Synthesis and characterization of inorganic-organic hybrid materials based on the intercalation of stable organic radicals into a fluoromica clay

Article abstract of DOI:10.1039/c3dt50627d

Hybrid materials, in which stable organic radical cations are intercalated into layered inorganic host materials, can be successfully synthesized via an ion exchange reaction in a layered fluoromica clay, to yield recyclable heterogeneous catalysts for oxidation of various alcohols. We have conducted systematic synthetic and structural studies on the intercalation of the radical cations 4-(diethylmethylammonium)-2,2,6,6-tetramethylpiperidin-1-oxyl (DEMTEP), 1-[2-(4-amino-2,2,6,6-tetramethyl-1-piperidinyloxyl)-2-oxoethyl] -1′-methyl-4,4′-bipyridinium (VIOTEP), and 2-(3-N-methylpyridinium)- 4,4,5,5-tetramethyl-4,5-dihydro-1H-imidazol-1-oxyl-3-N-oxide (m-MPYNN) into a synthetic fluoromica clay named Somasif ME 100, Na_2xMg _3.0-xSi₄O₁₀(F_yOH_1-y) ₂ (x = 0.33, y = 0.98). The guest-host interactions in these intercalation compounds have been characterized by X-ray powder diffraction and solid-state NMR of the constituent nuclei (²³Na, ¹⁹F, and ²⁹Si) of the Somasif structure. The intercalation process can be conveniently monitored using ²³Na MAS-NMR. Guest-guest interactions have been probed by magnetic susceptibility measurements as well as EPR and ¹H MAS NMR experiments. The ¹H MAS-NMR line widths and chemical shifts probe modifications in the electron spin density distributions and/or intermolecular interactions between the electron spins of the guest species. Despite these indications of weakly interacting spins, magnetic susceptibility measurements are consistent with the near-absence of cooperative magnetism. The VIOTEP and DEMTEP inclusion compounds demonstrate catalytic activity for the oxidation of benzyl alcohol, using NaOCl as a co-oxidant. Although the radical ion species is partially leached out under these conditions (ionic exchange with Na⁺ in solution) the catalytic activity remains for up to 40 subsequent cycles. Fully leached materials can be regenerated by catalyst re-loading and this process can be conveniently monitored by X-band EPR spectroscopy.

Full text of DOI:10.1039/c3dt50627d

Showcasing collaborative research from the Hellmut
Eckert and Armido Studer labs, SFB 858, Westfälische
Wilhelms-Universität Münster, Germany
As featured in:
Dalton
Transactions
Volume 42 | Number 24 | 28 June 2013 | Pages 8521–8962
An international journal of inorganic chemistry
www.rsc.org/dalton
Synthesis and characterization of inorganic–organic hybrid
materials based on the intercalation of stable organic radicals
into a fluoromica clay
Ion exchange of fluoromica clay with nitroxide radical cations
results in intercalation compounds that can serve as recyclable
oxidation catalysts. The guest molecular orientation is composition
dependent and controls the strength of intermolecular guest-host
and guest-guest interactions.
ISSN 1477-9226
COVER ARTICLE
Petzold et al.
Biphenyl bridged hexadentate N6-ligands – a rigid ligand backbone for
Fe(II)
spin crossover complexes
1477-9226(2013)42:24;1-M
See Studer, Dalton Trans., 2013, 42, 8585.
www.rsc.org/dalton
Registered Charity Number 207890

Dalton
Transactions
View Article Online
View Journal | View Issue
PAPER
Synthesis and characterization of inorganic–organic
hybrid materials based on the intercalation of stable
organic radicals into a fluoromica clay
Cite this: Dalton Trans., 2013, 42, 8585
a
b
b
c
Zhaoyang Zeng, David Matuschek, Armido Studer,* Christian Schwickert,
c
a
Rainer Pöttgen and Hellmut Eckert*
Hybrid materials, in which stable organic radical cations are intercalated into layered inorganic host
materials, can be successfully synthesized via an ion exchange reaction in a layered fluoromica clay, to
yield recyclable heterogeneous catalysts for oxidation of various alcohols. We have conducted systematic
synthetic and structural studies on the intercalation of the radical cations 4-(diethylmethylammonium)-
2
,2,6,6-tetramethylpiperidin-1-oxyl (DEMTEP), 1-[2-(4-amino-2,2,6,6-tetramethyl-1-piperidinyloxyl)-2-
oxoethyl]-1’-methyl-4,4’-bipyridinium (VIOTEP), and 2-(3-N-methylpyridinium)-4,4,5,5-tetramethyl-4,5-
dihydro-1H-imidazol-1-oxyl-3-N-oxide (m-MPYNN) into a synthetic fluoromica clay named Somasif ME
®
1
00, Na2xMg3.0−xSi
4
O
10(F
y
OH1−y
)
2
(x = 0.33, y = 0.98). The guest–host interactions in these intercalation
compounds have been characterized by X-ray powder diffraction and solid-state NMR of the constituent
2
3
19
29
nuclei ( Na, F, and Si) of the Somasif structure. The intercalation process can be conveniently moni-
2
3
tored using Na MAS-NMR. Guest–guest interactions have been probed by magnetic susceptibility
1
1
measurements as well as EPR and H MAS NMR experiments. The H MAS-NMR line widths and chemical
shifts probe modifications in the electron spin density distributions and/or intermolecular interactions
between the electron spins of the guest species. Despite these indications of weakly interacting spins,
magnetic susceptibility measurements are consistent with the near-absence of cooperative magnetism.
The VIOTEP and DEMTEP inclusion compounds demonstrate catalytic activity for the oxidation of benzyl
alcohol, using NaOCl as a co-oxidant. Although the radical ion species is partially leached out under
Received 6th March 2013,
Accepted 22nd April 2013
+
these conditions (ionic exchange with Na in solution) the catalytic activity remains for up to 40 sub-
DOI: 10.1039/c3dt50627d
sequent cycles. Fully leached materials can be regenerated by catalyst re-loading and this process can be
conveniently monitored by X-band EPR spectroscopy.
www.rsc.org/dalton
hydrogen-bond interactions can lead to orientationally ordered
Introduction
electron spin arrangements, offering the prospect of coopera-
1
,2
Inorganic–organic hybrid materials involving stable organic tive magnetism. One particular type of hybrid material can
radicals are a class of materials exhibiting a range of interest- be formed by intercalating the spin carrying molecule in the
ing and unusual optical, magnetic, and catalytic properties. form of a radical cation into an inorganic layered host struc-
Designing particular physical properties requires an under- ture via an ion exchange reaction, whereby various possible
standing and synthetic control of the two basic intermolecular orientation scenarios can be realized (Fig. 1). This principle
interaction types – on the one hand, those involving the host has been demonstrated for some layered crystalline host
and the guest species, and on the other hand, those among matrices such as the clay material saponite Na
x 3 x 10
Mg Si4−xAl O -
3
–7
the guest species themselves. For example, immobilization of (OH) ·nH O and cadmium phosphorus sulfide CdPS
2
2
3
.
