Sol-gel synthesis of hydrous ruthenium oxide nanonetworks from 1,2-epoxides

Article abstract of DOI:10.1016/j.jssc.2007.05.031

Hydrous ruthenium oxide (RuO₂·xH₂O) xerogels were synthesized through the addition of a 1,2-epoxide, propylene oxide, to commercial hydrated ruthenium chloride, RuCl₃·xH₂O, in ethanol. After a blue-black monolithic gel formed in 4 h, the samples were allowed to age for 24 h and were dried in ambient conditions. The dried samples were then characterized by XPS, XRD, DTA and TGA. XPS showed the Ru(3d_5/2) peak at a binding energy of 281.7 eV, corresponding to that of hydrous ruthenium oxide. XRD data revealed the synthesized material as amorphous. Heating the sample in inert atmospheres caused the complete reduction of the oxide to the zero-valent state, whereas heating the sample in air resulted in both crystalline anhydrous RuO₂ and zero-valent ruthenium, depending on the method of heating. DTA traces showed an endotherm ending at 150 °C, corresponding to the loss of coordinated water, as well as two higher temperature crystallization exotherms when the sample was heated in both inert and oxygen-rich atmospheres. TGA runs also confirmed the complete reduction of the hydrous oxide when heated in nitrogen below 270 °C and the formation of anhydrous ruthenium oxide when heated in air, confirming the XRD results.

Full text of DOI:10.1016/j.jssc.2007.05.031

ARTICLE IN PRESS
Journal of Solid State Chemistry 180 (2007) 2290–2297
www.elsevier.com/locate/jssc
Sol–gel synthesis of hydrous ruthenium oxide nanonetworks
from 1,2-epoxides
Ã
Jeremy Walkera, R. Bruce Kingb, Rina Tannenbauma,
aSchool of Materials Science and Engineering, Georgia Institute of Technology, Atlanta, GA 30332, USA
bDepartment of Chemistry, University of Georgia, Athens, GA 30602, USA
Received 22 February 2007; received in revised form 29 May 2007; accepted 31 May 2007
Available online 6 June 2007
Abstract
Hydrous ruthenium oxide (RuO2 ꢀ xH2O) xerogels were synthesized through the addition of a 1,2-epoxide, propylene oxide, to
commercial hydrated ruthenium chloride, ‘‘RuCl3 ꢀ xH2O,’’ in ethanol. After a blue-black monolithic gel formed in 4 h, the samples were
allowed to age for 24 h and were dried in ambient conditions. The dried samples were then characterized by XPS, XRD, DTA and TGA.
XPS showed the Ru(3d5/2) peak at a binding energy of 281.7 eV, corresponding to that of hydrous ruthenium oxide. XRD data revealed
the synthesized material as amorphous. Heating the sample in inert atmospheres caused the complete reduction of the oxide to the zero-
valent state, whereas heating the sample in air resulted in both crystalline anhydrous RuO2 and zero-valent ruthenium, depending on the
method of heating. DTA traces showed an endotherm ending at 150 1C, corresponding to the loss of coordinated water, as well as two
higher temperature crystallization exotherms when the sample was heated in both inert and oxygen-rich atmospheres. TGA runs also
confirmed the complete reduction of the hydrous oxide when heated in nitrogen below 270 1C and the formation of anhydrous ruthenium
oxide when heated in air, confirming the XRD results.
r 2007 Elsevier Inc. All rights reserved.
Keywords: Sol–gel chemistry; Hydrous ruthenium oxides; Ruthenium oxide nanonetworks
1. Introduction
resistance [13], and good electrochemical cyclability [14]
with maximum capacitance values of the material occur-
ring when xꢁ0.5 [15].
Anhydrous ruthenium oxide materials possess versatile
functionalities and properties that make them useful in
numerous industries. RuO2 is known as an exceptional
electrode material for high-power and high-energy density
capacitors as a result of its high capacitance [1] and exhibits
low resistivity and good thermal stability [2,3]. It is also
known to be a metallic conductor [4–6] and has been used
as a catalyst for organic and inorganic reactions [7,8], an
anode material in the production of chlorine [9], and in
cryogenic temperature sensors [10]. RuO2 is also a good
candidate for the bottom electrode for high dielectric
constant capacitors owing to its etching capability [11].
Hydrous ruthenium oxide, RuO2 ꢀ xH2O, can be used as an
electrochemical capacitor due to its high-energy density
[12]. These hydrous materials possess high capacitance, low
Ruthenium oxides have been prepared by sputtering [16],
cathodic electrosynthesis [17], and solution chemistry
including thermal decomposition of hydrated RuCl3 [18]
and sol–gel processes [19,20]. Suh et al. [13] have
demonstrated that the sol–gel epoxide addition method
can produce ruthenium oxide aerogels of high porosity
from hydrated ruthenium chloride solutions. The sol–gel
process is a very versatile synthesis method for the
formation of a porous oxide with controlled dimensions
of the particle sizes. Previous work involving iron(III)
oxides from hydrated iron salt precursors in the presence of
epoxides [21–25] has demonstrated that the sol–gel process
can generate a network material having particles with
nanoscale features, high porosity, high surface area, and
low density [22,23]. This epoxide addition in the sol–gel
formation of metal oxides has also been successful in the
synthesis of other technologically relevant oxides such as
Ã
Corresponding author. Fax: +1 404 894 9140.
