Experimental and theoretical investigation on the correlation between aqueous precursors structure and crystalline phases of zirconia

Article abstract of DOI:10.1016/j.molstruc.2003.12.006

Nanometer zirconia powders were prepared by the precipitation method at different pHs and different reaction temperatures. X-ray results show that monoclinic zirconia is favored at pH 4 while tetragonal zirconia is favored at pH 9.5 at room temperature, and monoclinic zirconia is also favored at pH 9.5 and 70°C reaction temperature, with the slow addition of alkali. Four models of zirconium complexes were applied to simulate the structural monomers in different pH solutions. Geometric parameters and Mulliken charge population were calculated by optimizing these complexes using the density functional theory (DFT/B3LYP). Theoretical analyses show that if Model I ([Zr(OH) ₂(H₂O)₄]²⁺ monomers) is favored in the aqueous precursor solution, it will be preferentially polymerized into monoclinic precursor structure irrespective of slow or quick alkali addition. Contrarily, if Model IV ([Zr(OH)₇]^3- monomers) is major in the aqueous precursor solution, tetragonal precursor structures are favored irrespective of slow or quick alkali addition. When Model II ([Zr(OH) ₄(H₂O)₂]⁰ monomers) and Model III ([Zr(OH)₆]^2- monomers), respectively, predominate in the aqueous precursor solution, they will be preferentially polymerized into tetragonal precursor structure at slow alkali addition, however, for quick alkali addition, they will be preferentially polymerized into monoclinic precursor structure. Our theoretical models well explain the present experimental results as well as previous experimental results, and allow building up a correlation between aqueous precursor structures and crystalline phases of zirconia.

Full text of DOI:10.1016/j.molstruc.2003.12.006

Journal of Molecular Structure 690 (2004) 181–187
www.elsevier.com/locate/molstruc
Experimental and theoretical investigation on the correlation between
aqueous precursors structure and crystalline phases of zirconia
a,
Shou-Gang Chen , Yan-Sheng Yin , Dao-Ping Wang
a
b
*
aKey Lab of Liquid Structure and Heredity of Materials, Ministry of Education, Key Lab of Engineering Ceramic of Shandong Province,
School of Materials, Shandong University, Jinan 250061, China
bSchool of Chemistry, Chemical Engineering and Materials Science, Shandong Normal University, Jinan 250014, China
Received 11 July 2003; revised 3 December 2003; accepted 4 December 2003
Abstract
Nanometer zirconia powders were prepared by the precipitation method at different pHs and different reaction temperatures. X-ray results
show that monoclinic zirconia is favored at pH 4 while tetragonal zirconia is favored at pH 9.5 at room temperature, and monoclinic zirconia
is also favored at pH 9.5 and 70 8C reaction temperature, with the slow addition of alkali. Four models of zirconium complexes were
applied to simulate the structural monomers in different pH solutions. Geometric parameters and Mulliken charge population were calculated
by optimizing these complexes using the density functional theory (DFT/B3LYP). Theoretical analyses show that if Model I
([Zr(OH)2(H2O)4]2þ monomers) is favored in the aqueous precursor solution, it will be preferentially polymerized into monoclinic precursor
structure irrespective of slow or quick alkali addition. Contrarily, if Model IV ([Zr(OH)7]32 monomers) is major in the aqueous precursor
solution, tetragonal precursor structures are favored irrespective of slow or quick alkali addition. When Model II ([Zr(OH)4(H2O)2]0
monomers) and Model III ([Zr(OH)6]22 monomers), respectively, predominate in the aqueous precursor solution, they will be preferentially
polymerized into tetragonal precursor structure at slow alkali addition, however, for quick alkali addition, they will be preferentially
polymerized into monoclinic precursor structure. Our theoretical models well explain the present experimental results as well as previous
experimental results, and allow building up a correlation between aqueous precursor structures and crystalline phases of zirconia.
q 2004 Elsevier B.V. All rights reserved.
Keywords: Zirconia; Precipitation; X-ray diffraction; Density functional theory/B3LYP; Mulliken charge population
1. Introduction
of the amorphous precursor and the tetragonal phase, and
other authors attributed it to the existence of a certain
number of oxygen vacancies [5].
