α-RuCl3/polymer nanocomposites: The first group of intercalative nanocomposites with transition metal halides

Article abstract of DOI:10.1021/ja9944610

Different types of polymers can be intercalated into α-RuCl₃ with different synthetic methodologies. Polyaniline/α-RuCl₃ nanocomposite was prepared by the in situ redox intercalative polymerization method, in which α-RuCl₃ was exposed to an aniline/acetonitrile solution in open air. Water-soluble polymers such as poly(ethylene oxide), poly(vinyl pyrrolidone), and polyethylenimine were intercalated by an encapsulative precipitation method using monolayer suspensions of α-RuCl₃. A modification of this method led to insertion of polypyrrole. Monolayer suspensions of α-RuCl₃ can be prepared from Li_xRuCl₃ (x ～ 0.2). The latter is produced by the reaction of α-RuCl₃ with 0.2 equiv of LiBH₄. The polymer insertion is topotactic and does not cause structural changes to the host. The metal chloride layers in these materials possess mixed valency. The reduction and polymer intercalation of α-RuCl₃ alters the intralayer and interlayer Ru³⁺ (low spin d⁵) magnetic coupling, so that interesting magnetic properties appear in the nanocomposites. In addition, the reduction brings in free hopping electrons to the RuCl₃ layers and the polymer intercalation builds up new electronic or ionic conducting channels in the galleries, so that the charge transport properties are changed dramatically. For example, Li_xRuCl₃ shows an electrical conductivity 3 orders of magnitude higher than pristine α-RuCl₃ at room temperature and Li_x(PEO)_yRuCl₃ has an ion conductivity comparable with the best (lithium salt)-polymer electrolytes. For a comprehensive understanding of the structure of the representative nanocomposite Li_x(PEO)_yRuCl₃, the arrangement of polymer chains inside the galleries was explored with analysis of its one-dimensional (00l) X-ray diffraction pattern. Calculated electron density maps along the stacking c-axis lead to a structural model that fills each gallery with two layers of polymer chains exhibiting a conformation found in type-II PEO-HgCl₂. The most consistent PEO arrangement in the gallery generates oxygen-rich channels in the middle of the gallery in which the Li ions can reside. The new nanocomposites were characterized with thermogravimetric analysis, infrared spectroscopy, powder X-ray diffraction, magnetic measurements, as well as electrical and ionic conductivity and thermopower measurements.

Full text of DOI:10.1021/ja9944610

J. Am. Chem. Soc. 2000, 122, 6629-6640
6629
R-RuCl3/Polymer Nanocomposites: The First Group of Intercalative
Nanocomposites with Transition Metal Halides
Lei Wang,† Melissa Rocci-Lane,‡ Paul Brazis,‡ Carl R. Kannewurf,‡ Young-Il Kim,§
Woo Lee,§ Jin-Ho Choy,§ and Mercouri G Kanatzidis*,†
Contribution from the Department of Chemistry and Center for Fundamental Materials Research,
Michigan State UniVersity, East Lansing, Michigan 48824, Department of Electrical and Computer
Engineering, Northwestern UniVersity, EVanston, Illinois 60208, and Department of Chemistry and
National Research Laboratory, College of Natural Sciences, Seoul National UniVersity,
Seoul 151-742, Korea
ReceiVed December 21, 1999
Abstract: Different types of polymers can be intercalated into R-RuCl3 with different synthetic methodologies.
Polyaniline/R-RuCl3 nanocomposite was prepared by the in situ redox intercalative polymerization method, in
which R-RuCl3 was exposed to an aniline/acetonitrile solution in open air. Water-soluble polymers such as
poly(ethylene oxide), poly(vinyl pyrrolidone), and polyethylenimine were intercalated by an encapsulative
precipitation method using monolayer suspensions of R-RuCl3. A modification of this method led to insertion
of polypyrrole. Monolayer suspensions of R-RuCl3 can be prepared from LixRuCl3 (x ∼ 0.2). The latter is
produced by the reaction of R-RuCl3 with 0.2 equiv of LiBH4. The polymer insertion is topotactic and does
not cause structural changes to the host. The metal chloride layers in these materials possess mixed valency.
The reduction and polymer intercalation of R-RuCl3 alters the intralayer and interlayer Ru3+ (low spin d5)
magnetic coupling, so that interesting magnetic properties appear in the nanocomposites. In addition, the
reduction brings in free hopping electrons to the RuCl3 layers and the polymer intercalation builds up new
electronic or ionic conducting channels in the galleries, so that the charge transport properties are changed
dramatically. For example, LixRuCl3 shows an electrical conductivity 3 orders of magnitude higher than pristine
R-RuCl3 at room temperature and Lix(PEO)yRuCl3 has an ion conductivity comparable with the best (lithium
salt)-polymer electrolytes. For a comprehensive understanding of the structure of the representative
nanocomposite Lix(PEO)yRuCl3, the arrangement of polymer chains inside the galleries was explored with
analysis of its one-dimensional (00l) X-ray diffraction pattern. Calculated electron density maps along the
stacking c-axis lead to a structural model that fills each gallery with two layers of polymer chains exhibiting
a conformation found in type-II PEO-HgCl2. The most consistent PEO arrangement in the gallery generates
oxygen-rich channels in the middle of the gallery in which the Li ions can reside. The new nanocomposites
were characterized with thermogravimetric analysis, infrared spectroscopy, powder X-ray diffraction, magnetic
measurements, as well as electrical and ionic conductivity and thermopower measurements.
Introduction
strength, which make them more suitable for applications. As
a result of recent efforts, a large variety of intercalative
nanocomposites has been prepared using hosts from most major
classes of layered inorganic compounds, i.e., clays, layered
transition metal chalcogenides, metal oxides, metal phosphates,
and metal thiophosphates.4
Layered transition metal halides, in contrast to layered metal
chalcogenides,5 oxides,6 and thiophosphates,7 have not received
much attention in the study of intercalation, because most of
them are not very stable, even under mild conditions, for soft
chemical reactions. Hydrolysis, dissolution, and other decom-
position reactions are often common properties of this group
The combination of two extremely different components, at
the molecular level, provides an avenue to designing new
nanocomposite hybrid materials as well as the ability to
modulate the properties of one or more of the components.1 In
some circumstances, it is also a unique way to generate materials
which have special properties that are unknown in the individual
constituents.2 Intercalation compounds with polymers as the
guest species are an important class of nanocomposite materials.3
Intercalative polymer hybrids have advantages over their small
molecular analogues in compositional stability and mechanical
† Michigan State University.
‡ Northwestern University.
(4) (a) Giannelis, E. P. AdV. Mater. 1996, 8, 29. (b) Ruiz-Hitzky, E.;
Aranda, P.; Casal, B.; Galva´n, J. C. AdV. Mater. 1995, 7, 180.
(5) (a) Hsu, C.-H.; Labes, M. M.; Breslin, J. T.; Edmiston, D. J.; Winter,
J. J.; Leupold, H. A.; Rothwarf, F. Nature, Phys. Sci., 1973, 246 (155),
122. (b) Kanatzidis, M. G.; Bissessur, R.; DeGroot, D. C.; Schindler, J. L.;
Kannewurf, C. R. Chem. Mater. 1993, 5, 595.
(6) (a) Liu, Y.-J.; DeGroot, D. C.; Schindler, J. L.; Kannewurf, C. R.;
Kanatzidis, M. G. AdV. Mater. 1993, 5, 369. (b) Nazar, L. F.; Zhang, Z.;
Zinkweg, D. J. Am. Chem. Soc. 1992, 114, 6239.
§ Seoul National University.
(1) Leroux, F.; Koene, B. E.; Nazar, L. F. J. Electrochem. Soc. 1996,
143 (9), L181. (b) Leroux, F.; Goward, G.; Power, W. P.; Nazar, L. F. J.
Electrochem. Soc. 1997, 144, 3886.
(2) Wang, Y.; Herron, N. Science 1996, 273 (5275), 632.
(3) (a) O’Hare, D. In Inorganic Materials; Bruce, D. W., O’Hare, D.,
Eds.; John Wiley & Sons Ltd: New York, 1992; p 164. (b) Ruiz-Hitzky,
E. AdV. Mater. 1993, 5, 334. (c) Alberti, G.; Bein, T., Eds. ComprehensiVe
Supramolecular Chemistry; Elsevier Science Ltd: New York, 1996; Vol.
7. (d) Giannelis, E. P. AdV. Mater. 1996, 8, 29.
(7) Lagadic, I.; Le´austic, A.; Cle´ment, R. J. Chem. Soc., Chem. Commun.
1992, 1396.
10.1021/ja9944610 CCC: $19.00 © 2000 American Chemical Society
Published on Web 06/29/2000