E-mail address: rinatan@mse.gatech.edu (R. Tannenbaum).
0022-4596/$ - see front matter r 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.jssc.2007.05.031

ARTICLE IN PRESS
J. Walker et al. / Journal of Solid State Chemistry 180 (2007) 2290–2297
2291
NiO [26], Al2O3 [27], Cr2O3, ZrO2, WO3, as well as
numerous other transition metals [28].
the ethanol. (The estimate of 1.6 mmol of Ru was
calculated by using the exact formula of the Ru precursor,
RuCl3.37 ꢀ 2.5H2O, having a molecular weight of 265.6 g/
mol, which was obtained after performing elemental
analysis on the as-bought RuCl3 ꢀ xH2O salt. However,
the commercially available hydrated ruthenium-(III) chlor-
ide is a mixture of several phases, including the hydrate of
ruthenium-(IV) oxide derivatives [29].) Then a 1.2 mL
allotment of the weak base, propylene oxide, was added as
the gelation chemical to the solution and stirred. The time
to gelation was monitored and recorded after the addition
of the gelation chemical. After gelation occurred, the gel
was covered and allowed to age for 24 h. Then the aged gel
was allowed to dry in a fume hood under ambient
conditions to allow solvents to evaporate. To complete
solvent evaporation, the sample was placed under vacuum
for an additional 48 h.
Our particular interest in exploiting these properties lies
in the design of metal oxide matrices for energetic
materials. Such energetic materials are comprised of a
particulate metal fuel (e.g., Al, Fe, Cr) closely mixed with
metal oxide particles, which, after a stress-induced oxida-
tion–reduction reaction, result in a substantial exothermic
heat release. In these materials, reactions between different
metal oxide networks (oxidants) and dispersed metallic
particles (fuels) will result in different energetic release
rates. The efficiency of the exothermic output of the
energetic reaction is also governed by the interfacial area
and reactivity between the metal fuel and the metal oxide
particles and hence, reducing the size of the metal oxide
(oxidant) network as well as the size of the particulate
metal fuel to the nanometer scale, will dramatically
increase the interfacial area and will directly translate into
a higher efficiency of the oxidation–reduction reaction.
In this work, we extend the sol–gel synthesis of hydrous
ruthenium oxide involving the addition of an epoxide, e.g.,
propylene oxide, to a hydrated ruthenium chloride
precursor solution, to generate a porous ruthenium oxide
xerogel with nanoscale dimensions of the oxide particles.
Unlike the work of Suh et al. [13] that is specifically
focusing on the formation of highly porous ruthenium
oxide aerogels by carbon dioxide supercritical drying, our
interest lies mainly in the formation and thermal behavior
of the denser xerogel material. In view of the potential
application of such oxides as energetic materials, the main
factor of interest in our case is the possibility of formation
of tight reactive interfaces between the ruthenium oxide
‘‘nanonetwork’’ and the fuel nanoparticles, and the
influence of the increase in the local temperature on
the chemistry of the reactants. Despite the fact that the
electronic configuration of Ru is similar to that of Fe,
the chemistry of its oxides bears little resemblance to that
of iron. While there is extensive chemistry associated with
the MxOy (M ¼ Fe or Ru) species for both elements, the
higher oxidation states of Ru are much more easily
obtained than for iron. Hence, in this paper, we set out
to probe the synthesis, thermal stability, and reactivity of
hydrous ruthenium oxide nanonetworks as the potential
oxidation moiety in energetic materials systems.
2.2. Physical characterization of synthesized material
The dried xerogel material was ground into a fine
powder using a ceramic mortar and pestle. XPS scans of
powder samples were taken using a Surface Science
Laboratories SSX-100 ESCA Spectrometer using mono-
chromatic AlKa radiation (1486.6 eV). The system oper-
ated at a pass energy of 50 eV. Powder samples were
housed in aluminum foil during analysis and a flood gun
was used at a voltage of 3 eV. The operating pressure of the
vacuum chamber was less than 3 ꢂ 10ꢃ8 Torr. General
scans covered the binding energy range of 0–1100 eV.
Thirty high-resolution C(1s)/Ru(3d5/2) scans were run with
a central binding energy (CBE) of 285 eV with a window
width of 20 eV at a spot size of 400 mm. Also, 30 high-
resolution O(1s) scans were run at a CBE of 532 eV using a
spot size of 400 mm and a window width of 20 eV. Each
high-resolution scan possessed a 0.1 eV per step interval.
Curve fitting of the data was accomplished using the
program Spectral Data Processor, Version 4.1.
XRD of the samples was performed on a Philips PW
1800 X-ray diffractometer. Patterns from 201 to 851 were
examined with a step size of 0.021 using monochromatic
˚
CuKa X-rays with a wavelength of 1.54056 A. Powder
samples were analyzed using a zero background sample
holder.