Zirconia-based ceramics have gained great attention as
engineering ceramics and catalysis materials because of
their diverse commercial applications. Research on the
metastable phases of these materials and their transitions
have been a subject of great scientific and technological
importance during recent years. A number of expla-
nations have been offered for the interpretation of the
presence of metastable tetragonal as well as cubic
zirconia structures at low temperature. Garvie attributed
it to crystallite size effects [1], Wu and Yu [2] attributed
it to strain and lattice defects, Clearfield [3] attributed it
to the rate of polymerization of the Zr4þ species, Livage
[4] attributed it to the short-range structural similarities
Precipitation and sol–gel methods are the most used
chemical methods to produce zirconia nanoparticles. The
precipitation-driven product is indeed more hydrous, and
consequently, its complex molecular structure contains
various amounts of water in the form of hydrous oxide
ZrO2·H2O or ZrO(OH)2. Sol–gel chemistry is also based
on inorganic polymerization reactions. This process
involves the use of molecular precursors, mainly alk-
oxides as starting materials. Hence, the aqueous chemistry
of zirconium is dominated by its tendency to form
inorganic polymers based on bridged-hydroxyl groups
whether using the precipitation method or the sol–gel
method [6]. During polymeric progress, the monomers
structures play a significant role in the final gel molecular
precursor structure [7], and the different structural
*
Corresponding author. Tel./fax: þ86-531-8392439.
E-mail address: sgchen2000@yahoo.com.cn (S.-G. Chen).
0022-2860/$ - see front matter q 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.molstruc.2003.12.006

182
S.-G. Chen et al. / Journal of Molecular Structure 690 (2004) 181–187
monomers will produce different reaction processes [8].
Recently, Caracoche et al. [9] synthesized nanometer
zirconia powders using the sol–gel method at different
pHs. Their experimental results indicate that the different
morphology and structure of precursor gels were deter-
mined by the extent of condensation, which is in turn
affected by the pH value of the precursor solution.
However, there exist some other arguments on the
formation of amorphous precursor structure for the
monoclinic and/or the tetragonal phases in different pH
solutions [10–12]. In this investigation, we synthesize
nanometer zirconia powders using a precipitation method
at different pHs and different reaction temperatures.
Interestingly, our experimental results show that mono-
clinic zirconia could be attained not only at a low pH but
also at high pH and a higher reaction temperature. The
detailed analysis shows that different precursor structures
in the solution are very important for the crystal phases
obtained. Thus, it is necessary to clarify the correlation
between the precursor structures of the aqueous solution
and the final crystalline phases of zirconia.
were calcined for 2 h at different temperatures (450 and
600 8C).
Crystalline species in zirconia powders were identified
using an X-ray diffraction (XRD: Model D/MAX-RB12X,
Rigaku Co., Tokyo, Japan). The scans were taken within the
range of 10–808 ð2uÞ using Cu Ka radiation. The volume
fraction of the monoclinic phase ðVmÞ was determined by
the empirical formula [14]
ꢀ
ꢀ
Vm ¼ ½Imð111Þ þ Imð111Þꢀ=½Imð111Þ þ Imð111Þ þ Itð111Þꢀ
where Im denotes the intensity of the monoclinic peak and It
denotes the intensity of the tetragonal peak. The mean
crystallite size was calculated by means of the line-
broadening method from the measurement of the full
width at half-maximum (FWHM) of the peak corresponding
to the (111) reflections of the tetragonal or cubic zirconia
phase. The mean value of crystallite size was estimated
using the following equation of Scherrer and Warren [15]:
Dt ¼ 0:9l=ðB2 2 b2Þ1=2 cos uB
where Dt is the mean crystallite size, l (0.15418 nm) and
uB; respectively, denote the wavelength of the X-ray and the
Bragg diffraction angle, and B and b denote the FWHMs
observed for the present sample and the standard sample,
respectively.
Computational quantum chemistry provides a detailed
description of the bonding, charge distribution, structure
preferences and geometry of molecular materials [7,13],
especially the development of Density Functional Theory
(DFT). Based on the idea that a theoretical study would help
us to further understand the correlation between the
precursor structure and the microstructure of prepared
samples, our research group decided to explore the
electronic structure of four zirconium complexes models
related to different pH solutions, namely, (I) [Zr(OH)2(H2-
O)4]2þ (pH , 7); (II) [Zr(OH)4(H2O)2]0 (pH < 7); (III)
[Zr(OH)6]22 (7 , pH # 10.5) and (IV) [Zr(OH)7]32
(10.5 # pH # 14).
