Tetragonal-(Zr,Co)O2 solid solution: Combustion synthesis, thermal stability in air and reduction in H2, H2-CH4 and H2-C2H4 atmospheres

Article abstract of DOI:10.1016/j.materresbull.2007.11.008

The synthesis of Co²⁺-stabilized zirconia by the nitrate/urea combustion route is investigated. Using seven times the so-called stoichiometric urea proportion allows to obtain for the first time the Zr_0.9Co_0.1O_1.9 solid solution fully stabilized in tetragonal form. The thermal stability in air and the reduction in H₂, H₂-CH₄ and H₂-C₂H₄ atmospheres are studied. The carbon forms obtained upon reduction are investigated. Reduction in H₂-CH₄ produces many carbon species including short carbon nanofibers, nanoribbons, hollow particles often forming bamboo structures, carbon-encapsulated Co particles and carbon nanotubes. Reduction in H₂-C₂H₄ produces 15-30 nm nanofibers.

Full text of DOI:10.1016/j.materresbull.2007.11.008

Materials Research Bulletin 43 (2008) 3088–3099
www.elsevier.com/locate/matresbu
Tetragonal-(Zr,Co)O solid solution: Combustion synthesis,
2
thermal stability in air and reduction in H , H –CH
2
2
4
and H –C H atmospheres
2
2 4
*
Felipe Legorreta Garcia, Alain Peigney, Christophe Laurent
Institut Carnot CIRIMAT, UMR CNRS-UPS-INP 5085, Université Paul-Sabatier, 31062 Toulouse Cedex 9, France
Received 16 April 2007; received in revised form 8 November 2007; accepted 10 November 2007
Available online 19 November 2007
Abstract
The synthesis of Co -stabilized zirconia by the nitrate/urea combustion route is investigated. Using seven times the so-called
2
+
stoichiometric urea proportion allows to obtain for the first time the Zr0.9Co0.1 1.9 solid solution fully stabilized in tetragonal form.
O
The thermal stability in air and the reduction in H , H –CH and H –C H atmospheres are studied. The carbon forms obtained upon
2
2
4
2
2 4
reduction are investigated. Reduction in H –CH produces many carbon species including short carbon nanofibers, nanoribbons,
2
4
2 2 4
hollow particles often forming bamboo structures, carbon-encapsulated Co particles and carbon nanotubes. Reduction in H –C H
produces 15–30 nm nanofibers.
#
2007 Elsevier Ltd. All rights reserved.
Keywords: A. Oxides; A. Composites; A. Nanostructures; B. Chemical synthesis
1
. Introduction
Cobalt oxides (CoO and Co O ) particles dispersed on the surface of zirconia (ZrO ) substrates are interesting
3
4
2
catalysts [1–8]. However, the incorporation of cobalt ions into the ZrO lattice and its possible role in the stabilization
2
of the high-temperature forms of ZrO is seldom mentioned. Indeed, pure ZrO is found as a monoclinic phase (m-
2
2
ZrO ) at room temperature, but it exhibits structural transformations upon an increase of the temperature, transforming
2
into tetragonal zirconia (t-ZrO ) at approximately 1170 8C and cubic zirconia (c-ZrO ) around 2370 8C [9]. It is well
2
2
2
+
2+
3+
known that metal cations (notably Ca , Mg , Y and some rare earth cations) stabilize either the tetragonal or cubic
form, at increasing contents. The substitution of Zr cations by lower valence cations is balanced by the formation of
oxygen vacancies. The influence of cobalt as a stabilizer was comparatively less studied. Lick et al. [7], using an
4
+
2
+
impregnation route, have reported that Co ions stabilizes the tetragonal form and they further proposed that the
3
+
incorporation of Co ions is less probable due to its very low ionic radius (54.5 pm). However, even for the low cobalt
fractions (0.2 and 2 wt%) used by these authors, the specimens still contain appreciable amounts of monoclinic phase
and it was proposed that higher loadings would result in the segregation of Co O . By contrast, Choudhary et al. [6]
3
4
used a coprecipitation/calcination route to prepare samples with a higher cobalt loading and reported a stabilized cubic
solid solution for a Co/Zr mole ratio equal to 0.25. Maglia et al. [10] prepared mixtures of the cubic and monoclinic
*
Corresponding author. Tel.: +33 5 61 55 61 22; fax: +33 5 61 55 61 63.
E-mail address: laurent@chimie.ups-tlse.fr (C. Laurent).
0025-5408/$ – see front matter # 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.materresbull.2007.11.008

