Crystallite size effect on the tetragonal-monoclinic transition of undoped nanocrystalline zirconia studied by XRD and Raman spectrometry

Article abstract of DOI:10.1006/jssc.1999.8565

Two powdered batches of undoped zirconia, prepared by spray pyrolysis and characterized by 6.9 and 13.2 nm average crystallite size were annealed for different temperatures and for different times. The tetragonal-monoclinic phase transformation was monitored by Raman spectroscopy and X-ray diffraction. The evolution of parameters such as crystallite size, c/a unit cell parameter ratio, content of monoclinic and tetragonal zirconia, and Raman wavenumbers was found to differ from one batch to the other. The destabilization of the tetragonal phase appears for the 10 to 40 nm crystallite size range and as a sudden phenomenon. The critical size range varies with preparation. The wavenumber shifts were also discussed on the basis of internal stresses associated with mechanical balance between the surface and volume states. (C) 2000 Academic Press.

Full text of DOI:10.1006/jssc.1999.8565

Journal of Solid State Chemistry 149, 399}407 (2000)
doi:10.1006/jssc.1999.8565, available online at http://www.idealibrary.com on
Crystallite Size Effect on the Tetragonal-Monoclinic Transition
of Undoped Nanocrystalline Zirconia Studied by XRD and
Raman Spectrometry
E. Djurado,ꢀ P. Bouvier, and G. Lucazeau
Laboratoire d+Electrochimie et de Physico-Chimie des Mate
H
riaux et des Interfaces, Institut National Polytechnique de Grenoble,
res Cedex, France
CNRS UMR 5631 associe l+UJF, B.P. 75, 1130 rue de la Piscine, 38402 St. Martin d+He
H
a`
`
Received August 4, 1999; in revised form October 21, 1999; accepted November 5, 1999
powder (8}10), water vapor in crystallite growth (11, 7), and
Two powdered batches of undoped zirconia, prepared by spray intrinsic defects such as vacancies in the anionic sublattice
pyrolysis and characterized by 6.9 and 13.2 nm average crystal-
lite size were annealed for di4erent temperatures and for di4erent
times. The tetragonal+monoclinic phase transformation was
monitored by Raman spectroscopy and X-ray di4raction. The
evolution of parameters such as crystallite size, c/a unit cell
parameter ratio, content of monoclinic and tetragonal zirconia,
and Raman wavenumbers was found to di4er from one batch to
the other. The destabilization of the tetragonal phase appears for
the 10 to 40 nm crystallite size range and as a sudden pheno-
and strain within the grains (12}15). Moreover, the existence
of a large stress gradient at the metal/oxide interface of
zircaloy oxidation "lms used in nuclear industry is believed
to stabilize the dense tetragonal form at the interface and to
limit the corrosion rate (16). Among the main factors in-
#uencing the stability of tetragonal zirconia, S. Lawson
reported that the purity of tetragonal powders is an inhi-
bitor of its transformation and that impurities act as
menon. The critical size range varies with preparation. The transformation nucleation points along grain boundaries
wavenumber shifts were also discussed on the basis of internal because they cause strain and thermal stresses.
stresses associated with mechanical balance between the surface
and volume states. ( 2000 Academic Press
Key Words: undoped tetragonal nanocrystalline zirconia;
monoclinic phase; crystallite size; Raman spectroscopy; XRD.
The purpose of this paper is to investigate the behavior of
two powdered batches, initially characterized by 6.9 and
1
3.2 nm average crystallite sizes, versus thermal treatment.
This present paper will not focus on the in#uence of water
content and the role of protons in the destabilization of
nanosized zirconia phases but on the crystallite size e!ect
and on the c/a parameters ratio. The tetragonal-to-mono-
clinic transformation is studied using XRD and Raman
spectroscopy and is aimed to model Zr oxidation. Indeed,
Zr oxidation can involve the formation of nanophases
located at the metal}oxide interface which are di$cult to
detect by X-ray di!raction, even under grazing incidence.
