Obtaining Ni nanoparticles on 3Y-TZP powder from nickel salts

Article abstract of DOI:10.1111/j.1551-2916.2005.00676.x

Non-agglomerated and homogeneously dispersed nickel nanoparticles were obtained on the surface of zirconia particles from the solution of three different nickel salts (nitrate, chloride, and acetate) in ethanol. After a drying process, the powders were calcined and reduced. Transmission electron microscopy studies only showed a uniform distribution of Ni nanoparticles on the zirconia grain surfaces in the case of nickel obtained from nitrate salt. The analysis of the magnetic hysteresis loop of these particles has revealed the absence of Ni particles larger than 50 nm. Mechanisms to justify the different Ni particle sizes obtained from several salts have been proposed based on X-ray photoelectron and infrared spectroscopy. As a result, only in the case of a good zirconia/nickel nanostructure in the green samples and using nitrate salt dense (98% th.) nanocrystalline ceramic composites were obtained as a precursor.

Full text of DOI:10.1111/j.1551-2916.2005.00676.x

J. Am. Ceram. Soc., 89 [1] 144–150 (2006)
DOI: 10.1111/j.1551-2916.2005.00676.x
r 2005 The American Ceramic Society
ournal
J
Obtaining Ni Nanoparticles on 3Y-TZP Powder from Nickel Salts
Fatima Esteban-Betegon, Sonia Lopez-Esteban, Joaquın Requena, Carlos Pecharroman, and Jose
´ ´ ´ ´ ´
S. Moya**,w
Instituto de Ciencia de Materiales de Madrid, CSIC, 28049 Madrid, Spain
Jose C. Conesa
´
Instituto de Catalisis y Petroleoquımica, CSIC, 28049 Madrid, Spain
´ ´
Non-agglomerated and homogeneously dispersed nickel nano-
particles were obtained on the surface of zirconia particles from
the solution of three different nickel salts (nitrate, chloride, and
acetate) in ethanol. After a drying process, the powders were
calcined and reduced. Transmission electron microscopy studies
only showed a uniform distribution of Ni nanoparticles on the
zirconia grain surfaces in the case of nickel obtained from ni-
trate salt. The analysis of the magnetic hysteresis loop of these
particles has revealed the absence of Ni particles larger than 50
nm. Mechanisms to justify the different Ni particle sizes ob-
tained from several salts have been proposed based on X-ray
photoelectron and infrared spectroscopy. As a result, only in the
case of a good zirconia/nickel nanostructure in the green sam-
ples and using nitrate salt dense (98% th.) nanocrystalline ce-
ramic composites were obtained as a precursor.
The choice of the system ZrO2/Ni is not arbitrary. In fact, the
thermal expansion coefficients of zirconia and nickel are nearly
identical so that the composites do not generate thermal stresses.
Zirconia ceramics are materials with fundamental properties re-
sponsible for the development of both functional and structural
ceramics of interest for engineers and designers. In addition, in
spite of the wide field of applications of zirconia ceramics in its
pure form, conductive particles incorporated into such an insu-
lating matrix yield new materials with appealing properties.
In previous works, different yttria-stabilized tetragonal zir-
conia polycrystals (Y-TZP)/nickel composites were prepared
following a wet processing pressureless sintering route, with Ni
particle size in the micrometer range. A wide range of metal
volume fractions were studied14,15 (10–40 vol%). These samples
had a large fraction of porosity (5–15 vol%). There are methods
to reduce this porosity, such as hot-press sintering.16 However,
this occurs at the cost of a large increase in nickel particle size
because of the superplastic character of Y-TZP at the sintering
temperatures. In this context, in the present work, we propose
an alternative route to obtain dense zirconia/Ni nanocomposites
by a conventional route starting from Ni salt precursors.
I. Introduction
ERAMIC/METAL composite materials (cermets) have attracted
significant attention because of a singular combination of
C
properties (including mechanical, electrical, and magnetic) that
makes them excellent candidates for the fabrication of multi-
functional devices. Moreover, the fact that one of the phases is
nanostructured substantially modifies the mechanical,1–3 electri-
cal,4 magnetoresistive,5 and magnetic properties.6–8
II. Experimental Procedure
(1) Starting Materials
The following commercially available salts have been used as
nickel source materials: nickel (II) nitrate hexahydrate (Merck,
Germany, 99.0% purity, Ni(NO3)2 Á 6H2O), nickel (II) acetate
tetrahydrate (Fluka, Switzerland, 99.0% purity, Ni(CH3-
COO)2 Á 4H2O), and nickel (II) chloride hexahydrate (Merck,
98.0% purity, NiCl2 Á 6H2O). The ceramic powders are tetrago-
nal zirconia polycrystals (3Y-TZP; 3 mol% Y2O3; TZ-3YS,
Tosoh Corp.), with an average particle size of d50 5 0.26 mm.