6630 J. Am. Chem. Soc., Vol. 122, No. 28, 2000
Wang et al.
such as polypyrrole (PPY). Here we report in detail the insertion
or encapsulation of polymers into R-RuCl3 to produce a new
group of lamellar nanocomposites with diverse physicochemical
properties. Two different types of polymers were examined,
conjugated and conventional nonconjugated systems. The former
include polyaniline (PANI) and polypyrrole (PPY) whereas the
latter involved poly(ethylene oxide) (PEO), poly(ethylene-imine)
(PEI) and poly(vinylpyrrolidone) (PVP). The Li/PEO/ RuCl3
system exhibits mixed electronic/ionic conductivity with a very
substantial ionic charge transport component.
Experimental Section
1. Reagents. LiBH4 (95%), PEO (100 000), PVP (10 000), and PEI
(25 000) were purchased from Aldrich Chemical Co., Inc. After the
polymers were dissolved, the polymer solutions were filtered to remove
insoluble residues. Aniline and pyrrole were purchased from Aldrich
and Mallinckrodt Inc., respectively, and were distilled before use.
2. Syntheses and Reactions.(a) Preparation of R-RuCl3. R-RuCl3
was synthesized as described in the literature.12 The setup for the
reaction (the gas supply, outlet tubing and the quartz reaction tube)
was first purged with nitrogen overnight. A sample of 10 g of Ru metal
powder spread in an alumina boat was reacted with flowing mixed Cl2
and CO gases at 800 °C for 8 h. The ratio of Cl2 to CO gas was adjusted
to about 1:2 or 1:3 and the flow was set at approximately 60-120
mL/min. The gas began to flow before heating and was turned off after
the reaction tube cooled to about 50 °C. Very thin platelike crystals of
R-RuCl3 were collected from the cooler end of the tube (down stream)
in a yield that exceeded 40%.
Figure 1. The layered structure of R-RuCl3. The coordination
environment of the Ru atoms is octahedral. (A) View down the c-axis.
(B) View parallel to the layers.
of materials. R-RuCl3 is a notable exception and is very stable
under these conditions.
(b) Synthesis of (PANI)xRuCl3. A sample of 0.1 g of R-RuCl3 in
10 mL of 4% aniline/acetonitrile solution was stirred in open air for 1
week. The product was washed with copious acetonitrile and dried in
a vacuum. Elemental analysis, done by Quantitative Technologies Inc.,
gave 15.33% C, 1.43% H, 3.11% N, which suggests the formula
(PANI)0.57(H2O)0.50RuCl3 (calculated 15.3% C, 1.43% H, 2.97% N).
The above formula contains 22 wt % water as determined from
thermogravimetric analysis (TGA) measurements.13 The formula
(PANI)xRuCl3 can vary from x ∼ 0.55 to 0.7 depending on preparation
conditions.
(c) Synthesis of LixRuCl3 and Preparation of RuCl3 Monolayer
Suspensions. In a typical reaction, 18 mmol of R-RuCl3 was mixed
with 3.6 mmol of LiBH4 in 40 mL of anhydrous ether under N2
atmosphere for 3 days. The product was collected by filtration, washed
with anhydrous ether, and pumped to dryness. It was stored in a
glovebox under N2 atmosphere. The product is designated as LixRuCl3,
where x ≈ 0.2. An aqueous [RuCl3]x- monolayer suspension was
prepared simply by stirring LixRuCl3 in water for 30 min. The value
of x was determined with ICP analysis. Compositions with x > 0.2 are
possible but they were not explored in this work.
(d) Synthesis of Lix(PEO)yRuCl3, Lix(PEI)yRuCl3, and Lix-
(PVP)yRuCl3. In the preparation of PEO and PVP nanocomposites, a
sample of 40 mL of 0.5 wt % LixRuCl3 monolayer suspension (0.96
mmol of LixRuCl3) was mixed with 40 mL of polymer solution
containing 3 mmol repeat-units of polymer. The mixture was stirred
for 2 days in ambient atmosphere. The nanocomposites of PEO and
PVP are still colloidal, and have to be collected from very concentrated
solutions. Therefore, the reaction mixture was condensed to less than
20 mL under vacuum before it was poured into an organic solvent
(e.g. ether) to precipitate the product. Acetonitrile and 2-propanol were
used as solvents for PEO and PVP, respectively. In the preparation of
the PEI nanocomposite, the sample of 40 mL of 0.5 wt % LixRuCl3
monolayer suspension was mixed with 40 mL of PEI solution containing
7 mmol repeat-units of polymer. The PEI nanocomposite precipitated
R-RuCl3 has a lamellar structure that can be regarded as defect
CdI2 type, see Figure 1. Intercalation chemistry of R-RuCl3 has
been reported with the insertion of simple cations and neutral
polar molecules.8 Cations can be intercalated reductively or
through ion-exchange, while neutral polar molecules can be
incorporated through solvent exchange. These properties are
similar to those of layered chalcogenides and oxides such as
TiS2, 2H-TaS2, and MoO3, suggesting the possibility of inter-
calating polymers in this material. In fact, as we describe here,
R-RuCl3 behaves as an excellent polymer-intercalation host. It
exhibits affinity for various polymers and is compatible with
several of the existing polymer intercalation methods. For
example, R-RuCl3/polyaniline (R-RuCl3/PANI) nanocomposites
were synthesized with in situ redox intercalative polymerization,9
while soluble polymers such as poly(ethylene oxide) (PEO),
poly(vinyl pyrrolidone) (PVP), and poly(ethylene-imine) (PEI)
are intercalated through encapsulative coprecipitation10 of
polymers and R-RuCl3 from solutions of exfoliated R-RuCl3.
A modification of the second method, in situ polymerization
coupled with encapsulative precipitation11 from suspensions of
exfoliated R-RuCl3, can be used to insert intractable polymers
(8) (a) Scho¨llhorn, R.; Steffen, R.; Wagner, K. Angew. Chem., Int. Ed.
Engl. 1983, 22, 555. (b) Steffen, R.; Scho¨llhorn, R. Solid State Ionics 1986,
22, 31. (c) Nonte, W.; Lobert, M.; Mu¨ller-Warmuth, W.; Scho¨llhorn, R.
Synth. Met. 1989, 34, 665.
(9) Wang, L.; Brazis, P.; Rocci, M.; Kannewurf, C. R.; Kanatzidis, M.
G. Chem. Mater. 1998, 10, 3298.
(10) The method of encapsulative precipitation from solutions of
exfoliated lamellar solids has be applied to MoS2, MoO3, TaS2, and
NbSe2: (a) Bissessur, R.; Kanatzidis, M. G.; Schindler, J. L.; Kannewurf,
C. R. J. Chem. Soc., Chem. Commun. 1993, 1582. (b) Wang, L.; Schindler,
J.; Kannewurf, C. R.; Kanatzidis, M. G. J. Mater. Chem. 1997, 7, 1277. (c)
Wang, L.; Schindler, J.; Kannewurf, C. R.; Kanatzidis, M. G. Manuscript
in preparation. (d) Tsai, H.-L.; Schindler, J. L.; Kannewurf, C. R.;
Kanatzidis, M. G. Chem. Mater. 1997, 9, 875.
(11) The method of in situ polymerization coupled with encapsulative
precipitation has been applied to MoS2, MoO3, and WS2: (a) Wang, L.;
Schindler, J. L.; Thomas, J. A.; Kannewurf, C. R.; Kanatzidis, M. G. Chem.
Mater. 1995, 7, 1753. (b) Kerr, T. A.; Wu, H.; Nazar, L. F. Chem. Mater.
1996, 8, 2005. (c) Wang, L. Intercalated Polymer-Layered Inorganic
Nanocomposites, Ph.D. Dissertation, 1999, Department of Chemistry,
Michigan State University.
(12) Brauer, G. Handbook of PreparatiVe Inorganic Chemistry; Academic
Press Inc.: New York, 1965; Vol. 2, p 1597.
(13) TGA indicated a total loss of 48.2% in oxygen flow at temperatures
up to 650 °C. By comparison, R-RuCl3 under the same conditions loses
33.07% of its weight and converts to RuO2. The amount of organics and
water inside the nanocomposite was determined assuming the final product
is RuO2. The amount of intercalates in Lix(PEO)yRuCl3, Lix(PVP)yRuCl3,
and Lix(PEI)yRuCl3 was determined similarly.