Raman spectrometry has been selected because of its ability
to identify amorphous phases (as, e.g., nanophases). A recent
study of nanometric size e!ect on monoclinic zirconia was
reported in the literature (17). Spectral changes reported in
this latter study and considered as speci"c of nanometric
character are quite small compared, for instance, to those
1
. INTRODUCTION
It is expected that nanostructured materials (i.e., grain
size less than 100 nm) present di!erent properties from those
of more conventional materials with grain structures on
a coarser scale. For instance, He et al. (1) have shown that
"
ne-grained zirconia exhibits better wear resistance than the
coarse-grained material. Tetragonal undoped zirconia
which is a metastable phase at lower temperature than the
well-known high-temperature zirconia in the 1170}23703C
temperature range (2, 3) could be stabilized by "ne crystal-
lites. Many controversies are reported in the literature
about the factors a!ecting the nucleation of tetragonal zir-
conia transformation. Stabilization of tetragonal zirconia
mainly depends on the amount of stabilizer (4, 5) and the
stabilizer distribution (6), the precipitation pH values during
the synthesis route (7), the grain size of the tetragonal
recently observed on nanocrystallites of SnO (18).
ꢁ
2. EXPERIMENTAL
2
.1. Powder Preparation
Undoped zirconia powders were prepared by the spray-
ꢀ
To whom correspondence should be addressed. Fax: 33 (0) 4 76 82 66
7
0. E-mail: Elisabeth.Djurado@lepmi.inpg.fr.
pyrolysis technique using an ultrasonic atomizer (19). The
3
99
0
022-4596/00 $35.00
Copyright ( 2000 by Academic Press
All rights of reproduction in any form reserved.

4
00
DJURADO, BOUVIER, AND LUCAZEAU
TABLE 1
Spray-Pyrolysis Conditions and Resulting Crystallite Sizes Measured from X-Ray Line Broadening
for Two Batches of ZrO Powders
2
Concentration
(mol L\ꢀ)
Frequency
(kHz)
Flow rate
(L min\ꢀ)
Temperature
Crystallite size
(nm)
Batch No.
(3C)
I
II
2.5;10\ꢁ
2.5;10\ꢁ
1.7;10ꢂ
850
6
3
600
900
6.9
13.2
precursor solution consisted of zirconyl nitrate hydrate zero shift and the instrumental resolution. Positions and
(
1
ZrO(NO ) , Aldrich) dissolved in distilled water at 2.5; FWHM of XRD peaks were determined by a decomposi-
ꢂ ꢁ
0\ꢁ mol L\ꢀ concentration. The aqueous solution was at- tion into pseudo-Voigt-shaped peaks with ABFit software
omized by a high-frequency ultrasonic mist generator. We (ILL, Grenoble, France). The crystallite size was calculated
used two piezoelectric ceramic transducers for which the using the Debye}Scherrer formula for the 111 and 222 peaks
frequencies were respectively 850 kHz or 1.7 MHz. The car- of the tetragonal phase (22): D"0.9j/b cos h, where D is the
rier gas (N #O mixture) carried the aerosol with #ow crystallite size (in nm), j the wavelength (in nm), b the
ꢁ
ꢁ
rates of 3 or 6 L min\ꢀ through a tubular furnace, heated at corrected FWHM from high-purity silicon (in radian), and
either 600 or 9003C after a very short pyrolysis time (typi- h the di!raction angle. Cell parameter re"nements were
cally, 8 s for a 6 L min\ꢀ #ow rate). Table 1 summarizes the performed using the software CELREF (by Laugier and
experimental conditions for the synthesis of two powder Filhol, ILL, Grenoble, France).
batches. At the di!erence with other methods of &&chimie
Room-temperature Raman spectra were recorded using
douce'' (20), the present synthesis route led to nanocrystal- a DILOR XY multichannel spectrometer equipped with
line powders of pure zirconia in which no amorphous phase a CCD detector refrigerated by liquid nitrogen. The 514.53
was detected, neither with X-ray di!raction nor with Raman nm (green) and the 488 nm (blue) excitation lines of an
spectrometry. It was not the case with the spray-pyrolysis argon ion laser were used. All experiments were carried out
route performed at 4503C and starting from metal}organic using a laser power of 20 mW in macroscopic back-scatter-
precursors (21). The presently prepared compounds are re- ing geometry. Low-temperature micro-Raman was carried
ferred to as I-6.9 and II-13.2, where I and II represent the down to the liquid nitrogen temperature (!1753C). Posi-
batch and 6.9 and 13.2 represent the average crystallite size tion and half width at half maximum (HWHM) of Raman
(in nm) which was estimated from X-ray line broadening bands were obtained by decomposition into Lorentzian-
measurements. shaped peaks with PeakFit software (v. 4.0, Jandel Scien-
In a second step, each powder batch, respectively I-6.9 ti"c).
