Characterization of one-electron oxidized copper(ii)-salophen-type complexes; Effects of electronic and geometrical structures on reactivities

Article abstract of DOI:10.1039/c3dt52338a

One-electron oxidized salophen-type complexes, [Cu(salophen)]⁺ (H₂salophen = N,N′-bis(3,5-di-tert-butylsalicylidene)-1,2- diaminobenzene), and its methoxy derivatives, [Cu(MeO-salophen)]⁺ and [Cu(salophen-(MeO)₂)]⁺ (H₂MeO-salophen = N,N′-bis(3-tert-butyl-5-methoxysalicylidene)-1,2-diaminobenzene, H ₂salophen-(MeO)₂ = N,N′-bis(3,5-di-tert- butylsalicylidene)-1,2-diamino-4,5-dimethoxybenzene), have been synthesized and structurally characterized, and their reactivities have been investigated. The solid state structures of the one-electron oxidized forms of these complexes suggested that [Cu(salophen)]⁺ and [Cu(MeO-salophen)]⁺ can be assigned to relatively localized Cu(ii)-phenoxyl radical complexes, while [Cu(salophen-(MeO)₂)]⁺ is the diiminobenzene radical complex. On the other hand, [Cu(salophen)]⁺ in solution showed a different electronic structure from that of the solid sample, the radical electron being delocalized over the whole π-conjugated ligand. The reaction of these oxidized complexes with benzyl alcohol has been investigated in the presence of a large excess of substrate, which revealed the difference in the kinetic behavior between the complexes. The mechanisms of the oxidation have been discussed on the basis of the electronic and geometrical structures and the reaction kinetics.

Full text of DOI:10.1039/c3dt52338a

Dalton
Transactions
View Article Online
View Journal | View Issue
PAPER
Characterization of one-electron oxidized copper(II)-
salophen-type complexes; effects of electronic
and geometrical structures on reactivities†
Cite this: Dalton Trans., 2014, 43,
2283
Kazutaka Asami,a Akiko Takashina,b Misato Kobayashi,c Satoshi Iwatsuki,c
Tatsuo Yajima,d Amélie Kochem,e Maurice van Gastel,e Fumito Tani,f
Takamitsu Kohzuma,b Fabrice Thomasg and Yuichi Shimazaki*a
One-electron oxidized salophen-type complexes, [Cu(salophen)]+ (H2salophen = N,N’-bis(3,5-di-tert-
butylsalicylidene)-1,2-diaminobenzene), and its methoxy derivatives, [Cu(MeO-salophen)]+ and [Cu(salo-
phen-(MeO)2)]+ (H2MeO-salophen = N,N’-bis(3-tert-butyl-5-methoxysalicylidene)-1,2-diaminobenzene,
H2salophen-(MeO)2 = N,N’-bis(3,5-di-tert-butylsalicylidene)-1,2-diamino-4,5-dimethoxybenzene), have
been synthesized and structurally characterized, and their reactivities have been investigated. The solid
state structures of the one-electron oxidized forms of these complexes suggested that [Cu(salophen)]+
and [Cu(MeO-salophen)]+ can be assigned to relatively localized Cu(II)-phenoxyl radical complexes, while
[Cu(salophen-(MeO)2)]+ is the diiminobenzene radical complex. On the other hand, [Cu(salophen)]+ in
solution showed a different electronic structure from that of the solid sample, the radical electron being
delocalized over the whole π-conjugated ligand. The reaction of these oxidized complexes with benzyl
alcohol has been investigated in the presence of a large excess of substrate, which revealed the difference
in the kinetic behavior between the complexes. The mechanisms of the oxidation have been discussed
on the basis of the electronic and geometrical structures and the reaction kinetics.
Received 26th August 2013,
Accepted 7th November 2013
DOI: 10.1039/c3dt52338a
www.rsc.org/dalton
tryptophanyl radicals, are now known to form in the course of
enzymatic reactions.6,7
Introduction
The redox chemistry of metal complexes has been widely de-
Phenoxyl radical formation from tyrosyl residues has been
veloped in recent years, affording deep insight into the reac- well established in metalloenzymes such as class I ribonucleo-
tion mechanisms for many useful homogeneous catalytic tide reductase and prostaglandin endoperoxide synthase.6,7 It
reactions and enzymatic reactions at the active site of was discovered in the early 1990s that the phenoxyl radical can
metalloenzymes.1–5 It has been increasingly recognized that bind to a metal ion and act as an organic radical cofactor in
radical species formed as a result of redox reactions are par- the metalloenzyme, galactose oxidase (GO), which is a single
ticularly important in biological systems.2–5 Thus, radicals of copper oxidase catalyzing the two-electron oxidation of
amino acid residues, such as tyrosyl, cysteinyl, and primary alcohols to aldehydes (Fig. 1).8 Various
aCollege of Science, Ibaraki University, Bunkyo, Mito 310-8512, Japan.
E-mail: yshima@mx.ibaraki.ac.jp; Fax: +81-29-228-8403; Tel: +81-29-228-8374
bInstitute of Applied Beam Science, Ibaraki University, Bunkyo, Mito 310-8512, Japan
cDepartment of Chemistry, Konan University, Higashinada-ku, Kobe 658-8501, Japan
dFaculty of Chemistry, Materials and Bioengineering, Kansai University, Suita,
Osaka 564-8680, Japan
eMax Planck Institute for Chemical Energy Conversion, Stifstrasse 34-36,
D-45470 Mülheim an der Ruhr, Germany
fInstitute for Materials Chemistry and Engineering, Kyushu University, Higashi-ku,
Fukuoka 812-8581, Japan
gDépartement de Chimie Moléculaire - Chimie Inorganique Redox Biomimétique
(CIRE) - UMR CNRS 5250, Université Joseph Fourier, B. P. 53, 38041 Grenoble
cedex 9, France
†Electronic supplementary information (ESI) available: Fig. S1–S18 and Tables
S1–S5. CCDC 957252 and 957253. For ESI and crystallographic data in CIF or
other electronic format see DOI: 10.1039/c3dt52338a
Fig. 1 Proposed catalytic alcohol oxidation mechanism of GO.8–10
Dalton Trans., 2014, 43, 2283–2293 | 2283
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
Dalton Transactions
physicochemical studies revealed the presence of the phenoxyl
In order to gain further insights into the relationship
•
radical in the active form of the enzyme (Tyr272 ) and evi- between electronic structure and reactivity, we have character-
denced the formation of a Cu(II)-phenoxyl radical bond in the ized one-electron oxidized Cu(salophen) complexes (Fig. 2).
active site.8–10 In the generally accepted catalytic pathway the The salophen-type ligands consist of two salicylideneimino
primary alcohol binds to the fourth equatorial position of the moieties linked by an o-phenylene group. Due to the presence
Cu(II) ion, which brings it close to the phenoxyl radical of a redox active o-phenylenediamine moiety, the salophen
•
residue. Subsequent hydrogen atom abstraction by Tyr272 and complexes may show various different electronic structures
electron transfer to the metal leads to the formation of the upon oxidation, such as the phenoxyl radical, the radical de-
aldehyde via a ketyl radical intermediate and the reduced form localized over the whole ligand and the o-diiminobenzene
of GO. With the aim of understanding the detailed catalytic radical.25 We have already reported that such an electronic
mechanism of GO and the fascinating properties of metal structure difference depends on the position of the methoxy-
complexes involving the coordinated phenoxyl radical, many substitution; oxidation of Cu(MeO-salophen) (2) having the
metal–phenolate complexes have been synthesized and p-methoxyphenolate moiety gave the methoxyphenoxyl radical
characterized.11,12
species while oxidation of the 4,5-dimethoxyphenylene salo-
Depending on the relative energies of the redox-active orbi- phen complex, Cu(salophen–(MeO)2) (3), gave the o-diimino-
tals, a metal complex with a non-innocent ligand exists in two benzene radical species.25 In this article, we report detailed
limiting descriptions, either a metal–ligand radical (Mn+(L•)) or characterization of the three one-electron oxidized Cu(II) salo-
a high valent metal (M(n+1)+(L−)) complex.13–23 Oxidized metal phen-type complexes and elucidation of the difference in the
salen-type complexes are known to exist in either form, and reactivity with benzyl alcohol between these complexes.