As
the radical species on surfaces and/or self-assembly via previously discussed, such materials can be used as recyclable
catalysts for the oxidation of alcohols. In the present study, we
report the intercalation of three different types of radical
cations, based on tetramethylpyridine N-oxide (TEMPO), a
nitronyl nitroxide, and a viologen into a fluoromica clay
a
Institut für Physikalische Chemie, WWU Münster, Corrensstrasse 28/30, 48149
Münster, Germany. E-mail: eckerth@uni-muenster.de
b
Organisch-Chemisches Institut, WWU Münster, Corrensstrasse 40, 48149 Münster,
material known under the trade name Somasif, Na2xMg3.0−x
-
Germany. E-mail: studer@uni-muenster.de
c
Si O (F OH ) (x = 0.33, y = 0.98). The materials have been
Institut für Anorganische und Analytische Chemie, WWU Münster, Corrensstrasse
4
10
y
1−y 2
28/30, 48149 Münster, Germany
thoroughly characterized with respect to their structural and
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 8585–8596 | 8585

View Article Online
Paper
Dalton Transactions
Fig. 1 Scheme showing nitroxide radicals intercalated into the layered Somasif host matrix. Depending on the orientation of the radical cations relative to the host
layer, the layer-to-layer distance is subject to significant variations. Parts a to c show three possible scenarios, corresponding to parallel, transversal, and longitudinal
orientations, respectively.
magnetic properties, using X-ray powder diffraction, magnetic cation exchange capacity (CEC) of 70 meq/100 g. The material
susceptibility measurements, multinuclear solid state NMR used in the present study has x = 0.33 and y = 0.98 according
8
and EPR spectroscopic techniques. In addition, the catalytic to chemical analysis. The three radical species used are
performance of these materials with respect to the oxidation 4-(diethylmethylammonium)-2,2,6,6-tetramethylpiperidin-1-oxyl
of benzyl alcohol has been tested.
(abbreviated as DEMTEP, Fig. 3a) iodide and 1-[2-(4-amino-
,2,6,6-tetramethyl-1-piperidinyloxyl)-2-oxoethyl]-1′-methyl-4,4′-
2
bipyridinium, chloride, iodide (1 : 1 : 1) (abbreviated as VIOTEP,
Experimental
Materials and synthesis
Fig. 3b), as well as 2-(3-N-methylpyridinium)-4,4,5,5-tetra-
methyl-4,5-dihydro-1H-imidazol-1-oxyl-3-N-oxide
iodide, (Fig. 3c, abbreviated as m-MPYNN).
(m-MPYNN)
The host matrix clay used in this study is Somasif® ME100, a
synthetic fluorohectorite clay produced by the CO-OP Chemical
Co., Japan. This material belongs to the 2 : 1 type layered
silicates (Fig. 2) and possesses a layer distance of 9.5 Å.
The chemical formula can be expressed as Na Mg
DEMTEP and m-MPYNN were prepared following the
10–12
reported methods.
VIOTEP was readily prepared as
described in Scheme 1. The educt 4-amino-2,2,6,6-tetramethyl-
8
1
2
-piperidinyloxyl and the precursor 4-[(2-chloroacetyl)amino]-
Si O -
4 10
2
x
3.0−x
,2,6,6-tetramethyl-1-piperidinyloxyl were synthesized using
9
(F
y
OH1−y
)
2
(x = 0.15–0.5, y = 0.8–1.0), which corresponds to a
13–15
standard literature procedures.
The second and third
steps of the reaction were prepared as follows.
Fig. 3 Lewis formulae of the radical species used in the present study: (a)
4
-(diethylmethylammonium)-2,2,6,6-tetramethylpiperidin-1-oxyl (DEMTEP) iodide,
b) m-MPYNN cation, 2-(3-N-methylpyridinium)-4,4,5,5-tetramethyl-4,5-dihydro-
H-imidazol-1-oxyl-3-N-oxide (m-MPYNN) iodide, (c) viologen tempo radical,
(
1
Fig. 2 Structure of the synthetic clay Somasif. Reprinted with permission from
1-[2-(4-amino-2,2,6,6-tetramethyl-1-piperidinyloxyl)-2-oxoethyl]-1’-methyl-4,4’-
ref. 8.
bipyridinium, chloride, iodide (1 : 1 : 1) (VIOTEP).
8
586 | Dalton Trans., 2013, 42, 8585–8596
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
Scheme 1 Synthesis of the VIOTEP radical.
1-[2-(4-Amino-2,2,6,6-tetramethyl-1-piperidinyloxyl)-2-oxoethyl]- Characterization
4
,4′-bipyridinium, chloride (1 : 1). In a sealed tube 4-[(2-chloro-
acetyl)amino]-2,2,6,6-tetramethyl-1-piperidinyloxyl (4.04 g,
6.35 mmol, 1.0 eq.) was dissolved in ACN and 4,4-bipyridine
2.80 g, 17.95 mmol, 1.1 eq.) was added. The reaction mixture
was stirred overnight at 60 °C and afterwards poured in Et
50 mL). The precipitate was filtered off and washed with
EtOAc (3 × 5 mL). The product was obtained as a brown solid
4.07 g, 10.09 mmol, 62%). IR (film): 3376w, 3337w, 3198w,
X-ray powder diffraction (XRD) was done on a Bruker D8
Advance diffractometer equipped with CuK
α
radiation (λ =
1
(
1
19
23
27
0.154 nm). H, F, Na and Al solid state NMR spectro-
scopic measurements were carried out on a Bruker DSX 500
spectrometer equipped with a 2.5 mm MAS NMR probe operat-
2
O
(
1
19
27
ing at a spinning speed of 30.0 kHz. H, F and Al
MAS-NMR spectra were recorded at 500.13 MHz, 470.56 MHz
and 130.33 MHz respectively, using 90° pulses of 1.6 to 4 μs
(
3
1
037m, 2932m, 1648s, 1639s, 1552s, 1461m, 1361m, 1303m,
23
length and a recycle delay of 1 s. Na MAS-NMR spectra were
−
1
242m, 1087m, 976m, 816s, 609m cm . HRMS (ESI) (m/z) cal-
measured at 132.2 MHz, using a 30° pulse of 0.9 μs length and
+
culated for [M] : 368.2212, found: 368.2202.