Different preparation processes of ceramic–metal nanocom-
posite powder can be found in the literature.9–12 Thin-film dep-
osition methods produce monodisperse nanoparticles but
generally small amounts of sample. Volumetric procedures al-
low larger amounts of sample to be obtained, although nano-
particles appear mostly agglomerated. Therefore, in these
methods, the synthesis of nanoparticles is not the main prob-
lem, but their manipulation is, in order to avoid the presence of
large aggregates. The aim of the present work is to study the
factors that affect the particle size of dense ceramic nanocrys-
talline composites prepared by a conventional chemical route, as
in the thermal decomposition of metallic salts. In this process,
the control of both grain size and agglomeration of nanoparti-
cles is mandatory. We have found that the mean metal particle
size is influenced by the anion size of the precursor salts, as also
shown in other systems.13 Indeed, the nucleation process of
nanoparticles on ceramic particle surfaces can prevent the
growth of the particles and, therefore, the nanocomposite prop-
erties can be improved. Besides, this simple method allows large
amounts of sample to be obtained.
(2) Nanopowder Fabrication
Nickel salt powders were weighed in order to fabricate 10 vol%
Ni composites. This metal volume fraction was selected because
it is high enough to detect spurious phases by X-ray diffraction
(XRD) but small enough to avoid metal coalescence.17 This
powder was initially dissolved in alcohol by ultrasonic agitation.
Subsequently, 3Y-TZP powder was mixed with the above-men-
tioned solution and ball milled for 24 h with ZrO2 balls in al-
cohol media. Depending upon the salt, a suitable amount of
alcohol was added to ensure the total dilution of the Ni salt. The
concentrations in ethanol of the nitrate, acetate, and chloride
salts were 0.96, 0.96, and 0.85 mol/L, respectively. The mixture
was dried at 2001C and then calcined at 6001C for 2 h in air to
obtain ZrO2/NiO mixed powders. Subsequently, the soft ag-
glomerates in the calcined powders were crushed by wet ball
milling18 for 24 h. The resulting powder was dried and sieved
through a 100 mm sieve. Finally, the nickel oxide previously ob-
tained was reduced to metallic Ni in a 90%Ar/10%H2 atmos-
phere at 5001C for 2 h to yield a ZrO2/Ni powder.
L. Klein—contributing editor
Manuscript No. 20198. Received November 5, 2003; approved July 15, 2005.
Supported by the Spanish Ministry of Education and Science, under Project Number
MAT 2003-04199-C02.
**Fellow, American Ceramic Society.
wAuthor to whom correspondence should be addressed. e-mail: jsmoya@icmm.csic.es
144

January 2006
Obtaining Ni Nanoparticles on 3Y-TZP Powder from Nickel Salts
145
(3) Nickel Particle Size Distribution
III. Results
XRD analysis was performed for phase identification (Philips
PW 1710) and for Ni crystallite size determination (Philips
X’Pert) based on the (111) peak (28.2231) linewidth. The aver-
age grain size of Ni was roughly estimated from X-ray line
broadening according to Scherrer’s equation, d 5 0.9l/b cos y,
where l is the wavelength of the incident X-ray radiation
(CuKa1, l 5 0.15405981 nm) and y is the scattering angle. The
corrected peak half-width, b, was obtained according to the
Warren formula (b2 5 Bm2 ÀBs2), where Bm is the measured peak
width and Bs is the instrumental width.19,20
The Ni particle size distribution in the specimens was meas-
ured using transmission electron microscopy (TEM) images
(JEOL microscope model FXII, JEM 2000 operating at 200
kV attached to a spectrometer AN 10000 Link with an Si–Li
detector for microanalysis) for powdered samples, and a JEOL
ARM microscope operating at 1250 keV for sintered samples.
The thermal decomposition of samples was measured by differ-
ential thermal analysis (DTA) and thermogravimetric analysis
(TGA, Stanton, STA 781).
(1) DTA–TGA
DTA and TGA curves of Ni salts precipitated on ZrO2 grains
were studied to determine the appropriate calcination tempera-
ture (Fig. 1). In all cases, thermal decomposition developed two
stages, dehydration and decomposition, to produce NiO. In the
case of Ni nitrate (Fig. 1(A)), dehydration took place from 1001
to 2001C, and a clearly defined endothermic peak corresponding
to the thermal decomposition of the salts into NiO was observed
at around 3301C. The thermal evolution of Ni acetate salt
(Fig. 1(B)) only shows two main features: the dehydration proc-
ess (B1001C) and the acetate decomposition to NiO (B3331C).
Finally, in Ni chloride (Fig. 1(C)), dehydration occurred be-
tween 1001 and 2001C. An endothermic peak indicating the
thermal decomposition of the salt appeared at B6721C.