R-RuCl3/Polymer Nanocomposites
J. Am. Chem. Soc., Vol. 122, No. 28, 2000 6631
out from water immediately upon mixing. It was collected with
centrifugation after the suspension was stirred for 2 days. The
nanocomposites were dried under vacuum. The compositions of the
atmosphere to prevent absorption of moisture, which causes peak
broadening and drifting. In the measurements, slits for different beam
width, as well as different data collection times, were used for different
2θ ranges, to achieve a compromise between peak broadening, peak
intensity, and experiment time. Specifically, 0.5° slits and a data
collection time of 12 s per step were chosen for measurements in the
range from 2° to about 29°; 1.0° slits and 60 s per step between 16°
and 71°; 2.0° slits and 60 s per step between 57° and 83.5°; and 4.0°
slits and 90 s per step from 72° to 136°. The step width was kept the
same (0.1°) for the entire 2θ range.
three nanocomposites were Lix(PEO)1.5RuCl3 (x ∼ 0.2), Lix(PVP)2.2
-
RuCl3 (x ∼ 0.2), and Lix(PEI)4.6RuCl3 according to TGA measurements
under oxygen flow.13 These compositions gave satisfactory C, H, N
elemental analyses. The amount of Li was determined with inductively
coupled plasma spectroscopy (ICP).
(e) Synthesis of (PPY)xRuCl3. A sample of 0.45 g (6.7 mmol) of
pyrrole was dissolved in 40 mL of water and mixed with 40 mL of
aqueous suspension which contained 0.20 g (0.96 mmol) of LixRuCl3.
The mixture was cooled in an ice bath before the dropwise addition of
a cold 10 mL aqueous solution of 0.11 g (0.67 mmol) of FeCl3. The
mixture was stirred in a closed, ice-cooled flask for 24 h. The product
was collected by centrifugation, washed with copious water, and dried
in air and under vacuum. Elemental analysis gave 13.88% C, 1.32%
H, and 3.95% N. This corresponds to (PPY)0.77(H2O)0.60RuCl3 (calcu-
lated 13.78% C, 1.31% H, and 4.01% N). The product did not contain
Li. TGA measurements (in air) showed that the material lost 50.2%
(theoretical 50.4%) of its weight in oxygen at temperatures higher than
450 °C. X-ray powder patterns of the final residue indicated pure RuO2.
3. Physicochemical Measurements. (a) Instrumentation. X-ray
diffraction (XRD) powder patterns were obtained on a Rigaku Ru-
200B X-ray diffractometer, at 45 kV and 100 mA, with a scintillation
counter detector and a graphite monochromator to produce a Cu KR
beam (λ ) 1.54184 Å). A continuous scanning mode with a speed of
1 deg/min in 2θ and an increment of 0.05° was chosen for general
purpose spectra. For 1-D ED calculations, we obtained XRD data from
highly oriented samples and a stepwise scanning mode with 0.1° per
step.
Infrared (IR) spectra were collected on solid samples as pressed KBr
pellets using a Nicolet IR/42 spectrometer with a 2.0 cm-1 resolution.
A typical 64 scans were applied for each sample. Thermal gravimetric
analysis (TGA) was performed with a Shimadzu TGA-50 under an
oxygen flow of 46 mL/min. The heating rate was 10 deg/min.
Scanning electron microscopy (SEM) and energy-dispersive X-ray
microanalysis (EDS) were done with a JEOL-JSM 35 CF microscope
at an accelerating voltage of 15 and 20 kV, respectively. Samples were
mounted on a sample stab with conductive tape. Electron diffraction
experiments were performed with a JEOL-100CX transmission electron
microscope (TEM) operating at 120 kV.
A full range XRD pattern (2° e 2θ e 136°) was obtained by merging
the data from the different 2θ ranges and scaling them based on the
overlapped regions. The overlapped data regions had at least two peaks
in common. The full data set was put into the XRD analysis program
PEAKOC16 to calculate the integrated peak area of each 00l reflection.
In pattern analysis, 00l peaks were fit with the split-pseudo-Voigt
function with a linear background subtraction for each peak. The
integrated peak area of the 00l peaks was used as their intensity. The
methodology used has been described in detail elsewhere.17
A
FORTRAN program was written to calculate the 1-D ED map. The
signs (phases) of the structure factors were assigned based on the signs
of the corresponding structure factors of the RuCl3 framework alone.
This reasonably assumes that the scattering contribution from the
intercalated polymer is relatively small compared to that of the RuCl3
component. After the model for the structure was established, the sign
of the Fl was checked by recalculating Flcalc from the scattering of all
atoms including those of the polymer.
(c) Impedance Spectroscopy Measurements. Impedance spectros-
copy measurements were carried out on a Hewlett-Packard HP4192A
low-frequency impedance analyzer to probe the impedance |Z*| and
the phase angle φ. Samples were pressed into disks under 50 kg/m2
pressure and assembled between two stainless steel electrodes in an
air-proof cell under a nitrogen atmosphere. The range of frequency ω
used in the measurements was 5-13 MHz. The cell was heated in a
silicon oil bath. Interpretation of the impedance spectra was made by
fitting the data with equivalent circuits. Nonlinear least-squares fitting,
which minimizes the sum ∑ω|Z*|-1{(Z′exp - Z′sim)2 + (Z′′exp - Z′′sim)2},
was used to determine the values of the electrical elements in the
equivalent circuit.18
Results and Discussion
Magnetic susceptibility measurements were done with a Quantum
Design MPMS2 SQUID magnetometer in the temperature region of
4-300 K. Samples were sealed in low-density polyethylene (LDPE)
bags under a nitrogen atmosphere. The magnetic moments of the bags
were acquired and subtracted from the data. To obtain the magnetic
susceptibility ømolar, the diamagnetic susceptibility ød and øTIP were
subtracted from the total susceptibility. The ød was derived by adding
up the diamagnetic susceptibility of each component, which was
obtained from the literature.14
In this section, the synthesis and characterization of several
families of hybrid polymer/RuCl3 lamellar intercalative nano-
composites are described. Next the structural arrangement of
the PEO member is explored and compared with those of other
PEO systems. Finally, it is shown that the conjugated polymers
are crucial to achieving high electronic conductivities whereas
in the Li/PEO/RuCl3 case the ionic transport is greatly enhanced,
compared to that of Li/RuCl3, and is as good as some of the
best Li/PEO electrolyte salts.
1. Preparation of (PANI)xRuCl3 by in Situ Redox Inter-
calative Polymerization. The in situ redox intercalative po-
lymerization reaction is the most direct method to intercalate
conductive polymers in open-framework hosts. Its topotactic
character least disturbs the crystalline structure of the host. In
the case of FeOCl/polyaniline, even single crystals of the
nanocomposite could be obtained.19 This type of reaction
requires a strongly oxidizing host to provide a driving force to
pull electrons from the monomers and oxidize them into
polymers. In addition, the host should be able to distribute
Electrical conductivity data were obtained with a computer-
automated system described elsewhere.15
(b) X-ray Diffraction and One-Dimensional Electron Density
Maps. One-dimensional electron density (1-D ED) maps were calcu-
lated from XRD data collected from highly oriented film samples in a
stepwise scanning mode. These film samples were made by casting
aqueous nanocomposite solutions so that the basal planes of the RuCl3
layers restacked parallel to the substrate. Several layers of a film were
loaded in the sample plate so as to obtain maximum diffraction intensity.
Immediately before the X-ray experiments, the samples were pumped
at 75 °C for 3 days to remove excess water from the nanocomposite.
This procedure ensures that the sample has only one phase with a single
basal spacing. The X-ray patterns were collected under a nitrogen
(16) PEAKOC is an XRD powder pattern analysis computer program
provided by Inel Inc. (Mail Adress in U.S.A.: P.O. Box 147, Stratham,
NH 03885.)
(17) Leung, S. Y.; Dresselhaus, M. S.; Underhill, C.; Krapchev, T.;
Dresselhaus, G.; Wuensch, B. J. Phys. ReV. B-Cond. Matter 1981, 24 (6),
3505-3518
(18) (a) Choy, J. H.; Park, N. G.; Kim, Y. I.; Hwang, S. H.; Lee, J. S.;
Yoo, H. I. J. Phys. Chem. 1995, 99, 7845. (b) Lee, J. S.; Yoo, H. I. Solid
State Ionics 1994, 68, 139.
(14) (a) Selwood, P. W. Magnetochemistry, 2nd ed.; Interscience
Publishers: New York, 1956; p 78. (b) Drago, R. S. Physical Methods for
Chemists; 2nd ed.; Saunders College Publishing: Philadelphia/San Diego/
New York, 1992; Chapter 11. (c) Epstein, A. J.; Ginder, J. M.; Richter, A.
F.; MacDiarmid, A. G. In Conducting Polymers; Alca´cer, L., Ed.; D. Reidel
Publishing Company: Dordrecht, The Netherlands, 1986; p 121.
(15) Lyding, J. W.; Marcy, H. O.; Mark T. J.; Kannewurf, C. R. IEEE.
Trans. 1nstrum. Meas. 1988, 37, 76.