-[2-(4-Amino-2,2,6,6-tetramethyl-1-piperidinyloxyl)-2-oxoethyl]-
′-methyl-4,4′-bipyridinium, chloride, iodide (1 : 1 : 1). In a
sealed tube 1-[2-(4-amino-2,2,6,6-tetramethyl-1-piperidinyloxyl)-
29
a recycle delay of 2 s. Si MAS-NMR spectroscopic measure-
1
ments were carried out on an Apollo 300 spectrometer (reson-
ance frequency 59.7 MHz) equipped with a 4 mm MAS NMR
probe (MAS spinning frequency 10.0 kHz), using 90° pulses of
1
2
1
-oxoethyl]-4,4′-bipyridinium, chloride (1 : 1) (2.10 g, 5.22 mmol,
.0 eq.) was dissolved in MeOH (5 mL) and methyliodide 19
29
5
μs length and relaxation delays between 1 and 80 s. Si and
F spin lattice relaxation times were measured by the saturation
recovery method. All the MAS-NMR lineshape deconvolutions
(1.6 mL, 25.69 mmol, 4.9 eq.) was added. The reaction
mixture was stirred overnight at 60 °C and afterwards
directly crystallized at −20 °C. The precipitate was filtered
16
were carried out using the DMFIT programme. Chemical shifts
19
3 3
are reported relative to CFCl ( F, using AlF as a secondary stan-
off and washed with Et
was obtained as a brown solid (2.85 g, 5.22 mmol, 99%). IR
film): 3436br, 3222w, 3039m, 2975m, 2931m, 1679s, 1638s,
2
O (3 × 10 mL). The product
23
1
29
dard), 1 M NaCl solution ( Na), and tetramethylsilane ( H, Si),
respectively. EPR measurements were conducted at room temp-
erature on an EMS 104 EPR Analyzer from Bruker with an X-band
(
1
5
553m, 1504w, 1449m, 1346m, 1240m, 1102w, 974w, 828m,
(9.7 GHz) spectrometer (magnetic field strength 0.34 T) using the
48m cm− . HRMS (ESI) (m/z) calculated for [M] : 191.62180,
1
2+
following conditions: microwave power between 10 and 25 mW,
+
found: 191.62185. Calculated for [M] : 383.23968, found:
3
−1
field sweep rate between 0.9 and 1.9 mT s , field sweep range
83.24420.
between 10 and 20 mT, modulation amplitude between 0.1 and
ESI-MS (m/z) and HRMS (m/z) were performed on a Bruker
0
.8 mT, modulation frequency 50 kHz. Magnetic susceptibility
measurements were carried out at a magnetic field strength of
T, in the temperature range of 3–305 K using a Quantum
MicroTof or a Thermo Scientific Orbitrap LTQ XL. Infrared
spectra were recorded on a Digilab Varian 3100 FT-IR Excalibur
Series.
The ion exchange intercalation was carried out by stirring
the clay in an aqueous solution of the radical salts for 4 hours
at room temperature. Typically for one synthesis 200 mg clay
was stirred in 20 mL of a 24 mM radical solution, which
means 0.48 mmol of the radical, corresponding to 300% CEC.
The product was filtered, washed until a colorless filtrate
was obtained, and subsequently dried under vacuum
1
Design Physical Property Measurement System (PPMS). All
measurements were carried out using the VSM option by packing
the powdered samples (9.437 mg Som(DEMTEP) 46% CEC,
6.761 mg Som(m-MPYNN) 71% CEC, 7.019 mg Som(m-MPYNN)
20% CEC and 6.918 mg Som(VIOTEP) 62% CEC) in polypropy-
lene capsules and attaching them to a sample holder.
Catalytic studies
(0.02 mbar) for 1–2 days. All the intercalation compounds are
denoted as “Som(Radical), x% CEC”, where x is the loading The immobilized TEMPO ammonium salt (1.5–20 mol% with
level.
respect to the amount of substrate) was added to a reaction
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 8585–8596 | 8587

View Article Online
Paper
Dalton Transactions
vessel which was equipped with a frit to filter off the reaction excess of the radical offered. Higher loadings (46% CEC) of
mixture directly. A solution of the alcohol (0.50 M, 1.0 equiv.) DEMTEP could be realized by using a 40 mM radical solution
in CH
2
Cl
2
, a solution of KBr (aq., 0.50 M, 10 mol%) and a solu- with doubly-distilled water as a solvent and stirring for
(approx. 10 mg, pH = 5 hours.
.0) were added at 0 °C. The reaction mixture was shaken for Guest species orientation inferred from X-ray powder
h at 0 °C. The catalyst was filtered off from the reaction diffraction. Fig. 5a–c summarize the X-ray powder diffraction
Cl (2 × 2 mL) and the conversion
tion of NaOCl (1.25 eq.) and NaHCO
9
3
3
mixture, washed with CH
2
2
was determined by means of gas chromatography using a
Hewlett-Packard GC HP6890-1 (30 m × 0.25 μm) column and a
GC Agilent Technologies 7890A (30 m × 0.25 μm) column with
hydrogen at about 1 bar. Subsequently the procedure was
repeated up to 40 times, every time by re-exposing the catalyst
material to a fresh reaction mixture.
Results and discussion
Guest–host interactions
Guest species uptake. The amount of the intercalated
radical (reported in units of % CEC) was quantified via CHN
elemental analysis. The level of intercalation x can be calcu-
lated from the measured radical content (in wt%) using the
5
following formula:
wt%radical Msomasif
x ¼
Mradical ꢀ wt%radical ðMradical ꢀ MNaÞ
Fig. 4 shows the uptake curve as a function of the amount
of the radical offered in the solution. The horizontal axis is
defined as the molar ratio of the offered radical cation in the
original solution to the total sodium content in the host
matrix (nradical/nNa+). The ordinate shows the amount x of the
intercalated radical in the host structure. Both curves in Fig. 4
demonstrate that by increasing the amount of the radical
offered, the amount of intercalated material can be increased
until a saturation value around 30% CEC for Som(DEMTEP)
and 70% CEC for Som(m-MPYNN) is reached, even with high
Fig. 4 Uptake curves of the radicals m-MPYNN and DEMTEP, determined by
CHN elemental analysis, intercalated into Somasif under the conditions
described in the text.