the factors that control the locus of oxidation in these com-
plexes are being actively pursued. Previously, we have reported
characterization and reactivity of the one-electron oxidized
Cu(salen)-type complexes (Fig. 2).21,22 The structure of the 2N-
donor chelate backbone of the salen-type ligands was effective
Experimental section
Materials and methods
in causing the electronic structure difference between the one-
electron oxidized complexes.23 Oxidation of Cu(II)-salen with a All the chemicals used were of the highest grade available and
five-membered chelate ring afforded the Cu(III)-phenolate were further purified whenever necessary.26 Solvents were also
ground state, which was shown to be in valence tautomerism purified before use by standard methods.26 All salophen and
with the Cu(II)-phenoxyl radical in CH2Cl2 solution at room salcn ligands were prepared in a manner similar to that pre-
temperature.21 On the other hand, the six-membered 2N- viously reported.25 UV-vis-NIR spectra were obtained using a
donor chelate complexes could be assigned to the Cu(II)-phen- JASCO V-670 spectrophotometer, equipped with a liquid nitro-
oxyl radical complex, which was maintained in the CH2Cl2 gen cooled cryostat, UNISOKU CoolSpeK UV USP-203-A, from
solution at room temperature.22 These two complexes showed −60 °C to 20 °C. Reflectance spectra were also obtained using
an oxidative reactivity toward benzyl alcohol to give benz- a JASCO V-670 spectrophotometer, equipped with an ISN-723
aldehyde.22,24 The reaction mechanism of these complexes integrating spheres unit. Cyclic voltammetry (CV) was per-
exhibited a similar rate determining step, which was assigned formed on a HOKUTO DENKO HZ-5000 automatic polarization
to the hydrogen abstraction of the α-CH2 position of benzyl system, equipped with a Ag wire reference electrode, a Pt disk
alcohol. However, the Cu(III)-ground state complex showed a working electrode, and a Pt counter electrode in 0.1 M
ca. 4-times lower reaction rate, mainly due to the equilibrium Bu4NClO4 solutions in CH2Cl2. Ferrocene was used as an exter-
between Cu(III)-phenolate and the Cu(II)-phenoxyl radical.22 nal standard. All EPR spectra were collected using a Bruker
Thus, the electronic structure difference leads to a different EMX plus spectrometer operating with an X-band microwave
reactivity of the complexes having a similar redox potential.
bridge and a high sensitivity cavity (∼9.4 GHz). EPR spectra
were simulated with a commercial Bruker XSophe software.
The X-ray photoelectron spectra (XPS) were recorded on a
Thermo VG Scientific Model Theta Probe spectrometer using
Al Kα radiation (1486.6 eV) operated at 15 kV and 3 mA as the
X-ray excitation source. The solid sample was dispersed on an
In film, and the carbon 1s binding energy (284.5 eV) was
used for calibration of the binding energy.27 Resonance
Raman spectra of CH2Cl2 solution samples were measured
with a JASCO NR-1800 triple polychromator equipped with a
liquid-nitrogen-cooled Princeton Instruments CCD detector.
Resonance Raman spectra of solid samples were excited by
the 457.9 nm Ar+ laser (Coherent I-301C). The laser power
at the sample point was adjusted to 5 mW. The Raman scatter-
ing light was collected by an objective lens and detected by
Fig. 2 Structures of some salen- and salophen-type complexes.
2284 | Dalton Trans., 2014, 43, 2283–2293
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
an LN-400B liquid nitrogen cooled CCD detector (Roper Scien- following the absorbance change with time at 582 nm for the
tific). Raman shifts were calibrated with indene and toluene, reaction of [1]+ and [3]+ and at 800 nm for [2]+, at a large
the accuracy of the peak positions of the Raman bands being excess concentration of benzyl alcohol over the concentration
1 cm−1
.
of oxidized complexes. The sample solutions for the kinetic
Synthesis of complexes. [Cu(salophen)]SbF6 ([1]SbF6). To a measurements were prepared by the following procedures:22
A
solution of Cu(salophen) (1) (0.060 g, 0.1 mmol) in CH2Cl2 CH2Cl2 solution of the oxidized complexes was prepared by fil-
(5 mL) was added AgSbF6 (0.034 g, 0.1 mmol), and after a few tering the equimolar mixture of the Cu(II) complexes and
minutes the reaction mixture was filtered. The filtrate was AgSbF6 in CH2Cl2. After the 0.2 or 0.4 mM solution of oxidized
evaporated in vacuo, and then 5 mL of toluene was added to complexes in CH2Cl2 in a 10 mm quartz cell was adjusted to
the residue, which was kept standing for a few hours below the appropriate reaction temperature, the reaction was
−10 °C to afford brown crystals. Elemental analysis (%) calcd initiated by adding the calculated amount of benzyl alcohol
for [1]SbF6·2(CH2Cl2) (C36H46N2O2CuSbF6·2(CH2Cl2)): C, 45.28; directly into the cell. The kinetic measurements for the reac-
H, 5.00; N, 2.78. Found: C, 45.62; H, 5.07; N, 2.94.
tion with α-benzyl-d2 alcohol and benzyl alcohol-d were carried
[Cu(MeO-salophen)]SbF6 ([2]SbF6). This complex was syn- out using freshly purified alcohol, since the reactions are
thesized similarly according to the procedure for [1]SbF6. highly sensitive to impurities of the deuterated alcohol as
Elemental analysis (%) calcd for [2]SbF6·0.5(CH2Cl2) reported in ref. 22 and 24.
(C30H34N2O4CuSbF6·0.5(CH2Cl2)): C, 44.22; H, 4.26; N, 3.38.
Found: C, 44.39; H, 4.40; N, 3.43.
Computational details
X-ray structure determination
All calculations were performed with the ORCA program
package.30 Full geometry optimizations were carried out for all
complexes using the GGA functional BP8631–33 in combination
with the Def2-TZV/P34 basis set for all atoms and by taking
advantage of the resolution of the identity (RI) approximation
in the Split-RI-J variant35 with the appropriate Coulomb fitting
sets.36 Increased integration grids (Grid4 in ORCA convention)
and tight SCF convergence criteria were used. The relative
energies were obtained from single-point calculations using
the B3LYP37,38 functional together with the Def2-TZV/P34 basis
set. They were computed from the gas-phase optimized struc-
tures as a sum of the electronic energy, thermal corrections to
the free energy, and free energy of solvation. Tight SCF conver-
gence criteria were used in these calculations as well.
The X-ray experiment was carried out on a well-shaped single
crystal of complex [2]SbF6 on a Rigaku R-AXIS imaging plate
area detector with graphite monochromated Mo Kα radiation
(λ = 0.71073 Å). The crystal was mounted on a glass fibre. In
order to determine the cell constants and orientation matrix,
3 oscillation photographs were taken for each frame with the
oscillation angle of 5° and the exposure time of 60 s.
Reflection data were corrected for both Lorentz and polari-
zation effects. The structure was solved by the heavy-atom
method and refined anisotropically for non-hydrogen atoms
by full-matrix least-squares calculations. All the disordered
carbons were also refined anisotropically. Each refinement was
continued until all shifts were smaller than one third of the
standard deviations of the parameters involved. Atomic scatter-
ing factors and anomalous dispersion terms were taken from
the literature.28 All hydrogen atoms were located at the calcu-
lated positions and were assigned a fixed displacement and
constrained to an ideal geometry with C–H = 0.95 Å. The
Results
Syntheses and crystal structures of Cu(salophen) complexes
thermal parameters of calculated hydrogen atoms were related The N2O2-donor salophen-type ligands with a bridging phenyl
to those of their parent atoms by U(H) = 1.2Ueq(C). All the cal- ring reacted with copper(II) acetate in C2H5OH to give com-
culations were performed by using the Crystal Structure plexes 1, 2, and 3 as brown crystals. The cyclic voltammograms
program package.29 Summaries of the fundamental crystal of all of the salophen complexes showed two reversible redox
data and experimental parameters for structure determination waves at E1/2a = 0.56 and E1/2b = 0.73 V for 1, E1/2a = 0.37 and
are given in Table S1.†
E1/2b = 0.52 V for 2, and E1/2a = 0.42 and E1/2b = 0.71 V for 3 vs.