HRTEM images were obtained from a Hitachi HF 2000
FE TEM operating at an accelerating voltage of 200.0 keV.
Elemental analysis of the material was performed by
Atlantic Microlabs, Norcross, GA. Quantitative values of
carbon, hydrogen, oxygen, and chlorine were determined
through experimental measures. The balance of the
compositional make-up of the sample was assigned as
ruthenium. Lower temperature DSC data were compiled
from a TA Instruments Q100 DSC up to temperatures of
550 1C. Samples were placed in aluminum pans and heated
at 10 1C/min with a nitrogen purge at 20.00 mL/min. High-
temperature DTA/TGA data (4600 1C) was determined
using a Netzsch STA 449-Jupiter TGA-DSC. Powder
2. Experimental
2.1. Synthesis of hydrous ruthenium oxide gels
Ruthenium-(III) chloride hydrate, RuCl3 ꢀ xH2O; and
propylene oxide, C3H6O, were purchased from Fisher
Scientific and used as received. Stock absolute ethanol was
obtained from Aldrich and used as received. The synthesis
was performed in a 20 mL glass scintillation vial under
ambient conditions by adding 0.42 g (1.6 mmol) RuCl3 ꢀ
xH2O to 3.5 mL of ethanol and stirring until the powdered
ruthenium chloride component was completely dissolved in

ARTICLE IN PRESS
J. Walker et al. / Journal of Solid State Chemistry 180 (2007) 2290–2297
2292
samples were placed in alumina crucibles and heated at
10 1C/min in both air and argon atmospheres.
performed by high-resolution scans of the C(1s)/Ru(3d)
and O(1s) regions. Fig. 2(b) shows the high-resolution
C(1s)/Ru(3d) region having adjusted the C(1s) peak to
285.0 eV to compensate for the presence of charging effects
in the material [30]. This region can be deconvoluted into
five separate peaks, three related to the C(1s) photoelec-
As a cautionary note, to the best of our knowledge, the
chemical, physical and toxicological properties of ruthe-
nium oxide have not been thoroughly investigated and
recorded. What is certain is the fact that it is irritating to
skin and mucous membranes. The oxidation of ruthenium
and/or ruthenium compounds may form the volatile, toxic
and highly irritating ruthenium-(VIII) oxide, and hence all
reactions are performed in a well-vented hood and
according to appropriately recommended safety proce-
dures.
trons and two associated with the Ru(3d5/2) and Ru(3d3/2
)
peaks: (A) at a binding energy (BE) of 281.7 eV corre-
sponding to the Ru(3d5/2) peak of the material, (B) at a BE
of 285.0 eV corresponding to aliphatic C–C bonding arising
from residual ethanol as well as the chloropropanol and
propanediol intermediates resulting from the sol–gel
formation process [21], (C) at a BE of 286.3 eV correspond-
ing to the Ru(3d3/2) peak of the material, (D) the more
electronegative C–Cl species at a binding energy of
287.5 eV originating from the chloropropanol species, and
(E) at a BE of 288.7 corresponding to the C–O bonding
from the same organic compounds mentioned in (A). (E) is
located at the highest BE owing to the highest electro-
negativity of oxygen. It is important to note the absence of
an epoxide peak in the C(1s)/Ru(3d) high-resolution scan.
This is due to the fact that propylene oxide is a key
component in the reaction and, unlike a traditional
catalyst, it actually gets consumed during the reaction
[21]. The Ru(3d5/2) peak of the material is located at a
3. Results and discussion
3.1. Synthesis and characterization of hydrous ruthenium
oxide gels
The addition of propylene oxide to a commercial
hydrated ruthenium chloride precursor solution (assumed
to contain mainly Ru(IV) as ‘‘Ru(OH)Cl3’’) results in a
heat-releasing reaction with the formation of a blue-black
monolithic wet gel in 4 h, as compared to 25 min for the
formation of an iron oxide gel from the FeCl3 ꢀ 6H2O
precursor hydrous salt. The slower kinetics in the forma-
tion of the ruthenium oxide nanonetwork as compared to
those of iron oxide nanonetworks may result from a greater
kinetic inertness of Ru(III) and Ru(IV) relative to Fe(III)
[29]. A porous xerogel of hydrous ruthenium oxide
(RuO2 ꢀ xH2O) forms after residual solvents in the gel are
allowed to evaporate in ambient conditions. Fig. 1 shows a
HRTEM image of the synthesized material, which reveals a
porous microstructure having clusters with diameters
ranging from 40 to 80 nm.
binding energy of 281.7 eV whereas the related Ru(3d5/2
)
binding energy peak of elemental ruthenium is at 280.0 eV
[7]. It is known that non-equivalent atoms of the same
element in a solid give rise to peaks with different BE with
increasing BE for increasing oxidation state [31]. The
Ru(3d5/2) peak of the material is located at a higher
BE than the same peak for anhydrous RuO2, i.e.,
280.7–281.0 eV [32], owing to the presence of –OH
functional groups on Ru [33]. In this context, it has been
reported that the Ru(3d5/2) peak for a hydrous ruthenium
oxide is located at a BE of 281.4–281.8 eV [34]. Therefore,
X-ray photoelectron spectroscopy was utilized to eval-
uate the composition of the ruthenium oxide material,
RuO2 ꢀ xH2O, synthesized from the weak base propylene
oxide. Fig. 2(a) shows the general survey spectrum of the
material. Further analysis of the XPS spectrum was
based on the analysis of the deconvolution of the Ru(3d5/2
)
region, it is possible to identify the synthesized material as
a hydrous ruthenium oxide, RuO2 ꢀ xH2O.