and II-13.2, was then heated in air, in a platinum crucible for
Initial water content of the as-prepared I-6.9 and II-13.2
various thermal treatments (in the 600}14003C temperature was obtained from thermal gravimetric analysis (TGA) and
range with heating and cooling rates of 53C/min and in the was found to be about 5 and 1 wt%, respectively. This was
1
5 min to 2 h time range) to increase the crystallite size. Two con"rmed by IR measurements from relative absorption at
di!erent sets of samples were obtained (Table 2). The repro- about 3250 to that at 580 cm\ꢀ. We should notice that the
ducibility of the crystallite size for a given thermal treatment
was checked.
TABLE 2
Referred Samples for Two Batches of Zirconia Powders
2
.2. Powder Characterization
Particle size, morphology of powders, and crystallite size Annealing conditions in air
Batch I, I-6.9
Batch II, II-13.2
were examined using transmission electron microscopy
No annealing
6.9
8.3
8.9
20.3
21.0
30.1
36.3
65.2
13.2
13.8
13.3
18.0
21.8
22.0
25.5
85.8
(JEOL 300CX).
6
6
003C, 15 min
003C, 2 h
Powder X-ray di!raction was carried out using a Siemens
D500 h/2h di!ractometer with the Bragg Brentano geo- 8003C, 2 h
metry from 203 to 803 in 2h (0.033 in 2h step, 12 s as a 9003C, 15 min
9
1
503C, 1 h
0003C, 1 h
counting time) equipped with a rear monochromator
CuKa radiation, j"1.5406 A). High-purity silicon was
used as the standard for all scans to precisely measure the
(
>
ꢀ
14003C, 2 h

STUDY OF UNDOPED NANOCRYSTALLINE ZIRCONIA
401
subsequent thermal treatments of these samples are of
course accompanied by a departure of water and this
process is under study (23).
3
. RESULTS
3
.1. TEM on I-6.9, II-13.2, and I-20.3 Powders
Microscopic examinations of the as-prepared powders
show submicron spherical particles consisting of aggregated
nanocrystallites. Figure 1 shows bright-"eld TEM micro-
graphs of as-prepared I-6.9 and II-13.2 powders (a and b) as
well as I-20.3 powder resulting from 8003C/2 h thermal
treatment (c). II-13.2 powder exhibits a homogeneous repar-
tition of crystallite size inside each agglomerate and the size
is close to that extracted from X-ray line broadening. This is
not clearly the case for I-6.9 powder and it is much more
di$cult to conclude the monodisperse aspect of the size
distribution.
3
.2. XRD Characterization
Figure 2a shows zoomed XRD patterns (203}453 in 2h) of
I-6.9 versus thermal treatment. The crystallite size increases
from 6.9 to 65.2 nm (see Table 2). The as-prepared I-6.9
di!ractogram was identi"ed with tetragonal zirconia crys-
tallizing in the P4 nmc space group which was obtained at
ꢁꢃ
1
2503C. An average crystallite size of 6.9 nm was deduced
from the broadening of the 111 line (FWHM"1.113 com-
pared to the 111 line of high-purity silicon FWHM"0.153).
At 2h"353 the absence of splitting into 200 and 002 di!rac-
tion lines, characteristic of the tetragonal structure, may be
attributed either to the broadening of di!raction lines due to
a small crystallite size or to the deviance from the tetra-
gonality toward a less anisotropic structure (c/a+1). Cell
parameter re"nement for I-6.9 leads to a"0.5092$9;
1
0\ꢄ nm and c"0.5165$8;10\ꢄ nm (c/a"1.014). At
this stage, one can assert that the broadening of XRD lines
is mainly due to the nanoscale size of crystallites and not to
the deviance from tetragonality. The simultaneous presence
of tetragonal and monoclinic ZrO was clearly detected
ꢁ
with increasing crystallite size. A continuous c/a ratio in-
crease is presented in Fig. 3 for an increase of crystallite
sizes. It is interesting to note that structural characteristics
of this tetragonal phase are close to those of other zirconia
of similar crystallite sizes and prepared via a hydrolysis
route from chlorides (20). For a larger crystallite size
FIG. 1. Bright-"eld TEM micrographs of I-6.9, II-13.2, and I-20
(
65.2 nm), monoclinic ZrO appears (Fig. 2a), (24). Cell powders (1 cm indicates 0.1 lm).