Differential thermal analysis (DTA) was also carried out
to study the transformation of hydrous ZrO2 samples using a
DTA-40M analyzer (Japan). Measurements were carried out
in the temperature range of 30–830 8C at a 20 8C/min
heating rate.
2.2. Computational method
Computations were performed using the GAUSSIAN 98
series of programs in Ref. [16]. Specifically, we used the
B3LYP gradient-corrected DFT program [17]. Zr atoms are
described using the Los Alamos Lanl2 effective core
potentials (ECP) and a valence double-j basis set [18]
while O and H atoms are described using the 6-31G(d) basis
set [19]. These geometries are fully optimized to find
stationary points followed by frequency calculations.
2. Experimental procedures and computational method
2.1. Experimental procedures
Two samples of hydrous ZrO2 were prepared using the
precipitation method from a ZrOCl2·8H2O (Tianda, China,
purity of 99%) solution (0.1 M) with 2 mass% polyglycol
(PEG, molecular weight 20,000 g/mol) by the slow addition
of the NH4OH (25%) solution, respectively, needed to
produce desired pHs of 4 and 9.5. The resulting precipitates
were repeatedly centrifugally washed with deionized water
at 8000 n/min until a negative test for Cl2, and then
centrifugally washed twice with anhydrous ethanol at
8000 n/min. Further, the obtained gels were dried at
100 8C for 24 h. In addition, we also prepared another
hydrous ZrO2 sample at 70 8C water bath by the slower
addition of the NH4OH (25%) solution needed to produce a
desired pH of 9.5 using the same precursor solution. The
washing and dried process was carried out as mentioned
above. Subsequently, the different hydrous ZrO2 powders
3. Results and discussion
3.1. Experimental results
Table 1 shows the crystallite size and phase volume
fraction of zirconia powders obtained from different pH
solutions and different reaction temperatures. The XRD
analysis of Fig. 1a was performed on ZrO2 xerogel obtained
from a pH 4 solution and thermally treated at 600 8C, for
which 90.91% volume fraction of monoclinic zirconia was
determined. The XRD analysis performed on ZrO2 xerogel
obtained from a pH 9.5 solution and thermally treated at
600 8C (Fig. 1b) led to only 35.48% volume fraction of

S.-G. Chen et al. / Journal of Molecular Structure 690 (2004) 181–187
183
Table 1
Crystallite size and phase volume fraction at different temperatures and different pH values
pHs
Temperature (8C)
Crystallite size (nm)
Tetragonal
Phase volume fraction
Tetragonal
Reaction
Calcining
Monoclinic
Monoclinic
4
25
25
25
70
450
600
600
600
11.530
17.243
15.141
17.382
15.768
43.199
23.763
33.145
17.30
9.09
82.70
90.91
35.48
81.02
4
9.5
9.5
64.52
18.98
monoclinic phase. It is obvious that monoclinic zirconia is
the preferential crystalline phase at the lower pH value and
tetragonal zirconia is the preferential crystalline phase at the
higher pH value, for the same rate of alkali addition. Using
Scherrer’s equation [15], the average crystallite size ðDt ¼
17:243 nmÞ of tetragonal zirconia was obtained from the
data of Fig. 1a. Based on Garvie’s viewpoint [1], this
crystalline size has not reached the critical size (30 nm) for
the tetragonal-to-monoclinic phase transformation, but it
has attained it according to other researchers’ viewpoints
[20–22]. Thus, we did not exclude the possibility that the
existing monoclinic zirconia could come from the tetra-
gonal-to-monoclinic phase transformation. In order to
explore the real reason, we then performed a thermal
analysis on the same zirconia powders with the sample of
Fig. 1a (Fig. 2). The DTA curve initially shows a small
endothermic effect (A–B) probably due to desorption of
water, and then an almost monotonic increase of heat output
due to the slow combustion (B–C) of the precursor.
