Proton Order-Disorder Phenomena in a Hydrogen-Bonded Rhodium-η5-Semiquinone Complex: A Possible Dielectric Response Mechanism

Article abstract of DOI:10.1002/chem.201500796

A newly synthesized one-dimensional (1D) hydrogen-bonded (H-bonded) rhodium(II)-η⁵-semiquinone complex, [CpRh(η⁵-p-HSQ-Me₄)]PF₆ ([1]PF₆; Cp=1,2,3,4,5-pentamethylcyclopentadienyl; HSQ=semiquinone) exhi

Full text of DOI:10.1002/chem.201500796

DOI: 10.1002/chem.201500796
Full Paper
&
Proton Dynamics
Proton Order–Disorder Phenomena in a Hydrogen-Bonded
5
Rhodium–h -Semiquinone Complex: A Possible Dielectric
Response Mechanism
[a]
[a]
[a]
[a]
[b]
Minoru Mitsumi,* Kazunari Ezaki, Yuuki Komatsu, Koshiro Toriumi, Tatsuya Miyatou,
[b]
[c]
[c]
[c]
Motohiro Mizuno,* Nobuaki Azuma, Yuji Miyazaki,* Motohiro Nakano,
[d]
[e]
[f]
[f]
Yasutaka Kitagawa,* Takayasu Hanashima, Ryoji Kiyanagi, Takashi Ohhara,* and
[g]
Kazuhiro Nakasuji
Chem. Eur. J. 2015, 21, 9682 – 9696
9682
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Abstract: A newly synthesized one-dimensional (1D) hydro-
that the proton exchange has already occurred in the low-
temperature (LT) phase. Rate constants for the proton ex-
5
gen-bonded (H-bonded) rhodium(II)–h -semiquinone com-
5
À4
À6
plex, [Cp*Rh(h -p-HSQ-Me )]PF ([1]PF ; Cp*=1,2,3,4,5-pen-
change are estimated to be 10 –10 s in the temperature
range of 240–270 K. DFT calculations predict that the proto-
4
6
6
tamethylcyclopentadienyl; HSQ=semiquinone) exhibits
a paraelectric–antiferroelectric second-order phase transition
at 237.1 K. Neutron and X-ray crystal structure analyses
reveal that the H-bonded proton is disordered over two
sites in the room-temperature (RT) phase. The phase transi-
tion would arise from this proton disorder together with ro-
+
nation/deprotonation of [1] leads to interesting hapticity
changes of the semiquinone ligand accompanied by reduc-
tion/oxidation by the p-bonded rhodium fragment, produc-
6
3+
6
ing the stable h -hydroquinone complex, [Cp*Rh (h -p-H Q-
2
2
+
2+
4
+
4
Me4)] ([2] ), and h -benzoquinone complex, [Cp*Rh (h -
À
tation or libration of the Cp* ring and PF6 ion. The relative
p-BQ-Me )] ([3]), respectively. Possible mechanisms leading
4
permittivity e ’ along the H-bonded chains reaches relatively
to the dielectric response are discussed on the basis of the
b
2
+
high values (ca., 130) in the RT phase. The temperature de-
migration of the protonic solitons comprising of [2] and
[3], which would be generated in the H-bonded chain.
13
pendence of C CP/MAS NMR spectra demonstrates that
the proton is dynamically disordered in the RT phase and
Introduction
transfer, so-called proton-coupled electron transfer (PCET) reac-
[2]
tions.
Proton transfer in hydrogen bonds (H-bonds) is one of the
most fundamental reactions in many chemical reactions and
plays an important role in acid–base and enzymatic reactions
The proton potential in OÀH···O-type H-bonds with ordinary
medium strength (d(O···O)>2.7 Š) is an asymmetric double-
minimum potential, but it changes into a symmetric double-
minimum potential for O···O distances ranging from 2.7 to
2.5 Š. For very strong H-bonds (d(O···O)<2.5 Š), it becomes
a single-minimum potential and the proton is stabilized near
the center of the H-bond. The symmetric double-minimum po-
tential plays a central role in investigations of proton transfer
dynamics in H-bonds, which reflects the shape of the potential
surface. Proton dynamics in H-bonds presents two interesting
features that depend on the lightest nucleus: The particle
property of protons moving as classical particles beyond an
energy barrier and the wave property corresponding to quan-
[
1]
in the natural world. Moreover, enzymatic reactions occurring
in the living body, such as those involving cytochrome c ox-
idase in mitochondrial and bacterial membranes, are highly
achieved by proton transfers closely cooperating with electron
[3]
[
a] Dr. M. Mitsumi, K. Ezaki, Y. Komatsu, Prof. Dr. K. Toriumi
Department of Material Science, Graduate School of Material Science
University of Hyogo and, Research Center for New Functional Materials
Graduate School of Material Science, University of Hyogo
3
-2-1 Kouto, Kamigori-cho, Ako-gun, Hyogo 678-1297 (Japan)
tum mechanical proton tunneling across this energy barrier.
E-mail: mitsumi@sci.u-hyogo.ac.jp
À11
Moreover, the timescale of proton transfer may reach 10
–
[
b] Dr. T. Miyatou, Prof. Dr. M. Mizuno
À12
Department of Chemistry
10 s for quasi-symmetric proton potentials, leading to ultra-
Graduate School of Natural Science & Technology
Kanazawa University, Kakuma, Kanazawa, Ishikawa 920-1192 (Japan)
E-mail: mizuno@se.kanazawa-u.ac.jp
[4]
fast proton transfer in H-bonds. In recent years, proton dy-
namics in H-bonds has attracted much attention in materials
[5]
science, because it induces interesting ferroelectric and
[
c] N. Azuma, Prof. Dr. Y. Miyazaki, Prof. Dr. M. Nakano
Research Center for Structural Thermodynamics
Graduate School of Science, Osaka University, 1-1 Machikaneyama
Toyonaka, Osaka 560-0043 (Japan)
[6]
proton-conducting properties. In particular, proton-conduct-
ing materials are expected to be used for application as fuel
cell electrolytes.
E-mail: miyazaki@chem.sci.osaka-u.ac.jp
p-Conjugated electronic systems bearing a b-diketone enol
skeleton, such as 3-hydroxyenone derivatives, can form H-
bonded one-dimensional (1D) chains, in which head-to-tail
[
d] Prof. Dr. Y. Kitagawa
Department of Materials Engineering Science
Graduate School of Engineering Science, Osaka University
1
-3 Machikaneyama, Toyonaka, Osaka 560-8531 (Japan)
linkages generate polarity through the intermolecular H-
E-mail: kitagawa@cheng.es.osaka-u.ac.jp
[5b,7]
bonds.
Interestingly, proton transfer in cooperation with
[
e] Dr. T. Hanashima
the switching of p-bond alternation, that is, tautomerism, can
cause polarization reversal between two energetically equiva-
lent states in these H-bonded p-conjugated systems. In a typi-
cal H-bonded p-conjugated electronic system, such as the
squaric acid (3,4-dihydroxy-3-cyclobutene-1,2-dione) crystal,
molecules are connected by strong H-bonds (O···O=2.554 Š)
to form two-dimensional sheets, which display polarization in
Research Center for Neutron Science and Technology
CROSS Tokai, Ibaraki 319-1106 (Japan)
[
f] Dr. R. Kiyanagi, Dr. T. Ohhara
J-PARC center, Japan Atomic Energy Agency
Tokai, Ibaraki 319-1195 (Japan)
E-mail: takashi.ohhara@j-parc.jp
[g] Prof. Dr. K. Nakasuji
School of Materials Science, Fukui University of Technology
[
8]
the plane and antiferroelectric stacking. Because of thermally
disordered protons, this crystal undergoes an antiferroelectric–
paraferroelectric phase transition at about 373 K to exhibit
3
-6 Gakuen, Fukui 910-8505 (Japan)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201500796.
Chem. Eur. J. 2015, 21, 9682 – 9696
www.chemeurj.org
9683
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[
9]
5
2+
a high dielectric constant reaching 400. Sugawara and Mochi-
da et al. have found three proton dynamics phenomena result-
ing from the controlled dimensionality of the H-bonded p-con-
jugated electron system and the intervention of water mole-
the iridium(II)–h -semiquinone complex monocation [Cp*Ir
5
+
6
(h -p-HSQ)] , which produce the stable h -hydroquinone com-
3
+
6
2+
4
plex dication, [Cp*Ir (h -p-H Q)] , and h -benzoquinone com-
2
+
4
[18]
plex, [Cp*Ir (h -p-BQ)], respectively. Furthermore, the coordi-
[
7b]
[10]
cules in the H-bonds: A protonic soliton and proton relay
in which the proton moves over a barrier as a classic particle
nation of the metal fragments makes these complexes quite
6
acidic. For example, [(h -p-H Q)Mn(CO) ] presents a pK value
2
3
a
[
11]
and a proton tunneling. Furthermore, PCET is not restricted
to biological systems but also occurs in molecular mat-
of about 3, which is significantly smaller than that of free hy-
[19]
5
droquinone (pK ca., 10). As a result, these h -semiquinone
a
[
2a,12]
erials.
Quinhydrone, a H-bonded charge-transfer complex
complexes form remarkably shorter OÀH···O H-bonds (ca.,
[15,17]
[20]
composed of p-benzoquinone (p-BQ) and hydroquinone (p-
2.44–2.47 Š)
than those of quinhydrone (2.74 Š), gener-
H Q), has exhibited a cooperative proton–electron transfer
ating infinite H-bonded 1D chains. Based on these features, we
2
[12a,b]
5
(PET) phenomena under pressures exceeding 30 kbar.
Be-
chose rhodium(II)–h -semiquinone complexes, which are capa-
cause the semiquinone (p-HSQ) neutral radical results from the
ble of exhibiting proton dynamics along with proton-coupled
intramolecular redox ability, as the target compounds.
removal of one proton and an electron from p-H Q or the ad-
2
dition of one proton and an electron to p-BQ, quinhydrone has
been proposed to produce a semiquinone neutral radical by
cooperative proton and electron transfers under high pressure.
