+

[Cu 2(t 2pN8)(OH)2 ]

2+

→ [Cu 2(t 2pN8)X]

2−

(aq) + X

Figure 7. Calculated spin ladder of the [Cu3(t2pN8)(OH)2(oxa)]2+ dication showing the relative position of the spin state functions (in dark bold, one quartet and two doublets) and the (nonprojected) broken symmetry determinant energies (in gray) with respect to the highest spin state. All the values were calculated at the B3LYP-D3/ TZP+DZP level of theory.

+

(aq) + 2H3O

(aq) + 4H 2O ΔE

X 1

−464.9 −442.6 −418.0

Cu(κ2-C2O4) Cu(κ2-O2CCH2CO2) Cu(κ2-O2C(CH2)2CO2)

in comparison to the classic copper(II) tetraacetate dihydrate (J = −149 cm−1). The positioning of the microstates may be derived by employing the Heisenberg−Dirac−Van Vleck Hamiltonian (Figure 7). The spin densities of BS#1 and BS#2 are displayed in Figure 8, and they provide insight into the mechanism of the

2 C2O42− [O2CCH2CO2]2− (O2C(CH2)2CO2)2−

−481.3 −444.6 −425.1

The most notable aspect is that both processes 1 and 2 become less favorable in the series of anions with the increase of their aliphatic chain, reflecting a better steric fit for the smaller oxalate. This conclusion is consistent with the values of the binding constants listed in Table 3. The same trend is observed when the ligand bridging the two copper(II) centers is both the dianion or the square planar copper anion complex, due to the widening bite angle of the bulkier succinate. Interestingly, when the succinate dianion coordinates to [Cu2(t2pN8)]4+, the copper centers recognizably adopt an octahedral geometry (see Figure S7 in Supporting Information) since there is now a better steric fit of the dicarboxylate groups to the cavity left by the framework spacer. Overall, the [Cu(dicarboxylate)] complex competes with the dicarboxylate anions for coordination to the [Cu2(t2pN8)(OH)2]2+ cryptate. We also computationally analyzed the spin coupling in the [Cu3(t2pN8)(OH)2(oxa)]2+ complex to interpret the magnetic measurement data and to examine the nature of the ground state. Since it has C2 symmetry, the two copper(II) ions Cu1 and Cu1′ are equivalent (Figure 7). The DFT broken symmetry (BS) approach17−19 was employed to calculate the magnetic exchange couplings 2J and 2J′ (see eq 1 and Experimental Section). The spin coupling between both Cu1 centers will surely be smaller so that their magnetic exchange coupling (J′) will be weaker than the principal Cu1−Cu1 exchange route (J). Since as many J values as broken symmetry determinants can be obtained, a system of two equations with two unknowns may be solved, equating the J values with energy differences from the highest spin state, the latter generally represented by a single determinant. The two BS configurations are BS#1 = |Cu1↓ Cu2↑ Cu1′↑| and BS#2= |Cu1↑ Cu2↓ Cu1′↑|, with energies 459 and 782 cm−1 below the energy of the quartet state, respectively (Figure 7). The mean expectation value of the spin squared operator (⟨S2⟩) in both BS configurations was 1.7. These energies represent 2J + 2J′ and 4J, respectively, thus yielding the values J = −196 cm−1 and J′ = −34 cm−1. The value of J in particular is remarkably high

Figure 8. Calculated spin densities of spin states BS#1 (left) and BS#2 (right).

magnetic coupling. Considering the crystal field splitting of the trigonal bipyramidal (D3h) ligand field in the lateral Cu1 ions as e″4e″4a11, the 3dz2 SOMO in the a1 representation will be the magnetic orbital at the two Cu1/Cu1′ sites. Moreover the square planar Cu2 has a local eg4a1g2b2g2b1g(3dx2−y2)1 electronic configuration. In the context of the Goodenough−Kanamori rules,20−22 the sign of J is dictated by the overlap between these magnetic orbitals which also depends on the angle between these magnetically active centers. If the Cu1−Cu2−Cu1′ angle were 180°, the predicted J would be positive (ferromagnetic) since the direct overlap between the 3dz2 (Cu1) and the 3dx2−y2 (Cu2) orbitals would be zero and potential exchange (the intersite exchange integral) would be the only mechanism at play. But in [Cu3(t2pN8)(OH)2(oxa)]2+, the angle between the copper sites is 136°, and thus in this geometry these overlaps are far from zero. In fact, this overlap can be estimated from the local spin densities of the Cu1 and Cu2 ions as expounded in the seminal paper of Caballol et al.:23 √(ρHS2 − ρBS#22), with regards to BS#2 where only the Cu1−Cu2 coupling is present. The calculated values of ρHS = 0.647 and ρBS#2 = −0.610 yield SCu1Cu2 = 0.217, which is a considerable value. The calculated values of the magnetic exchange coupling J and J′ reproduce very well the experimental trends, respectively −210 and −70 cm−1, with some underestimation (by 36 cm−1) of the weak coupling between the two equivalent Cu1 centers. G


