Giant photoconductivity in NMQ[Ni(dmit)2]

Article abstract of DOI:10.1002/ejic.201402035

The simple molecular salt NMQ[Ni(dmit)₂] (NMQ = N-methylquinolinium, dmit = 1,3-dithiol-2-thione-4,5-dithiolate) functions as a diamagnetic insulator with an activation energy E_a(dark) of 0.20 eV. However, at 300 K, it exhibits ca. 40 times higher conductivity (σ_UV) under UV irradiation [(375±5) nm, 15.7 mWcm ^-2] than it does under dark conditions (σ_dark). The ratio σ_UV/σ_dark rapidly increases with decreasing temperature and reaches ca. 880 at 200 K. From the temperature dependence of σ_UV, the activation energy under irradiation E_a(UV) is 0.12 eV. These observations cannot be explained as the result of sample heating during UV irradiation. Rather, the X-ray photoelectron spectra of the sulfur and nickel atoms, the calculated band structure, and the UV/Vis spectra of the salt can all be explained consistently as follows: charge transfer between the Ni(dmit)₂ moieties upon exposure to 375 nm UV light induces melting of the charge-ordered state and produces the unusually large photoconductivity of NMQ[Ni(dmit)₂].

Full text of DOI:10.1002/ejic.201402035

FULL PAPER
DOI:10.1002/ejic.201402035
Giant Photoconductivity in NMQ[Ni(dmit)2]
CLUSTER
ISSUE
Toshio Naito,*[a] Tomoaki Karasudani,[a] Naoki Nagayama,[a]
Keishi Ohara,[a] Kensuke Konishi,[a] Shigeki Mori,[b]
Takahiro Takano,[c] Yukihiro Takahashi,[c,d] Tamotsu Inabe,[c,d]
Shota Kinose,[e] Sadafumi Nishihara,[e] and Katsuya Inoue[e]
Keywords: Photoconductors / Molecular crystals / Charge transfer / S ligands / Simple salts / Semiconductors
The simple molecular salt NMQ[Ni(dmit)2] (NMQ = N-meth-
ylquinolinium, dmit 1,3-dithiol-2-thione-4,5-dithiolate)
functions as a diamagnetic insulator with an activation en-
under irradiation Ea(UV) is 0.12 eV. These observations can-
not be explained as the result of sample heating during UV
irradiation. Rather, the X-ray photoelectron spectra of the
sulfur and nickel atoms, the calculated band structure, and
the UV/Vis spectra of the salt can all be explained consist-
ently as follows: charge transfer between the Ni(dmit)2 moie-
ties upon exposure to 375 nm UV light induces melting of
the charge-ordered state and produces the unusually large
photoconductivity of NMQ[Ni(dmit)2].
=
ergy Ea(dark) of 0.20 eV. However, at 300 K, it exhibits ca. 40
times higher conductivity (σUV
) under UV irradiation
[(375Ϯ5) nm, 15.7 mWcm–2] than it does under dark condi-
tions (σdark). The ratio σUV/σdark rapidly increases with
decreasing temperature and reaches ca. 880 at 200 K. From
the temperature dependence of σUV, the activation energy
(dmit)2]2– salts, which is why they are readily oxidized to-
ward the neutral state. As the radical anions of M(dmit)2
(M = Ni, Pd, Pt, Au, etc.) do not have any peripheral
hydrogen atoms or substituents, they have fully delocalized
π-conjugated systems. This is a unique feature and is advan-
tageous with regard to the formation of intermolecular in-
teractions. In a metal complex molecule, the amplitude of
the energy gap between the highest occupied molecular or-
bital (HOMO) and the lowest unoccupied molecular orbital
(LUMO)[5] depends on the strength of the crystal field.
Introduction
Among the large number of molecular conducting solids
reported to date,[1] Ni(dmit)2 and its derivative dithiolene
complex salts[2,3] are unique. Although Ni(dmit)2 is readily
oxidized, it tends to form stable anions rather than cations.
The resultant [Ni(dmit)2]n– anions exhibit various charges
(0 Յ n Յ 2) in the solid state depending on the countercat-
ions and the crystal structure of the material. The salts of
[Ni(dmit)2]2– and [Ni(dmit)2]– with onium ions are particu-
larly stable and can be recrystallized under ambient atmo-
sphere to obtain well-developed single crystals.[4] These
unique chemical properties originate from the electronic
structure of the Ni(dmit)2 molecule (Scheme 1), in which
the frontier orbitals consist almost exclusively of the π-con-
jugated systems of the dmit ligands, which are primarily the
electron-donating sulfur orbitals. The two highest-energy
electrons occupy an antibonding orbital in the [Ni-
[a] Graduate School of Science and Engineering, Ehime
University,
Bunkyo-cho, Matsuyama, Ehime 790-8577, Japan
E-mail: tnaito@ehime-u.ac.jp
http://chem.sci.ehime-u.ac.jp/~physchem01/
[b] Integrated Center for Sciences, Ehime University,
Bunkyo-cho, Matsuyama, Ehime 790-8577, Japan
[c] Graduate School of Science,
Kita-ku, Sapporo, Hokkaido 060-0810, Japan
[d] Graduate School of Chemical Sciences and Engineering,
Kita-ku, Sapporo, Hokkaido 060-0810, Japan
[e] Department of Chemistry, Graduate School of Science,
Hiroshima University,
Higashi-Hiroshima 739-8526, Japan
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.201402035.
Scheme 1. Molecular structures of Ni(dmit)2, MV, BPY, and NMQ.
Eur. J. Inorg. Chem. 2014, 4000–4009
4000
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Therefore, the gap in such compounds is far smaller than cal doping or the application of pressure, this technique is
those of similar π-conjugated compounds, as energy separa- facile, instantaneous, and reversible. Thus, it may be pos-
tions produced by crystal fields are generally much smaller sible to discover new electronic states or physical behaviors
than those of covalent bonds. Owing to the dimerized crys- simply by irradiating an insulating, nonmagnetic simple
tal structure, the relative energy levels of the antibonding molecular salt. On the basis of this idea, the title compound
HOMO (Φ–HOMO) and bonding LUMO (Φ+LUMO) of the has been synthesized, and its physical properties have been
dmit complex molecules can sometimes reverse order subse- examined under both dark and irradiated conditions. Here,
quent to the application of pressure or replacement of the we report the results and discuss our findings in terms of
metal center.[6,7] As a result, the Fermi energy moves from the “giant photoconductivity” mechanism that occurs in
the Φ+
to the Φ–HOMO. As the Φ–
and Φ+
this salt, which is different from optical doping.