Fc/Fc+ in the range of 0–1.0 V.39 The first potential of complex
1 is the highest in the series of salophen complexes, and the
Kinetic measurements
Kinetic measurements for the reaction of oxidized complexes potential decreased in the order 1 > 3 > 2. In comparison to
with benzyl alcohol were carried out using a Shimadzu UV2550 the Cu(1,2-salcn) and Cu(1,3-salcn) complexes, the potential
or a JASCO V-660 or V-670 spectrophotometer. The tempera- differences are within ca. 0.1 V (E1/2a = 0.45 and 0.48 V, respect-
ture of the sample solutions was held constant within 0.1 °C ively). These redox waves were found to be one electron pro-
using a thermostated cell unit. All the sample solutions for cesses from the current analysis for each wave. Based on these
UV-vis measurements were prepared using purified CH2Cl2. potentials, the one-electron oxidized complexes can be gener-
Anhydrous benzyl alcohol (99.8%, Aldrich) was used without ated by the addition of an equimolar amount of chemical oxi-
further purification. Preparations of α-d2-benzyl alcohol dants such as AgSbF6. Accordingly, the addition of AgSbF6 to
(C6H5CD2OH) and benzyl alcohol-d (C6H5CH2OD) were these complexes in CH2Cl2 caused a color change to dark
reported previously.22 Rate constants were measured by brown, indicating the formation of one-electron oxidized
This journal is © The Royal Society of Chemistry 2014
Dalton Trans., 2014, 43, 2283–2293 | 2285

View Article Online
Paper
Dalton Transactions
forms, which can be described as a monocationic species, [1]+, be isolated, and the structure was determined successfully.
[2]+, and [3]+ (vide infra).25,39
[1]SbF6 was shown to be a typical localized phenoxyl radical
The solutions of these oxidized complexes kept standing at complex on one side of the phenolate moieties similar to
low temperature gave crystals of [1]SbF6, [2]SbF6, and [3]SbF6, [2]SbF6, while it has a large distortion of one of the phenolate
respectively. X-ray photoelectron spectroscopy (XPS) of the crys- moieties and the coordination plane (see Fig. S1†).
tals exhibited the same binding energies of the copper atom as
Solution chemistry of the one-electron oxidized Cu(salophen)
complexes
those of the neutral complexes, indicating that the oxidized
complexes have the same oxidation state of the central copper
ion. Such characteristics were also observed in the other phen-
oxyl radical species,16b,22,23 and these complexes can therefore
be assigned to the metal(II)-radical complexes in the solid
state.
Although oxidized complexes [2]SbF6 and [3]SbF6 did not show
significant color change upon dissolution in CH2Cl2, the
CH2Cl2 solution of [1]SbF6 gave a slightly different color from
the pale-greenish brown color of the solid state.39 The UV-vis-
NIR spectrum of [1]+ in CH2Cl2 at 273 K exhibited some
absorption bands at 5300 (ε = 2900 M−1 cm−1), 10 400 (1500),
17 200 (4000), 22 650 (30 000), and 24 500 cm−1 (20 000), while
the reflectance spectrum of the solid state complex [1]SbF6
showed different absorption features, especially the intensity
of the NIR and the 17 200 cm−1 band (Fig. 4). The decrease of
the absorption intensity at 17 200 cm−1 in CH2Cl2 solution can
be considered to be due to regeneration of the π-conjugation
of the ligand and as a result delocalization of the radical elec-
tron on the ligand as a whole. Indeed, a temperature-depen-
dent UV-vis-NIR absorption spectrum of complex [1]SbF6 was
observed; the intensities of the absorption bands in CH2Cl2
solution increased with decreasing temperature, and at the
same time the 10 400 cm−1 band at 273 K was slightly blue-
shifted (Fig. 4, inset). The spectrum was close to the reflec-
tance spectrum of the solid sample. The temperature-depen-
dent spectral change was reversible at least three times. A
similar spectral feature was also observed for the Cu(III)-
phenolate ground state complex, [Cu(1,2-salcn)]+, which was
assigned to a ca. 1 : 1 mixture of Cu(II)-phenoxyl radical and
Cu(III)-phenolate in CH2Cl2 at room temperature.21 However,
[1]SbF6 was previously revealed to have a magnetic interaction
between the copper and radical electron spins in CH2Cl2 from
The X-ray molecular structures of these oxidized complexes
revealed that they consist of the monocation of the salophen
complexes and SbF6− as a counter anion (Fig. 3). Previously we
reported the X-ray structure of [3]SbF6, which showed that the
two C–N bonds in the o-phenylenediamine moiety were longer
in comparison to those of the neutral complex and that the
SbF6− counter anion was located in the proximity of the o-phe-
nylenediamine moiety.25 These features suggested that the
oxidized complex [3]SbF6 can be assigned to an o-diiminoben-
zene radical Cu(II) complex. The structure of the oxidized
complex [2]SbF6 is different from [3]SbF6, and the C–N bond
elongation of the o-diaminophenyl moiety observed in [3]SbF6
could not be detected. The Cu coordination sphere geometry
in [2]SbF6 is asymmetric, indicating that one salicylideneimino
moiety becomes a weaker donor with a longer bond to
Cu (Cu–O(1) 1.962(7) vs. Cu–O(2) 1.879(6) Å). In addition, the
−
counter anion SbF6 is located at the proximal position of the
phenoxyl radical site with the closest distance of 3.316(11) Å.
This situation differs from that encountered in [3]SbF6, where
the counter anion contacts with the o-diiminobenzene radical
site.25 These results are consistent with radical localization on
one side of the molecule in [2]SbF6, which is also in agreement
with the bonding pattern observed for other localized phenoxyl
radical complexes.40 In comparison with the phenoxyl radical
complex [Cu(1,3-salcn)]SbF6, the angle between the two
phenolate moieties in [2]SbF6 (7.5(1)°) is smaller than that of
[Cu(1,3-salcn)]SbF6 (51.8(2)°).22 The smaller distortion of
[2]SbF6 may suggest that this complex has a rigid framework of
the ligand. The geometrical structure of complex [1]SbF6 could
not be determined by X-ray analysis, since the crystals were
rapidly broken outside the mother liquid. However, crystals
of the 2 : 1 mixture of [1]SbF6 with the neutral complex 1 could
1
EPR studies,39 and the H-NMR spectrum of [1]SbF6 in CD2Cl2
at room temperature exhibits a paramagnetic behavior. These
results suggested that the CH2Cl2 solution of [1]+ could not be
Fig. 4 UV-vis-NIR absorption spectra and reflectance spectrum of
complex [1]SbF6: (A) UV-vis absorption spectrum in CH2Cl2 at 293 K;
inset, temperature-dependent spectral changes in CH2Cl2 and (B) reflec-
tance spectrum of the solid sample. The reflectance spectra of all the
oxidized complexes in the NIR region showed a very broad band, which
may be due to an intermolecular CT (see ESI and Fig. S2–S4†).
Fig. 3 Crystal structure of complex [2]SbF6.