6632 J. Am. Chem. Soc., Vol. 122, No. 28, 2000
Wang et al.
character. The d6 electron configuration maximizes the Ligand
Field Stabilization Energy (LFSE) and this probably acts as a
powerful driving force for the oxidation of aniline. As has been
demonstrated in the intercalative polymerization of aniline in
FeOCl and V2O5, the presence of oxygen is key to a successful
outcome of the reaction. This was verified by control experi-
ments where in the absence of air or oxygen no intercalation
reaction occurred after 23 days. Ambient oxygen serves as an
important electron acceptor so that it relieves the pressure on
the host to accept and distribute large numbers of electrons
which could destabilize the host structure.
2. Nanocomposites of R-RuCl3 with Water Soluble Poly-
mers. To prepare nanocomposites with water soluble polymers,
the method of encapsulative precipitation from solutions of
exfoliated lamellar solids was used.6,10,22 To a stable and
concentrated aqueous LixRuCl3 monolayer suspension, water-
soluble polymers PEO, PVP, and PEI were added and subse-
quently encapsulated. The XRD reflectivity patterns of the
products show that the interlayer spacings of Lix(PEO)yRuCl3,
Lix(PVP)yRuCl3, and Lix(PEI)yRuCl3 increase by 8.5, 23.0, and
3.6 Å, respectively, see Figure 2.23 The existence of PEO, PVP,
and PEI in the nanocomposites was confirmed by IR spectros-
copy, see IR spectra C and D of Lix(PEO)yRuCl3 and Lix(PVP)y-
RuCl3 in Figure 3.
Figure 2. X-ray diffraction patterns (reflection mode) of R-RuCl3 and
nanocomposites.
3. r-RuCl3/Polypyrrole Nanocomposites. The exfoliating
property of LixRuCl3 offers the possibility to synthesize nano-
composites with insoluble polymers, by in situ polymerization-
encapsulative precipitation.11 In this method, a solution of a
suitable monomer (such as pyrrole) is mixed with an exfoliated
suspension of the inorganic host. When an initiator is added,
the polymerization causes the coprecipitation of the polymer
and host monolayers, forming a nanocomposite. This happens
because the growing polymer (in this case polypyrrole) is a
positively charged species and adheres to the negatively charged
inorganic layers mainly due to electrostatic forces. The initiator
in this case is the Fe3+ ions.
efficiently those electrons throughout the structure. Because of
the scarcity of such highly oxidizing and conducting hosts, the
reaction has been limited to FeOCl,19 V2O5,20 and VOPO4.21
Now we have discovered that R-RuCl3 is also a suitable host,
and can form intercalative nanocomposites with polyaniline.
The reaction of an aniline/CH3CN solution with R-RuCl3 in
air results in the formation of polyaniline (PANI) within the
gallery space of RuCl3. The reflection-mode powder X-ray
diffraction (XRD) patterns of the product show a 6.2 Å increase
in the separation of the RuCl3 layers, see Figure 2. This
expansion is reasonable for insertion of a monolayer of PANI
molecules, and comparable to the 5.94 Å observed in
(PANI)xFeOCl19 and 5.2 Å in (PANI)xV2O5.20 The transmission-
mode powder XRD patterns show that the hk0 reflections of
(PANI)xRuCl3 remain the same as those of R-RuCl3, indicating
that the inorganic host structure (i.e. its framework) is preserved.
This result is also supported by electron diffraction experiments,
which provide essentially identical two-dimensional hk0 dif-
fraction patterns for (PANI)xRuCl3 and R-RuCl3.
The formation of lamellar (PPY)xRuCl3 (x ∼ 0.7) is indicated
by XRD patterns which show basal spacings of ∼11.7 Å, see
Figure 2. The existence of the conductive form of polypyrrole
in the galleries is confirmed by its characteristic IR spectrum,
which exhibits peaks at 1541, 1314, 1150 1043, and 963 cm-1 24
,
see Figure 3B. The 6 Å spacing between the RuCl3 layers
corresponds to a monolayer of polypyrrole arranged with its
pyrrole rings inclined ∼40° to the RuCl3 layers. This situation
is similar to that of (PPY)xMoS2 obtained from pyrrole-rich
conditions.11b,c The packing density of polypyrrole in (PPY)x-
RuCl3, 0.77 pyrrole-unit per RuCl3, is comparable to that in
(PPY)xMoS2 prepared under similar conditions, 0.50 pyrrole-
unit per MoS2.25
The formation of polyaniline between the RuCl3 layers is
supported by IR spectroscopy. Almost all peaks in the IR spectra
of (PANI)xRuCl3 are associated with polyaniline (emeraldine
salt) and only a few with the anilinium cation (at 744 and 687
cm-1), see Figure 3A. The peaks corresponding to the polymer
are much more intense than those of anilinium. Heating the
product at 120 °C in air for 5 days removes all anilinium peaks.
The appearance of monomer peaks implies that the polyaniline
formed in the galleries does not have very high molecular
weight. This is reasonable since both the concentration and
mobility of the aniline in the gallery are limited. We speculate
the polymer MW to be ∼5000, as was found in the (PANI)x-
FeOCl system.
In the reaction to form (PPY)0.77RuCl3, only 0.7 equiv of
FeCl3 was used, which could oxidize, at maximum, ∼0.35 equiv
(22) Bissessur, R.; Schindler, J. L.; Kannewurf, C. R.; Kanatzidis, M.
G. Mol. Cryst. Liq. Cryst. 1993, 245, 249-254
(23) Electron diffraction experiments showed that the exfoliated single
RuCl3 layers have the same hexagonally symmetric hk0 diffraction pattern
as R-RuCl3. The same electron diffraction pattern was also observed in
(PANI)xRuCl3, suggesting no change in intralayer structure during intercala-
tion and thus a truly topotactic process.
In the process of intercalation, a fraction of Ru3+ centers is
reduced to Ru2+ giving a mixed-valence compound. The Ru2+
centers are very stable due to their low spin, diamagnetic d6
(24) Polypyrrole exhibits peaks at 1540, 1300, 1150, 104, and 900 cm-1
:
Kang, E. T.; Neoh, K. G.; Tan, T. C.; Ong, Y. K. J. Macromol. Sci.-
Chem. 1987, A24 (6), 631.
(19) Wu, C.-G.; DeGroot, D. C.; Marcy, H. O.; Schindler, J. L.;
Kannewurf, C. R.; Bakas, T.; Papaefthymiou, V.; Hirpo, W.; Yesinowski,
J. P.; Liu, Y.-J.; Kanatzidis, M. G. J. Am. Chem. Soc. 1995, 117, 9229.
(20) Wu, C.-G.; Marcy, H. O.; Kannewurf, C. R.; Kanatzidis, M. G. J.
Am. Chem. Soc. 1989, 111, 4139.
(25) The (PPY)xMoS2, prepared under similar conditions, has x ∼ 0.5.
Considering that the unit area of RuCl3 is 1.75 times that of MoS2, the
amount of polypyrrole in the galleries is quite comparable. (For a single
MoS2 layer one unit cell contains one Mo and two S atoms; its a-axis
dimension is 3.159 Å. By comparison, for one unit cell of two Ru and six
Cl atoms has 5.87 Å as its a-axis dimension.)
(21) Nakajima, H.; Matsubayashi, G. J. Mater. Chem. 1994, 4 (8) 1325.

R-RuCl3/Polymer Nanocomposites
J. Am. Chem. Soc., Vol. 122, No. 28, 2000 6633
Figure 3. Infrared spectra: (A) (PANI)xRuCl3, PANI, and aniline; (B) (PPY)xRuCl3 and PPY; (C) Lix(PEO)yRuCl3 and PEO; and (D) Lix(PVP)y-
RuCl3 and PVP. (The peak at 1385 cm-1 is due to an impurity in KBr.)
of pyrrole to polypyrrole. Therefore, the amount of polypyrrole
is greater than what would be expected if FeCl3 alone were the
oxidant. This discrepancy is explained by the fact that the
ambient oxygen participates in the reaction as an electron
acceptor, as discussed above. It has been proven that ambient
oxygen can oxidize and polymerize pyrrole and its oligomers
when FeCl3 is present as a catalyst.26 Ambient oxygen has been
shown to be the ultimate electron acceptor in the formation of
(PPY)xV2O5‚nH2O nanocomposites as well.27
For comparison we also tried the in situ redox intercalative
polymerization to prepare a RuCl3/PPY nanocomposite. This
procedure uses R-RuCl3 rather than an exfoliated suspension.
When R-RuCl3 was stirred in an aqueous pyrrole solution28 in
open air, intercalation occurred in two weeks to form a product
with 11.2 Å basal spacing. However, in contrast to the PANI/
RuCl3 system, IR spectroscopy showed that the polymer formed
inside the gallery was not ordinary polypyrrole but a highly
defective form. Further investigations were not performed.
4. Charge Transport Properties. The insertion of polymers
or Li causes large changes in the properties of R-RuCl3. Because
of the formation of Ru2+ centers, the resulting materials are
now mixed-valent compounds with both Ru2+ and Ru3+ in the
structure. This provides free hopping electrons between the metal
atoms and explains the relatively high electrical conductivity
of LixRuCl3 (x ∼ 0.2) at room temperature of ∼0.3 S/cm, about
3 orders of magnitude higher than that of pristine R-RuCl3, 5
× 10-4 S/cm.29 Intercalation of insulating polymers reduces the
conductivity (with respect to that of LixRuCl3) to 4.5 × 10-3
and 1.7 × 10-3 S/cm for PEO and PVP, respectively. The charge
transport in these two materials is substantially enhanced by
the presence of conductiVe polymers. Thus (PANI)xRuCl3 shows
a room temperature conductivity of ∼1 S/cm, whereas
(PPY)xRuCl3 shows 23 S/cm.
(26) Toshima, N.; Ihata, O. Synth. Met. 1996, 79, 165.
(27) (a) Wu, C. G.; DeGroot, D. C.; Marcy. H. O.; Schindler, J. L.;
Kannewurf, C. R.; Liu, Y. J.; Hirpo, W.; Kanatzidis, M. G. Chem Mater
1996, 8 (8) 1992. (b) Wu, C. G. Ph.D. Dissertation, Michigan State
University, 1992.
(28) The choice of solvent is important. For example, a pyrrole/CH3CN
solution significantly decreases the reaction rate and does not produce single-
phase (PPY)xRuCl3 even after one month.
Variable-temperature measurements on pressed pellets reveal
that the electrical conductivities of R-RuCl3, LixRuCl3 (x ∼ 0.2),
(29) Binotto, L.; Pollini, I.; Spinolo, G. Phys. Status Solidi B 1971, 44,
245.

