Metathesis-like splitting reactions of metallated [36]octaphyrins (1.1.1.1.1.1.1.1): Experimental and computational investigations

Article abstract of DOI:10.1002/chem.200900454

meso-Octakis(pentafluoro-phenyl)-substituted [36]octaphyr-in(1.1.1.1.1.1.1. 1) (1) is a figure-of-eight nonaromatic macrocycle that serves as a unique platform to induce a metathesis-like splitting reaction upon bis-Cu^II metalation. To get a be

Full text of DOI:10.1002/chem.200900454

DOI: 10.1002/chem.200900454
Metathesis-like Splitting Reactions of Metallated
[36]Octaphyrins(1.1.1.1.1.1.1.1): Experimental and Computational
Investigations
Yasuo Tanaka, Hiroshi Shinokubo, Yosuke Yoshimura, and Atsuhiro Osuka*[a]
5674
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 5674 – 5685

FULL PAPER
À
Abstract:
phenyl)-substituted
meso-Octakis(pentafluoro-
[36]octaphyr-
Activation parameters of the splitting
subsequent C21 C40 bond formation
reactions were determined: Ea =
to give a spirocyclobutane intermediate
(INT2), followed by a radical reverse
cycloaddition reaction to give two met-
alloporphyrins. Inherent instability of
1–CuCu, which may arise from its
strongly distorted structure, was indi-
cated to be a main factor in the smooth
splitting reaction. Finally, a new bis-
PdII complex (5–PdPd) was isolated in
in(1.1.1.1.1.1.1.1) (1) is a figure-of-eight
nonaromatic macrocycle that serves as
a unique platform to induce a metathe-
sis-like splitting reaction upon bis-CuII
metalation. To get a better understand-
ing of this splitting reaction, we exam-
ined the metalation of 1 with several
metal ions. In contrast with the smooth
and quantitative splitting reaction of
bis-CuII complex 1–CuCu, free-base 1,
mono-CuII complex 1–Cu, and bis-ZnII
complex 1–ZnZn do not undergo the
splitting reaction. Mono-CoII complex
1–Co was selectively produced from
metalation with the CoII ion, from
which hybrid complex 1–CoCu was
synthesized. The hybrid complex 1–
CoCu undergoes the splitting reaction
to give 2–Co and 2–Cu quantitatively.
104 kJmolÀ1,
DH° =101 kJmolÀ1,
DS° =À25.0 JmolÀ1 KÀ1, and DV° =
18 cm3 molÀ1 for 1–CuCu and Ea =
105 kJmolÀ1, DH° =102 kJmolÀ1, and
DS° =À29.9 JmolÀ1 KÀ1 for 1–CoCu. A
marked difference between the split-
ting reaction reactivity of 1–CuCu and
1–ZnZn has been examined by DFT
calculations at the B3LYP/631SDD//
B3LYP/LANL2DZ level, which re-
vealed that the reaction proceeds
through a stepwise route involving ini-
the metalation of 1 with PdACTHNUTRGENUN(G OAc)2 in a
9:1 mixture of 2,2,2-trifluoroethanol
and methanol as a manifestation of the
transannular electronic interaction in
metalated octaphyrin complexes. Col-
lectively, these results underscore the
importance of the transannular elec-
tronic interactions that are enhanced
by metalation, depending upon the co-
ordinated metal ions.
À
tial C1 C20 bond formation to give
INT1 as the rate-determining step and
Keywords: metalation
noids reaction mechanisms
splitting reaction
· porphyri-
·
·
Introduction
rangement to a bis-NiII complex of spirodicorrole through
the extrusion of two carbon dioxide molecules.[11] They also
reported that the PdII metalation of [34]octaphyr-
in(1.1.1.0.1.1.1.0) caused the formation of a bis-PdII complex
of a quite unprecedented bis-spirodiporphyrin skeleton,
again through the transannular interaction.[11]
In recent years, increasing attention has been focused on ex-
panded porphyrins[1] that are pyrrolic conjugated macrocy-
cles larger than porphyrin because of their large conjugated
p systems,[2] multiple oxidation states,[3] conformational flexi-
bilities,[4] anion-binding abilities,[5] and novel aromatic char-
acters.[3,6,7] Aside from these intriguing properties, the metal-
ation of expanded porphyrins is quite effective for produc-
ing unique metal complexes of novel structures and proper-
ties, as illustrated by the interesting examples, such as the
bis-CuII complex of amethyrin,[8] the bis-AuIII complexes of
[26]hexaphyrin(1.1.1.1.1.1),[3b,d] the heterometallic complexes
of [40]nonaphyrin(1.1.1.1.1.1.1.1.1),[9] and Mçbius aromatic–
metal complexes.[10]
With increasing molecular size, expanded porphyrins tend
to take distorted structures, such as figure-of-eight confor-
mations or helical structures as a consequence of a fully con-
jugated and cyclic structure.[2,9] In some cases, such distorted
molecular systems lead to unique irreversible chemical reac-
tions aided by the transannular interactions. Vogel et al.
found that the metalation of 5,24-dioxo-octaphyr-
in(1.1.1.0.1.1.1.0) with NiII ions caused an unexpected rear-
We have previously reported the synthesis of a series of
meso-pentafluorophenyl-substituted expanded porphyrins
under the modified Rothemund–Lindsey protocol for por-
phyrin synthesis.[12] This synthetic method allowed us to ex-
plore the new chemistry of these expanded porphyrins. For
instance, we found unexpected cross-bridging reactions of
5,20-bis-ethynyl-substituted [26]hexaphyrins to provide vi-
nylene-bridged hexaphyrins[13a] and the [3+2] annulation re-
action of 5-aryl-20-ethynyl-substituited [26]hexaphyrins to
provide 1,3-indenylene-bridged hexaphyrins.[13b] These two
reactions are apparently triggered by the transannular inter-
actions, which are enforced during dynamic conformational
changes.[13] As a more surprising reaction, we found the
clean splitting reaction of the bis-CuII complex of meso-pen-
tafluorophenyl-substituted [36]octaphyrin(1.1.1.1.1.1.1.1) (1–
CuCu) that quantitatively produced two molecules of CuII–
porphyrins 2–Cu.[14] This reaction mode is quite unprece-
dented because it requires the rupture and formation of two
carbon–carbon double bonds similar to a metathesis reac-
tion. Another interesting example is the extrusion of meso-
pentafluorophenyl-substituted BIII-subporphyrin from the
corresponding [32]heptaphyrin(1.1.1.1.1.1.1) upon treatment
of its mono-CuII complex with the BIII ion.[15] These two re-
actions are also initiated by the transannular electronic in-
teractions in distorted expanded porphyrins.
[a] Y. Tanaka, Prof. H. Shinokubo, Prof. Y. Yoshimura, Prof. A. Osuka
Department of Chemistry
Graduate School of Science
Kyoto University
Sakyo-ku, Kyoto 606-8502 (Japan)
Fax : (+81)75-753-3907
E-mail: osuka@kuchem.kyoto-u.ac.jp
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200900454.
Chem. Eur. J. 2009, 15, 5674 – 5685
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5675