2286 | Dalton Trans., 2014, 43, 2283–2293
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
assigned to the Cu(III)-phenolate. On the other hand, the NIR semiquinone radical, respectively. Therefore, the oxidized com-
band at 5300 cm−1 of the CH2Cl2 solution of [1]+ showed a plexes [2]SbF6 and [3]SbF6 can be assigned to a Cu(II)-phenoxyl
small intensity and a large band width (ε = 2900 M−1 cm−1 and and a Cu(II)–o-diiminobenzene radical, respectively.25 The
Δν1/2 = 3100 cm−1) in comparison to some other group results further indicate that the methoxy substitution of salo-
10 metal salen-type complexes,17,23 which could be assigned to phen-type complexes can localize the radical electron on the
the class III delocalized species based on the Robin and Day substituted moiety.
classification.41–45 These results suggest that [1]SbF6 in CH2Cl2
can be assigned to a ligand radical delocalized over the whole
DFT calculations of complex [1]+
Previously we have reported DFT calculations of complexes [2]+
aromatic ligand framework, rather than a simple class III
mixed-valent phenoxyl radical delocalization system only
through the central metal ion. Thus, [1]SbF6 is a localized
and [3]+.25 In order to complete the series, we performed DFT
calculation on complex [1]+. Geometry optimizations were per-
phenoxyl radical species at the ground state, which is held in
equilibrium with the ligand delocalized radical species in
CH2Cl2 solution at around room temperature. Such a tempera-
ture dependent change and spectral difference between the
solid and solution states could not be detected in complexes
[2]SbF6 and [3]SbF6 (Figs. S3 and S4†).
formed at the BP86 level of the theory for the neutral com-
pound 1 and the corresponding cation [1]+ (Fig. S5–S9†). In 1,
the coordination bond distances calculated by geometry
optimization match those of the X-ray structures within 0.02
and 0.04 Å for the Cu–N and Cu–O bonds, respectively, while
the bond distances within the organic part are well repro-
duced.20 This slight over-estimation of the coordination bond
distances is not surprising and is typical of current DFT func-
tionals.47 For the corresponding cation [1]+, geometry optimi-
zations were performed with and without the counter ion
SbF6−. In both cases, the optimized geometries for the singlet
state predict symmetric structures with shortened metal–
ligand bonds as compared to 1 (the Cu–O and Cu–N bonds are
shortened by ca. 0.04 Å). In addition, the bond lengths within
the phenolic rings are not significantly affected by oxidation
(Table 1). Thus, the geometric parameters calculated for the
singlet states of [1]+ and [1]SbF6 markedly differ from the
experimentally observed bond lengths and asymmetric coordi-
nation sphere. Moreover, the triplet state is found to be
favored in energy by 6 and 7 kcal mol−1 for [1]SbF6 and [1]+,
respectively, over the singlet state.48
The optimized geometry of [1]SbF6 in the triplet state repro-
duces the asymmetry of the complex (Fig. 6 and Table 1). One
Cu–O bond is significantly elongated (+0.048 Å) in comparison
to 1, while the other is only moderately elongated (+0.017 Å).
Detailed investigation of the C–C and C–O bond lengths reveals
a quinoid structure for the ring containing the O atom of the
elongated Cu–O bond. The C–O bond length is shortened by
0.021 Å as compared to 1, consistent with a significant phen-
oxyl radical character of this ring. Interestingly, the geometric
Resonance Raman spectra of [1]SbF6 in CH2Cl2 and in the
solid state supported the radical delocalization over the ligand
as a whole (Fig. 5). The Raman spectra of [1]SbF6 in CH2Cl2 at
413, 441 and 458 nm excitations showed no significant change
from those of the complex before oxidation, and no typical
phenoxyl radical Raman band was observed around
1500 cm−1 46
However, the solid sample of [1]SbF6 exhibited
.
the typical phenoxyl radical ν7a band at 1508 cm−1, clearly indi-
cating that [1]SbF6 in the solid state can be assigned to a phen-
oxyl radical. Such features of the Raman spectra suggest that
perturbation of the ligand by one-electron oxidation is rather
small in comparison to the other phenoxyl radical species.
Therefore, [1]SbF6 can be assigned to the delocalized ligand
radical species in CH2Cl2 solution.
The Raman spectra of the oxidized complexes [2]SbF6 and
[3]SbF6 showed a significant change from those of the com-
plexes before oxidation.25 The characteristic phenoxyl radical
Raman bands were observed at 1491 and 1608 cm−1 for
[2]SbF6, corresponding to the phenoxyl radical ν7a and ν8a
bands, respectively. The oxidized complex [3]SbF6 showed
sharp intense bands at 1373 and 1542 cm−1, which are very
similar to the C–O and C–C stretching vibrations of the
Table 1 Calculated first coordination sphere bond lengths (Å) for 1, [1]+
a
and [1]SbF6
Bond
1
[1]+
Triplet
[1]SbF6
Triplet
[1]+
Singlet
[1]SbF6
Singlet
Doublet
Cu–N1
Cu–N2
Cu–O1
Cu–O2
1.959
1.958
1.937
1.939
1.950
1.954
1.935
1.935
1.968
1.969
1.985
1.956
1.924
1.928
1.898
1.899
1.916
1.915
1.903
1.898
a Atom numbering used:
Fig. 5 Resonance Raman spectra of 1 and [1]+: (A) spectra of solid
sample of 1 (black) and [1]SbF6 (red), inset, differential spectrum from [1]
SbF6 to 1 and (B) spectra in CH2Cl2 solution of 1 (black) and [1]SbF6 (red)
at room temperature. The asterisks in (A) denote the reflectance peak of
the Raman cell.
This journal is © The Royal Society of Chemistry 2014
Dalton Trans., 2014, 43, 2283–2293 | 2287

View Article Online
Paper
Dalton Transactions
a total spin S = 1. Broken symmetry calculations indeed show
that the triplet state is favoured by 7.5 and 1.5 kcal mol−1 over
the singlet for [1]SbF6 and [1]+, respectively.48 The ZFS para-
meters are expected to be larger than 0.3 cm−1 in such radical
complexes,20 which makes it difficult to detect them experi-
mentally by X-band EPR.
Reactivity of one-electron oxidized Cu(salophen) complexes
with benzyl alcohol
The oxidized salophen complexes, [1]+, [2]+ and [3]+, were
stable in CH2Cl2 solution at 293 K. They gradually decayed
approximately in a first-order process, whose decay constants
were estimated to be 5.3 × 10−5 s−1 for [1]+, 6.6 × 10−6 s−1 for
[2]+ and 4.0 × 10−5 s−1 for [3]+ at 293 K, with the half-lives of
3.6 h, 29 h and 4.8 h for [1]+, [2]+ and [3]+, respectively.
However, the decay processes of these oxidized complexes were
different. The oxidized complex [1]+ decayed in a first-order
process simply to give the neutral complex 1, while the
methoxy-substituted oxidized complexes [2]+ and [3]+ decom-
posed into insoluble products. The addition of benzyl alcohol
to a solution of the oxidized salophen-type complexes in
CH2Cl2 caused a simple color change, but no insoluble
decomposition products were detected for [2]+ and [3]+. GC-MS
analyses showed that benzaldehyde was formed in 50% yield
relative to the initial amount of the oxidized CuII complex in
regard to all salophen-type complexes. This reaction was also
observed under a nitrogen atmosphere, indicating that the
presence of air did not affect the benzyl alcohol oxidation. The
UV-vis absorption spectral change revealed that the complexes
were converted to the neutral complexes, 1–3. Therefore, the
overall stoichiometry of each reaction observed is expressed as
follows:
Fig. 6 Localized SOMOs for (A) [1]SbF6 and (B) [1]+ in the triplet state.
parameters of the opposite ring are intermediate between
those of the phenoxyl moiety of [1]SbF6 and the phenolate
moieties of 1. The calculated angle between the plane defined
by the coordinated 2N2O atoms and the phenoxyl ring is 22°,
which deviates significantly from zero, but is found to be
slightly under-estimated from the experimentally observed
value (25°). The localization of the radical in [1]SbF6 is likely
due to the close contacts that exist between the negatively
charged counter ion and the electron-deficient radical ring. In
order to confirm this assumption, we have run a geometry
optimization of [1]+ in the triplet state in the absence of the
counter ion. Calculations predict a symmetric structure with
similar Cu–O bond lengths and do not evidence significant
changes in the metal–ligand bond distances in comparison
with 1. These features are typical of a phenoxyl radical deloca-
lized over the two rings but not of the localized radical
observed experimentally in the solid state. Rather, they are
representative of a delocalized phenoxyl radical form, which is
2½CuðSalophen•Þꢀþ þ C6H5CH2OH
! 2CuðSalophenÞ þ C6H5CHO þ 2Hþ
ð1Þ
The reaction stoichiometries of all of the oxidized com-
in good agreement with the spectra of [1]+ in CH2Cl2 solution. plexes with benzyl alcohol are in good agreement with those of
The localized versus delocalized phenoxyl radical character oxidized Cu(II) salen-type complexes reported previously, while
may thus be interpreted by differences in association between the resulting salophen-type complexes in the reaction are
[1]+ and the SbF6 counter ion, which result from changes in different from those found for the reactions of the oxidized
−
the experimental conditions, in particular the phase and Cu-salcn-type complexes, which were converted to the proton
temperature.