As a compromise between the results of DTA and TGA
curves, and the sintering process of zirconia and NiO, the cal-
cination temperature of the three Ni salts on ZrO2 in air was
chosen to be 6001C. It should be noted that according to DTA/
TGA data, the Ni chloride will not totally decompose at such
temperature. However, the duration of time for the decompo-
sition (2 h) to occur ensured that no trace of NiCl2 remained in
the powder. This fact was proved with the XRD patterns shown
(Fig. 2). After calcination in air at 6001C for 2 h, only peaks
corresponding to tetragonal zirconia, monoclinic zirconia, and
nickel oxide were found by XRD in all samples (Fig. 2(A)).
(4) Adsorption of Nickel Salts on Zirconia Surface
In order to study the adsorption of nickel salts on the surface of
zirconia, two different suspensions with the same weight percent
of solid content were prepared. First of all, Ni salts were dis-
solved in ethanol and mixed with 3Y-TZP powder. Secondly,
pure 3Y-TZP powder was mixed with the corresponding diluted
acid (nitric, acetic, and hydrochloric acid, respectively) in the
same Ni concentration as for the salts. After 24 h stirring, the
mixtures were filtered, washed four times with alcohol, and sub-
sequently dried. Self-supporting pellets were prepared by press-
ing the samples diluted in KBr, and then recorded by an infrared
(IR) spectrophotometer (Bruker IF66 v/s). These analyses were
compared with the one corresponding to the Ni salt and the one
corresponding to zirconia. X-ray photoelectron spectra (XPS)
were determined for the powders obtained from the suspensions
prepared with the salts. XPS spectra were recorded with a VG
ESCALAB 200R system equipped with a hemispherical electron
analyzer and an MgKa, 120 W, X-ray source. Atomic ratios
were estimated by calculating the integral of each peak (or a set
of them) after subtraction of an ‘‘S-shaped’’ (Shirley) back-
ground and using standard relative sensitivity factors.21 Binding
energy (BE) values were referenced to the C(1s) line at 284.6 eV
(estimated accuracy 5 70.1 eV). The magnetic hysteresis loops
were measured by a vibrant sample magnetometer (ML-VSM9)
with a low- and high-temperature chamber.
(2) XRD and TEM
The XRD technique was used to determine the Ni particle size
as well. In the case of powders reduced, obtained from the
(5) Sintering and Characterization
Nickel micrometer particle size 3Y-TZP/Ni (10 vol% Ni) com-
posites, obtained by a wet-processing route, were used for a sake
of comparison with nanometer nickel composites. Both micro-
and nanometer-size composites conformed following the same
procedure: dried powders were isostatically pressed at 200 MPa;
the resulting cylindrical rods with 5 mm diameter were fired us-
ing a tubular furnace on a 90%Ar/10%H2 atmosphere in two
steps: (i) at 5001C for 2 h in order to reduce the NiO present in
the surface of the Ni particles and (ii) at 14301C for 2 h for final
sintering. The heating and cooling rate were maintained at 601C/
h. The bulk densities of the sintered compacts were determined
by the Archimedes method. The microstructures of fired spec-
imens were examined using optical microscopy after polishing
the surfaces with a 1 mm diamond paste. Thin foils for TEM
observations were obtained following a standard procedure for
grinding (GATAN 656) and ion thinning (Bal-Tec, Balzers,
model RES-010, 5 kV Ar1 ions, incident angle 101). Cermets
were characterized by TEM using an electron microscope JEOL
2010 F (200 kV).
Fig. 1. Differential thermal analysis (DTA) (solid line) and therm-
ogravimetric analysis (TGA) (dash line) curves for tetragonal zirconia
polycrystals (3Y-TZP) powder milled in alcohol media with (A) nickel
(II) nitrate hexahydrate, (B) nickel (II) acetate tetrahydrate, and (C)
nickel (II) chloride hexahydrate, and then dried. Dehydration tempera-
tures ( & ) and thermal decomposition (&) are indicated in the figure.

´
Journal of the American Ceramic Society—Esteban-Betegon et al.
146
Vol. 89, No. 1
Fig. 2. X-ray diffraction profiles of (A) tetragonal zirconia polycrystals
(3Y-TZP)/NiO powder obtained after calcination and (B) 3Y-TZP/Ni
powder obtained after calcination and subsequent reduction. Sample
powders were obtained from nickel (II) nitrate hexahydrate (NIT), nick-
el (II) acetate tetrahydrate (ACET), and nickel (II) chloride hexahydrate
(CHL).
nickel chloride salt (Fig. 2(B)), the Ni(111) peak was sharp,
corresponding to large particles (around 1 mm). The same peak
was broader in the case of samples prepared from acetate (B60
nm) or nitrate (B50 nm) as the precursor and, therefore, nickel
particles had a smaller size.