LUMO
HOMO
LUMO
differ in symmetry, which significantly affects the effective
molecular orbital overlap between adjacent Ni(dmit)2 mo-
lecules, it is important for the design of M(dmit)2-based
conductors to control or predict the molecular orbitals
(MOs) that will accommodate unpaired electrons. In ad-
dition to control over the HOMO–LUMO energy level in-
version, gradual band mixing between the HOMO and
LUMO and, thus, band filling has been realized by the ap-
plication of progressively increasing pressure.[7] However, it
has been determined that control of band filling by doping
of the molecular crystals with various chemical species is
not overly versatile, and only a limited number of pioneer-
ing works have reported successful cases.[8,9]
In keeping with the typical behavior of multi-sulfur-con-
taining, electron-donor, π-conjugated molecules, [Ni-
(dmit)]n– species (0 Ͻ n Ͻ 2) usually aggregate and interact
with one another in the solid state to form an inter-sulfur-
atom short-contact network. Owing to the strong antiferro-
magnetic interactions between [Ni(dmit)]n– molecules, the
charge-transfer (CT) salts of [Ni(dmit)2]n– often function as
narrow-band diamagnetic or paramagnetic insulators and
can be regarded as band or Mott insulators, respectively.
The doping of Mott insulators with small numbers of carri-
ers is interesting, because such modification is often known
to produce high-TC superconductors, as well as to confer
other unusual physical properties on the insulator, as has
been demonstrated with additives such as cuprates[10] and
fullerides.[11] Our work has recently revealed that one can
realize metallic photoconductors with localized spins by se-
lecting appropriate countercations (Cat) for Ni(dmit)2 salts
to form Cat[Ni(dmit)2]n.[12,13] As an example, the conduc-
tivity of the methyl viologen (MV) salt MV[Ni(dmit)2]2 is
enhanced by three orders of magnitude under UV irradia-
tion, and this compound also exhibits metallic behavior
down to low temperatures. Various analyses have demon-
strated that these marked changes in the electrical and mag-
netic behavior under UV irradiation are closely related to
charge (electron) transfer between MV and Ni(dmit)2. This
conclusion is made on the basis of assessments of diffuse
Results and Discussion
Crystal Structure
The title compound was originally prepared by Cor-
nelissen et al.,[14] and the crystal structure of our product
agrees with that given in this previous report. The unit cell
of NMQ[Ni(dmit)2] is shown in Figure 1. The asymmetric
unit contains two cationic (NMQ+) and two anionic ([Ni-
(dmit)2]–) species as crystallographically independent con-
stituents. Therefore, the crystal may be better described as
the 2:2 salt NMQ2[Ni(dmit)2]2.[15] All of the crystallograph-
ically independent NMQ and Ni(dmit)2 species are nearly
completely planar. The adjacent Ni(dmit)2 species are al-
most parallel to each other (dihedral angle 1.57°), and their
interplanar distance is 3.60–3.68 Å. There are many inter-
molecular sulfur–sulfur contacts (Ͻ3.60 Å) between the
neighboring Ni(dmit)2 species within the same bc plane but
not between those located in different bc planes. This mo-
lecular arrangement can be explained in several different
ways: (1) the cations and anions form a mixed stacking
structure along the b axis to generate a typical one-dimen-
sional molecular crystal, (2) the structure consists of an
ionic crystal formed from alternating cations and anions
in three dimensions, or (3) the [Ni(dmit)2]– species form a
honeycomb structure in the bc plane that accommodates
NMQ+ species in each cell. Considering the calculated
tight-binding band structure, intermolecular overlap inte-
grals, and the observed physical properties (see below), the
last interpretation (3) of the crystal structure appears to be
the most appropriate.
Electronic Structure and Spectra
Figure 2a and b shows the tight-binding band structure.
reflectance spectra (in powder form), ESR spectra (under A series of bands over a wider energy range are shown in
dark and UV-irradiated conditions), calculated band struc- the Supporting Information (Figure S2). The contributions
tures, and the selective response of conducting and mag- of the Ni, dmit, and NMQ fragments of selected bands are
netic properties only to UV light. By utilizing an appropri- also tabulated in Table S2. The calculations indicate that
ate donor–acceptor pair, it may be possible to stabilize the this material should function as a band insulator. The cal-
resultant CT excited state, which otherwise could not be culated band gap [Eg(calc) ≈ 0.4 eV] quantitatively agrees
accomplished by either chemical synthesis or thermo- with the observed band gap in the resistivity measurement
dynamic means. Accordingly, our strategy corresponds to without irradiation [Eg(obs) = 2Ea(dark) = (0.40Ϯ0.02) eV;
“optical doping” of Mott (band) insulators. Unlike chemi- Ea(dark) is the activation energy in the dark]. All of the
Eur. J. Inorg. Chem. 2014, 4000–4009
4001
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Figure 2. Electronic band structure of NMQ[Ni(dmit)2]: (a) empty
bands located at ca. 3.3 eV (ca. 375 nm) above the Fermi level (EF)
and (b) energy bands around EF (EF = –11.01 eV). The relative
contributions from NMQ and Ni(dmit)2 in each band are indicated
by cyan (NMQ) and magenta [Ni(dmit)2] mixed at the actual band
mixing ratios. Γ, X, M, Y, L, Z, and R designate the following
points in reciprocal space: (0,0,0), (0.5,0,0), (0.5,0.5,0), (0,0.5,0),
(0,0.5,0.5), (0,0,0.5), and (0.5,0.5,0.5), respectively.
Figure 1. Crystal structure of NMQ[Ni(dmit)2]: (a) molecular ar-
rangement in the bc plane, (b)–(e) enlarged views indicating inter-
molecular interactions S1–S38 (for details, see Table 1). The yellow,
brown, and grey atoms are sulfur, carbon (filled)/hydrogen (open),
and nitrogen (smaller)/nickel (larger), respectively. In (c) and (e),
the faint molecules are in the lower sheet. In (d) and (e), the
hydrogen atoms are omitted for clarity.