The O(1s) region can be deconvoluted into three peaks:
(F) 529.5 eV corresponding to the lattice oxygen within the
ruthenium oxide; (G) 530.8 eV corresponding to the oxygen
within the chemisorbed water in RuO2 ꢀ xH2O [35]; and (H)
532.5 eV corresponding to the oxygen–carbon bonding of
the residual organic diol species from the synthesis as well
as residual ethanol. The relative ratios between the three
experimental O(1s) peaks corresponding to RuO2,
RuO2 ꢀ xH2O, and O–C are 31.1:50.6:18.3, respectively.
This indicates that the majority of the synthesized product
(ꢁ81.7%) is a hydrated ruthenium oxide, with the balance
of the sample oxygen originating from residual organic
species. The reported O(1s) peak location for a hydrated
ruthenium oxide is 529.3 eV for the oxide and 530.5 for the
oxygen within the adsorbed H2O [7]. These O(1s) values are
comparable to the values reported in the synthesized
material here. Whereas the O(1s) peak of elemental oxygen
is centered at 532.0 eV, the BE of the oxygen peak of an
Fig. 1. High-resolution transmission electron microscopy (HRTEM)
image of the as-synthesized hydrous ruthenium oxide gel, exhibiting the
presence of clusters on the order of 40–80 nm.

ARTICLE IN PRESS
J. Walker et al. / Journal of Solid State Chemistry 180 (2007) 2290–2297
2293
A
C(1s)/Ru(3d)
B
C
D
E
1000
800
600
400
200
0
292 290 288 286 284 282 280 278
Binding Energy (eV)
Binding Energy (eV)
O(1s)
G
F
H
536
534
532
530
528
526
Binding Energy (eV)
Fig. 2. The X-ray photoelectron spectra (XPS) of a sol–gel synthesized hydrated ruthenium oxide from the weak base propylene oxide. (a) The general
scan of the oxide material. (b) The high-resolution C(1s)/Ru(3d) core level electrons and its deconvolution into five peaks. (c) The O(1s) core level emission
spectrum and its deconvolution into three peaks.
oxide is lower than pure oxygen–again proving the
existence of a ruthenium oxide material [36]. The O(1s)
region is shown in Fig. 2(c). Also observed in the general
scan of the material is the existence of the Cl(2s) and Cl(2p)
peaks originating from the chloride species of the initial
precursor and the chloropropanol species at BE of 269.7
and 199.3 eV, respectively, and the Al(2s1/2), Al(2p1/2) and
Al(2p3/2) peaks originating from the Al foil used to house
the powder sample in the XPS instrument. The XPS data
for the as-synthesized material are summarized in Table 1.
Table 1
The binding energies, full-width half-maximum values, percentage of total
peak area, and assignment of atoms of the X-ray photoemission spectrum
(XPS) of the O(1s) and C(1s)/Ru(3d) core electron regions for a hydrous
ruthenium oxide sample synthesized by the sol–gel process from propylene
oxide, corresponding to Fig. 2
Peak label
BE (eV)
FWHM
(eV)
% of total
area
Assignment
of atoms
C(1s)/Ru(3d)
A
B
C
D
E
281.7
285.0
286.3
287.5
288.7
2.77
2.15
1.91
1.64
1.71
49.7
23.1
15.2
7.2
Ru(3d5/2
)
)
C–C
Ru(3d3/2
C–Cl
C–O
3.2. Temperature-dependent formation of ruthenium oxide
species
4.7
O(1s)
F
Figs. 3 and 4 show X-ray diffraction data for the as-
synthesized sample as well as patterns for the sample after
various heat treatments. Figs. 3(a) and (b) display the
diffraction pattern of the as-prepared gel and the gel heated
to 125 1C for 2 h under nitrogen, respectively. These
patterns show that the synthesized material is amorphous
with crystallization not occurring until temperatures great-
er than 125 1C. Fig. 3(c) shows the diffraction pattern of a
gel specimen heated to 250 1C and held for 2 h under
nitrogen. The onset of low levels of crystallinity may be
observed at this temperature, generating broad peaks in the
vicinity of Bragg angles of 381, 441, and 701. Fig. 3(d)
shows the XRD pattern of a sample heated to 550 1C at
10 1C/min under nitrogen using a DSC. At this tempera-
529.5
530.8
532.5
1.39
1.51
1.06
31.1
50.6
18.3
RuO2
RuO2 ꢀ xH2O
G
H
O–C
ture, the material exhibits a higher degree of crystallinity
with peaks at 38.431, 42.211, 44.111, 58.521, 69.481, 78.531,
and 84.751. The broad peaks in Figs. 3(c) and (d) are
indicative of the presence of zero-valent ruthenium metal,
while the peaks listed in Fig. 3(d) correspond directly and
unequivocally to the (1 0 0), (0 0 2), (1 0 1), (1 0 2), (1 1 0),
(1 0 3), (1 1 2), and (2 0 1) planes of Ru(0), respectively.