adducts.22,24
In both cases of [1]+ and [1]SbF6 (Fig. 6), it is clear that one
In the presence of a large excess of benzyl alcohol, conver-
SOMO (SOMO 1 in Fig. 6) has a main metal dx2−y2 character sion of [1]+ to 1 can be determined as a first-order process
contribution of 64% and 62% for [1]SbF6 and [1]+, respectively, (Fig. S10†); the rate law is expressed by −d[[1]+]/dt
=
with small contributions of the coordinating atoms of the k1osbts[[1]+], where k
is the conditional pseudo first-order
1st
obs
ligand due to the covalency of the metal–ligand bonds. The rate constant. This suggests that only one molecule of complex
1st
other SOMO (SOMO 2 in Fig. 6) is mainly localized (95%) on [1]+ is involved in the rate-determining step. The k
value
obs
−
the aromatic ring close to SbF6 in [1]SbF6, while it is equally was not changed by varying the concentration of the complex
shared between the phenol rings in the case of [1]+. Thus, (k
= (1.10–1.19) × 10−3 s−1 for 0.2 and 0.4 mM solutions of
1st
obs
from this analysis it is evident that the cation has a Cu(II)- [1]+, respectively, at [C6H5CH2OH] = 1.0 M and 293 K,
phenoxyl radical character. It has to be emphasized that in Table S2†), while it depended on the benzyl alcohol concen-
1st
both optimized structures the dx2−y2 and the π-radical orbitals tration. The plot of k
against the concentration of benzyl
obs
1st
are essentially orthogonal. The magnetic coupling is ferro- alcohol exhibited a linear correlation (Fig. S11†), i.e., k
=
obs
magnetic and the ground spin state of the radical complex has k2[C6H5CH2OH], giving the second-order rate constant
2288 | Dalton Trans., 2014, 43, 2283–2293
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
k2 = (1.21 0.02) × 10−3 M−1 s−1 at 293 K for the rate-determin- [C6H5CH2OH] = 1.0 M. The Eyring plots of complex [1]+
ing process. We could not determine the association constant showed a linear fit (Fig. S16†), and the activation parameters
of complex [1]+ with benzyl alcohol, due to the linear corre- of [1]+ were estimated to be ΔH‡ = 10.0 kcal mol−1 and ΔS‡ =
lation. In contrast to [1]+, the reaction of [2]+ with excess of −37.8 cal mol−1 K−1, which are very similar to those of oxi-
benzyl alcohol was regarded as
a second-order process dized salcn complexes reported previously (Table S5†). The
(Fig. S12†); the rate law was expressed by –d[[2]+]/dt = temperature dependence of the reaction rate for the p-methoxy-
k2onbsd[[2]+]2, where k
is the conditional pseudo second- phenoxyl radical complex [2]+ was rather small in comparison
2nd
obs
order rate constant, suggesting that two molecules of complex to complex [1]+, and the activation parameters of complex [2]+
[2]+ are involved in the rate-determining step. The reaction of were estimated to be ΔH‡ = 5.8 kcal mol−1 and ΔS‡ = −39.2 cal
complex [2]+ showed also the benzyl alcohol concentration mol−1 K−1. The significantly low enthalpy and negatively high
dependence, and the conditional second-order constant of this entropy are in good agreement with those of the other electron
reaction exhibited a linear relationship with the benzyl alcohol transfer mechanisms49 and suggest a different reaction mech-
2nd
obs
concentration (Fig. S13†), i.e., k
= k3[C6H5CH2OH], giving anism from that of complex [1]+. The activation parameters of
the third-order rate constant k3 = 0.84 0.03 M−2 s−1 at 293 K. complex [3]+, estimated to be ΔH‡ = 16.2 kcal mol−1 and ΔS‡ =
On the other hand, complex [3]+ showed a pseudo first-order −20.6 cal mol−1 K−1, were different from those of complex [1]+,
process for the aldehyde formation, followed by side decompo- indicating that complex [3]+ exhibited a different reaction
sition reactions (detailed descriptions on the kinetic analyses mechanism in comparison to the other salophen-type
to determine the rate constants for the aldehyde formation are complexes.
provided in ESI and Fig. S14†). The plot of the conditional
first-order rate constant (k1osbts) against the benzyl alcohol
concentration revealed that benzyl alcohol oxidation proceeds
Discussion
1st
in a substrate saturation system (Fig. S15†), i.e., k
=
obs
Kk1[C6H5CH2OH]/(1 + K[C6H5CH2OH]), so that the rate con- Metal complexes of tetradentate N2O2 Schiff-base ligands,
stant (k1) and association constant (K) were estimated to be salen and its derivatives, have been extensively studied as cata-
k1 = (1.9 0.1) × 10−4 s−1 and K = 3.7 0.7 M−1 at 293 K. Thus, lysts for small molecule transformations. Recently we reported
the reaction mechanism and reaction rate are different a relationship between the size of the chelate ring involving
between the salophen-type complexes.
the two imino nitrogens and the reactivity towards benzyl
In order to determine whether C–H bond cleavage is alcohol oxidation. The complex involving a 6-membered
involved in the rate determining step, we carried out kinetic chelate ring showed a reaction rate ca. 4-times faster than the
analyses using α-benzyl-d2 alcohol (1.0 M). The KIE value (= k(H)/ analog complex involving a 5-membered chelate ring.22 On the
k(D)) of complex [1]+ was estimated to be 13 at 293 K other hand, the reaction mechanisms of these two complexes
(Table S2†), which indicates that the C–H bond scission is were found to be very similar; the rate determining step is
included in the rate-determining step.22,24 As for complex [2]+, hydrogen abstraction from the α-position of the benzyl
the rate constant did not significantly change upon using alcohol, and the activation parameters are also similar. From
α-benzyl-d2 alcohol (KIE = 1, Table S3†). This result suggests these results, the difference in chelate ring size is mainly
that the C–H bond activation at the α-position of benzyl responsible for the reaction rate, the larger chelate ring being
alcohol was not included at least in the rate determining step more effective for the oxidation of the substrate. The main
of complex [2]+, while in the case of complex [3]+ a small rate reason for the difference in reaction rate is a change in the
constant was observed, giving the KIE value of 1.5 at 293 K electronic structure of the active species; indeed, the oxidized
(Table S4†). On the other hand, the reaction rates of 5-membered chelate complex exhibits a Cu(III)-phenolate in
C6H5CH2OD for all the complexes were very similar to those of the solid state, while it exists both in the Cu(III)-phenolate and
C6H5CH2OH (KIE = 1.0–1.1, Tables S2–S4†), indicating that the Cu(II)-phenoxyl radical states in CH2Cl2 solution at room temp-
benzyl alcohol oxidation does not involve the O–H hydrogen erature. In contrast, the oxidized 6-membered chelate complex
abstraction in the rate determining step. Thus, a hydroxyalkyl is a relatively localized Cu(II)-phenoxyl radical, and this elec-
radical such as RC•HOH is generated as the reaction inter- tronic structure is maintained both in the solid state and in
mediate in the case of complex [1]+, while complex [2]+ acts CH2Cl2 solution at room temperature. The decreased reaction
only as an electron transfer reagent without any hydrogen rate in the case of the oxidized 5-membered chelate ring
abstraction. Therefore, the substrate may be two-electron oxi- complex can therefore be assigned to the existence of the equi-
dized by two equivalents of [2]+, because second-order reac- librium between the Cu(III)-phenolate and Cu(II)-phenoxyl
tions with respect to [2]+ were observed under the present radical states under the reaction conditions, the Cu(III)-phenol-
conditions as mentioned above. The KIE value of 1.5 for ate state being less favorable for hydrogen abstraction from
complex [3]+ obeying pseudo first-order kinetics indicates a benzyl alcohol in comparison to the Cu(II)-phenoxyl radical
different mechanism from that for complexes [1]+ and [2]+.