TEM micrographs (Fig. 3(A)) and local microanalysis on se-
lected particles revealed that metal nanoparticles obtained from
nitrate appeared homogeneous and well dispersed on the zir-
conia surface, and the Ni average particle size was r50 nm. As
can be observed in Fig. 3(B), the Ni nanoparticles obtained from
acetate have a heterogeneous size distribution. Mainly two par-
ticle sizes have been observed: on the one hand, particles with a
size much lower than 50 nm and, on the other, particles in the
80–150 nm range.
Conversely, the sample obtained from chloride exhibited the
largest Ni particle size even larger than that of the ceramic ma-
trix particles (B200 nm). It should be noted that these results
have been inferred not only from the analysis of the TEM mi-
crography (Fig. 3(C)), but also from the results of microanalysis
on large selected particles (data not shown). Thus, it can be
concluded that although Ni particles seem to be agglomerated
because of growth during the drying or calcination processes it is
difficult to determine a reliable data size distribution only by
using TEM results.
Fig. 3. (A) Transmission electron microscopy (TEM) image showing
tetragonal zirconia polycrystals (3Y-TZP)/Ni, 10 vol%, obtained from
nitrate Ni salt. (B) and (C) TEM images of samples with the same com-
position obtained from acetate and chloride, respectively.
(3) Magnetic Measurements
Figure 4 shows the magnetic hysteresis loop for 3Y-TZP/10
vol% Ni powders using nitrate and chloride as the starting ma-
terials, which have the smallest Ni size and a good homogeneous
dispersion of the Ni particles on the surface of the 3Y-TZP
grains. The magnetization curves were obtained for applied
magnetic fields ranging over 7200 kA/m. Two sets of measure-
ments were made at two different temperatures (90 and 300 K,
respectively). The coercive force for samples obtained from ni-
trate salts was found to be Hc 5 25 kA/m at a temperature of
90 K and Hc 5 15 kA/m at 300 K. However, Ni powders ob-
tained from chloride salts presented narrower hysteresis loops at
the same temperatures (Hc 5 0.4 and 1.1 kA/m for 90 and 300
K, respectively). As expected, the coercive force decreased as the
temperature increased. The saturation magnetization was 510
kA/m at a temperature of 90 K, and 480 kA/m at 300 K for all
Ni samples. These results are in good agreement with the tab-
ulated value of Ni saturation magnetization (520 kA/m at a
temperature of 90 K and 490 kA/m at 300 K).

January 2006
Obtaining Ni Nanoparticles on 3Y-TZP Powder from Nickel Salts
147
Fig. 4. Hysteresis loop corresponding to tetragonal zirconia polycrys-
tals (3Y-TZP)/Ni 10 vol% obtained from nickel. Nitrate and nickel
chloride measured at different temperatures (90 and 300 K).
(4) Infrared Spectroscopy and XPS
Infrared absorption spectroscopy (Fig. 5) and XPS (Fig. 6) have
been used to elucidate the nature of the interaction of Ni salts
with the powdered ZrO2 substrate. In order to investigate the
degree of physiadsorption of each type of anion, and for com-
parison purposes, we have recorded the IR spectra of the three
Ni salts, as well as those of the starting zirconia powder, and
zirconia powder mixed with the corresponding acid (nitric, ace-
tic, and hydrochloric, respectively).
There is a peak located at 1384 cmÀ1 in the commercial zir-
conia sample (Figs. 5(A)–(C), curve (2)). This peak corresponds
to weakly adsorbed NOÀ3 species. The origin might be found in
the preparation method of stabilized zirconia powders starting
from Y(NO3)3. The ZrO2 powder stirred with nickel nitrate in
alcohol media, and subsequently filtered and washed (Fig. 5(A),
curve (4)) presented this band, stronger than that observed in the
zirconia original powder.22 The same band was found in the case
of using nitric acid (Fig. 5(A), curve (3)), although weaker than
that of nickel nitrate. It is noteworthy that this peak is weaker in
the case of the zirconia substrate once treated with acetic acid
and nickel acetate (Fig. 5(B), curves (3) and (4), respectively),
and it disappears from the zirconia treated with hydrochloric
acid and nickel chloride (Fig. 5(C), curves (3) and (4), respec-
tively).
In the case of acetate, there is no experimental evidence of any
adsorption of (COOÀ) groups on the zirconia surface (Fig. 5(B),
curve (4)), probably because of the presence of zirconia additives
that hide the weak bands of acetate. In the case of nickel chlo-
ride, no impurity bands were detected (Fig. 5(C), curve (4)),
but probably because the possible vibrations corresponding to
modes of ZrCl4 fall in the same spectral area as those of the
zirconia.
In order to gain more information about the presence of the
Ni salts on the zirconia surface, Ni-containing samples prepared
in the same way as for the IR spectroscopy measurements (i.e.,
the same solids as in Fig. 5, curve (4)) were examined with XPS.