Fig. 3(e) shows the continued trend of the complete
reduction process of the hydrated ruthenium oxide

ARTICLE IN PRESS
J. Walker et al. / Journal of Solid State Chemistry 180 (2007) 2290–2297
2294
(e) 1400°C TGA Ar
(d) 550°C DSC N2
(c) 250°C TGA N2
(d)
1400°C
Alr
(b) 125°C TGA N2
(c)
600°C
Alr
(a) as prepared
(b) 300°C
Alr
(a) as prepared
20
30
40
50
60
70
80
20
30
40
50
60
70
80
2 Theta
2 Theta
(f) RuO2
(e) RuO2
(g) Ru[o]
(f) Ru[o]
20
30
40
50
60
70
80
20
30
40
50
60
70
80
2 Theta
2 Theta
Fig. 3. X-ray diffraction (XRD) patterns of the sol–gel synthesized
hydrous ruthenium oxide samples heated in inert atmospheres to (a) room
temperature, (b) 125 1C, (c) 250 1C, (d) 550 1C, and (e) 1400 1C and
Fig. 4. X-ray diffraction (XRD) patterns of the sol–gel synthesized
hydrous ruthenium oxide samples heated in air at (a) room temperature,
(b) 300 1C, (c) 600 1C, and (d) 1400 1C and published diffraction patterns
published diffraction patterns for (f) anhydrous RuO2 and (g) Ru(0)
.
for (e) anhydrous RuO2 and (f) Ru(0)
.
material to ruthenium metal in an inert atmosphere when
heated to 1400 1C. Figs. 3(f) and (g) show the diffraction
patterns for Ru(0) and RuO2. Ji et al. [34] have stated that
the reduction of hydrous ruthenium oxide to ruthenium
metal is due to the presence and affinity of chloride ions,
which are generated from the ruthenium chloride precursor
salt involved in the sol–gel reaction, to act as intramole-
cular reducing agents, as shown by
thermodynamically driven propensity of ruthenium oxide
to decompose to the zero-valent state under low oxygen
partial pressure, as is the case within the instrument
housings of the TGA and DSC under inert gas purging.
Hence, a possible reaction sequence that could account for
the products observed in this sol–gel reaction is as follows:
4RuIIICl3
ꢃ!
EtOH; Propylene oxide Ruð0Þ þ 3RuIVO2;
(3)
Air
D; Inert Gas
RuIVO2 þ 4Cl1ꢃ ꢃ! Ruð0Þ þ 2Cl2 þ 2O2ꢃ
:
(1)
D; Inert Gas
RuIVO2 ꢃ! Ruð0Þ þ O2:
(4)
Pagnaer et al. [37] have postulated also the possibility
that the organic matrix, stemming from the organic solvent
and gelation agent, is oxidized during heating, forming
carbon monoxide, which is another potential reducing
agent, as shown by
As shown in Eq. (3), RuIII undergoes disproportionation
to Ru(0) and RuIV. The absence of oxygen, or the low
oxygen partial pressure during heating cycles under inert
atmosphere, destabilizes the ruthenium oxide compound
and drives the reduction reaction shown in Eq. (4). The
presence of Ru(0) immediately upon the completion of the
sol–gel reaction, as indicated by Eq. (3), cannot be
confirmed because of three main reasons: (1) the ambiguity
in the assignment of the XPS peaks, owing to the overlap
RuIVO2 þ 2CO D; Iꢃne!rt Gas Ruð0Þ þ 2CO2:
(2)
The reduction of the ruthenium oxide that ensues at
temperatures below 250 1C is most likely due to the

ARTICLE IN PRESS
J. Walker et al. / Journal of Solid State Chemistry 180 (2007) 2290–2297
2295
of the Ru(3d5/2) and Ru(3d3/2) regions with those of C(1s),
100
90
80
70
60
50
40
30
(2) the very large FWHM of the 281.7 eV band assigned to
Ru(3d5/2), which could also be deconvoluted into an
additional band in the vicinity of 280.0 eV corresponding
to Ru(0), and (3) the amorphous character of the product
that does not afford any conclusion based on the X-ray
diffraction below 125 1C.
Air
Indirect evidence of the initial formation of Ru(0) from
the sol–gel reaction may be obtained when examining the
behavior of the product mixture upon heating in air.
Whereas heating the sample in an inert environment results
in the complete reduction of the hydrated ruthenium oxide
to the zero-valent state, heating the sample in air results in
the formation of an anhydrous ruthenium oxide material.