ꢁ
parameter re"nement leads to a"0.53144$5;10\ꢅ nm,
b"0.52100$4;10\ꢅ nm, c"0.51476$5;10\ꢅ nm,
and b"99.218$0.009. Figure 2b shows zoomed XRD lines clearly appeared for an +22 nm crystallite size, giving
patterns (203}453 in 2h) of II-13.2 versus thermal treatment. rise to a pure monoclinic phase. A striking change of c/a
The crystallite size increases from 13.2 to 85.8 nm (see values for II-13.2 was observed in a smaller crystallite size
Table 2). The pure tetragonal crystalline phase was charac- domain, +22 nm (Fig. 3), with the simultaneous appear-
terized by an average +13 nm crystallite size. Monoclinic ance of monoclinic ZrO , c/a ratio reaching a plateau for
ꢁ

4
02
DJURADO, BOUVIER, AND LUCAZEAU
FIG. 2. X-ray di!ractograms of (a) I-6.9 and (b) II-13.2 versus crystallite size increase. Monoclinic and tetragonal ZrO are indexed according to
ꢁ
C37-1484 and C42-1164 JCPDS cards, respectively.
both batches and for the 25 nm crystallite size. An +5% belonging to a small proportion of monoclinic ZrO , one
ꢁ
relative variation of the unit cell volume is observed during observes the six Raman active modes of the tetragonal form
phase transformation in agreement with the literature data (Dꢀꢅ). For a large crystallite size (65.2 nm) the structure is
ꢄF
(
25).
purely monoclinic (Cꢅ ); it can be con"rmed by the observa-
ꢁ
F
tion of 16 Raman modes from the 18 (9A #9B ) expected
g
g
ꢀ
ꢀ
by symmetry analysis (26). As the crystallite size increased
from 8.3 to 36.3 nm, tetragonal bands around 600, 480, and
3
.3. Raman Spectroscopy
The continuous transformation of the tetragonal-to- 320 cm\ꢀ are progressively replaced by monoclinic bands at
monoclinic phase was also observed with Raman spectro- 630}618, 476, and 347}335}306 cm\ꢀ, respectively.
scopy. Figure 4a shows Raman spectra of I-6.9 powder Figure 4b represents Raman spectra of II-13.2 powder
versus crystallite size. For a small crystallite size, except for versus thermal treatment. Here, the transformation is ob-
the presence of two weak bands at 179 and 189 cm\ꢀ served in a much narrower range of crystallite sizes com-
prised between 21.8 and 25.5 nm.
Raman frequency variation versus crystallite size of
noninterfering modes belonging to tetragonal and mono-
clinic forms are presented in Fig. 5. The critical size R
above which tetragonal zirconia disappears is equal to
#
2
3 nm in the II-13.2 batch and is much more di$cult to
de"ne in batch I. For the I-6.9 batch, the 150 and 270 cm\ꢀ
tetragonal modes strongly decrease when the crystallite size
increases (Figs. 5a and 5d, respectively) while on the con-
trary, no substantial frequency variation is detected for the
II-13.2 batch below R . If we focus on the monoclinic phase
#
below R , we observe a positive or constant variation of 178
#
and 189 cm\ꢀ modes when the crystallite size increases
(Figs. 5c and 5d) contrary to the tetragonal phase. The
small proportion of the monoclinic phase in the II-13.2
batch generates large error bars on the wavenumber
FIG. 3. c/a ratio measured by X-ray re"nements versus crystallite size
increase.

STUDY OF UNDOPED NANOCRYSTALLINE ZIRCONIA
403
FIG. 4. Raman spectra of (a) I-6.9 and (b) II-13.2 powder versus crystallite size increase.
measurement. When the crystallite size increases, a discon-
The HWHM of Raman modes at 150 and 260 cm\ꢀ for
tinuity is observed at R and above R the monoclinic the two di!erent batches are depicted versus the crys-
#
#
wavenumbers increase to reach the bulk frequency.
tallite size in Fig. 6. A continuous decrease of HWHM
for all the modes was observed with crystallite size ex-
pansion.