F. Legorreta Garcia et al. / Materials Research Bulletin 43 (2008) 3088–3099
3089
forms of zirconia by reacting metallic Zr and CoO. The respective proportion of these phases depends on the reaction
temperature, but no information was given on the cobalt content in each phase. A study of the thermal stability in O2
suggested that the initially cubic phase may transform into the tetragonal form upon heating and then forms m-ZrO2
together with an unspecified cobalt oxide. The process is all the more pronounced when the calcination temperature is
increased in the range 800–1050 8C.
In the present paper, it is proposed to investigate this topic by preparing a stabilized Zr0.9Co0.1O1.9 solid solution
using the nitrate/urea combustion route, which uses cheap reactants, produces gram-scale quantities of powders at the
laboratory scale and can be readily upscaled. Ringuede et al. [11] have used this method for the synthesis of mixtures
of CoO and Co O with yttria-stabilized zirconia (50/50 volume fractions). The present authors have investigated the
3
4
influence of the urea proportion on the formation of a nanocrystalline tetragonal Zr0.9Fe0.1O1.95 solid solution [12]. In a
second part of the study, the thermal stability in air of a t-Zr0.9Co0.1O1.9 solid solution is investigated. Finally, the
selective reduction of the tetragonal-stabilized Zr0.9Co0.1O1.9 solid solution in H , H –CH and H –C H atmospheres
2
2
4
2
2 4
is studied with the aim to evaluate the formation of cobalt nanoparticles and thus the potential of such a material for the
formation of carbon nanotubes (CNTs). Indeed, this method has given very interesting results in the case of
Mg1 Co Al O [13] and Mg Co O [14] solid solutions. To the best of our knowledge, only some results using the
Àx
x
2
4
1Àx
x
Fe–ZrO [15] and Ni–ZrO [16,17] systems have been reported in that respect.
2
2
2
. Experimental
2
.1. Combustion synthesis
The appropriate amounts of Co(NO ) Á9H O and ZrO(NO ) ÁxH O were dissolved in deionized water (20 mL,
3
3
2
3 2
2
7
0 8C) with varying amounts of urea, in order to produce 5 g of Zr Co
O
0.1 1.9
solid solution. The urea proportions
0
.9
were calculated by considering the total oxidizing valencies (V ) and reducing valencies (V ) of the different species
O
R
[
18]. The total oxidizing valencies of the nitrates is V = 10.5. The reducing valence of urea is V = 6. The molar
O R
quantity of urea (m) is first calculated so that the so-called stoichiometric ratio F = mV /V is equal to unity, which
e
R
O
gives m = 1.75. However, as discussed by Zhang and Stangle [19], this method involves some approximations. Firstly,
O from the air atmosphere is not taken into account, which affects V . Secondly, the nitrogen-containing products are
2
O
considered as being only N (thus, zero is taken as the valence of nitrogen element) despite that nitrogen oxides and/or
2
NH can be formed in usually undetermined proportions. This affects both V and V . Hence it is necessary to
3
O
R
experimentally determine the urea proportion suitable for a particular synthesis and therefore the present syntheses
were performed for 10 different values of F (1–10), i.e. different values of m (1.75–17.5). However, for the sake of
e
simplicity, we use the integer n (n = 1, 2, 3, 4, 5, 6, 7, 8, 9 and 10), which represents the multiplying factor rather than
the molar quantity. The dish containing the solution was placed in a furnace pre-heated at 600 8C, keeping the door of
the furnace open. After water evaporation, a combustion reaction takes place according to a redox reaction between the
nitrates and urea, producing an oxide powder (henceforward named the as-prepared powder). The as-prepared
powders will be designated as ZC-nU, with n = 1, 2, 3, 4, 5, 6, 7, 8, 9 and 10.
2
.2. Thermal stability of the Zr0.9Co0.1O1.9 solid solution in air
A selected specimen (ZC-7U) was chosen to perform the study of the thermal stability in air. Differential thermal
analysis (DTA Netzsch 404S) was performed in flowing air (heating and cooling rates 20 8C/min up to 1300 8C).
Batches of 80 mg of the powder were calcined in flowing air (4.8 L/h) at 625, 725, 825, 855, 880, 900, 925 and
1
030 8C, yielding powders henceforward referred to as C625, . . ., C1030. A heating and cooling rate of 20 8C/min was
applied (no dwell time at the target temperature).
2
.3. Reduction of the Zr0.9Co0.1O1.9 solid solution in H , H –CH and H –C H
2 2 4 2 2 4
Batches of 1.5 g of the ZC-7U powder were reduced in H atmosphere at 600, 700, 800, 900, 950 and 1000 8C,
2
producing Co–ZrO nanocomposite powders henceforward referred to as R600, . . ., R1000. Batches of 1.5 g of the
2
powder were also reduced in H –CH atmosphere (20 mol.% CH , 1000 8C) and in H –C H atmosphere (40 mol.%
2
4
4
2
2 4