Ueda and Mori et al. have recently observed the interesting H-
bond dynamics-based switching of conductivity and magnet-
ism in a purely organic catechol-fused ethylenedithiotetrathia-
Herein, we report a comprehensive study of the synthesis,
heat capacity, X-ray and neutron crystal structure analyses,
density functional theory (DFT) calculations, dielectric proper-
13
ties, and C cross polarization/magic angle spinning (CP/MAS)
NMR spectroscopy performed on
a new one-dimensional (1D) H-
[13]
5
fulvalene (EDT-TTF) conductor crystal, k-H (Cat-EDT-TTF) .
bonded
rhodium(II)–h -semiqui-
3
2
5
The deuteration of its H-bond gives rise to a phase transition
at 185 K, accompanied by significant switching of the p-elec-
tronic structure (Dimer-Mott vs. charge order), electrical con-
ductivity (semiconducting vs. insulating), and magnetism (para-
magnetic vs. nonmagnetic). These changes stem from deuteri-
um transfer or displacement within the H-bond accompanied
by electron transfer between H-bonded p-systems.
none complex [Cp*Rh(h -p-HSQ-
Me )]PF6 ([1]PF6; p-HSQ-Me =
4
4
2,3,5,6-tetramethyl-p-semiquinone;
see Scheme 2) that exhibits a para-
electric-to-antiferroelectric second-
order phase transition at 237.1 K.
We succeeded in determining the
5
Scheme 2. [Cp*Rh(h -p-HSQ-
Me )]PF ([1]PF ).
4
6
6
1
3
The exploration of molecular materials displaying multifunc-
tionality, such as magnetic, electrical, dielectric, or optical prop-
erties, is an important but challenging theme in materials sci-
rate constants for proton exchange in the H-bonds by the C
CP/MAS NMR spectroscopy and demonstrated the relationship
between the dielectric response of the H-bonded proton and
proton exchange rates for the first time. Furthermore, a possi-
ble dielectric response mechanism in the room-temperature
(RT) phase was discussed on the basis of the rapid migration
[
14]
ence. We are developing multifunctional materials showing
interesting magnetic properties in addition to proton dynamics
and ferroelectricity or unique solid-state properties resulting
from cooperative proton and electron transfers. Semiquinone
3
+
6
of the protonic solitons consisting of [Cp*Rh (h -p-H Q-
2
5
2+
2+
typically form stable h -coordinated complexes with metal
Me4)]
([2] ; p-H Q-Me =2,3,5,6-tetramethyl-hydroquinone)
2 4
2
+
+
4
fragments, such as Cp*M (M=Co, Rh, Ir, Cp*=1,2,3,4,5-pen-
and [Cp*Rh (h -p-BQ-Me )] ([3]; p-BQ-Me =2,3,5,6-tetrameth-
4
4
[
15]
+
tamethylcyclopentadienyl),
[Rh(cod)] (cod=1,5-cycloocta-
yl-p-benzoquinone) that would be generated in the H-bonded
[
16]
+ [17]
diene), and [Mn(CO)3]
Protonation and deprotonation of
chain.
.
Results and Discussion
Heat capacity
5
The heat capacity of [Cp*Rh(h -p-HSQ-Me )]PF ([1]PF ) was
4
6
6
measured from 8 to 302 K by adiabatic calorimetry (Figure 1).
A thermal anomaly was observed around 237 K and the meas-
urements in the transition region were repeated four times to
determine the transition temperature. The transition tempera-
ture Ttrs was determined to be 237.1 K. Because the thermal
anomaly exhibits a widespread tail in the low temperature (LT)
region and no supercooling was detected around its peak, the
anomaly is attributed to a second-order phase transition. The
excess heat capacity was calculated by subtracting the normal
heat capacity from the overall heat capacity. The transition en-
thalpy and entropy that were obtained by integrating the
excess heat capacity with respect to T and ln T amounted to
Scheme 1. Representative protonation/deprotonation-intramolecular redox
processes for a iridium(II)–h -semiquinone complex.
5
[18]
these complexes cause interesting stepwise hapticity changes
of the semiquinone ligand accompanied with its reduction/oxi-
dation by the p-bonded metal fragment to satisfy the 18-elec-
[
16,18]
tron rule.
Scheme 1 shows the representative protonation/
deprotonation-intramolecular redox processes observed for
Chem. Eur. J. 2015, 21, 9682 – 9696
www.chemeurj.org
9684
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
sulting from phase transition-induced variations in vibrational
states, the actual transition entropy originating from the pure
order–disorder phase-transition of the protons would be even
À1
À1
smaller than 5.216 JK mol .
X-ray and neutron crystal structure analyses
To ascertain how crystal structure changes with temperature
variations, variable-temperature crystallographic studies of
[
1]PF were performed between 101 and 345 K by using X-ray
6
diffraction. Crystallographic data and intensity data collection
are listed in Table S1 (the Supporting Information) for all struc-
tures and crystal structures are shown in Figure S1. A crystal
structure of [1]PF including exact positions of the hydrogen
6
Figure 1. Molar heat capacity (*), estimated normal heat capacity (c), and
atoms was determined in the LT phase by neutron diffraction
at 83 K. Structures are shown in Figure 2a–2c. Compound
5
excess heat capacity (*) of [Cp*Rh(h -p-HSQ-Me
6 6
)]PF ([1]PF ) as a function
4
of temperature.
[
1]PF6 in the LT phase crystallizes in the monoclinic space
group P2 /c, and the asymmetric unit contains one independ-
1
À1
À1
À1
5
1
.035 kJmol and 5.216 JK mol , respectively. When an indi-
ent [Cp*Rh(h -p-HSQ-Me )]PF unit. Individual monocationic
4
6
5
vidual proton that is among the ordered ones is perfectly dis-
ordered in the H-bonded chain, the transition entropy is
complexes adopt a sandwich configuration consisting of h -
2
+
semiquinone-Rh -Cp*. These cations are arranged along the
b-axis, which are related by the two-fold screw axis, and are
connected through strong OÀH···O H-bonds (O1···O2*=
2.514(5) Š), forming a 1D zigzag chain parallel to the b axis. In
NA
À1
À1
kB ln2 =Rln2=5.763 JK mol . The observed transition en-
tropy is smaller than the Rln2 value. Furthermore, because the
observed transition entropy includes the entropy change re-
5
Figure 2. Crystal structures of [Cp*Rh(h -p-HSQ-Me
4
)]PF
6
([1]PF
6
) in the LT phase determined by neutron diffraction at 83 K a)–c) and in the RT phase character-
ized by X-ray diffraction at 256 K d)–f). a) Structure of the infinite H-bonded zigzag chain parallel to the b-axis with an atomic numbering Scheme (thermal el-
lipsoid set at the 50% probability level). Symmetry operation: *=1Àx, 1/2+y, 1/2Àz, ’=1Àx, À1/2+y, 1/2Àz; b) and c) Packing diagrams projected down to
a- and b-axes, respectively; d) Structure of the infinite H-bonded zigzag chain parallel to the b-axis with an atomic numbering Scheme (thermal ellipsoid set
at the 50% probability level). Symmetry operation: *=Àx, À1/2+y, Àz, ’=Àx, 1/2+y, Àz; e) and f) Packing diagrams projected down to c- and b-axes, re-
spectively. Hydrogen atoms of methyl groups are omitted for clarity in b)–f). Dashed lines represent OÀH···O H-bonds and C···O contacts. The simple vectorial
relationship between unit cells is defined as follows: aRT =1/2cLT, bRT =bLT (1D chain direction), and cRT =aLT
.
Chem. Eur. J. 2015, 21, 9682 – 9696 www.chemeurj.org
9685
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
addition to the strong H-bonds, two types of C···O contacts
and c =a . The temperature dependence of the lattice con-
RT
LT
shorter than the sum of the van der Waals radii of these atoms
stants did not show any jump upon phase transition, excluding
a possibility of first-order phase transition (the Supporting In-
formation, Figure S3). The crystal structure in the RT phase is
(3.22 Š) are detected between adjacent semiquinone ligands
(O1···C8*=3.145(5) and O2···C7’=2.976(5) Š), indicating the
5
presence of CÀH···O H-bonds. The observed O1···O2 H-bond
observed as a superposition of [Cp*Rh(h -p-HSQ-Me )]PF units
4
6
5
length is slightly longer than those observed in [Cp*M(h -p-
having two different orientations related by the glide plane in
the LT phase. The refined site occupancies of the two Cp*
rings are 0.64(1) and 0.36(1) at 256 K, which become closer to
each other with increasing temperature to reach 0.56(2) and
0.44(2), respectively, at 345 K. Similarly to the LT phase, the
monocationic complexes related by the two-fold screw axis are
connected through strong OÀH···O H-bonds to form a 1D
zigzag H-bonded chain parallel to the b-axis. The O1···O2* dis-
tance at 256 K is 2.495(12) Š, which is 0.019 Š shorter than in
the LT phase. The O1···C8* and O2···C7’ contact distances are
3.117(13) and 3.075(14) Š, respectively. The O1ÀC1 and O2ÀC4
HSQ)]CF SO (for M=Rh, 2.46 Š on average; for M=Ir, 2.44 Š
3
3
[
15a]
on average).
This may originate from a decrease in acidity
upon introduction of the methyl groups in [1]PF and/or steric
6
hindrance of the C···O contacts. Nonetheless, this H-bond
[
20]
length is considerably shorter than for quinhydrone (2.74 Š)
[
21]
and duroquinhydrone (2.846(9) Š),
in which 2,3,5,6-tetra-
methyl-p-benzoquinone (p-BQ-Me ) and 2,3,5,6-tetramethyl-hy-
4
droquinone (p-H Q-Me ) are linked by OÀH···O H-bonds. The
2
4
observed strong H-bonds result from an increase in acidity of
the semiquinone ligand upon coordination to the Cp*Rh frag-
ment, in addition to resonance effects occurring in p-conjugat-
ed chains of alternated single and double bonds. The hydro-
gen atom in the H-bond was confirmed to be located near the
O1 atom (O1ÀH1=1.036(8) Š and O2*ÀH1=1.480(8) Š). Two
distances of the p-HSQ-Me ligand at 256 K are 1.330(11) and
4
1.240(12) Š, respectively. The difference between these CÀO
distances is smaller than in the LT phase and decreases with in-
creasing temperature until these distances become almost
identical (the Supporting Information, Table S2). Difference
Fourier maps across the OÀH···O H-bond in the RT phase show
doubly split electron density distribution (the Supporting Infor-
mation, Figure S4). Neutron data confirmed that the hydrogen
atom in the H-bond is disordered over two positions H1 and
H2 with site-occupancy factors of 0.56(5) and 0.44(5), indicative
of a slightly asymmetric double-minimum proton potential for
the H-bond at 256 K, in which the H1 site exhibits a lower
energy than the H2 site. This proton potential approximates
a symmetric double-minimum potential with increasing tem-
perature. In the semiquinone boat-like conformation, dihedral
angles a1 and a2 across C2···C6 and C3···C5 are 13.8 and
10.58, respectively, and show a smaller difference than in the
LT phase.