However, at 474.3 8C a large exothermic peak (C–D) is
observed. The exothermic enthalpy (,18.8 kJ/mol) is larger
than the enthalpy (,5.9 kJ/mol) difference between mono-
clinic and tetragonal zirconia [23]. So, we can predict that
this change is not due to only a phase transition. To say the
least, if this exothermic peak represents a phase transition
process, it should occur above 455 8C. However, the XRD
Fig. 1. X-ray diffraction patterns for zirconia powders derived from different pH precursor solutions, different reaction temperature and different calcination
temperatures. (a) pH 4 and 25 8C; 600 8C, (b) pH 9.5 and 25 8C; 600 8C, (c) pH 4 and 25 8C; 450 8C, (d) pH 9.5 and 70 8C water bath; 600 8C.

184
S.-G. Chen et al. / Journal of Molecular Structure 690 (2004) 181–187
Fig. 2. DTA curve of pure zirconia xerogel for pH 4 at 20 8C/min.
analysis of the sample calcined at 450 8C (Fig. 1c) still
contains 82.70% volume fraction of monoclinic zirconia,
and the crystallite size (11.53 nm) is even smaller.
According to the previous reports [1,20–22], the tetra-
gonal-to-monoclinic phase transformation is impossible at
this crystallite size. In a word, our experimental results
indicate that the monoclinic zirconia is not transformed
from the tetragonal zirconia obtained at low pHs, thus
further pointing out the important role of the precursor
structure. As it was interpreted in Ref. [24], the exothermic
peak at 474.3 8C will be assigned to mass combustion
of carbonaceous materials along with the crystallization
of ZrO2.
Fig. 3. Optimized cluster structures for models named I to IV.
In order to further investigate the effect of the reaction
rate on the zirconia precursor structure, we also prepared
zirconia powders in a 70 8C water bath at the same rate of
alkali addition. The ZrO2 xerogel obtained from a pH 9.5
solution was thermally treated at 600 8C, and the corre-
sponding XRD pattern is shown in Fig. 1d. Surprisingly, the
volume fraction of monoclinic zirconia reached 81.02%,
much higher than the value (35.48%) determined from Fig.
1b. It is obvious that monoclinic zirconia increases at high
reaction temperature, though the crystallite size
(17.382 nm) does not exhibit a marked change. This result
shows that the high reaction temperature could help the
monoclinic phase precursor structure to form, and also
shows that the polymerization rate of monomer could
produce important role for the crystalline phases of zirconia.
The present experimental results indicate that both the
pH and the reaction rate of precursor solution apparently
affect the final structure of zirconia, which, in turn,
influences the crystalline phases of the zirconia product.
These results also accord well the previous experimental
results [3,4,7,8,10,12]. In order to verify this by theoretical
calculations, we then studied in detail the correlation
between the aqueous precursor structure and the final
crystalline phases structure of zirconia via selected four
models based on experimental analyses and studies.
presursor structure. Table 2 shows the interatomic distances
and Mulliken charge population for the optimized cluster
structures.
Model I denotes the preponderant zirconium coordi-
nation monomers in the lowest pH (,7) solutions. Because
of the low hydroxyl ions concentration, water molecules are
the mainly coordination body and two hydroxyl ions
localize at the axial ends of the tetragonal bipyramid. A
partial charge of þ1.672920 a.u. was observed on zirco-
nium atom whereas the negative charge is distributed on
oxygen atoms. Charges of 20.765551 and 20.764250 a.u.
3.2. Theoretical results
Fig. 3 shows the optimized cluster structures for models
named I to IV. Fig. 4 shows the ordered polymerization
Fig. 4. The ordered structure of polymerization precursors.

S.-G. Chen et al. / Journal of Molecular Structure 690 (2004) 181–187
185
Table 2
Interatomic distances and Mulliken charge populations for the optimized structure of four models used
˚
Interatomic distance (A)
Mulliken charge population (a.u.)