Article

Inorganic Chemistry

■

for 0.025 M malonic acid aqueous solution. Turquoise-blue single crystals suitable for X-ray crystallographic determinations were obtained within a month. Yield: 15%. X-ray Crystallography. Crystals of [Cu3t2pN8(OH)2(oxa)][NO3]2·2H2O·2MeOH and [Cu3t2pN8(OH)2(mal)][NO3]2·6H2O suitable for X-ray diffraction study were mounted with Fomblin in a cryoloop. Data were collected on a Bruker AXS-KAPPA APEX II diffractometer with graphite-monochromated radiation (Mo Kα, λ = 0.71073 Å) at 150 K. The X-ray generator was operated at 50 kV and 30 mA, and the X-ray data collection was monitored by the APEX231 program. All data were corrected for Lorentzian, polarization, and absorption effects using SAINT25 and SADABS25 programs. SIR9726 and SHELXS-9727 were used for structure solution, and SHELXL-9727 was used for full matrix least-squares refinement on F2. These three programs are included in the package of programs WINGX-Version 1.80.05.28 Non-hydrogen atoms were refined anisotropically. A full matrix least-squares refinement was used for the non-hydrogen atoms with anisotropic thermal parameters. All of the hydrogen atoms were inserted in idealized positions and allowed to refine in the parent carbon atom. It was not possible to locate the water hydrogen atoms in [Cu3t2pN8(OH)2(mal)][NO3]2·6H2O, and therefore they were not inserted in the structure. Water molecules are also disordered, justifying the short distances between some of the O atoms, but it was not possible to model it. Molecular diagrams presented are drawn with PyMOL.29 PLATON30 was used to calculate bond distances and angles as well as hydrogen bond interactions. Table 5 summarizes the data

CONCLUSIONS Dicopper(II) complexes of the known ditopic octaazacryptand t2pN8 were expected to form a selective binding with oxalate (oxa2−) or malonate (mal2−) anions by bridging the two copper centers. However, the binding studies in aqueous solution revealed a tendency for the formation of a trinuclear copper cryptate for oxa2− and mal2−, and in the solid state, single crystal X-ray diffraction structures of [Cu3(t2pN8)(OH)2(oxa)][NO3]2· 2H2O·2MeOH and [Cu3(t2pN8)(OH)2(mal)][NO3]2·6H2O were found. On the other hand, the solution studies also demonstrated the much stronger tendency for succinate (suc2−) to form the expected bridge between the copper centers of the [Cu2(t2pN8)(H2O)2]4+ receptor. In the trinuclear complexes, each of two copper atoms binds four nitrogen donor atoms of the t2pN8 cryptand and one additional hydroxide group, which bridges to the third copper. The square planar environment of this one is complete with two oxygen atoms from the oxalate (or the malonate). The two copper centers bound to the tren heads are ∼6.5 Å apart, and each one at about 3.5 Å from the third Cu center (Cu2). The magnetic behavior of the cation [Cu3(t2pN8)(OH)2(oxa)]2+ was studied in detail, and the magnetic susceptibility measurements showed a strong magnetic coupling (J = − 210 cm−1) between Cu1 and Cu2 (bound to oxalate), while the coupling between the two equivalent Cu1 atoms was only −70 cm−1. This result was reproduced by DFT calculations. Additionally, a computational thermodynamic analysis of the addition of Cu(κ2-dianion) to the [Cu2(t2pN8)(OH)2]2+ cryptate or substitution of two OH− anions by each of the three dianions revealed that, with the increasingly longer aliphatic chain of the dicarboxylate anion, both reactions become less favorable.

■

Table 5. Crystal Data and Structure Refinement Details for [Cu3(t2pN8)(OH)2(oxa)][NO3]2·2H2O·2MeOH and [Cu3(t2pN8)(OH)2(mal)]][NO3]2·6H2O

formula fw cryst form, color cryst size, mm cryst syst space group a, Å b, Å c, Å Z V, Å3 T, K Dc, g cm−3 μ(Mo Kα), mm−1 θ range (deg) reflns collected independent reflns Rint R1,a wR2b [I ≥ 2σ(I)] GOF on F2