However, the presence of zero-valent Ru upon heating is
still observed under certain conditions. Fig. 4(a) shows the
XRD pattern of the as-prepared amorphous material and
Fig. 4(b) shows the sample heated to 300 1C for 20 h in air,
where the peaks corresponding to anhydrous ruthenium
oxide begin to emerge. Broad peaks can be seen at Bragg
angles of 28.41, 35.31, 44.21, 54.41, and 69.71. The first three
of these peaks match the most intense peaks of anhydrous
RuO2 corresponding to the (1 1 0), (1 0 1), and (2 1 1)
planes. Fig. 4(c) shows an XRD pattern of the material
heated at 600 1C for 20 h in air showing a more crystalline
RuO2. However, peaks related to zero-valent ruthenium
clearly remain in these patterns. Fig. 4(d) displays the XRD
pattern of the sample heated at 10 1C/min under air to a
temperature of 1400 1C. It is not until this temperature that
the complete oxidation of all ruthenium within the sample
occurs resulting in a pure anhydrous ruthenium oxide,
shown by
Argon
0
300
600
900
1200
1500
Temperature (°C)
Fig. 5. Thermogravimetric analysis (TGA) thermal profile for the
decomposition of hydrated ruthenium oxide samples heated to 1400 1C
at a heating rate of 10 1C/min in two different atmospheres, air and argon.
Note an almost zero-weight loss for the sample heated in argon above
300 1C, whereas the sample heated in air retains a larger mass percentage
above this temperature until high-temperature ruthenium oxide thermal
decomposition causes mass loss above 1000 1C.
different atmospheres, namely air and argon. Both samples
lose a substantial fraction of their mass up to ꢁ270 1C. The
sample heated in air loses ꢁ45% and the sample heated in
argon loses ꢁ55% of their respective total weights. This
loss is attributed to loss of adsorbed water in the material
as well as the loss of residual organics remaining from the
initial ambient drying procedure. For the sample heated in
an inert atmosphere, weight loss essentially ceases at this
temperature until 1400 1C. The TGA results also confirm
that the as-prepared hydrous ruthenium oxide is reduced to
the zero-valent state when heated in an inert atmosphere, in
accord with the XPS and XRD data. If the weight loss data
in both atmospheres, air and argon, is normalized between
the temperatures 300 and 1000 1C (the temperature range
where almost no weight loss is observed), a simple weight
ratio of Ru metal to RuO2 (0.76) is equal to the normalized
ratio of the weights of the samples heated in argon and air
(0.77). This value shows that the sample heated in argon
does not have the associated weight of the two oxygens in
RuO2 and the additional 10% weight loss below 300 1C can
be considered a result of the liberation of oxygen formerly
coordinated with the oxide. Also, at 1400 1C, the total
remaining weight percent is 37.8%. This corresponds well
to the total weight percent of 35.5% of ruthenium of the
synthesized material found by elemental analysis. There-
fore, the sample heated in argon can be considered Ru(0)
and the sample heated in air RuO2.
1400 oC; O2
Ruð0Þ ꢃ! RuIVO2:
(5)
Another possible explanation for the presence of Ru(0) in
the samples is that some of the surface oxygen atoms in the
samples heated in air to 300 and 600 1C, generated by the
thermodynamic instability of the ruthenium oxide with
increased temperature, may have desorbed owing to the
lack of a sufficient supply of atmospheric oxygen to
replenish the lost oxygen. This decomposition results in a
Ru(0) surface and bulk anhydrous RuO2, consistent with
the presence of both phases in the XRD patterns. However,
heating the sample to 1400 1C in the TGA can replenish the
surface oxygen liberated from this high-temperature
thermal decomposition due to a continuous source of
oxygen flowing over the sample surface. This maintains a
high oxygen partial pressure and prevents this spontaneous
decomposition, thereby resulting in a single-phase material,
anhydrous RuO2, as seen in the XRD pattern.
Above 300 1C, the weight loss of the sample heated in air
is minimal up to 1000 1C, at which point a drastic drop in
weight is seen. This weight loss can be explained as
spontaneous decomposition of the oxide at a temperature
greater than 600 1C as the ruthenium–oxygen bonds break,
and the oxygen is liberated as O2 [38]. This decomposition
may also explain the existence of Ru(0) in the XRD
patterns of the samples heated in air, but not in the TGA,
owing to insufficient flow of oxygen that could drive the
This change in material composition after heating the
sample to different temperatures under different atmo-
spheres can be further evaluated from thermo-gravimetric
analysis (TGA) profiles. Fig. 5 shows the changes in the
mass fraction of the two samples upon heating under two

ARTICLE IN PRESS
J. Walker et al. / Journal of Solid State Chemistry 180 (2007) 2290–2297
2296
thermodynamic equilibrium completely towards the oxide,
as shown by
overall lower level of crystallization. This relative increase
in crystallinity for samples in both atmospheres is
confirmed by the corresponding XRD patterns.
T4600 o
RuO2 ꢃ! C RuðsÞ þ O2ðgÞ " .