I
Rð1; 2Þ ¼ 2:2398; Rð2; 5Þ ¼ 1:9311; Rð2; 6Þ ¼ 2:2394; Rð2; 7Þ ¼ 1:9301;
Rð2; 8Þ ¼ 2:2397; Rð2; 9Þ ¼ 2:2397; Rð1; 3Þ ¼ 0:9769; Rð1; 4Þ ¼ 0:9769;
Rð5; 10Þ ¼ 0:9711; Rð6; 11Þ ¼ 0:9768; Rð6; 12Þ ¼ 0:9769;
Zr2 ¼ 1.672920; O1 ¼ 20.804272; O5 ¼ 20.765551; O6 ¼ 20.804085;
O7 ¼ 20.764250; O8 ¼ 20.804280; O9 ¼ 20.804103; H3 ¼ 0.514257;
H4 ¼ 0.514244; H10 ¼ 0.479695; H11 ¼ 0.514196; H12 ¼ 0.514198;
Rð7; 13Þ ¼ 0:9712; Rð8; 14Þ ¼ 0:977; Rð8; 15Þ ¼ 0:9769; Rð9; 16Þ ¼ 0:977; H13 ¼ 0.480016; H14 ¼ 0.514393; H15 ¼ 0.514172; H16 ¼ 0.514278;
Rð9; 17Þ ¼ 0:9769
H17 ¼ 0.514170
II Rð1; 2Þ ¼ 2:4399; Rð2; 5Þ ¼ 2:024; Rð2; 6Þ ¼ 2:0205; Rð2; 7Þ ¼ 2:0208;
Rð2; 8Þ ¼ 2:0241; Rð2; 9Þ ¼ 2:4412; Rð1; 3Þ ¼ 0:9773; Rð1; 4Þ ¼ 0:9767;
Rð5; 10Þ ¼ 0:9667; Rð6; 11Þ ¼ 0:9683; Rð7; 12Þ ¼ 0:9683; Rð8; 13Þ ¼
0:9667; Rð9; 14Þ ¼ 0:9772; Rð9; 15Þ ¼ 0:9767
Zr2 ¼ 1.3221191; O1 ¼ 20.786074; O5 ¼ 20.824608; O6 ¼ 20.830272;
O7 ¼ 20.830363; O8 ¼ 20.824525; O9 ¼ 20.7861063; H3 ¼ 0.470554;
H4 ¼ 0.469041; H10 ¼ 0.416345; H11 ¼ 0.424013; H12 ¼ 0.424025;
H13 ¼ 0.416344; H14 ¼ 0.470488; H15 ¼ 0.469018
III Rð1; 2Þ ¼ 2:1951; Rð2; 4Þ ¼ 2:1371; Rð2; 5Þ ¼ 2:149; Rð2; 6Þ ¼ 2:0977;
Rð2; 7Þ ¼ 2:1326; Rð2; 8Þ ¼ 2:1274; Rð1; 3Þ ¼ 0:971; Rð4; 9Þ ¼ 0:9701;
Rð5; 10Þ ¼ 0:9716; Rð6; 11Þ ¼ 0:9728; Rð7; 12Þ ¼ 0:9726; Rð8; 13Þ ¼
0:9725
Zr2 ¼ 0.9091041; O1 ¼ 20.889489; O3 ¼ 20.861415; O5 ¼ 20.860193;
O6 ¼ 20.8449103; O7 ¼ 20.866479; O8 ¼ 20.862843; H3 ¼ 0.383405;
H9 ¼ 0.377786; H10 ¼ 0.380153; H11 ¼ 0.375040; H12 ¼ 0.377374;
H13 ¼ 0.382468
IV Rð1; 2Þ ¼ 2:1565; Rð2; 4Þ ¼ 2:2278; Rð2; 5Þ ¼ 2:218; Rð2; 6Þ ¼ 2:1882;
Rð2; 7Þ ¼ 2:2066; Rð2; 8Þ ¼ 2:3973; Rð2; 9Þ ¼ 2:1778; Rð1; 3Þ ¼ 0:9775;
Rð4; 10Þ ¼ 0:9804; Rð5; 11Þ ¼ 0:9745; Rð6; 12Þ ¼ 0:9735; Rð7; 13Þ ¼
0:9721; Rð8; 14Þ ¼ 0:9752; Rð9; 15Þ ¼ 0:9816
Zr2 ¼ 0.510373; O1 ¼ 20.859210; O4 ¼ 20.867244; O5 ¼ 20.865347;
O6 ¼ 20.860163; O7 ¼ 20.886562; O8 ¼ 20.895826;
O9 ¼ 20.894994; H3 ¼ 0.364385; H10 ¼ 0.371605; H11 ¼ 0.373557;
H12 ¼ 0.373863; H13 ¼ 0.377215; H14 ¼ 0.381845; H15 ¼ 0.376501
correspond to the hydroxyl oxygens, with values being
lower than the mean charge on the oxygen atoms
(20.804185 a.u.) of coordination water. Hence, at low pH
conditions, the polymerization reaction of zirconium
coordination monomer mainly occurs between Zr–H2O
and Zr–H2O. Of course, this reaction also partially occurs
between Zr–H2O and Zr–OH or between Zr–OH and Zr–
OH. As listed in Table 2, the mean bond length of Zr–OH is
because of the relatively longer distance involved. More-
over, the flux of charge between zirconium ion and its
coordination oxygens also decreases the polymerization
reaction speed. For slow alkali addition, hydroxyl ions
substitution is probably enough to cause the polymerization
reaction mainly between Zr–OH and Zr–OH. The almost
equivalent bond lengths of Zr–OH also promote the
formation of a netlike ordered precursor structure, which
is close to that of tetragonal zirconia. However, for rapid
alkali addition, hydroxyl ions substitution may be not
sufficient and will cause Models I and II describe the
precursor solution. Hence, polymerization reactions will
occur between Zr–OH and Zr–OH or Zr–H2O and Zr–OH.
As in addition, these polymerization reactions lead to plenty
of structural water in the precursor structure, so they not
only introduce inequivalent bond lengths, but also disturb