EXPERIMENTAL SECTION

Materials and Methods. All solvents and chemicals were commercially purchased, reagent grade quality, and used as supplied without further purification. The compound t2pN8 was synthesized according to a previously reported procedure.24 The purity of the compound was confirmed by 1H NMR (Figures S8 and S9 in the Supporting Information) and elemental analysis. For the compound t2pN8, 1H NMR (400 MHz, CDCl3, 298 K; Me4Si): δ = 6.81, 3.61, 2.76, 2.59, 1.63. Anal. Calcd for C36H54N8: C, 72.2; H, 9.1; N, 18.7. Found: C, 72.3; H, 8.7; N, 18.7. Microanalyses and electrospray mass spectra (ESI-MS) were carried out by the ITQB Analytical Services Unit. The 1H and 13C{1H} NMR spectra were recorded on a Bruker Avance II+ 400 (1H at 400.13 MHz and 13C at 100.61 MHz) or a Bruker Avance DRX 300 (1H at 300.13 MHz) spectrometer at a probe temperature of 298.2 K. The reference used for the 1H NMR measurements in CDCl3 was tetramethylsilane (TMS). Syntheses of the Trinuclear Copper(II) Cryptates. Synthesis of [Cu3(t2pN8)(OH)2(oxa)][NO3]2·2H2O·2MeOH. An aqueous solution of 0.050 M Cu(NO3)2 (1.64 mL, 0.082 mmol) was added to a 2.05 × 10−3 M aqueous solution of the cryptand H6t2pN8 (20.00 mL, 0.041 mmol). The pH was adjusted to 6.00 with 0.10 M aqueous KOH and the solution stirred at room temperature (rt) for 24 h. Then, an aqueous solution of 0.025 M oxalic acid was added to the mixture and the pH readjusted to 6.00. The solution was stirred at rt for another 24 h and the pH set to 6.00. The solution was evaporated to dryness. The complex was dissolved in methanol, and the insoluble matter was filtered off. This procedure was repeated until no more precipitation was observed. The blue solution was left slowly evaporating at rt. Blue single crystals suitable for X-ray crystallographic determinations were obtained within a few days. Yield: 65%. ESI−MS (MeOH): m/z 455 [Cu3(t2pN8)(OH)2(oxa)]2+ (see Figure S5 in the Supporting Information). Synthesis of [Cu3(t2pN8)(OH)2(mal)][NO3]2·6H2O. A procedure similar to that described before was used, replacing the oxalic acid

a

[Cu3(t2pN8)(OH)2(oxa)]] [NO3]2·2H2O·2MeOH

[Cu3(t2pN8) (OH)2(mal)]][NO3]2· 6H2O

C20H29Cu1.5N5O8 562.79 block, blue 0.15 × 0.10 × 0.06 orthorhombic Pbcn 9.8537(6) 17.0537(11) 27.5549(17) 8 4630.4(5) 150(2) 1.615 1.447 2.499−26.480 65040 4690 0.0660 0.0624, 0.1364 1.160

C39H50Cu3N10O18 1137.51 block, blue 0.13 × 0.10 × 0.08 orthorhombic Pna21 17.244(5) 10.274(3) 27.349(7) 4 4845(2) 150(2) 1.559 1.387 2.362−26.376 20691 7415 0.0868 0.0537, 0.1071 1.025

R1 = Σ∥Fo| − |Fc∥/Σ|Fo|. bwR2 = [Σ[w(Fo2 − Fc2)2]/Σ[w(Fo2)2]]1/2.

collection and refinement details. CCDC-1474060−1474061 contain the supplementary crystallographic data for this Article. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif. Potentiometric Measurements. Equipment and Working Conditions. The potentiometric setup for conventional titrations consisted of a 50 mL glass-jacketed titration cell sealed from the atmosphere, connected to a separate glass-jacketed reference electrode cell by a Wilhelm type salt bridge containing 0.10 M KNO3 solution. An Orion SA720 measuring instrument fitted with a Metrohm 6.0150.100 glass electrode and a Metrohm 6.0733.100 Ag/AgCl reference electrode was used for the measurements. The ionic strength H