Figure 3. Diffuse reflectance spectra of powdered NMQ[Ni(dmit)2]
(20 °C, black) and absorption spectra of TBA[Ni(dmit)2]
(7.2ϫ10–6 moldm–3 in CH3CN, 20 °C, red) and NMQ·I
(3.56ϫ10–5 moldm–3 in CH3CN, 20 °C, blue): (a) entire spectra
(190–2500 nm) and (b) enlarged view (190–1000 nm).
Eur. J. Inorg. Chem. 2014, 4000–4009
4002
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Table 1. Overlap and transfer integrals in NMQ[Ni(dmit)2].
Interacting pair
Molecular orbitals involved[a]
Overlap integrals S
Transfer integrals H [eV]
S1
S2
S3
S4
S5
S6
S7
S8
D:L–D:L
D:L–D:L
D:L–D:L
D:L–D:L
D:L–D:L
D:L–D:L
D:L–D:L
D:L–D:L
0.0058
–0.0058
–0.0005
0.0007
0.0005
0.0005
0
–0.121
0.121
0.0113
–0.0167
–0.0113
–0.0106
–0.0009
0
0
S9
D:L–D:L
D:L–D:L
D:L–D:L
D:L–D:L
D:L–D:L
D:L–D:L
D:L– A:L
D:L+1–A:H
D:L–A:L
D:L+1–A:H
D:L–A:L
D:L+1–A:H
D:L–A:L
D:L+1–A:H
D:L–A:L
D:L+1–A:H
D:L–A:L
D:L+1–A:H
D:L–A:L
D:L+1–A:H
D:L–A:L
D:L+1–A:H
D:L–A:L
D:L+1–A:H
D:L–A:L
D:L+1–A:H
D:L–A:L
D:L+1–A:H
D:L–A:L
D:L+1–A:H
D:L–A:L
D:L+1–A:H
D:L–A:L
D:L+1–A:H
D:L–A:L
D:L+1–A:H
D:L–A:L
D:L+1–A:H
D:L–A:L
D:L+1–A:H
D:L–A:L
D:L+1–A:H
D:L–A:L
D:L+1–A:H
D:L–A:L
D:L+1–A:H
D:L–A:L
D:L+1–A:H
D:L–A:L
D:L+1–A:H
D:L–A:L
0.0001
0
0
0
0
–0.0027
0
S10
S11
S12
S13
S14
S15
0
0
0
0
0
0.0028
–0.005
–0.0067
0.001
0.0025
–0.0025
0.003
0.0008
ca. 0
–0.0002
0.0002
0.0001
0.0001
0
–0.0598
0.1192
0.1448
–0.0182
–0.0496
0.0415
–0.0688
–0.0129
0.0007
0.0051
–0.004
–0.0012
–0.0025
0.0008
0.0015
0.0044
–0.0019
0.0002
–0.0095
0.0055
0.0054
–0.0056
0.0046
0.0098
–0.0022
0.0099
–0.0098
–0.0056
–0.0053
–0.0188
–0.0001
0.0115
0.0013
0.0062
–0.0005
0.0008
0.0013
0.0062
0.0005
–0.0008
0.0053
0.0188
–0.0001
–0.0115
–0.028
–0.0013
0.0073
0.0003
S16
S17
S18
S19
S20
S21
S22
S23
S24
S25
S26
S27
S28
S29
S30
S31
S32
S33
S34
S35
S36
S37
S38
–0.0001
–0.0002
0.0001
ca. 0
0.0004
–0.0003
–0.0002
0.0003
–0.0002
–0.0004
0.0001
–0.0005
0.0004
0.0002
0.0002
0.0008
ca. 0
–0.0005
ca. 0
–0.0003
ca. 0
ca. 0
ca. 0
–0.0003
ca. 0
ca. 0
–0.0002
–0.0008
ca. 0
0.0005
0.0012
0.0001
–0.0003
ca. 0
D:L+1–A:H
D:L–A:L
D:L+1–A:H
[a] A = NMQ, D = Ni(dmit)2, L = LUMO, H = HOMO. L+1 designates the next higher energy level/band of LUMO. Note that the
“HOMO” is empty in NMQ+ and the “LUMO” is singly occupied in [Ni(dmit)2]– in the molecular orbital designation rule here.
Eur. J. Inorg. Chem. 2014, 4000–4009
4003
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
bands are poorly dispersed, and many of the bands are mechanism that is unique to this salt and is discussed in
nearly degenerate, both of which suggest that there are only detail below. As these three processes have different tem-
weak intermolecular interactions in the solid state.
perature dependences, the overall behavior of this com-
The interpretation of the electronic structure thus far is pound is different from both its original semiconducting
qualitatively consistent with the powder UV/Vis diffuse re- behavior and its nearly metallic (the contribution of dI3 be-
flectance spectra (Figure 3). In the powder spectra, a series low) properties. The activation energy under irradiation
of intense bands form an envelope that covers the entire [Ea(UV) = 0.12 eV] is different from that in the dark
UV/Vis region. As a broad band such as this is not ob- [Ea(dark) = 0.20 eV; Figure 4c]. In NMQ[Ni(dmit)2], the
served in the absorption spectra of either TBA[Ni(dmit)2] ratio between the dark conductivity (σdark) and the photocon-
(TBA = tetrabutylammonium) or NMQ·I in solution, it is ductivity under UV irradiation (σUV) in vacuo is unusually
assigned to CT bands between NMQ and Ni(dmit)2 or be- large: σUV/σdark is ca. 40 at 300 K and ca. 880 at 200 K.
tween Ni(dmit)2 species. This assignment is semiquantita- The latter value is comparable to that of MV[Ni(dmit)2]2
tively consistent with the calculated band structures (Fig- [ca. 90 at 300 K (13.6 mWcm–2), ca. 1000 at 200 K in vacuo
ures 2 and S2).