3090
F. Legorreta Garcia et al. / Materials Research Bulletin 43 (2008) 3088–3099
C H , 650 8C). For all reductions, a heating and cooling rate of 5 8C/min was applied with no dwell time at the target
2
4
temperature. The total gas flow rate was equal to 15 L/h and the gas was dried on P O .
2
5
2
.4. Characterization
Room temperature powder X-ray diffraction (XRD) patterns were obtained with a Brüker D4 endeavor
diffractometer using Cu Ka radiation. The proportions of tetragonal and monoclinic zirconia phases were evaluated
from the XRD patterns using a method similar to that reported by Garvie and Nicholson [20]. Specific surface area
measurements were performed using N adsorption at liquid N temperature in a Micromeritics Flow Sorb II 2300
2
2
apparatus (BET method). This instrument gives a specific surface area value from one point (i.e. one adsorbate
pressure) and requires calibration. The reproducibility of the results was determined to be in the Æ3% range. The
carbon content in the as-prepared powders and in the powders obtained after reduction in H –CH and H –C H was
2
4
2
2 4
determined by flash combustion with an accuracy of Æ2%. Selected oxide powders were observed by scanning
electron microscopy (SEM JEOL JSM 6400). Selected reduced powders were observed by field-emission gun
scanning electron microscopy (JEOL JSM 6700F) and transmission electron microcopy (JEOL 200CX).
3
. Results and discussion
3
.1. Combustion synthesis
The dish containing the nitrate/urea solution is introduced in the furnace pre-heated at 600 8C. Water evaporation is
observed, the furnace temperature diminishing by about 40 8C before increasing again to 600 8C. Then the combustion
takes place according to a redox reaction between the nitrates and urea. Neither red fumes characteristic of NO2
formation nor the characteristic smell of ammonia was detected for any synthesis. For ZC-1U, the combustion reaction
is not violent and is accompanied by a red flame. For ZC-2U, and notably for ZC-3U, the reaction is much more violent
and is accompanied by a strong yellow flame, whereas for ZC-4U, the reaction is milder but still with a yellow flame.
All four combustion products are not homogeneous in color. For samples with n = 5–10, the process is much milder
and there is simply an incandescence of the material and no flame. All six combustion products are homogeneous in
color. The measured temperature of the furnace during the combustion (Fig. 1a) varies very little except for n = 3,
reflecting the visual observations. Note that the maximum temperature reached in the vessel is probably higher than
that recorded for the furnace. The total time from dish introduction to end of reaction increases with n (Fig. 1b). There
is a gap between ZC-4U (2.5 min) and ZC-5U (6.8 min). This time increase reflects the increase of the water
evaporation duration due to a more sluggish solution when the urea content is increased and also the fact that the
process is slower when there is no flame (i.e. for n = 5–10). Similar results have been obtained for the synthesis of
Zr0.9Fe0.1O1.95 [12].
It was verified that the as-prepared powders contain only trace amounts of carbon. The BET specific surface area
(
S , Fig. 2) tends to increase when the urea content is increased, but it could also be possible to separate the specimens
w
2
2
into two groups, with S in the range 7–17 m /g for n = 1–5 and S in the range 23–30 m /g for n = 6–10. These groups
w
w
almost match those observed above except that powder ZC-5U now falls in the first one. The higher S for the second
w
group of powders could reflect a smaller grain size, due to both the milder nature of the combustion process and the
increasing expansion of the powder due to the higher release of gases for higher urea contents.
The XRD patterns of the as-prepared powders are shown in Fig. 3. Tetragonal zirconia is the major phase detected
for n = 1–4 and the only one for n = 5–10. Monoclinic zirconia is also detected, extremely faintly for ZC-1U and ZC-
4
U but very clearly for ZC-2U and ZC-3U, where it accounts for 40 and 23%, respectively, with respect to all zirconia
phases. Avery wide (3 1 1) peak of Co O is faintly detected for ZC-2U, ZC-3U and ZC-4U, and also possibly for ZC-
3
4
9
U and ZC-10. The (1 0 1) peak of the tetragonal phase appears to be slightly shifted to the high 2u-side for ZC-7U
2
+
compared to the other powders, which could reveal a higher Co content in the tetragonal zirconia phase for ZC-7U. It
is further noticed that the diffraction peaks of the tetragonal phase are quite narrow for n = 1–3 and much broader for
n = 4–10, ZC-7U showing the broadest peaks. This could reflect a trend for a decreasing crystallite size upon the
increase in urea proportion, in agreement with the BET S data.
w
Typical low-magnification SEM images of the as-prepared powders are shown in Fig. 4a and b. The powders are
made up of large grains, measuring several tens of micrometers across, as well as micro- or submicrometric grains. It