A. Osuka et al.
Herein, we examine the ther-
mal splitting reactions of metal
complexes of 1. Among these,
only bis-CuII complex 1–CuCu
and hybrid complex 1–CoCu
undergo the splitting reaction.
To determine the reaction pro-
file, the activation parameters
(Ea, DH°, DS°, and DV°) of the
splitting reactions were deter-
mined and discussed. Theoreti-
cal calculations were also per-
formed for the splitting reac-
tions of 1–CuCu and 1–ZnZn at
the
LANL2DZ level. Finally, in the
metalation with Pd(OAc)2, we
B3LYP/631SDD//B3LYP/
ACHTUNGTRENNUNG
isolated an interesting complex,
À
5–PdPd, which has C1 C20 and
À
C2 N5 (pyrrole E) bonds as
evidence for the effective trans-
annular interactions. Collective-
ly, these results provide valua-
ble insight into the molecular
mechanism of the splitting reac-
tion.
Results and Discussion
Free-base
octaphyrin:
The
solid-state structure of octa-
phyrin 1 exhibits a figure-of-
eight conformation with C2
symmetry (Figure 1a).[16] The
two hemi-porphyrin-like seg-
ments are roughly planar and Figure 1. Crystal structures of a) 1, b) 1–CuCu, c) 1–ZnZn, and d) 1–Co; top views (left) and side views (right).
The thermal ellipsoids are scaled to the 50% probability level. meso-Aryl substituents are omitted for clarity.
arranged in a roughly parallel
manner with an interplanar dis-
tance of 3.4 ꢁ. Accordingly, the
distances from C1 to C20 and from C19 to C40, which are
considered to form bonds during the metathesis-like split-
ting reaction, are 4.127 and 4.176 ꢁ. The free-base 1 does
not undergo the splitting reaction, even under harsh condi-
tions.
CuII complexes: To obtain 1–Cu and 1–CuCu, the metala-
tion should be conducted at 08C to avoid the thermal split-
ting reaction of 1–CuCu. A solution of 1 in a mixture of di-
chloromethane and methanol was stirred in the presence of
CuACHTUNGTRENNUNG(OAc)2 (3 equiv) and NaOAc (3 equiv) for 1 h to provide
1–Cu (77%) and 1–CuCu (9%) (see Scheme 1). The com-
plex 1–CuCu was selectively prepared by stirring 1 in the
presence of CuACHTUNGTRENNUNG(OAc)2 (10 equiv) and NaOAc (5 equiv) for
7 h to give 81% yield. Similarly to 1, the mono-CuII complex
1–Cu is not split upon prolonged heating, highlighting the
crucial importance of the two coordinated CuII ions for the
Scheme 1. Metalation of 1 with Cu
ting reaction of 1–CuCu.
ACHTUNGRTEN(NUNG OAc)2 and subsequent thermal split-
5676
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 5674 – 5685

Splitting Reactions of Metallated [36]Octaphyrins(1.1.1.1.1.1.1.1)
FULL PAPER
splitting reaction. The splitting
reaction of 1–CuCu was con-
ducted by stirring a solution of
1–CuCu in toluene at 508C.
The progress of the reaction
was monitored by UV/Vis ab-
sorption spectroscopy. More
simply, almost the same absorp-
tion spectral changes were ob-
served by stirring a solution of
1 in toluene in the presence Cu-
ACHTUNGTRENNUNG(OAc)2 (10 equiv) and NaOAc
(5 equiv) at 508C. Absorption
bands at 410 and 643 nm due to
1 changed to the absorption
bands at 412 and 668 nm due to
1–Cu after 2 h and then to the
absorption bands at 351 and
522 nm due to 1–CuCu (Fig-
ure 2a) after 4 h. These spectral
assignments were based on re-
ferring to the respective au-
thentic absorption spectra (see
the Supporting Information).
Surprisingly, the absorption
spectrum of this reaction mix-
ture changed dramatically upon
prolonged heating, and finally
reached the absorption spec-
trum with a sharp absorption
band at 408 nm and two small
absorption bands at 535 and
570 nm after 2 weeks. The final
absorption spectrum of the re-
action mixture clearly indicated
the exclusive formation of 2–
Cu, which was actually isolated
in 67% yield from this reaction
mixture. Elevating the reaction
temperature to 1118C complet-
ed the splitting reaction within
2 h, to provide 2–Cu in a better
yield of 91%. Single-crystal X-
ray diffraction analysis of 1–
CuCu indicated
a
distorted
structure of C2 symmetry, in
which the two Cu atoms are
bound to the four pyrrolic ni-
trogen atoms but the coordina-
tion features are far from those
of usual square-planar complex-
es.[14a] As shown in Figure 1b,
pyrrole rings C, B, and D and F,
G, N, and H constitute relative-
ly planar segments, but pyrrole
Figure 2. Absorption spectra of a) 1–CuCu, b) 1–ZnZn, c) 1–Co, d) 1–CoZn, and e) 1–CoCu in CH2Cl2 (left)
and in acetonitrile (right).
rings A and H are significantly tilted from these segments.
relief of the strain in 1–CuCu may be a main driving force
This severely distorted structure of 1–CuCu suggests that a
for this splitting reaction.
Chem. Eur. J. 2009, 15, 5674 – 5685
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5677

A. Osuka et al.
The rates of the splitting reaction were determined by
monitoring the increasing absorbance of 2–Cu at 410 nm at
various temperature to be 5.1ꢂ10À5 sÀ1 at 608C, 7.8ꢂ10À5 sÀ1
at 658C, 1.4ꢂ10À4 sÀ1 at 708C, 2.5ꢂ10À4 sÀ1 at 758C, and
4.1ꢂ10À4 sÀ1 at 808C. By using these results, the activation
parameters of the splitting reaction have been determined
as the following: activation energy, Ea =104 kJmolÀ1; activa-
tion enthalpy, DH° =101 kJmolÀ1; and activation entropy,
DS° =À24.4 JmolÀ1 KÀ1 (see the Supporting Information).
The reaction rates and activation parameters were also mea-
sured in the more polar acetonitrile solution: 4.3ꢂ10À5 sÀ1 at
608C, 7.7ꢂ10À5 sÀ1 at 658C, 1.4ꢂ10À4 sÀ1 at 708C, 2.4ꢂ
10À4 sÀ1 at 758C, and 3.5ꢂ10À4 sÀ1 at 808C, and Ea =
Scheme 3. A possible mechanism of the thermal splitting reaction of 1–
CuCu to 2–Cu. meso-Pentafluorophenyl substituents were omitted for
clarity. Newly formed bonds are indicated as red lines.
104 kJmolÀ1,
DH° =101 kJmolÀ1,
and
DS° =
À25.0 JmolÀ1 KÀ1, which indicates no significant solvent-po-
larity effects for the splitting reaction. In spite of the disso-
ciative nature of the splitting reaction, the observed negative
activation entropy is notable, since it suggests an increased
ordering in the transition state from the initial state. We also
determined an activation volume (DV°) of the splitting reac-
tion by measuring the reaction rate at high pressure at
758C.[17] With increasing pressure, the reaction was deceler-
ated (2.7ꢂ10À4 sÀ1 at 1 bar, 1.9ꢂ10À4 sÀ1 at 500 bar, and 1.1ꢂ
10À4 sÀ1 at 1000 bar), which led to an evaluation of DV° =
18 cm3 molÀ1. The positive value of the activation volume in-
dicated an increased volume in the transition state from the
initial state. Taking these results into consideration, one pos-
sible mechanism of the splitting reaction is [2+2] cycloaddi-
tion to give a spirocyclobutane intermediate, which is split
into two molecules of 2–Cu through a reverse cycloaddition
reaction (Scheme 2), since the proposed spirocyclobutane
structure is more ordered and larger in volume than the ini-
tial state. In accordance with this consideration, the distance
between C1 and C20, which will be connected during the re-
action, is 3.728 ꢁ in 1–CuCu, which is distinctly shorter than
that of 1 (4.127 ꢁ).
formation). The UV/Vis absorption spectrum of 1–ZnZn
changes depending on the solvent (Figure 2b), and the solu-
tion color also changes; red in CH2Cl2 and green in acetoni-
trile probably due to the coordination of acetonitrile mole-
cules to the ZnII ions. The structure of 1–ZnZn has been re-
vealed to be a severely distorted figure-of-eight conforma-
tion, in which the two ZnII ions are each bound to the por-
phyrin-like four nitrogen atoms in a similar manner to that
À
of 1–CuCu (Figure 1b). The Zn N bond lengths are 2.08,
À
2.01, 2.02, and 2.03 ꢁ, and the Zn Zn distance is 5.21 ꢁ.
Pyrrole rings B, C, and D constitute a relatively flat tripyr-
rolic segment with the mean plane deviation of 0.22 ꢁ, to
which pyrrole A is held with a large tilting angle of 628. De-
spite the distorted structure comparable to that of 1–CuCu,
complex 1–ZnZn is thermally robust and does not undergo
splitting reactions even by prolonged heating at reflux. Im-
portantly, the distance between C1 and C20 in 1–ZnZn is
3.859 ꢁ, which is slightly, but distinctly longer than that of
1–CuCu (3.728 ꢁ).
ZnII complex: Zinc(II) metalation of 1 was conducted by
heating a solution of 1 in methanol at reflux in the presence
CoII complexes: A solution of 1 in dichloromethane was
heated at reflux in the presence of CoACTHNUTRGENUG(N OAc)2 (10 equiv),
of excess amounts of Zn
(OAc)2 and Na
E
NaOAc (10 equiv), and a small amount of methanol for 2 h
to give mono-CoII complex 1–Co almost quantitatively. In-
terestingly, under these forcing conditions, the formation of
1–CoCo was not detected. The HR ESI-TOF mass spectrum
of 1–Co shows the parent ion peak at m/z 2005.0354
([MÀH]À, calcd for C88H17F40N8Co1 =2005.0307). The struc-
ture of 1–Co was determined by single-crystal X-ray diffrac-
tion analysis to be a distorted figure-of-eight conformation
(Figure 1d), which is similar to
gave 1–ZnZn almost quantitatively (Scheme 3). High-resolu-
tion electrospray ionization time-of-flight (HR ESI-TOF)
mass analysis of 1–ZnZn showed the parent ion peak at m/z
2110.9160 ([M+Cl]À, calcd for C88H16F40N8Zn2Cl=
1
2110.9117). The H NMR spectrum exhibited eight signals in
the range from 6.1 to 7.5 ppm, suggesting its C2 symmetric
structure and nonaromatic character (see the Supporting In-
those of 1, 1–CuCu, and 1–
ZnZn. In this complex, the CoII
ion is bound to the four nitro-
gen atoms of the pyrroles A, B,
C
and D with distances of
2.149, 2.024, 2.115, and 2.116 ꢁ
and to the nitrogen atom of the
crystallizing solvent acetonitrile
with a distance of 2.222 ꢁ. The
Scheme 2. Metalation of 1 with Zn
ACHTUNGTRENNUNG
5678
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 5674 – 5685