(6)
This high-temperature weight loss can also be explained
by the reactions that occur when RuO2 is heated in the
presence of oxygen, as shown by [7]
4. Conclusions
The objective of this work was to evaluate the use of a
1,2-epoxide in the sol–gel synthesis of hydrous ruthenium
oxide from a ruthenium chloride solution. After verifying
that gelation successfully occurs 4 h after the addition of
the epoxide, characterization of the material was carried
out by XPS, XRD, DTA, and TGA. Through deconvolu-
tion of the C(1s)/Ru(3d) region, XPS established that the
synthesized gel material was a hydrous ruthenium oxide,
RuO2 ꢀ xH2O, which also contains residual organic species
not removed during the drying process. After heating the
hydrous ruthenium oxide in inert and oxygen-containing
atmospheres to temperatures up to 1400 1C, evaluation of
the XRD patterns showed that the as-prepared material
was amorphous and samples heated in inert atmospheres
were completely reduced to Ru(0) below 250 1C. XRD
patterns of samples heated in air revealed the existence of
anhydrous RuO2 as well as Ru(0), unless a continuous
supply of oxygen was present in the system, ensuring the
presence of the high oxygen partial pressure required to
maintain the oxide as the thermodynamically stable phase.
DTA traces show an endothermic loss of water below
150 1C, with exotherms for samples heated in both air and
argon owing to crystallization. The lower temperature
exotherm of the sample heated in argon is smaller than that
of the corresponding sample heated in air owing to the
competing endothermic decomposition of the hydrous
oxide into ruthenium metal and oxygen gas in the same
temperature interval. TGA data also illustrate that the
synthesized material contained approximately 45% organic
material by weight. This organic material is evolved below
250 1C with the complete reduction of the original material
to Ru(0) below 250 1C for the sample heated in argon. Also,
TGA demonstrates that the sample heated in air becomes
an anhydrous ruthenium oxide by this same temperature.
Further, high-temperature weight loss is attributed to the
well-known formation of gaseous higher valency ruthe-
nium oxide species as well as high-temperature thermal
decomposition of the RuO2 species into ruthenium metal
and oxygen gas.
Tꢁ800 o
RuO2ðsÞ þ O2ðgÞ ꢃ! C RuO4ðgÞ " ,
T41000 o
RuO2ðsÞ þ O2ðgÞ ꢃ! C RuO3ðgÞ " .
ð7Þ
1
2
In these reactions, higher valency gaseous ruthenium
oxide species form on the surface of the sample and are
eventually purged from the sample, resulting in a decrease
in sample weight at high temperatures. RuO3(g) is the
dominant species found at temperatures greater than
1000 1C and RuO4(g) is dominant around 800 1C [7]. As
the sample heated in argon does not show a decrease in
sample weight above 1000 1C, neither the ruthenium–
oxygen bond decomposition nor the high-temperature
formation of higher valency gaseous ruthenium oxide
species occurs, indicating a complete absence of oxygen in
the system and, therefore, the sole existence of Ru(0)
.
Fig. 6 shows the differential thermal analysis (DTA) data
of the samples heated in air and argon from room
temperature to 1400 1C. Both samples display small
endotherms below 150 1C resulting from the loss of water
[17]. The sample heated in air displays a sharp exotherm at
250 1C owing to crystallization of the as-synthesized
amorphous material as well as a broad exotherm starting
at 300 1C owing to a continued increase in crystallinity of
the sample over the entire temperature range. The sample
heated in argon displays a smaller exotherm at 235 1C
owing to the competing endothermic decomposition of the
hydrous oxide to Ru(0) with the exothermic crystallization
of these zero-valent atoms in this temperature range. As
with the sample heated in air, a broad exotherm beginning
at 300 1C can be seen. Its magnitude is smaller owing to an
5.0
4.0
Air
3.0
2.0
Argon
1.0
0.0
Acknowledgments
This work was supported by the Air Force/Bolling AFB/
DC MURI project on Energetic Structural Materials,
award No. F49620-02-1-0382. The authors are indebted to
Professors Thomas Sanders and Naresh Thadhani of
Georgia Institute of Technology for their insightful and
helpful comments.
0
300
600
900
1200
1500
Temperature (°C)
Fig. 6. Differential thermal analysis (DTA) data for the synthesized
hydrous ruthenium oxide heated at 10 1C/min to 1400 1C in two different
atmospheres, air and argon.

ARTICLE IN PRESS
J. Walker et al. / Journal of Solid State Chemistry 180 (2007) 2290–2297
2297
References
[20] V. Panic, A. Dekanski, V.B. Miskovic-Stankovic, S. Milonjic,
B. Nikolic, J. Electroanal. Chem. 579 (2005) 67–76.
[21] A.E. Gash, T.M. Tillotson, J.H. Satcher Jr., J.F. Poco, L.W.
Hrubesh, R.L. Simpson, Chem. Mater. 13 (2001) 999–1007.
[22] A.E. Gash, J.H. Satcher Jr., R.L. Simpson, Chem. Mater. 15 (2003)
3268–3275.
[1] I.-H. Kim, K.-B. Kim, Electrochem. Solid-State Lett. 4 (2001)
A62–A64.