F. Legorreta Garcia et al. / Materials Research Bulletin 43 (2008) 3088–3099
3091
Fig. 1. Temperature reached by the furnace during combustion in a furnace pre-heated at 600 8C (a) and time for completion of reaction (b) vs. the
proportion of urea (n) used for the combustion synthesis.
was observed in the case of (Zr,Fe)O powders [12], the proportion of the latter ones tends to increase upon the
2
increase of the urea proportion due to the higher expansion occurring during the combustion reaction. Furthermore, the
zirconia grains are quite porous, reflecting the escaping of gases during the combustion. Higher magnification SEM
images (Fig. 4c and d) illustrate typical areas found in the powders: fairly dense flat surfaces, partly broken due to the
escaping of gases (Fig. 4c) and much more porous areas, revealing crystallites not larger than 100 nm (Fig. 4d). A
previous study [12] revealed a trend towards a decrease in the average crystallite size upon the increase of the urea
proportion. This is in broad agreement with the present SEM, XRD and S results.
w
Using a low urea proportion (n = 1–4) produces powders which are neither homogeneous samples nor monophased
zirconia, possibly because the maximal temperature reached during the combustion is not the same throughout
the sample. For higher values of n (n = 5–8), the milder and significantly longer combustion allows to obtain the
Zr0.9Co0.1O1.9 solid solution stabilized in the tetragonal form, no free cobalt oxide being detected. It is possible that the
crystallite size of the zirconia species plays an important role as larger crystallites will more easily end up in the form
of the stable monoclinic form. Increasing the urea proportion favors the emission of gases and therefore the expansion
of the solid phase, which favors the formation of smaller crystallites. This could explain why no monoclinic zirconia is
detected for n = 5–10. These experimental conditions could also favor the formation of suitable gel-like species by
formation of complex of the zirconium and cobalt ions with urea, which could in turn trigger the direct incorporation of

3092
F. Legorreta Garcia et al. / Materials Research Bulletin 43 (2008) 3088–3099
Fig. 2. Specific surface area of the as-prepared powders vs. the proportion of urea (n) used for the combustion synthesis.
Fig. 3. XRD patterns of the as-prepared powders. The proportion of urea used for the combustion is indicated as 1U, 2U, . . ., 10U.
the cobalt ions in the zirconia phase and therefore favor crystallization in stabilized form at moderate temperature.
Further studies, notably on all intermediate species, would be required to totally elucidate the involved mechanisms.
The powder ZC-7U, which is considered to be the desired tetragonal Zr0.9Co0.1O1.9 solid solution, was selected for
further studies.

