Redox potential of Cu(II) and group 10 metal complexes
The oxidation chemistry of metal complexes has been widely developed in recent years, affording deep insights into the reaction mechanisms for many useful homogeneous catalytic reactions and enzymatic reactions at the active site of metalloenzymes . In the course of the studies, a large number of novel complexes have been synthesized and wellcharacterized [2-18], and especially high valent metal complexes formed as a result of redox reactions have become important in catalytic and biological systems. In general high valent metal complexes have been meant to show thecomplexes oxidized at the metalcenter, and the formal oxidation state is identical with the oxidation state of the central metal ion reported in [19-23]. For example, one-electron oxidation of potassium hexacyanoferrate(II), K4[Fe(CN)6], gives potassium hexacyanoferrate(III), K3[Fe(CN)6], whose valence state of the central iron ion is +III andagrees with the experimental valence state of the ion. In contrast to this, the formal oxidation number of the central metal ion of the complexes of iminophenolate dianion, (LAP)2-is not always identical with the experimental valence state reported in references [24,25]. In the case of [Ni(LAP)2]0, the formal oxidation state of the central nickel ion is +IV, but the experimental valence state of nickel can be assigned to be +II, and two iminophenolate dianions are oxidized to iminosemiquinonate radical anions (LSQ)- (Figure 1).
Such a difference between the formal oxidation number and the experimental oxidation state is also observed in biological systems. Recently, it has been reported that various radicals can be generated at a proximal position of the metal center in metalloproteins, and the radical can sometimes interact with the central metal ion as shown in reference . Galactose oxidase (GO) is a single copper oxidase, which catalyzes a two-electron oxidation of a primary alcohol to the corresponding aldehyde [21-24]. The active site structure of the inactive form of GO is shown in Figure 2, where two imidazole rings of histidine residues, two phenol moieties of tyrosine residues, and an acetate ion are coordinated to the copper(II) ion . One of two phenol moieties is in the deprotonated form and is coordinated to the copper ion at an equatorial position, and another phenol moiety is protonated and located at an apical position.
Conversion to the active form of GO occurs upon one-electron oxidation and deprotonation from the apical phenol moiety, which gives the structure with the two phenolate moieties coordinated to the copper center. This active form should act as a two-electron oxidant, causing the conversion from primary alcohol to aldehyde. Therefore, the formal oxidation state of the active form of GO can be described as a Cu(III)-phenolate species. Actually, the active form of GO had been considered to be a copper(III) species [25,26],but early in 1990’s, various spectroscopic studies of the active form of GO revealed the formation of the phenoxyl radical species and the Cu(II)–phenoxyl radical bond. The free phenoxyl radical is very unstable with half life estimated to be 2.4 ms at ambient conditions, while the Cu(II)–phenoxyl radical in the active form of GO has a long life; the radical is not quenched for more than one week at room temperature [22,23,27,28]. Thus, properties of the metal coordinated phenoxyl radical show a significant change from those of the free phenoxyl radical.
The proposed reaction mechanism of GO is that the primary alcohol is coordinated to the Cu(II)-phenoxyl radical species generated by molecular oxygen and oxidized by an intramolecular two-electron redox reaction with the hydrogen atom scission from the alcohol moiety to give the aldehyde and the Cu(I)–phenol species (Scheme 1) reported in [22,23]. The Cu(I)-phenol complex is oxidized by molecular oxygen to regenerate the Cu(II)-phenoxyl radical species.
For understanding the detailed mechanism of GO and the properties of the metal complexes with the coordinated phenoxyl radical, many metal−phenolate complexes have been synthesized, and their oxidation behavior and properties of the oxidized forms have been characterized in [12,13, 29-36]. Salen (Salen = di(salicylidene)ethylenediamine) and its family are one of the most important and famous ligands having two phenol moieties (Figure 3) [37,38]. One of the characteristics of Salen is its preference for a square planar 2N2O coordination environment, while the distortion from the square plane can be introduced by changing the diamine backbone. A number of metal-Salen complexes have been reported to be the very important catalysts for oxidation and conversion of various organic substrates from early 1990’s . While studies on oxidative reaction intermediates are in progress, the oxidation state of the metal ions in the active species hasnot been fully understood until now. Detailed descriptions of the oxidation state of the intermediate are sometimes complicated, because the oxidation locus on oxidized metal complexes is often different from the “formal” oxidation site [19,20]. Although “formal” and “experimental” oxidation numbers are identical in many cases, they are often used as synonyms, since the term of the physical or experimental oxidation state has not been accepted in some areas of chemistry.