state. Therefore, the reaction mechanism should be similar,
As a further step toward a detailed mechanism of benzyl since the “real active metal complex” in the reaction is the
alcohol oxidation by oxidized Cu(II)-salophen-type complexes, Cu(II)-phenoxyl radical, not the Cu(III)-phenolate. From these
temperature dependence studies were carried out at results and discussions, it is clear that the electronic structure
This journal is © The Royal Society of Chemistry 2014
Dalton Trans., 2014, 43, 2283–2293 | 2289

View Article Online
Paper
Dalton Transactions
of the intermediates and active species is very important for [3]+, indicating that their electronic structures are maintained
understanding the nature of the reaction and further develop- in solution at room temperature. Thus, methoxy-substitution
ment of better catalysts.22
led to the localization of the oxidation locus on the substituted
In a preliminary report, we reported that the geometrical moiety (see below). The electronic structure of [2]+ was
and electronic structures of the oxidized salophen-type com- assigned to a relatively localized phenoxyl radical state, i.e. the
plexes strongly depend on the position of the methoxy substi- CuII(phenolate)(phenoxyl radical) form on the basis of X-ray
tuent.25 Complex [2]+, containing two p-methoxyphenolate diffraction. In addition, the UV-Vis-NIR spectrum of complex
moieties, was assumed to have a phenoxyl character by spec- [2]+ did not exhibit significant changes from the reflectance
troscopy. We herein unequivocally established its phenoxyl spectrum, which indicates that the electronic structure is iden-
radical character by X-ray diffraction, and further demon- tical both in the solid state and in CH2Cl2 solution. The weak
strated that the unpaired electron was localized on a single and broad band in the NIR region in the solution spectrum of
phenoxyl moiety. Complex [3]+, having a 4,5-dimethoxy-1,2-di- [2]+ suggests that this complex belongs to the Class II mixed
aminobenzene bridging moiety, was assigned to a diimino- valence compound according to the Robin and Day classifi-
benzene π-radical complex. On the other hand, complex [1]+, cation. In other terms it is a relatively localized CuII(phen-
without methoxy substituents, was ascribed to a Cu(II)-phen- olate)(phenoxyl radical) complex in CH2Cl2 solution similar to
oxyl radical, especially on the basis of EPR spectroscopy.39 the oxidized Cu(1,3-salcn) complex. Thus, these three salo-
This indirect evidence for phenoxyl radical formation was sub- phen-type complexes exhibit different electronic structures: a
stantiated by the present study. Information from a detailed relatively localized CuII(phenolate)(phenoxyl radical) in equili-
investigation of the electronic structure of [1]+ in CH2Cl2 solu- brium with a delocalized radical form in CH2Cl2 at room temp-
tion is provided in this study with the help of reflectance, Res- erature for complex [1]+, a relatively localized CuII(phenolate)
onance Raman and UV-vis-NIR spectroscopies. In particular (phenoxyl radical) complex independent of the temperature
we evidenced the temperature-dependent electronic structure in solution for complex [2]+, and o-diiminobenzene radical
for the radical complex [1]+. While the reflectance spectrum of for [3]+.
complex [2]+ in the solid state was very similar to the absorp-
Although the reactions of all these oxidized complexes with
tion spectrum in CH2Cl2 except the intensity of the NIR band benzyl alcohol gave the same yield in benzaldehyde, ca. 50%,
that may include the intermolecular charge transfer, complex and showed no significant change of the reaction rates using
[1]+ showed different features in the solid state in comparison C6H5CH2OD, the reactivities of these complexes showed a dra-
to those in solution: the intensities of the absorption bands matic difference in the reaction mechanism. Previously we
below 17 000 cm−1 for the CH2Cl2 solution of [1]+ decreased reported a difference in the reaction rate between the CuIII-
significantly in comparison to the bands in the solid state. phenolate and CuII-phenoxyl radical forms of copper-salcn
Further, the intensities of the absorption bands in solution complexes.22 The difference in reaction rate mainly arises from
depended on the temperature, increasing with the temperature the existence of a tautomeric equilibrium between the CuIII-
decrease. Such a temperature dependent change in solution salcn phenolate and CuII-salcn radical forms.21,22 Complex [1]+
was not detected for complexes [2]+ and [3]+. The resonance showed a simple bimolecular reaction process without sub-
Raman spectra of complexes [2]+ and [3]+ exhibited the charac- strate saturation, which is quite similar to that observed for
teristic bands for each radical moiety, phenoxyl radical bands the CuIII-salcn complex. The kinetic data of [1]+ were fitted
for [2]+ and o-diiminobenzene radical bands for [3]+. Complex accordingly, giving the activation parameters and KIE values
[1]+ in CH2Cl2 showed no significant change from neutral (using α-d2-benzyl alcohol) very similar to those of the oxidized
complex 1. These results suggest that the electronic structural CuII-salcn complex.22 These results clearly indicate that the
changes from 1 to [1]+ in CH2Cl2 are rather small in compari- reaction mechanisms are similar for both [1]+ and the pre-
son to the other salophen-type complexes. In addition, the NIR viously reported one-electron oxidized Cu(1,2-salcn) complex;
band of [1]+ is of very small intensity as compared with the while the active species was the CuII-phenoxyl radical complex
other Class III delocalization systems,42–45 such as Ni and Pt for both complexes, the fact that the oxidized Cu(1,2-salcn)
salen-type complexes.17,23 These results could be rationalized was in thermal equilibrium with its CuIII-phenolate redox
by considering that complex [1]+ has a localized Cu(II)-phen- isomer led to a decreased reactivity. The equilibrium observed
oxyl radical ground state, while in solution at room tempera- for [1]+ instead involves the localized CuII-phenoxyl radical and
ture there is an equilibrium between the localized radical and the delocalized CuII–ligand radical forms. The equilibrium
the radical delocalized over the conjugated salophen frame- caused a slower reaction rate in comparison to CuII-localized
work. Such a characteristic is suggested by the DFT calculation phenoxyl radicals.
showing that complex [1]+ in solution at room temperature has
Complex [2]+ showed a different reaction mechanism. [2]+
a delocalized radical electron, which is different from the solid has a relatively localized CuII-phenoxyl radical electronic struc-
state complex [1]SbF6 due to the effect of the existence of the ture, but its reaction with benzyl alcohol proceeds slowly in
proximal counter anion to one side of the molecule. Close comparison to complex [1]+. In addition we could not detect a
contact promotes radical localization ([1]SbF6), whereas the significant KIE value using α-d2-benzyl alcohol. The results
absence of contact rather favors a delocalized radical character indicate that the rate determining step in the course of the oxi-
([1]+). This behavior was not detected for complexes [2]+ and dation of benzyl alcohol by complex [2]+ does not involve
2290 | Dalton Trans., 2014, 43, 2283–2293
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
hydrogen abstraction at the benzylic position of the substrate.
On the other hand, [2]+ showed very low activation enthalpy and
negatively high entropy, whose values are very similar to those
of the other compounds showing the electron transfer mechan-
ism.49 Therefore, complex [2]+ can be considered as a one elec-
tron transfer reagent towards benzyl alcohol, without hydrogen
abstraction. Though complex [2]+ has an electronic structure
similar to that of the previously reported CuII-localized phen-
oxyl radical complex, it shows a different reaction rate and
mechanism.20,22,24 This may be due to the flexibility of the
ligand in comparison to other Cu(II)-phenoxyl radical species.
Indeed, the crystal structure of the [Cu(1,3-salcn)]SbF6 revealed
that the torsion angle between the two phenolate moieties of
the complex is larger than that of complex [2]+, and although
we have preliminary data only, the trans N–Cu–O(phenoxyl)
angle in complex [1]+ (163.5°) is smaller than that in complex
[2]+ (174.6°). The distortions from the square plane may
suggest a trend in the flexibility of the Cu-phenoxyl radical
bond, especially for the salophen-type complexes. In addition,
it has been reported that the reactivity of the Cu(II)–p-methoxy-
phenoxyl radical of the salcn complex is higher for benzyl
alcohol oxidation in comparison to the Cu(II)–tert-butyl-
phenoxyl radical of salcn complexes.20,22 These results suggest
that the origin of the different reactivity of the p-methoxy-
phenoxyl Cu(II) complex is not only the electronic structure of
the oxidized complex but also its flexibility.
Scheme 1 Proposed mechanism of benzyl alcohol oxidation by the
relatively localized phenoxyl radical complex [Cu(1,3-salcn)]+.
Scheme 2 Proposed mechanisms for benzyl alcohol oxidation in the
case of electronic structure difference between ground state and
CH2Cl2 solution at room temperature: (A) the case of [Cu(1,2-salcn)]+
having the equilibrium between Cu(III)-phenolate and Cu(II)-phenoxyl
radical and (B) the case of complex [1]+ showing the equilibrium
between Cu(II)-relatively localized phenoxyl radical and Cu(II)-deloca-
lized ligand radical.
The reaction mechanism of complex [3]+ is different from
that of the other salophen-type complexes. Complex [3]+ also
exhibited the slowest reaction rate for one-electron oxidized
copper complexes involving salen-type ligands. The reaction
with α-d2-benzyl alcohol allowed the determination of a KIE
value of 1.5, which suggests that hydrogen abstraction from
the α-position of benzyl alcohol is not included in the rate-
determining step. Therefore, benzyl alcohol oxidation proceeds
mainly via the electron transfer mechanism. However, the acti-
vation parameters of the reaction (activation enthalpy and
entropy changes) are not consistent with a simple electron
transfer pathway. The activation parameters of [3]+ are large
compared with the values of complex [1]+, and the substrate
saturation kinetics revealed that the reaction of complex [3]+
proceeds via association with benzyl alcohol.24 Further the
relatively high enthalpy value of the reaction suggests that the
oxidation of benzyl alcohol is energetically less favorable in
nature. The lower reactivity of complex [3]+ for oxidation of
benzyl alcohol indicates that this complex is stable in compari-
son to the other oxidized copper(II)-salen type complexes.22,24
From these results and discussion, the reaction mecha-
nisms for the benzyl alcohol oxidation, including the pre-
viously reported two [Cu(salcn)]+ species, were summarized in
Schemes 1–3. The reaction mechanisms of [1]+ and the two
[Cu(salcn)]+ species are very similar. The rate determining step
of these complexes involves the hydrogen atom abstraction
from the α-position of the benzyl alcohol.22 [Cu(1,3-salcn)]+
existing in a temperature-independent, relatively localized phe-
noxyl radical species shows a slightly faster reaction rate in
comparison to [1]+ and [Cu(1,2-salcn)]+ (Scheme 1).
Scheme 3 Proposed simple electron transfer mechanisms for benzyl
alcohol oxidation by use of complexes [2]+ and [3]+: (A) the mechanism
for complex [2]+ and (B) the substrate-adduct formation pathway for
complex [3]+.
In the case of [Cu(1,2-salcn)]+, we have previously concluded
that its slower reaction rate is caused by an equilibrium
between the less reactive Cu(III)-phenolate and the active Cu(II)-
radical for [Cu(1,2-salcn)]+ (Scheme 2(A)).21,22,24 Complex [1]+
can be considered to be in a similar situation, which is
described as an equilibrium between the localized phenoxyl
radical and the delocalized π-conjugated ligand radical in
CH2Cl2 at room temperature (Scheme 2(B)).
Complexes [2]+ and [3]+ exhibit different reaction mecha-
nisms, both of which have an electron transfer pathway
without the hydrogen abstraction (Scheme 3). In the case of
This journal is © The Royal Society of Chemistry 2014
Dalton Trans., 2014, 43, 2283–2293 | 2291

View Article Online
Paper
Dalton Transactions
complex [3]+ displaying the o-diiminobenzene radical, hydro- Grants-in-Aid for Scientific Research (no 22550055 and
gen abstraction is less favorable due to the rigid radical 25410060 to Y. S.) from the Ministry of Education, Culture,
species and inability of the coordinated imine nitrogen atoms Sports, Science, and Technology of Japan, and by the Coopera-
to bind a proton. Though complex [2]+ has a relatively localized tive Research Program of Network Joint Research Center for
phenoxyl radical species similar to [Cu(1,3-salcn)]+, the differ- Materials and Devices (Institute for Materials Chemistry and
ence in the reaction mechanism may be ascribed to a flexibility Engineering, Kyushu University). Solid state resonance Raman
difference of the complexes as discussed above. Thus, the elec- spectroscopic measurements were supported in part by the
tronic structure and flexibility differences cause the difference Research Promotion Bureau of the Ministry of Education,
in the reaction mechanism.
Culture, Sports, Science and Technology (MEXT), Japan, under
the contract no 19-198 (to T. K.).
Conclusions
Notes and references
We have synthesized and characterized one-electron oxidized
copper(II)-salophen complexes with methoxy substitution on
the phenolate and o-diaminobenzene moieties. The electronic
structures of the one-electron oxidized complexes are different;
complexes [1]+ and [2]+ showed a Cu(II)-phenoxyl radical
ground state, while complex [3]+ has a o-diiminobenzene
radical character. Further, the CH2Cl2 solutions of complexes
[1]+ and [2]+ exhibit differences in their electronic structures at
room temperature. Complex [1]+ undergoes an equilibrium
between the relatively localized phenoxyl radical on one side of
the phenolate moiety and the delocalized radical on the ligand
in solution at room temperature. In contrast, complex [2]+ can
be described as a relatively localized phenoxyl radical species
on one side of the molecule in solution. The reaction of these
oxidized complexes with benzyl alcohol gave benzaldehyde in
the same yield, but the mechanisms are different. Complex
[1]+ showed a reaction mechanism similar to that observed for
the CuIII-phenolate complex of the salcn ligand. In contrast,
the methoxy-substituted complexes [2]+ and [3]+ exhibit
different reaction mechanisms, wherein the rate-determining
step is mainly electron transfer, as reflected in the much
smaller KIE values in the presence of α-d2-benzyl alcohol. The
localized phenoxyl radical complex [2]+ exhibits activation and
kinetic parameters that suggest that the flexibility of the ligand
system has a role in determining the reactivity: the lower the
flexibility of the complexes is, the less effective they are for
proton abstraction. Complex [3]+ also showed mainly an elec-
tron transfer mechanism from isotope labeled experiments,
but this complex includes a substrate saturation mechanism.
The difference in kinetic behavior between complexes [3]+ and
1 R. H. Holm, P. Kennepohl and E. I. Solomon, Chem. Rev.,
1996, 96, 2239.
2 W. Nam (Ed), Acc. Chem. Res., 2007, 40, 465–634 (Special
issue on Dioxygen Activation by Metalloenzymes and
Models).
3 (a) H. Sono, M. P. Roach, E. D. Coulter and J. H. Dawson,
Chem. Rev., 1996, 96, 2841; (b) B. Meunier, S. P. de Visser
and S. Shaik, Chem. Rev., 2004, 104, 3947.
4 (a) M. Baik, M. Newcomb, R. A. Friesner and S. J. Lippard,
Chem. Rev., 2003, 103, 2385; (b) E. Y. Tshuva and
S. J. Lippard, Chem. Rev., 2004, 104, 987; (c) M. Costas,
M. P. Mehn, M. P. Jensen and L. Que, Jr., Chem. Rev., 2004,
104, 939; (d) E. I. Solomon, T. C. Brunold, M. I. Davis,
J. N. Kemsley, S.-K. Lee, N. Lehnert, F. Neese, A. J. Skulan,
Y.-S. Yang and J. Zhou, Chem. Rev., 2000, 100, 235;
(e) A. Gunay and K. H. Theopold, Chem. Rev., 2010, 110,
1060.
5 (a) L. Que, Jr. and W. B. Tolman, Angew. Chem., Int. Ed.,
2002, 41, 1114; (b) L. M. Mirica, X. Ottenwaelder and
T. D. P. Stack, Chem. Rev., 2004, 104, 1013; (c) E. A. Lewis
and W. B. Tolman, Chem. Rev., 2004, 104, 1047.
6 J. Stubbe and W. A. van der Donk, Chem. Rev., 1998, 98,
705.
7 P. A. Frey, A. D. Hegeman and G. H. Reed, Chem. Rev.,
2006, 106, 3302.
8 (a) J. W. Whittaker, Met. Ions Biol. Syst., 1994, 30, 315;
(b) J. W. Whittaker, Chem. Rev., 2003, 103, 2347.
9 K. Clark, J. E. Penner-Hahn, M. M. Whittaker and
J. W. Whittaker, J. Am. Chem. Soc., 1990, 112, 6433.
[2]+ shows that the reaction mechanism is affected by not only 10 M. L. McGlashen, D. D. Eads, T. G. Spiro and
the difference in geometry but also the difference in electronic
J. W. Whittaker, J. Phys. Chem., 1995, 99, 4918.
structure. Further studies on the correlation between reactivity 11 B. A. Jazdzewski and W. B. Tolman, Coord. Chem. Rev.,
and electronic structure are currently under way in our
laboratories.
2000, 200–202, 633.
12 (a) P. Chaudhuri and K. Wieghardt, Prog. Inorg. Chem.,
2001, 50, 151; (b) F. Thomas, Eur. J. Inorg. Chem., 2007,
2379; (c) Y. Shimazaki and O. Yamauchi, Indian J. Chem.,
Sect. A: Inorg., Bio-inorg., Phys., Theor. Anal. Chem., 2011,
50, 383 and references cited therein; (d) K. Butsch,
T. Günther, A. Klein, K. Stirnat and A. Berkessel, Inorg.
Chim. Acta, 2013, 394, 237.
Acknowledgements
We would like to thank Prof. Satoru Karasawa of the Graduate
School of Pharmaceutical Sciences, Kyushu University for dis-
cussions and suggestions regarding the magnetic properties of 13 (a) P. J. Chirik and K. Wieghardt, Science, 2010, 327, 794;
oxidized salophen complexes. This work was supported by
(b) P. L. Holland, Acc. Chem. Res., 2008, 41, 905; (c) B. de
2292 | Dalton Trans., 2014, 43, 2283–2293
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
Bruin, D. G. H. Hetterscheid, A. J. J. Koekkoek and 25 A. Kochem, O. Jarjayes, B. Baptiste, C. Philoze, H. Vezin,
H. Grützmacher, Prog. Inorg. Chem., 2007, 55, 247.
K. Tsukidate, F. Tani, M. Orio, Y. Shimazaki and
14 (a) P. Chaudhuri, C. N. Verani, E. Bill, E. Bothe,
F. Thomas, Chem.–Eur. J., 2012, 18, 1068.
T. Weyhermüller and K. Wieghardt, J. Am. Chem. Soc., 2001, 26 D. D. Perrin, W. L. F. Armarego and D. R. Perrin, Purifi-
123, 2213; (b) D. Herebian, E. Bothe, E. Bill,
T. Weyhermüller and K. Wieghardt, J. Am. Chem. Soc., 2001,
123, 10012.
cation of Laboratory Chemicals, Pergamon Press, Elmsford,
1966.
27 J. Chastain and R. G. King, Jr., Handbook of X-ray Photo-
electron Spectroscopy, Physical Electronics, Inc., Minnesota,
1995.
15 (a) K. Ray, T. Petrenko, K. Wieghardt and F. Neese, Dalton
Trans., 2007, 1552; (b) K. Ray, S. D. George, E. I. Solomon,
K. Wieghardt and F. Neese, Chem.–Eur. J., 2007, 13, 2783; 28 J. A. Ibers and W. C. Hamilton, International Tables
(c) K. Ray, T. Weyhermuller, F. Neese and K. Wieghardt,
Inorg. Chem., 2005, 44, 5345.
for X-Ray Crystallography, Kynoch, Birmingham, 1974,
vol. IV.
16 (a) Y. Shimazaki, F. Tani, K. Fukui, Y. Naruta and 29 Crystal Structure Analysis Package, Molecular Structure Cor-
O. Yamauchi, J. Am. Chem. Soc., 2003, 125, 10512; poration, The Woodlands, TX, 1985 and 1999.
(b) Y. Shimazaki, T. Yajima, F. Tani, S. Karasawa, K. Fukui, 30 F. Neese, WIREs Comput. Mol. Sci., 2012, 2, 73.
Y. Naruta and O. Yamauchi, J. Am. Chem. Soc., 2007, 129, 31 J. P. Perdew, Phys. Rev. B: Condens. Matter, 1986, 33, 8822.
2559.
32 J. P. Perdew, Phys. Rev. B: Condens. Matter, 1986, 34, 7406.
17 Y. Shimazaki, T. D. P. Stack and T. Storr, Inorg. Chem., 33 A. D. Becke, Phys. Rev. A, 1988, 38, 3098.
2009, 48, 8383.
34 F. Weigend and R. Ahlrichs, Phys. Chem. Chem. Phys., 2005,
7, 3297.
18 (a) T. Storr, E. C. Wasinger, R. C. Pratt and T. D. P. Stack,
Angew. Chem., Int. Ed., 2007, 46, 5198; (b) T. Storr, 35 F. Neese, J. Comput. Chem., 2003, 24, 1740.
P. Verma, Y. Shimazaki, E. C. Wasinger and T. D. P. Stack, 36 F. Weigend, Phys. Chem. Chem. Phys., 2006, 8, 1057.
Chem.–Eur. J., 2010, 16, 8980.
19 (a) O. Rotthaus, O. Jarjayes, F. Thomas, C. Philouze, C. P. Del 38 C. T. Lee, W. T. Yang and R. G. Parr, Phys. Rev. B: Condens.
Valle, E. Saint-Aman and J.-L. Pierre, Chem.–Eur. J., 2006, 12, Matter, 1988, 37, 785.
2293; (b) O. Rotthaus, F. Thomas, O. Jarjayes, C. Philouze, 39 F. Thomas, O. Jarjayes, C. Duboc, C. Philouze, E. Saint-
E. Saint-Aman and J.-L. Pierre, Chem.–Eur. J., 2006, 12, 6953; Aman and J.-L. Pierre, Dalton Trans., 2004, 2662.
37 A. D. Becke, J. Chem. Phys., 1993, 98, 1372.
(c) L. Chiang, A. Kochem, O. Jarjayes, T. J. Dunn, H. Vezin, 40 Y. Shimazaki, Adv. Mater. Phys. Chem., 2013, 3, 60.
M. Sakaguchi, T. Ogura, M. Orio, Y. Shimazaki, F. Thomas 41 M. B. Robin and P. Day, Adv. Inorg. Chem., 1967, 10, 247.
and T. Storr, Chem.–Eur. J., 2012, 18, 14117.
42 C. Creutz and H. Taube, J. Am. Chem. Soc., 1969, 91, 3988.
20 M. Orio, F. Jarjayes, H. Kanso, C. Philouze, F. Neese and 43 N. S. Hush, Prog. Inorg. Chem., 1967, 8, 391.
F. Thomas, Angew. Chem., Int. Ed., 2010, 49, 4989.
21 T. Storr, P. Verma, R. C. Pratt, E. C. Wasinger, Y. Shimazaki
and T. D. P. Stack, J. Am. Chem. Soc., 2008, 130, 15448.
44 D. M. D’Alessandro and F. R. Keene, Chem. Soc. Rev., 2006,
35, 424.
45 N. S. Hush, Coord. Chem. Rev., 1985, 64, 135.
22 K. Asami, K. Tsukidate, S. Iwatsuki, F. Tani, S. Karasawa, 46 A. Mukherjee, M. L. McGlashen and T. G. Spiro, J. Phys.
L. Chiang, T. Storr, F. Thomas and Y. Shimazaki, Inorg.
Chem., 2012, 51, 12450.
Chem., 1995, 99, 4912.
47 F. Neese, J. Biol. Inorg. Chem., 2006, 11, 702.
23 Y. Shimazaki, A. Arai, T. J. Dunn, T. Yajima, F. Tani, 48 It is noticed that with such values, however, the Cu(III) state
C. F. Ramogida and T. Storr, Dalton Trans., 2011, 40, 2469.
24 (a) R. C. Pratt and T. D. P. Stack, J. Am. Chem. Soc., 2003,
may be accessible under different phase or solvent con-
ditions. See for example: ref. 21 and 24.
125, 8716; (b) R. C. Pratt and T. D. P. Stack, Inorg. Chem., 49 A. G. Lappin, Redox Mechanisms in Inorganic Chemistry,
2005, 44, 2367.
Ellis Horwood, Chichester, UK, 1994.
This journal is © The Royal Society of Chemistry 2014
Dalton Trans., 2014, 43, 2283–2293 | 2293