Fig. 5 X-ray powder diffraction pattern of (a) Som(m-MPYNN), (b) Som-
(DEMTEP) (bottom) at different levels of intercalation, and (c) of Som(VIOTEP)
before and after 20 subsequent catalytic conversions.
8
588 | Dalton Trans., 2013, 42, 8585–8596
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
Table 1 Intercalation degree [±2ꢀ CEC], layer-to-layer distance d(001) and assume a considerable extent of disorder. The broad reflection
interlayer spacing (±0.2 Å) of the intercalation compounds investigated
(
tailing) of the 20% CEC sample indicates a distribution of
domains with parallel and transversal orientations in the
manner as shown in Fig. 1a and b as well as domains with
incomplete loading.
Intercalation
degree
Interlayer Corresponding
d(001) [Å] spacing [Å] arrangement
[
% CEC]
Fig. 5b shows the results for the DEMTEP intercalation
compounds. Again, the layer-to-layer distance tends to increase
with increasing radical loading level, even though the changes
appear less gradual. At the lowest loading (7% CEC) the layer-
to-layer distance is 12.1 Å, corresponding to an interlayer gap
size of 5.1 Å. Within the concentration range 21–33% CEC, a
constant layer-to-layer distance of 14.6 Å is observed, corres-
ponding to an interlayer gap of 7.6 Å. As the DEMTEP radical
ion has the estimated molecular dimensions of 8.7 × 6.3 ×
Som(m-MPYNN) series
+
1
8
9.6
2.2
14.8
15.0
15.2
2.6
5.2
7.8
8.0
8.2
11.9
10.2
(Na )
1
Parallel
2
3
4
20
42
51
Transversal
Transversal
Transversal
Longitudinal
Disordered
transversal/
longitudinal)
Longitudinal
18.9
5
61
17.2
(
6
71
18.9
11.9
Som(DEMTEP) series
4.6 Å, these two interlayer gap sizes are most consistent with a
1
2
3
4
5
6
7
7
12.1
14.6
14.6
14.7
14.6
14.6
14.7
5.1
7.6
7.6
7.7
7.6
7.6
7.7
10.5
8.9
5.6
Parallel
parallel arrangement at the lowest concentration and a trans-
versal orientation in the intermediate concentration range.
The Som(DEMTEP) sample with the highest degree of inter-
calation (46% CEC) shows in addition to the 14.6 Å distance
also larger distances up to 17.5 Å, which suggests certain
layers with transversal and other layers with longitudinal inter-
calation, similar to the 51% CEC m-MPYNN sample.
21
30
32
33
35
46
Transversal
Transversal
Transversal
Transversal
Transversal
Transversal
Longitudinal
17.5
Som(VIOTEP)
56
15.9
12.6
After 20 catalytic ∼10
cycles
Fig. 5c shows the XRD data of the VIOTEP radical inter-
calate. Here the expansion of the layer-to-layer distance of
15.9 Å (interlayer gap 8.9 Å) allows no straightforward con-
data obtained on all intercalated compounds. For the pure clusion about the orientation, as the molecular dimensions
host compound, the (001) reflection appears at d = 9.5 Å, are not known owing to the absence of a crystal structure. The
which corresponds to the layer-to-layer distance. From this diffraction pattern observed for the material following 20 suc-
value, the size of the interlayer gap can be deduced to be 2.5 Å. cessive runs is also shown. For this material, the expansion is
As illustrated in Fig. 5 and Table 1, the (001) reflection is sig- seen to be considerably diminished (12.6 Å), consistent with a
nificantly shifted in the intercalated materials, revealing sub- significantly altered orientation distribution of those VIOTEP
stantial expansions of the layer-to-layer distance for the molecules still retained within the host structure at this stage.
materials studied. In the case of the m-MPYNN intercalation The very broad diffraction peaks indicate a distribution of
compounds, Fig. 5a shows that with increasing intercalated domains with different degrees of intercalation, a consequence
radical contents the layer-to-layer distance changes continu- of the leaching (vide infra).