The present argument is focused on recent advancesin the chemistry related with the synthesis, characterization, and reactivity of some of the one-electron oxidized metal(II)–salen type complexes, especially the complexes of group 10 metal(II) ions, Ni(II), Pd(II), and Pt(II) [39-43], and Cu(II) complexes [44-47] (Figure 3).
Various useful and interesting salen-type complexes have been synthesized and characterized until now , and detailed electronic structural studies have recently been reported fora dinuclear chelating salen-type ligand having a catechol and a tetra(amino)tri(hydroxy)phenyl moiety [49, 50]. In view of the interest in the redox properties of metal-phenolate complexes, detailed electronic structures of the one-electron oxidized complexes are discussed in order to understand the electronic structure difference between the metal-centered and ligand-centered oxidation products and its dependence on the properties of metal ions and substituents of the phenolate moiety. For this reason, this argument is described only for a few examples whose detailed electronic structures have been clarified.
2. Redox potentials of phenol and metal(II) salen-type complexes
One-electron oxidation of closed-shell organic molecules is generally difficult, and the one-electron oxidized products may be unstable. Oxidation of the free phenol similarly gives the unstable phenoxyl radical described in the introduction and references [21-23,27,28]. The potential of formation of the phenoxyl radical is predicted to be high, and actually the potential of a tri(
The oxidation chemistry of metal phenolate complexes has shown that the oxidation potentials of phenolate complexes are intermediate between those of free phenolate and free phenol . Such a trend is also applied to the salen-type complexes, and most of the Cu(II) and group 10 metal(II) salen-type complexes exhibited two reversible redox waves in the range of 0 to 1.5 V vs NHE, due to having two phenolate moieties [29-31, 39-47]. On the other hand, the metal centered one- and two-electron oxidized complexesmay possiblybe generated in the similar potential range. Some copper(III) [12,13] and nickel(III) [54,55] complexes have been reported, andtwo-electron oxidized species of the group 10 metal ions, especially some Pd(IV) and Pt(IV) complexes, have been reported.Some of the metal complexes such as K2PdCl6 and K2PtCl6are commercially available . Therefore, the experimental oxidation state can not be determined from the oxidation potential only. The redox potentials of some salen-type complexes are listed in Table 1, and the voltammograms of two copper(II) complexes are shown as examples in Figure 4 [39-47].
(E / V vs. Fc / Fc+ )
(E / V vs. Fc / Fc+ )
|Ni(Salcn)||0.46, 0.80||Ni(Salphen)||0.58, 0.80|
|Pd(Salcn)||0.45, 0.80||Cu(Salcn)||0.45, 0.65|
|Pt(Salen)||0.35, 0.94||Cu(MeO-Salcn)||0.28, 0.44|
|Ni(Salpn)||0.43, 0.69||Cu(Salphen)||0.65, 0.83|
|Pd(Salpn)||0.52, 088||Cu(MeO-Salphen)||0.38, 0.49|
|Pt(Salpn)||0.44, 0.99||Cu(Salphen-OMe)||0.41, 0.70|
From the list of Table 1, the ranges of both first and second redox potentials are relatively narrow, especially the range of the first redox potentialsbeing from 0.28 to 0.65 V. Such a narrow range may be generally ascribed to the fact that the all the complexes have a similar oxidation locus.The small potential differences are due to the substitution of the phenolate moiety and the ligand structure. However, there are some different electronic structures among the one-electron oxidized forms of the complexes in Table 1, i.e., a MIII-phenolate ground state complex, a MII-phenoxyl radical where the radical electron is fully delocalized on two phenolate moieties, a relatively localized MII(phenolate)(phenoxyl) complex, and a MII(dinitrogen ligand radical)(phenolate)2 species [39-47]. The similarity of the potentials indicates that formation of all of the oxidized species is due to a simple electron transfer without significant structural changes. Thus, the experimental oxidation state of the oxidized complexes cannot be determined only from the redox potentials.
3. Isolation and solid state characterization of one-electron oxidized complexes
One-electron oxidized complexes have been synthesized by reaction with a one-electron oxidant, such as CeIV, AgI, NO+, some organic reagents, and so on . For most complexes shown in Figure 3, one-electron oxidized species were generated by addition of AgSbF6 to the CH2Cl2 solution of metal(II) salen-type complexes. AgI with a potential of 0.799 V vs. NHE can act as a one-electron oxidant for complexes [57,58]. The oxidation method using Ag ion is useful for generation of the relatively stable oxidized complexes, since Ag0is an easily removable product formed in the course of the oxidation. Some solutions of one-electron oxidized complexes were kept standing for a few days to give the products as crystals.