(15.7 mWcm–2)]. NMQ[Ni(dmit)2] also exhibits two unique
To quantitatively examine the intermolecular interac- behaviors with regard to its photoconductivity: (A) sharp
tions, it is helpful to consider the calculated overlap and wavelength selectivity and (B) nonlinear dependence on the
transfer integrals summarized in Table 1. The calculations light intensity (Figure 4d). These behaviors are not ob-
indicate that the adjacent Ni(dmit)2 anions interact with served in existing photoconductors and, thus, indicate that
one another through their LUMOs;[5] this is consistent with the photoconduction mechanism in NMQ[Ni(dmit)2] has
the experimental results as the LUMO has unpaired elec- some features that differ from those of the currently known
trons in the [Ni(dmit)2]n– anion (0 Ͻ n Ͻ 2). The interdimer photoconductors.
interactions within a sheet (S3–S6) are approximately one
To elucidate these differences, the results in Figure 4d
tenth of the intradimer interactions (S1, S2), and all the re- (except for the data at 300 K in air) were analyzed by apply-
maining Ni(dmit)2–Ni(dmit)2 interactions are negligible.
ing Equation (1).
σsq = a + bI + cI2 + dI3
(1)
a, b, c, and d are fitting parameters, and I is the light inten-
sity. In standard photoconductors, the relationship between
a, c, and d may be summarized by Equation (2).
Electrical and Magnetic Properties
a ≈ c ≈ d ≈ 0
(2)
The electrical behavior under different conditions as ob-
tained from various measurements is shown in Figure 4,
The predicated results from Equation (1) are in good
namely, the temperature dependence of the resistivity in the agreement with the observed photoresponse behavior
dark (Figure 4a), the photoresponse to a UV laser (375 nm; shown in (Figure 4d), and the experimentally obtained val-
Figure 4b), a comparison of activation energies under dark ues for parameters a to d are tabulated in the Supporting
and UV-irradiated conditions (Figure 4c), the light intensity Information (Table S3). The temperature dependences of a
dependence of the photocurrent at different temperatures to d (Figure 4e) demonstrate the unique nature of the pho-
(Figure 4d), the temperature dependencies of the coeffi- toconductivity exhibited by this salt. The first coefficient,
cients in Equation (1) (Figure 4e), and the contribution to a, is independent of I and, thus, corresponds to the contri-
photoconduction of each term in Equation (1). In the ab- bution of the dark conductivity, σdark. The experimentally
sence of UV irradiation, NMQ[Ni(dmit)2] is highly insulat- derived temperature dependence of a gives an activation en-
ing (surface resistivity Rsq Ͼ 108 Ω) and exhibits thermally ergy of 0.22 eV, which agrees with the value obtained from
activated behavior (Figure 4a), as expected from its crystal the dark resistivity measurements [Ea(dark) = 0.20 eV]. The
and electronic band structures. The surface resistivity at second coefficient, b, is associated with the contribution of
room temperature in the dark Rsq(room temp.) is standard photoconductivity, because this conductivity is
7.9ϫ108 Ω . NMQ[Ni(dmit)2] exhibits hardly any pho- linearly dependent on I. The third coefficient, c, is negligible
toconduction in response to irradiation at wavelengths at all measurement temperatures applied in this study. The
other than 375 nm (Figure S3 in the Supporting Infor- fourth coefficient, d, represents the most unique feature of
mation). This is readily explained by the narrow band the photoconductivity of this salt. Its contribution to pho-
widths (Figures 2 and S2), which narrow the line widths of toconductivity σsq undergoes only a slight decrease with
the CT transitions; as a result, there is a sharp wavelength decreasing temperature (Figure 4f), whereas all the remain-
dependence. A rapid, clear, stable, and reproducible re- ing contributions rapidly drop to zero. Therefore, it is this
sponse was observed (Figure 4b). Unlike MV[Ni(dmit)2] contribution (dI3) that makes σUV/σdark unusually large
[12]
,
NMQ[Ni(dmit)2] exhibits semiconducting behavior and, thus, is responsible for the rapid increase to increas-
2
(Figure 4c) under UV irradiation. This occurs, because the ingly larger values with decreasing temperature. For this
observed behavior results from three different phenomena: reason, the mechanism of photoconduction of this material
(1) photoconduction based on the standard mechanism, is different from that of any existing substance, including
(2) dark conductivity, and (3) photoconductivity due to a Cat[Ni(dmit)2]2 (Cat = MV,[12] BPY).[13]
Eur. J. Inorg. Chem. 2014, 4000–4009
4004
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Figure 4. Electrical behavior of NMQ[Ni(dmit)2]: (a) temperature-dependent surface resistivity in the dark [inset: Arrhenius plot of the
same data; Ea(dark) = 0.20 eV]; (b) photoresponse (300 K, 375 nm, 11.6 mWcm–2, applied voltage = 2 V, direct current, two-probe
method); (c) comparison of activation energies under dark [black; Ea(dark) = 0.20 eV] and UV-irradiated [red; Ea(UV) = 0.12 eV] condi-
tions; (d) surface conductivity vs. light intensity (375 nm) at various temperatures: 300 K in air (black squares), 300 K (red circles),
280 K (blue triangles), 260 K (red triangles), 240 K (green diamonds), 220 K (black triangles), 200 K (purple triangles) under vacuum;
(e) temperature dependences of coefficients a to d in Equation (1): a (black circles), b (red squares), c (purple triangles), d (green squares);
(f) contribution to photoconduction of each term in Equation (1) in the case of I = 15.71 mWcm–2: a (black circles), bI (red squares),
dI3 (green squares). In (d), the solid curves are the best fits generated by using the parameters given in Table S3, whereas the broken lines
and curves in (e) and (f) are simply included as visual aids. Equation (1) was used in the curve-fitting analysis of the results in (d), except
for the data obtained at 300 K in air. In (d), measurements were carried out at 300 K under vacuum (red circles) and under ambient
atmosphere (i.e. in air) (black squares), while all other measurements were performed under vacuum. In (f), the contributions of a and b
are zero at T Ͻ 220 K and T Ͻ 240 K, respectively. Note that the vertical dotted line at 240 K is not meant to signify a phase transition.