F. Legorreta Garcia et al. / Materials Research Bulletin 43 (2008) 3088–3099
3093
Fig. 4. SEM images of some as-prepared powders synthesized with different urea proportions (n): (a) n = 4; (b) n = 7; (c) n = 4; (d) n = 10.
3
.2. Thermal stability of the tetragonal Zr0.9Co0.1O1.9 solid solution in air
The thermal stability of the tetragonal Zr0.9Co0.1O1.9 solid solution was investigated by DTA. The DTA curve (not
shown) revealed endothermic peaks at 940 and 1194 8C upon heating and exothermic peaks at 915 and 800 8C upon
cooling. The DTA curve of a cobalt-free tetragonal ZrO powder prepared in the same conditions for the sake of
2
comparison showed an endothermic peak at 1193 8C upon heating and a exothermic peak at 857 8C upon cooling. The
room temperature XRD patterns of the powders after calcination in air at different temperatures (625, 725, 825, 855,
8
80, 900, 925 and 1030 8C) are shown in Fig. 5. For powders C625 and C725, only the peaks of the tetragonal solid
solution are detected, as in the as-prepared powder. For higher calcination temperatures, Co O and m-ZrO are
3
4
2
detected as well, the proportion of the latter phase markedly increasing at the expense of the tetragonal phase upon the
increase of calcination temperature (Fig. 6). For C625–C855, the (1 0 1) peak accounting for the tetragonal phase is
progressively more narrow, indicating some crystallite growth. Moreover, for C625, C725 and C825, the position of
this peak gradually shifts to lower 2u angles, i.e. revealing that the corresponding inter-reticular distance is
2
+
progressively higher, which could indicate that the tetragonal phases contains less and less Co ions. This could reveal
that the phase partitioning from the cobalt-stabilized tetragonal solid solution to the mixture of m-ZrO and Co O
3 4
2
takes place at temperatures as low as 625 8C. Free Co O could remain undetected on the XRD patterns because of a
3
4
very small size. The presence of free Co O thus accounts for the DTA peaks at 940 8C (heating) and 915 8C (cooling),
3
4
which represent the reversible transformation of this phase into CoO. This is in agreement with previous results [14].
The DTA peaks at 1194 8C (heating) and 800 8C (cooling) therefore, respectively represent the monoclinic-to-
tetragonal (heating) and tetragonal-to-monoclinic (cooling) transformations of ZrO . As could be expected, the
2
specific surface area of the powders decreases with increasing calcination temperature (ca. 24, 12.7, 3.5, 3.1, 1.8 and
2
1
.4 m /g for the as-prepared, C625, C725, C825, C855 and C1030 powders, respectively). The main part of the
decrease principally takes place prior to both the formation of Co O in detectable amounts and the transformation of
3
4
the zirconia phase and could reflect the crystallite growth of the tetragonal phase, as indicated by XRD data. The above
data show that the thermal stability in air of the Zr0.9Co0.1O1.9 solid solution is lower than that of a Zr0.9Fe0.1O1.95 solid
solution, which was shown [12] to be stable up to about 875 8C.

3094
F. Legorreta Garcia et al. / Materials Research Bulletin 43 (2008) 3088–3099
Fig. 5. Room temperature XRD patterns of the powders after calcination in air at different temperatures.
3
.3. Reduction of the Zr0.9Co0.1O1.9 solid solution in H , H –CH and H –C H
2 2 4 2 2 4
The tetragonal Zr0.9Co0.1O1.9 solid solution was reduced in H atmosphere at 600, 700, 800, 900, 950 and 1000 8C
2
producing Co–ZrO nanocomposite powders (R600, . . ., R1000). The XRD pattern (not shown) of R600 is identical to
2
that of the starting powder, i.e. only the tetragonal solid solution is detected. For R700 and R800, the (1 1 1) peak of
face-centered-cubic cobalt (e-Co) is very faintly detected besides the tetragonal solid solution. For powders R900,
R950 and R1000, the e-Co(1 1 1) peak is detected with an increasing intensity upon the increase in reduction
temperature and m-ZrO is also detected. The latter phase represents 14, 67 and 83% of all zirconia phases for R900,
2
R950 and R1000, respectively, which is slightly less to what is observed upon calcinations in air (Fig. 6). These results
2
+
could reflect that the reduction of the Co ions to form metallic Co, in H atmosphere, takes place at slightly higher
2
2
+
3+
temperatures than does the phase partitioning in air because the latter involves the oxidation of Co in Co which is
probably easy. It is nevertheless probable that some Co nanoparticles are formed when the reduction is performed at
6
00 8C, even though e-Co is not detected on the XRD pattern for R600. Indeed, a Mössbauer spectroscopy study of the