Typical Raman spectra of II-18 powder at room tempera-
ture (253C) and liquid nitrogen temperature (!1753C) are
presented in Fig. 7. The decreasing temperature triggered o!
a spectacular narrowing of all Raman bands and a positive
frequency shift without simultaneous destabilization of the
FIG. 5. Wavenumbers versus crystallite sizes of tetragonal Raman
bands at (a) 150 and (d) 260 cm\ꢀ and monoclinic Raman bands at (b) 180
and (c) 190 cm\ꢀ of I-6.9 and II-13.2 batches. A vertical line is reported for
FIG. 6. hwhm versus crystallite sizes of Raman bands at 150 and
260 cm\ꢀ for I-6.9 and II-13.2 batches.
the critical size R above which tetragonal zirconia disappears.
#

4
04
DJURADO, BOUVIER, AND LUCAZEAU
tetragonal structure. This thermal narrowing e!ect is stron-
ger for low-frequency modes.
3
.4. Comparative Quantitative XRD and Raman Analysis
Quantitative XRD and Raman analysis of the tetragonal/
monoclinic two-phase zirconia were already reported by
several authors (27}31). Calibration curves for mixtures of
monoclinic and tetragonal powders were respectively con-
structed by Evans et al. from X-ray di!raction results (28)
and Kim et al. from Raman data (29). The monoclinic
content, referred to as f , was obtained (i) from XRD pat-
ꢀ
terns using the following equation,
f "I (111)/+I (111)#I (111),,
ꢀ
ꢀ
ꢁ
ꢀ
or (ii) from Raman spectra using the following equation (30),
FIG. 8. Monoclinic content versus crystallite size determined respec-
tively by X-ray di!raction and Raman spectroscopy for I-6.9 and II-13.2
batches. A zoom of low monoclinic content versus low crystallite range is
also given.
0
.13
f " 0.19!
!0.56,
ꢀ
ꢀ
X !1.01
ꢀ
with X "+I (180 cm\ꢀ)#I (192 cm\ꢀ),/+I (180 cm\ꢀ)
ꢀ
ꢀ
ꢀ
ꢀ
#
I (192 cm\ꢀ)#I (148 cm\ꢀ),.
ꢀ ꢁ
The monoclinic ZrO content versus crystallite size is
ꢁ
presented in Fig. 8 from XRD and Raman data. The global micrographs (Fig. 1) showed the true geometric size of
behavior is similar for the two techniques. For II-13.2 pow- crystallites, i.e., the true physical value, whereas crystallite
der batch, a drop from t-ZrO to m-ZrO for a crystallite diameters calculated from peak broadening of XRD pat-
size '21 nm was observed while a continuous phase trans- terns are the measurements of a mean sphere diameter.
formation was detected for I-6.9. Moreover, as it was "rst Moreover, the crystallite size of tetragonal and monoclinic
stated by Kim et al. (29), we should notice that Raman forms which appear after annealing are found to be quite
spectroscopy is much more appropriated to detect traces of similar.
ꢁ
ꢁ
the monoclinic phase (see zoom in Fig. 8).
According to XRD and Raman data, the increase of
crystallite size with annealing treatments of undoped ZrO
induces a tetragonal-to-monoclinic phase transformation
ꢁ
4
. DISCUSSION
(see Fig. 8). These results are in good agreement with those
Crystallite sizes determined by XRD and TEM are found reported by Gupta et al. (10). Stefanic et al. (7) reported that
to be of the same order of magnitude. Dark "eld TEM the destabilization of undoped zirconia by thermal treat-
ment is believed to be due to dehydroxylization for anneal-
ing above 4003C. In the present study, the destabilization
rate di!ers from one batch to the other. For I-6.9 powder
which presents the smallest crystallite size, the highest con-
tent of the monoclinic phase (5 wt%) and of water (5 wt%),
the structural transformation is continuous with a coexist-
ence of the monoclinic and tetragonal phases in a broad
domain of crystallite sizes extending up to 45 nm. How-
ever, for II-13.2 which is characterized by larger crystal-
lites, 1 wt% of the monoclinic phase and 1 wt% of water,
the transformation appears more critical and the tetragonal
phase suddenly disappears above +23 nm. This latter
critical phenomenon can be well explained by Garvie et al.'s
theory (32). The authors assumed that an excess of surface
energy in smaller crystallites should stabilize the pure tetra-
FIG. 7. Raman spectra of II-18 at decreasing temperatures from 253C
to !1753C.
gonal phase. Here, the critical size is R "23 nm, whereas
#

STUDY OF UNDOPED NANOCRYSTALLINE ZIRCONIA
405
Garvie et al. estimated it to be around 30 nm (32). Below R ,
The examination of c/a ratios is directly correlated with
#
the departure of water due to thermal treatments does relative tetragonal proportion for increasing crystallite sizes.
not in#uence the tetragonal stability. The thermal behavior For II-13.2 critical behavior is once again observed (Fig. 3).
of I-6.9 cannot be adequatly understood by the surface At higher sizes, c/a reaches a plateau at c/a"1.022 which is
energy theory reported by Garvie et al. Here, the size disper- very close to the 1.025 reported for undoped zirconia from
sion and the global content of water and of the monoclinic the XRD pattern taken at 12503C. The larger c/a ratio
phase of the powders must play a major role in these (1.022) found for larger crystallites should lead to larger
two di!erent behaviors. Mitsuhashi et al. (12) showed anisotropic stress relaxation.
from mechanical treatment that a tetragonal and mon-
Let us compare the Raman study of our zirconia with
oclinic mixture are much more di$cult to transform into a recent work (17) for similar sizes. Contrary to Ref. (17),
a monoclinic than a single tetragonal domain. The authors zirconia nanocrystallites do not present any surface modes,
suggested that the domain boundaries may suppress the the crystallites likely being too large for allowing a su$cient
tPm transformation. Note that water could be located in proportion of surface atoms. For instance, for a 7 nm
these boundaries. Stefanic et al. (7) assumed that the rate spherical particle, there are only 4% of surface atoms. These
of the tPm transformation decreases with an increase of authors mentioned the presence of a strong band close to
m-ZrO content. Let us compare the in#uence of thermal 1030 cm\ꢀ which they assigned to a surface second order
ꢁ
treatment on phase fractions of nanocrystalline zirconia mode. We also observe the presence of intense peaks be-
prepared by di!erent routes. A recent study carried out by tween 1000 and 1400 cm\ꢀ (Fig. 9) which can be unambigu-
Benfer et al. (33) focused on zirconia prepared by (i) CVD ously attributed to #uorescence peaks due to presumably
from organic precursors and (ii) #ame hydrolysis from defects resulting from the initial preparation. Notice that
ZrCl (Degussa). Our zirconia were prepared by pyrolysis this #uorescence is the sum of two di!erent #uorescence
ꢄ
of nitrates at 6003C (I-6.9) and 9003C (II-13.2). In all patterns (34). The larger one is related to tetragonal zirconia
cases, a much higher tetragonal phase fraction is asso- while the weaker one belongs to the monoclinic phase which
ciated with larger crystallite sizes. However, it is clear that is a minority in the sample. The patterns are di!erent
di!erences in the monoclinic appearance rate with temper- because the cationic coordination number changes from
ature occur depending on the synthesis conditions (i.e., raw 8 to 7 with the structural transition from tetragonal to
products, temperature, atmosphere, etc.). Our new experi- monoclinic. This #uorescence deserves to be more precisely
mental data provide a valid comparison of two batches analyzed in the future.
which were prepared by the same procedure, playing
The bulk frequencies exhibit some substantial variation
with only the properly controlled temperature parameter. with increasing size. The size dependence of wavenumbers
The di!erent annealings were also carried out in the same can be discussed in terms of a phonon con"nement model
furnace.
(35) and also in terms of internal stresses which can result
FIG. 9. Raman spectra of I-20.3 powder recorded with the 514.5 nm (green) and 488.8 nm (blue) excitation line of an argon ion laser.

4
06
DJURADO, BOUVIER, AND LUCAZEAU
from the defects linked to the large speci"c area of the rystalline tetragonal zirconia and their transformation into
crystallites. For I-6.9, the wavenumbers of volume modes of monoclinic phase versus thermal treatments. The rate of
tetragonal zirconia observed at 148 and 270 cm\ꢀ decrease transformation is found quite similar by both techniques.
with increasing crystallite size. These trends are not so A slightly higher Raman sensitivity is found for the detec-
clearly evidenced for II-13.2. Considering the gru
mode constants obtained in a recent high pressure study of
nanometric tetragonal zirconia (36) (c "#1.75 cm\ꢀ/ modest as long as the surface contribution is smaller than
K
neisen tion of low monoclinic content.
The crystallite size e!ect on Raman spectra remains
ꢀ
ꢄꢆ
GPa, c "! 3.59 cm\ꢀ/GPa), the same red shift of the the volume contribution. Tetragonal Raman wavenumber
ꢇꢈ
48 and 270 cm\ꢀ modes cannot be explained by an iso- shifts are negative and lower than 4 cm\ꢀ in the 6 to 15 nm
ꢁ
1
static stress variation when the crystallite size increases. crystallite size range. The tetragonal HWHM is reduced by
Actually, this could be due to the fact that the internal a factor of 2 in this crystallite size range. These results did
pressure is anisotropic as it can be deduced from the in- not favour the con"nement model nor a mechanical stress
crease of the c parameter with the crystallite size while the model.
a parameter is nearly invariant. Thus, it is di$cult to con-
The critical crystallite size for the transformation is well
clude on the sign of the stresses. Assuming that vacancies de"ned (23 nm) for the sample which contains only 1 wt%
are preferably located in the near surface region and that the monoclinic zirconia and 1 wt% water in its initial state
accomodation of oxygen at the surface activates the (powder prepared at higher temperature and lower speed).
transition, one can predict that the transformation induces This observation is in good agreement with Garvie's ther-
a volume expansion in the near surface layer and thus modynamic model. Reciprocally, for the compound initially
a compression in the core of the crystallite. As in Ref. (17), rich in monoclinic phase (5 wt%) and in water (5 wt%), the
the wavenumbers of monoclinic crystallites exhibit a discon- transformation is delayed.
tinuity at R and then those above R exhibit a slight
Contrary to the literature about monoclinic nanometric
increase and tend toward those measured in a bulk sample powders (17), Raman modes have not been observed around
relaxed monoclinic zirconia). This discontinuity could cor- 1000 cm\ꢀ. On the other hand, monoclinic and tetragonal
#
#
(
respond to anisotropic stress relaxation which is di$cult to phases present di!erent #uorescence bands in this frequency
quantify. Moreover, from general considerations such as the range.
anisotropic stress gradient in each crystallite, the size distri-
bution inside the particule, presence of defects, or "nally
phonon con"nement, one can expect a general broadening
of Raman bands for small crystallite sizes. The Raman lines
of II-18 powder surprisingly become quite narrow (at the
di!erence of 3 mol% Yꢂ> stabilized zirconia) with decreas-
ing temperature. This observation is a clear proof that both
the con"nement model and stress distribution cannot ex-
plain the HWHM of Raman bands at room temperature.
Consequently, this remarkable behavior in nanometric crys-
tallites is typical of crystallites without defects presenting
harmonic behavior with similar Raman HWHM. More-
over, the small HWHM con"rms the homogeneity of crys-
tallite sizes. On the contrary, the Raman HWHM of I-6.9
are rather broad and we observe a substantial decrease with
thermal treatment. No such important narrowing of Raman
bands was observed at low temperature (37). Here, the
relative in#uence of intrinsic defects, con"nement e!ects,
and size distribution cannot be easily decorrelated.
ACKNOWLEDGMENTS
The authors would like to thank N. Rosman for his technical assistance
for Raman measurements.
REFERENCES
1
. Y. J. He, A. J. A. Winnubst, A. J. Burggraaf, H. Verweij, P. G. van der
Varst, and B. G. De With, J. Am. Ceram. Soc. 79, 3090 (1996).
. H. J. Scott, J. Mater. Sci. 10, 1527 (1975).
. G. Teufer, Acta Crystallogr. 15, 1187 (1962).
. K. Haberko and R. Pampuch, Ceram. Int. 9, 8 (1983).
5. K. Tsukuma, Y. Kubota, and T. Tsukidate, &&Advances in Ceramics,
Science and Technology of Zirconia II'' (N. Claussen, M. Ruehle, and
A. H. Heuer, Eds.), Vol. 12, p. 382. The American Ceramic Society Inc.,
Columbus, OH, 1984.
. S. Lawson, J. Eur. Ceram. Soc. 15, 485 (1995).
. G. Stefanic, S. Music, B. Grzeta, S. Popovic, and A. Sekulic, J. Phys.
Chem. Solids 59, 879 (1998).
2
3
4
6
7
8
. R. C. Garvie, R. H. Hannink, and R. T. Pascoe, Nature (¸ondon) 258,
7
03 (1975).
To evidence the nature, the quantity, and the distribution
of defects in the crystallites such as dislocations, surstruc-
tures, microdomains, TEM, and low-temperature Raman
spectrometry are in progress.
9
1
. F. F. Lange, J. Mater. Sci. 17, 225 (1982).
0. T. K. Gupta, J. H. Bechtold, R. C. Kuznicki, L. H. Cado!, and B. R.
Rossing, J. Mater. Sci. 12, 2421 (1977).
11. Y. Murase and E. Kato, J. Am. Ceram. Soc. 66, 196 (1982).
1
1
1
2. T. Mitsuhashi, M. Ichihara, and U. Tatsuke, J. Am. Ceram. Soc. 57, 97
1974).
3. A. P. Mirgorodsky, M. B. Smirnov, and P. E. Quintard, Phys. Rev. B
5, 19 (1997).
(
5
. CONCLUSION
5
This present work is focused on X-ray di!raction and
4. A. H. Heuer, M. Ruhle, and D. B. Marshall, J. Am. Ceram. Soc. 73, 1084
Raman spectrometry characterizations of undoped nanoc-
(1990).

STUDY OF UNDOPED NANOCRYSTALLINE ZIRCONIA
407
1
1
5. K. Negita and H. Takao, J. Phys. Chem. Solids 50, 325 (1989).
6. J. Godlewski, P. Bouvier, G. Lucazeau, and L. Fayette, &&12th Interna-
26. D. Michel, M. Perez, Y. Jorba, and R. Collongues, J. Raman Spectrosc.
5, 163 (1976).
tional Symposium &Zirconium in Nuclear Industry,''' Toronto, 1998, 27. E. D. Whitney, ¹rans. Faraday Soc. 61, 1191 (1965).
AST-STP 1343, G. D. Moan and G. P. Sabol, Eds., American Society 28. P. A. Evans, R. Stevens, and J. G. P. Binner, Br. Ceram. ¹rans. J. 83, 39
for testing and material, 1999.
(1984).
1
1
7. G. G. Siu, M. J. Stokes, and Y. Liu, Phys. Rev. B 59, 4 (1999).
8. L. Abello, B. Bochu, A. Gaskov, S. Koudryavtseva, G. Lucazeau, and
M. Roumyantseva, J. Solid State Chem. 135, 78 (1998).
9. E. Djurado and E. Meunier, J. Solid State Chem. 141, 191 (1998).
29. B. K. Kim, J. W. Hahn, and K. R. Han, J. Mater. Sci. ¸ett. 16, 669
(1997).
30. B. Alzyab, C. H. Perry, and R. P. Ingel, J. Am. Ceram. Soc. 70, 760
1
2
(1987).
0. X. Bokhimi, A. Morales, A. Garcia-Ruiz, T. D. Xiao, H. Chen, and 31. C. G. Kontoyannis and M. Orkoula, J. Mater. Sci. ¸ett. 29, 5316
P. R. Strutt, J. Solid State Chem. 142, 409 (1999). (1994).
1. P. Murugavel, M. Kalaiselvam, A. R. Raju, and C. N. R. Rao, J. Mater. 32. R. C. Garvie, J. Phys. Chem. 69, 1238 (1965).
Chem. 7(8), 1433 (1997). 33. S. Benfer and E. Knozinger, J. Mater. Chem. 9, 1203 (1999).
2. A. Guinier, &&Theorie et Technique de la radiocristallographie'' (Dunod, 34. I. M. Asher, B. Papanicolaou, and E. Anastassakis, J. Phys. Chem.
Ed.), 3rd ed., Chap. 12, p. 482. Paris, 1964. Solids 37, 221 (1976).
3. E. Djurado, C. Roux, and L. Dessemond, &&12th International Confe- 35. H. Richter, Z. P. Wang, and L. Ley, Solid State Commun. 39, 625
2
2
2
K
H
rence on Solid State Ionics,'' Thessaloniki, 1999, June 6}12, 1999.
4. D. K. Smith and H. W. Newkirk, Acta Crystallogr. 18, 983 (1965).
(1981).
2
2
36. P. Bouvier and G. Lucazeau, J. Phys. Chem. Solid, 1999, in press.
5. R. A. Cutler, J. R. Reynolds, and A. Jones, J. Am. Ceram. Soc. 75, 2173 37. P. Bouvier, E. Djurado, C. Ritter, and G. Lucazeau, to be
1992). published.
(