Ni was detected in all of the spectra (Fig. 6). First of all, the
position of the main Ni(2p3/2) peak was found to be at values of
BE of 855.8, 855.2, and 855.4 eV for the specimens prepared
from nitrate, acetate, and chloride salts, respectively. Moreover,
the presence of a strong satellite at about 6 eV higher clearly
agrees with a basically Ni (II) redox state in all cases. The in-
tensities of the Ni features varied among the samples, corre-
sponding to Ni/Zr atomic ratios, averaged within the explored
depth (ca. 2–3 nm) of 0.15, 0.04, and 0.035 for samples obtained
from nitrate, acetate, and chloride, respectively. These values
were calculated from the 2p3/2 components only, as in the latter
two samples the satellite of the Ni(2p1/2) peak was weak and
could not be quantified.
Fig. 5. Study of anion adsorption on tetragonal zirconia polycrystals
(3Y-TZP) particles by infrared spectroscopy for different Ni salts: (A)
Ni(NO3)2 Á 6H2O, (B) Ni(C2H3O2)2 Á 4H2O, and (C) NiCl2 Á 6H2O. (1) Ni
corresponding salt, (2) as-received zirconia powders, (3) dried and
washed suspension of zirconia in ethanol/acid, and (4) dried and washed
suspension of zirconia in ethanol/salt. Corresponding acids are (A)
HNO3, (B) HC2H3O2, and (C) HCl. Curves (2)–(4) share the same scale,
but they have been shifted vertically for convenience.
Observation of the N(1s) feature was expected in the range
just above BE5 400.0 eV. However, it was impeded by its rel-
atively low sensitivity and its overlap with the Ni Auger peak
occurring in these spectra in a position close to BE5 407.0 eV.
Some diffuse shoulders (not shown) were observed around
BE5 405.5 eV, a position that corresponds to nitrate-type spe-
cies, but the mentioned overlap did not allow to ensure a rela-
tionship between them and nitrogen species, especially as the
shoulder was more intense in the sample obtained from nickel
acetate salt, in which, as previously mentioned, the Ni contri-
bution was stronger. Thus, from the intensity of this shoulder,
an upper limit corresponding to an atomic ratio (averaged with-
in the explored depth) of N/Zr5 0.05 could be set for the sample
obtained from nickel acetate, and approximately half of that for
the other two. Certainly, the value for the sample obtained from
nickel nitrate did not show any higher intensity with respect to
the other samples.
On the other hand, the Zr and O contributions (not shown)
were as expected for zirconia. It is worth mentioning that an

´
Journal of the American Ceramic Society—Esteban-Betegon et al.
148
Vol. 89, No. 1
Fig. 7. X-ray photoelectron spectra (XPS) spectra in the C(1s) range,
obtained for tetragonal zirconia polycrystals (3Y-TZP)/Ni samples pre-
pared from Ni (II) (A) nitrate, (B) acetate, and (C) chloride salts after
washing.
Fig. 6. X-ray photoelectron spectra (XPS) spectra in the Ni(2p) range,
obtained from tetragonal zirconia polycrystals (3Y-TZP)/Ni samples
prepared from (A) Ni (II) nitrate, (B) acetate, and (C) chloride salts after
washing. Note the three times higher magnification of curves (A)
and (B).
porosity was located around the micrometer-sized metal parti-
cles (Fig. 9(B)).
IV. Discussion
important contribution from Cl (not shown) appeared in all
spectra. From the intensity of the Cl(2p) peak at BE ꢀ 195.5 eV,
atomic ratios (averaged within the explored depth) Cl/Zr5 0.59,
0.65, and 0.56 were measured for these specimens. It was also
noted that the intensity of the ‘‘adventitious carbon’’ (hydro-
carbon-like contamination adsorbed from air after sample prep-
aration) was significantly lower for the specimen obtained from
nickel acetate (ratio C/Zr5 0.47) than for the other two (C/
Zr 5 0.76 in both), suggesting a lower tendency of organic mat-
ter adsorption in the former case (Fig. 7).
The study of powder TEM micrographs reveals that nickel par-
ticles obtained from nitrate present a homogeneous distribution.
The Ni particle size in these samples is the smallest one com-
pared with samples obtained from acetate and chloride salts.
However, in the case of samples prepared from acetate as a
precursor, nickel nanoparticles that precipitated on the zirconia
surface coexist with large nickel clusters. In the case of the pow-
ders obtained from a chloride salt, the nickel particles are, at
least, as large as those of the zirconia matrix.