Splitting Reactions of Metallated [36]Octaphyrins(1.1.1.1.1.1.1.1)
FULL PAPER
spin state of CoII ion is considered to be the high-spin state
from its temperature-dependent magnetic susceptibility (see
the Supporting Information). The complex 1–Co is thermal-
ly quite robust and does not undergo the splitting reaction.
Because 1–Co holds a vacant hemi-porphyrin-like tetra-
pyrrolic cavity, further metalation was attempted to synthe-
size hetero-metal complexes of [36]octaphyrin. First, hetero
bis-metal complex 1–CoZn was prepared by the metalation
in dichloromethane is similar to that of 1–CuCu, showing
broad bands at 347 and 522 nm and weaker bands at 688,
821, and 984 nm (Figure 2e). The complex 1–CoCu is red in
color in dichloromethane.
The thermal splitting reaction of 1–CoCu was examined.
While heating 1–CoCu in toluene led to its decomposition,
the clean splitting reaction of 1–CoCu was observed in ace-
tonitrile, which gave an equal mixture of 2–Co and 2–Cu as
indicated by UV/Vis absorption and 19F NMR spectroscop-
ies (see the Supporting Information). We indeed isolated
both 2–Co and 2–Cu in more than 90% yields from the ther-
molysis of 1–CoCu by heating at reflux in acetonitrile for
24 h. This contrasting solvent effect on the thermal reactivity
of 1–CoCu is difficult to explain at this stage. The red color
of 1–CoCu in dichloromethane changes to green in acetoni-
trile. In fact, the absorption spectrum of 1–CoCu shows a
relatively intense band at 697 nm along with smaller bands
at 440, 909, and 1030 nm in acetonitrile (Figure 2e). These
spectral changes may be ascribed to additional coordination
of an acetonitrile molecule on the CoII ion, as in the crystal
structure of 1–Co (Figure 1d). The rates of the splitting re-
action were measured by monitoring the UV/Vis absorbance
of a mixture of 2–Co and 2–Cu at variable temperatures.
The reaction rates thus determined were 2.1ꢂ10À5 sÀ1 at
608C, 3.2ꢂ10À5 sÀ1 at 658C, 6.3ꢂ10À4 sÀ1 at 708C, 1.1ꢂ
10À4 sÀ1 at 758C, and 3.5ꢂ10À4 sÀ1 at 808C. The splitting re-
action rates of 1–CoCu were slower than those of 1–CuCu
at the same temperature. By analyzing these data with Ar-
rhenius and Eyring plots, an activation energy (Ea =
105 kJmolÀ1) and activation parameters (DH° =102 kJmolÀ1
and DS° =À29.9 JmolÀ1 KÀ1) were estimated. Importantly
Ea and DH° values of the splitting reaction of 1–CoCu are
almost the same as those in 1–CuCu, but the DS° value of
1–CoCu was more negative by 4.9 JmolÀ1 KÀ1 than that of
1–CuCu. Thus, the slow splitting reaction of 1–CoCu can be
ascribed mainly to the entropy term. The splitting reaction
of 1–CoCu is probably more ordered and includes the coor-
dinated acetonitrile molecule in the transition state.
of 1–Co with ZnACHTUNGTRENNUNG(OAc)2 (Scheme 4). Although some deme-
Scheme 4. Metalation of 1 with Co
splitting reaction of 1–CoCu.
ACHTUNGERTN(NUNG OAc)2 and the synthesis and thermal
talation of the ZnII ion occurred during work up, pure 1–
CoZn was obtained in 77% yield along with recovery of 1–
Co (22%). The parent ion peak of 1–CoZn was observed at
m/z 2103.9173 ([M+Cl]À, calcd for C88H16F40N8CoZnCl=
2103.9179) in the HR ESI-TOF mass spectrum. The absorp-
tion spectrum of 1–CoZn was revealed to be solvent depen-
dent. In dichloromethane, the spectrum is ill-defined feature
with broad bands at 346, 516, 689, and 817 nm (Figure 2d),
whereas that in acetonitrile exhibits rather similar features
to that of 1–ZnZn, in which there is a relatively intense
band at 700 nm along with weaker bands at 357, 434, and
1025 nm. Measurement of the magnetic susceptibility of 1–
CoZn suggested that the CoII ion is in the high-spin state.
The 1H NMR spectrum in CDCl3 shows b-pyrrolic proton
signals at 95.42, 90.63, 83.21, 76.00, 62.15, 55.67, 54.07, 24.67,
12.58, 6.89, 6.31, 3.39, 0.09, À2.77, À2.96, and À25.55 ppm,
suggesting the effect of a paramagnetic shift by CoII. The
splitting reaction of 1–CoZn was attempted by heating solu-
tions in toluene or acetonitrile, which, however, led to facile
liberation of the ZnII ion and thus further examinations
were abandoned. In the next step, hetero bis-metal complex
1–CoCu was efficiently prepared by treatment of 1–Co with
Computational investigations: At this stage, we investigated
the reaction mechanism by computational methods with the
hope of understanding the marked difference in the splitting
reactivity between 1–CuCu and 1–ZnZn. All calculations
were carried out by using the Gaussian 03 program. DFT
calculations were performed starting from the model sys-
tems of 1–CuCu and 1–ZnZn. Initial geometries were ob-
tained from the crystal structures of 1–CuCu and 1–ZnZn.
To reduce the calculation time, we replaced two pentafluoro-
phenyl groups at the 20-, and 40-positions with 2,6-difluoro-
phenyl groups, which seem to have significant influence on
the reaction path, and other six pentafluorophenyl groups,
which are probably not critical, were replaced by hydrogen
atoms. The structures of the intermediates and transition
states were obtained by estimates. The spin states were des-
ignated as triplet for all bis-copper complexes because the
triplet ground state (S=1/2ꢂ2) for 1–CuCu was elucidated
by the magnetic susceptibility measurement, in which no
Cu
was also synthesized by the metalation of 1–Cu with Co-
(OAc)2. The hetero bis-metal composition was confirmed by
ACHTUNGTRENNUNG(OAc)2 under similar conditions. The complex 1–CoCu
ACHTUNGTRENNUNG
the parent ion peak at m/z 2065.9507 ([M]À, calcd for
C88H16F40N8CoCu=2065.9493) in the HR ESI-TOF mass
spectrum. The spin state of the CoII ion was again revealed
to be high spin from its magnetic susceptibility (see the Sup-
porting Information). The absorption spectrum of 1–CoCu
Chem. Eur. J. 2009, 15, 5674 – 5685
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5679