2
3
19
29
ously from 12.2 Å to 18.9 Å, corresponding to a continuous
Na, F, and Si solid state NMR spectroscopy. Solid state
increase of the size of the interlayer gap from 5.2 Å to 11.9 Å, NMR spectroscopic studies of the constituent nuclei in the
at the lowest and the highest radical contents, respectively. By host structure were conducted to probe the interactions
comparing the extent of the interlayer expansion with the between the paramagnetic guest species and the Somasif
2
3
guest species’s molecular dimensions, information about the slabs. Fig. 6a shows the Na MAS-NMR signals of the pure
guest species orientation relative to the layers can be deduced. host material and the three hybrid materials Som(DEMTEP),
To this end, the molecular dimensions were estimated from Som(m-MPYNN) and Som(VIOTEP), all of which have the
furthest interatomic distances in each dimension, as measured highest degree of intercalation realized. Somasif samples have
from the crystal structure, using the program DIAMOND. In been shown previously to yield three spectroscopically distinct
2
3
8
doing so, the molecules were approximated as rigid species
Na resonances: the signal near 38 ppm (isotropic chemical
iso
without internal mobility. For the m-MPYNN radical, with esti- shift δCS = 41.7 ppm, SOQE = 1.6 MHz) is assigned to non-
mated molecular dimensions of 9.8 × 5.8 × 4.6 Å, the orien- exchangeable sodium ions on the Mg positions within the
iso
tation of the radicals within the layers can be considered as aluminosilicate slabs. The resonance near −22.0 ppm (δCS
=
almost longitudinal, i.e. with the long axis of the molecule per- −19.8 ppm, SOQE = 1.5 MHz) arises from exchangeable non-
+
pendicular to the layer (Fig. 1c) for the samples with 71% CEC hydrated Na ions found in the interlayer space. Furthermore,
iso
while for the samples with lower degrees of intercalation, the a weaker signal characterized by δCS = −2.3 ppm, SOQE =
parallel and transversal orientations appear to be preferred. 2.9 MHz can be assigned to (at least partially) hydrated
+
The 51% CEC sample shows almost equal contents of exchangeable Na ions in the interlayer space. Fig. 6a shows
domains with longitudinal and transversal orientations. The that in the intercalation compounds having the maximum
+
d spacing of the 61% CEC sample is in between both degree of intercalation, only the non-exchangeable Na ions
orientations. In view of the enhanced half width, one can remain. Fig. 6a also includes a spectrum of a Som(VIOTEP)
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 8585–8596 | 8589

View Article Online
Paper
Dalton Transactions
surface adsorbed sodium ions species from the NaOCl electro-
lyte solution used in the catalytic reaction. Fig. 6b and c show
2
3
the
Na MAS-NMR spectra of the Som(m-MPYNN)- and
Som(DEMTEP)-series as a function of the amount of the inter-
calated radical. As expected the resonance line near
−
22.0 ppm decreases monotonically as the degree of intercala-
23
tion is increased. The data illustrate that Na MAS-NMR spectro-
scopy could be a convenient tool for monitoring the progress
of the ion exchange reaction. In the case of the DEMTEP and
m-MPYNN intercalates the near-disappearance of the low-fre-
quency signal at the highest intercalation levels suggests that
virtually no exchangeable sodium ions remain in the interlayer
space. This is not, however, what is expected based on the CEC
values. Indeed, closer inspection of the signals reveals a broad
+
signal component due to hydrated Na ions, which is difficult
+
to quantify. It is also possible that part of the residual Na ions
in the interlayer space do not contribute to the spectrum, as
their line shapes may be broadened beyond detectability by
paramagnetic interactions with nitroxide radical ions in their
vicinity. In contrast, the line shapes of the non-exchangeable
+
Na ions appear completely unaffected by paramagnetic broad-
ening, suggesting that these ions are indeed spatially remote
from the interlayer space. These uncertainties make it difficult
to quantify the degree of intercalation on the basis of the inte-
2
3
grated Na NMR peak area ratio of the non-exchangeable and
+
the exchangeable Na ions. While the corresponding analysis
was attempted and indeed reveals the expected general trend,
the scatter observed by this analysis is significantly stronger
than that obtained from the chemical analysis data. Hence the
23
latter method must be considered superior to the Na NMR
MAS peak integration in determining the extent of ion
exchange achieved.
1
9
Fig. 7 presents the F MAS-NMR spectra of the intercala-
tion compounds with m-MPYNN and DEMTEP. As previously
discussed, the complex spectrum that is observed in the
pure host material reflects the presence of multiple F environ-
2
+
+
ments with different numbers of Mg ions, Na ions and
8
vacancies within their first coordination spheres. For the
m-MPYNN-series (Fig. 7a) the presence of the radical species
in these samples has no strong spectroscopic consequences
aside from a significant shortening of the spin-lattice relax-
1
ation times T (Table 2). In contrast, Fig. 7b shows that in the
Na MAS-NMR spectra of the intercalation compounds: (a) comparison case of the DEMTEP-intercalates the F MAS NMR line shapes
23
19
Fig. 6
between the different systems at their maximum respective intercalation levels,
b) comparison between different intercalation levels for the Som(m-MPYNN)
series, (c) comparison between different intercalation levels for the Som-
DEMTEP) series.
are significantly broadened and a drastic shortening of the T1
values is also observed. Thus, these results suggest that elec-
tron spin–host structure interactions in the DEMTEP intercala-
tion compound are significantly stronger than in the
m-MPYNN intercalate, despite the fact that the DEMTEP
(
(
sample exposed to 20 catalytic cycles. From chemical analysis species are present at considerably lower concentrations.
2
7
it is known that this sample has been partially leached and
The Al MAS-NMR spectra (data not shown) reveal single
thus contains significantly lower concentrations of radical peaks, with isotropic chemical shifts (67 ppm vs. 1 M aqueous
2
3
ions. Indeed, the Na NMR spectrum indicates that this leach- Al(NO3)3 solution) typical of four-coordinate Al species. No
ing effect has restored part of the sodium occupancies in the changes are detectable upon nitroxide intercalation. Finally,
2
9
exchangeable interlayer sites. The spectroscopic parameters the principal Si MAS-NMR signals (Fig. 8) of the hybrid
indicate that this sodium is mostly of the hydrated type. In materials also show no noticeable changes in the position of
addition, a strong signal near 7.0 ppm can be attributed to the dominant line (−95 ppm), whereas the minority line
8
590 | Dalton Trans., 2013, 42, 8585–8596
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
2
9
Fig. 8
Si MAS-NMR spectra of pure Somasif and the intercalation com-
pounds with different systems.
1
spin lattice relaxation times T (Table 2). Again these effects
are more pronounced for the DEMTEP than for the m-MPYNN
intercalates. However, in the case of the m-MPYNN intercalates,
Table 2 reveals a remarkable anomaly: with increasing radical
ion content the value of T
1
initially shows the expected
decrease, passes through a minimum near 42–51% CEC and
increases again towards higher loading levels, despite the
increased spin concentration. A plausible explanation for this
anomaly is based on Fig. 9, which illustrates that the three dis-
tinct intercalation scenarios present different distances
between the spin-carrying nitroxide groups of the m-MPYNN
radical cations and the silicon layers in the host matrix.
Specifically, the N–O groups are most closely situated towards
the layers in the transversal intercalates, while in the parallel
and longitudinal orientation they are much more remote.
1
9
Fig. 7
F MAS-NMR spectra and spin-lattice relaxation times of (a) the Som-
(
m-MPYNN) and (b) the Som(DEMTEP) series with different compositions.
2
9
Thus, the influence of the electron spin centre on the Si
layers is expected to depend on the arrangement of the radical
cations. Indeed, Table 2 reveals that the spin-lattice relaxation
times go through a minimum at those compositions at which
the fraction of molecules with transversal orientation reaches a
maximum. In contrast, in the DEMTEP intercalation com-
pounds, where no close approach of the nitroxide group to the
silicate slabs is expected for the transversal intercalation, the
2
9
19
Table 2
1
Si and F spin-lattice relaxation times T of the pure host material
Somasif and the hybrid materials Som(m-MPYNN), Som(DEMTEP) and Som-
VIOTEP)
(
1
9
29
Sample
F T
1
[ms]
1
Si T [ms]
Somasif
8635s (±226)
29 679 (±2506)
Som(m-MPYNN)
8
2
4
5
6
7
% CEC
472 (±77)
60 (±11)
22 (±4)
25 (±5)
22 (±5)
45 (±8)
0% CEC
2% CEC
1% CEC
1% CEC
1% CEC
355 (±35)
116 (±9)
119 (±23)
216 (±18)
575 (±60)
29
experimental trend in the Si spin lattice relaxation times just
reflects (roughly) the expected dependence on overall spin
concentration.