6634 J. Am. Chem. Soc., Vol. 122, No. 28, 2000
Wang et al.
Figure 4. (A) Variable-temperature electrical conductivity measurements for pressed pellets of R-RuCl3 and nanocomposites. (B) Thermopower
measurements for pressed pellets of LixRuCl3 (x ∼ 0.2) and several nanocomposites.
Lix(PEO)yRuCl3, (PANI)xRuCl3, and (PPY)xRuCl3 are thermally
activated, see Figure 4A. In the case of R-RuCl3, LixRuCl3, and
Lix(PEO)yRuCl3, the log(σ) versus 1/T plots are almost linear.
The activation energies, which are calculated according to the
formula σ ) σ0e-∆E/2kT, are 0.36, 0.30, and 0.35 eV, respectively.
Since the pellets used in measurements have interparticle
boundaries, the activation energies do not necessarily correspond
to the intrinsic band gaps. In (PANI)xRuCl3 and (PPY)xRuCl3,
the data do not form straight lines in the log(σ) versus 1/T plots,
which suggests that more than one kind of electrical barriers
exist in these materials, including electron hopping barriers
associated with transport through and across the chains of the
conjugated polymer. At temperatures higher than 100 K, the
data of (PANI)xRuCl3 and (PPY)xRuCl3 almost fall in straight
lines in the log(σ) versus 1/T plots. The apparent activation
energies are 0.18 eV for (PANI)xRuCl3 (>125 K) and 0.10 eV
for (PPY)xRuCl3 (>160 K).
5. Magnetic Susceptibility. In R-RuCl3 the intralayer Ru3+
ions are ferromagnetically coupled; however, between layers
the coupling is antiferromagnetic (AF).30 The interlayer AF
coupling is observable at low temperatures (2-20 K), and causes
a well-defined ordering at about 15.6 K, see Figure 5A. At
higher temperatures (50-300 K), the magnetic susceptibility
follows Curie-Weiss law. The Weiss constant θ is positive
which reflects the weak effect of the intralayer ferromagnetic
coupling.
In LixRuCl3 (x ∼ 0.2), the conspicuous AF ordering disap-
pears and the magnetic susceptibility follows Curie-Weiss law
to temperatures as low as 2 K, see Figure 5B. We find that a
temperature-independent diamagnetic correction (øTIP) is neces-
sary before the Curie-Weiss law behavior is clearly revealed.
The Weiss constant θ becomes negative, indicating a change
to a weak overall AF coupling. These changes in magnetic
behavior are due to the replacement of Ru3+ centers for
diamagnetic Ru2+ centers in the Ru sublattice, which alters the
magnetic couplings both in the layers and between the layers.
The magnetic susceptibility data for Lix(PEO)yRuCl3, (PANI)x-
RuCl3, and (PPY)xRuCl3 are similar to those of LixRuCl3, see
Figure 5C,D. Since the reduction of the RuCl3 layers has already
disrupted the original interlayer AF coupling, the subsequent
insertion of the polymers does not have any additional effect
on the magnetic susceptibility.
The magnetic moment µeff for R-RuCl3 is 2.32 µB (literature
value 2.25 µB),30 while those for LixRuCl3 and nanocomposites
Lix(PEO)yRuCl3, (PANI)xRuCl3, and (PPY)xRuCl3 range from
1.6 to 1.74 µB. The drop in magnetic moment is due to the
decrease in the number of unpaired electrons in RuCl3 layers
because of the presence of diamagnetic low-spin Ru2+ centers.
The paramagnetic moments µeff and Weiss constants θ for these
compounds are listed in Table 1. The magnetic susceptibilities
ødia (diamagnetic contribution of all atoms) and øTIP (temper-
ature-independent paramagnetism) used in the manipulation of
the magnetic data are given in Table 1.
The thermopower, which is less sensitive to grain boundary
effects, was measured for LixRuCl3 (x ∼ 0.2), (PANI)xRuCl3,
and (PPY)xRuCl3, see Figure 4B. All three materials have
positive Seebeck coefficients suggesting that the dominant
carriers are holes. The p-type charge-transport behavior is
consistent with the fact that the reduced RuCl3 layers, the
polyaniline (emeraldine salt), and polypyrrole are all p-type
conductors. Both the high Seebeck coefficient value of ∼60
µV/K for LixRuCl3 and its decreasing trend with rising tem-
perature indicate that LixRuCl3 is a type of semiconductor. The
hole type transport in the [RuCl3]x- layer arises from the
partially empty band composed of t2g type Ru orbitals that has
5
an electron configuration between t2g and t2g6. (PANI)xRuCl3
and (PPY)xRuCl3 have lower thermopower values than LixRuCl3,
which increase as the temperature increases. This trend is usually
seen in metallic conductors. Considering that the reduced RuCl3
layers are thermally activated hopping conductors, the bulk
metal-like conductivity, as suggested by the small thermopower
of (PPY)xRuCl3 and (PANI)xRuCl3, indicates that charge
transport in these materials is controlled by the conductive
polymers.
(30) Kobayashi, Y.; Okada, T.; Asai, K.; Katada, M.; Sano, H.; Ambe,
F. Inorg. Chem. 1992, 31, 4570.