A. Osuka et al.
magnetic interaction was observed between two isolated
spins (see the Supporting Information). The structures were
fully optimized with Beckeꢃs three-parameter hybrid ex-
change functional and the Lee–Yang–Parr correlation func-
tional (B3LYP), employing LANL2DZ basis set for all
atoms. Zero-point and thermal energy corrections were con-
ducted for all optimized structures. The obtained transition
states gave single imaginary frequencies and intrinsic reac-
tion coordinate (IRC) calculations supported the transition
structures. At the optimized geometries, energies were fur-
ther calculated by the B3LYP method with SDD basis set
for copper and zinc and 6-31G* for the rest of atoms (de-
noted as 631SDD).
C20 gradually, we found the transition state TS1, which then
changes to the singly bridged intermediate INT1 (Figure 3
and Scheme 2). The activation energy of this initial process
was calculated to be Ea =41.7 kcalmolÀ1 at the B3LYP/
631SDD//B3LYP/LANL2DZ level (Figure 4a). Then we lo-
The optimized structure of the model compound 1–CuCu
is in good agreement with the X-ray crystal structure
(Figure 3). It is noteworthy that the distance between C1
Figure 4. Calculated reaction coordinates for the splitting reaction of
a) 1–CuCu to 2–Cu and b) 1–ZnZn to 2–Zn. Energies were calculated at
the B3LYP/631SDD//B3LYP/LANL2DZ level.
cated the transition state TS2 with very small activation
À1
À
energy (Ea =2.0 kcalmol ) for the second C C bond forma-
tion between C21 and C40 from INT1 to a spirocyclobutane
intermediate INT2.
A reverse cycloaddition reaction of the cyclobutane in
INT2 would provide two porphyrins, but the concerted
mechanism is thermally prohibited for retro [2+2] cycload-
dition. Consequently, a stepwise pathway involving two se-
À
quential homolytic cleavages of two C C bonds is reasona-
bly expected. In fact, transition-state TS3 with an activation
energy of 11.8 kcalmolÀ1 was obtained for the initial C1
À
C40 bond scission of the spirocyclobutane intermediate.
Transition-state TS3 exhibits a significant radical character
with <S2 >=2.372 (B3LYP/631SDD//B3LYP/LANL2DZ),
whereas all other bis-CuII species have very small spin densi-
ty on the macrocycle (<S2 >=2.006–2.008, see the Support-
ing Information), which only resulted from the spin on CuII.
Here, <S2 > denotes the expectation value of the total spin
angular momentum. In INT2, C1 and C20 have Mulliken
spin densities of 0.00242 and 0.00084, while the spin densi-
ties at those carbon atoms in TS3 are 0.136 and À0.271, re-
spectively (see the Supporting Information). This implies a
Figure 3. Optimized structures of intermediates and transition states for
the splitting reaction of 1–CuCu to 2–Cu at the B3LYP/LANL2DZ level.
Hydrogen atoms were omitted for clarity.
and C20 is only 3.519 ꢁ, indicating very close proximity be-
tween a- and meso-carbon atoms in the crossing point of the
figure-of-eight structure. This is important for the reaction
mechanism because bond formation between a- and meso-
carbon atoms is required for the formation of porphyrin
from octaphyrin. By reducing the distance between C1 and
À
partial singlet biradical character of the C C bond cleavage
process in TS3. All CuII species including 2–Cu posses Mul-
5680
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 5674 – 5685

Splitting Reactions of Metallated [36]Octaphyrins(1.1.1.1.1.1.1.1)
FULL PAPER
liken spin densities on copper centers ranging from 0.614 to
0.635, thus indicating no significant change of the valence
state of copper during the splitting reaction. No further in-
termediates could be found from TS3 to 2–Cu, and the IRC
and the Mulliken spin densities of C21 and C40 are 0.150
and À0.336, respectively.
Then, the question remains as to the reason for such a
high activation energy for the initial step. Since the structur-
al factors are similar between 2–Cu and 2–Zn, it is meaning-
ful to compare relative energies of intermediates and transi-
tion states on the basis of the energies of 2–Cu and 2–Zn
(Figure 3). This clearly highlights the relative instability of
1–CuCu in comparison with 1–ZnZn. This inherent instabili-
ty of 1–CuCu may be a main reason for its smooth splitting
reaction. The instability of 1–CuCu largely arises from the
À
calculation confirmed that the second C C bond cleavage
between C1 and C20 occurs successively from TS3 without
any intermediate. Probably such an intermediate species
with a strong singlet biradical nature is too unstable to be
optimized and provides two porphyrins directly.
À
These DFT calculations indicate that the initial C C bond
formation from 1–CuCu to INT1 is the rate-determining
step. On the basis of the frequency calculation, the activa-
tion entropy for this process is estimated to be DS° =
À24.2 JmolÀ1 KÀ1, which is matched well with the experi-
mental value À24.4 JmolÀ1 KÀ1 in toluene.
À
ring strain. Characteristically, the calculated C1 C20 dis-
tance in 1–CuCu is 3.519 ꢁ, which is distinctly shorter that
À
that (3.797 ꢁ) in 1–ZnZn (Figure 5). Such a short C1 C20
distance forces a larger distortion for the ring system of 1–
À
To investigate the origin of the different reactivity be-
tween 1–CuCu and 1–ZnZn, we also calculated the reaction
pathway for the similar splitting of 1–ZnZn to 2–Zn. Similar
intermediates and transition states (1–ZnZn, TS1, INT1,
TS2, INT2, TS3, and 2–Zn) were obtained (Figure 4b). In
the case of ZnII complexes, the activation energy (Ea =
53.0 kcalmolÀ1) for the initial step from 1–ZnZn to TS1 is
significantly higher than that of 1–CuCu (Ea =41.7 kcal
molÀ1). Again, TS3 also has significant biradical character
with <S2 >=0.464 (B3LYP/631SDD//B3LYP/LANL2DZ)
CuCu. In addition, a shorter N metal bond length causes
shrinkage of the octaphyrin macrocycle to decrease the dis-
tance between a- and meso-carbon atoms and thus induce
À
more strain. In fact, the calculated Cu N bond lengths are
À
in the range of 2.013–2.096 ꢁ in 1–CuCu, whereas Zn N
bond lengths are 2.092–2.162 ꢁ in 1–ZnZn. These tenden-
cies are confirmed in the crystal structures of the real sys-
À
tems 1–CuCu and 1–ZnZn: the C1 C20 distance is 3.728 ꢁ
À
and Cu N bond lengths are 1.974–2.059 ꢁ in 1–CuCu,
À
À
whereas the C1 C20 distance is 3.859 ꢁ and Zn N bond
lengths are 2.007–2.079 ꢁ in 1–ZnZn. In this context, it is in-
teresting to note that the C1–C20 distances of the mono-
metal complexes are calculated to be 3.889 ꢁ for 1–Cu and
4.048 ꢁ for 1–Co.
PdII complexes: Recently, we reported the formation of bis-
PdII complexes 3–PdPd and 4–PdPd in yields of 51 and
20%, respectively, in the metalation of 1 with PdACHTUNGTRENNUNG(OAc)2
(Scheme 5). These two products have been identified as a
twisted Hꢄckel antiaromatic molecule and a twisted Mçbius
aromatic molecule, respectively.[10a] The complex 4–PdPd is
one of the first Mçbius aromatic molecules with distinct aro-
maticity.[7,18,19] We also found that a new complex, 5–PdPd,
was formed in 21% yield along with 3–PdPd (28%) and 4–
Figure 5. Calculated structures of 1–CuCu and 1–ZnZn at the B3LYP/
PdPd (trace), when the metalation of 1 with Pd
ACHTUNGRTEN(NUNG OAc)2 was
À
À
LANL2DZ level. Several metal N and C1 C20 bond lengths are also in-
dicated in ꢁ.
conducted in a 9:1 mixture of 2,2,2-trifluoroethanol and
Scheme 5. Metalation of 1 with Pd
Chem. Eur. J. 2009, 15, 5674 – 5685
ACHTUNGERTN(NUNG OAc)2. A part of 5–PdPd is highlighted, where newly generated bonds are indicated in grey.
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5681

A. Osuka et al.
methanol. The HR ESI-TOF mass analysis of 5–PdPd
showed the parent ion peak at m/z 2157.8976 ([M]À, calcd
for C88H16N8F40Pd2, 2157.8967), which is the same as those
of 3–PdPd and 4–PdPd. The structure of 5–PdPd was re-
vealed by X-ray analysis, as shown in Figure 6.[20] The com-
hydrogen atoms. In agreement with this structure, the
1H NMR spectrum of 5–PdPd showed fourteen doublet sig-
nals due to the b-pyrrolic protons and two singlet signals at
d=2.87 ppm due to the saturated b-pyrrolic protons. The
structure was also confirmed by measuring the absorption
spectrum, which is shown in Figure 7.
Figure 7. Absorption spectrum of 5–PdPd in CH2Cl2.
The three complexes 3–PdPd, 4–PdPd, and 5–PdPd are
each thermally stable and no interconversions were ob-
served, even when heated at reflux in toluene. A plausible
reaction pathway to 5–PdPd is shown in Scheme 6, in which
Figure 6. Crystal structure of 5–PdPd; top view (left) and side view
(right). Newly generated bonds are indicated in red. The thermal ellip-
soids are scaled to the 50% probability level. meso-Aryl substituents are
omitted for clarity.
complex 6–PdPd is postulated as a precursor for 5–PdPd.
II
À
Probably due to the intrinsic C H activation ability of Pd
ion when coordinated in expanded porphyrins, the NNNN
metalation as observed for 1–CuCu and 1–ZnZn is impossi-
ble for PdII metalation. As a result, symmetric NNNC metal-
ation led to the formation of 3–PdPd and non-symmetric
NNNC and NNNN metalation provided 6–PdPd. In 3–PdPd,
the distances between the meso-carbon and the pyrrolic a-
carbon are relatively long (4.835 and 4.824 ꢁ),[10a] which
may explain the lack of transannular reactivity. On the other
hand, the transannular interaction is enhanced in 6–PdPd,
plex 5–PdPd takes a unique structure, in which the Pd1
atom is bound to the four nitrogen atoms of pyrrole rings A,
B, C, and D with distances of 1.994, 1.974, 1.950, and
2.000 ꢁ, and the Pd2 atom is bound to the three nitrogen
atoms of pyrrole rings F, G, and H with distances of 2.078,
2.032, and 1.993 ꢁ and the b-carbon atom of pyrrole ring A
with a distance of 1.956 ꢁ. In addition, the meso-carbon
(C20) and the a-carbon of pyrrole ring A (C1) are directly
connected and the b-carbon of pyrrole ring A (C3) has two
À
which causes the C1 C20 bond formation reaction to give
À
7–PdPd and subsequent C2 N5 (pyrrole ring E) bond for-
mation with concurrent hydrogen transfer from pyrrole ring
Scheme 6. A possible metalation pathway to 5–PdPd and comparison with splitting reaction pathway of 1–CuCu. Newly generated bonds are indicated in
red.
5682
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 5674 – 5685

Splitting Reactions of Metallated [36]Octaphyrins(1.1.1.1.1.1.1.1)
FULL PAPER
raphy on silica gel gave 1–ZnZn (20.0 mg, 96%). 1H NMR (600 MHz,
CDCl3, 298 K): d=7.50 (t, J=4.14 Hz, 2H), 6.60 (d, J=4.14 Hz, 2H),
6.58 (brs, 2H), 6.48 (brs, 2H), 6.38 (d, J=4.14 Hz, 2H), 6.31 (d, J=
4.14 Hz, 2H), 6.26 (d, J=4.8 Hz, 2H), 6.12 ppm (d, J=4.8 Hz, 2H);
19F NMR (564.73 MHz, CDCl3, 298 K): d=À131.87 (d, J=22.02 Hz, 2F;
o-F), À132.90 (d, J=23.4 Hz, 2F; o-F), À136.63 (d, J=23.46 Hz, 2F; o-
F), À137.05 (d, J=19.02 Hz, 2F; o-F), À138.34 (t, J=24.9 Hz, 4F; o-F),
À140.37 (d, J=22.02 Hz, 2F; o-F), À141.54 (brs, 2F; o-F), À151.37 (t,
J=19.02 Hz, 2F; p-F), À153.10 (t, J=21.96 Hz, 2F; p-F), À153.30 (t, J=
21.96 Hz, 2F; p-F), À153.37 (t, J=22.02 Hz, 2F; p-F), À160.13 (dt, J1 =
5.88 Hz, J2 =20.52 Hz, 2F; m-F), À160.73 (d, J=19.08 Hz, 2F; m-F),
À161.39 (m, 8F; p-F), À161.71 ppm (m, 4F; p-F); UV/Vis (CH2Cl2): lmax
(e)=350 (59200), 517 (71300), 678 (30200), 807 (17500), 958 nm
(10000 mÀ1 cmÀ1); UV/Vis (CH3CN): lmax (e)=359 (44400), 435 (71000),
693 (135000), 787 (sh, 24500), 1030 nm (6200 mÀ1 cmÀ1); HR ESI-TOF-
MS (negative mode): m/z calcd for C88H16F40N8Zn2Cl [M+Cl]À:
2110.9117; found: 2110.9160 (100).
E to the b-carbon of pyrrole ring A affords 5–PdPd. It is
noteworthy that the reaction sequence from 6–PdPd to 7–
PdPd is similar to that from 1–CuCu to INT1 and the struc-
ture of 5–PdPd is the manifestation of the transannular elec-
tronic interaction.
Conclusions
Metalation of 1 with CuII, ZnII, CoII, and PdII ions was exam-
ined. CuII, ZnII, and CoII ions favor NNNN coordination to
form figure-of-eight complexes 1–Cu, 1–CuCu, 1–ZnZn, and
1–Co, whereas PdII prefers NNNC and NNNN coordination.
Among these, 1–CuCu and 1–CoCu undergo the splitting re-
action to two porphyrins upon heating, but the reaction rate
was faster for 1–CuCu. The large negative reaction entropies
observed for the splitting reactions of 1–CuCu and 1–CoCu
and DFT calculations suggested a stepwise route via inter-
mediates INT1 and INT2 (spirocyclobutane) followed by a
radical reverse cycloaddition reaction. Bis-PdII complex 5–
PdPd was newly isolated as the manifestation of the transan-
nular electronic interactions in the metalated octaphyrins.
Collectively, these results indicate that octaphyrin 1 is a
unique scaffold to effect the topological splitting reactions
through the transannular electronic interactions that are en-
hanced upon metalation.
CCDC-716724 contains 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.
Mono-cobalt(II) complex of [36]octaphyrin (1–Co): A solution of Co-
AHCTUNGRTEG(NNNU OAc)2 (55.9 mg, 10 equiv) and NaOAc (26.4 mg, 10 equiv) in methanol
(10 mL) was added to a solution of 1 (58.6 mg, 0.03 mmol) in CH2Cl2
(30 mL) and the resulting mixture was heated at reflux for 2 h. The resi-
dues were washed with water, dried over Na2SO4, and the solvent was re-
moved. Purification by chromatography on silica gel gave 1–Co (55.2 mg,
92%).
UV/Vis
(CH2Cl2):
lmax
(e)=411
(75500),
682 nm
(102000 mÀ1 cmÀ1); UV/Vis (CH3CN): lmax (e)=411 (73600), 678 nm
(99000 mÀ1 cmÀ1); HR ESI-TOF-MS (negative mode): m/z calcd for
C88H17F40N8Co1 [MÀH]À: 2005.0307; found: 2005.0354.
CCDC-716725 contains 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.
Cobalt(II)–zinc(II) hetero complex of [36]octaphyrin (1–CoZn): A solu-
Experimental Section
tion of ZnACHTUNGRTENNG(U OAc)2 (18.3 mg, 10 equiv) and NaOAc (8.2 mg, 10 equiv) in
methanol (10 mL) was added to a solution of 1–Co (20.4 mg, 0.01 mmol)
in CH2Cl2 (10 mL), and the resulting mixture was stirred for 1 h. The res-
idues were washed with water, dried over Na2SO4, and the solvent was
removed. Purification by chromatography on silica gel gave 1–CoZn
(16.0 mg, 77%) and demetalated 1–Co (4.4 mg, 22%). 1H NMR
(600 MHz, CDCl3, 298 K): d=95.42 (s, 1H), 90.63 (s, 1H), 83.21 (s, 1H),
76.00 (s, 1H), 62.15 (s, 1H), 55.67 (s, 1H), 54.07 (s, 1H), 24.67 (s, 1H),
12.58 (s, 1H), 6.89 (s, 1H), 6.31 (s, 1H), 3.39 (s, 1H), 0.09 (s, 1H), À2.77
(s, 1H), À2.96 (s, 1H), À25.55 ppm (s, 1H); 19F NMR (564.73 MHz,
CDCl3, 298 K): d=À75.70 (s, 1F), À107.34 (s, 1F), À108.87 (s, 1F),
À119.99 (s, 1F), À134.18 (d, J=22.02 Hz, 1F), À136.47 (d, J=22.02 Hz,
1F), À137.71 (s, 2F), À139.15 (s, 1F), À140.33 (t, J=29.34 Hz, 2F),
À140.77 (s, 2F), À140.94 (s, 2F), À141.12 (s, 2F), À146.68 (s, 1F),
À148.06 (s, 2F), À150.25 (s, 1F), À150.50 (s, 1F), À151.13 (s, 1F),
À152.66 (t, J=22.02 Hz, 1F), À153.44 (t, J=21.96 Hz, 1F), À154.73 (s,
1F), À155.80 (s, 1F), À155.95 (s, 1F), À160.41 (t, J=22.02 Hz, 1F),
À161.10 (t, J=22.02 Hz, 1F), À161.23 (t, J=22.02 Hz, 1F), À161.43 (t,
J=22.02 Hz, 1F), À162.33 (t, J=29.40 Hz, 1F), À162.38 (s, 1F), À165.76
(s, 1F), À168.56 (s, 1F), À176.64 (s, 1F), À195.97 (s, 1F), À201.57 ppm (s,
1F); UV/Vis (CH2Cl2): lmax (e)=346 (53000), 416 (39700), 516 (47400),
689 (41400), 817 nm (16400 mÀ1 cmÀ1); UV/Vis (CH3CN): lmax (e)=357
(42000), 434 (58000), 700 (125000), 1025 nm (5700 mÀ1 cmÀ1); HR ESI-
TOF-MS (negative mode): m/z calcd for C88H16ClCoF40N8Zn [M+Cl]À:
2103.9179; found: 2103.9173.
General: 1H and 19F NMR spectra were recorded on a JEOL ECA-600
spectrometer (600 MHz for 1H and 565 MHz for 19F). Chemical shifts
were reported on the delta scale in ppm relative to the residual solvent
as the internal reference for 1H (d=7.260 ppm), and hexafluorobenzene
was used as external reference for 19F (d=À162.9 ppm). NMR signals
1
were assigned from the H–1H COSY spectra, and from comparison with
the spectra in the presence of D2O (signals assigned for NH protons dis-
appear in the presence of D2O) and the simulated spectra. UV/Vis spec-
tra were recorded on a Shimadzu UV-3100PC spectrometer. MALDI-
TOF mass spectra were recorded on a Shimadzu KRATOS KOMPACT
MALDI4 spectrometer by using the positive or negative MALDI
method. High-resolution FAB mass spectra were recorded on a JEOL
HX110 spectrometer by using the fast atom bombardment method in the
positive ion mode with a 3-nitrobenzylalcohol matrix. High-resolution
ESI-TOF mass spectra of samples in acetonitrile were recorded on a
BRUKER microTOF instrument in the positive or negative ion mode.
X-ray data were recorded on a Rigaku-Raxis imaging plate system, or on
a BRUKER-APEX X-Ray diffractometer equipped with a large area
CCD detector. The magnetic susceptibility was measured for the powder
samples with the temperature range from 2 to 300 K at 0.1 T magnetic
field by using Quantum Design MPMS-5. Unless otherwise noted, mate-
rials obtained from commercial suppliers were used without further pu-
rification. Silica gel column chromatography was performed on Wakogel
C-300 and C-400. Thin-layer chromatography (TLC) was carried out on
aluminum sheets coated with silica gel 60 F254 (Merck 5554).
Cobalt(II)–copper(II) hetero complex of [36]octaphyrin (1–CoCu): A so-
lution of Cu
ACHTUNGTRENNUNG
in methanol (10 mL) was added to
a
Bis-zinc(II) complex of [36]octaphyrin (1–ZnZn): ZnACTHUNRGTNEUNG(OAc)2 (18.9 mg,
0.01 mmol) in CH2Cl2 (30 mL), and the resulting mixture was stirred for
1 h. The residues were washed with water, dried over Na2SO4, and the
solvent was removed. Purification by chromatography on silica gel gave
1–CoCu (19.0 mg, 92%). UV/Vis (CH2Cl2): lmax (e)=349 (60300), 521
(63900), 688 (16600), 834 (15200), 993 nm (10200 mÀ1 cmÀ1); UV/Vis
(CH3CN): lmax (e)=350 (54700), 439 (58000), 524 (32000), 697 nm
0.10 mmol; 10 equiv) and NaOAc (8.7 mg, 0.10 mmol; 10 equiv) were
added to a solution of 1 (19.5 mg, 0.01 mmol; 1.0 mm in MeOH) and the
resulting mixture was heated at reflux for 2 h under a nitrogen atmos-
phere in the dark. After removing the solvent, the residues were dis-
solved in CH2Cl2, washed with water, dried over Na2SO4, and the solvent
was removed under reduced pressure. Purification by column chromatog-
Chem. Eur. J. 2009, 15, 5674 – 5685
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5683

A. Osuka et al.
(98400 mÀ1 cmÀ1); HR ESI-TOF-MS (negative mode): m/z calcd for
C88H16F40N8Co63Cu [M]À: 2065.9493; found: 2065.9507.
Jendrny, L. Zander, H. Schmickler, J. Lex, Y.-D. Wu, M. Nendel, J.
Chen, D. A. Plattner, K. N. Houk, E. Vogel, Angew. Chem. 1995,
107, 2709–2711; Angew. Chem. Int. Ed. Engl. 1995, 34, 2515–2517;
e) J. Setsune, Y. Katakami, N. Iizuna, J. Am. Chem. Soc. 1999, 121,
8957–8958; f) J. Setsune, S. Maeda, J. Am. Chem. Soc. 2000, 122,
12405–12406; g) V. G. Anand, S. K. Pushpan, S. Venkatraman, A.
Dey, T. K. Chadrashekar, B. S. Joshi, R. Roy, W. Teng, K. R. Senge,
J. Am. Chem. Soc. 2001, 123, 8620–8621; h) S. Shimizu, N. Aratani,
A. Osuka, Chem. Eur. J. 2006, 12, 4909–4918; i) Y. Tanaka, J.-Y.
Shin, A. Osuka, Eur. J. Org. Chem. 2008, 1341–1349.
Bis-palladium(II) complex of [36]octaphyrin (5–PdPd): PdACTHNUTRGNEUNG(OAc)2
(24.0 mg, 10 equiv) and NaOAc (9.2 mg, 10 equiv) was added to a solu-
tion of 1 (19.6 mg, 0.01 mmol) in a 9:1 mixture of CF3CH2OH (9 mL)
and MeOH (1 mL), and the resulting mixture was heated at reflux for
2 h. After removal of the solvent under reduced pressure, the residues
were dissolved in AcOEt, washed with water, dried over Na2SO4, and the
solvent was again removed. Purification by chromatography on silica gel
gave 3–PdPd (6.1 mg, 28%) and 5–PdPd (4.5 mg, 21%) along with a
trace of 4–PdPd.
[3] a) M. G. P. M. S. Neves, R. M. Martins, A. C. Tomꢅ, A. J. D. Silves-
tre, A. M. S. Silva, V. Fꢅlix, M. G. B. Drew, J. A. S. Cavaleiro, Chem.
Commun. 1999, 385–386; b) S. Mori, A. Osuka, J. Am. Chem. Soc.
2005, 127, 8030–8031; c) S. Mori, J.-Y. Shin, S. Shimizu, F. Ishikawa,
H. Furuta, A. Osuka, Chem. Eur. J. 2005, 11, 2417–2425; d) S. Mori,
K. S. Kim, Z. S. Yoon, S. B. Noh, D. Kim, A. Osuka, J. Am. Chem.
Soc. 2007, 129, 11344–11345.
[4] a) M. Suzuki, A. Osuka, Chem. Commun. 2005, 3685–3687; b) S.
Shimizu, W.-S. Cho, J. L. Sessler, H. Shinokubo, A. Osuka, Chem.
Eur. J. 2008, 14, 2668–2678; c) J. Setsune, M. Mori, T. Okawa, S.
Maeda, J. M. Lintuluoto, J. Organomet. Chem. 2007, 692, 166–174.
[5] a) J. L. Sessler, M. J. Cyr, V. Lynch, E. McGhee, J. A. Ibers, J. Am.
Chem. Soc. 1990, 112, 2810–2813; b) J. L. Sessler, T. Morishima, V.
Lynch, Angew. Chem. 1991, 103, 1018–1020; Angew. Chem. Int. Ed.
Engl. 1991, 30, 977–980; c) S. Shimizu, R. Taniguchi, A. Osuka,
Angew. Chem. 2005, 117, 2265–2269; Angew. Chem. Int. Ed. 2005,
44, 2225–2229.
5–PdPd: 1H NMR (600 MHz, CDCl3, 298 K): d=6.98 (t, J=5.9 Hz, 2H),
6.95 (d, J=5.0 Hz, 1H), 6.84 (d, J=5.5 Hz, 1H), 6.78 (d, J=5.0 Hz, 1H),
6.66 (t, J=5.0 Hz, 1H), 6.64 (d, J=4.6 Hz, 1H), 6.61 (d, J=5.0 Hz, 1H),
6.43 (d, J=4.1 Hz, 1H), 6.37 (d, J=5.0 Hz, 1H), 6.16 (d, J=5.0 Hz, 1H),
6.14 (d, J=4.1 Hz, 1H), 6.12 (d, J=4.6 Hz, 1H), 2.87 ppm (s, sp3-H, 2H);
19F NMR (564.73 MHz, CDCl3, 298 K): d=À121.45 (dd, J1 =26 Hz, J2 =
266 Hz, 1F; o-F), À127.90 (dd, J1 =24 Hz, J2 =265 Hz, 1F; o-F), À135.67
(dd, J1 =26.4 Hz, J2 =65.9 Hz, 1F; o-F), À136.28 (dd, J1 =25.3 Hz, J2 =
66.5 Hz, 1F; o-F), À136.62 (dd, J1 =24.2 Hz, J2 =64.9 Hz, 1F; o-F),
À136.96 (dd, J1 =8.8 Hz, J2 =26.4 Hz, 1F; o-F), À137.30 (dd, J1 =6.6 Hz,
J2 =28.6 Hz, 1F; o-F), À137.84 (dd, J1 =5.5 Hz, J2 =24.2 Hz, 1F; o-F),
À138.57 (dd, J1 =6.6 Hz, J2 =23.0 Hz, 1F; o-F), À138.94 (d, J=22.0 Hz,
1F; o-F), À139.13 (dd, J1 =19.8 Hz, J2 =67.1 Hz, 1F; o-F), À139.43 (d,
J=23.1 Hz, 1F; o-F), À139.61 (d, J=27.5 Hz, 1F; o-F), À139.82 (d, J=
19.7 Hz, 1F; o-F), À140.1 (dd, J1 =22.0 Hz, J2 =63.7 Hz, 1F; o-F),
À140.63 (dd, J1 =24.2 Hz, J2 =63.8 Hz, 1F; o-F), À151.08 (t, J=23.0 Hz,
1F; p-F), À151.77 (m, 1F; p-F), À152.39 (t, J=22.0 Hz, 1F; p-F),
À152.52 (t, J=22.0 Hz, 1F; p-F), À153.07 (t, J=22.0 Hz, 1F; p-F),
À153.27 (t, J=22.0 Hz, 1F; p-F), À153.41 (m, 1F; p-F), À153.92 (t, J=
23.0 Hz, 1F; p-F), À160.97 (dt, J1 =8.8 Hz, J2 =23.6 Hz, 1F; m-F),
À161.25 (dt, J1 =8.8 Hz, J2 =22.5 Hz, 1F; m-F), À161.49 (m, 5F; m-F),
À161.69 (m, 4F; m-F), À161.85 (dt, J1 =7.7 Hz, J2 =23.6 Hz, 1F; m-F),
À161.98 (dt, J1 =7.7 Hz, J2 =23.1 Hz, 1F; m-F), À162.59 (m, 2F; m-F),
À162.83 (t, J=22.0 Hz, 1F; m-F), À163.73 (t, J=22.0 Hz, 1F; m-F),
À165.19 ppm (t, J=22.0 Hz, 1F; m-F); UV/Vis (CH2Cl2): lmax (e)=306
(43200), 347 (37000), 415 (53800), 494 (56700), 749 (19400), 831 nm
(27600 mÀ1 cmÀ1); HR ESI-TOF-MS (negative mode): m/z calcd for
C88H16N8F40Pd2 [M]À: 2157.8967; found: 2157.8976.
[6] a) D. Seidel, V. Lynch, J. L. Sessler, Angew. Chem. 2002, 114, 1480–
1483; Angew. Chem. Int. Ed. 2002, 41, 1422–1425; b) T. Kçhler, D.
Seidel, V. Lynch, F. O. Arp, Z. Ou, K. M. Kadish, J. L. Sessler, J.
Am. Chem. Soc. 2003, 125, 6872–6873.
´
´
˙
[7] a) M. Ste˛pien, L. Latos-Graynski, N. Sprutta, P. Chwalisz, L. Szter-
enberg, Angew. Chem. 2007, 119, 8015–8019; Angew. Chem. Int. Ed.
2007, 46, 7869–7873; b) N. Jux, Angew. Chem. 2008, 120, 2577–
2581; Angew. Chem. Int. Ed. 2008, 47, 2543–2546.
[8] S. J. Weghorn, J. L. Sessler, V. Lynch, T. F. Baumann, J. W. Sibert,
Inorg. Chem. 1996, 35, 1089–1090.
[9] Y. Kamimura, S. Shimizu, A. Osuka, Chem. Eur. J. 2007, 13, 1620–
1628.
CCDC-716726 contains 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.
[10] a) Y. Tanaka, S. Saito, S. Mori, N. Aratani, H. Shinokubo, N. Shiba-
ta, Y. Higuchi, Z. S. Yoon, K. S. Kim, S. B. Noh, J. K. Park, D. Kim,
A. Osuka, Angew. Chem. 2008, 120, 693–696; Angew. Chem. Int.
Ed. 2008, 47, 681–684; b) J. K. Park, Z. S. Yoon, M.-C. Yoon, K. S.
Kim, S. Mori, J.-Y. Shin, A. Osuka, D. Kim, J. Am. Chem. Soc. 2008,
130, 438–489; c) S. Mori, S. Shimizu, R. Taniguchi, A. Osuka, Inorg.
Chem. 2005, 44, 4127–4129.
Acknowledgements
[11] E. Vogel, M. Michels, L. Zander, J. Lex, N. S. Tuzun, K. N. Houk,
Angew. Chem. 2003, 115, 2964–2969; Angew. Chem. Int. Ed. 2003,
42, 2857–2862.
[12] J.-Y. Shin, H. Furuta, K. Yoza, S. Igarashi, A. Osuka, J. Am. Chem.
Soc. 2001, 123, 7190–7191.
This work was partially supported by Grants-in-Aid for Scientific Re-
search (No. 19205006, and 20108001 “pi-Space”) from the Ministry of
Education, Culture, Sports, Science and Technology, Japan. Y.T. thanks
the JSPS fellowship for the young scientist.
[13] a) M. Suzuki, A. Osuka, J. Am. Chem. Soc. 2007, 129, 464–465;
b) M. Suzuki, A. Osuka, Angew. Chem. 2007, 119, 5263–5266;
Angew. Chem. Int. Ed. 2007, 46, 5171–5174.
[14] a) Y. Tanaka, W. Hoshino, S. Shimizu, K. Youfu, N. Aratani, N. Mar-
uyama, S. Fujita, A. Osuka, J. Am. Chem. Soc. 2004, 126, 3046–
[1] a) B. Franck, A. Nonn, Angew. Chem. 1995, 107, 1941–1957; Angew.
Chem. Int. Ed. Engl. 1995, 34, 1795–1811; b) A. Jasat, D. Dolphin,
Chem. Rev. 1997, 97, 2267–2340; c) J. L. Sessler, D. Seidel, Angew.
Chem. 2003, 115, 5292–5333; Angew. Chem. Int. Ed. 2003, 42, 5134–
5175; d) H. Furuta, H. Maeda, A. Osuka, Chem. Commun. 2002,
1795–1804; e) T. K. Chandrashekar, S. Venkatraman, Acc. Chem.
Res. 2003, 36, 676–691; f) S. Shimizu, A. Osuka, Eur. J. Inorg.
Chem. 2006, 1319–1335; g) R. Misra, T. K. Chandrashekar, Acc.
Chem. Res. 2008, 41, 265–279.
[2] a) J. L. Sessler, S. J. Weghorn, V. Lynch, M. R. Johnson, Angew.
Chem. 1994, 106, 1572–1575; Angew. Chem. Int. Ed. Engl. 1994, 33,
1509–1512; b) J. L. Sessler, S. J. Weghorn, Y. Hisaeda, V. Lynch,
Chem. Eur. J. 1995, 1, 56–67; c) E. Vogel, M. Brçring, J. Fink, D.
Rosen, H. Schmickler, J. Lex, K. W. K. Chan, Y.-D. Wu, D. A. Platt-
ner, M. Nendel, K. N. Houk, Angew. Chem. 1995, 107, 2705–2709;
Angew. Chem. Int. Ed. Engl. 1995, 34, 2511–2514; d) M. Brçring, J.
´
˙
3047; b) L. Latos-Graynski, Angew. Chem. 2004, 116, 5234–5238;
Angew. Chem. Int. Ed. 2004, 43, 5124–5128.
[15] S. Saito, K. S. Kim, Z. S. Yoon, D. Kim, A. Osuka, Angew. Chem.
2007, 119, 5687–5689; Angew. Chem. Int. Ed. 2007, 46, 5591–5593.
[16] Re-analysis of the crystal structure of 1 gave better crystal data than
that reported in ref. [12], thus new crystal data is indicated.
[17] The activation volume DV° can be defined by the following equa-
tion: DV° =ÀRT
ACHTUNGTRENNUNG
[18] a) E. Heilbronner, Tetrahedron Lett. 1964, 5, 1923–1928; b) H. E.
Zimmerman, J. Am. Chem. Soc. 1966, 88, 1564–1567; c) H. S.
Rzepa, Chem. Rev. 2005, 105, 3697–3715; d) R. Herges, Chem. Rev.
2006, 106, 4820–4842.
5684
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 5674 – 5685

Splitting Reactions of Metallated [36]Octaphyrins(1.1.1.1.1.1.1.1)
FULL PAPER
[19] a) D. Ajami, O. Oeckler, A. Simon, R. Herges, Nature 2003, 426,
819–821; b) D. Ajami, K. Hess, F. Kçhler, C. Nꢆther, O. Oeckler, A.
Simon, C. Yamamoto, Y. Okamoto, R. Herges, Chem. Eur. J. 2006,
12, 5434–5445.
[20] In the crystal of 5–PdPd two n-heptane molecules were severely dis-
ordered. Thus, electron density due to the disordered n-heptane was
removed by the use of the utility SQUEEZE in PLATON software
contains 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. PLATON/SQUEEZE is an effective tool for taking the contri-
bution of disordered solvent into account in crystal structure refine-
ment; see: a) A. L. Spek, PLATON, A Multipurpose Crystallograph-
ic Tool, Utrecht University, Utrecht, 2005; b) P. van der Sluis, A. L.
Spek, Acta Crystallogr. A 1990, 46, 194–201.
package,
giving
the
following
better
crystal
data:
¯
C88H16F40N8Pd2·C7H8; Mw =2250.02; triclinic; P (No. 2); a=13.03(3),
b=17.31(4), c=21.91(5) ꢁ; a=67.23(10), b=84.84(9), g=88.24(8)8;
V=4540 (18) ꢁ3; Z=2; 1calcd =1.646 gcmÀ3; R1 =0.0975 (I>2.0s(I));
wR2 =0.2633 (all data); GOF=0.848 (I>2.0s(I)). CCDC-716727
Received: February 19, 2009
Published online: April 28, 2009
Chem. Eur. J. 2009, 15, 5674 – 5685
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5685