Som(DEMTEP)
7
2
3
3
3
4
% CEC
285 (±77)
31 (±7)
15 (±4)
15 (±4)
12 (±3)
8 (±2)
Guest–guest interactions
1% CEC
0% CEC
2% CEC
3% CEC
6% CEC
40 (±3)
23 (±3)
24 (±4)
75 (±21)
16 (±2)
Magnetic susceptibility and EPR measurements. To test for
cooperative magnetic ordering effects the bulk magnetic prop-
erties have been measured for four representative intercalation
compounds. To eliminate the contribution from the Somasif
host, the experimental χT versus T curve for the latter has been
subtracted from the data obtained on the intercalation com-
pounds. The resulting curves are shown in Fig. 10 and fitted to
SOM(VIOTEP)
6% CEC
10% CEC
5
∼
109 (±16)
872 (±48)
89 (±7)
7075 (±1150)
(
−98 ppm) is not visible in the intercalates. There are no sys- the expression χT = χ T + CT/(T − Θ). In this expression χ com-
0
0
2
9
tematic trends in the Si NMR line widths (about 5 ppm) prises the effect of the temperature independent diamagnetic
upon intercalation, but there is a significant shortening of the and van-Vleck paramagnetic contributions, C is the Curie
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 8585–8596 | 8591

View Article Online
Paper
Dalton Transactions
Fig. 9 A schematic description showing the different distances between the nitroxide groups on the m-MPYNN radical molecules and the silicon layer in the host
matrix depending on the orientation of the radical cations relative to the host layer: (a) parallel, (b) transversal, (c) longitudinal.
Table 3 Magnetic properties of representative Som(Radical) intercalation
compounds
(emu mol−
1
)
Compound
µ
eff (µ
B
/f.u.)
θ
P
(K)
χ
0
−
6
5
4
4
Som(DEMTEP) 46% CEC
Som(m-MPYNN) 71% CEC
Som(m-MPYNN) 20% CEC
Som(VIOTEP) 62% CEC
1.71(1)
1.66(1)
1.77(1)
1.10(1)
−0.4(5)
−1.9(5)
0.6(5)
9.32(1) × 10
6.63(1) × 10
−
−
−
−1.48(1) × 10
1.21(1) × 10
2.0(5)
Representative intercalation compounds were also studied
by EPR spectroscopy. The spectra measured on the m-MPYNN
intercalation compounds show a symmetric signal at g = 2.007
(
see Fig. 11). In all of the materials studied, the peak structure
expected from the anisotropic magnetic hyperfine coupling
1
4
with the N nuclei is averaged out owing to rapid spin-state
fluctuations caused by strong intermolecular electron–electron
interactions. The line width of this resonance shows a systema-
tic dependence on radical content, suggesting that the rate of
these spin state fluctuations increases with increasing levels
Fig. 10 Plot of the product of molar magnetic susceptibility and temperature,
χT, versus temperature for representative Somasif intercalation compounds
subject to the present study.
constant, and Θ is the Weiss constant characterizing coopera- of intercalation (motional narrowing effect). Such behaviour
tive ordering effects. Table 3 summarizes the parameters would be expected for the case of a random distribution
obtained. All the data are consistent with dominant Curie– (absence of clustering) of the radical species. An interesting
Weiss behavior. Based on the known radical ion concen- deviation from this monotonic behaviour is observed in the
trations the molecular magnetic moments of the radical DEMTEP series at the lowest intercalation level studied. Here
species can be determined from the C values obtained: they the line width appears narrower than expected on the basis of
B B
fall between 1.77µ and 1.66µ
, for the DEMTEP and the a random spin distribution, suggesting some extent of cluster-
m-MPYNN intercalates, respectively, corresponding closely to ing. This observation may be related to the parallel orientation
1
7
the theoretically expected value of 1.68µ
B
.
A substantially of the radical ion relative to the Somasif layers seen at this low
lower value of 1.11µ
B
is measured for the VIOTEP intercalate. ion concentration, which is distinctly different from the orien-
This result could arise from diamagnetic impurities or from tation at higher loadings. Finally, the spin state fluctuations
some clustering of the intercalated VIOTEP radical species. For among neighboring electron spins are the fastest in the pure
all the intercalated systems studied, the dominant Curie– radical ion salts, producing the narrowest lines within each of
Weiss behavior observed indicates the absence of cooperative the compositional series investigated.
1
magnetic interactions. We thus conclude that the strategy of
confining these radical cations into the interlayer space of clay
H MAS-NMR spectroscopy. Fig. 12a and b show high speed
1
H MAS-NMR spectra of the pure host material, the pure
materials does not lead to cooperative magnetic phenomena radical ion salts, as well as of their intercalation compounds
in the bulk, consistent with earlier results from our as a function of composition, for both the m-MPYNN and the
4
,5
laboratory.
DEMTEP series. For the pure host material, one sharp signal
8
592 | Dalton Trans., 2013, 42, 8585–8596
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
Fig. 11 X-Band EPR spectra of (a) Som(m-MPYNN) and (b) Som(DEMTEP) intercalation compounds depending on the degree of intercalation, (c) and (d) EPR line
widths as a function of the degree of intercalation for both systems.
1
9,20
at 4.0 ppm is observed, which arises from H O molecules molecule.
Only minor variations in resonance shifts are
2
present in the interlayer space. The spectra of the intercalation observed for a given guest species moiety. This indicates that
compounds (and of the pure radical molecular ions) are sig- the molecular spin distribution is not much different from
nificantly affected by paramagnetic shift and broadening that observed in the pure radical cation salt, and is also not
effects, reflecting the spin density distributions in the mole- much influenced by the guest species orientation. The biggest
cule as well as intermolecular interaction effects. Large para- chemical shift discrepancy is observed for the N-methyl group
magnetic shifts have been observed for paramagnetic (13.2 ppm in the intercalate vs. 15.3 ppm in the pure radical
materials of great chemical diversity including organic cation iodide), but this effect could also be attributed to a
1
9,20
21
radicals,
biological solids
and doped inorganic chemical shift contribution arising from the interaction of the
2
2,23
materials.