R-RuCl3/Polymer Nanocomposites
J. Am. Chem. Soc., Vol. 122, No. 28, 2000 6635
Figure 5. Magnetic susceptibility data as a function of temperature for (A) R-RuCl3, (B) LixRuCl3 (x ∼ 0.2), (C) Lix(PEO)yRuCl3, and (D)
(PANI)xRuCl3.
Table 1. Magnetic Properties of R-RuCl3 and Nanocomposites
µeff
106ødia 106 øTIP
structural issues, and due to the fact that the PEO conformation
and packing varies from host to host.
θ
Many structures for PEO and PEO complexes are known.
The helical conformation is the most common for bulk PEO,32
and exists in pure spherulite samples.33 The planar zigzag
conformation has been observed in stretched PEO samples.34
Two more types of conformations have been found in PEO-
HgCl2 complexes.35,36 There are many PEO-(alkali metal salt)
complexes with versatile morphologies and the determination
of their structures has been attempted. The conformation of PEO
in these alkali metal salt complexes is thought to be a double
helix,37 a waving single helix, which accommodates itself around
the alkali ion lattice,38 or one similar to that in the type-II PEO-
HgCl2 complex.39 In addition, many PEO complexes form with
organic molecules: PEO-urea complex,40 PEO-thiourea com-
plex,41 PEO-p-dibromobenzene complex,42 and PEO-pa-
nitrophenol complex.43 Several common PEO conformations are
shown in Figure 6. Are any of these structures and conforma-
sample
(µB/mol Ru) (K) (L/mol Ru) (L/mol Ru)
R-RuCl3
2.32
1.60
1.63
1.73
1.74
11
-18
-20
-10
-15
-101
-101
-133.
-167
-142
0
176
250
667
247
LixRuCl3 (x ∼ 0.2)
(PANI)xRuCl3
Lix(PEO)yRuCl3
(PPY)xRuCl3
6. Polymer Structure: One-Dimensional Electron Density
Calculations and Arrangement of Polymer Chains in Lix-
(PEO)yRuCl3. In lamellar inorganic/polymeric nanocomposites,
the atomic arrangement in the inorganic layers is well-defined.
However, the arrangement of polymer chains in the interlayer
galleries is not so clear and in general cannot be deduced form
the extent of the interlayer expansion alone. The structural issue
in these materials is critical if we are to gain better understanding
of the properties. Because the interlayer gallery provides a well-
defined two-dimensional confinement space for polymer chains,
their arrangement, conformation, and dynamic behavior in the
gallery is an outstanding issue in fundamental polymer physics.
Many have explored such systems with different approaches.31
Because of the simple repeat structure and flexible chain of
PEO, its intercalative nanocomposites have been the subject of
most investigations,31c-f yet the structure and conformation of
the confined polymer remains an important question. This is
due to the relatively few studies aimed at specifically answering
(32) (a) Takahashi, Y.; Tadokoro, H. Macromolecules 1973, 6, 672. (b)
Brandrup, J.; Immergut, E. H. Polymer Handbook, 3rd ed.; John Wiley &
Sons: New York, 1989, pp VI/72.
(33) Price, F. P.; Kilb, R. W. J. Polym. Sci. 1962, 57, 395.
(34) Takahashi, Y.; Sumita, I.; Tadokoro, H. J. Polym. Sci., Polym. Phys.
Ed. 1973, 11, 2113.
(35) (a) Blumberg, A. A.; Pollack, S. S. J. Polym. Sci., Part A 1964, 2,
2499. (b) Ishihara, H.; Saito, Y.; Ishihara, H.; Tadokoro, H. J. Polym. Sci.,
A-2 1968, 6, 1509.
(36) Yokoyama, M.; Ishihara, H.; Ishihara, H.; Tadokoro, H. Macro-
molecules 1969, 2, 184.
(37) Parker, J. M.; Wright, P. V.; Lee, C. C. Polymer 1981, 22, 1305.
(38) (a) Chatani, Y.; Okamura, S. Polymer 1987, 28, 1815. (b) Lightfoot,
P.; Mehta, M. A.; Bruce, P. G. Science 1993, 262, 883. (c) Lightfoot, P.;
Nowinski, J. L.; Bruce, P. G. J. Am. Chem. Soc. 1994, 116, 7469.
(39) Chatani, Y.; Fujii, Y.; Takayanagi, T.; Honma, A. Polymer 1990,
31, 2238.
(31) (a) Nazar, L. F.; Zhang, Z.; Zinkweg, D. J. Am. Chem. Soc. 1992,
114, 6239. (b) Aranda, P.; Ruiz-Hitzky, E. Chem. Mater. 1992, 4, 1395.
(c) Nazar, L. F.; Wu, H.; Power, W. P. J. Mater. Chem. 1995, 5, 1985. (d)
Liu, Y.-J.; Schindler, J. L.; DeGroot, D. C.; Kannewurf, C. R.; Hirpo, W.;
Kanatzidis, M. G. Chem. Mater. 1996, 8, 525. (e) Matsuo, Y.; Tahara, K.;
Sugie, Y. Carbon 1997, 35, 113. (f) Kerr, T. A.; Wu, H.; Nazar, L. F.
Chem. Mater. 1996, 8, 2005.
(40) Tadokoro, H.; Yoshihara, T.; Chatani, Y.; Murahashi, S. J. Polym.
Sci. B 1964, 2, 363.

6636 J. Am. Chem. Soc., Vol. 122, No. 28, 2000
Wang et al.
Figure 8. One-dimensional electron density map for Lix(PEO)y-
RuCl3: (A) calculated from XRD data (B) using a model with
intragallery double helical PEO that takes an orientation as shown in
Figure 6e and (C) using an intragallery double helical PEO model that
assumes an orientation of 45° rotation relative to the one shown in
Figure 6e and (D) a model with randomly rotated intragallery double
helical PEO.
Figure 6. Experimentally observed PEO conformations: (a) single
heLix (7/2), (b) zigzag, (c) type-I PEO-HgCl2, (d) type-II PEO-HgCl2,-
and (e) double helix.
calculation for the 1-D ED map, the phases of the structure
factors are needed. Approximate starting phases for the calcula-
tion were obtained from the positions of the Ru and Cl atoms.
Due to the minor contribution of the polymer to the total
scattering intensity, the inorganic layer plays a critical role here
in phasing the structure factors and consequently in the
computation of the electron density map. Figure 8B,C,D simply
provides expected 1D ED profiles based on known models of
PEO conformation. Figure 8A shows the 1D ED map calculated
from the diffraction data using starting phases.
The 1-D ED map profile for Lix(PEO)yRuCl3, projected along
the stacking c-axis, shows clearly that the electron density due
to the guest molecules forms two asymmetric peaks placed 1.4
Å above and below the center of the gallery (Figure 8A). The
1-D ED map suggests that two layers of PEO exist in the gallery,
one above and the other below the middle of the gallery. In
each layer, the PEO chains seem to adopt a conformation very
similar to that of the type-II HgCl2 complex.45 (The model is
shown in Figure 6.) The wide valley in the 1-D ED map in the
center of the gallery immediately excludes the possibility of a
layer of helically formed PEO, which must have appreciable
electron density in the central region of the gallery. Considering
that the gallery height of Lix(PEO)yRuCl3 is 8.0 Å, the
accommodation of two layers of helical PEO in such a gallery
is also physically impossible, because the van der Waals
diameter of one PEO helix is ∼8.0 Å.32 Furthermore, neither
the arrangement of PEO in the zigzag conformation (which was
proposed in (PEO)xV2O5 31d) nor that of the type-I HgCl2
complex (which was proposed for Na0.3(PEO)xMoO3 31c) is
consistent with the hereby observed 1-D ED map profile.
Figure 7. X-ray powder diffraction pattern of an oriented Lix(PEO)y-
RuCl3 film used for the calculation of one-dimensional electron density
maps. Due to the preferential orientation of the layers on the sample
holder, the reflection-mode powder XRD patterns show predominantly
the 00l reflections.
tions relevant to the PEO-nanocomposites described here? To
answer this question we performed a careful analysis of the XRD
reflectivity patterns of well-oriented samples. From these
patterns we obtained one-dimensional electron density map
profiles projected along the stacking axis (defined as the c-axis)
and compared them with those profiles expected from the known
conformations.
Oriented films of Lix(PEO)yRuCl3 have well-defined sharp
XRD patterns with up to fourteen 00l reflections corresponding
to a resolution of 0.98 Å, see Figure 7. The 1-D ED maps for
Lix(PEO)yRuCl3, which can provide information about the
projected structure on the c-axis, were calculated from the X-ray
powder pattern of oriented films (see Figure 844). In the
(44) 1-D ED maps were also calculated for models with PEO in zigzag
and type-I PEO-HgCl2 conformations. These calculated maps do not match
the experimental data and are not presented.
(41) Tadokoro, H. Macromol. ReV. 1966, 1, 119.
(42) Point, J. J.; Coutelier, C. J. Polym. Sci., Polym. Phys. Ed. 1985, 23,
231.
(43) (a) Point, J. J.; Damman, P. Macromolecules 1992, 25, 1184. (b)
Damman, P.; Point, J. J. Macromolecules 1994, 27, 3919.
(45) This conformation was also suggested for poly(ethylene glycol)
(PEG) in kaolinite/PEG nanocomposites, because its hydrophilic side has
higher affinity for the highly polar gibbsite Al(OH)3 side of the clay, whereas
its hydrophobic side prefers the not so polar tetrahedral SiO2 side. Tunney,
J. J.; Detellier, C. Chem. Mater. 1996, 8, 927.