[2] J.M.E. Harper, S.E. Hoernstroem, O. Thomas, A. Charai, L. Krusin-
Elbaum, J. Vac. Sci. Technol. A 7 (1989) 875–880.
[3] L. Krusin-Elbaum, M. Wittmer, D.S. Yee, Appl. Phys. Lett. 50
(1987) 1879–1881.
[23] L. Duraes, B.F.O. Costa, J. Vasques, J. Campos, A. Portugal, Mater.
Lett. 59 (2005) 859.
[24] J. Livage, M. Henry, C. Sanchez, Prog. Solid State Chem. 18 (1988)
259–341.
[4] M.M. Jevtic, E.V. Jelenkovic, K.Y. Tong, G.K.H. Pang, Thin Solid
Films 496 (2006) 214–220.
[5] D.A. McKeown, P.L. Hagans, L.P.L. Carette, A.E. Russell,
[25] B. Dobinson, W. Hofmann, B.P. Stark, The Determination of
Epoxides, Pergamon Press, Oxford, 1969.
K.E. Swider, D.R. Rolison, J. Phys. Chem.
4825–4832.
B 103 (1999)
[26] A.E. Gash, J.H. Satcher Jr., R.L. Simpson, J. Non-Crystal. Solids
350 (2004) 145–151.
[6] K.E. Swider-Lyons, C.T. Love, D.R. Rolison, J. Electrochem. Soc.
152 (2005) C158–C162.
[27] T.F. Baumann, A.E. Gash, S.C. Chinn, A.M. Sawvel, R.S. Maxwell,
J.H. Satcher Jr., Chem. Mater. 17 (2005) 395–401.
[28] A.E. Gash, T.M. Tillotson, J.H. Satcher Jr., L.W. Hrubesh, R.L.
Simpson, J. Non-Cryst. Solids 285 (2001) 22–28.
[29] E.A. Seddon, K.R. Seddon, The Chemistry of Ruthenium, Elsevier,
Amsterdam, 1984.
[7] K.S. Kim, N. Winograd, J. Catal. 35 (1974) 66–72.
[8] D.J. Suh, J. Non-Crystal. Solids 350 (2004) 314–319.
[9] A.J. Terezo, E.C. Pereira, Mater. Lett. 53 (4/5) (2002) 339–345.
[10] R. Sahul, V. Tasovski, T.S. Sudarshan, Sensors and Actuators A 125
(2006) 358–362.
[30] S. Schamm, R. Berjoan, P. Barathieu, Mater. Sci. Eng. B 107
(2004) 58.
[11] S. Bhaskar, P.S. Dobal, S.B. Majumder, R.S. Katiyar, J. Appl. Phys.
89 (2001) 2987–2992.
[31] D. Briggs, M.P. Seah, Practical Surface Analysis, Wiley, Chicester,
1983, p. 533.
[12] W. Sugimoto, K. Yokoshima, K. Ohuchi, Y. Murakami, Y. Takasu,
J. Electrochem. Soc. 153 (2006) A255–A260.
[32] P. Froment, M.J. Genet, M. Devillers, J. Electron Spectrosc. Relat.
Phenom. 104 (1999) 119.
[13] D.J. Suh, T.-J. Park, W.-I. Kim, I.-K. Hong, J. Power Sources 117
(2003) 1–6.
[33] D.R. Rolison, P.L. Hagans, K.E. Swider, J.W. Long, Langmuir 15
(1999) 774–779.
[14] M. Min, K. Machida, J.H. Jang, K. Naoi, J. Electrochem. Soc. 153
(2006) A334–A338.
[34] L. Ji, J. Lin, H.C. Zeng, Chem. Mater. 13 (2001) 2403–2412.
[35] R. Tannenbaum, C. Hakanson, A.D. Zeno, M. Tirrell, Langmuir 18
(2002) 5592–5599.
[15] J.P. Zheng, P.J. Cygan, T.R. Jow, J. Electrochem. Soc. 142 (1995)
2699–2703.
[16] E. Kolawa, F.C.T. So, W. Flick, X.A. Zhao, E.T. Pan, M.A. Nicolet,
Thin Solid Films 173 (1989) 217–224.
[36] P. Mills, J.L. Sullivan, J. Phys. D 16 (1983) 723–732.
[37] J. Pagnaer, D. Nelis, D. Mondelaers, G. Vanhoyland, J. D’Haen,
M.K. Van Bael, H. Van den Rul, J. Mullens, L.C. Van Poucke,
J. Eur. Ceram. Soc. 24 (2004) 919–923.
[17] I. Zhitomirsky, L. Gal-Or, Mater. Lett. 31 (1997) 155–159.
[18] S. Ardizzone, A. Daghetti, L. Franceschi, S. Trasatti, Colloids Surf.
35 (1989) 85–96.
[38] P.F. Campbell, M.H. Ortner, C.J. Anderson, Anal. Chem. 33 (1961)
58–61.
[19] H.B. Suffredini, V. Tricoli, L.A. Avaca, N. Vatistas, Electrochem.
Commun. 6 (2004) 1025–1028.