Deconvolution results, including the corres- positively charged nitrogen atom with the anionic Somasif
ponding peak assignments, are summarized in Table 4. Note slabs. While paramagnetic and chemical shift contributions
that the spectra of the pure radical ion salts appear to be sub- can be separated in principle by variable temperature
1
9–21
stantially better resolved than those of the intercalation com- measurements,
the spectroscopic resolution observed in
pounds. This effect can be explained on the basis of the the present intercalation compounds is not sufficient to
motional narrowing effects evident in the EPR spectra. Very provide the necessary measurement precision. Similar con-
fast spin state fluctuations such as those present in the pure clusions were previously reached for a number of radical inter-
4
radical ion salts produce an average electronic environment calation compounds in saponites. For the DEMTEP
for the proton nuclei, reflecting the local spin density distri- intercalate (Fig. 12b), the biggest difference with respect to the
bution in the molecule, averaged according to the electron DEMTEP iodide salt is the absence of the low-frequency
spin state populations governed by the Boltzmann distri- shifted signals attributed to the β-methyl groups, except at the
bution. The spectra of the intercalation compounds are signifi- highest intercalation level (46% CEC) where an extremely
cantly less-well resolved and reveal substantial line broadening broad component may be perceived. Again, we attribute the
effects, particularly in the m-MPYNN intercalate at low load- disappearance of these signals to a slowing down of the elec-
ings. For this series of intercalation compounds, there appears tronic spin state fluctuations evidenced from the EPR spectra,
to be a clear correlation between line broadening effects and because the intermolecular electron–electron interactions are
guest species orientation, suggesting that intermolecular weakening as the loading level is decreased. Unfortunately a
radical–radical interactions are weaker in the transversal inter- detailed analysis of the peak positions is not possible in this
calates. The peak positions (Fig. 12a and Table 4) reflect the case, because of strong overlap with the background signal
spin density distributions within the radical cation from the host structure.
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 8585–8596 | 8593

View Article Online
Paper
Dalton Transactions
Scheme 2 Catalyzed oxidation of benzyl alcohol studied in the present
manuscript.
host material is unreactive under comparable experimental
conditions (<26% conversion), and Som(m-MPYNN) inter-
calates were found to be unsuitable as they resulted in
decomposition reactions.
Fig. 13 shows the conversion during the first 20 subsequent
catalytic reactions using both hybrid materials at a catalyst
concentration of 4 mol% with respect to the substrate. Within
experimental error, the performances of both catalysts are
comparable up to 20 cycles. Experiments with lower catalyst-to-
substrate ratios gave similar results; however, catalytic activity
tends to cease more quickly (data not shown). With higher
ratios, up to 40 cycles could be reached with the VIOTEP
system (data not shown). The catalytic activity of spent catalytic
materials after 20–40 reaction cycles can be restored by re-
exposing them to a fresh aqueous solution of the radical
cation salt. As shown in Fig. 14, this process can be monitored
by EPR spectroscopy. While the spectra of the original and re-
loaded catalysts are broadened by strong dipolar and exchange
interactions, reflecting the considerably enhanced magnetic
dipole–dipole interactions in these higher-concentrated
materials, the spectra observed in the spent catalysts, where
the radical species are rather dilute, are typical of isolated nitr-
oxide radicals, allowing the resolution of anisotropic g-values
1
24
Fig. 12 H MAS-NMR spectra of (a) the Som(m-MPYNN) series, (b) the Som-
DEMTEP) series. Each plot includes reference spectra of the pure host and the
and hyperfine coupling constants. From the spectrum shown
in Fig. 14b for Som(VIOTEP), the following parameters can be
(
pure radical cation salts. Spinning sidebands are marked with asterisks.
1
8
determined using Matlab Easy Spin library:
yy = 2.005302, gzz = 2.001715 and for the A-tensor: Axx
2.65204 mT, Ayy = 26.1093 mT, Azz = 96.63939 mT. Clear differ-
ences between leached and re-loaded catalysts are also seen in
gxx = 2.006681,
g
1
=
Catalytic activity
2
3
1
The hybrid materials Som(DEMTEP) and Som(VIOTEP) can be the Na NMR spectra (see Fig. 9) and the H MAS-NMR
used as recyclable catalysts for the oxidation reaction of benzyl spectra (data not shown) in agreement with previous studies of
6
alcohol to benzaldehyde (Scheme 2). In contrast, the empty TEMPO-loaded saponites.
1
Table 4 H MAS-NMR fits and peak assignments (in ppm ± 0.5 ppm unless specified otherwise) obtained for the spectra of the pure radical ion salt and representative
intercalation compounds
Sample
Aromatic
N-Methyl
β-Methyl
m-MPYNN
Som(m-MPYNN) 71% CEC
50.5; 44.9 (2); −4.5
47.1 (2); 38.8; −2.5
15.3
13.2
−8.1; −12.8 (2); −14.2
−11.1; −12.7
N–R-group
Ring-CH
2
β-Methyl
DEMTEP
Som(DEMTEP) 46% CEC
3.6; 2.9; 2.0; 1.4
5.0; 2.1
−8.9; −15.2
—
−25.2; −27.5
−27 ± 2
Viologen-group
N–R-group
Ring-CH
2
β-Methyl
VIOTEP
10.2; 9.2
9.1; 8.6
9.1; 7.9
4.5; 3.5; 2.0; 0.8
4.8; 3.1; 1.7; 0.9
4.5; 3.6; 1.1
−16.3
—
—
−27.5
—
—
Som(VIOTEP) 56% CEC
Som(VIOTEP) after 20 cycles ∼10% CEC
8
594 | Dalton Trans., 2013, 42, 8585–8596
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
Fig. 13 Catalysis conversion degree against the cycle number for the two hybrid samples (a) Som(DEMTEP) 46ꢀ CEC and (b) Som(VIOTEP) 56ꢀ CEC.