NMQ[Ni(dmit)2] exhibits magnetic susceptibility (χdark
=
net sample mass irradiated was unknown, the suscep-
–5.65ϫ 10–3 emumol–1) that is both negative and tempera- tibility is shown in arbitrary units (A.U.). During irradia-
ture-independent, such that it is diamagnetic at all measure- tion, the susceptibility is clearly enhanced compared with
ment temperatures (from 2 to 315 K), with the exception of the value measured in the dark, and nearly identical results
a sudden increase at T Յ 10 K (Figure 5a).[16] The observed were obtained when different levels of UV intensity were
diamagnetism is consistent with its electrical behavior. In applied (Figure 5b). These results indicate that the irradia-
addition, the observed monotonic behavior also indicates tion effect is saturated above 0.30 mWcm–2 as the net irra-
that there is no phase transition within the temperature diated mass remained essentially constant above
range 2–315 K, which is in good agreement with the in- 0.30 mWcm–2. The temperature dependence of χUV is
terpretation that the crystal structure is essentially un- nearly identical to that of χЈdark, in that both are almost
changed between 296 and 93 K.
temperature-independent. Yet, UV irradiation results in an
The magnetic susceptibility of the compound under dif- evident enhancement of magnetic susceptibility (χUV Ͼ
ferent intensities of UV laser light (χUV) and the back- χЈdark). Thus, the majority of the excited electrons are
ground susceptibility in the dark (χЈdark) are shown in Fig- somewhat delocalized and behave as carriers rather than
ure 5b. The differences in the magnetic susceptibilities Δχ localized Curie spins. This is consistent with the observed
(Δχ = χUV – χЈdark) between the dark and the UV-irradi- high photoconductivity under the same irradiation condi-
ated conditions are summarized in Figure 5c and d. As the tions.
Eur. J. Inorg. Chem. 2014, 4000–4009
4005
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
corresponds to CT between Ni(dmit)2 species. When there
is only limited CT between the Ni(dmit)2 moieties, the dif-
ference in the XPS spectra of each Ni(dmit)2 moiety will
also be minimal. Furthermore, when two Ni(dmit)2 species
form part of the conduction pathways, more rapid electron
exchange between the two species would be expected than
otherwise. This situation obscures the differences in elec-
tronic states and averages the XPS spectra of each Ni-
(dmit)2 species. Thus, considering its energy resolution,
XPS could potentially generate almost the same spectra be-
tween dark and UV-irradiated conditions unless the
amount of CT is sufficiently large. The reasons why such a
small amount of CT can occur in tandem with a high
Figure 5. Magnetic susceptibility vs. temperature as determined
from measurements under (a) dark (χdark and χЈdark) and (b) UV-
irradiated (χUV) conditions (raw data; blue = 17 mWcm–2, red =
0.30 mWcm–2) together with χЈdark (black) for comparison; and dif-
ferences in susceptibility between dark and UV-irradiated condi-
tions (Δχ = χUV – χЈdark) at (c) 0.30 and (d) 17 mWcm–2. The red
closed circles and black crosses in (a) designate the results of mea-
surements with polycrystalline samples of ca. 1 mg in a gelatin cap-
sule and ϽϽ1 mg on Kapton tape, respectively. Note that χЈdark
includes contributions from the Kapton tape and other background
signals, whereas χdark (emumol–1) should contain lattice defects but
no background contributions.
X-ray Photoelectron Spectroscopy
The X-ray photoelectron spectroscopy (XPS) patterns
are presented in Figure 6a–c. As the escape depth of the
photoelectrons is much less than the penetration depth of
the UV light, these spectra enable us to selectively observe
the electronic states of the portion of the sample that is
photoexcited by UV irradiation. During UV irradiation, all
the XPS patterns exhibit no significant changes. These re-
sults can be interpreted as evidence that the optical transi-
Figure 6. X-ray photoelectron spectra of NMQ[Ni(dmit)2] under
dark (black) and irradiation with 375 nm UV laser (red). (a) S 2s,
tion that occurs in response to UV irradiation at 375 nm (b) S 2p1/2 and S 2p3/2, (c) Ni 2 p3/2, and (d) valence band spectra.
Eur. J. Inorg. Chem. 2014, 4000–4009
4006
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
(photo)conductivity ratio, σUV/σdark, will be discussed fur-
ther on.
As the next step in this field of research, it would be
interesting to try to control the doping level, although it is
To provide further information, the valence band spec- not clear if it is possible to do so by modifying irradiation
trum was also measured and is shown in Figure 6d. This conditions such as duration or intensity. Preliminary results
spectrum clearly shows that the density of states at the suggest that the temperature and duration of irradiation
Fermi level (ca. 0 eV–1) remains unchanged under UV irra- can alter the resultant electrical resistivity in some Ni-
diation. This is consistent with the observation that this ma- (dmit)2 simple salts and will be reported in due course. The
terial remains a semiconductor when it is subjected to UV doping level should be closely related to the selection of the
irradiation, as has been discussed in the previous section.
redox pair, similarly to the physical properties in the ground
states. In general, optical doping may represent a new
means of developing organic molecular super- or semi-
conductors, and we have identified a variety of materials
Origin of Giant Photoconductivity
In most insulators, all the electrons below EF form elec- that we intend to study with regard to their electrical prop-
tron pairs in bound states, which represent closed-shell erties under irradiation. These compounds may indeed exhi-
structures or so-called charge-ordered states. For this bit totally different behaviors under irradiation, as photoex-
reason, insulators are generally diamagnetic, as is cited states generally have rather different wave functions
NMQ[Ni(dmit)2]. However, in NMQ[Ni(dmit)2], the major (i.e., electronic states) compared with the thermally access-
intermolecular interactions (S1 and S2) through which un- ible or ground states with which we are more familiar.
paired electrons on each Ni(dmit)2 species are paired are
weak and generate only a small degree of stabilization en-
ergy (ca. 0.24 eV). Weak interactions such as these are often
Conclusions
and uniquely observed in molecular salts and may be re-
NMQ[Ni(dmit)2] exhibits an unusually high conductivity
lated to the frequent observation of the melting of charge- ratio (σUV/σdark) exclusively under irradiation at 375 nm.
ordered states in such salts on photoirradiation.[17] Photo- This is considered to originate from CT between the Ni-
irradiation produces unpaired electrons and holes at both (dmit)2 species. This CT provides a suitable amount of car-
excited and original energy levels/bands. In these salts, the riers on the Ni(dmit)2 moieties in a process that can be re-
slightest change in the number of carriers may sometimes garded as melting of the charge-ordered state. The condi-
completely break down the charge-ordered state in a phen- tions required to achieve this melting state have not yet
omenon similar to a phase transition.[17] In NMQ[Ni- been fully elucidated. The high photoconductivity ratio
(dmit)2], the energy bands are close to one another (Fig- seen here is more important in terms of practical applica-
ures 2 and S2) and, thus, excited states generated by optical tions than the absolute value of photoconductivity, σUV
.