R-RuCl3/Polymer Nanocomposites
J. Am. Chem. Soc., Vol. 122, No. 28, 2000 6637
Figure 10. One-dimensional electron density map for Agx(PEO)yRuCl3
using a model similar to that shown in Figure 9. (Solid line, calculated
from experimental data; dash line, from model.)
examined by conducting a control experiment in which the Li+
ions were replaced with Ag+ ions. We reasoned that because
both ions are single-charged the heavier Ag+ may behave
similarly to Li+ in the nanocomposite and at the same time
substantially contribute to the diffraction pattern (i.e. isomor-
phous replacement). The 1-D ED map for the Agx(PEO)yRuCl3
(x ∼ 0.2; y ∼ 0.7)46 sample indicated substantially higher
electron density in the central region, see the central peak in
Figure 10. To calculate the expected 1D ED map (dashed line),
0.2 equiv of Ag+ ions were placed in the center of the gallery.
The good agreement between the two maps supports the
argument that the ions tend to occupy the center of the gallery.
Figure 9. (A) Structural model for Lix(PEO)yRuCl3 derived from the
one-dimensional electron density maps of Lix(PEO)yRuCl3. For clarity,
in the two galleries the directions of the PEO are drawn orthogonal to
each other. This model should not be taken to mean that a regular
periodic arrangement is present in the ab-plane. Certainly substantial
disorder and lack of registry between these chains is expected. (B)
Comparison between the electron density map derived from the
proposed model and a map calculated from XRD data of Lix(PEO)y-
RuCl3.
There exists some similarity between the present model and
the structure of the PEO-NaSCN complex, in which PEO also
has a conformation similar to that found in type-II PEO-
HgCl2.38 In PEO-NaSCN, the chains are arranged with the CH2-
CH2 units in proximity of the S atoms of NaSCN, which are
less hydrophilic, whereas O atoms coordinate with the Na+ ions,
which are much more hydrophilic. By comparison, PEO in
Lix(PEO)yRuCl3 behaves similarly by placing its -CH2CH2-
segments close to the Cl ion surface of the layers while
maximizing Li- - -O interactions. In addition, this particular form
of PEO-NaSCN complex exists under high tension and
converts to another form with a helical PEO conformation if
the tension is released. In Lix(PEO)yRuCl3, the polymer chains
extend themselves over a 2-D environment, which of course
represents a form of tension as well.
On the basis of the above considerations, the model of Figure
9 can be proposed in which the PEO chains are placed at a van
der Waals distance from the RuCl3 layers, i.e., it is in touch
with RuCl3. The van der Waals thickness of a layer of R-RuCl3
is 5.73(3) Å. According to the crystal structure of the type-II
HgCl2 complex,34 the distance of the outmost C atom in PEO
to the van der Waals boundary of the molecule should be about
1.73 Å. Therefore, the distance of the C atoms should be 4.60
Å from the Ru layer, and the distance of the O atoms from the
middle of the gallery is ∼1.20 Å. As shown in Figure 8A, the
1-D ED map profile of the model matches well that calculated
from experimental data in all regions except the center of the
gallery. This model should not be taken to mean that a regular
periodic arrangement is present in the ab-plane. Certainly
substantial disorder and lack of registry between these chains
is expected. Furthermore, the data are not adequate to preclude
small contributions from other PEO forms.
The proposed Lix(PEO)yRuCl3 model of Figure 9 matches
well the observed 1-D ED map and is geometrically and
chemically reasonable. This model is different from the one
proposed for the Li/PEO/montmorillonite nanocomposite in
which the Li+ ions are thought to be located near the surface
oxides of the silicate layer.47 Presumably the lower affinity of
Li+ for Cl- causes it to select the middle of the gallery, which
presents an oxygen-rich coordination environment. Such a PEO
arrangement should provide a good diffusion environment for
ions in directions parallel to the layers, which could make
Lix(PEO)yRuCl3 a comparable or even better ion-conductor than
the amorphous PEO-(alkali-metal) complexes which have
segments of helical PEO chains in the structure48 (see below).
In the center of the gallery, the electron density of the model
is obviously lower than that suggested from the experimental
data due to the fact that disordered water molecules and lithium
ions accumulate in this region between the PEO layers. The
fact that all the O atoms in the PEO face toward the center of
the gallery makes this region hydrophilic, suggesting that it is
the most probable location for the Li+ ions.
(46) Agx(PEO)xRuClx was prepared from LixRuCl3 with a procedure
similar to that used for the Li analogue Lix(PEO)xRuClx using AgNO3.
This is an ion-exchange (Ag+ for Li+) reaction where both PEO and AgNO3
are added to the reaction. The Ag/RuCl3 ratio was determined to be 0.2
with EDS analysis. The amount of PEO was determined from C, H, N
analysis.
(47) Yang, D.-K.; Zax, D. B. J. Chem. Phys. 1999, 110, 5325.
(48) Gray, F. M. Solid Polymer Electrolytes; VCH Publisher: New York,
1991; Chapter 5, p 83.
Electron density due to the Li+ ions themselves is not
observable in these ED maps due to the small quantities of this
very light metal in the material. The validity of the argument
that the lithium ions could be in the center of the gallery was

6638 J. Am. Chem. Soc., Vol. 122, No. 28, 2000
Wang et al.
Table 2. A Comparison of IR Absorptions (cm-1) of PEO in
7. PEO Conformation from IR Spectra. IR spectroscopy
is not a decisive technique to probe the PEO conformation in a
nanocomposite. The weakness of IR spectroscopy derives from
the uncertainty in correctly assigning the observed signals.
Nevertheless, some discussion here is warranted to highlight
the differences between the various known forms and the
difficulties associated with an unambiguous assignment. The
spectral region from 800 to 1000 cm-1 is supposed to be
sensitive to the conformation of PEO chains. A trans O-C-
C-O conformation would have absorption peaks around 773
and 992 cm-1, while a gauche O-C-C-O conformation would
Various Forms
Lix(PEO)1.5RuCl3
pure
PEO
(x ∼ 0.5)
PEO‚HgCl2(I) PEO‚HgCl2(II) region
1465 (w)
1455 (w)
1445 (w)
1352 (w)
1303 (w)
1462 (s)
1456 (w)
1468 (s)
1451 (s)
I
1445 (s)
1356 (m)
1312 (s)
1355 (w)
1338 (s)
1278 (s)
1350 (s)
1328 (m)
1278 (s)
II
1262 (m)
III
IV
V
1248 (w)
1246 (s)
have absorptions around 880 and 944 cm-1 49
. All known PEO
1237 (m)
1143 (vs)a
1110 (vs)a
1239 (s)
1154 (m)
1100 (vs)
conformations, however, only have absorptions around 850 and
950 cm-1, even those in type-I PEO-HgCl2 complex35b and
PEO-p-nitrophenol complex,43b in which the trans O-C-C-O
exists. Less is known about amorphous versions.
1094 (vs)
1060 (sh)
1110 (vs)
1064 (s)
1044 (s)
1032 (s)
945 (s)
1057 (s)
950 (s)
1046 (vs)
1015 (vs)
944 (w)
923 (s)
1031 (m)
950 (m)
920 (sh)
In addition, some controversy in the PEO IR peak assignment
arises from the disagreement among the type-I PEO-HgCl2 IR
data reported.35a,b The peaks observed at 1324, 1309, 1014, and
815 cm-1 35a were assigned to the trans O-C-C-O conform-
ation,49,31c because there were no corresponding peaks in the
IR spectra of bulk PEO. The more recent IR spectra, reported
for type-I PEO-HgCl2, 35b did not have peaks at 1309, 1014,
and 815 cm-1, while the spectra of type-II PEO-HgCl2 had
VI
VII
892 (m)
876 (s)
859 (m)
831 (m)
846 (m)
837 (s)a
835 (s)
a The peaks around 1148, 1112, and 851 cm-1 are characteristic of
the helical PEO conformation: Matsuura, H.; Miyazawa, T. Spectro-
chim. Acta 1967, 23A, 2433.
peaks at 1312 and 1015 cm-1 35b
Recalling the fact that the
.
type-I PEO-HgCl2 converts to type-II PEO-HgCl2 spontan-
eously,35b one might suspect that the type-I PEO-HgCl2 sample
used by Blumberg and Pollack35a also contained type-II PEO-
HgCl2. If this is so, the peaks at 1309 and 1014 cm-1 do not
belong to the trans O-C-C-O conformation. Therefore, 1324
cm-1 may be the only peak that can be assigned to the trans
O-C-C-O conformation, as is done in some references.31b,e,f
The spectrum of Lix(PEO)yRuCl3 does not have a peak near
1324 cm-1, see Figure 4C, which suggests that no trans O-C-
C-O conformation exists. This agrees with the choice of the
type-II PEO-HgCl2 conformation proposed in the previous
section. However, IR spectroscopy alone is not adequate to
unambiguously select which conformation is present.
Absorption peaks of Lix(PEO)yRuCl3 are compared in Table
2 with those of helical PEO, type-I PEO-HgCl2, and type-II
PEO-HgCl2 read from the spectra in red 35b.50 For easier
comparison the peaks have been grouped into regions I through
VII in Table 2. The IR spectrum of Lix(PEO)yRuCl3, exhibits
broader peaks resulting in some adjacent overlap. In regions II
and VII, the peaks of Lix(PEO)yRuCl3 match those of type-II
PEO-HgCl2. The absence of strong peaks at ∼1278 and ∼875
cm-1 in the spectrum of Lix(PEO)yRuCl3 (see regions III and
VII in type-I PEO-HgCl2) suggests that a type-I-like form of
the polymer is probably not present in significant amounts in
the nanocomposite.
In region IV, the broad peak at 1094 cm-1 matches the 1100
cm-1 peak of the type-II PEO-HgCl2. The peak is assigned to
C-O stretching and its slight shift to the lower frequency is
caused by the coordination of O atoms to Li+ ions. The peak at
1154 cm-1 in type-II PEO-HgCl2 is weak, so it does not appear
in the IR spectrum of Lix(PEO)yRuCl3. Type-II PEO-HgCl2
has a pair of peaks in both regions III and V (see Table 2),
while Lix(PEO)yRuCl3 has only one peak in each of them. In
the spectrum of type-II PEO-HgCl2, the two peaks in each pair
Figure 11. Impedance plots for Lix(PEO)yRuCl3 (top) and Li0.2RuCl3
(bottom) at different temperatures.
have a difference of 31 cm-1 or less. Each pair might merge to
give a single peak in the nanocomposite spectrum. Some of the
peaks observed in the type-II conformation may be due to modes
associated with Hg2+ ion binding (absent in the nanocomposite)
and may not appear in the spectrum of Lix(PEO)yRuCl3.51
8. Ionic Conductivity. Intercalation electrodes and polymer
electrolytes have been central topics in solid-state ionics for
more than 20 years.52 Ion conductivity plays an important role
(49) Papke, B. L.; Ratner, M. A.; Shriver, D. F. J. Phys Chem. Solids
1981, 42, 493.
(50) Reference 35b provides the spectra of two orientations with the
electric vector perpendicular and parallel to the fiber axes. A spectrum from
an un-oriented sample should correspond to the superposition of the two
spectra.