The X-ray crystal structure analyses of one-electron oxidized group 10 metal salen-type complexes are shown in Figure 5 [41-43]. The structures of all these complexes were foundto be similar to those of the corresponding complexes before oxidation, whichsupports the CV results that significant structure changes did not occur in the course of the oxidation.
However, a close look into the details of the crystal structures reveals that there are subtle differences between them, and especially the oxidized Pd(II) complexes are different from the other complexes [42,43]. Comparison of the 5-membered dinitrogen chelate backbones of the Salcn and Salen complexes indicates that upon oxidation the Ni and Pt complexes exhibited a clear coordination sphere contraction due to shortening of the M–O and M–N bond lengths. On the other hand,the Pd complex showed an unsymmetrical contraction : One of the Pd–O bonds (2.003 Å) is longer than the other (1.963 Å), and the C–O bond (1.263 Å) of the phenolate moiety with the longer Pd–O bond is shorter than the other C–O bond (1.317 Å). The phenolate moiety with a shorter C–O bond length has the lengthening of the ring
The Ni and Pt 5-membered dinitrogen chelate complexes also exhibited a clear symmetrical coordination sphere contraction in both two M–O and two M–N bond lengths (ca. 0.02 Å) upon oxidation, and the C–O bond distances of these complexes are also shorter than the same bonds before oxidation [41,43]. These observations suggest that the complexes have the phenoxyl radical characteristics and that the radical electron is delocalized on the two phenolate moieties. Indeed, the XPS and K-edge XANES of an oxidized Ni complex showed the same binding energies and pre-edge peak of nickel ion as those of the complex before oxidation . These results supported that the valence state of the nickel ion is +II. In the case of Pt complexes, the XPS of the oxidized complex was slightly different from that before oxidation.The binding energiesof the Pt ion in the oxidized complex were +0.2 eV higher, and LIII-edge XANES exhibited an increasing white line . Such spectral features suggest that the oxidation state of the Pt ion in the oxidized complex is higher than +II but that the differences are rather small . Therefore, [Pt(Salen)]SbF6 can be described mainly as the Pt(II)-phenoxyl radical species, but the radical electron is fully delocalized over the whole molecule including the central metal ion . On the other hand, the six-membered NiII andPtII Salpn chelate complexes are slightly different from the 5-membered dinitrogenchelate Salcn and Salen complexes . Crystal structures of both oxidized Salpn complexes exhibited two crystallographically independent molecules in the unit cells. The M–O and M–N bond lengths do not differ substantially between the two molecules in the unit cell. The bond lengths in the coordination plane are ca. 0.02 Å shorter than those of the neutral complexes, and this contraction upon oxidation is in good agreement with the 5-membered dinitrogen chelate complexes. However, the C–O bond lengths of the two phenolate moieties differed for the two independent molecules; one of the molecules showed very similar C–O bond lengths, while the bond lengths in the other molecule were slightly different, showing a similar tendency to that of the oxidized Pd complexes. Therefore, the 6-membered Ni and Pt chelate complexes can be considered to be closer to the localized phenoxyl radical metal(II) complexes in comparison with the 5-membered chelate complex, Salcn and Salpn due to the chelate effect of the dinitrogen backbone . However, determination of the detailed electronic structure of oxidized salen-type complexes, especially 5-membered Ni and Pt chelate complexes, only from the X-ray crystal structure analysis may be difficult, since we can detect mainly the contraction of the coordination sphere, which is predicted to be also observed in high valent metal salen-type complexes.
On the other hand, the electronic structure determination of the Cu complexes are clearlymade by X-ray structure analyses. Structures of three one-electron oxidized CuII salen-type complexes are shown in Figure 6 [44,45,47]. X-ray analyses of all these complexes established that their structuresare similar to those before oxidation, indicating a simple one-electron transfer from these precursors. However, the three Cu complexes have different electronic structures.