transitions cannot be singled out even with excitation by The findings in this work demonstrate that it is possible to
laser. As the three different types of contribution to conduc- realize a high photoconductivity ratio with this kind of typ-
tion coexist under UV irradiation (Figure 6f), only a par- ical insulator with characteristics of both molecular and
ticular optical transition may lead to melting of the charge- ionic crystals.
ordered state. Additionally, it should be noted here that the
photoconduction mechanism is not explained by any kind
of phase transition, as, unlike phase transitions, a high pho-
Experimental Section
Materials: Acetonitrile (the purest grade available) was purchased
from Wako Pure Chemicals and used as received. N-Methylquinoli-
nium iodide (NMQ·I) was prepared as follows. A solution of quin-
oline (1.29 g, 10 mmol; Eastman Synthetic) and a large excess of
methyl iodide (3.24 g, 22.8 mmol; Tokyo Chemical Industries) in
methanol (20 mL) was heated to reflux for 1.5 h. The resultant yel-
low solution was concentrated to leave a dark reddish-orange tar,
which was cooled in an ice bath to yield a colorless crystalline solid.
Further cooling solidified the product to form a yellow cake. The
crude material was recrystallized from methanol/diethyl ether and
dried under vacuum to produce pale yellow needles. Yield: 0.91 g
(34%): C10H10NI (271.10): calcd. C 44.30, H 3.72, N 5.17; found
toconductivity ratio (σUV/σdark) is observed at all tempera-
tures, and additionally because the original dark conduc-
tion continues to exist in parallel. These points distinguish
the behavior of NMQ[Ni(dmit)2] from that of known pho-
toconductors. Further study and comparison with related
salts is expected to provide a deeper understanding of this
unique photoconduction mechanism.
Significance of Giant Photoconductivity
An evident advantage of the technique reported here
over existing methods to control conductivity is the pos-
sibility of realizing junction structures from a single crystal
by using photolithography.[18] Junction structures are com-
posed of materials that exhibit qualitatively different con-
ducting properties from one another and are required in
electronic and semiconducting devices. Such fine control of
conducting properties within a given molecular crystal can-
not be realized by applying pressure or magnetic/electric
fields or by adding dopants during crystallization.
C 43.34, H 3.90, N 5.20. IR (KBr): ν = 1525, 810, 775 cm–1 (for
˜
details, see the spectrum in Figure S1). MS (FAB +): m/z = 144
[C10H10N]+ (144.19 calcd. for [C10H10N+]). UV/Vis (CH3CN): λ
(logε) = 316.0 (3.67), 238.0 (4.44), 196.0 (4.58) nm (for details, see
the spectrum in Figure 3). (nC4H9)4N[Ni(dmit)2] (abbreviated as
TBA[Ni]) was prepared as reported previously.[4] Single crystals of
the title compound were obtained from double decomposition in
acetonitrile as follows. Acetonitrile solutions of NMQ·I and
TBA[Ni] were separately prepared (10 mg of solute was added to
20 mL of the solvent in both cases) and filtered. The former solu-
Eur. J. Inorg. Chem. 2014, 4000–4009
4007
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
tion was added to the latter, and the resulting mixture was loosely
sealed with plastic (polyvinylidene chloride) film and allowed to
stand undisturbed at room temp. to allow gradual evaporation of
the solvent until precipitation of the crystals was observed (from
one day to several weeks). The single crystals were collected by
filtration, washed thoroughly with acetonitrile and acetone, and
then dried in vacuo. C16H10NNiS10 (595.60): calcd. C 32.27, H
1.69, N 2.35; found C 32.24, H 1.99, N 2.38.
with a quartz optical window. The apparatus was a cryostat (made
in-house) consisting of a picoammeter/voltage source (Keithley
6487), a Daikin Cryotec U104CWA helium compressor, a digital
temperature controller (Scientific Instruments, Model 9650), and a
turbo molecular pump (Adixen). Photoconductivity was similarly
measured by using a UV laser [(375Ϯ5) nm, Neoark] equipped
with focus and power adjusters. The incident beam was finely tuned
and focused on the sample by using a six-axis (XYZ–ωθφ) precise
positioning stage (Suruga Seiki). The light intensity at the sample
surface was measured with a power meter (Ophir, NOVA, PD-300-
UV). The sample was set on an Si diode thermometer (LakeShore
DT-470-SD) with Apiezon N grease to allow real-time measure-
ment of the sample temperature as accurately as possible. As it was
difficult to ascertain the penetration depth of the UV light accu-
rately, the resistivity under irradiation was reported as the surface
resistance Rsq [Ω] rather than the resistivity ρ [Ωcm].
Crystal Structural Analysis: X-ray structural analysis was per-
formed with the obtained single crystals of NMQ[Ni(dmit)2] at
296 K. A black platelet crystal of C16H10NNiS10 with approximate
dimensions of 0.500ϫ0.250ϫ0.025 mm was mounted on a glass
fiber. All measurements were performed with a Rigaku R-AXIS
RAPID diffractometer with graphite-monochromated Cu-Kα radi-
ation. The crystal-to-detector distance was 127.40 mm. Among the
47580 reflections collected, 8065 were unique (Rint = 0.1752), and
equivalent reflections were merged. The data were collected and
processed by using CrystalClear software (Rigaku). The linear ab-
sorption coefficient, μ, for the Cu-Kα radiation was 101.274 cm–1.