the netlike ordered precursor structure. Finally, this
disordered precursor structure will preferentially crystallize
into the monoclinic phase.
˚
˚
1.93 A, and that of Zr–OH2 is 2.24 A; it was also observed
to be similar to that of bulk monoclinic zirconia [25]. Thus,
these complex polymerization reactions will introduce two
kinds of inequivalent bond lengths in the precursor
structure, so causing an asymmetric and disordered
precursor structure. The similitude between this long-
distance disordered structure and that of bulk monoclinic
zirconia structure probably facilitates the formation of the
monoclinic phase.
Model II corresponds to preponderant zirconium coordi-
nation monomers in pH < 7 solutions. Hydroxyl ions have a
larger binding energy with zirconium ion than water
molecules [26], so that water molecules are gradually
replaced by hydroxyl ions with the increase of pH value.
The tightly coordinated hydroxyl ions localize at the four
acmes of tetrahedron, and the water molecules localize at
the near neighbor sites of zirconium ion. With the increase
of coordination hydroxyls around the zirconium ion, the
Mulliken charge of zirconium atom decreases to
þ1.3221191 a.u., whereas the mean negative charge,
distributed on hydroxyl oxygen atoms increases to
20.827442 a.u., being higher than the mean charge
distributed on oxygen atoms (20.78609015 a.u.) of the
coordinated water. Hence, at pH < 7 conditions, the
polymerization reaction of zirconium coordination mono-
mers mainly occurs between Zr–OH and Zr–OH since the
polymerization reaction between Zr–H2O is not favored
Model III identifies the preponderant zirconium coordi-
nation monomers in pH < 7–10.5 solutions. At these high
pH solutions the nearest water molecules are completely
replaced by hydroxyl ions, which fully occupy the six acmes
of a tetragonal bipyramid. With the increase in coordination
hydroxyls around the zirconium ion, the Mulliken charge of
zirconium atom further decreases to þ0.9091041 a.u.,
whereas the mean negative charge of hydroxyl oxygen
atoms is further increased to 20.86422155 a.u. The flux of
charge between zirconium ion and its coordination oxygens
will help the zirconium coordination monomer to be fully
replaced by hydroxyl ions, and further decreases the
polymerization reaction speed between Zr–OH and Zr–
OH. This fact promotes the order of the precursor structure
and further forms the precursor structure of Fig. 4. In
˚
addition, the mean bond length of Zr–OH, of 2.1398 A, and

186
S.-G. Chen et al. / Journal of Molecular Structure 690 (2004) 181–187
being close to the Zr–O length in bulk tetragonal zirconia
[25], again promotes the formation of netlike ordered
tetragonal precursor structure. Thus, for slow alkali
addition, hydroxyl ions substitution is complete and
polymerization reaction fully occurs between Zr–OH and
Zr–OH. Once again, like Model II, hydroxyl ions
substitution also may be not sufficient and easily causes
Models II and III describe the precursor solution for rapid
alkali addition. Hence, the netlike precursor structure will
involve plenty of structural and free water, which will
disturb the order in the precursor structure. The disordered
precursor structure will easily crystallize into the mono-
clinic phase.
for the present experimental results and also accord well
with the previous experimental studies [10,12]. This fact
shows that our theoretical model and theoretical analyses
are reasonable and believable. Thus, the present theoretical
study allows building up a correlation between aqueous
precursor structures and crystalline phases of zirconia.
Furthermore, it would be possible to make some reasonable
predictions about the precipitation, hydrolysis and conden-
sation reactions of these complexes during the chemical
processes of the precursor aqueous solution.
4. Conclusions
Model IV describes the preponderant zirconium coordi-
nation monomers in pH , 10.5–14 solutions. Here, not
only the first neighbor water molecules but also the next
neighbor ones are completely replaced by hydroxyl ions (O8
in Fig. 3). With the continuous increase of coordination
hydroxyls around the zirconium ion, the Mulliken charge
population of zirconium atom is further decreased to
þ0.510373 a.u., and the mean negative charge of hydroxyl
oxygen atoms is slightly increased (20.872253 a.u.).
Hence, at the highest pH conditions, the polymerization
reaction of zirconium coordination monomers again occurs
between Zr–OH and Zr–OH. As in Model III, the slow
polymerization reaction speed will promote the ordering of
the precursor structure leading to the array shown in Fig. 4.
(1) Nanometer zirconia powders were prepared at different
pHs and different reaction temperatures using the
precipitation method. X-ray results show that mono-
clinic zirconia is favored at pH 4 while tetragonal
zirconia is favored at pH 9.5 at room temperature, and
monoclinic zirconia is also favored at pH 9.5 and 70 8C
reaction temperature, with the slow addition of alkali.
(2) If Model I ([Zr(OH)2(H2O)4]2þ monomers) is favored
in the low pH (,7) aqueous precursor solutions, it will
be preferentially polymerized into monoclinic precur-
sor structure irrespective of slow or quick alkali
addition. Contrarily, if Model IV ([Zr(OH)7]32 mono-
mers) is major in the high pH (.10.5) aqueous
precursor solutions, tetragonal precursor structures are
favored irrespective of slow or quick alkali addition.
(3) When monomers of Model II ([Zr(OH)4(H2O)2]0
monomers) and Model III ([Zr(OH)6]22 monomers),
respectively, predominate in the precursor solution, the
rate of alkali addition will produce a great effect on the
preferential precursor structure: tetragonal precursor
structure is favored at slow alkali addition and
monoclinic precursor structure is favored at rapid
alkali addition.
˚
Besides, the mean bond length of Zr–OH, now of 2.1958 A,
will also favor the formation of the netlike ordered
tetragonal precursor structure. Thus, the present monomer
of zirconium coordination is the most favorable for forming
the tetragonal precursor structure. Moreover, the substi-
tution of the next neighbor hydroxyl ions (O8) will prohibit
the formation of Models II and III, and synchronously
reduce the structural and free water of the ordered precursor
structure irrespective of slow or quick alkali addition, thus
leading always to the ordered cubic/tetragonal zirconia
precursor structure.
The above theoretical analyses (summarized in
Scheme 1) using selected models are a good explanation
Acknowledgements
This work was sponsored by the National Science
Foundation (No. 50242008) and the Education Ministry
Doctorate Foundation (No. 20020422001) of China.
References
[1] (a) R.C. Garvie, J. Phys. Chem. 69 (1965) 1238. (b) R.C. Garvie,
J. Phys. Chem. 82 (1978) 218.
[2] F. Wu, S. Yu, J. Mater. Sci. 25 (1990) 970.
[3] A. Clearfield, J. Mater. Res. 5 (1990) 161.
[4] J. Livage, K. Doi, C. Mazieres, J. Am. Ceram. Soc. 51
(1968) 349–353.