Article

Inorganic Chemistry of the experimental solutions was kept at 0.10 ± 0.01 M with KNO3; the temperature was maintained at 298.2 ± 0.1 K using a PolyScience910 thermostat. Atmospheric CO2 was excluded from the titration cell at the beginning of experiments by slightly bubbling purified nitrogen on the experimental solution. During titrations, a continuous nitrogen flux was used to ensure the absence of CO2. Titrant solutions were added through capillary tips at the surface of the experimental solution by a Metrohm Dosimat 665 automatic buret. The titration procedure is automatically controlled by software, allowing for long unattended experimental runs. In cases where automatic titrations could not be performed, out-of-cell (batch) titrations were carried out, and the electromotive force was measured with a Metrohm 6.0234.100 combined pH electrode previously calibrated by in-cell titrations. Measurements. Purified water was obtained from a Millipore Milli-Q demineralization system. Stock solutions of t2pN8 were prepared at ca. 2.00 × 10−3 M. Metal ion solutions were prepared in water at 0.025−0.050 M from analytical grade nitrate salts of the metal ions and standardized by titration with Na2H2edta.31 Carbonate-free solutions of the titrant KOH were prepared from a Merck ampule diluted with 1 L of water (freshly boiled for about 2 h and allowed to cool under nitrogen). These solutions were standardized by application of Gran’s method.32 A 0.100 M standard solution of HNO3 prepared from a commercial ampule was used for backtitrations. The [H+] of the solutions was determined by measurement of the electromotive force of the cell, E = E°′ + Q log[H+] + Ej. The term pH is defined as −log [H+]. E°′ and Q were determined by titrating a solution of known hydrogen-ion concentration at the same ionic strength in the acid pH region. The liquid-junction potential, Ej, was found to be negligible under the experimental conditions used. The value of Kw = [H+][OH−] was found to be equal to 10−13.78 by titrating a solution of known hydrogen-ion concentration at the same ionic strength in the alkaline pH region, considering E°′ and Q valid for the entire pH range. Measurements were carried out with ca. 0.040 mmol of ligand in a total volume of ca. 40 mL, in the absence of metal ions and in the presence of Cu2+ ions at ca. 0.5, 1.0, 1.5, and 2.0 equiv. ratio. A backtitration was always performed at the end of each direct titration in order to check if equilibrium was attained throughout the full pH range. Each titration curve typically consisted of 70−90 points at the 2.5−11.5 pH range, and a minimum of two replicate titrations were performed for each system. Batch titrations were carried out for systems displaying very slow complexation kinetics, which happened for oxalate and malonate anions in the presence of t2pN8 and Cu2+. Each of those titrations consisted of a set of independent points at different pH values prepared under the same experimental conditions used for the conventional titrations, but at 1/10 of the total volume. The tubes containing each point were tightly closed under nitrogen and kept at 298.2 K until equilibrium was reached, which was verified by pH measurement of a test tube each week. The equilibrium was generally reached after 2−3 weeks of stabilization. Thermodynamic Equilibrium Constants Determination. Overall protonation constants, βiH, of t2pN8 (L), the overall stability constants of complexes, βCumHhL, and the overall association constants of the complexes of t2pN8 with the dicarboxylates, βCumHhLAa, were calculated by fitting the potentiometric data obtained for all of the performed titrations in the same experimental conditions with the HYPERQUAD program.33 At least two titration curves for each system were fitted together. The initial computations were obtained in the form of overall constants, βHhL = [HhL]/[H]h[L], βCumHhL = [CumHhL]/ [Cu]m[H]h[L], βCumHhLAa = [CumHhLAa]/[Cu]m[H]h[L][A]a, and βCumHh−1L = βCumL(OH) × Kw. The errors quoted are standard deviations of the overall constants given directly by the program for the input data, which include all of the experimental points of all titration curves. The HYSS program34 was used to calculate the concentration of equilibrium species from the calculated constants from which distribution diagrams were plotted. The species considered in a particular model were those that could be justified by the principles of coordination and supramolecular chemistry.

SQUID Magnetometry. Magnetization measurements as a function of temperature (from 2 to 300 K) under different applied magnetic fields (up to 5.5 T) were performed using a SQUID magnetometer (Quantum Design MPMS). The magnetic susceptibility values were corrected for diamagnetism of the constituent atoms, estimated from Pascal constants. Software program PHI was used to fit an isotropic spin Hamiltonian (exchange coupling and Zeeman effect components) to the χm, χm−1, and χmT vs temperature curves, where χm is the molar susceptibility. DFT Calculations. The Amsterdam Density Functional (ADF) program package, version 2013.0135 has been used employing Becke’s36 three parameter gradient corrected exchange and Lee, Yang, and Parr’s correlation37 density functionals with Grimme’s third generation dispersion correction38 (B3LYP-D3) using Becke−Johnson damping39−41 in all of the calculations. The ZORA42,43 scalar relativistic Hamiltonian was employed with triple-ζ Slater type orbitals44 (STO) augmented with one polarization function (TZP) for copper and double-ζ STO type functions augmented with d functions (DZP) on nitrogen, carbon, and oxygen and a double-ζ (DZ) set on the hydrogen atoms. The optimizations employed the local point group symmetry of the complex (C2) as this did not change from subsequent reoptimizations of the high spin complex without symmetry constraints. The COSMO45 implicit solvation scheme was used to calculate electronic energies in aqueous solution employing the default solute radii. In the magnetic coupling calculations, the broken symmetry technique of Noodleman and Norman17 was used with the nonspin projected formula of Ruiz et al.46,47 to take full advantage of the self-interaction error present in the quantitative evaluation of the J constant with standard density functionals.48 This formula has been generalized by Bencini and Totti49 for polynuclear clusters as

ΔE(S − Smax ) =

∑ 2Jij (2sisj + sj)

(2)

i

Di- versus Trinuclear Copper(II) Cryptate for the ... - ACS Publications

Recommend Documents