F. Legorreta Garcia et al. / Materials Research Bulletin 43 (2008) 3088–3099
3095
Fig. 6. Proportion of monoclinic zirconia (with respect to all zirconia phases) evaluated from the XRD patterns vs. the calcination temperature.
powder formed by H reduction at 600 8C of a tetragonal Zr0.9Fe0.1O1.95 solid solution revealed the presence of a-Fe,
2
which was not detected on the corresponding XRD pattern despite a size in the range 10–50 nm as evaluated from
2
+
FEG-SEM observations [21]. Moreover, Choudhary et al. [6] showed that mobility of lattice oxygen in Co -stabilized
3
+
cubic zirconia is higher than in Fe -stabilized cubic zirconia, thus favoring the reduction to the metallic phase.
The tetragonal Zr Co O solid solution was reduced in H –CH atmosphere (20 mol.% CH ) at 1000 8C. The
0
.9
0.1 1.9
2
4
4
XRD pattern (not shown) is similar to that of powder R1000, monoclinic zirconia representing 88% of all zirconia phases
andthee-Co(1 1 1)peakshowinganintensitycomparabletothatobtainedforR1000. Nopeak correspondingtocarbonis
2
detected. The carbon content is equal to 3.7 wt%. The specific surface is equal to 7.9 m /g. FEG-SEM and TEM
observations (Fig. 7) reveal many different carbon species, including short carbon nanofibers, nanoribbons, hollow
particles often forming bamboo structures, carbon-encapsulated Co particles and CNTs, albeit in a fairly low proportion.
This is by and large similar to what is reported when using a ZrO -supported Fe catalyst [15]. The selectivity towards the
2
formation of CNTs is far less than what is obtained using a Mg Co O solid solution as starting material, the reduction
0
.9
0.1
of which provokes mainly the formation of single- and double-walled CNTs [14] and less than what is obtained using a
Mg Co Al O solid solution as starting material [13]. The formation of so many different carbon species upon
0
.9
0.1
2 4
reduction is known to reflect that the catalytic material is prone to the formation of metal particles with a large or even
multimodal size distribution and a fairly high average size [13,22,23], well above the 3–5 nm cutoff size [24,25] for the
formation of CNT. The propensity of the present catalytic material to form too large Co particles is thought to arise
2
+
because of the relatively poor thermal stability of the solid solution and possibly because the Co ions are not diluted
enough in the zirconia lattice. In both cases, using lower cobalt content could be beneficial.
In the following part of the study, it was attempted to perform the reduction at a much lower temperature (650 8C) in
2
+
order to obtain a lower amount of metallic phase through a lower reduction rate of the Co ions, which would favor the
formation of smaller Co particles. The lower temperature would also help in avoiding their coalescence. Since the
catalytic decomposition of CH is fairly low at 650 8C, the tetragonal Zr0.9Co0.1O1.9 solid solution was reduced in H2–
4
C H atmosphere (40 mol.% C H ). The XRD pattern (not shown) presents the peaks of tetragonal zirconia as well as
2
4
2 4
a very wide e-Co(1 1 1) peak with a very weak intensity. No peak corresponding to carbon is detected. The carbon
2
content is equal to 16.3 wt% and the specific surface is equal to 36.3 m /g. Interestingly, both values are over four
times what is obtained for the powder reduced in H –CH at 1000 8C. At least part of specific surface area increase
2
4
2
compared to the starting solid solution (23.7 m /g) could correspond to the formation of some filamentous carbon
species. FEG-SEM and TEM observations (Fig. 8) reveal the presence of carbon filaments with a diameter in the range
1
5–30 nm. Although some larger nanofibers can be observed in some areas, the powder reduced in H –C H is quite
2 2 4
different and much more homogeneous than the one reduced in H –CH . Some filaments show some degree of
2
4
branching and some of them do not appear to be hollow (Fig. 8c). The tip of some filaments contains a particle,
probably cobalt, appearing as dark dots in Fig. 8c. These carbon filaments resemble those obtained by heat-treatment
of Fe/Ni-Al O catalysts at 650 8C in N –C H (17%C H ) [26] or Fe O -based catalysts in H –C H [27], although
2
3
2
2
4
2
4
2
3
2
2 4