An empirical absorption correction was applied and resulted in
transmission factors ranging from 0.316 to 0.776; the data were
also corrected for Lorentz and polarization effects. The structure
was solved by direct methods and expanded by using Fourier tech-
niques. The non-hydrogen atoms were refined anisotropically, and
hydrogen atoms were refined by using a riding model. The final
cycle of full-matrix least-squares refinement on F2 was based on
8063 observed reflections and 507 variable parameters and con-
verged [the largest parameter shift was 0.00 times its estimated stan-
dard deviation (esd)] with the unweighted and weighted agreement
Magnetic Susceptibility Measurements: The temperature depen-
dence of the magnetic susceptibility was measured between 2 and
315 K by using the DC mode of a superconducting quantum inter-
ference device (SQUID) susceptometer (Quantum Design MPMS-
XL7). During these tests, the temperature was varied at a rate of
0.5–2 K/min. The applied field was 1 T, and measurements were
performed under reduced-pressure He under dark conditions. The
measurements were repeated several times with different samples
from different batches to check reproducibility. A polycrystalline
sample (typically ca. 1 mg; 1ϫ10–6 mol) was packed in a gelatin
capsule with quartz wool, which was subsequently set in the middle
of a straw fixed to the end of a sample rod purchased from Quan-
tum Design. The diamagnetic contributions from the capsule and
quartz wool were estimated from independent measurements prior
to the loading of the sample into the capsule. Although diamagne-
tism was expected to be observed, field and zero-field cooling pro-
cesses were conducted to confirm that both results were identical
within experimental error and, thus, demonstrate that there were
no artifacts from magnetic contamination. Similarly, a magnetiza-
tion versus magnetic field (M–H) curve was generated to confirm
the linearity of the susceptibility around 1 T. To ascertain the
susceptibility under UV irradiation, a quartz rod (3.1 mm diameter,
1300 mm length) was utilized as the sample support, and the
375 nm laser head was attached to the top of the rod with a
multimode optical fiber. The finely ground powder sample was
spread on a section of Kapton® tape, which was set in the middle
of the polyethylene straw such that the tape intercepted the laser
beam. The sample was continuously irradiated during the zero-field
cooling process (to 2 K) and also during successive susceptibility
measurements from 2 to 300 K under 1 T in the SQUID magne-
tometer (Quantum Design MPMS-5S). The measurements were
initially performed without irradiation to obtain background val-
ues, following which the same measurements were repeated under
UV irradiation at 375 nm. The magnetic susceptibility was exam-
ined at different laser intensity levels (17 and 0.30 mWcm–2 as mea-
sured at the end of the quartz rod). The distance between the end
of the quartz rod and the sample was ca. 2.5 cm, which assured
that the diamagnetic signal from the quartz rod did not affect the
results.
2 2
factors R1 = Σ||Fo| – |Fc||/Σ|Fo| = 0.0870 and wR2 = {Σ[w(Fo2 – Fc ) ]/
Σw(Fo2)2}1/2 = 0.2376. The standard deviation of an observation of
unit weight was 1.05. Unit weights were used. The maximum and
minimum peaks on the final difference Fourier map corresponded
to 0.77 and –0.88 eÅ3, respectively. Illustrations of the molecular
and crystal structures were produced by using Ortep 3 software for
Windows or VESTA.[19] CCDC-957799 contains the supplemen-
tary 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.
Tight-Binding Band Calculation: By using the atomic parameters
obtained from the X-ray structural analysis, the band structure was
calculated on the basis of a tight-binding band approximation. All
the calculations (molecular orbitals, overlap integrals, band disper-
sion, and Fermi surface) were performed by using the Caesar soft-
ware package (versions 1.0 and 2.0, PrimeColor Software). The
atomic orbital parameters used in the calculation are tabulated in
Table S1 in the Supporting Information.
Resistivity Measurements: The crystals were typically platelets, and
the most-developed crystal facet was often the ab or bc plane. As
the electrical resistivity of the title compound is high (surface resis-
tivity Rsq Ͼ 105 Ω at any temperature) whether it is UV-irradiated
or not, the resistivity was measured along the b axis by using a
direct current two-probe method with gold paste (Tokuriki Chemi-
cal Research Co., Ltd. No. 8560) and gold wires (25 μm diameter)
as the electrical contacts. The crystal surface was freshly cleaved
and coated with gold by vapor deposition, and gold wires were
attached to the deposited gold thin film immediately before every
measurement. The gap between the two electrodes was maintained
at 50 μm by using the same mask during gold deposition. A con-
stant voltage of 2 V was applied, and the resultant current was
measured. All the measurements were performed in vacuo (ca.
10–5 Pa) by using a commercially available refrigerator equipped
Absorption and Diffuse Reflectance Spectroscopy: To determine the
CT wavelengths between NMQ and Ni(dmit)2, diffuse reflectance
spectra were measured with a Hitachi U-4000 spectrophotometer.
Single crystals were briefly examined by obtaining X-ray photo-
graphs to exclude the possibility of different phases, that is, crystals
with different stoichiometries and/or different crystal structures.
The samples were then finely ground in an agate mortar along with
dry KBr to produce diluted powder specimens. Each powdered
Eur. J. Inorg. Chem. 2014, 4000–4009
4008
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
sample was sandwiched between a pair of quartz glass plates, and
its reflectance spectrum was acquired over the range 220–2500 nm
at room temp. under an ambient atmosphere. No noise peaks or
dispersion was observed in the baseline spectrum obtained by scan-
ning quartz glass plates in air over 220–2500 nm. As it was difficult
to distinguish small peaks from noise and artifacts, each spectrum
was acquired five times by using independently prepared samples,
and the results were averaged. To identify the intramolecular transi-
tions in NMQ and Ni(dmit)2, the solution spectra of (TBA)n[Ni-
(dmit)2] (n = 1,2) and NMQ·I (190–1100 nm) in acetonitrile were
also acquired by using a JASCO V-630 spectrophotometer at 20 °C
at a resolution of 2 nm. A trace amount of triethylamine was added
to the Ni(dmit)2 solution to prevent air oxidation of the [Ni-
(dmit)2]n– (n = 1,2), except for spectral measurements between 200
and 250 nm, in which region triethylamine has an intense absorp-
tion.
ure S1), calculated band structures over a wider energy range (Fig-
ure S2), photoresponses to light of various wavelengths (Figure S3),
crystal structure at 93 K (Figure S4), XPS data (Figure S5).
Acknowledgments
The elemental analysis and some of the spectroscopic measure-
ments were carried out at the Integrated Center for Sciences, Ehime
University.
[1] H. Mori, J. Phys. Soc. Jpn. 2006, 75, 051003.
[2] P. Cassoux, L. Valade, H. Kobayashi, A. Kobayashi, R. A.
Clark, A. E. Underhill, Coord. Chem. Rev. 1991, 110, 115–160.
[3] A. Kobayashi, Y. Okano, H. Kobayashi, J. Phys. Soc. Jpn.