Magnetic hysteresis loops were recorded in order to confirm
whether the average nickel particle size in samples obtained
from nitrate salts is in the nanometer range, as well as their dis-
persion within the matrix. When the particle size of a magnetic
material decreases, its magnetic structure varies from a multi-
domain to a single-domain state.23 The particle size for which
nickel remains in the single-domain state ranges from 3 to 50
nm.24 The magnetic properties, determined through hysteresis
loops of single-domain particles, are different from those of
multidomain particles in the sense that the coercive force in-
creases notably for lower particle sizes. Moreover, the presence
of large aggregates in a set of nanometer-size particles will also
narrow the hysteresis loop. As a consequence, magnetic meas-
urements can determine the particle size of a powdered sample
in the nanometric scale. Under these conditions, the sample
obtained from nickel nitrate exhibits a hysteresis loop with a
coercive force of the order of 25 kA/m at a temperature of
90 K, and 15 kA/m at 300 K. Because large polycrystalline par-
ticles will narrow the hysteresis loop completely, we can conclude
that this sample does not have large agglomerates. Therefore, all
nanoparticles are well dispersed in the zirconia matrix.
(5) Sintered Samples
Figure 8 shows optical micrographs of the cross-sectional micro-
structure of composites with 10 vol% Ni obtained from nickel
nitrate, acetate, and chloride, respectively. The darker phase
corresponds to zirconia and the light phase corresponds to nick-
el. At the magnification used, the Ni particles are difficult to
distinguish in the sample obtained from nitrate. This is not the
case for the sample corresponding to nickel chloride, where the
metal particle size can reach values as high as 12 mm. The sample
obtained from acetate remains in an intermediate position
(o5 mm).
The density of the sintered 3Y-TZP/Ni composites was found
to be 93%–98% theoretical. The procedure presented in this
work has allowed us to obtain dense nanocrystalline ceramic
composites obtained from nitrate salts (Fig. 9(A)). However, in
former works, this group has obtained different zirconia/nickel
monolithic samples via a wet processing route by mixing nickel
metal powder (Ni average particle size of 1.6 mm) and zirconia
powder directly within a wide range of metal volume fractions.
The final density of such composites after sintering ranged from
85% to 95% with regard to the theoretical value. Generally, the
No direct conclusion can be obtained about the adsorption of
Ni salt anions on the zirconia surface from XPS data, because of

January 2006
Obtaining Ni Nanoparticles on 3Y-TZP Powder from Nickel Salts
149
Fig. 9. (A) Transmission electron microscopy (TEM) image showing
nanosized Ni particles in a dense sample obtained from Ni nitrate salt
following the method presented in this work. (B) TEM image of sample
prepared following a wet processing pressureless sintering route, where
micrometer-sized Ni particles and an extremely high porosity are
evident.
present for all three samples (Figs. 5(A)–(C), curve (2)), because
of its presence in the initial zirconia material, and again a direct
conclusion is not possible. However, a quantitative analysis of
the 1384 cmÀ1 peak reveals that it disappears from the zirconia
substrate once treated with an acid solution of some other an-
ion, such as ClÀ (Fig. 5(C), curve (3)). On the contrary, this peak
notably increases in the case of the zirconia powder treated with
Ni nitrate salt (Fig. 5(A), curve (4)), being even larger than that
of the powder treated with HNO3 (Fig. 5(A), curve (3)). This
suggests that the whole molecule of the nickel nitrate is adsorbed
on the zirconia surface.
However, markedly different Ni/Zr ratios were detected by
XPS in the three types of washed precursor samples, even
though the volume fraction of nickel was the same for each
sample. It was found to be the lowest in the sample obtained for
the nickel nitrate case, and especially high in that obtained from
the nickel acetate salt. This latter observation suggests that, in
the presence of acetate, the adsorption properties of zirconia
(which probably controls the amount of Ni retained after wash-
ing) have changed. Such an indication may be coupled with the
observation that the amount of XPS-detected surface carbon-
aceous species is lowest in the case of acetate, in spite of the fact
that an organic anion is used. A lower tendency to adsorb or-
ganic impurities from the ambient atmosphere suggests a less
acidic character of the surface; this agrees with the more basic
character of acetate anion in comparison with nitrate or chlo-
ride. Thus, it could be postulated that, in this case, the zirconia
surface may be more populated by basic OHÀ groups, which
could lead to a higher amount of adsorbed Ni21 cations during
Fig. 8. Optical micrographs corresponding to polished cross-sections of
tetragonal zirconia polycrystals (3Y-TZP)/10 vol% Ni-sintered samples
obtained from (A) Ni (II) nitrate, (B) Ni (II) acetate, and (C) Ni (II)
chloride.
contamination problems (in the case of Cl and C) or overlapping
with other bands (N). In this case, IR spectroscopy should help
to detect the presence of the NOÀ3 ion weakly adsorbed on
the surface of zirconia grains.23 Unfortunately, such a band is

´
Journal of the American Ceramic Society—Esteban-Betegon et al.
150
Vol. 89, No. 1
present a heterogeneous distribution. This result has been ex-
plained through the formation of Ni(OH)2 in the suspension,
which is a consequence of the lowest acidity. Finally, the results
obtained in the present work indicate that the main factors that
allow Ni nanoparticles to precipitate on a ceramic matrix are the
following: the metallic precursor may be a salt only if its anion is
a large one and if the metallic cations remain in solution without
reacting with other chemical species.
Acknowledgments
Thanks are due to E. Pardo for assistance in recording the XPS spectra, as well
as to L. Gremillard for his help with the observation of the samples by TEM.
References
1H. Hwang, K. Tajima, M. Sando, and M. Toriyama, ‘‘Fatigue Behavior of
PZT-Based Nanocomposites with Fine Platinum Particles,’’ J. Am. Ceram. Soc.,
81 [12] 3325–8 (1998).
2M. Lieberthal and W. Kaplan, ‘‘Processing and Properties of Al2O3 Nano-
composites Reinforced with Sub-Micron Ni and NiAl2O4,’’ Mater. Sci. Eng. A,
302, 82–91 (2001).
Fig. 10. Infrared spectra of nickel (II) nitrate on zirconia surface (con-
tinuous line). Ni(OH)2 infrared spectra drawn for reference (dashed
line). Inset: details of the band owing to the presence of Ni(OH)2.
3S. Oh, T. Sekino, and K. Niihara, ‘‘Fabrication and Mechanical Properties of 5
vol% Copper Dispersed Alumina Nanocomposite,’’ J. Eur. Ceram. Soc., 18, 31–7
(1998).
preparation. This might contribute to the presence of a fraction
of highly dispersed Ni (within a very heterogeneous size distri-
bution, as detected by TEM) in that obtained from the nickel
acetate salt sample. Taking into account the possible effects of
the basic character of the acetate by these XPS data, one may
think that Ni(OH)2 might be formed, together with Ni(CH3
COO)2, during the preparation of the material. The presence of
Ni(OH)2 has been demonstrated by the 3645 cmÀ1 mode as-
signed to Ni(OH)2 (Fig. 10). This might also provide an expla-
nation for the heterogeneous size distribution in samples
obtained from acetate salts. Since the OHÀ ion has a smaller
size than CH3COOÀ ion, the electrostatic Madelung stabilizat-
ion energy of Ni(OH)2 will be higher than that of Ni(CH3
COO)2. Thus, Ni particles obtained from Ni(OH)2 have a high-
er particle size than those obtained from Ni(CH3COO)2, giving
rise to a heterogeneous particle size distribution.
4H. Hyuga, Y. Hayashi, T. Sekino, and K. Niihara, ‘‘Fabrication Process and
Electrical Properties of BaTiO3/Ni Nanocomposites,’’ Nanostruct. Mater., 9,
547–50 (1997).
5J. S. Helman and B. Abeles, ‘‘Tunneling of Spin-Polarized Electrons and Mag-
netoresistance in Granular Ni Films,’’ Phys. Rev. Lett., 37 [21] 1429–32 (1976).
6D. Kumar, H. Zhou, A. Kvit, and J. Narayan, ‘‘Improved Magnetic Properties
of Self-Assembled Epitaxial Nickel Nanocrystallites in Thin-Film Ceramic
Matrix,’’ J. Mater. Res., 17 [4] 738–42 (2002).
7M. Nawa, T. Sekino, and K. Niihara, ‘‘Fabrication and Mechanical Behavior
of Al2O3/Mo Nanocomposites,’’ J. Mater. Sci., 29 [12] 3185–92 (1994).
8T. Sekino and K. Niihara, ‘‘Microstructural Characteristics and Mechanical
Properties for Al2O3/Metal Nanocomposites,’’ Nanostruct. Mater, 6 [5–8] 663–6
(1995).
9K. Hiroki, S. Tohru, K. Takafumi, N. Tadachika, and K. Niihara, ‘‘Ceramic
Processing Science VI’’; pp. 867–72 in Ceramic Transactions, Vol. 12, Edited by
S. -I. Hirano, G. L. Messing, and N. Claussen. The American Ceramic Society,
Westerville, OH, 1993.
10CH. Laurent, A. Peigney, O. Que
´
nard, and A. Rousset, ‘‘Synthesis and
Mechanical Properties of Nanometric Metal Particles–Ceramic Matrix Nanocom-
posites,’’ Sil. Ind., 63 [5–6] 77–84 (1998).
Using the same reasoning, the large NOÀ3 anion would lead to
deposited salt crystals smaller than the ClÀ or OHÀ anions.
Thus, the lower Ni particle size obtained using nitrate salt can be
justified.
These results can be confirmed in the sintered samples, as can
be observed in Fig. 8. We could only obtain agglomerate-free
nanostructured composites in the case of starting from nano-
metric Ni particles. The best results were obtained from powders
using a nitrate solution as a precursor. Moreover, this kind of
composites presented porosity values lower than 2 vol% even
though they were sintered by a conventional pressureless route.
11C. A. Schuh, T.G Nieh, and T. Yamasaki, ‘‘Hall-Petch Breakdown Manifest-
ed in Abrasive Wear Resistance of Nanocrystalline Nickel,’’ Scripta Mater., 46,
735–40 (2002).
12E. Breval, G. Dodds, and C. G. Pantano, ‘‘Properties and Microstructure of
Ni-Alumina Composite Materials Prepared by the Sol/Gel Method,’’ Mater. Res.
Bull., 20 [10] 1191–205 (1985).
13L. A. Dı
az, A. F. Valdes, C. Dıaz, A. M. Espino, and R. Torrecillas, ‘‘Alu-
´ ´ ´
mina/Molybdenum Nanocomposites Obtained in Organic Media,’’ J. Eur. Ceram.
Soc., 23, 2829–34 (2003).
14C. Pecharroma
n, S. Lopez-Esteban, J. F. Bartolome, and J. S. Moya, ‘‘Evi-
´ ´ ´
dence of Nearest-Neighbor Ordering in Wet-Processed Zirconia–Nickel Compos-
ites,’’ J. Am. Ceram. Soc., 84 [10] 2439–41 (2001).
15S. Lo
pez-Esteban, J. F. Bartolome, and J . S. Moya, ‘‘Mechanical Perform-
´ ´
ance of 3Y-TZP/Ni Composites: Tensile, Bending, and Uniaxial Fatigue Tests,’’ J.
Mater. Res., 17 [7] 1592–600 (2002).
16H. Kondo, T. Sekino, Y.-H. Choa, T. Kusunose, T. Nakayama, M. Wada, T.
Adachi, and K. Niihara, ‘‘Mechanical and Magnetic Properties of Nickel-Dis-
persed Tetragonal Zirconia Nanocomposites,’’ J. Nanosci. Nanotech., 2 [5] 485–90
(2002).
V. Conclusions
In the present work, the processing of 3Y-TZP/10 vol% Ni
nanopowders via calcination and subsequent reduction from
three different Ni salts (Ni (II) nitrate, chloride, and acetate) to
obtain nanodispersed Ni particles has been established. It has
been found that the size of the anion is probably the factor that
mostly determines the final Ni particle size. The smallest anions
lead to the formation of larger crystals during preparation,
which influence the size of the final product. This is the case
for Ni particles obtained from chloride salt, which have a small
precursor anion, and the resultant Ni particle size is larger than
in the other two cases. The Ni particles obtained from nitrate are
the smallest ones, and they are better dispersed in the matrix.
Magnetic measurements of samples obtained from nitrate in-
dicate a magnetic single-domain structure. Thus, only in the case
where single nanoparticles appear perfectly dispersed and at-
tached to the zirconia particles will it be possible to obtain truly
zirconia-Ni nanocomposites, with Ni particle size around 50 nm.
In the case of using acetate as a precursor, the Ni particles
17C. Pecharroma
n, F. Esteban-Betegon, J. F. Bartolome, G. Richter, and J. S.
´ ´ ´
Moya, ‘‘Theroretical Model of Hardening in Zirconia–Nickel Nanoparticle Com-
posites,’’ Nanoletters, 4 [4] 747–51 (2004).
18S. Oh, M. Sando, and K. Niihara, ‘‘Processing and Properties of Ni–Co Alloy
Dispersed Al2O3 Nanocomposites,’’ Scripta Mater., 39 [10] 1413–8 (1998).
19V. Srdic, M. Winterer, A. Mo
¨
ller, G. Miehe, and H. Hahn, ‘‘Nanocrystalline
Zirconia Surface-Doped with Alumina: Chemical Vapor Synthesis, Characteriza-
tion, and Properties,’’ J. Am. Ceram. Soc., 84 [12] 2771–6 (2001).
20A. R. West, Solid State Chemistry and Its Applications, pp. 173–5. John Wiley
and Sons, Great Britain, 1987.
21C. D Wagner, L. E. Davis, M. V. Zeller, J. A. Taylor, R. M. Raymond, and L.
H. Gale, ‘‘Empirical Atomic Sensitivity Factors for Quantitative Analysis by Elec-
tron Spectroscopy for Chemical Analysis,’’ Surf. Interface Anal., 3, 211 (1981).
22K. I. Hadjiivanov, ‘‘Identification of Neutral and Charged NxOy Surface Spe-
cies by IR Spectroscopy,’’ Catal. Rev.-Sci. Eng., 42 [1–2] 71–144 (2000).
23C. Kittel, J. K. Galt, and W. E. Campbell, ‘‘Crucial Experiment Demonstrat-
ing Single Domain Property of Fine Ferromagnetic Powders,’’ Phys. Rev., 77, 725
(1950).
24B. D. Cullity, Introduction to Magnetic Materials, pp. 117–9 and 309–11.
Addison-Wesley, California, 1972.
&