3096
F. Legorreta Garcia et al. / Materials Research Bulletin 43 (2008) 3088–3099
2 4
Fig. 7. FEG-SEM images (a, b) and TEM image (c) of the powder prepared by reduction in H –CH at 1000 8C.
their diameter is significantly smaller (15–30 nm vs. over 100 nm). Those filaments that are not hollow could be
identified as the so-called platelet nanofibers, where the graphene sheets are parallel to each other and are
perpendicular to the direction of the axis of the filament [28]. Interestingly, McCaldin et al. [27] showed that the
formation of the various carbon forms (platet nanofibers, herrinbone nanofibers, multi-walled CNTs and encapsulated

F. Legorreta Garcia et al. / Materials Research Bulletin 43 (2008) 3088–3099
3097
2 2 4
Fig. 8. FEG-SEM images (a, b) and TEM image (c) of the powder prepared by reduction in H –C H at 650 8C.

3098
F. Legorreta Garcia et al. / Materials Research Bulletin 43 (2008) 3088–3099
metal particles) was strongly depending on the temperature (400–800 8C) and H –C H gas mixture composition. The
2
2 4
present experimental conditions (650 8C, 40% C H ) are just outside the temperature-composition space (both on the
2
4
high-side) for which they report the formation of the platelet nanofibers, but this probably reflects the difference in
catalytic material. Thus, in order to principally obtain the carbon as multi-walled CNTs, future studies will be devoted
to fine tuning some materials characteristics such as the nature of the catalyst metal (Co, Fe, Ni or alloys), its
proportion with respect to ZrO and the stabilization of ZrO by a non-reactive oxide such as Y O . Some reaction
2
2
2 3
parameters such as temperature, duration and gas composition and flow rate will also be investigated. As noted by
Ferlauto et al. [16], such CNT–metal–ZrO composites could be interesting due to mixed electronic-ionic transport
2
properties.
4
. Conclusions
The synthesis of a Zr0.9Co0.1O1.9 solid solution fully stabilized in tetragonal form is reported for the first time. The
nitrate/urea combustion route was used and it is shown that the appropriate amount of urea is seven times the so-called
stoichiometric proportion. A study of the thermal stability in air revealed that the phase partitioning from the cobalt-
stabilized tetragonal solid solution to the mixture of m-ZrO and Co O takes place at temperatures as low as 625 8C. A
2
3 4
study of the reduction of the Zr0.9Co0.1O1.9 solid solution in H , H –CH and H –C H atmospheres indicated that the
2
2
4
2
2 4
Co particles formed upon reduction are probably of a too large size, and size distribution, to be selective towards the
formation of single- and double-walled CNTs. In particular, reduction at 1000 8C in H –CH produced many carbon
2
4
species including short carbon nanofibers, nanoribbons, hollow particles often forming bamboo structures, carbon-
encapsulated Co particles and CNTs. By contrast, reduction at 650 8C in H –C H produced 15–30 nm nanofibers. For
2
2 4
at least some of them, the graphene sheets are parallel to each other and are perpendicular to the direction of the axis of
the filament. Future studies will be devoted to fine tuning both some materials characteristics and some reaction