2006, 75, 051002.
X-ray Photoelectron Spectroscopy: X-ray photoelectron spec-
troscopy (XPS) patterns were obtained by using Mg-Kα radiation
(1253.6 eV, 10 kV, 10 mA) with a JPS-9200 (JEOL) spectrometer at
room temp. The single crystals were placed on a small piece (ca.
5ϫ5 mm2) of conductive carbon tape (Nisshin EM) on a titanium
sample stage, and the glancing angle of the X-ray beam to the sam-
ple surface was ca. 35°. This setup enabled us to analyze the elec-
tronic states at a depth of a few nanometers (ca. a few unit cells)
from the surface, which is much shallower than the penetration
depth of the UV/Vis/NIR light (in the order of 1 μm) in this mate-
rial, yet sufficiently deep to allow observations of the bulk elec-
tronic state, as shown in the Supporting Information. The X-ray
beam had a square section of ca. 5ϫ5 mm2 at the sample surface,
and the area from which photoelectrons were collected (the field of
view) was ca. 3 mm in diameter. The pass energy and dwell time
were 10 eV and 100 ms, respectively. Including these parameters,
the measurement conditions (diffraction angle and selected area in
the electron lens system, sample position relative to X-ray and UV
laser incident beams, and laser focus) were optimized to provide
the best possible signal/background (S/B) ratio and signal/noise
(S/N) ratio. To maintain the intensity of the incident X-ray beam,
a monochromator was not used, and the energy resolution was ca.
0.75 eV as judged from the Ag 3d5/2 spectra measured under iden-
tical conditions. As sulfur-containing compounds can be unstable
under exposure to soft X-rays, the reliabilities of the spectra were
confirmed by a series of XPS measurements, in which spectral
changes were followed on progressively prolonged exposure of the
samples to the X-ray beam. Similarly, the effects of Ar+-etching
before XPS measurements was carefully examined, as the [Ni-
(dmit)2]n– (0 Ͻ n Յ 2) compounds are generally sensitive to oxi-
dants such as Ar+. Details of the means by which we determined
the best measurement conditions are described in the Supporting
Information (Figure S5). The XPS spectra of samples under UV
irradiation were measured by using the same laser setup as that
applied in other experiments in this work, and the sample was irra-
diated through a quartz window attached to the XPS chamber.
Spectral measurements under dark and irradiated conditions were
performed in series, and the order in which the measurement condi-
tions (irradiated or dark) were applied was not found to affect the
results. In addition to assessing the reproducibility of measure-
ments, the O 1s peaks were used to judge whether there was a
charge-up effect. All of the spectra shown in Figure 6 were con-
firmed to have good reproducibility and to be free of artifacts due
to decomposition/degradation of the samples.
[4] G. Steimecke, H.-J. Sieler, R. Kirmse, E. Hoyer, Phosphorus
Sulfur Relat. Elem. 1979, 7, 49–55.
[5] In this paper, the HOMOs and LUMOs are always based on
the neutral-molecule states.
[6] H. Tajima, T. Naito, M. Tamura, A. Kobayashi, H. Kuroda,
R. Kato, H. Kobayashi, R. A. Clark, A. E. Underhill, Solid
State Commun. 1991, 79, 337–341.
[7] R. Kato, Y.-L. Liu, S. Aonuma, H. Sawa, Synth. Met. 1997,
86, 2087–2088.
[8] R. Kato, Bull. Chem. Soc. Jpn. 2000, 73, 515–534.
[9] H. Mori, M. Kamiya, M. Haemori, H. Suzuki, S. Tanaka, Y.
Nishio, K. Kajita, H. Moriyama, J. Am. Chem. Soc. 2002, 124,
1251–1260.
[10] For recent progress in high-TC cuprates, see: “SPECIAL TOP-
ICS: Recent Development in Superconductivity”, J. Phys. Soc.
Jpn. 2012, 81, 011001–015002.
[11] J. E. Han, O. Gunnarsson, V. H. Crespi, “Superconductivity in
Fullerides” in Frontiers in Superconducting Materials (Ed.:
A. V. Narlikar), Springer, Berlin, 2005, pp. 231–253.
[12] T. Naito, T. Karasudani, K. Ohara, T. Takano, Y. Takahashi,
T. Inabe, K. Furukawa, T. Nakamura, Adv. Mater. 2012, 24,
6153–6157.
[13] T. Naito, T. Karasudani, S. Mori, K. Ohara, K. Konishi, T.
Takano, Y. Takahashi, T. Inabe, S. Nishihara, K. Inoue, J. Am.
Chem. Soc. 2012, 134, 18656–18666.
[14] J. P. Cornelissen, E. J. Creyghton, R. A. G. de Graaff, J. G.
Haasnoot, J. Reedijk, Inorg. Chim. Acta 1991, 185, 97–102.
[15] Owing to the heavy disorder of the quinoline ring atoms at
296 K, the single-crystal X-ray structural analysis was also per-
formed at 93 K. Although the thermal factors of the quinoline
ring atoms become much smaller, the crystal retains its essen-
tial features such as the overall molecular arrangement, the
space group (P21/n), and the lattice parameters (for details, see
Supporting Information).
[16] Note that difference in the absolute values of χ and its tempera-
ture dependence at the lowest temperatures should originate
from the difference in contamination of each sample.
[17] For recent reviews, see: T. Naito (Ed.), Molecular Electronic
and Related Materials
– Control and Probe with Light,
Transworld Research Network, Kerala, 2010.
[18] This method would transform the crystals into semiconducting
devices only during the irradiation, and the crystal would re-
turn to the initial diamagnetic insulators (“initialized”) on
ceasing the irradiation.
[19] a) Ortep 3 for Windows: L. J. Farrugia, J. Appl. Crystallogr.
1997, 30, 565; the program is available from http://
www.chem.gla.ac.uk/_louis/software/ortep3/; b) Visualization
for Electronic and STructural Analysis, http://jp-minerals.org/
vesta/jp/.
Supporting Information (see footnote on the first page of this arti-
Received: February 10, 2014
cle): Parameters of atomic orbitals (Table S1), IR spectra (Fig-
Published Online: May 27, 2014
Eur. J. Inorg. Chem. 2014, 4000–4009
4009
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim