Aromatic hydroxylation in a copper bis(imine) complex mediated by a μ-η2:η2 peroxo dicopper core: A mechanistic scenario

Article abstract of DOI:10.1002/chem.200800799

Detailed mechanistic studies on the ligand hydroxylation reaction mediated by a copper bis(imine) complex are presented. Starting from a structural analysis of the Cu¹ complex and the Cu^II product with a hydroxylated ligand, the optical absorption and vibrational spectra of starting material and product are analyzed. Kinetic analysis of the ligand hydroxylation reaction shows that O₂ binding is the ratelimiting step. The reaction proceeds much faster in methanol than in acetonitrile. Moreover, an inverse kinetic isotope effect (KIE) is evidenced for the reaction in acetonitrile, which is attributed to a sterically congested transition state leading to the peroxo adduct. In methanol, however, no KIE is observed. A DFT analysis of the oxygenation reaction mediated by the μ-η²:η² peroxo core demonstrates that the major barrier after O₂ binding corresponds to electrophilic attack on the arene ring. The relevant orbital interaction occurs between the σ* orbital of the Cu ₂O₂ unit and the HOMO of the ligand. On the basis of the activation energy for the rate-limiting step (18.3 kcal mol^-1) this reaction is thermally allowed, in agreement with the experimental observation. The calculations also predict the presence of a stable dienone intermediate which, however, escaped experimental detection so far. Reasons for these findings are considered. The implications of the results for the mechanism of tyrosinase are discussed.

Full text of DOI:10.1002/chem.200800799

DOI: 10.1002/chem.200800799
Aromatic Hydroxylation in a Copper Bis(imine) Complex Mediated by a
A
m-h2:h2 Peroxo Dicopper Core: A Mechanistic Scenario
Ole Sander,[a] Anja Henß,[b] Christian Näther,[a] Christian Würtele,[b]
Max C. Holthausen,*[c] Siegfried Schindler,*[b] and Felix Tuczek*[a]
Dedicated to Professor Helmut Schwarz on the occasion of his 65th birthday
Abstract: Detailed mechanistic studies
on the ligand hydroxylation reaction
the reaction in acetonitrile, which is at-
tributed to a sterically congested transi-
tion state leading to the peroxo adduct.
In methanol, however, no KIE is ob-
served. A DFT analysis of the oxygena-
tion reaction mediated by the m-h2:h2
peroxo core demonstrates that the
major barrier after O2 binding corre-
sponds to electrophilic attack on the
arene ring. The relevant orbital interac-
tion occurs between the s* orbital of
the Cu2O2 unit and the HOMO of the
ligand. On the basis of the activation
energy for the rate-limiting step
(18.3 kcalmolꢀ1) this reaction is ther-
mally allowed, in agreement with the
experimental observation. The calcula-
tions also predict the presence of a
stable dienone intermediate which,
however, escaped experimental detec-
tion so far. Reasons for these findings
are considered. The implications of the
results for the mechanism of tyrosinase
are discussed.
mediated by a copper bis(imine) com-
A
plex are presented. Starting from a
structural analysis of the CuI complex
and the CuII product with a hydroxylat-
ed ligand, the optical absorption and
vibrational spectra of starting material
and product are analyzed. Kinetic anal-
ysis of the ligand hydroxylation reac-
tion shows that O2 binding is the rate-
limiting step. The reaction proceeds
much faster in methanol than in aceto-
nitrile. Moreover, an inverse kinetic
isotope effect (KIE) is evidenced for
Keywords: copper · hydroxylation ·
ꢀ
N ligands · O O activation · reac-
tion mechanisms
Introduction
Recent publication of the first crystal structure determina-
tion of a tyrosinase has opened a new perspective for under-
standing the chemical reactivity of this class of enzymes at a
molecular level.[1,2] Tyrosinases (Ty) are ubiquitous copper
enzymes mediating the hydroxylation of monophenols to o-
diphenols and subsequent two-electron oxidation to o-qui-
nones.[3] Specifically, tyrosine is converted to dopaquinone,
the first step of melanine synthesis.[4] Two-electron oxidation
of o-diphenols (catechols) to o-quinones is also catalyzed by
the related enzyme catechol oxidase (CO), which, however,
lacks monooxygenase activity.[5,6] The active sites of Ty and
CO exhibit two copper atoms both of which are coordinated
by three histidine residues (type 3 copper). The third group
of proteins with type 3 copper active sites is that of hemo-
cyanins (Hc), which serve as oxygen carriers in some arthro-
pods and mollusks and exhibit highly cooperative oxygen-
binding characteristics.[7]
[a] Dipl.-Chem. O. Sander, Priv. Doz. C. Näther, Prof. Dr. F. Tuczek
Institut für Anorganische Chemie
Christian Albrechts Universität Kiel
Max-Eyth Strasse 2, 24098 Kiel (Germany)
Fax : (+49)4318801520
E-mail: ftuczek@uni-kiel.de
[b] Dipl.-Chem. A. Henß, Dipl.-Chem. C. Würtele, Prof. Dr. S. Schindler
Institut für Anorganische und Analytische Chemie
Justus-Liebig-Universität Gießen
Heinrich-Buff-Ring 58, 35392 Gießen (Germany)
Fax : (+49)6419934149
E-mail: siegfried.schindler@chemie.uni-giessen.de
[c] Prof. Dr. M. C. Holthausen
Institut für Anorganische und Analytische Chemie
Johann Wolfgang Goethe-Universität Frankfurt
60438 Frankfurt am Main (Germany)
Fax : (+49)79829417
E-mail: max.holthausen@chemie.uni-frankfurt.de
There have been numerous attempts to reproduce and un-
derstand the chemical reactivity of tyrosinase with small-
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200800799.
9714
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 9714 – 9729

FULL PAPER
molecule model systems, both on the basis of the hydroxyl-
ation or oxidation of external substrates and on hydroxyl-
ation of the ligand.[8,9,10,11] In the latter case part of the
ligand coordinating one or both copper centers is hydroxy-
lated after exposure of the CuI complex to dioxygen. The
relevance of these reactions to tyrosinase has intensively
been discussed, and a molecular mechanism of tyrosinase
functionality has been suggested based on DFT calcula-
tions.[12,13] It has further been established that both aliphatic
and aromatic parts of the ligand can be hydroxylated in this
way (aliphatic and aromatic ligand hydroxylation).[8,14] The
latter reactivity was discovered by Karlin and co-workers in
Herein we present spectroscopic, kinetic, and theoretical
investigations on the Cu2 bis
A
AHCTREUNG
nomethyl]benzene; Scheme 1) with the goal of developing a
mechanistic scenario for the ligand hydroxylation reaction
occurring in this system. To this end the kinetics of the reac-
tion were studied in different solvents and at different tem-
peratures, and a deuterium-substituted ligand was also em-
ployed. In order to monitor the time course of the hydroxyl-
ation spectroscopically the UV/Vis spectra of the CuI pre-
cursor I and the hydroxylated CuII product II were analyzed
and compared with each other. Moreover, the vibrational
spectroscopic properties of the reactant and the product
were determined. In particular, the IR and Raman spectra
of the CuII complex II are analyzed with the help of 18O sub-
stitution and quantum chemical calculations. Reactant I and
product II were further characterized by X-ray crystallogra-
phy. Density functional theory was employed to identify po-
tential reaction pathways leading from the initially formed
m-h2:h2 peroxo dicopper intermediate to the hydroxylated
product II. The quantum chemical results are discussed in
light of the experimental findings, and implications for the
reactivity of tyrosinase are considered.
their study on [Cu2(XYL)] (XYL=tetrakis[2-(pyridin-2-
G
yl)ethyl]benzene-1,3-diamine).[15] Reaction of the CuI2 pre-
cursor with O2 was found to lead to hydroxylation of the
bridging xylylene spacer in the 2-position. It was later
shown that the CuI2 species binds O2 in a side-on bridging
fashion and that most probably the Cu2 m-h2:h2-peroxo unit
mediates electrophilic attack on the aromatic ring, leading
[16]
ꢀ
to O O cleavage and hydroxylation. A mechanistic alter-
native to this scenario is full or partial conversion of the
CuII m-h2:h2 peroxo species to the CuIII bis
ACHTREUNG
2
2
ACHTREUNG
ation mediated by a Cu2 bis
lished.[18]
A
Experimental and Computational Details
A simplified version of the Karlin system is provided by
Materialsand techniques : The reagents isophthalaldehyde and 3-dime-
thylaminopropylamine were used as received from Aldrich Chemical Co.
Tetrakis(acetonitrile)copper(I) hexafluorophosphate was either obtained
commercially from Aldrich or synthesized from copper(I) oxide accord-
ing to a method described in the literature.[25] Solvents used were all of
reagent grade and were further purified by refluxing over drying agents
CuII bis
ACHTREUNG
of the bridging ligand on reaction with O2. This reaction has
been intensively studied as well,[19,20,21,22] and its relevance to
the tyrosinase reaction has been stressed.[23,24] The binucleat-
ing ligand contains a bridging xylylene group carrying two
arms which provide two nitrogen donors (one imine and
one terminal amine) each. On exposure of the CuI precursor
to dioxygen, the central phenylene spacer is hydroxylated,
in analogy to the XYL complex (Scheme 1). In the latter
and distilling under argon. Methanol was distilled from Mg(OCH3)2, di-
A
ethyl ether from LiAlH4, and acetonitrile from CaH2. The NMR spectra
were recorded at 300 K on a Bruker Avance 400 Pulse Fourier Transform
spectrometer operating at a 1H frequency of 400.13 MHz and 13C fre-
quency of 100.62 MHz relative TMS as standard. Elemental analysis was
performed using an Euro Vector CHNS-O element analyzer (Euro EA
3000). Samples were burned in sealed tin containers in a stream of
oxygen. FTIR spectra were recorded in KBr pellets on a Mattson Gene-
sis Type I spectrometer. Optical absorption spectra of solutions were re-
corded on a Cary 5 UV/Vis/NIR spectrometer equipped with a CTI cryo-
cooler. Raman spectra were recorded on a Bruker IFS 66 FT spectrome-
ter equipped with a Raman assembly.
system, however, the bis(imine) ligation enforces an almost
G
planar molecular geometry, which highly restricts the config-
uration space involved in the ligand hydroxylation reaction.
Although it has been speculated that in this reaction a
peroxo or bis(m-oxo) intermediate is involved, no such inter-
H
Variable-temperature stopped-flow measurements allowed the collection
of time-resolved UV/Vis spectra for the fast reaction of I with dioxygen
in methanol. Solutions of the complexes were prepared in a glove box
(MBraun, Garching, Germany) and transferred by syringe to the low-
temperature stopped-flow instrument. A dioxygen-saturated solution was
prepared by bubbling dioxygen through methanol in a syringe (solubility
of dioxygen in MeOH at 258C: 8.5 mm).[26] Lower dioxygen concentra-
tions were obtained by mixing these solutions with argon-saturated sol-
vents. The reaction was studied under pseudo-first-order conditions
([complex]![O2]), and time-resolved UV/Vis spectra of the reactions of
dioxygen with copper(I) complexes were recorded with a modified Hi-
Tech SF-3L low-temperature stopped-flow unit (Salisbury, UK) equipped
with a J & M TIDAS 16-500 photodiode-array spectrophotometer (J &
M, Aalen, Germany). Data fitting was performed with the integrated J &
M software Kinspec. Details of such studies have been described previ-
ously.[27]
mediate has ever been detected. Accordingly, key features
of the reaction course for this important class of tyrosinase
model systems have remained unclear to date.
Scheme 1. Hydroxylation of the Cu2
II (charges are omitted).
A
Single-crystal structure analysis: The X-ray crystallographic data for com-
plex I were collected at 173 K on a STOE IPDS diffractometer equipped
Chem. Eur. J. 2008, 14, 9714 – 9729
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9715

M. C. Holthausen, S. Schindler, F. Tuczek et al.
with a low-temperature system (Karlsruher Glastechnisches Werk). MoKa
radiation (l=0.71069 Š) and a graphite monochromator were used. Cell
parameters were refined by using up to 5000 reflections. No absorption
corrections were applied. The structure was solved by direct methods
with SHELXS97, and refined as a racemic twin by using full-matrix least-
squares techniques in SHELXL97 (BASF parameter: 0.542). An attempt
to solve and refine the structure in the centrosymmetric space group P21/
m was unsuccessful. The hydrogen atoms were positioned geometrically
and all non-hydrogen atoms were refined anisotropically, if not men-
tioned otherwise. Details of the structure determination are given in
Table 1.
were filtered off and dried. Elemental analysis (%) calcd for
CuC8H12N4ClO4: C 29.37, H 3.7, N 17.12, Cl 10.84; found: C 29.2, H 3.66,
N 17.4, Cl 10.81.
1,3-bis{[3-(N,N-dimethyl)propyl]iminomethyl}benzene (DAPA): Iso-
phthalaldehyde (400 mg, 2.98 mmol) and (610 mg, 5.96 mmol) 3-
dimethylaminopropylamine were dissolved in methanol (40 mL) and the
A
solution heated to reflux for 1 h. The solvent was removed on a rotary
evaporator and the remaining yellow oil dried in vacuo. The product was
purified by chromatography on silica gel with methanol as eluant (Rf =
0.4). Elemental analysis (%) calcd for C18H30N4: C 71.48, H 10.0, N
18.52; found: C 70.93, H 10.35, N 18.53; 1H NMR (400 MHz, CD2Cl2/
TMS): d=8.3 (s, 2H; imine H), 8.03 (s, 1H; ArH), 7.76 (dd, 2H; ArH),
7.44 (t, 1H; ArH), 3.61 (dt, 4H,=NCH2), 2.3 (t, 4H; -CH2N), 2.18 (s,
12H; CH3), 1.81 ppm (q, 4H; -CH2-); 13C NMR (100.6 MHz, CD2Cl2/
TMS): d=160.2, 137.0, 129.6, 128.7, 127.5, 59.4, 57.4, 42.2, 29.0 ppm; MS
(EI, 70 eV): m/z (%): 303.4 (100) [M+]; calcd: 303.46.
Table 1. Crystal data and structural refinement for I and II.
Parameter
I
II
empirical formula
formula weight
temperature [K]
crystal system
space group
wavelength [Š]
a [Š]
b [Š]
c [Š]
a [8]
b [8]
C24H39N7Cu2P2F12 C19H32N4Cu2Cl4O10
A
2(DAPA)
(CH3CN)]2+ (I): The complex was prepared under argon
N
842.64
745.37
173(2)
monoclinic
P21
170(2) K
triclinic
P1
atmosphere, either according to the published procedure as the PF6 salt
(Ia)[28] or as the ClO4 salt (Ib) by using the following modified proce-
dure: DAPA (230 mg, 0.76 mmol) was dissolved in dry, degassed metha-
nol (20 mL). Tetrakis(acetonitrile)copper(I) perchlorate 296 mg
(1.52 mmol) was added. The resulting yellow solution was heated for 1 h.
After concentrating the solution to 10 mL a yellow solid precipitated,
which was filtered off and washed with degassed methanol (2”5 mL).
1H NMR (400 MHz, CD3CN/TMS): d=8.37 (s, 2H, imine H), 8.25 (s,
1H, ArH), 7.93 (d, 2H, ArH), 7.55 (t, 1H, ArH), 3.72 (t, 4H,=NCH2),
2.51 (t, 4H, -CH2N), 2.23 (s, 12H, CH3), 1.81 ppm (q, 4H, -CH2-);
13C NMR (100.6 MHz, CD3CN/TMS): d=163.2, 137.0, 131.8, 130.0, 129.1,
62.7, 60.7, 47.2, 29.3 ppm. Crystals suitable for diffraction studies were
obtained by diffusion of diethyl ether into a solution of Ia in acetonitrile.
¯
0.71073
11.491(2)
25.391(5)
12.564(3)
0.71073
7.8818(9)
13.955(2)
14.649(2)
62.06(2)
b=84.90(2)
81.66(2)
1407.9(3)
2
105.59(3)
g [8]
V [Š3]
3530.8(12)
4
Z
1calcd [Mgmꢀ3
]
1.585
1.758
abs. coeff. [cmꢀ1
]
1.386
1.948
Synthesis of the tetraphenylborate salt of [CuI
2(DAPA)] (Ic): Salt Ib
E
crystal size [mm]
q range for data collection [8]
index ranges
0.25”0.1”0.12
2.13–26.04
ꢀ14ꢁhꢁ13,
ꢀ31ꢁkꢁ31,
ꢀ14ꢁlꢁ15
24832
13443
6125
0.1061
0.780
0.2”0.2”0.1
2.61–25.03
ꢀ9ꢁhꢁ9,
ꢀ16ꢁkꢁ16,
ꢀ17ꢁlꢁ17
12085
4681
3304
0.1250
0.998
(200 mg, 0.32 mmol) was dissolved in dry, degassed acetonitrile (10 mL).
Potassium tetraphenylborate (228 mg, 0.64 mmol) dissolved in acetoni-
trile (15 mL) was added. After concentrating the yellow solution to
10 mL a colorless solid precipitated and was filtered off. The solution was
evaporated to dryness and the yellow residue was used without further
purification. The conversion of the perchlorate to the tetraphenylborate
salt was checked for completeness by IR spectroscopy. The perchlorate
bands were absent in the product.
reflns collected
independent reflns
reflns with I>2s(I)
Rint
GOF on F2
A
N
R1 [I>2s(I)]
wR2 [I>2s(I)]
R1 (all data)
wR2 (all data)
0.0532
0.0845
0.1389
0.1065
0.0519
0.1172
0.0805
0.1319
solved in dry dichloromethane (40 mL). Dioxygen was bubbled through
the solution for 5 min. The color changed from yellow to green. The solu-
tion was concentrated to 5 mL. On adding 20 mL of diethyl ether a green
solid precipitated, which was filtered off. The product was recrystallized
from dichloromethane by diffusing diethyl ether into the solution. Green
crystals were obtained. Elemental analysis (%) calcd for C18H30N4Cu2O2-
max./min. residual electron den- 0.489/ꢀ0.324
0.646/ꢀ1.242
ꢀ3
sity [eŠ
]
A
N
UV/Vis (CH3CN) lmax (e)=255 (35753), 357 (9135), 628 nm
(253 mꢀ1 cmꢀ1). The 18O isotopomer of II was prepared analogously by
employing 18O2 instead of 16O2.
Data collection on complex II was performed with an Imaging Plate Dif-
fraction System (IPDS-1) from STOE & CIE. Structure solution was
done with SHELXS-97, and structure refinement against F2 with
SHELXL-97. All non-hydrogen atoms were refined with anisotropic dis-
placement parameters. All CH hydrogen atoms were positioned with ide-
alized geometry and were refined isotropically by using a riding model.
The OH H atom was located in the difference map; its bond length was
set to ideal values and afterwards it was refined by using a riding model.
CCDC-686492 (I) and CCDC-686493 (II) contain the supplementary
crystallographic data for this paper. These data can be obtained free of
charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
[D6]Isophthalaldehyde: m-[D10]xylene 2.5 g (21.55 mmol) was dissolved
in CCl4 (250 mL). N-Bromosuccinimide (22.55 g, 126.7 mmol) and one
drop of bromine were added. The reaction was started with small
amounts of 2,2’-azobisisobutyronitrile and the mixture was heated for
15 h under reflux. The succinimide was filtered off and after removal of
the solvent the product was used without further purification. The red-
dish product was dissolved in H2SO4 (30 mL) at 1108C. After the forma-
tion of bromine the solution was hydrolyzed with ice (100 mL) and ex-
tracted with methyl tert-butyl ether (MTBE). The combined organic
phases were neutralized with NaHCO3 solution and the aqueous phase
was again extracted with MTBE. The combined organic phases were
dried over MgSO4 and after removal of the solvent the product was puri-
fied by chromatography on silica gel with dichloromethane as eluant
(Rf =0.35). 1H NMR showed no signals; 13C NMR (100.6 MHz, CD2Cl2/
TMS): d=190.7 (deutero-t), 136.9 (s), 134.1 (deutero-t), 130.2 (deutero-
t), 129.4 ppm (deutero-t); MS (EI, 70 eV): m/z (%): 141.1 (100) [M+];
calcd: 141.2.
Tetrakis(acetonitrile)copper(I) perchlorate: CuCO3 was added to
perchloric acid (10 mL) until the solution was saturated. The remaining
precipitate was filtered off and the solution was concentrated in vacuo.
After cooling the solution overnight, blue crystals precipitated, which
were filtered off and dissolved in acetonitrile. The blue solution was re-
fluxed with copper turnings under argon till the color disappeared. After
cooling of the colorless solution colorless crystals precipitated, which
9716
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 9714 – 9729

Aromatic Hydroxylation in a Copper Bis(imine) Complex
A
FULL PAPER
A
(CH3CN)3]
U
lematic, depending on the strength of the spin–spin coupling. The
broken-symmetry (BS) approach has been identified as an efficient
means to include the dynamic and static correlation effects underlying
these magnetic interactions to a large extent, and it has been applied to
related bioinorganic problems with considerable success.[38] Following the
suggestion of Noodleman et al.[39] we applied spin-unrestricted broken-
symmetry calculations for the antiferromagnetically coupled singlet
states. The overlap integrals hajbi of the magnetic orbitals obtained for
the BS wave functions vary significantly (Table S1, Supporting Informa-
tion), which indicates strongly varying coupling strengths between the
spin centers involved. We therefore applied the formalism of Yamaguchi
et al.[40] to obtain the Heisenberg coupling parameter J related to the
phenomenological Heisenberg Hamiltonian H=ꢀ2JSASB [Eq. (1)].
2.82 mmol) was dissolved in dry methanol (40 mL). 3-Dimethylaminopro-
pylamine (0.75 mL, 5.9 mmol) was added and the solution heated to
reflux for 1 h. After cooling tetrakis(acetonitrile)copper(I) perchlorate
(1.1 g, 5.9 mmol) was added to the orange solution, which was stirred for
1 h. After concentrating the solution to 10 mL, a yellow solid precipitat-
ed, which was filtered off and washed with degassed methanol (2”5 mL).
1H NMR (400 MHz, CD3CN/TMS): d=3.74 (t, 4H,=NCH2), 2.53 (t, 4H;
CH2N), 2.24 (s, 12H; CH3), 1.82 ppm (q, 4H; CH2 ); 2H NMR
(61.4 MHz, CD3CN/TMS) between 7.4 and 8.4 ppm four peaks with inte-
gration ratio 2:2:1:1 were observed; 13C NMR (100.6 MHz, CD3CN/
TMS): d=61.8, 59.8, 46.2, 28.2 ppm.
ꢀ
[CuII2([D6]DAPA-O)(OH)], perchlorate salt (IID): Salt ID (60 mg) was
dissolved in dry dichloromethane (40 mL). Dioxygen was bubbled
through the solution for 5 min. The color changed from orange to green.
The solution was concentrated to 5 mL. By adding diethyl ether (20 mL),
a green solid precipitated which was filtered off and recrystallized from
dichloromethane. Elemental analysis (%) calcd for C18H12D6N4Cu2O2-
EHSꢀEBS
J ¼ ꢀ
ð1Þ
hS2iHSꢀhS2iBS
This approach covers the range from weak to strong coupling situations.
The BS wave functions were obtained by employing the corresponding
orbital transformation procedure implemented in ORCA.[41] In the pres-
ent context, a negative value for J corresponds to an “antiferromagneti-
cally coupled” or “open-shell” singlet ground state.
A
Caution! Although the compounds reported in this paper seem to be
stable to shock and heat, extreme care should be used in handling them
because of the potential explosive nature of perchlorate salts.
Computational methods: Quantum-chemical calculations on reaction
pathways were performed at the DFT level with the three-parameter
hybrid functional B3LYP/G[29] as implemented in the ORCA program.[30]
Geometry optimizations and harmonic frequency calculations were per-
formed by employing the SVP basis set of Ahlrichs and co-workers[31] for
all atoms. In all calculations we used the TightSCF, NoFinalGrid, and
Grid4 options/cutoffs, and the RIJONX approach[32] was used together
with the SV/J auxiliary basis set[33] for enhanced numerical efficiency via
the RI approximation.[34] Solvation effects were included in these calcula-
tions by employing the COSMO continuum model (solvent acetonitrile,
dielectric constant at room temperature e=36.6; the following radii were
used for construction of the cavity: H: 1.300, C: 2.000, N: 1.830 Š, O:
1.720, Cu: 2.223, solvent: 1.300 Š).[35] The nature of stationary points lo-
calized (minima or transition structures) were identified by Hessian cal-
culations based on numerical evaluation of energies and analytical gradi-
ents, which were also used to obtain zero-point vibrational energy
(ZPVE) and thermal contributions to Gibbs free energies at 298.15 K.
We verified the connections between minima and transition structures
implied in Figure 12 below by intrinsic reaction coordinate (IRC) calcula-
tions. For transition-state searches, IRC calculations, and numerical Hes-
sian calculations we used the Gaussian 03[36] external driver facility in
combination with a Gau_External module that we developed to extract
energies and gradients from ORCA calculations, which were then fed
into geometry optimization driver routines of the Gaussian 03 pro-
gram.[37] Improved final energies were obtained by single-point calcula-
tions employing the B3LYP/G functional in combination with the TZVP
basis of Ahlrichs and co-workers and the COSMO continuum solvent
model (together with the TZV/J auxiliary basis sets, with all other pro-
gram parameters and options as described above).
For some species involved in the reaction pathways discussed below we
found triplet ground states (J=+25 to +534 cmꢀ1, see Table S1, Support-
ing Information). In some instances we reoptimized the corresponding
triplet structures, but we did not observe any significant energy lowering
or structural change (e.g., 1 kcalmolꢀ1 in the case of 1 with essentially un-
altered structural features). Analysis of spin densities for all species in-
vestigated revealed that all magnetic interactions are caused by interac-
tions of spin densities essentially localized on the copper ions with their
formal d9 electronic configurations. These CuII-based spin systems exhibit
antiferromagnetically coupling in most cases, which gives rise to singlet
ground states, but triplet ground states result in some instances as a con-
sequence of ferromagnetic spin coupling. The intricacies of magnetic cou-
pling in related Cu2O2 systems due to the interplay between electronic
and structural properties are the subject of ongoing research,[42] and, in
view of the general tendency of the B3LYP functional to overestimate
the stability of high-spin over low-spin states for transition metal ions,[43]
they are outside the scope of the present study. The small energy differ-
ences between singlet and triplet species documented in Table S1
(<2 kcalmolꢀ1 in all cases) are insignificant in the context of our investi-
gation of reaction pathways. Hence, for the present system we can safely
exclude the possibility that copper-based spin coupling gives rise to a
prominent two-state reactivity scenario[44] in the sense that spin crossover
phenomena could provide alternative reaction pathways that significantly
alter the mechanistic scenario proposed below.[45]
Additional calculations were performed to support analysis of experi-
mental IR and resonance-Raman spectra. Here we used the UBP86 func-
tional[46] to optimize (BS) the singlet structure for the full molecular
model of the hydroxylated product, starting from the X-ray structure of
II. It is well established that this level of DFT quite generally provides vi-
brational frequencies in good agreement with experimentally determined
spectra.[47,48] Calculated frequencies were therefore used without further
scaling.
In several instances frequency analyses of stationary points obtained for
the full molecular model showed spurious imaginary modes related to ro-
tation of the methyl groups at the tert-amine N donor atoms, which seri-
ously affects the use of computed ZPVE and thermal contributions to
obtain Gibbs free energies. In view of the unjustifiably large numerical
effort necessary for repeated reoptimizations of geometries and numeri-
cal frequency calculations, we decided to perform investigations on reac-
tion pathways based on a simpler molecular model, in which we replaced
the methyl groups by hydrogen atoms.
Results
X-ray structure analysis
In all species studied here (but TS8 with its closed-shell singlet ground
state wave function, see Figure 12 below and Table S1, Supporting Infor-
mation), the presence of two coupled CuII ions with their formal d9 elec-
tronic configuration gives rise to spin–spin coupling phenomena, that is,
the two unpaired electrons can couple to yield singlet or triplet states.
While treatment of the triplet states is straightforward within the spin-un-
restricted Kohn–Sham framework, a description of the corresponding sin-
glet states by spin-restricted Kohn–Sham calculations can be highly prob-
CuI complex Ia: The DAPA ligand and the corresponding
copper complex were prepared according to the literature.[25]
Despite the facile synthesis of I and related bis(imine) com-
A
plexes, crystallographic characterizations of these com-
pounds are rare. Recently a copper(I) complex with a ligand
similar to DAPA (ethylene instead of propylene bridges)
Chem. Eur. J. 2008, 14, 9714 – 9729
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9717

M. C. Holthausen, S. Schindler, F. Tuczek et al.
was structurally characterized, but in this case a dinuclear
complex with a CuI:ligand ratio of 2:2 was obtained.[49]
After optimizing conditions in our crystallization experi-
late ring size and a different Cu-to-ligand ratio are also re-
sponsible for significant differences in bond lengths and
angles between Ia and a binuclear CuI Schiff base complex
characterized recently by Mukherjee and co-workers.[49] Fur-
thermore, comparison of I with a related dinuclear macrocy-
clic Schiff base copper(I) complex previously described by
Utz et al. as well as by Rieger and co-workers also shows
differences in bond lengths and angles around the tetracoor-
dinate Cu3 center of I.[21,52] Most likely this is a consequence
of the macrocyclic ligand, which enforces a bowl shape of
the complex, and compared to I it has a much smaller
Cu···Cu separation of 4.250(3) Š. Dinuclear copper(I) com-
plexes of the nonmacrocyclic ligand derivative have been
described recently, but for these no hydroxylation reactions
were observed.[53]
ments we succeeded in obtaining crystals of [CuI2
ACHTREUNG
A
ACHTREUNG
CuII complex II: Oxygenation of the CuI precursor Ib in di-
chloromethane leads to the CuII complex II. Single crystals
2
suitable for X-ray analysis were obtained by diffusion of di-
ethyl ether into a solution of II in dichloromethane. The
structure of the complex is shown in Figure 2. Compound II
Figure 1. Molecular structure of the cation of I. Selected bond lengths [Š]
¯
crystallizes in the triclinic space group P1 with Z=2 and all
ꢀ
ꢀ
ꢀ
ꢀ
and angle [8]: Cu1 N3 1.85(1), Cu1 N1 1.96(9), Cu1 N2 2.055(8), Cu3
atoms in general positions. Each of the two crystallographi-
cally independent copper atoms is coordinated by two nitro-
gen atoms and one oxygen atom of the ligand as well as one
hydroxyl oxygen atom within a strongly distorted square-
planar geometry (Figure 2). There are two additional con-
tacts to oxygen atoms of perchlorate anions of 2.4497(4) Š
ꢀ
ꢀ
ꢀ
N7 1.96(1), Cu3 N6 2.02(1), Cu3 N4 2.019(9), Cu3 N5 2.11(1); Cu3-
Cu1-N1 138.9(3).
Compound Ia crystallizes in the monoclinic space group
P21 with two complex cations and four anions per unit cell.
The two copper(I) centers are bridged by a m-xylyl group
with a Cu1···Cu3 separation of 7.1(1) Š. Each copper ion is
coordinated by one imine nitrogen atom (N1/N4) and one
aliphatic amine nitrogen atom (N2/N5) per bidentate DAPA
arm. As additional coligands one or two acetonitrile mole-
cules are ligated to the copper ions. One N-bonded MeCN
molecule completes the almost trigonal-planar coordination
geometry of Cu1. With two bound acetonitrile molecules
the coordination geometry around Cu3 is best described as
distorted tetrahedral. The formation of six-membered che-
late rings leads to respective N1-Cu1-N2 and N4-Cu3-N5
angles of 96.7(3) and 95.2(3)8 which substantially deviate
from ideal trigonal or tetrahedral geometry. The dihedral
angle between the N1-Cu1-N2 and N4-Cu3-N5 planes is
ꢀ
ꢀ
(Cu1 O11) and 2.6724(3) Š (Cu2 O13). If these contacts
are taken into account the coordination around the copper
atoms can be described as strongly distorted octahedral. In
II the copper atoms are connected by the perchlorate anions
to form chains which extend in the direction of the crystallo-
graphic a axis.
ꢀ
53.098. As expected, the Cu N bond lengths of the tricoor-
dinate Cu1 center are shorter than that of Cu3 (see Support-
ing Information). The “harder” amine nitrogen atoms N2
and N5 are bound more weakly to the copper(I) ion than
the imine nitrogen atoms N1 and N4.
Feringa and co-workers have reported the synthesis and
characterization of a structurally related binuclear CuI com-
plex, in which each copper center is coordinated to one bi-
dentate ligand arm and one acetonitrile molecule.[50] Com-
parison of this crystal structure with that of Ia reveals quite
ꢀ
similar Cu N bond lengths, all in the range typical for
three-coordinate CuI complexes,[50,51] while the two com-
plexes differ in the bond angles around the copper ions be-
cause of their different donor-atom environments (aliphatic
amine versus pyridine nitrogen donor atoms). A larger che-
Figure 2. Molecular structure of the cation of II. Selected bond
ꢀ
ꢀ
ꢀ
lengths [Š] and angles [8]: Cu1 O2 1.899(3), Cu1 O1 1.986(4), Cu2 O2
ꢀ
ꢀ
ꢀ
ꢀ
1.899(4), Cu2 O1 1.991(3), Cu1 Cu2 3.0546(11), O1 C1 1.324(6), Cu1
ꢀ
ꢀ
ꢀ
N1 1.938(4), Cu1 N2 2.025(5), Cu2 N3 1.948(5), Cu2 N4 2.031(4); N1-
Cu1-N2 95.13, N3-Cu2-N4 97.8, Cu1-O1-Cu2 100.39, Cu2-O2-Cu1 107.07.
9718
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 9714 – 9729

Aromatic Hydroxylation in a Copper Bis(imine) Complex
A
FULL PAPER
Evidently the phenyl ring of the bis
A
the spectra of I and II in methanol; the spectrum of II was
obtained after bubbling O2 through a solution of I. As in the
spectrum of II in acetonitrile, an intense band is observed at
360 nm and a ligand-field band at 700 nm.
been hydroxylated on the reaction of Ib with O2 to form a
phenoxo group that bridges the two CuII centers; a second
bridge is provided by a hydroxo ligand. The m-hydroxo m-
phenoxo Cu2 unit is coordinated by the two terminal amine
and two imine nitrogen atoms of the bis
A
Vibrational spectroscopy: Infrared and Raman spectra of I
and II were obtained from solid samples at room tempera-
ture and are presented in Figure 4–7. We also studied the
18O isotopomer of II, which was prepared by reaction of I
with 18O2. Spectral assignments were facilitated by compari-
give an almost perfectly planar dinuclear complex with
square-planar coordination of each CuII center. The Cu···Cu
distance is 3.055 Š and the average Cu···N distances are
1.94 Š for the imine nitrogen atoms (N1 and N3) and
2.03 Š for the amine nitrogen atoms (N2 and N4). Similar
to the CuI complex the N1-Cu1-N2 and N3-Cu2-N4 angles
are 95.1(3) and 97.88, due to formation of a six-membered
chelate ring. As expected, the molecular structure of II is
very similar to that of this complex published by Drew
et al.[28] The Cu···Cu distance in II is slightly longer (3.055 in-
stead of 3.015 Š), most likely as a consequence of the water
molecule that is additionally coordinated to CuB in the
complex of Drew et al. Moreover, two perchlorate anions
are coordinated at about 2.5 Š (vide supra).
Spectroscopic investigations
UV/Vis spectroscopy: The UV/Vis spectrum of II in acetoni-
trile is shown in Figure 3. It exhibits two intense bands in
the region between 200 and 300 nm and one band at 357 nm
(e=10000mꢀ1 cmꢀ1) The latter feature is absent in the spec-
trum of CuI precursor I, which only shows a rising slope
below 300 nm with a couple of weak shoulders. The 357 nm
band is therefore the most conspicuous UV/Vis spectroscop-
ic signature of the oxygenated complex. It has been assigned
to a charge-transfer transition from the CuII centers to the
Figure 4. IR spectra of Ib (dotted) and II (solid).
son with computed harmonic frequencies at the UBP86/SVP
level, which show an overall pleasing agreement with the ex-
perimental results (Figure 5). In contrast to our initial ex-
pectation 16O/18O isotopic substitution in II does not lead to
the identification of a unique vibrational signature for a
hydroxylated bis
(imine) ligand.[22] At higher concentration a
A
weak absorption band can also be detected at 663 nm, which
is assigned to a ligand-field transition of the square-planar
CuII centers. Oxygenation of I to II proceeds only slowly in
acetonitrile (see below); therefore, we employed methanol
as solvent as well. Figure S1 (Supporting Information) shows
ꢀ
C O stretching vibration. Detailed analysis of the computed
spectra reveals instead that there are several normal modes
ꢀ
with varying C O stretching contributions. Here we report
Figure 3. UV/Vis spectra of Ib (dotted line, 8”10ꢀ5 molLꢀ1) and II (solid
Figure 5. Experimental (bottom) and calculated (top) IR spectra of II
(solid lines) and 18O2-II (dotted lines) with isotope-sensitive bands.
line, 11”10ꢀ5 molLꢀ1) in acetonitrile.
Chem. Eur. J. 2008, 14, 9714 – 9729
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9719

M. C. Holthausen, S. Schindler, F. Tuczek et al.
only assignments of the most prominent vibrational features
as a result of a careful correlation between measured and
computed spectra. Figure 8 shows dominant atomic contri-
butions to normal modes for the vibrations discussed below.
The IR (Figure 4 and Figure 5) and Raman (Figure 6 and
Figure 7) spectra of I and II show an intense signal at
1630 cmꢀ1 which correlates with a normal mode in the com-
puted spectra dominated by the symmetric stretching
motion of the Schiff base C=N bonds (1622 cmꢀ1, see
Figure 8); this band exhibits a second feature due the corre-
sponding antisymmetric vibration (computed at 1615 cmꢀ1,
not shown in Figure 8) with much lower intensity. On oxy-
genation of I an intense band appears in the spectrum of II
at 1568 cmꢀ1 (Figure 4). This band corresponds to a vibra-
tion at 1551 cmꢀ1 in the computed spectrum (Figure 8) and
can be assigned to an asymmetric deformation mode in the
aromatic ring in II. As consistently revealed by both experi-
Figure 6. FT Raman spectra of Ib (dotted line) and II (solid line).
ꢀ
mental and computed spectra, neither band shows any C O
participation. Buried among a series of bands between 1443
and 1394 cmꢀ1 due to C H stretching and bending vibra-
ꢀ
tions of all parts of the ligand, the computed spectra exhibit
an intense peak at 1424 cmꢀ1 (shifted to 1420 cmꢀ1 on 18O
ꢀ
substitution) that contains some C O stretching component.
In the experimental spectra this signal occurs on oxygena-
tion of I at 1450 cmꢀ1 (Figure 4) but does not show any sig-
nificant isotopic shift in the spectrum of 18O-II. One band at
1357 cmꢀ1 in the spectrum of I moves to 1339 cmꢀ1 in the
spectrum of II. Comparison of the IR spectra of 16O-II and
18O-II (Figure 5) reveals a significant 18O shift of this mode
from 1339 cmꢀ1 in 16O-II to 1329 cmꢀ1 in the spectrum of
18O-II. Almost the same isotopic shift (11 cmꢀ1) is found in
the computed spectra (1332 cmꢀ1!1321 cmꢀ1), in line with a
ꢀ
significant C O contribution to this normal mode (see
Figure 8).
A
lower-energy band at 866 cmꢀ1 shifts to
861 cmꢀ1 on isotopic substitution. This is reproduced by a
computed vibration at 863 cmꢀ1 which shifts to 857 cmꢀ1 and
involves significant O motion (see Figure 8).
Figure 7. FT Raman spectra of II (solid line) and 18O-II (dotted line).
The FT Raman spectra of I and II obtained with lexc
=
contribution in the spectral range investigated. Some other
1064 nm are shown in Figure 6. A comparison of the FT
Raman spectra of 16O-II and 18O-II (Figure 7) reveals only
small isotopic shifts. After oxygenation of I two prominent
new peaks appear at 1450 and 1257 cmꢀ1, but only the latter
vibrations show minor isotope shifts on the order of ꢀ3 to
ꢀ1
ꢀ
ꢀ5 cm , depending on the contribution of the C O stretch-
ing motion.
ꢀ
has some minor C O contribution: In the region around
Kinetic investigations: Yellow solutions of Ia in methanol
turn immediately green when exposed to dioxygen with for-
mation of II. Time-resolved spectra could be obtained by
low-temperature stopped-flow techniques; a typical example
of the oxidation reaction in methanol is shown in Figure 9.
The spectra resemble those obtained previously by some of
us in kinetic studies on the related imine systems [Cu2-
1250–1280 cmꢀ1, where normally the (phenolate) C O
ꢀ
stretching vibration is found,[16,54] only the intense peak at
1257 cmꢀ1 shifts to 1254 cmꢀ1 upon 18O substitution, and
DFT calculations indeed reveal some minor contribution
ꢀ1
ꢀ
from the C O stretch to this vibration (1249!1247 cm
;
not shown in Figure 8). Three less intense peaks with small
isotope shifts are observed at 1313 cmꢀ1 (shifting to
1308 cmꢀ1), 1330 cmꢀ1 (shifting to 1327 cmꢀ1) and 1371 cmꢀ1
(shifting to 1366 cmꢀ1). The lowest-energy band may corre-
spond to the vibration calculated at 1317 cmꢀ1, shifting to
1315 cmꢀ1 (see Figure 8).
A
(CH3CN)2]
U
U
U
(HBPB=1,3-bis[N-(2-pyridylethyl)formimidoyl]benzene;
mac=3,6,9,17,20,23-hexaazatricyclo[2.3.3.1.1]triaconta-1(29),
2,9,11(30),12(13),14,16,23,25,27-decaene).[27] Again, similar
to these systems, we could not detect the buildup of a dioxy-
To conclude, the only vibration with a strong isotope shift
gen adduct. This is in contrast to the complex [Cu2(XYL)]
reported by Karlin and co-workers, for which spectroscopic
detection of such an intermediate was possible at low tem-
A
(ꢀ10 cmꢀ1) is located at 1339 cmꢀ1 (exptl; calcd: 1332 cmꢀ1)
ꢀ
and has mostly IR intensity. It has the most prominent C O
9720
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 9714 – 9729

Aromatic Hydroxylation in a Copper Bis(imine) Complex
A
FULL PAPER
The absorbance versus time
traces could be fitted by using
one exponential or the sum of
two or three exponential func-
tions at different temperatures
(see Figure 9). Acceptable fit-
ting over a larger temperature
range was, however, not possi-
ble. Moreover, we observed a
linear dependence on dioxygen
concentration for two rate con-
stants with an intercept. This
result was not surprising insofar
as similar difficulties had been
encountered for the complexes
[Cu
G
G
AHCTREUNG
discussion on the different pos-
sible reaction pathways has
been given) and [Cu
A
A
ACHTREUNG
Figure 8. Eigenvectors of the most important vibrations of structure 3 (corresponding to complex II).
From the present findings we
can at least state that one part
of the rate law should contain
the term kobs[O2], consistent with the occurrence of a dioxy-
gen adduct as an intermediate. The reaction proved to be
faster than observed for [Cu
A
G
N
and for [Cu2(mac)(CH3CN)2](ClO4)2 and therefore required
A
N
ACHTREUNG
low-temperature stopped-flow techniques. This is under-
standable, as the ligands HBPB-H and mac stabilize the
copper(I) complexes better than DAPA.
Hydroxylation of Ib leading to II was also investigated in
acetonitrile. In this case, however, a much slower reaction
was observed. This can be attributed to the fact that acetoni-
trile binds fairly strongly to the CuI centers of Ib and must
be replaced by O2 in the course of the hydroxylation reac-
tion. Loss of acetonitrile ligands bound to I is, of course, im-
peded if the reaction is performed in this solvent. If, in con-
trast, the reaction is performed in methanol, the acetonitrile
ligands of Ia obviously are exchanged in a first reaction
step, that is, prior to binding of O2. This follows from the ob-
servation that in this solvent no isotope effect on the bind-
ing of O2 is observed (see below).
Figure 9. Spectral changes during reaction of I with dioxygen in methanol
(T=ꢀ60.18C, [complex]=0.1 mm, [O2]=4.25 mm, t=10.605 s). Inset: Ab-
sorbance versus time trace at 353 nm and fit to the sum of two exponen-
tials (k1obs =(3.1ꢂ0.2), ko2bs =(0.41ꢂ0.03) sꢀ1
.
In the course of the present study, DFT calculations indi-
cated several thermally accessible pathways from a m-h2:h2
peroxo adduct to the hydroxylated final product (see below)
involving proton-transfer steps that should show prominent
primary kinetic isotopic effects on ligand deuteration. In
order to experimentally check these predictions, further ki-
netic investigations based on deuterated DAPA complex ID
were performed. Replacement of Ib by ID was expected to
decrease the reaction rate in the presence of rate-determin-
ing H-atom or proton-transfer steps or leave it unchanged in
the absence of such reactions.[55] Importantly, the rate of the
oxygenation reaction of Ia in methanol was unchanged, that
is, no significant KIE is present (Figure 10). In acetonitrile,
on the other hand, the reaction of ID was found to be ap-
peratures.[15,16] The lack of a detectable O2 intermediate
most likely is the consequence of rate-determining forma-
tion of the reactive intermediate and much faster consecu-
tive reactions according to the general Equation (2).
k1
k2
½Cu2LŠ2þ þ O2
½Cu LðO ފ
product
ð2Þ
ƒƒ!
ƒ!
2
2
slow
fast
Immediately after the dioxygen adduct is formed it further
reacts to give the product(s) and therefore cannot be ob-
served spectroscopically. As a consequence, no kinetic data
for conversion of the peroxo complex to the hydroxylated
product complex could be obtained.
Chem. Eur. J. 2008, 14, 9714 – 9729
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9721

M. C. Holthausen, S. Schindler, F. Tuczek et al.
However, this is not observed, and this can be attributed to
the fact that this solvent coordinates more weakly than ace-
tonitrile, thus the steric congestion in the transition state
leading to the dioxygen adduct is diminished, as is accord-
ꢀ
ingly the influence of the C H stretching motion on forma-
tion of this adduct.
Quantum chemical investigations of reaction pathways: As
discussed in the Computational Methods section above, we
assume throughout this study that all elementary steps in-
volved in the reaction pathways investigated take place on a
ground-state singlet potential-energy surface. In other
words, we inherently suppose that the spin flip that evident-
ly occurs on binding of O2 in its triplet ground state to the
bare copper(I) complex occurs as part of the initial steps
leading to formation of peroxo complex 1. These steps,
which include solvent exchange at the CuI sites and poten-
tially prominent spin-flip phenomena, are not studied
here.[56] Instead we concentrate on the fate of the initially
formed peroxo Cu2O2 species 1 and the elementary steps
that are involved in reaction pathways leading to the prod-
uct of the aromatic hydroxylation (3), which corresponds to
complex II. All relevant intermediates and transition struc-
tures are compiled in Figure 12; corresponding energies are
collected in Table S1 of the Supporting Information.
Figure 10. Absorbance during oxygenation of Ia at 353 nm in methanol.
T=ꢀ60.18C; [DAPA]=0.1 mm; [[D6]DAPA]=0.1 mm; [O2]=4.25 mm;
t=10.6 s.
proximately 2.6 times faster (KIE=0.38) than of Ib
(Figure 11).
Several minima resulted from geometry optimizations per-
formed for the starting point of our quantum chemical in-
vestigation, the side-on m-h2:h2-peroxo intermediate 1; for
the sake of brevity we only consider the most stable isomers
directly involved in the reaction pathways discussed below.
Intermediate 1 exhibits significant butterfly distortion of the
Cu2O2 moiety, indicative of substantial strain introduced by
the ligand framework, which obviously does not allow for
the formation of the planar arrangement of this subunit
identified experimentally in unstrained systems.[57] Interest-
ingly, we were unable to localize a bis(m-oxo) isomer for the
N
present ligand environment, although such a species should
generally be more stable than the peroxo intermediate for
bidentate N-donor ligand environments of Cu2O2 cores such
Figure 11. Absorbance at 358 nm for Ib (solid), 1D (dotted), and Ic
(dash-dotted) during oxygenation in acetonitrile.
as that present here.[8,58] Even carefully preoptimized bis
(m-
A
oxo) structures obtained by constrained-geometry optimiza-
tions in which the ideal core substructure was kept fixed
while the rest of the structure was optimized fell back into a
peroxo structure in subsequent unconstrained optimization
runs. We take this finding as another consequence of the sig-
nificant strain introduced by the ligand framework, which is
not flexible enough to allow for the formation of a tighter
ꢀ
The lack of a primary KIE in MeOH and the observation
of an inverse deuterium effect in CH3CN are not compatible
with the presence of a rate-determining proton-transfer step
in the reaction phase after formation of the s complex. The
inverse deuterium effect in CH3CN is rather attributed to
the reaction phase before formation of the s complex, that
is, binding of O2: Due to the geometry of the complex the
bis(m-oxo) core with its closer Cu Cu contact (2.74–2.79 Š)
G
ꢀ
compared to the larger Cu Cu distance found in the peroxo
ꢀ
ꢀ
O O axis in the peroxo adduct points toward the C H
cores (3.37–3.56 Š).[57]
Starting from 1 we localized a transition structure (TS1)
bond of the aromatic ring; O2 binding thus is hindered by
ꢀ
ꢀ
the C H stretching motion of the H atom at C2 of the phen-
for O O bond cleavage that resembles to a large extent the
ꢀ
ylene spacer, and due to the smaller C D vibrational ampli-
corresponding transition structure for the peroxo/bis
A
tude O2 binding proceeds faster in the deuterated than in
the nondeuterated complex. Since O2 binding is also the
rate-limiting process in methanol, the inverse deuterium
effect should in principle appear in this solvent as well.
core isomerization identified in our previous study on the
aliphatic hydroxylation reactivity of a Cu2O2 complex bear-
ing unstrained bidentate N-donor ligands.[59] In the present
system, however, TS1 is a multicenter transition state for si-
9722
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 9714 – 9729

Aromatic Hydroxylation in a Copper Bis(imine) Complex
A
FULL PAPER
Figure 12. Reaction pathways identified for the hydroxylation reaction starting from m-h2:h2 peroxo complex 1 (Gibbs free energies at 298 K in kcalmolꢀ1
relative to 1; UB3LYP/TZVP+COSMO//UB3LYP/SVP+COSMO results).
ꢀ
ꢀ
multaneous O O bond cleavage and C O bond formation.
IRC calculations confirm direct connection of this transition
state with peroxo minimum 1 and arenium-like s complex 2,
16.3 kcalmolꢀ1 results for this step, in spite of the highly
strained nature of this transition state. Clearly this elementa-
ry step profits from significant stabilizing Coulomb interac-
tions between the transferred proton and the high negative
charge of the m-oxo atom (formally O2ꢀ). Also, using quali-
tative Hammond arguments, the large exothermicity of this
step (DGR =ꢀ87.7 kcalmolꢀ1 with respect to 2) might be
seen as another factor contributing to lowering of the barri-
er. In fact, already at this point product formation appears
feasible under the thermal conditions of the experiment:
once the barrier of 18.3 kcalmolꢀ1 connected with TS1 is
surmounted, subsequent proton transfer via TS2 should be
efficient, as its barrier is even lower, by 2 kcalmolꢀ1.
The situation is, however, more complicated than that.
The arenium intermediate 2 can actually rearrange almost
without a barrier (DGꢀ =3.9 kcalmolꢀ1) via TS3, the transi-
tion structure of a [1,2] H shift across the phenyl ring, to
form the thermodynamically rather stable dienone 4 (DGR =
ꢀ24.1 kcalmolꢀ1). Hence, any amount of the s intermediate
2 formed will immediately decay to yield the dienone rather
than passing the significantly larger barrier TS2 that would
lead to direct product formation. Decay of the dienone in-
without occurrence of an intermediate bis(m-oxo) species.
A
This electrophilic attack of the aromatic ring is connected
with a barrier of 18.3 kcalmolꢀ1, and formation of s inter-
mediate 2 is exoergic by 5.7 kcalmolꢀ1.
We identified three pathways for decay of s complex 2. A
somewhat unexpected but conceptually strikingly simple
path is direct, highly exoergic formation of the final product
of the overall reaction, that is, formation of m-hydroxo m-
phenolato complex 3 (DGR =ꢀ93.4 kcalmolꢀ1 with respect
to 1) is possible in a single step after passage of TS2. In this
transition structure the proton at the tetrahedral phenyl
carbon atom of s complex 2 is abstracted by the m-oxo atom
bridging the two copper ions. The unusually low imaginary
frequency of i206 cmꢀ1 for this proton transfer step is a con-
sequence of the participation of several atoms in the transi-
tion normal mode resulting from the excessive deformation
about the Cu2O2 moiety necessary to bend the m-oxo atom
over into bonding range with the transferred proton. Some-
what counterintuitively, however, a rather modest barrier of
Chem. Eur. J. 2008, 14, 9714 – 9729
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9723

M. C. Holthausen, S. Schindler, F. Tuczek et al.
termediate via a [1,3] H shift onto the phenolate oxygen
atom to form the more stable phenol intermediate 5 (DGR =
ꢀ10.1 kcalmolꢀ1) is unlikely to occur because of the exces-
sively high barrier connected with this step via TS4 (DGꢀ =
51.4 kcalmolꢀ1). Yet another route to phenol formation was
found via TS5, which connects s complex 2 with 5, but this
path has a much higher barrier (DGꢀ =19.8 kcalmolꢀ1) than
TS2 and is the kinetically least favorable step among the
three routes identified for the decay of 2. Thus, even though
formation of intermediate 5 would provide a strong thermo-
dynamic driving force and subsequent formation of the
product complex 3 via TS6 could occur with a moderately
low barrier of 16.3 kcalmolꢀ1, this intermediate is unlikely
to play any role in the course of the overall reaction, be-
cause unfavorably high barriers preclude its formation from
2.
With the somewhat unexpected nature of TS2 in mind we
actually located a corresponding transition structure TS9
leading directly from 4 to the final product 3, which obvi-
ously profits from the same driving force provided by the
large proton affinity of the bridging m-oxo atom. This step is
the most favorable pathway we could find for product for-
mation from dienone 4, and its barrier of 26.6 kcalmolꢀ1 ap-
pears high enough to prevent an immediate decay of the di-
enone intermediate. Because all other barriers surrounding
4 are even higher (i.e., reaction back to 2 via TS3 with
DGꢀ =28.0 kcalmolꢀ1, or phenol formation via TS4 with
DGꢀ =51.4 kcalmolꢀ1), the dienone should be thermody-
namically and kinetically stable enough to have a significant
lifetime under the experimental conditions. Because the bar-
riers in question are related to proton-transfer transition
states, these elementary steps should be subject to substan-
tial KIEs upon deuteration of the 2-position of the phenyl
ring. Indeed, H/D exchange of the respective hydrogen
atom in the optimized stationary points 4 and TS9 yields an
increase of the corresponding barrier by 1.8 kcalmolꢀ1 for
the most favorable route for dienone decay. This corre-
sponds to a classical KIE of 21.4 at 298.15 K. Qualitative
consideration of quantum mechanical tunneling[60] by em-
ploying a simple one-dimensional model[61] significantly in-
creases the predicted KIE to 27.1 (298.15 K). We thus pre-
dict a situation in which it may actually be possible to iden-
tify this species by experimental means even at room tem-
perature by employing a deuterated ligand framework.
Given the fact that it is the only relevant nonaromatic inter-
mediate in the mechanistic scenario established so far, we
suggest NMR spectroscopy as a promising tool for its exper-
imental identification.
lated product 3. With formation of two thermodynamically
unfavorable intermediates and two energetically demanding
reaction barriers, this reaction sequence can effectively be
seen as a unimolecular decomposition of 1 (assuming reac-
tion kinetics with pre-equilibrium) with an overall barrier of
44.6 kcalmolꢀ1 related to TS8. Thus, with the highest effec-
tive barrier identified in our entire study, this reaction chan-
nel is unlikely to contribute to product formation.
In summary, we identified four alternative reaction path-
ways leading from m-h2:h2 peroxo species 1 to hydroxylated
product 3. The three thermally accessible pathways all in-
volve decay of s intermediate 2. Formation of 2 requires
passing an activation barrier of 18.3 kcalmolꢀ1 via TS1 cor-
responding to electrophilic attack on the aromatic ring by
the peroxo Cu2O2 core. Intermediate 2 can rearrange almost
without activation barrier to dienone intermediate 4, which
we propose as a key intermediate in the course of the aro-
matic hydroxylation reaction. For the energetically least de-
manding route leading from 4 to product formation via TS9
we predict an effective barrier of 26.6 kcalmolꢀ1. Conse-
quently, for the overall mechanistic scenario established
here, this last proton transfer is the rate-limiting step for
product formation. The preceding electrophilic attack via
TS1 is energetically significantly less demanding, by
8.3 kcalmolꢀ1. For the rate-limiting step we predict a large
H/D KIE, which may allow future characterization of the di-
enone intermediate by NMR techniques.
Why then did the dienone intermediate escape any exper-
imental detection in our hands so far? An intuitively strik-
ing explanation for the experiments performed in methanol
is participation of this protic solvent in the proton-transfer
steps in the latter phase of the reaction sequence. In this
case intermolecular proton transfer steps will most likely
provide much lower barriers than those identified in our
computations, which were performed with a continuum sol-
vent model but without explicit consideration of solvent
molecules. But also in experiments in aprotic solvents, we
searched in vain for any trace of this intermediate. Spurred
by recent reports on the potential importance of counterions
for related systems[53,62] we investigated the influence of a
ꢀ
single ClO4 counterion on the rate-limiting step of the
mechanistic scenario suggested above. Indeed, compared to
the computed barrier height of 26.6 kcalmolꢀ1 for the step
4!TS9 a dramatically lower barrier of only 7.5 kcalmolꢀ1
results for the corresponding process 4·ClO4!TS9·ClO4
(Figure 13). This result implies a high reaction rate for this
elementary step at room temperature, which straightfor-
wardly explains the lack of any experimental evidence for
the occurrence of a stable dienone intermediate prior to
product formation or any prominent KIE in acetonitrile or
CH2Cl2.
To substantiate this quantum chemical result the hydrox-
ylation reaction was repeated in the aprotic solvent acetoni-
trile, with the CuI DAPA complex in the form of its salt
with the noncoordinating anion BPh4 (Ic). Under these con-
ditions a significantly slower reaction than with perchlorate
as counterion (Ib) was observed (see Figure 11). This
As an alternative to the reaction paths considered above,
we identified a route to product formation commencing
with initial rearrangement of the m-h2:h2 peroxo species 1 to
m-h2:h1 coordinated isomer 6, which is less stable than 1 by
6.1 kcalmolꢀ1. From 6 electrophilic attack of the aromatic
ring via TS7 leads to formation of s complex 7, a thermody-
namically highly unfavorable counterpart of 2 (less stable by
33.7 kcalmolꢀ1). From 7, proton transfer with simultaneous
ꢀ
O O bond cleavage via TS8 leads to formation of hydroxy-
9724
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 9714 – 9729

Aromatic Hydroxylation in a Copper Bis(imine) Complex
A
FULL PAPER
peroxo intermediate that subsequently attacks the aromatic
ring. In the present system, however, this pathway is associ-
ated with an overall barrier of 44.6 kcalmolꢀ1, which renders
it less likely for product formation. In this context it is inter-
esting to note that Siegbahn has suggested a related se-
ꢀ
quence (i.e., peroxide attack followed by O O bond cleav-
age) as key steps for the ortho-hydroxylation of phenolate
in the catalytic cycle of tyrosinase, but with an overall barri-
er of only 14.4 kcalmolꢀ1.[13] Alternative pathways were not
reported in this study. In contrast, all energetically favorable
pathways identified in our study involve direct decay of the
m-h2:h2 peroxo intermediate 1 without rearrangement to an
alternative peroxide coordination mode.
The energetically most favorable route commences with
direct electrophilic attack of the aromatic ring by the
peroxo moiety to form an arenium ion, which subsequently
undergoes an almost barrierless proton shift to form a
rather stable dienone intermediate (4). A second proton-
transfer step from the dienone directly leads to product for-
mation. We have identified several alternative pathways
with moderate barrier heights that would actually be ther-
mally accessible under the experimental conditions. None of
these pathways, however, can compete with the energetically
most favorable pathway, even assuming potential errors in
computed relative energies as large as 7 kcalmolꢀ1, and so
we feel safe in suggesting a single pathway as a general
mechanistic scenario for the system under study.
Figure 13. Dienone 4·ClO4 and transition structure TS9·ClO4 leading to
ꢀ
product formation optimized in the presence of a coordinating ClO4
counterion (Gibbs free energies at 298 K in kcalmolꢀ1 relative to 4·ClO4;
UB3LYP/TZVP+COSMO//UB3LYP/SVP+COSMO results).
indeed supports the presence of a rate-limiting proton-trans-
fer step leading from the dienone to the final product, which
is markedly accelerated in the presence of a coordinating
counterion.
Discussion and Conclusion
Structural, spectroscopic, kinetic, and quantum chemical in-
vestigations of the ligand hydroxylation reaction mediated
by a Cu bis
A
This scenario involves as a key intermediate in the conver-
sion of the peroxo complex to the hydroxylated product the
dienone intermediate 4, which should eventually be detecta-
ble experimentally. We attribute the fact that we (and
others) were unable so far to identify such an intermediate
to several factors: first, explicit participation of protic (or
Lewis basic) solvent molecules (which have not been consid-
ered in our calculations) in the proton transfer steps after
formation of s complex 2 may lead to drastically lowered
barriers. Second, the key proton-transfer steps may occur in
an intermolecular fashion, involving initial deprotonation of
the s complex by another copper dioxygen intermediate. A
third factor contributing to the rapid decay of the dienone
the counterion was explicitly identified in the computations:
Compared to the calculated barrier height of 26.6 kcalmolꢀ1
for the decay of 4 via TS9, a dramatically lowered barrier of
7.5 kcalmolꢀ1 was found for the corresponding process
4·ClO4!TS9·ClO4!product (3). This hypothesis is support-
from a structural analysis of the CuI complex I and the CuII
product II exhibiting a hydroxylated ligand the optical ab-
sorption and vibrational spectra were analyzed. Special at-
tention was directed to reproducing the structural and spec-
troscopic properties of complex II by quantum chemical
means. Kinetic analysis of ligand hydroxylation provided
evidence for O2 binding as the rate-limiting step in the over-
all reaction. Conversion of I to II was found to proceed
much faster in methanol than in acetonitrile, which was at-
tributed to the fact that the acetonitrile ligands of the CuI
precursor I are displaced in the course of O2 binding. More-
over, an inverse KIE was evidenced for the reaction in ace-
tonitrile, which was rationalized by a sterically congested
transition state leading to the peroxo adduct. In methanol,
however, no KIE was observed. Finally, a DFT analysis of
the oxygenation reaction demonstrated that the dominant
barrier after O2 binding is represented by electrophilic
attack of m-h2:h2 peroxo intermediate 1 on the arene ring.
Nevertheless, on the basis of the activation energy (18.3 kcal
molꢀ1) this reaction is thermally allowed, in agreement with
the experimental observation. Given that the B3LYP func-
tional employed has been shown to overestimate barrier
heights by 4–7 kcalmolꢀ1,[63] the actual barrier height may be
in the range of 10–15 kcalmolꢀ1.
ed by kinetic measurements on the BPh4 salt of the CuI2
ꢀ
DAPA complex (Ic), for which a significantly lower reaction
rate was found compared to corresponding perchlorate salt
Ib. To directly detect dienone intermediate 4 further experi-
mental studies must be performed. The present system, how-
ever, appears less suitable for these investigations, as bind-
ing of O2 is also associated with a thermal barrier which
probably precludes accumulation of a dienone species at
low temperatures.
The DFT investigation of the reactivity of the CuII peroxo
intermediate 1 formed by reaction of the CuI precursor and
O2 indicated the presence of four pathways to the hydroxy-
lated final product 3. One of the pathways studied involves
rearrangement of the m-h2:h2 peroxo structure to a m-h2:h1
Another notable and somewhat surprising result of our
quantum chemical analysis is the finding that no stable bis-
(m-oxo) isomer is formed in the ligand environment studied
Chem. Eur. J. 2008, 14, 9714 – 9729
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9725

M. C. Holthausen, S. Schindler, F. Tuczek et al.
here. Based on the experience gathered in related investiga-
tions by others, we would actually expect preferential forma-
tion of a bis(m-oxo) species within the chelating bidentate
G
N-donor ligand environment of the system studied here.
Yet, the initial steps take place without intermediate forma-
tion of a bis(m-oxo) species. While the initial transition state
U
ꢀ
TS1 exhibits structural features nearly identical to O O
bond-breaking transition states leading to the formation of a
bis
(m-oxo) species in related ligand environments,[59] in the
U
ꢀ
present case it is a multicenter transition state for O O
bond cleavage and simultaneous s attack of the aromatic
ring by the peroxo moiety. We view this finding as a conse-
quence of the rather tight coordination sphere of the bis-
ꢀ
A
Cu2O2 subunit in TS1 into close proximity to the aromatic
ring, ideally oriented for direct orbital interactions with the
aromatic p system to generate s complex 2 in a single step.
The mechanistic scenario established here has implica-
tions for the ligand hydroxylation reaction occurring in the
prototype binuclear copper complex [Cu2(XYL)] of Karlin
A
et al.[15] on reaction with O2. In full agreement with our pres-
ent results, Pidcock et al. implied a m-h2:h2 peroxo complex
as active species and excluded the occurrence of a corre-
sponding bis(m-oxo) species in a spectroscopic analysis of
N
this reaction.[16] They rationalized the electrophilic attack of
the Cu2O2 moiety on the aromatic ring in terms of qualita-
tive frontier orbital arguments. Within this picture, it was
convincingly argued that the HOMO of the arene system
overlaps with an unoccupied orbital of the electrophile,
*
which can either be the ps or the s* orbital of the side-on
Figure 14. Geometries in the transition states leading to the s complex.
a) Complex of Karlin et al. b) Bis
(imine) complex (see text).[16]
A
peroxo dicopper core (Figure 14a). Based on the particulari-
ties of the coordination geometry present in the tridendate
ligand environment in that study, an attack via the s* perox-
ide orbital was considered unlikely, and a preferred, symme-
formation of a dienone as a key intermediate, in analogy to
the mechanistic scenario developed here. A subsequent
*
try allowed pathway involving the ps orbital was suggested.
ꢀ
ꢀ
The geometry of the initial transition state TS1 optimized
C N cleavage, in analogy to the cleavage of the methylene
(amine) bond in the XYL system, would, however, appear
unlikely in a bis(imine) complex due to the double-bond
for the copper bis
A
N
ACHTREUNG
sentially from the assumed transition state geometry dis-
ACHTREUNG
ꢀ
cussed by Pidcock et al., and it is in fact the s* orbital of
character of the C N linkage.
*
the peroxo group, and not the ps orbital, that is involved in
The present results are also relevant to the enzyme tyrosi-
nase, specifically the ortho-hydroxylation of tyrosine mediat-
ed by the oxy form of this enzyme. Of crucial importance
for this reaction is the orientation of phenolic substrates
with respect to the binuclear copper active site. Experimen-
tal information on this point has mostly been derived from
spectroscopic studies on the bonding of inhibitors to tyrosi-
nase.[65] Alternatively, it has been proposed that an external
tyrosine substrate is oriented at the active site of Ty in the
same way as Phe49 in the Limulus oxy Hc structure.[66] The
recently solved structure of Streptomyces castaneoglobispo-
rus tyrosinase revealed a very similar arrangement with
Tyr98 (provided by the associated caddie protein) extending
into the active-site pocket like a potential substrate.[1,2] An
external phenolic substrate may be preoriented at the active
site in a similar geometry (Scheme 2). Then a slight shift to-
wards CuA is necessary to bind in the position trans to
His63, after which hydroxylation of the aromatic ring can
the electrophilic attack on the arene ring. In particular, the
rather rigid ligand framework studied in the present work
ꢀ
does not allow tilting of the aromatic ring along the O O
axis of the Cu2O2 group in the course of this elementary
step, which would indeed exclude any constructive orbital
interaction between the s* peroxo orbital and the aromatic
p system. Instead the arene ring and the Cu2O2 group rotate
ꢀ
in the transition state about axes perpendicular to the O O
vector such that the s* orbital now interacts with the aro-
matic p system from below in a symmetry-allowed fashion
that mediates the initial step along the hydroxylation path-
way (Figure 14b).
Another interesting implication for the mechanism of the
XYL system relates to the observation of an NIH shift in
studies employing a modified xylyl ligand that was methylat-
ed at the 2-position of the aromatic core.[64] This observation
could in fact straightforwardly be rationalized by assuming
9726
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 9714 – 9729

Aromatic Hydroxylation in a Copper Bis(imine) Complex
A
FULL PAPER
Scheme 2. Hydroxylation/oxidation of monophenolic substrates in tyrosinase from Streptomyces castaneoglobisporus (adapted from ref. [2], see text).
occur. Importantly, the proximity of the ortho position of
the phenolic ring to the side-on-coordinated peroxo group
enables electrophilic attack of the Cu2O2 moiety on the aro-
the wide occurrence of oxygenation reactions in binuclear
copper–dioxygen systems exhibiting widely different ligand
structures and explains why these reactions are not limited
to a few systems with special geometries.
matic ring.[2,16] To this end, the O O axis of the peroxo
ꢀ
ligand may rotate to point towards the substrate
(Scheme 2). This geometry would correspond to the transi-
tion state TS1 in our complex leading to s complex 2 and
would involve an orbital interaction between the peroxide
s* orbital and the HOMO of the aromatic ring, in analogy
to our model system.
Acknowledgement
F.T. thanks CAU Kiel and COST D21 for financial support and F. Studt
for preliminary calculations. O.S. acknowledges help of U. Cornelissen
with the Raman and IR measurements. Computer time provided by the
CSC Frankfurt, the HHLR Darmstadt, and the HRZ Marburg is grate-
fully acknowledged.
The analysis of the hydroxylation pathway in the Cu2 bis-
(imine) complex thus has shown that besides the previously
*
discussed ps pathway a second pathway for the oxygenation
chemistry mediated by binuclear copper site does exist, that
is, electrophilic attack on the substrate by the s* orbital of
side-on bound peroxide. For systems exhibiting more struc-
[1] Y. Matoba, T. Kumagai, A. Yamamoto, H. Yoshitsu, M. Sugiyama, J.
Biol. Chem. 2006, 281, 8981.
[2] H. Decker, T. Schweikhardt, F. Tuczek, Angew. Chem. 2006, 118,
4658; Angew. Chem. Int. Ed. 2006, 45, 4546.
tural flexibility than the Cu bis(imine) systems, both path-
U
ways may be operative. Importantly, the presence of a
second orbital pathway provides additional flexibility for the
position of the aromatic substrate with respect to the Cu2O2
unit in the transition state leading to the s complex. This is
of relevance both for the enzyme tyrosinase and for copper
model systems which are active in the hydroxylation of ex-
ternal substrates.[67] In all of these cases the s* orbital direct-
[3] E. I. Solomon, U. M. Sundaram, T. E. Machonkin, Chem. Rev. 1996,
96, 2563.
[4] A. Sanchez-Ferrer, J. N. Rodríguez-López, F. García-Cµnovas, F.
García-Carmona, Biochim. Biophys. Acta 1995, 210, 1.
[5] C. Eicken, C. Gerdemann, B. Krebs, Handbook of Metalloproteins,
Vol. 2 (Eds.: A. Messerschmidt, R. Huber, T. Poulos, K. Wieghardt),
Wiley, New York, 2001, pp. 1319.
[6] M, Tr›moli›res, J. B. Bieth, Phytochem. 1984, 23, 501.
[7] a) K. E. van Holde, K. I. Miller, H. Decker, J. Biol. Chem. 2001, 276,
15563; b) H. Decker in Encyclopedia of Inorganic Chemistry, Vol. 2,
2nd ed. (Ed.: R. B. King), Wiley, New York, 2006, pp. 1159–1173.
[8] E. A. Lewis, W. B. Tolman, Chem. Rev. 2004, 104, 1047.
[9] L. Li, A. N. Sarjeant, K. D. Karlin, Inorg. Chem. 2006, 45, 7160.
ꢀ
ed along the prolonged O O vector just has to “hit” the p
system of the substrate (except for the nodal plane of the C
orbitals) from an arbitrary angle to mediate hydroxylation.
This structural flexibility in the transition state accounts for
Chem. Eur. J. 2008, 14, 9714 – 9729
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9727

M. C. Holthausen, S. Schindler, F. Tuczek et al.
[10] D. Ghosh, R. Mukherjee, Inorg. Chem. 1998, 37, 6597.
[11] S. B. Foxon, D. Utz, J. Astner, S. Schindler, F. Thaler, F. W. Heine-
mann, G. Liehr, J. Mukherjee, V. Balamurugan, D. Ghosh, R. Mu-
kherjee, Dalton Trans. 2004, 2321
[12] P. E. M. Siegbahn, M. Wirstam, J. Am. Chem. Soc. 2001, 123, 11819.
[13] P. E. M. Siegbahn, J. Biol. Inorg. Chem. 2003, 8, 567.
[14] S. Itoh, M. Taki, H. Nakao, P. L. Holland, W. B. Tolman, L. Que, Jr.,
S. Fukuzumi, Angew. Chem. 2000, 112, 409; Angew. Chem. Int. Ed.
2000, 39, 398.
[15] a) K. D. Karlin, P. L. Dahlstrom, S. N. Cozzette, P. M. Scensny, J. Zu-
bieta, J. Chem. Soc. Chem. Commun. 1981, 881; b) K. D. Karlin,
J. C. Hayes, Y. Gultneh, R. W. Cruse, J. W. McKown, J. P. Hultchin-
son, J. Zubieta, J. Am. Chem. Soc. 1984, 106, 2121–2128.
[16] E. Pidcock, H. V. Obias, C. X. Zhang, K. D. Karlin, E. I. Solomon, J.
Am. Chem. Soc. 1998, 120, 7841.
G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R.
Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao,
H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian, J. B. Cross,
C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev,
A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, P. Y. Ayala, K.
Morokuma, G. A. Voth, P. Salvador, J. J. Dannenberg, V. G. Zakr-
zewski, S. Dapprich, A. D. Daniels, M. C. Strain, O. Farkas, D. K.
Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V. Ortiz,
Q. Cui, A. G. Baboul, S. Clifford, J. Cioslowski, B. B. Stefanov, G.
Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. L. Martin, D. J. Fox,
T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, M. Challa-
combe, P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, C. Gonza-
lez, J. A. Pople, Gaussian, Inc, Pittsburgh, PA, 2003.
[38] L. Noodleman, T. Lovell, W.-G. Han, J. Li, F. Himo, Chem. Rev.
2004, 104, 459.
[17] a) L. M. Mirica, M. Vance, D. Jackson Rudd, B. Hedman, K. O.
Hodgson, E. I. Solomon, T. D. P. Stack, Science 2005, 308, 1890;
b) L. M. Mirica, D. J. Rudd, M. A. Vance, E. I. Solomon, K. O.
Hodgson, B. Hedman, T. D. P. Stack, J. Am. Chem. Soc. 2006, 128,
2654; c) A. Company, S. Palavicini, I. Garcia-Bosch, R. Mas-Bal-
lestØ, L. Que, Jr., E. Rybak-Akimova, L. Casella, X. Ribas, M.
Costas, Chem. Eur. J. 2008, 14, 3535.
[39] a) L. Noodleman, J. Chem. Phys. 1981, 74, 5737; b) L. Noodleman,
E. J. Baerends, J. Am. Chem. Soc. 1984, 106, 2316; c) L. Noodleman,
E. R. Davidson, Chem. Phys. 1986, 109, 131; d) L. Noodleman, D. A.
Case, Adv. Inorg. Chem. 1992, 38, 423; e) L. Noodleman, J. Li, X.
-G. Zhao, W. H. Richardson in Density-functional Methods in
Chemistry and Materials (Ed.: M. Springborg), Wiley, New York,
1997.
[18] P. L. Holland, K. R. Rodgers, W. B. Tolman, Angew. Chem. 1999,
111, 1210; Angew. Chem. Int. Ed. 1999, 38, 1139.
[40] a) K. Yamaguchi, Y. Takahara, T. Fueno in Applied Quantum
Chemistry (Eds.: V. H. Smith, Jr., H. F. Schaefer III, K. Morokuma),
Reidel, Dordrecht, 1986, p. 155. b) T. Soda, Y. Kitagawa, T. Onishi,
Y. Takano, Y. Shigeta, H. Nagao, Y. Yoshioka, K. Yamaguchi,
Chem. Phys. Lett. 2000, 319, 223–230.
[19] B. L. Feringa in Bioinorganic Chemistry of Copper (Eds.: K. Karlin,
Z. Tyeklµr), Chapman and Hall, New York, 1993, pp. 306–324.
[20] P. Amudha, P. Akilan, M. Kandaswamy, Polyhedron 1999, 18, 1355.
[21] D. Utz, F. W. Heinemann, F. Hampel, D. T. Richens, S. Schindler,
Inorg. Chem. 2003, 42, 1430.
[22] L. Casella, M. Gullotti, G. Pallanza, L. Rigoni, J. Am. Chem. Soc.
1988, 110, 4221.
[23] S. Ryan, H. Adams, D. E. Fenton, M. Becker, S. Schindler, Inorg.
Chem. 1998, 37, 2134–2140.
[24] R. Menif, A. E. Martell, P. J. Squattrito, A. Clearfield, Inorg. Chem.
1990, 29, 4723–4729.
[25] G. J. Kubas, B. Monzyk, A. L. Crumbliss, Inorg. Synth. 1979, 19, 90.
[26] S. V. Kryatov, E. V. Rybak-Akimova, S. Schindler, Chem. Rev. 2005,
105, 2175.
[27] M. Weitzer, M. Schatz, F. Hampel, F. W. Heinemann, S. Schindler,
Dalton Trans. 2002, 686.
[28] M. G. B. Drew, J. Trocha-Grimshaw, K. P. McKillop, Polyhedron
1989, 8, 2513.
[41] F. Neese, J. Phys. Chem. Solids 2004, 65, 781.
[42] a) T. Saito, Y. Kitagawa, M. Shoji, Y. Nakanishi, M. Ito, T. Kawaka-
mi, M. Okumura, K. Yamaguchi, Chem. Phys. Lett. 2008, 456, 76;
b) I. de P. R. Moreira, R. Costa, M. Filatov, F. Illas, J. Chem. Theory
Comput. 2007, 3, 764; c) F. Illas, I. de P. R. Moreira, J. M. Bofill, M.
Filatov, Theor. Chem. Acc. 2006, 116, 587; d) I. Ciofini, C. A. Daul,
Coord. Chem. Rev. 2003, 238–239, 187; e) A. Rodríguez-Fortea, P.
Alemany, S. Alvarez, E. Ruiz, Inorg. Chem. 2002, 41, 3769; f) C.
Despanches, E. Ruiz, A. Rodríguez-Fortea, S. Alvarez, J. Am.
Chem. Soc. 2002, 124, 5197; g) J. Cano, E. Ruiz, P. Alemany, F.
Lloret, S. Alvarez, J. Chem. Soc. Dalton Trans. 1999, 1669; h) E.
Ruiz, J. Cano, S. Alvarez, P. Alemany, J. Am. Chem. Soc. 1998, 120,
11122; i) F. Bernardi, A. Bottoni, R. Casadio, P. Fariselli, A. Rigo,
Inorg. Chem. 1996, 35, 5207; j) A. Bencini, D. Gatteschi J. Am.
Chem. Soc. 1986, 108, 5763.
[29] The B3LYP/G functional corresponds to that implemented in the
Gaussian 03 program: a) A. D. Becke, J. Chem. Phys. 1993, 98, 5648;
b) A. D. Becke, Phys. Rev. A 1988, 38, 3098; c) C. Lee, W. Yang,
R. G. Parr, Phys. Rev. B 1988, 37, 785; d) P. J. Stephens, F. J. Devlin,
C. F. Chabalowski, M. J. Frisch, J. Phys. Chem. 1994, 98, 11623.
[30] F. Neese, ORCA—An ab initio, DFT and semiempirical SCF-MO
package, Version 2. 5. 00, Universität Bonn, Germany, 2006.
[31] A. Schaefer, H. Horn, R. Ahlrichs, J. Chem. Phys. 1992, 97, 2571.
[32] a) F. Neese, J. Comput. Chem. 2003, 24, 1740; b) F. Neese, G. Ol-
brich, Chem. Phys. Lett. 2002, 362, 170.
[43] M. C. Holthausen J. Comput. Chem. 2005, 26, 1505.
[44] a) S. Shaik, D. Danovich, A. Fiedler, D. Schrçder, H. Schwarz, Helv.
Chim. Acta 1995, 78, 1393; b) D. Schrçder, S. Shaik, H. Schwarz,
Acc. Chem. Res. 2000, 33, 139; c) H. Schwarz, Int. J. Mass Spectrom.
2004, 237, 75.
[45] a) D. Schrçder, M. C. Holthausen, H. Schwarz, J. Phys. Chem. B
2004, 108, 14407; b) J. N. Harvey, Phys. Chem. Chem. Phys. 2007, 9,
331.
[46] a) A. D. Becke, Phys. Rev. A 1988, 38, 3098; b) J. B. Perdew, Phys.
Rev. B 1986, 33, 8822.
[33] a) K. Eichkorn, O. Treutler, H. Ohm, M. Häser, R. Ahlrichs, Chem.
Phys. Lett. 1995, 240, 283; b) K. Eichkorn, F. Weigend, O. Treutler,
R. Ahlrichs, Theor. Chem. Acc. 1997, 97, 119.
[47] M. Schatz, V. Raab, S. P. Foxon, G. Brehm, S. Schneider, M. Reiher,
M. C. Holthausen, J. Sundermeyer, S. Schindler, Angew. Chem.
2004, 116, 4460; Angew. Chem. Int. Ed. 2004, 43, 4360.
[34] R. A. Kendall, H. A. Früchtl, Theor. Chem. Acc. 1997, 97, 158.
[35] a) A. Klamt, G. Schürmann, J. Chem. Soc. Perkin Trans. 2 1993, 799;
b) A. Klamt in Encylopedia of Computational Chemistry (Eds.: N. L.
Allinger, T. Clark, J. Gasteiger, P. Kollmann, H. F. Schaefer III,
P. von R. Schleyer, P. R. Schreiner), Wiley, Chichester, 1998, p. 604.
[36] We successfully repeated some of the transition-state searches with
identical results (to numerical accuracy) with the extended geometry
optimization routines implemented in the most recent ORCA Ver-
sion 2.6.35, which was made available to us by the Bonn developer
group after this project was about to be finished.
[48] Compare Chapter 8 in W. Koch, M. C. Holthausen, A Chemistꢀs
Guide to Density Functional Theory, 2nd ed., Wiley-VCH, Wein-
heim, 2001.
[49] S. Mandal, R. Mukherjee, Inorg. Chim. Acta 2006, 359, 4019.
[50] O. J. Gelling, F. v. Bolhuis, A. Meetsma, B. L. Feringa, J. Chem. Soc.
Chem. Commun. 1988, 552.
[51] K. D. Karlin, Y. Gultneh, J. P. Hutchinson, J. Zubieta, J. Am. Chem.
Soc. 1982, 104, 5240.
[52] H. Ma, M. Allmendinger, U. Thewalt, A. Lentz, M. Klinga, B.
Rieger, Eur. J. Inorg. Chem. 2002, 11, 2857–2867.
[37] Gaussian 03, Revision B3, M. J. Frisch, G. W. Trucks, H. B. Schlegel,
G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. J. A. Montgomery,
T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J.
Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega,
[53] A. Company, L. Gomez, R. Mas-Balleste, I. V. Korendovych, X.
Rabas, A. Poater, T. Parella, X. Fontodora, J. Benet-Buchholz, M.
Sola, L. Que, Jr., E. V. Rybak-Akimova, M. Costas, Inorg. Chem.
2007, 46, 4997.
9728
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 9714 – 9729

Aromatic Hydroxylation in a Copper Bis(imine) Complex
A
FULL PAPER
[54] J. W. Pyrz, K. D. Karlin, T. N. Sorrell, G. C. Vogel, L. Que, Jr., Inorg.
Chem. 1984, 23, 4581.
Karlin, E. I. Solomon, J. Am. Chem. Soc. 2003, 125, 5186; h) T. D. P.
Stack, Dalton Trans. 2003, 881.
[59] P. Spuhler, M. C. Holthausen, Angew. Chem. 2003, 115, 6143;
Angew. Chem. Int. Ed. 2003, 42, 5961.
[55] S. Mahaptra, J. A. Halfen, W. B. Tolman, J. Am. Chem. Soc. 1996,
118, 11575.
[60] R. Berger, Angew. Chem. 2004, 116, 402; Angew. Chem. Int. Ed.
2004, 43, 398.
[56] a) M. Metz, E. I. Solomon, J. Am. Chem. Soc. 2001, 123, 4938;
b) E. I. Solomon, P. Chen, M. Metz, S. -K. Lee, A. E. Palmer,
Angew. Chem. 2001, 113, 4702; Angew. Chem. Int. Ed. 2001, 40,
4570; c) Y. Takano, K. Yamaguchi, Int. J. Quant. Chem. 2007, 107,
3103; d) P. Chen, D. E. Root, C. Campochiaro, K. Fujisawa, E. I. So-
lomon, J. Am. Chem. Soc. 2003, 125, 466; e) C. J. Cramer, W. B.
Tolman, K. H. Theopold, A. L. Rheingold, Proc. Natl. Acad. Sci.
USA 2003, 100, 3635; f) D. A. Pantazis, J. W. McGrady, Inorg. Chem.
2003, 42, 7734; g) C. J. Cramer, W. B. Tolman, Acc. Chem. Res. 2007,
40, 601; h) B. F. Gherman, C. J. Cramer, Inorg. Chem. 2004, 43,
7281.
[61] R. T. Skodje, D. G. Truhlar, J. Phys. Chem. 1981, 85, 624.
[62] X. Ottenwaelder, D. J. Rudd, M. C. Corbett, K. O. Hodgson, B.
Hedman, T. D. P. Stack, J. Am. Chem. Soc. 2006, 128, 9268.
[63] S. S. Yi, E. L. Reichert, M. C. Holthausen, W. Koch, J. C. Weisshaar,
Chem. Eur. J. 2000, 6, 2232.
[64] a) K. D. Karlin, B. I. Cohen, R. R. Jacobson, J. Zubieta, J. Am.
Chem. Soc. 1987, 109, 6194; b) M. S. Nasir, B. I. Cohen, K. D.
Karlin, J. Am. Chem. Soc. 1992, 114, 2482.
[65] a) R. S. Himmelwright, N. C. Eickman, C. D. LuBien, E. I. Solomon,
K. Lerch, J. Am. Chem. Soc. 1980, 102, 7339; b) A. W. Tepper, L.
Bubacco, G. W. Canters, J. Am. Chem. Soc. 2005, 127, 567.
[66] a) H. Decker, R. Dillinger, F. Tuczek, Angew. Chem. 2000, 112,
1656; Angew. Chem. Int. Ed. 2000, 39, 1591; b) H. Decker, F.
Tuczek, Trends Biochem. Sci. 2000, 25, 392.
[67] a) L. Casella, M. Gulotti, R. Radaelli, P. Di Gennaro, J. Chem. Soc.
Chem. Commun. 1991, 22, 1611; b) L. Santagostini, M. Gullotti, E.
Monzani, L. Casella, R. Dillinger, F. Tuczek, Chem. Eur. J. 2000, 6,
519; c) G. Battaini, M. De Carolis, E. Monzani, F. Tuczek, L. Case-
lla, Chem. Commun. 2003, 726; d) G. Battaini, E. Monzani, A. Per-
otti, C. Para, L. Casella, L. Santagostini, M. Gullotti, R. Dillinger,
C. Näther, F. Tuczek, J. Am. Chem. Soc. 2003, 125, 4185.
[57] a) N. Kitajima, K. Fujisawa, Y. Moro-Oka, K. Toriumi, J. Am.
Chem. Soc. 1989, 111, 8975; b) L. Mirica, X. Ottenwaelder, D. T. P.
Stack, Chem. Rev. 2004, 104, 1013.
[58] a) W. B. Tolman, Acc. Chem. Res. 1997, 30, 227; b) V. Mahadevan,
Z. Hou, A. P. Cole, D. E. Root, T. K. Lal, E. I. Solomon, T. D. P.
Stack, J. Am. Chem. Soc. 1997,119 11996; c) H. V. Obias, Y. Lin,
N. N. Murthy, E. Pidcock, E. I. Solomon, M. Ralle, N. J. Blackburn,
Y. -M. Neuhold, A. D. Zuberbühler, K. D. Karlin, J. Am. Chem. Soc.
1998, 120, 12960; d) E. Pidcock, S. DeBeer, H. V. Obias, B. Hedman,
K. O. Hodgson, K. D. Karlin, E. I. Solomon, J. Am. Chem. Soc.
1999, 121, 1870; e) V. Mahadevan, M. J. Henson, E. I. Solomon,
T. D. P. Stack, J. Am. Chem. Soc. 2000, 122, 10249; f) M. Taki, S. Ter-
amae, S. Nagatomo, Y. Tachi, T. Kitagewa, S. Itoh, S. Fukuzumi, J.
Am. Chem. Soc. 2002, 124, 6367; g) M. J. Henson, C. Liang, K. D.
Received: April 25, 2008
Published online: September 11, 2008
Chem. Eur. J. 2008, 14, 9714 – 9729
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9729