Figure 6 shows that the oxidized Cu(II) complexes have the SbF6- counterion at different positions. The structures of the same dinitrogen backbone complexes, [Cu(Salcn)]SbF6 and [Cu(MeO-Salcn)]SbF6, indicate that a weak axial Cu-F interaction (2.76 Å) exists between the counterion and the metal center in [Cu(Salcn)]+ , whereas [Cu(MeO-Salcn)]+ has a weak F–C interaction betweenthe counterion and oneside of the phenolate moieties . This difference suggests that the oxidation locus of these complexes are different; [Cu(Salcn)]SbF6 has a Cu(III) character, while [Cu(MeO-Salcn)]SbF6 is a Cu(II)-phenoxyl radical complex. [Cu(Salcn)]SbF6 showed contraction of the coordination sphere without shortening of the C–O bonds of both phenolate moieties and distortion of the coordination plane was substantially reduced from that in Cu(Salcn) . Such structural features are in good agreement with those of the low-spin d8 Cu(III) complexes [12,13]. Indeed, the XAF and XPS studies of [Cu(Salcn)]SbF6 reported the Cu(III)-phenolate ground state, because the pre-edge of oxidized complex was more than 1 eV higher, and the binding energies of Cu ion in the oxidized complex was also 1 eV higher as compared with the neutral complex Cu(Salcn) . These characteristics are in good agreement with the Cu(III) valence state [12,13]. On the other hand, the MeO-substituted complex [Cu(MeO-Salcn)]SbF6 exhibits that the C-O bond of one-side of the phenolate moieties is shortened and that the Cu–O bond with the shortened phenolate moiety becomes longer . The counterion is close to the phenolate moiety of the shortened C-O bond. Such characteristics are in good agreement with those of the oxidized Pd(Salpn) having a relatively localized PdII(phenoxyl)(phenolate) structure. Therefore, [Cu(MeO-Salcn)]SbF6 can be described asa localized CuII(phenoxyl)(phenolate) complex .
Another one-electron oxidized complex, [Cu(Salphen-OMe)]SbF6,hasa different electronic structure from the previous two Salcn complexes . The C–C and C–O bond lengths within the phenolate rings do not differ significantly from those of the complex before oxidation, Cu(Salphen-OMe). Although there is only a slight contraction of the Cu-O and Cu-N bonds, the copper ion geometry is significantly distorted toward a tetrahedral geometry in our case (the dihedral angle of 22°between the O1-Cu-N1 and O2-Cu-N2 planes). These structural features indicate that the oxidation locus of [Cu(Salphen-OMe)]+ is neither the two phenolates nor the central copper ion. A striking feature upon oxidation is the change in the bond lengths within the phenyl ring. Further the counterion SbF6-is located close to the
4. Magnetic properties of one-electron oxidized metal salen-type complexes
The behavior of the phenoxyl radical bound to metal ions with an open-shell configuration such as Cu(II) having the d9 configuration is different from that bound to Cu(III) ground state complex, [Cu(Salcn)]SbF6. In general, Cu(III) complexes have a square-planar geometry and are diamagnetic and EPR silent due to the low–spin d8 configuration [12,13]. The Cu(II)–phenoxyl radical complex, on the other hand, has two electron spins on different nuclei, and therefore the spin–spin interaction should be considered in [29-31]. In the case of the oxidized Cu(II) salen-type complexes, since correlation between ligand
Magnetic properties of [Cu(Salcn)]SbF6in the solid state have been reported to be a temperature independent diamagnetic species with the effective magnetic moment
On the other hand, ligand oxidation and metal oxidation can be distinguished in one-electron oxidation of the group 10 metal salen-type complexes. One–electron oxidized group 10 metalsalen-type complexes have an unpaired electron with
Such a trend is also observed for the Pd and Pt complexes. The one-electron oxidized Pd complexes, [Pd(Salen)]+ and [Pd(Salpn)]+, have a similar characteristic; the band width of the signal is narrow, and the g value is close to the free electron value (
5. Characterization of one-electron oxidized M(II) salen-type complexes in solution
The electronic structure of oxidized complexes in solution is sometimes different from that in the solid state, due to some additional factors such as interactions with solvent molecules and removal of restriction from the lattice energy. One of the important examples is shown for [Cu(Salcn)]SbF6 in . This complex can be assigned to the Cu(III)-phenolate complex in the solid state described above, but in CH2Cl2 solution it is different. The absorption spectrum of [Cu(Salcn)]SbF6 exhibited two intense bands at 18600 and 5700 cm-1 in the visible and NIR region (Figure 10). These two bands were assigned by the time-dependent density functional theory calculation (TD-DFT) to the HOMO-4 to LUMO and the HOMO to LUMO transition, respectively. The HOMOs mainly consist of the ligand molecular orbitals, while contribution of the copper
One-electron oxidized Ni(Salcn) complexes also exhibit a valence state change. The one-electron oxidation of Ni(Salcn) in DMF caused a color change to purple, exhibiting a new absorption band at 476 nm. In CH2Cl2, however, a different UV-vis absorption spectrum was detected with the bands at 1100, 900 and 415(shoulder) nm (Figure 12) . The CH2Cl2 solution showed an EPR signal similar to that of the solid sample, indicating the Ni(II)-phenoxyl radical species, while the DMF solution of the oxidized species showed the characteristic Ni(III) signals. The resonance Raman spectra of both solutions are different, the CH2Cl2 solution showed the phenoxyl radical
One of the important characteristics of the oxidized group 10 metal salen-type complexes, [MII(Salen)]+, [MII(Salcn)]+, and [MII(Salpn)]+, is the appearance ofan intense NIR band at ~5000 cm-1 [42,43]. The band is assigned to the phenolate to phenoxyl radical (ligand-to-ligand) charge transfer (LLCT) by TD-DFT calculation. The analyses of the low energy NIR LLCT band reveal the degree of the radical delocalization,which can be estimated by Robin-Day classification for understanding the mixed valence system. The classification is categorized in three systems as follows: (1) a fully localized system (class I), (2) a fully delocalized system (class III), and (3) a moderately coupled system (class II) . In the case of the fully localized system, there is no characteristic LLCT band in the NIR region, while the fully delocalized system showsan intense NIR LLCT band. On the other hand, the moderate coupling system exhibits a less intense NIR band, which depends on small perturbations such as solvent polarity. UV-vis-NIR spectra of group 10 metal salen-type complexes are shown in Figure 14 and reference .