CÀO distances in p-HSQ-Me ligand are O1ÀC1=1.326(4) and
4
O2-C4=1.258(4) Š, respectively, which differ slightly from
those of duroquinhydrone (CÀO=1.373(6) Š in p-H Q-Me and
2
4
1
.219(4) Š in p-BQ-Me ) because of the strong H-bonding inter-
4
action. A distinct difference of 0.068 Š between the CÀO dis-
tances is consistent with the H-bonded hydrogen atom posi-
tion. Therefore, H-bonded chain is polarized along the b axis.
5
Furthermore, the h -semiquinone ligand adopts a boat-like
conformation with dihedral angles a1 and a2 of 8.44 and
1
8.148 across C2···C6 and C3···C5, respectively. In contrast,
4
both dihedral angles of a1 and a2 in h -benzoquinone com-
4
4
plexes [CpRh(h -p-BQ-Me )] and [(C H )Rh(h -p-BQ-Me )] (Cp=
4
9
7
4
cyclopentadienyl, C H =indenyl) are 23 and 258, respective-
9
7
[
22]
ly.
The average distance of four RhÀC(diene) is 2.23(2) Š,
whereas RhÀC(1) and RhÀC(4) distances are 2.289(4) and
In addition to the disorder accompanying the phase transi-
À
2
.456(4) Š, respectively. The RhÀC(4) distance is especially
tion, Cp* ring and PF6 anion display significantly larger ther-
about 0.2 Š longer than the other five RhÀC distances, indica-
mal ellipsoids in the RT phase than in the LT phase. To extract
information about the thermal motions of these moieties, tem-
perature dependences of atomic displacement ellipsoids and
anisotropic atomic displacement parameters (ADPs, U , U ,
5
tive of an h -p-semiquinone coordination mode. The dihedral
angle between Cp* and p-HSQ-Me ligands is 0.248. The aver-
4
age distance for five RhÀC(Cp*) is 2.172(8) Š. The H-bonded
chains arranged along the c-axis are related by the glide plane.
Therefore, polarizations generated in the chains canceled each
other out in the antiferroelectric arrangement.
11
22
and U ) were evaluated (the Supporting Information, Figur-
33
es S5–S10). The Cp* ring exhibits slightly larger thermal ellip-
soids than the p-HSQ-Me ligand even at 101 K. These ellip-
4
X-ray and neutron crystal structures of [1]PF in the room-
soids are increased continuously from the LT phase instead of
the transition point (the Supporting Information, Figures S7
and S9), indicating the potential for libration or rotation of the
Cp* ring around the RhÀCp* ring centroid vector. The Cp*
ligand is known to undergo rotational motion, called 2p/5
6
temperature (RT) phase determined at 256 K are shown in Fig-
ures 2d–2 f and S2 (the Supporting Information), respectively.
The discussion mainly focuses on X-ray diffraction data. Neu-
tron diffraction results could not be analyzed with sufficient ac-
curacy because of very weak diffraction intensities in the large
sinq/l region originating from the significant disorder of the
[23]
13
jumping motion.
Also, temperature-independent C reso-
nances corresponding to this rotational motion have been ob-
À
PF6 anion and the Cp* ring, especially methyl hydrogens in
served in [(Cp*Rh) (m-CH ) (m-O SSO )], which presents thermal
2
2 2
2
2
the Cp* ring. Upon second-order phase transition, the space
ellipsoids of almost the same size as those observed in [1]PF
6
[24]
group of compound [1]PF changes from P2 /c in the LT phase
at 101 K. These observations suggest that the Cp* ring of
6
1
to P2 in the RT phase and the unit-cell volume in the RT
[1]PF starts rotating at low temperature, and this rotation is
1
6
phase becomes half of the LT phase by losing a glide plane. A
simple vectorial relationship is defined between the unit cells,
in which a =1/2c , b =b (H-bonded 1D chain direction),
accelerated with increasing temperature. Investigation of this
13
dynamic behavior by using C CP/MAS NMR spectroscopy will
À
be discussed below. The thermal ellipsoids of the PF6 anion
RT
LT
RT
LT
Chem. Eur. J. 2015, 21, 9682 – 9696
www.chemeurj.org
9686
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
2
are also increased continuously with increasing temperature
from the LT phase instead of the phase transition (the Support-
ing Information, Figures S5 and S10). Therefore, the structural
phase transition observed in compound [1]PF6 would result
from the loss of the c-glide plan symmetries by the proton dis-
order-related averaging of the semiquinone ligand structure
ferromagnetic and the <S > value for the singlet state ap-
proximates zero, indicating a closed-shell electronic structure
(the Supporting Information, Table S4). The highest occupied
(HOMO) and lowest unoccupied molecular orbitals (LUMO) are
delocalized over the entire molecule and comprise of an in-
phase and out-of-phase combination of the p orbitals of the
Cp* and semiquinone ligands, respectively (Figure 3).
À
and by the rotation or libration of Cp* ring and PF6 ion.
The LUMO and LUMO+1 are almost energetically equivalent
and are doubly degenerate. In contrast, the HOMOÀ1 predom-
inantly comprises a semiquinone p orbital centered on the car-
bonyl group with some contributions from a Rh 4dx2Ày2 orbital.
DFT calculations
DFT calculations were performed to elucidate the electronic
structures of the monocationic semiquinone complex
The HOMO–LUMO gap amounts to 3.697 eV. The optimized
5
+
+
+
[
Cp*Rh(h -p-HSQ-Me )] ([1] ) and its protonated/deprotonat-
structure of [Cp*Rh(p-HSQ-Me
4
)] exhibits almost the same
4
5
+
ed states. Frontier molecular orbitals and their energies (eV) for
structure as the h -semiquinone monocationic complex [1]
+
[
1] obtained using neutron diffraction data simulated at the
determined by neutron diffraction (the Supporting Informa-
[25]
B3LYP/6-31G* level using the Gaussian 09 program
are
tion, Table S3). However, the dihedral angles a
a boat-like conformation of p-HSQ-Me4 ligand are 3.74 and
3.078, respectively, differing from neutron diffraction data by
1 2
and a of
2
Æ58. The dihedral angle between Cp* and p-HSQ-Me ligands
4
is 5.598, exceeding neutron data by about 58. These differen-
ces between experiment and theory may be attributed to the
strong H-bonding interactions in the actual crystal structure.
The HOMO and HOMOÀ1 in the optimized structure is slightly
stabilized due to the difference between observed and opti-
mized structures. The CÀC and CÀO distances in the p-BQ-Me
4
ligand of [Cp*Rh(p-BQ-Me )] show a p-quinoid-type distortion.
4
Specifically, average distances of 1.419, 1.483, and 1.236 Š are
obtained for two alternating short CÀC bonds, four longer CÀC
bonds, and two CÀO bonds, respectively, consistent with ben-
zoquinone
complexes
(the
Supporting
Information,
[18b,26]
Table S3).
The dihedral angles a and a of a boat-like
1 2
conformation of the p-BQ-Me ligand are 22.3 and 22.48, in
4
4
agreement with those observed for the h -benzoquinone com-
4
[22]
2+
plex [CpRh(h -p-BQ-Me )] (238). In [Cp*Rh(p-H Q-Me )] , the
4
2
4
p-H Q-Me ligand presents CÀC and CÀO distances of 1.429–
2
4
1.435 Š (six-membered ring) and 1.338 Š (average for two
bonds), respectively, coinciding with those of the hydroqui-
[
27]
none state. The dihedral angles a and a of the p-H Q-Me
4
1
2
2
6
ligand are 7.30 and 7.538, approximating those of h -hydroqui-
none complex (nBu N) [Rh(cod)(h -p-H Q)](BF ) (11.3 and
6
4
4
2
4 3
[
27a]
1
2.38).
The HOMOs and LUMOs in benzoquinone complex
+
4
[
[
Cp*Rh (h -p-BQ-Me )] ([3]) and hydroquinone complex
4
3
+
6
2+
2+
Cp*Rh (h -p-H Q-Me )] ([2] ) are also delocalized over the
5
2
4
Figure 3. Frontier molecular orbitals and energies [eV] for [Cp*Rh(h -p-HSQ-
Me )] ([1] ) obtained using neutron diffraction data at the B3LYP/6-31G*
4
level (MIDI+polarization function for Rh ion).
+
+
entire molecule and the Cp* and benzoquinone/hydroquinone
ligands in each compound interacts through an in-phase and
out-of-phase combination, respectively. Their LUMO and
LUMO+1 are almost energetically equivalent and are doubly
+
shown in Figure 3. Geometrical structures of hydroquinone di-
degenerate. These tendencies are identical to those of [1] . In
2
+
cationic complex, [Cp*Rh(p-H Q-Me )] , and benzoquinone
contrast, the HOMOÀ1 and HOMOÀ2 in [3] are mainly benzo-
quinone p orbitals with an out-of-phase and in-phase combi-
nation of two carbonyl orbitals, respectively. As expected, the
2
4
complex, [Cp*Rh(p-BQ-Me )], together with that of [Cp*Rh(p-
4
+
HSQ-Me4)] were also optimized by using B3LYP-D, in which
2
+
dispersion force corrections are applied to take account of p–p
interactions between Cp* and semiquinone. Their frontier mo-
lecular orbitals are shown in Figure S11 and S12, and their
bond lengths are listed in Table S3 (the Supporting Informa-
hydroquinone complex [2] did not show any frontier orbital
involving the carbonyl orbital. As previously observed in
0
5
2+
5
[16,18]
[Rh (cod)(h -p-HSQ)] and [Cp*Ir (h -p-HSQ)]BF ,
these DFT
4
results clearly demonstrate that the protonation/deprotonation
5
tion). DFT calculations confirmed that the spin coupling state
of the h -semiquinone complex accompanies the simultaneous
+
between the rhodium ion and p-HSQ-Me ligand in [1] is anti-
reduction/oxidation by the p-bonded rhodium fragments,
4
Chem. Eur. J. 2015, 21, 9682 – 9696
www.chemeurj.org
9687
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
r
Figure 6. Temperature dependence of the real part e ’ of the complex dielec-
tric permittivity measured at 100 kHz along [100], [010] (H-bonded chain di-
rection), and [001] directions.