parameters in order to improve the selectivity towards CNTs.
Acknowledgments
The authors would like to thank Mr. L. Datas and Mr. L. Weingarten for their assistance in the TEM observations,
which were performed at TEMSCAN, the ‘‘Service Commun de Microscopie Electronique à Transmission’’,
Université Paul-Sabatier.
References
[
[
[
[
[
[
[
[
[
1] L.A. Boot, M.H.J.V. Kerkhoffs, B.T. van der Linden, A.J. van Dillen, J.W. Geus, F.R. van Buren, Appl. Catal. A 137 (1996) 69–86.
2] T.-C. Xiao, S.-F. Ji, H.-T. Wang, K.S. Coleman, M.L.H. Green, J. Mol. Catal. A 175 (2001) 111–123.
3] V.G. Milt, M.A. Ulla, E.A. Lombardo, J. Catal. 200 (2001) 241–249.
4] V.G. Milt, E.A. Lombardo, M.A. Ulla, Appl. Catal. B 37 (2006) 63–73.
5] M. Kantcheva, A.S. Vakkasoglu, J. Catal. 223 (2004) 364–371.
6] V.R. Choudhary, S. Banerjee, S.G. Pataskar, Appl. Catal. A 253 (2003) 65–74.
7] I.D. Lick, A. Carrascull, M. Ponzi, E.N. Ponzi, I.L. Botto, Mater. Chem. Phys. 92 (2005) 327–332.
8] P. Soisuwan, P. Praserthdam, J. Panpranot, D.L. Trimm, Catal. Commun. 7 (2006) 761–767.
9] M. Yashima, T. Hirose, S. Katano, Y. Suzuki, M. Kakihana, M. Yoshimura, Phys. Rev. B 51 (1995) 8018–8025.
[
[
[
[
[
[
[
[
[
[
[
[
[
10] F. Maglia, U. Anselmi-Tamburini, G. Spinolo, A.J. Munir, Mater. Synth. Process. 7 (1999) 327–332.
11] A. Ringuede, J.A. Labrincha, J.R. Frade, Solid-State Ionics 141–142 (2001) 549–557.
12] F. Legorreta Garcia, V. Gonzaga de Resende, E. De Grave, A. Peigney, A. Barnabé, C. Laurent, Solid-State Ionics, submitted for publication.
13] A. Govindaraj, E. Flahaut, C. Laurent, A. Peigney, A. Rousset, C.N.R. Rao, J. Mater. Res. 14 (1999) 2567–2576.
14] E. Flahaut, A. Peigney, C. Laurent, A. Rousset, J. Mater. Chem. 10 (2000) 249–252.
15] M.A. Ermakova, D.Y. Ermakov, A.L. Chuvilin, G.G. Kuvshinov, J. Catal. 201 (2001) 183–197.
16] A.S. Ferlauto, D.Z. Deflorio, F.C. Fonseca, V. Esposito, R. Muccillo, E. Traversa, L.O. Ladeira, Appl. Phys. A 84 (2006) 271–276.
17] S. Kurasawa, S. Iwamoto, M. Inoue, Mol. Cryst. Liq. Cryst. 387 (2002) 123–128.
18] K.C. Patil, Bull. Mater. Sci. 16 (1993) 533–541.
19] Y. Zhang, G.C. Stangle, J. Mater. Res. 9 (1994) 1997–2004.
20] R.C. Garvie, P.S. Nicholson, J. Am. Ceram. Soc. 55 (1972) 303–305.
21] V. Gonzaga de Resende, F. Legorreta Garcia, A. Peigney, E. De Grave, C. Laurent, Solid-State Ionics, submitted for publication.
22] Ch. Laurent, A. Peigney, A. Rousset, J. Mater. Chem. 8 (1998) 1263–1271.

F. Legorreta Garcia et al. / Materials Research Bulletin 43 (2008) 3088–3099
3099
[
[
23] P. Coquay, A. Peigney, E. De Grave, R.E. Vandenberghe, C. Laurent, J. Phys. Chem. B 106 (2002) 13199–13210.
24] J.H. Hafner, M.J. Bronikowski, B.R. Azamian, P. Nikolaev, A.G. Rinzler, D.T. Colbert, K.A. Smith, R.E. Smalley, Chem. Phys. Lett. 296
1998) 195–202.
(
[25] A. Peigney, P. Coquay, E. Flahaut, R.E. Vandenberghe, E. De Grave, Ch. Laurent, J. Phys. Chem. B 105 (2001) 9699–9710.
[26] W. Qian, T. Liu, Z. Wang, H. Yu, Z. Li, F. Wei, G. Luo, Carbon 41 (2003) 2487–2493.
[27] S. McCaldin, M. Bououdina, D.M. Grant, G.S. Walker, Carbon 44 (2006) 2273–2280.
[28] W.B. Downs, R.T.K. Baker, J. Mater. Res. 10 (1995) 625–633.