[5] A. Bogicevic, C. Wolverton, G.M. Crosbie, E.B. Stechel, Phys. Rev.
B 64 (2001) 014106.
Scheme 1. Outline of the theoretical results.

S.-G. Chen et al. / Journal of Molecular Structure 690 (2004) 181–187
187
´ ´
[6] J.A. Navio, G. Colon, P.J. Saachez-soto, M. Macias, Chem. Mater. 9
Ochterski, G.A. Petersson, P.Y. Ayala, Q. Cui, K. Morokuma,
N. Rega, P. Salvador, J.J. Dannenberg, D.K. Malick, A.D. Rabuck,
K. Raghavachari, J.B. Foresman, J. Cioslowski, J.V. Ortiz, A.G.
Baboul, B.B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I.
Komaromi, R. Gomperts, R.L. Martin, D.J. Fox, T. Keith, M.A. Al-
Laham, C.Y. Peng, A. Nanayakkara, M. Challacombe, P.M. W. Gill,
B. Johnson, W. Chen, M.W. Wong, J.L. Andres, C. Gonzalez, M.
Head-Gordon, E.S. Replogle, J.A. Pople, Gaussian Inc., Pittsburgh
PA, 1998.
(1997) 1256–1261.
´
[7] (a) A. Clearfield, Inorg. Chem. 3 (1964) 146. (b) M. Angeles Dıaz-
´
´ ´ ´
Dıez, A. Macıas-Garcıa, G. Silvero, R. Gordillo, R. Caruso, Ceram.
Int. 29 (2003) 471–475.
[8] J. Sun, L. Gao, J. Am. Ceram. Soc. 85 (2002) 2382–2384.
´
[9] M.C. Caracoche, P.C. Rivas, M.M. Cervera, R. Caruso, E. Benavadez,
O. de Sanctis, S.R. Mintzer, J. Mater. Res. 18 (2003) 208–215.
[10] R. Srinivasan, M.B. Harris, S.F. Simpson, R.J. De Angelis, B.H.
Davis, J. Mater. Res. 3 (1988) 787–797.
[17] A.D. Becke, J. Chem. Phys. 98 (1993) 1372.
[18] P.C. Hariharan, J.A. Pople, Theor. Chim. Acta 28 (1973) 213.
[19] M.S. Gordon, Chem. Phys. Lett. 76 (1980) 163.
[20] J.G. Yu, M. Pan, L.M. Zhang, China Sci. Bull. 36 (1991) 1270 in
Chinese.
[11] R. Srinivasan, C.R. Hubbard, O.B. Cavin, B.H. Davis, Chem. Mater. 5
(1993) 27–31.
[12] H.J. Noh, D.S. Seo, H. Kim, J.K. Lee, Mater. Lett. 57
(2003) 2425–2431.
[13] Y. Widjaja, C.B. Musgrave, J. Chem. Phys. B 117 (2002) 1931–1934.
[14] H. Toraya, M. Yoshimura, S. Somiyam, J. Am. Ceram. Soc. 67
(1984) 119.
[21] G. Baldinozzi, D. Simeone, D. Gosset, M. Dutheil, Phys. Rev. Lett. 90
(2003) 216103.
[22] T. Chrsaka, A.H. King, C.C. Berndt, Mater. Sci. Engng A 286
(2000) 169.
[15] B.D. Cullity, Elements of X-Ray Diffraction, Addison-Wesley,
Reading, MA, 1978.
[23] J.P. Coughlin, E.G. King, J. Am. Ceram. Soc. 72 (1950) 2262.
[24] J.C. Ray, A.B. Panda, C.R. Saha, P. Pramanik, J. Am. Ceram. Soc. 86
(2003) 514–516.
[16] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb,
J.R. Cheeseman, V.G. Zakrzewski, J.A. Montgomery, Jr., R.E.
Stratmann, J.C. Burant, S. Dapprich, J.M. Millam, A.D. Daniels,
K.N. Kudin, M.C. Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi,
R. Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J.
[25] D.N. Wang, Y.Q. Guo, K.M. Liang, K. Tao, Sci. China A28
(1998) 823 in Chinese.
[26] A.W. Thomas, H.S. Owens, J. Am. Chem. Soc. 57 (1935) 1825.