+
4
Figure 4. Comparison between frontier orbital energy levels of [Cp*Rh (h -
sured to probe the dynamics of the H-bonded protons. These
samples were prepared by cutting the single crystals perpen-
dicularly to the crystallographic [100] and [010] directions. Re-
sults are shown in Figure 6 and S13 (the Supporting Informa-
2
+
5
+
+
3+
6
4 4 2
p-BQ-Me )] ([3]), [Cp*Rh (h -p-HSQ-Me )] ([1] ), and [Cp*Rh (h -p-H Q-
2
+
2+
4
Me )] ([2] ).
which act as internal electron donors/acceptors, to produce
tion). The value of e ’ measured by applying an electric field
b
6
4
the stable h -hydroquinone/h -benzoquinone complexes.
along the H-bonded chain (Ej j[010]) is 127 at 300 K, which is
+
Figure 4 compares the frontier orbital energy levels of [3], [1] ,
much larger than e ’ (5.0) and e ’ values (3.3) measured along
a
c
2
+
and [2] . These energy levels decrease in the order of benzo-
quinone, semiquinone, and hydroquinone complexes instead
of increasing, consistent with charge effects. However, no sig-
nificant difference in HOMO–LUMO gaps and orbital splitting
widths were observed between these compounds. Molecular
dipole moments obtained by B3LYP-D/6-31G* calculations are
shown in Figure 5. Despite the absence of significant differ-
ence in orbital splitting patterns, the three complexes present
dramatically altered dipole moments. Especially, because the
[100] and [001] directions, respectively, indicative of the high
anisotropy of dielectric response. When the temperature is de-
creased, e ’ gradually increases to 138 at 251 K before exhibit-
b
ing a relatively sharp transition near 238 K, which is in good
agreement with specific heat measurements and variable-tem-
perature crystal structure analyses. The value of e ’ is decreased
b
to about 6.6 at 12 K in the LT phase. Moreover, the transition
temperature rises to 274 K upon deuterium substitution of the
H-bonded hydrogen atoms (the Supporting Information, Fig-
ure S14). This isotope effect stems from a decrease in the zero-
point energy of deuterium, which is heavier than hydro-
+
electric dipole of [1] (5.63 D) is oriented along the chain di-
rection, proton transfer and bond alternation exchange would
significantly change polarization in this direction.
[7b,9b,13,28]
gen.
This reveals that the proton (deuteron) order–dis-
order transition in the H-bonds plays a major role in the phase
transition mechanism. The e ’ value in the RT phase is only
b
2
slightly decreased when the frequency is increased from 10 to
6
1
0 Hz (the Supporting Information, Figure S13), indicating that
the motion of the electric dipoles easily follow electric fields
oscillating at extremely high frequency.
The electric field (E) dependence of electric polarization (P)
was measured by applying an electric field along the H-
bonded chains (Ej j[010]) above and below the phase-transi-
tion temperature (T =237.1 K) to examine the potential for
trs
ferroelectricity of [1]PF in the LT phase (Figure 7). No hystere-
6
sis loops were observed regardless of temperature under an
+
4
À1
4
Figure 5. Calculated molecular dipole moments of [Cp*Rh (h -p-BQ-Me )]
applied electric field up to Æ45 kVcm , indicating that this
2
+
5
+
+
3+
6
2+
(
(
4 2 4
[3]), [Cp*Rh (h -p-HSQ-Me )] ([1] ), and [Cp*Rh (h -p-H Q-Me )]
compound did not show ferroelectricity. The opening of the P–
E loop is not attributed to ferroelectricity but to the heat-in-
duced increase in conductivity. Although the H-bonded chain
is polarized itself because of the localization of protons in the
LT phase, this absence of ferroelectricity would result from the
mutual cancelation of the polar chains related by glide plane
in the antiferroelectric arrangement.
2
+
[2] ).
Dielectric properties
The temperature dependence of the real part of the complex
dielectric permittivity e ’ of crystal samples of [1]PF was mea-
r
6
Chem. Eur. J. 2015, 21, 9682 – 9696
www.chemeurj.org
9688
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Carbonyl and phenolic carbon resonance regions are shown
in Figure 8. Compound [1]PF at 145 K exhibits single C=O and
6
CÀOH carbon resonances at d=159.1 (A2) and 134.9 (A1) ppm,
respectively, which is in agreement with their different environ-
ments observed by crystal structure analysis. When the tem-
perature is increased, these signals gradually approach each
other in the LT phase, consistent with a typical line-shape
change resulting from a two-site exchange, and coalesce into
a single peak at d=146.3 ppm at 242 K. This demonstrates
that the disorder of H-bonded hydrogen atom observed in the
crystal structure is not static but dynamic in a double mini-
mum potential on the NMR timescale and moreover the
proton exchange has already occurred in the LT phase. This
agrees with the second-order phase transition observed in
heat capacity measurements. The averaged chemical shifts be-
tween resonances A1 and A2 do not match that of resonance
A. Therefore, proton localization in the H-bonds causes a distor-
tion of the chemical structure of cationic complex and this
phase transition is an order–disorder phase transition accom-
panied by displacive transition. Previous variable-temperature
Figure 7. Temperature dependence of polarization versus electric field (P–E)
curve of [1]PF measured at 10 Hz along [010] at different temperatures.
6
13
C CP/MAS NMR spectroscopy
In general, the crystal structure analysis cannot distinguish
static and dynamic disorders because it only detects time-aver-
aged atomic positions. In contrast, high-resolution solid-state
NMR spectroscopy is a powerful tool for investigating thermo-
dynamics and kinetics in dynamic systems. Variable-tempera-
13
C CP/MAS NMR investigations in squaric acid have also re-
vealed that the phase transition consists of a mixed order–dis-
[29]
order plus displacive transition. However, kinetic analysis of
the order–disorder transition was not performed in these stud-
1
3
ture C CP/MAS NMR spectra of the compound [1]PF were
ies. Observed spectra for [1]PF were simulated by using an in-
6
6
measured in the temperature range of 145–293 K to elucidate
the proton transfer dynamics in the H-bonds and the dynamic
house FORTRAN program, in which the model assumes chemi-
cal exchange between two sites. The temperature dependence
of rate constants k derived from this line shape simulation is
shown in Figure 11 (see later). The computed rate constants re-
vealed that proton transfer in the temperature range of 240–
13
behavior of the Cp* ligand. Temperature dependences of
C
CP/MAS NMR spectra, divided into three resonance regions
and their chemical shifts, are shown in Figures 8–10 and the
À4
À6
2
70 K occurs on a timescale of 10 to 10 s. The rate constant
À8
at RT is estimated to 10 s by extrapolation. Therefore, the
almost frequency-independent behavior of e ’ observed in the
b
RT phase results from the ultrafast proton transfer. Rate con-
stants in the RT phase can be fitted using the Arrhenius law
À1
with activation energy of 76 kJmol . The rate constants rapid-
À2
ly drops to 4.0”10 s at about 240 K accompanying the
second-order phase transition, consistent with the temperature
dependence of the dielectric response. The activation energy
of the proton exchange in the LT phase is estimated to
À1
3
4 kJmol from the Arrhenius plot, suggests that proton ex-
change occurs in the LT phase despite its extremely slow rate.
The reason for the higher activation energy in the averaged RT
phase than in the ordered LT phase is unclear. Although pro-
tons may tunnel from one oxygen atom to the other because
of the presence of strong H-bonds (O···O=2.514(5) Š), the ex-
perimental detection of this proton tunneling was not feasible.
The methyl carbon resonance region is shown in Figure 9.
At 293 K, compound [1]PF only displays one sharp resonance
6
13
Figure 8. Experimental (black lines) and simulated (gray lines) C CP/MAS
NMR spectra of carbonyl and phenolic carbon resonance regions for [1]PF
A color version of this Figure is available in the Supporting Information.
B at d=7.7 ppm assigned to the methyl carbon of the Cp*
ligand and two slightly broad resonances C and D at d=10.8
and 12.1 ppm attributed to the methyl carbons of the p-HSQ-
6
.
Me ligand. Peak B is essentially temperature independent, in-
4
13
Supporting Information, Figure S15, respectively. Isotropic
C
dicating that the Cp* ring is rotated at a high speed or its five
13
chemical shifts calculated by DFT for a dimer model by using
neutron diffraction data at 83 K are shown in Figure S16 (the
Supporting Information).
C atoms are exposed to equivalent environments because of
the high symmetry of the Cp* ring surroundings. However,
these actual environments are not equivalent. As mentioned
Chem. Eur. J. 2015, 21, 9682 – 9696
www.chemeurj.org
9689
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
1
3
13
Figure 10. Experimental (black lines) and simulated (gray lines) C CP/MAS
Figure 9. Experimental (black lines) and simulated (gray lines) C CP/MAS
6
NMR spectra of the aromatic carbon resonance region for [1]PF . A color ver-
6
NMR spectra of the methyl carbon resonance region for [1]PF . A color ver-
sion of this Figure is available in the Supporting Information.
sion of this Figure is available in the Supporting Information.
above, the anisotropic atomic displacement parameters of the
Cp* ring carbons increased continuously with increasing tem-
perature in the LT phase, suggesting that the Cp* ring is al-
ready rotated around the RhÀCp* ring centroid vector in this
phase. Therefore, it is natural to consider that the Cp* ring is
resonance E and strong overlapped resonances F+G at d=
99.3 and 102.8 ppm, respectively. These resonances split into
five signals in the LT phase and can be simulated by using
a model that assumes two kinds of exchanges. According to
these simulations, the overlapped resonances F+G comprise
5
rotating faster than the fast rotation limits (kꢀ10 Hz) at 145 K.
one Cp* ring resonance G and two equivalent CÀCH carbon
3
Whereas, as described in Crystal Structure section, two methyl
groups of the C7 and C8 atoms, which are formed short inter-
molecular C···O contacts with the neighboring O2* and O1’
atoms, are assumed to exhibit nearly equivalent chemical envi-
ronments in the RT phase because of dynamic proton disorder.
Also, their surrounding environments clearly differ from two
methyl groups of the C9 and C10 atoms. According to the DFT
calculations, the resonances of the methyl carbons presenting
short C···O contacts appear in the lower magnetic field than
those without C···O contacts. Therefore, resonances D and C
are attributed to C7, C8 and C9, C10 carbon resonances, re-
spectively. Moreover, the separation to phenolic and carbonyl
oxygen atoms caused by the localization of protons in the LT
phase is expected to generate environmental differences be-
tween the two methyl groups of C7 and C8 atoms as well as
those of the C9 and C10 atoms. The electron-withdrawing C=O
(diene carbon) resonances F for the p-HSQ-Me ligand. Reso-
4
nance G is essentially temperature independent and corre-
sponds to the rotation motion of the Cp* ligand, whereas reso-
nance F further splits into two signals F1 and F2 in the LT
phase. Resonance E, which is assigned to the remaining two
equivalent CÀCH3 carbon (diene carbon) resonances, is also
split into two signals (E1 and E2) upon localization of the H-
bonded protons. The splitting of the CÀCH carbon (diene car-
3
bons) resonances in the RT phase may be related to the non-
equivalence caused by the short C···O contacts of the methyl
groups. Also, DFT calculations suggest that the diene carbon
resonances of sites forming short C···O contacts are shifted
downfield compared with those without short C···O contacts.