Fig. 14 (a) EPR spectra of (from top to bottom, a1 to a5): (a1) DEMTEP iodide, (a2) Som(DEMTEP) 33ꢀ CEC, before catalytic use, (a3) sample (a2) following 20 cata-
lytic runs (catalyst level 4 molꢀ of the amount of substrate samples), (a4) sample (a2) following 40 catalytic runs (catalyst level 20 molꢀ of the amount of substrate
samples) and (a5) sample (a4) following re-loading by exposure to 10 mL of a 16 mM solution of DEMTEP iodide for 4 h. (b) EPR spectra of (from top to bottom b1
to b4): VIOTEP iodide chloride, (b2) Som(VIOTEP), 56ꢀ CEC, before catalytic use, (b3) sample (b2) following 20 catalytic runs (catalyst level 20 molꢀ of the amount
of substrate samples) and (b4) sample (b3) following re-loading by exposure to 15 mL of a 16 mM solution of DEMTEP iodide for 4 h.
ordering is observed, consistent with only weak intermolecular
Conclusions
interactions among the electron spins. For the oxidation of
In summary, a series of new inorganic–organic hybrid benzyl alcohol to benzaldehyde, a typical model reaction, the
materials based on a layered synthetic fluoromica host DEMTEP and VIOTEP intercalates reveal high reactivity during
material with three different stable radicals has been prepared the first twenty cycles.
via an ion exchange reaction and has subsequently been
2
3
characterized. Na MAS-NMR spectroscopy is well-suited to
+
monitor the depletion of the Na ions in the interlayer space
caused by the ion exchange process. The expansion of the host
Acknowledgements
matrix layer distance can be correlated with the molecular Z.Z. thanks for a personal fellowship of the International
dimensions of the radical cations, suggesting various different Research Training Group (IRTG) “Complex Functional Systems
orientations of these species in the interlayer space. While in Chemistry”, which is a programme from the Deutsche
2
9
19
solid state Si and F MAS-NMR spectra indicate that radical Forschungsgemeinschaft (DFG). Additional funding by SFB
intercalation does not affect the local environments of the 858 “Cooperative Systems” is also acknowledged. We thank
atoms in the layered slabs of the host structure, their spin Dr. Alice Cattaneo (University of Milan, Italy) for providing the
relaxation times probe the proximity of the spin carrying moi- host material Somasif.
eties and are therefore sensitive to molecular orientation
effects. Systematic compositional trends in the line widths of
1
the EPR spectra and the solid-state H MAS-NMR spectra indi-
References
cate that the intermolecular electron spin exchange rates are
concentration dependent. In addition, trends in the resonance
shifts suggest that the spin density distributions are modified
by intermolecular electron–electron interactions among neigh-
boring radical cation species. However, no bulk magnetic
1 Y. Hosokoshi, M. Tamura, H. Sawa, R. Kato and
M. Kinoshita, J. Mater. Chem., 1995, 5, 41.
2 M. Deumal, J. Cirujheda, J. Veciana and J. J. Novoa,
Chem.–Eur. J., 1999, 5, 1631.
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 8585–8596 | 8595

View Article Online
Paper
Dalton Transactions
3
4
5
6
7
8
W. Fujita and K. Awaga, J. Chem. Soc., Chem. Commun., 14 C. Wagner and A. Studer, Eur. J. Org. Chem., 2010, 5782.
1
995, 739.
15 Y. V. Kokhanov, É. G. Rozantsev, L. N. Nikolenko and
L. A. Maksimova, Chem. Heterocycl. Compd., 1971, 7, 1421.
16 D. Massiot, F. Fayon, M. Capron, I. King, S. Le Calvé,
B. Alonso, J.-O. Durand, B. Bujoli, Z. Gan and G. Hoatson,
Magn. Reson. Chem., 2002, 40, 70.
W. L. Hemme, W. Fujita, K. Awaga and H. Eckert, J. Solid
State Chem., 2009, 182, 3330.
W. L. Hemme, W. Fujita, K. Awaga and H. Eckert, Dalton
Trans., 2009, 7995.
C. Röben, A. Studer, W. L. Hemme and H. Eckert, Synlett, 17 K. Awaga, T. Inabe, Y. Maruyama, T. Nakamura and
010, 1110. M. Matsumoto, Synth. Met., 1993, 55–57, 3311.
W. L. Hemme, W. Fujita, K. Awaga and H. Eckert, Solid 18 S. Stoll and A. Schweiger, J. Magn. Reson., 2006, 178, 42.
2
State Nucl. Magn. Reson., 2011, 39, 106.
19 G. Maruta, S. Takeda, A. Yamaguchi, T. Okuno and
A. S. Cattaneo, S. Bracco, A. Comotti, M. Galimberti,
K. Awaga, Polyhedron, 2003, 22, 1989.
P. Sozzani and H. Eckert, J. Phys. Chem. C, 2011, 115, 20 H. Heise, F. H. Köhler, F. Mota, J. J. Novoa and J. Veciana,
2517. J. Am. Chem. Soc., 1999, 121, 9659.
H. He, J. Duchet, J. Galy and J. F. Gerard, J. Colloid Interface 21 N. P. Wickramasinghe, M. A. Shaibat, C. R. Jones,
1
9
Sci., 2005, 288, 171.
0 P. Rey and A. Rassat, Nouv. J. Chim., 1983, 7, 531.
L. B. Casabianca, A. C. de Dios, J. S. Harwood and Y. Ishii,
J. Chem. Phys., 2008, 128, 052210, and references therein.
1
1
1 E. F. Ullmann, J. H. Osiecki, D. G. B. Boocok and R. Darcy, 22 C. P. Grey, M. E. Smith, A. K. Cheetham, C. M. Dobson and
J. Am. Chem. Soc., 1972, 94, 7049. R. Dupree, J. Am. Chem. Soc., 1990, 112, 4670.
2 K. Awaga, T. Inabe, U. Nagashima, T. Nakamura, 23 H. Trill, H. Eckert and V. I. Srdanov, Phys. Rev. B: Condens.
M. Matsumoto and Y. Maruyama, Chem. Lett., 1991, 1777. Matter, 2005, 71, 014412.
3 J. Zakrzewski, J. Hupko and K. Kryczka, Monatsh. Chem., 24 M. Tabak, A. Alonso and O. R. Nascimento, J. Chem. Phys.,
003, 134, 843. 1983, 79, 1176.
1
1
2
8
596 | Dalton Trans., 2013, 42, 8585–8596
This journal is © The Royal Society of Chemistry 2013