R-RuCl3/Polymer Nanocomposites
J. Am. Chem. Soc., Vol. 122, No. 28, 2000 6639
in both types of materials. Since lamellar nanocomposites which
contain PEO can be used for both intercalation electrodes and
polymer electrolytes, ionic conductivity is an important property
to investigate.
The complex impedance (Z* ) Z* - iZ**) spectra for
Li0.2(PEO)xRuCl3 and for comparison Li0.2RuCl3 were measured
between 20 and 130 °C. The data are presented in Figure 11.
The spectra for Li0.2(PEO)xRuCl3 (Figure 11) are composed of
two well-separated semicircles whose overall size decreases with
increasing temperature. Li0.2(PEO)xRuCl3 exhibits both ionic and
electronic conductivity due to the intercalated Li+ ions and
reduced [RuCl3]x- layers, so its impedance derives from the
combination of the two types of charge transport. The depression
of semicircles in Figure 11 (top), with their centers below the
Z* axis, indicates distributed relaxation time constants (τ)
involved in the Li+ conduction, as is expected from the
composite structure of Li0.2(PEO)xRuCl3.
The electronic (σel) and ionic (σion) conductivities of
Li0.2(PEO)xRuCl3, converted from the fitted Rel and Rion values,
are presented in Figure 12A. As shown in the figure, the ionic
conductivity of Li0.2(PEO)xRuCl3 falls in the range 10-4.2-10-3.0
S/cm between 20 and 130 °C, and is comparable to those of
the best (lithium salt)/polymer electrolytes.48 The electronic
conductivity is of the same order of magnitude in the same
temperature range. The contribution of the ionic conductivity
to the total electrical conductivity, which is defined as the
transference number (tion ) σion/(σion + σel)), is estimated at
0.55-0.43 in the above temperature range. The deflection in
the line of the Arrhenius plot, which appears in both the ionic
and electrical conductivity, indicates a phase transition at about
80 °C. The corresponding activation energies below and above
the transition temperature are 0.146 and 0.485 eV for σion and
0.082 and 0.488 eV for σel.
The corresponding impedance spectra for Li0.2RuCl3 indicate
different transport properties for this material. First, their
significantly lower impedance values correspond to electrical
conductivities about 2 orders of magnitude higher, see Figure
11 (bottom). Second, they are composed of only one semicircle
and the vertical tail extends across the Z* axis, in contrast to
those of Li0.2(PEO)xRuCl3.
Figure 12. (Top) Arrhenius plots of the ionic (right) and electronic
(left) conductivities of Lix(PEO)yRuCl3. (Bottom) Comparison of the
electronic conductivity of LixRuCl3 with the electrical (σTot) and
electronic conductivities of Lix(PEO)yRuCl3, measured by impedance
spectroscopy.
could not be extracted from σTot. As a consequence, σion is
estimated to be much lower than that of Li0.2(PEO)xRuCl3.53
The σel (σTot) of Li0.2RuCl3, which is also thermally activated,
is plotted together with the σel and σTot of Li0.2(PEO)xRuCl3 in
Figure 12B. The agreement with the electrical conductivity
values determined independently with the four-probe method,
as discussed above, is excellent. The comparison shows that
the encapsulation of insulating PEO results in a substantial
increase in ionic mobility of Li. The decrease in σel (with respect
to Li0.2RuCl3) is attributed mainly to the increased separation
in the [RuCl3]x- layers. The presence of PEO in the galleries
provides for low activation energy barriers for Li+ transport,
which is believed to occur primarily through two-dimensional
channels.
In Li0.2RuCl3, σel is so much higher than σion that the total
electrical conductivity σTot is practically equal to σel. Thus σion
(51) The IR absorption of Lix(PEO)yRuCl3 in regions I and VI is different
from that of type-II PEO-HgCl2 than in other regions. The peaks of
Lix(PEO)yRuCl3 at 1465, 1455, and 950 cm-1, which are close to those of
the helical PEO, may come from PEO chains dangling outside the layers.
(Other absorption peaks of free PEO chains may be buried in the absorption
of Lix(PEO)yRuCl3.) In region VI, the absence of the 892 and 944 cm-1
peak in the Lix(PEO)yRuCl3 spectrum might be explained by its low
intensity. The absorption in this region is contributed mostly by the 950
cm-1 peak of the free PEO chains and the 923 cm-1 peak of the type-II
PEO-HgCl2 conformation. In any case, the IR absorption profile of
Lix(PEO)yRuCl3 seems to match the spectrum derived from a superposition
of the absorption peaks of type-II PEO-HgCl2 conformation and free PEO
chains.
Concluding Remarks
The robustness of R-RuCl3, in acidic or basic aqueous
solutions, and its good stability against reducing and oxidizing
conditions makes it an excellent compound for intercalation
reactions. This is a rare set of properties among transition metal
halides. The insertion of conducting polymers PANI and PPY
and the water-soluble polymers PEO, PVP, and PEI in RuCl3
gives a new class of lamellar (metal halide)/polymer intercalative
nanocomposites. The synthesis of these nanocomposites is
enabled mainly by the successful exfoliation of R-RuCl3 after
controlled lithiation, which allows the use of the encapsulative
precipitation method.10 On the other hand, R-RuCl3 is also one
of the few layered hosts that is suitable for in situ redox
intercalative polymerization, and provides an interesting and
unique metal halide component for new nanocomposites.
R-RuCl3 nanocomposites contain reduced layers in which free
electron hopping, associated with Ru2+/Ru3+ couples, raises the
(52) Bruce, P. G. In Fast Ion Transport in Solids; Scrosati, B., et al.,
Eds.; NATO ASI Ser. Ser. E; Kluwer Academic Pub.: Dordrecht, The
Netherlands, 1993; Vol. 250, p 87.
(53) As an alternative method to determine the ionic conductivity of Lix-
RuCl3, a DC relaxation technique could be useful, since in principle it can
provide the impedance spectrum information at the extremely low ω region
(,5 Hz). DC polarization-relaxation measurements for Li0.2RuCl3 were
tried with a cell in the configuration “stainless steel/Lix/sample/stainless
steel”, with various polarization voltages and times, and with various lithium
salts (LiI, Li(CF3SO2)2N, and LiClO4). Nevertheless, we could not observe
a proper relaxation behavior for the specimen, which is probably due to a
systematic limitation arising from the very low σion of LixRuCl3. An ideal
DC relaxation experiment can be performed by an application of the
electron-blocking compartment with very high σion, but as far as we know,
such Li+ conductors are not readily available yet.

6640 J. Am. Chem. Soc., Vol. 122, No. 28, 2000
Wang et al.
electrical conductivity by 2-3 orders of magnitude. The
intercalation of conducting polymers can further increase the
conductivity. The dominant carriers in the [RuCl3]x- layers are
holes residing in a narrow t2g-type band. The reduction of
R-RuCl3 and polymer insertion profoundly affect the intralayer
and interlayer Ru3+ magnetic couplings, so that new magnetic
properties appear in the nanocomposites.
troscopy. In fact, the ionic conductivity is comparable to those
of the best (lithium salt)/polymer electrolytes. Given that the
materials in the general class of polymer/RuCl3 combine high
electrical conductivity (e.g. conjugated polymer guests) with
the potential catalytic properties of RuCl3, it is intriguing to
contemplate new systems with valuable electrocatalytic proper-
ties.
On the basis of X-ray scattering and IR spectroscopy, a
structural model is proposed for Lix(PEO)yRuCl3, in which each
gallery contains two layers of PEO chains in a conformation
similar to that found in type-II PEO-HgCl2. The Li+ ions seem
to reside exactly in the middle of the interlayer space sandwiched
between two monolayers of PEO. The model suggests that
Lix(PEO)yRuCl3 should have good two-dimensional ionic
conductivity, which is indeed confirmed by impedance spec-
Acknowledgment. Financial support from the National
Science Foundation is gratefully acknowledged. This work made
use of the magnetic measurement facilities of the Department
of Physics and the SEM and TEM facilities of the Center for
Electron Optics at Michigan State University.
JA9944610