Therefore, resonances F and E are assigned to the C2, C3, and
C5, C6 carbons, respectively. The separation into the C=O and
CÀOH groups caused by the localization of the H-bonded pro-
tons in the LT phase may also lead to intramolecular environ-
mental differences between their ortho carbon atoms. Accord-
ing to DFT calculations, diene carbon resonances in the ortho
positions with respect to the C=O group shift to a higher mag-
netic field compared to those in the ortho positions with re-
spect to the CÀOH group. Therefore, resonances F1 and F2 are
tentatively assigned to C3 and C2 carbons, respectively, where-
as resonances E1 and E2 are attributed to C5 and C6 carbons,
respectively. The rate constants k determined using diene
carbon resonances in the LT phase are remarkably small com-
pared with those derived from the carbonyl and phenolic
carbon resonances (Figure 11). However, this phenomenon is
unclear.
group is expected to shift CÀCH carbon resonances connected
3
to the carbon atoms in its ortho positions downfield, which is
in agreement with DFT results. Taking these environmental dif-
ferences into consideration, resonances D2 and D1 in the LT
phase are attributed to C8 and C7 carbons, respectively. The
broad resonance C slightly splits into resonances C2 and C1
below 170 K, which are assigned to C9 and C10 carbons, re-
spectively. Although rate constants k determined from the
methyl carbon resonances only comprise two points, their
values closely matches those derived from the carbonyl and
phenolic carbon resonances.
The aromatic carbon resonance region is displayed in
Figure 10. At RT, this aromatic region shows a slightly broad
Chem. Eur. J. 2015, 21, 9682 – 9696
www.chemeurj.org
9690
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 11. Temperature dependence of rate constants k derived from the
line shape simulation of chemical exchange between two sites in three dif-
ferent resonance regions: Carbonyl carbon and phenolic carbon resonance
region (*), aromatic carbon resonance region (+), and methyl carbon reso-
nance region ( ).
&
Proton dynamics and dielectric response
As mentioned above, the relative permittivity of compound
[1]PF6 along the 1D chain direction increased rapidly upon
second-order phase-transition, in association with proton hop-
ping in the H-bond. Moreover, the transition temperature rose
to 274 K upon deuterium substitution. These facts clearly dem-
onstrate that the observed dielectric response would result
from proton dynamics in the H-bonds. Because semiquinones
belonging to p-conjugated systems undergo internal conver-
sions depending on proton positions, the dielectric response
would more precisely derive from the dipole moments gener-
ated not only by H-bonded protons but also by internal con-
versions of p-conjugated system including the p-bonded rho-
dium fragment.
Now, we discuss the relationship between the proton dy-
namics and dielectric response. As an influence of the ordered
proton arrangement on the dielectric response, two cases of
long-range and short-range protonic correlations are expected
despite the exceptionally short timescale of proton hopping.
First, let us consider the case in which all protons in the or-
dered H-bonded chain collectively move in the same direction
Figure 12. Proton arrangements in H-bonded chains and proton potential
energies in each 1D chain. a) H-bonded monocationic chain comprising or-
dered protons, b)–d) short-range protonic correlations in H-bonded mono-
3
+
6
2+
cationic chains; b) a pair of kink defects involving [Cp*Rh (h -p-H
2
Q-Me
4
)]
2
+
+
4
(
[2] ) and [Cp*Rh (h -p-BQ-Me )] ([3]) formed by proton transfer between
4
2
+
adjacent molecules; c) and d) kink defects involving [2] or [3]. Polarities
on both sides of the kink adopt opposite directions from one another for
(Figure 12a). The ground state of this system is degenerated
2+
short-range orders. c), c’) d) and d’) Migration of [2] or [3] protonic soliton
doubly with respect to the proton orientations in the bistable
H-bonded chain. The activation energy required for proton
transfer corresponds to the barrier height in the double mini-
mum hydrogen potential. However, the collective movement
of a large number of protons in the same direction requires
much more energy to overcome thermal fluctuation. The inter-
conversion of two ground states is also considered energetical-
ly disadvantageous because, unlike the inside of the chain, the
environments of H-bonded chain ends are not reversible. Fur-
thermore, when protons move collectively in the same direc-
tion in response to an alternating current electric field, the rel-
ative permittivity is expected to exceed several thousands.
However, even though H-bonds are inclined by about Æ608
generated in the 1D monocationic chain by applying an electric field;
b’) Separation of a pair of kink defects. All Cp*Rh fragments, except for a)
and b), and all methyl groups are omitted for clarity. A color version of this
Figure is available in the Supporting Information.
against the H-bonded chain, the relative permittivity of com-
pound [1]PF does not reach such high values. Therefore, the
6
observed dielectric response is clearly attributed to the short-
range order of the electric dipoles instead of their long-range
order.
Next, we discuss the case in which proton arrangements in
the H-bonds depend on the short-range protonic correlation
at a very short timescale. As mentioned for Scheme 1 and DFT
Chem. Eur. J. 2015, 21, 9682 – 9696
www.chemeurj.org
9691
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
+
calculations, the protonation and deprotonation of the [1]
in the RT phase are considered to present a symmetric or
nearly symmetric double-minimum potential surface, because
protons are dynamically disordered between two sites. In con-
trast, proton potential surfaces in kink defects in Figure 12c
and 12d are expected to become slightly unstable and stable,
respectively, compared with that of the short-range ordered
protons because of addition or elimination of a proton carry-
monocation are expected to cause stepwise hapticity changes
5
of the h -semiquinone ligand coupled with the reduction/oxi-
dation by the p-bonded rhodium fragment to give the stable
6
3+
6
2+
2+
h -hydroquinone complex, [Cp*Rh (h -p-H Q-Me )]
([2] ),
2
4
4
+
4
and h -benzoquinone complex, [Cp*Rh (h -p-BQ-Me )] ([3]), re-
4
spectively. Therefore, in this system, proton transfer should be
considered in conjunction with this hapticity changes of the
semiquinone ligand coupled with redox reactions instead of
accounting for a simple proton transfer that generates unsta-
[7b,31a]
ing a positive charge.
These reduced barrier heights are
considered to facilitate proton transfer to the neighboring site
in Figure 12c or from the neighboring site in Figure 12d.
When electric fields are applied to the H-bonded chains, the
generated protonic solitons (charged kink solitons) migrate
rapidly in the H-bonded chain through successive proton
transfers accompanied by polarity inversion. Therefore, the die-
lectric response is considered to occur by expanding the short-
range ordered domain possessing a stable polarity with re-
spect to the applied electric field. In this case, the effect of
charged soliton, that is, proton conductivity, is expected to be
small in this system because [3], carrying a negative charge in
5
ble protonated and deprotonated h -semiquinone complexes,
2
+
5
2+
2+
5
[
Cp*Rh (h -p-H SQ-Me )] and [Cp*Rh (h -p-SQ-Me )]. Thus,
2
4
4
kink defects as shown in Figure 12b–12d are expected to form
topologically at the boundary of short-range order regions
when the arrangement of the protons is kept short-range
orders.
A pair of kink defects in the H-bonded chain (Figure 12b)
formed by proton transfer between adjacent molecules in the
2
+
ordered chain comprises [2] and [3] and polarities of short-
range orders on both sides of the kink are the same. In this
case, the polarization reversal requires a collective movement
of protons in the same direction, which is energetically disad-
vantageous, similar to that in the long-range protonic correla-
tion (Figure 12a). When kink defects in the chains consist of
2
+
the cationic chain (Figure 12c), and [2] , carrying a positive
charge in the cationic chain (Figure 12d), have to move against
the applied electric field. The loss tangent (tan d) of [1]PF6
could not be measured properly because of the very low
proton conductivity. Even in the case shown in Figure 12b, the
dielectric response is expected to result from the migration of
protonic solitons when the protonic solitons (Figure 12c and
12d) are generated by the separation of kink defect pairs in
the H-bonded chain (Figure 12b’). In addition, the relative per-
mittivity in the RT phase would be less dependent on the ac
electric field frequency, because the proton exchange rate is
2
+
[
2] (Figure 12c) and [3] (Figure 12d), polarities of short-range
orders on both sides of the kink are of opposite directions. To
maintain the short-range protonic correlation, these kink de-
fects are assumed to coexist topologically within the same H-
bonded chain. According to DFT calculations, because benzo-
quinone and hydroquinone complexes formed by deprotona-
tion and protonation of the semiquinone complex receive de-
stabilization and stabilization of the same degree compared
with their precursor, their simultaneous generation would not
be energetically disadvantageous. These kink defects are
known as protonic solitons in the 1D H-bonded system. The
soliton model for proton transfer (the ADZ model) was first
proposed by Antonchenko, Davydov, and Zolotaryuk to de-
4
5
extremely fast above 10 –10 Hz and the rapid migration of
the protonic solitons easily follows this electric field. The short-
range protonic correlation operating in the RT phase is expect-
ed to gradually disappear with increasing temperature because
of thermal fluctuations. Therefore, the relative permittivity in
the RT phase is considered to decrease with increasing temper-
ature. Conversely, the thermal energy responsible for proton
hopping becomes insufficient in the LT phase and dipolar fluc-
[
1a,30]
scribe collective proton transport in ice.
This protonic soli-
ton mechanism has been proposed to explain the dielectric re-
sponse observed in 1D H-bonded chain systems of bisquaric
tuations induced by proton hopping slow down below T , as
trs
13
observed in C CP/MAS NMR spectra. As a result, the relative
À
acid and dabcoHX (dabco=1,4-diazabicyclo[2.2.2]octane, X =
permittivity decreases rapidly below Ttrs.
À
À
À
[7b,31]
Br , I , and BF ).