The NIR band intensity is in the order, Pt > Ni > Pd. The NIR spectrum of the oxidized Pd complex showed a less intense band in comparison to the bands of the oxidized Ni and Pt complexes, and the band wasthe most solvent dependent. From these results, the oxidized Pd complexesmay be slightly closer to the class II moderately coupled system among the oxidized group 10 metal-salen type complexes,that is, the oxidized Pd complex is a more localized system . On the other hand, the NIR band of the Pt complexes was the highest in intensity and less solvent dependent. These results strongly support that the Pt complexes belong to the class III delocalization system, which isalso supported by all other experimental results in . As compared with 5-membered and 6-membered chelate dinitrogen backbone complexes, intensity of all of the NIR bands of 6-membered complexes is decreased, indicating that the 6-membered dinitrogen chelate leads to radical localization on the ligand . It is therefore obvious that there is a clear difference between the delocalized and localized systems of the phenoxyl radical species.
The oxidation chemistry of the group 10 metal(II) and copper(II) salen-type complexes is discussed in this chapter, to show that the oxidized salen-type complexes have a variety of the oxidation products. The CV of these salen-type complexes exhibited two reversible redox waves of one-electron process and the values of the first redox potential of complexes were in a narrow range. However, one-electron oxidized complexes have different electronic structures, which are dependent onthe central metal ion, aromatic ring substituents, and the chelate effect of the dinitrogen backbone.
The electronic structure of the one-electron oxidized group 10 metal salen-type complexes is mainly a metal(II)-radical species, but there is a subtle difference in the detailed electronic structures. The oxidized nickel complexes are the delocalized phenoxyl radical species, but the valence state changesupon addition of an exogenous ligand to form the Ni(III)-phenolate species. Such a valence state change could not be detected for Pd and Pt complexes. The oxidized Pd complexes are relatively localized phenoxyl radical species, which can be described as the Pd(phenoxyl)(phenolate) complexes. On the other hand, the oxidzed Pt(II) complexes are regarded as the fully delocalized phenoxyl radical species and have a large distribution of the radical electron spin on the central Pt ion. Among the group 10 metal complexes, the Pd complexes have the most localized electronic structure.The energy of the
The one-electron oxidized complexes have different electronic structures showing at least three sorts of complexes described asM(III)-phenolate, M(II)-phenoxyl radical, and M(II)-ligand radical except the phenoxyl radical. These electronic structure differencesgive rise to the crystal structure differences, especially the position of the counterion in the proximity of the oxidation locus. The bond lengths and angles of oxidized complexes also revealthe electronic structure difference. The magnetic susceptibility, UV-vis-NIR measurements, and other physicochemical data substantiate the electronic structure difference and afford further insightsinto the novel properties, such as valence tautomerism between Cu(III)-phenolate and Cu(II)-phenoxyl radical in CH2Cl2 solution of oxidized Cu(Salcn) complex. It is now obvious that detailed electronic structures of one-electron oxidized complex should be concluded on the basis of the results of various physical measurements.
Many salen-type complexes have been reported also in organic chemistry as the catalysts for organic molecular conversion . However, the detailed reaction mechanism has not been discussed, and especially the electronic structure of the catalyst has been unclear. Information on the detailed electronic structure of the metal ion in complexes may lead to construction of more efficient catalysts and discovery of further interesting phenomena.
The author is grateful to Prof. Dr. Osamu Yamauchi, Kansai University,for helpful comments and suggestions during preparation of this manuscript.