In strongly H-bonded 1D solids of 3-hy-
4
droxyenone derivatives, their IR spectra exhibit a low-lying
Conclusion
À1
weak band around 1800 cm , called the S band, which be-
5
comes sharper with decreasing temperature, in addition to
In this work, a new 1D H-bonded rhodium–h -semiquinone
À1 [7]
5
a broad OÀH stretching mode around 2500 cm . The S band
complex, [Cp*Rh(h -p-HSQ-Me )]PF ([1]PF ), which exhibits
a second-order phase transition at T =237.1 K was synthe-
4
6
6
has been assigned to an OÀH stretching vibration arising from
around kink-type defects formed in H-bonded chains, because
the force constant of this vibration decreases around these de-
trs
sized. Neutron and X-ray crystal structure analyses clearly re-
vealed that the H-bonded proton is disordered over two sites
in the RT phase. This proton disorder together with the rota-
[
7a]
fects. In addition to a broad OÀH stretching mode around
À1
À
2
400 cm , compound [1]PF also exhibits a weak and broad
tion or libration of the Cp* ring and PF6 ion caused the loss
6
À1
band at 1720 cm , assigned to the S band, supporting the ex-
istence of kink defects (the Supporting Information, Figur-
es S17 and S18). Furthermore, the transition entropy observed
of the c-glide plane symmetry, resulting in the second-order
phase transition. The dielectric property of compound [1]PF6
was highly anisotropic, and its e ’ values along the H-bonded
b
for [1]PF was smaller than R ln 2, which corresponds to fully
independent proton disorders, consistent with the presence of
short-range ordered protons. The short-range ordered protons
chains reached about 130 in the RT phase. These dielectric per-
mittivity values showed little frequency dependence between
6
2
6
10 and 10 Hz, indicating that the motion of the electric di-
Chem. Eur. J. 2015, 21, 9682 – 9696
www.chemeurj.org
9692 ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
poles easily follows a high-frequency oscillating electric field.
ganic Elemental Microanalysis of the Graduate School of Pharma-
ceutical Sciences, Kyoto University.
13
The temperature dependence of the C CP/MAS NMR spectra
clearly demonstrated that the H-bonded hydrogen atom is dy-
namically disordered in a double minimum potential on the
NMR timescale in the RT phase. Moreover, proton exchange
has already occurred in the LT phase, which is in agreement
with the second-order phase transition. The estimated rate
2,3,5,6-Tetramethyl-hydroquinone (p-H
synthesized according to a previously published procedure.
solution of [SnCl ]·2H O (10.977 g, 48.65 mmol) dissolved in 2m
Q-Me ): p-H Q-Me was
2
4 2 4
[34]
A
2
2
HCl (100 mL) was added to a vivid yellow suspension of 2,3,5,6-tet-
ramethyl-p-benzoquinone (4.005 g, 24.39 mmol) in ethanol (20 mL),
with stirring. The resulting dark-grayish brown suspension was
stirred at 958C for 1 h until its color changed to light orange, and
then was allowed to cool to room temperature. The resulting pre-
cipitate was filtered off, washed two times with water, and dried
under vacuum to give a light orange powder. This crude product
À4
constants k for the proton exchange are ranged within 10 to
À6
10
s in the temperature range of 240–270 K, suggesting that
the frequency independent behavior of e ’ observed in the RT
b
phase stems from the very fast proton transfer. These con-
À2
(3.956 g) was dissolved in ethanol (140 mL) containing a solution
stants rapidly dropped below 10 s in the LT phase. This tem-
of [SnCl ]·2H O (364 mg, 1.61 mmol) in 2m HCl (7 mL) with heating
2
2
perature dependence of the rate constants coincides well with
the dielectric response of the H-bonded protons. Thus, we suc-
ceeded in demonstrating the relationship between the dielec-
tric response of the H-bonded proton and proton exchange
rates for the first time.
at 908C, filtered using a polytetrafluoroethylene (PTFE) cannula
equipped with a glass-fiber filter, and recrystallized to afford the
product as colorless needle crystals (3.055 g, 18.38 mmol, 75%
1
yield). H NMR (600 MHz, [D ]DMSO, TMS): d=2.02 (s, 12H; CH ),
6
3
7
.32 ppm (s; 2H, OH).
Long-range protonic correlations are energetically disadvan-
tageous in cases where a very fast exchange of protons occurs.
Therefore, the observed dielectric response should be attribut-
ed to short-range protonic correlations. To maintain short-
range protonic correlations, kink defects known as protonic
solitons are expected to form topologically on the boundary of
short-range order areas. In agreement with previous reports on
5
[
Cp*Rh(h -p-HSQ-Me )]PF ([1]PF6): Silver hexafluorophosphate
4
6
(493.7 mg, 1.953 mmol) was added to a vigorously stirring suspen-
sion of [Cp*Rh(m-Cl)Cl] (301.7 mg, 488 mmol) in acetone (45 mL)
to give an orange solution of [Cp*Rh(OCMe
precipitate of AgCl. The reaction mixture was stirred only for
2
) ](PF ) with a white
2 3 6 2
À
À [35]
1
0 min at 208C to avoid the partial solvolysis of PF6 to PO2F2 .
The resulting orange suspension was filtered through a PTFE can-
nula equipped with a glass-fiber filter and the precipitate was
washed three times with acetone (9 mL in total). Combined filtrate
and washings were added to a colorless solution of p-H Q-Me
(324.5 mg, 1.952 mmol) in acetone (15 mL) using the PTFE cannula
and the reaction mixture was vigorously stirred for 2 h at 208C.
The solvent was removed under vacuum and the vivid reddish
yellow residue was dissolved in acetone/methanol (4:1, v/v, 13 mL),
filtered through a PTFE cannula equipped with a glass-fiber filter,
and crystallized by slow vapor diffusion of ether (ca. 75 mL) into
the filtrate under slightly reduced pressure (ca., 350 mmHg) to
give yellow crystals. Yield: 31% (165 mg, 301 mmol). For analysis,
the compound was recrystallized once more from the same solvent
5
h -semiquinone complexes, DFT calculations revealed that in-
5
stead of generating simply protonated/deprotonated h -semi-
2
4
+
quinone complexes, the protonation/deprotonation of [1] led
to hapticity changes of semiquinone ligand accompanied by
reduction/oxidation by the p-bonded rhodium fragment,
6
2+
4
giving stable h -hydroquinone complex [2] and h -benzoqui-
none complex [3], respectively. Therefore, kink defects presum-
6
4
ably consist of these stable h - and h -coordinated complexes,
but the proof is not easy to establish. The generated protonic
solitons (charged kink solitons) migrate rapidly in the H-
bonded chain through successive proton transfer coupled with
polarity inversion, suggesting that the dielectric response
occurs by expanding the short-range ordered domain exhibit-
ing a stable polarity with respect to the applied electric field.
The potential use of semiquinone and/or metal ion spins in
future complex designs may create interesting magnetic or
solid-state properties exploiting cooperative proton and elec-
tron transfers in addition to proton dynamics and ferroelectrici-
ty. Our team is currently searching for such new H-bonded
multifunctional materials.
1
system using the same vapor diffusion techniques. H NMR
(600 MHz, [D ]DMSO, TMS): d=1.68 (s, 15H; Cp*), 1.89 (s, 12H;
semiquinone-CH ), 3.77 ppm (brs; 1H, OH); C NMR (150.92 MHz,
3
6
13
[
9
D ]DMSO, TMS): d=7.81 (Cp*, -CH ), 11.48 (semiquinone, ÀCH ),
6
3
3
9.12 (semiquinone, ÀC=CÀ), 101.42 (Cp*, ÀC=CÀ), 129.39 (semi-
quinone, CÀOH), 158.04 ppm (semiquinone, C=O); elemental analy-
sis calcd (%) for C H F O PRh: C 43.81, H 5.15; found: C 43.65; H
20
28
6
2
5
.12.
Calorimetry
Heat-capacity measurements were performed from 8 to 302 K in
a laboratory-made low-temperature adiabatic calorimeter for small
[
36]
samples. After buoyancy correction, the sample (0.15455 g) was
loaded into a gold-plated copper cell and sealed under helium gas
at ambient pressure using an indium gasket. The helium gas acted
as a heat-exchange medium. Thermometry was performed using
Experimental Section
Syntheses
a
rhodium-iron alloy resistance thermometer (nominal 27 W,
All syntheses were performed at room temperature under an
argon atmosphere by using standard vacuum-line and Schlenk
techniques. All solvents were dried using appropriate drying
Oxford Instruments) calibrated based on the international tempera-
ture scale of 1990 (ITS-90).
[32]
agents and were freshly distilled under argon before use.
Cp*Rh(m-Cl)Cl] was prepared according to a published proce-
Single-crystal X-ray diffraction
[
2
[33]
1
13
dure. General H and C NMR spectra were obtained on a JEOL
ECA-600 spectrometer using tetramethylsilane (TMS) as an internal
standard. Elemental analyses were performed at the Center for Or-
A summary of the crystallographic data and intensity data collec-
tion for all structures of [1]PF is given in Table S1 (the Supporting
6
Information). Crystals were mounted on fine nylon loops, which
Chem. Eur. J. 2015, 21, 9682 – 9696
www.chemeurj.org
9693
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
were attached to the copper mounting pin, using petroleum jelly.
X-ray diffraction measurements were performed using a Rigaku
RAXIS-RAPID imaging plate diffractometer with graphite-mono-
chromated MoKa radiation (l=0.71075 Š). The diffractometer was
equipped with a Rigaku variable-temperature apparatus. Collec-
tion, indexing, peak integration, cell refinement, and scaling of the
diffraction data were performed using the RAPID AUTO software.
Lorentz, polarization, and empirical absorption corrections were
applied to the intensity data. Structures were solved by direct
methods using SHELXS-97 and expanded using Fourier tech-
DFT calculations
DFT calculations were performed to elucidate the electronic struc-
tures of isolated benzoquinone, semiquinone, and hydroquinone
complexes. Geometries were first optimized at the B3LYP level of
theory along with Grimme’s long-range dispersion correction
[42]
(
B3LYP-D2). The initial geometry was obtained using the neutron
crystal structure of the semiquinone complex. During the geometry
optimization, Huzinaga’s MIDI plus polarization function (MIDI+P)
and 6-31G* basis sets were used for Rh and other atoms, respec-
[43] 13
[37]
2
tively.
C NMR chemical shifts were also simulated for an H-
niques. All refinements were made on F by a full-matrix least-
[
37]
bonded dimer of two adjacent semiquinone complexes. The geom-
etry of the dimer model was fixed to the neutron crystal structure.
The NMR simulation was conducted by using the gauge including-
atomic-orbital (GIAO) method at the B3LYP-D2 level of theory with
the larger MIDI+P and 6-311G(2d,p) basis sets for Rh and other
atoms, respectively. Calculated chemical shifts were calibrated
using the TMS reference values at the same theoretical levels in
GaussView 5.0. All DFT calculations were performed using the
squares method using the SHELXL-2014
implemented on the
[38]
free GUI software Yadokari-XG. All non-hydrogen atoms in the LT
and RT phases were refined anisotropically. Methyl hydrogen
atoms were placed in geometrically calculated positions and re-
fined with isotropic temperature factors U (H) fixed at 1.5 times
iso
the Ueq (C) of their parent carbon atoms. The hydroxyl hydrogen
atoms were located using a difference Fourier map and refined as
riding in their as-found position with U (H)=1.5U (O). The Cp*
iso
eq
[25]
À
Gaussian 09 program package.
ring, PF6 anion, and hydroxyl hydrogen atoms in the RT phase
were disordered over two orientations. The disordered Cp* moiety
was modeled using a rigid group. The occupancy factors of the
FTIR spectroscopy
À
Cp* ring and PF6 were refined, whereas those of hydroxyl hydro-
FTIR spectra at RT were recorded on a JASCO FT/IR-4100 spectro-
photometer using KBr pressed disks. Variable-temperature FTIR
spectra were measured on a JASCO FT/IR-6100 spectrophotometer
equipped with an Oxford Instruments Optistat CF cryostat.
gen atoms were set to 0.5. CCDC-1051327 (345K), CCDC-1051328
(
1
286K), CCDC-1051329 (256K), CCDC-1051330 (197K), CCDC-
051331 (149K), and CCDC-1051332 (101K) contain the supple-
mentary crystallographic data for this paper. These data can be ob-
tained free of charge from The Cambridge Crystallographic Data
Centre via www.ccdc.cam.ac.uk/data_request/cif.
Dielectric measurements
Electric measurements were performed on crystal samples using
painted gold or silver paste as the electrodes. These samples were
Single-crystal neutron diffraction
prepared by cutting single crystals of [1]PF perpendicularly to the
6
crystallographic [100] and [010] directions. The real and imaginary
parts of the dielectric constant were measured using an Agilent
Single-crystal neutron diffraction data were collected using the
[39]
single-crystal diffractometer SENJU at the beamline BL18 of the
Materials and Life-Science Facility (MLF), Japan Proton Accelerator
Research Complex (J-PARC), using the wavelength-resolved time-
4
284 A precision LCR meter from 30 Hz to 1 MHz for a measuring
À1
ac field of 3 to 5 Vcm . Electric polarization vs. electric field curves
were recorded using an aixACCT Easy Check 300 ferroelectric
tester equipped with a high voltage amplifier (Matsusada Precision
Inc., HEOP-1.5B20). The sample temperature was controlled be-
tween 300 and 12 K using a closed-cycle helium refrigerator (Iwata-
ni, CryoMini D105).
of-flight Laue diffraction method for wavelength ranging from 0.4
3
to 4.4 Š. A single crystal of [1]PF (2.0”0.7”0.3 mm ) was mounted
6
on an aluminum pin and attached to a closed-cycle helium cryo-
stat. Diffraction data were acquired at 83 and 256 K under vacuum
conditions. Intensities were collected using 37 two-dimensional
scintillation detectors to cover a wide area of reciprocal lattice
space. Each measurement with 16 crystal orientations was con-
ducted over four days under an accelerator power of 300 kW. Data
Solid-state NMR spectroscopy
13
[
40]
C CP/MAS NMR spectra were acquired using a JEOL ECA-300
reduction was performed using the STARGazer
and Jana2006
[41]
spectrometer operating at a carbon frequency of 74.17484 MHz.
The sample (95 mg) was packed into a 4 mm probe and subjected
to a spinning rate of 4 kHz. An optimum contact time of 2 ms was
determined at ambient temperature, and the same value was used
for other measurement temperatures. The total suppression of side
bands sequence was used to suppress spinning side bands.
SPINAL-64 decoupling was employed during carbon signal acquisi-
tion using a minimum proton decoupling power of 31.25 kHz. Line
shape simulations were performed using an in-house FORTRAN
program assuming that chemical exchange occurs between two
sites.
programs to obtain hkl indexes and the corresponding integrat-
ed intensities of reflections corrected for the detector efficiency,
Lorentz factor, and scaling factor for each crystal orientation. A
[37]
least-squares refinement was performed using SHELXL-2014 with
I>3s(I) reflections. Atomic coordinates resulting from X-ray crystal
structure analyses at 101 and 256 K were used as initial structural
models. All hydrogen atoms at 83 K were determined using the nu-
clear difference Fourier map of the neutron data and refined aniso-
À
tropically. Cp* ring and PF6 anion, which were highly disordered
in the RT phase, were defined as rigid bodies and their anisotropic
parameters were fixed once they reached maximum convergence
using a conjugate-gradient least-squares (CGLS) refinement. The H-
bonded hydrogen atom in the RT phase was disordered over posi-
tions H1 and H2 with site-occupancy factors of 0.56(5) and 0.44(5).
CCDC-1051325 (256K) and CCDC-1051326 (83K) contain the sup-
plementary crystallographic data. These data can be obtained free
of charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
Acknowledgements
This work was financially supported by Grant-in-Aid for Scien-
tific Research (C) (24550162) by JSPS, Grant-in-Aid for Scientific
Research on Innovative Areas “New Frontier of Materials Sci-
Chem. Eur. J. 2015, 21, 9682 – 9696
www.chemeurj.org
9694
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
ence Opened by Molecular Degree of Freedom” (23110723) by
MEXT, and Hyogo Science and Technology Association. Single
crystal neutron diffraction measurements were carried out as
a general proposal of J-PARC/MLF (2014A0315).
ji, Angew. Chem. Int. Ed. 2007, 46, 5938–5942; Angew. Chem. 2007, 119,
6042–6046; f) Y. Kobayashi, M. Yoshioka, K. Saigo, D. Hashizume, T.
Ogura, J. Am. Chem. Soc. 2009, 131, 9995–10002; g) Y. Morita, T.
Murata, K. Nakasuji, Bull. Chem. Soc. Jpn. 2013, 86, 183–197; h) S. C.
Lee, A. Ueda, A. Nakao, R. Kumai, H. Nakao, Y. Murakami, H. Mori, Chem.
Eur. J. 2014, 20, 1909–1917.
[
[
13] A. Ueda, S. Yamada, T. Isono, H. Kamo, A. Nakao, R. Kumai, H. Nakao, Y.
Murakami, K. Yamamoto, Y. Nishio, H. Mori, J. Am. Chem. Soc. 2014, 136,
12184–12192.
Keywords: density functional calculations
bonding · NMR spectroscopy · proton transport · quinones ·
rhodium
·
hydrogen
14] a) M. Mitsumi, T. Nishitani, S. Yamasaki, N. Shimada, Y. Komatsu, K. Toriu-
mi, Y. Kitagawa, M. Okumura, Y. Miyazaki, N. Górska, A. Inaba, A. Kanda,
N. Hanasaki, J. Am. Chem. Soc. 2014, 136, 7026–7037; b) B. Zhou, Y. Ido-
bata, A. Kobayashi, H. Cui, R. Kato, R. Takagi, K. Miyagawa, K. Kanoda, H.
Kobayashi, J. Am. Chem. Soc. 2012, 134, 12724–12731; c) N. Hoshino, F.
Iijima, G. N. Newton, N. Yoshida, T. Shiga, H. Nojiri, A. Nakao, R. Kumai, Y.
Murakami, H. Oshio, Nat. Chem. 2012, 4, 921–926; d) H. Miyasaka, N.
Motokawa, T. Chiyo, M. Takemura, M. Yamashita, H. Sagayama, T. H.
Arima, J. Am. Chem. Soc. 2011, 133, 5338–5345; e) P. Jain, V. Ramachan-
dran, R. J. Clark, H. D. Zhou, B. H. Toby, N. S. Dalal, H. W. Kroto, A. K.
Cheetham, J. Am. Chem. Soc. 2009, 131, 13625–13627; f) O. Sato, J. Tao,
Y.-Z. Zhang, Angew. Chem. Int. Ed. 2007, 46, 2152–2187; Angew. Chem.
2007, 119, 2200–2236; g) H. Cui, Z. Wang, K. Takahashi, Y. Okano, H. Ko-
bayashi, A. Kobayashi, J. Am. Chem. Soc. 2006, 128, 15074–15075; h) E.
Coronado, J. R. Galµn-Mascarós, C. J. Gómez-García, V. Laukhin, Nature
2000, 408, 447–449; i) O. Sato, T. Iyoda, A. Fujishima, K. Hashimoto, Sci-
ence 1996, 272, 704–705.
[
1] a) Proton Transfer in Hydrogen-Bonded Systems, Plenum Press, New York,
992; b) G. A. Jeffrey, An Introduction to Hydrogen Bonding, Oxford Uni-
1
versity Press, New York, 1997; c) T. Elsaesser, H. J. Bakker, Ultrafast Hy-
drogen Bonding Dynamics and Proton Transfer Processes in Condensed
Phase, Kluwer Academic Publishers, Dordrecht, The Netherlands, 2002;
d) G. Gilli, P. Gilli, The nature of the hydrogen bond, outline of a compre-
hensive hydrogen bond theory, Oxford University Press, New York, 2009.
2] a) M. H. V. Huynh, T. J. Meyer, Chem. Rev. 2007, 107, 5004–5064; b) S. Y.
Reece, D. G. Nocera, Annu. Rev. Biochem. 2009, 78, 673–699; c) S.
Hammes-Schiffer, A. A. Stuchebrukhov, Chem. Rev. 2010, 110, 6939–
[
6
960; d) D. R. Weinberg, C. J. Gagliardi, J. F. Hull, C. F. Murphy, C. A. Kent,
B. C. Westlake, A. Paul, D. H. Ess, D. G. McCafferty, T. J. Meyer, Chem. Rev.
012, 112, 4016–4093.
2
[
[
[
3] a) S. Scheiner in Proton Transfer in Hydrogen-Bonded Systems (Ed.: T.
Bountis), Premium Press, New York, 1992, pp. 29–47; b) W. C. Hamilton,
J. A. Ibers, Hydrogen Bonding in Solid, W. A. Benjamin, Inc., New York,
[15] a) J. Moussa, C. Guyard-Duhayon, P. Herson, H. Amouri, M. N. Rager, A.
Jutand, Organometallics 2004, 23, 6231–6238; b) N. A. Bailey, H. Adams,
Cryst. Struct. Comm. 1980, 9, 1213–1221.
[16] S. U. Son, S. B. Kim, J. A. Reingold, G. B. Carpenter, D. A. Sweigart, J. Am.
Chem. Soc. 2005, 127, 12238–12239.
[17] M. Oh, G. B. Carpenter, D. A. Sweigart, Organometallics 2002, 21, 1290–
1295.
[18] a) J. L Bras, H. Amouri, J. Vaissermann, Organometallics 1998, 17, 1116–
1121; b) J. Le Bras, H. Amouri, J. Vaissermann, J. Organomet. Chem.
1998, 553, 483–485.
1968; c) A. Novak, Struct. Bonding 1974, 18, 177–216.
4] a) J. M. Lopez del Amo, U. Langer, V. Torres, G. Buntkowsky, H.-M. Vieth,
M. PØrez-Torralba, D. Sanz, R. M. Claramunt, J. Elguero, H.-H. Limbach, J.
Am. Chem. Soc. 2008, 130, 8620–8632; b) Y. Masuda, T. Nakano, M. Su-
giyama, J. Phys. Chem. A 2012, 116, 4485–4494.
5] a) S. Horiuchi, F. Kagawa, K. Hatahara, K. Kobayashi, R. Kumai, Y. Muraka-
mi, Y. Tokura, Nat. Commun. 2012, 3, 1308; b) S. Horiuchi, R. Kumai, Y.
Tokura, Adv. Mater. 2011, 23, 2098–2103; c) S. Horiuchi, Y. Tokunaga, G.
Giovannetti, S. Picozzi, H. Itoh, R. Shimano, R. Kumai, Y. Tokura, Nature
[19] J. A. Reingold, S. U. Son, S. B. Kim, C. A. Dullaghan, M. Oh, P. C. Frake,
G. B. Carpenter, D. A. Sweigart, Dalton Trans. 2006, 2385–2398.
[20] a) T. Sakurai, Acta Crystallogr. 1965, 19, 320–330; b) T. Sakurai, Acta Crys-
tallogr. B 1968, 24, 403–412.
2010, 463, 789–792.
[6] a) J. A. Hurd, R. Vaidhyanathan, V. Thangadurai, C. I. Ratcliffe, I. L. Mou-
drakovski, G. K. H. Shimizu, Nat. Chem. 2009, 1, 705–710; b) S. Buree-
kaew, S. Horike, M. Higuchi, M. Mizuno, T. Kawamura, D. Tanaka, N.
Yanai, S. Kitagawa, Nat. Mater. 2009, 8, 831–836; c) H. O¯ kawa, M. Sada-
[21] W. T. Pennington, A. O. Patil, D. Y. Curtin, I. C. Paul, J. Chem. Soc. Perkin
Trans. 2 1986, 1693–1700.
kiyo, T. Yamada, M. Maesato, M. Ohba, H. Kitagawa, J. Am. Chem. Soc.
[22] G. G. Aleksandrov, Y. T. Struchkov, V. S. Khandarova, S. P. Gubin, J. Orga-
nomet. Chem. 1970, 25, 243–247.
[23] D. Braga, Chem. Rev. 1992, 92, 633–665.
2
013, 135, 2256–2262.
7] a) Y. Moritomo, Y. Tokura, T. Oohashi, T. Kojima, A. Itsubo, J. Chem. Phys.
992, 96, 8507–8513; b) I. Takasu, T. Sugawara, T. Mochida, J. Phys.
[
1
[24] H. Nakai, T. Nonaka, Y. Miyano, M. Mizuno, Y. Ozawa, K. Toriumi, N. Koga,
T. Nishioka, M. Irie, K. Isobe, J. Am. Chem. Soc. 2008, 130, 17836–17845.
[25] Gaussian 09, Revision C.01, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E.
Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Men-
nucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian,
A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara,
K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O.
Kitao, H. Nakai, T. Vreven, J. Montgomery, J. A., J. E. Peralta, F. Ogliaro,
M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Ko-
bayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyen-
gar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B.
Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann,
O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin,
K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg,
S. Dapprich, A. D. Daniels, Ö. Farkas, J. B. Foresman, J. V. Ortiz, J. Cio-
slowski, D. J. Fox, Gaussian, Inc., Wallingford, CT, Wallingford CT, 2009.
[26] a) G. G. Aleksandrov, Y. T. Struchkov, J. Struct. Chem. 1970, 11, 653–659;
b) G. G. Aleksandrov, Y. T. Struchkov, J. Struct. Chem. 1971, 12, 953–958;
c) G. G. Aleksandrov, Y. T. Struchkov, J. Struct. Chem. 1971, 12, 99–104.
[27] a) R. E. Marsh, D. A. Clemente, Inorg. Chim. Acta 2007, 360, 4017–4024;
b) S. Sun, G. B. Carpenter, D. A. Sweigart, J. Organomet. Chem. 1996,
512, 257–259.
Chem. B 2004, 108, 18495–18499.
8] a) D. Semmingsen, F. J. Hollander, T. F. Koetzle, J. Chem. Phys. 1977, 66,
[
4
405–4412; b) F. J. Hollander, D. Semmingsen, T. F. Koetzle, J. Chem.
Phys. 1977, 67, 4825–4831; c) D. Semmingsen, Acta Chem. Scand. 1973,
7, 3961–3972; d) D. Semmingsen, Z. Tun, R. J. Nelmes, R. K. McMullan,
2
T. F. Koetzle, Z. Kristallogr. 1995, 210, 934–947.
[
9] a) J. Feder, Ferroelectrics 1976, 12, 71–83; b) G. A. Samara, D. Semming-
sen, J. Chem. Phys. 1979, 71, 1401–1407; c) Y. Moritomo, Y. Tokura, H.
Takahashi, N. M o¯ ri, Phys. Rev. Lett. 1991, 67, 2041–2044; d) Y. Moritomo,
T. Katsufuji, Y. Tokura, J. Chem. Phys. 1991, 95, 2244–2251.
10] H. Terao, T. Sugawara, Y. Kita, N. Sato, E. Kaho, S. Takeda, J. Am. Chem.
Soc. 2001, 123, 10468–10474.
11] a) T. Mochida, A. Izuoka, T. Sugawara, Y. Moritomo, Y. Tokura, J. Chem.
Phys. 1994, 101, 7971–7974; b) Y. Moritomo, Y. Tokura, T. Mochida, A.
Izuoka, T. Sugawara, J. Phys. Soc. Jpn. 1995, 64, 1892–1895; c) T.
Matsuo, K. Kohno, M. Ohama, T. Mochida, A. Izuoka, T. Sugawara, Euro-
phys. Lett. 1999, 47, 36–41.
12] a) T. Mitani, G. Saito, H. Urayama, Phys. Rev. Lett. 1988, 60, 2299–2302;
b) K. Nakasuji, K. Sugiura, T. Kitagawa, J. Toyoda, H. Okamoto, K. Okani-
wa, T. Mitani, H. Yamamoto, I. Murata, A. Kawamoto, J. Tanaka, J. Am.
Chem. Soc. 1991, 113, 1862–1864; c) M. Felderhoff, I. Steller, A. Reyes-
Arellano, R. Boese, R. Sustmann, Adv. Mater. 1996, 8, 402–405; d) T.
Kubo, M. Ohashi, K. Miyazaki, A. Ichimura, K. Nakasuji, Inorg. Chem.
[
[
[
[28] S. Horiuchi, R. Kumai, Y. Tokunaga, Y. Tokura, J. Am. Chem. Soc. 2008,
130, 13382–13391.
2
004, 43, 7301–7307; e) M. Tadokoro, T. Inoue, S. Tamaki, K. Fujii, K.
[29] a) A. N. Klymachyov, N. S. Dalal, Z. Phys. B 1997, 104, 651–656; b) R. Fu,
A. N. Klymachyov, G. Bodenhausen, N. S. Dalal, J. Phys. Chem. B 1998,
Isogai, H. Nakazawa, S. Takeda, K. Isobe, N. Koga, A. Ichimura, K. Nakasu-
Chem. Eur. J. 2015, 21, 9682 – 9696
www.chemeurj.org
9695
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
1
02, 8732–8735; c) N. Dalal, A. Klymachyov, A. Bussmann-Holder, Phys.
[38] a) K. Wakita, Yadokari-XG, Software for crystal structure analyses, 2001;
b) C. Kabuto, S. Akine, T. Nemoto, E. Kwon, J. Crystallogr. Soc. Jpn. 2009,
51, 218–224.
Rev. Lett. 1998, 81, 5924–5927.
30] a) V. Y. Antonchenko, A. S. Davydov, A. V. Zolotariuk, Phys. Status Solidi B
[
[
1
983, 115, 631–640; b) A. S. Davydov, Solitons in Molecular Systems, D.
[39] T. Kawasaki, K. Kaneko, N. Aso, A. Nakamura, M. Hedo, T. Nakama, Y.
Onuki, T. Ohhara, R. Kiyanagi, K. Oikawa, I. Tamura, A. Nakao, K. Munaka-
ta, T. Hanashima, JPS Conf. Proc. 2014, 1, 014009.
Reidel Publishing, Dordrecht, The Netherlands, 1985.
31] a) M. Szafra n´ ski, J. Phys. Chem. B 2009, 113, 9479–9488; b) M. Szafra n´ ski,
A. Katrusiak, J. Phys. Chem. B 2008, 112, 6779–6785; c) M. Szafra n´ ski, A.
Katrusiak, J. Phys. Chem. B 2004, 108, 15709–15713.
[
40] T. Ohhara, K. Kusaka, T. Hosoya, K. Kurihara, K. Tomoyori, N. Niimura, I.
Tanaka, J. Suzuki, T. Nakatani, T. Otomo, S. Matsuoka, K. Tomita, Y. Nishi-
maki, T. Ajima, S. Ryufuku, Nucl. Instrum. Methods Phys. Res. A 2009, 600,
[
[
32] D. D. Perrin, W. L. F. Armarego, Purification of laboratory chemicals, Per-
gamon Press, New York, 1988.
33] C. White, A. Yates, P. M. Maitlis, D. M. Heinekey, Inorg. Synth. 1992, 29,
195–197.
[
[
41] V. Pet rˇ i cˇ ek, M. Du sˇ ek, L. Palatinus, Z. Kristallogr. 2014, 229, 345–352.
42] a) A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652; b) S. Grimme, J.
Comput. Chem. 2006, 27, 1787–1799.
228–234.
[
[
34] A. D. Thomas, L. L. Miller, J. Org. Chem. 1986, 51, 4160–4169.
35] C. White, S. J. Thompson, P. M. Maitlis, J. Organomet. Chem. 1977, 134,
[
43] S. Huzinaga, Physical science data 16, Gaussian Basis Sets for Molecular
Calculations, Elsevier, 1983.
3
19–325.
36] Y. Kume, Y. Miyazaki, T. Matsuo, H. Suga, J. Phys. Chem. Solids 1992, 53,
297–1304.
37] G. M. Sheldrick, Acta Crystallogr. A 2008, 64, 112–122.
[
[
1
Received: February 27, 2015
Published online on June 1, 2015
Chem. Eur. J. 2015, 21, 9682 – 9696
www.chemeurj.org
9696
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim