Formation mechanism of CaTiO3 hollow crystals with different microstructures

Article abstract of DOI:10.1021/ja106461u

The crystal growth of CaTiO₃ hollow crystals with different microstructures has been investigated. In a water-free poly(ethylene glycol) 200 (PEG-200) solution, CaTiO₃ nanocubes formed first. The nanocubes underwent an oriented self-assembly into spherical particles, enhanced by the surface-adsorbed polymer molecules. Since the growth of nanocubes and their aggregation took place simultaneously, the nanocubes in the outer shells were larger than those in the cores. Disappearance of the small nanocubes in the cores of the spheres during an Ostwald ripening process led to spherical hollow crystals. Addition of a small amount of water (1.25 vol %) in the polymer solution enhanced surface recrystallization of the aggregated spheres, forming a cubic morphology. The orthorhombic distortion of the perovskite CaTiO ₃ structure did not have a significant effect on the nanocube aggregation, resulting in a domain structure in the shells. Single-crystalline hollow cubes were produced with a slightly higher water content, e.g., 5 vol %. This process of (1) aggregation of nanocubes and (2) surface crystallization followed by (3) surface-to-core extension of recrystallization gives a good example of the reversed crystal growth route in ceramic materials. The proposed formation mechanism of the hollow CaTiO₃ crystals would enable us to control the microstructures of these materials and to explain the formation of many other hollow crystals.

Full text of DOI:10.1021/ja106461u

Published on Web 09/15/2010
Formation Mechanism of CaTiO Hollow Crystals with
3
Different Microstructures
†
†
†
†
†
Xianfeng Yang, Junxiang Fu, Chongjun Jin, Jian Chen, Chaolun Liang,
,
†
,‡
Mingmei Wu,* and Wuzong Zhou*
State Key Laboratory of Optoelectronic Materials and Technologies, MOE Key Laboratory of
Bioinorganic and Synthetic Chemistry, School of Chemistry and Chemical Engineering, School
of Physics and Engineering, and Instrumental Analysis and Research Centre, Sun Yat-Sen
(
Zhongshan) UniVersity, Guangzhou 510275, P. R. China, and School of Chemistry, UniVersity
of St Andrews, St Andrews, Fife KY16 9ST, United Kingdom
Received July 29, 2010; E-mail: ceswmm@mail.sysu.edu.cn; wzhou@st-andrews.ac.uk
Abstract: The crystal growth of CaTiO
In a water-free poly(ethylene glycol) 200 (PEG-200) solution, CaTiO
3
hollow crystals with different microstructures has been investigated.
nanocubes formed first. The nanocubes
3
underwent an oriented self-assembly into spherical particles, enhanced by the surface-adsorbed polymer
molecules. Since the growth of nanocubes and their aggregation took place simultaneously, the nanocubes
in the outer shells were larger than those in the cores. Disappearance of the small nanocubes in the cores
of the spheres during an Ostwald ripening process led to spherical hollow crystals. Addition of a small
amount of water (1.25 vol %) in the polymer solution enhanced surface recrystallization of the aggregated
3
spheres, forming a cubic morphology. The orthorhombic distortion of the perovskite CaTiO structure did
not have a significant effect on the nanocube aggregation, resulting in a domain structure in the shells.
Single-crystalline hollow cubes were produced with a slightly higher water content, e.g., 5 vol %. This
process of (1) aggregation of nanocubes and (2) surface crystallization followed by (3) surface-to-core
extension of recrystallization gives a good example of the reversed crystal growth route in ceramic materials.
The proposed formation mechanism of the hollow CaTiO
3
crystals would enable us to control the
microstructures of these materials and to explain the formation of many other hollow crystals.
Introduction
inorganic micro-/nanomaterials with different shell structural
features, including their shapes, porosities, thicknesses, and
crystallinity, were fabricated in the past few years.
Control of crystalline phases, microstructures, and morphol-
ogies of inorganic functional materials continues to be a
fundamental issue in modern materials science and technology.
3
1
Besides traditional spherical hollow architectures composed
of randomly aggregated or radially arranged building blocks,
nonspherical hollow structures with well-defined regular mor-
phologies have recently attracted more attention for their specific
physical and chemical performances. Compared to the former,
the nonspherical shapes are much more difficult to achieve by
conventional templating methods, because suitable templates
for a specific shape are often not available and the microstruc-
Inorganic nanomaterials with hollow interiors have attracted
great attention in recent years due to their widespread potential
applications in many fields, e.g., new materials for sensors/
probes, drug-delivery carriers, biomedical diagnosis agents,
2
,3
catalysts for size-selective reactions, etc.
To satisfy the
different requirements of such applications, various hollow
†
4
Sun Yat-Sen University.
St Andrews University.
tures of the shells are less controllable. Consequently, self-
‡
templating methods were extensively adopted in the synthesis
of nonspherical hollow structures. For example, Yang et al.
prepared rhombododecahedral silver cages through the Kirk-
(
1) (a) Jun, Y. W.; Choi, J. S.; Cheon, J. Angew. Chem., Int. Ed. 2006,
5, 3414–3439. (b) Sayle, D. C.; Seal, S.; Wang, Z.; Mangili, B. C.;
4
Price, D. W.; Karakoti, A. S.; Kuchibhatla, S.; Hao, Q.; Mobus, G.;
Xu, X.; Sayle, T. X. T. ACS Nano 2008, 2, 1237–1251. (c) Wu, C. Z.;
Xie, Y. Chem. Commun. 2009, 5943–5957. (d) Yang, H. G.; Sun,
C. H.; Qiao, S. Z.; Zou, J.; Liu, G.; Smith, S. C.; Cheng, H. M.; Lu,
G. Q. Nature 2008, 453, 638–641. (e) Lee, I.; Delbecq, F.; Morales,
R.; Albiter, M. A.; Zaera, F. Nat. Mater. 2009, 8, 132–138. (f) Qi,
L. M. Coord. Chem. ReV. 2010, 254, 1054–1071.
endall effect by using rhombododecahedral Ag
3
PO
4
crystals as
5
sacrificial self-templates. Polyhedral Cu
2
O crystals have been
widely used as sacrificial self-templates for growing copper
6
sulfide hollow cages with nonspherical morphologies. A variety
(
2) (a) Ma, Y. R.; Qi, L. M. J. Colloid Interface Sci. 2009, 335, 1–10. (b)
Zeng, H. C. J. Mater. Chem. 2006, 16, 649–662. (c) Skrabalak, S. E.;
Chen, J. Y.; Sun, Y. G.; Lu, X. M.; Au, L.; Cobley, C. M.; Xia, Y. N.
Acc. Chem. Res. 2008, 41, 1587–1595. (d) Zhang, H. G.; Zhu, Q. S.;
Zhang, Y.; Wang, Y.; Zhao, L.; Yu, B. AdV. Funct. Mater. 2007, 17,
of novel metal hollow architectures were fabricated by Xia and
(3) (a) Zhao, Y.; Jiang, L. AdV. Mater. 2009, 21, 3621–3638. (b) An, K.;
Hyeon, T. Nano Today 2009, 4, 359–373. (c) Zhang, Q.; Wang, W. S.;
Goebl, J.; Yin, Y. D. Nano Today 2009, 4, 494–507.
(4) Lou, X. W.; Archer, L. A.; Yang, Z. C. AdV. Mater. 2008, 20, 3987–
4019.
2
766–2771. (e) Yavuz, M. S.; Cheng, Y. Y.; Chen, J. Y.; Cobley,
C. M.; Zhang, Q.; Rycenga, M.; Xie, J. W.; Kim, C.; Song, K. H.;
Schwartz, A. G.; Wang, L. H. V.; Xia, Y. N. Nat. Mater. 2009, 8,
9
2
35–939. (f) Ye, L. N.; Wu, C. Z.; Guo, W.; Xie, Y. Chem. Commun.
006, 4738–4740.
(5) Yang, J. H.; Qi, L. M.; Lu, C. H.; Ma, J. M.; Cheng, H. M. Angew.
Chem., Int. Ed. 2005, 44, 598–603.
10.1021/ja106461u  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 14279–14287 9 14279

A R T I C L E S
Yang et al.
7
co-workers also via self-templating routes. These self-templat-
ing processes consist of two steps, first fabrication of precursors
as sacrificial templates and then transformation to final hollow
structures via the Kirkendall effect.
To avoid complicated operations, one-step template-free
methods have also been widely employed for efficient produc-
tion of nonspherical hollow structures. A general procedure of
the template-free strategy can be described as (1) surfactant-
assisted self-aggregation of primary building blocks and (2)
development of the hollow interior via an Ostwald ripening
8
process as observed in the formation of PbTe nanoboxes,
9
10
CaTiO
3
, and Cu
2
O
hollow cubes. According to the previous
reports, the shells of the as-obtained inorganic cage-like hollow
architectures were mostly polycrystalline or built up by oriented
aggregration of nanocrystallites. Detailed formation mechanisms
of the aggregation and the microstructures of the shells have
not been investigated extensively. In the present work, we focus
Figure 1. Powder XRD patterns of the CaTiO
3
specimens produced with
different water contents in the synthetic solutions: (a) water-free, (b) 1.25
vol %, and (c) 5 vol %. The bottom pattern with vertical bars is derived
from the JCPDS card (No. 82-0229, space group Pbnm) of orthorhombic
on the formation mechanism of hollow crystals of CaTiO
different microstructures in order to enrich our knowledge of
this field of materials science.
3
with
CaTiO
and pattern (c) is indexed to the orthorhombic unit cell with a ) 5.4033, b
5.4406, and c ) 7.6653 Å.
3
. Pattern (a) is indexed to the pseudocubic subcell with a ≈ 3.84 Å,
)
The mineral perovskite CaTiO
and practical importance in many disciplines such as mineral-
3
is of both fundamental interest
early stages of the crystal growth are studied, and a detailed
formation mechanism of these hollow crystals is established.
1
1
12
13
ogy, solid-state chemistry, materials sciences, electronic
Experimental Section
1
4
15
engineering, and even biotechnology due to its unique
structure, easy fabrication, high stability, and biocompatibility.
Much effort has been devoted to its synthesis, structural analysis,
and application as an electronic or optical material. In our
3
Synthesis. The synthetic method for CaTiO hollow crystals was
9
similar to that described in our previous report with a minor
modification. For sample I, 1.0 mmol of solid calcium nitrate,
3 2
Ca(NO ) , was directly dissolved in 19.67 mL of PEG -200 solvent
without adding any water. An ultrasonic treatment was performed to
enhance the dissolution of the calcium nitrate powder, and then 0.33
preliminary research, one type of CaTiO
3
hollow cage has been
produced. However, our knowledge of the formation of this
hollow CaTiO and other hollow crystals is still very limited.
Herein, we report three different types of microstructured
CaTiO hollow crystals synthesized via a simple poly(ethylene
9
4 9 4
mL of tetrabutyltitanate [titanium n-butoxide, Ti(OC H ) , TNB] was
3
added dropwise into the solution under vigorous stirring, followed by
addition of 22 mmol of sodium hydroxide powder under stirring to
serve as the mineralization reagent. Subsequently, the mixture feed-
stock was transferred into a Teflon-lined stainless steel autoclave for
solvothermal treatment at 180 °C for 15 h. The vessel was then cooled
to ambient temperature. The precipitate was recovered by centrifuga-
tion, washed with acetone, diluted acetic acid, and distilled water, and
then dried in a desiccator at ambient temperature.
3
glycol) 200 (PEG-200)-assisted solvothermal procedure but with
varying small amounts of water in the solution. In particular,
(
(
6) (a) Cao, H. L.; Qian, X. F.; Wang, C.; Ma, X. D.; Yin, J.; Zhu, Z. K.
J. Am. Chem. Soc. 2005, 127, 16024–16025. (b) Jiao, S. H.; Xu, L. F.;
Jiang, K.; Xu, D. S. AdV. Mater. 2006, 18, 1174–1177. (c) Zhang,
W. X.; Chen, Z. X.; Yang, Z. H. Phys. Chem. Chem. Phys. 2009, 11,
For samples II and III, the precursor for Ca, Ca(NO
was replaced by aqueous Ca(NO solution (4.0 and 1.0 M,
respectively). The volume ratios of Ca(NO solution to PEG-200
3 2
) powder,
3 2
)
3 2
)
6
263–6268.
are adjusted to 0.25 mL (4.0 M):19.42 mL for sample II and 1.00
mL (1.0 M):18.67 mL for sample III, corresponding to water
contents of 1.25 and 5 vol %, respectively, in the final solutions
after addition of 0.33 mL of TNB. For examination of early-stage
crystals, specimens after shorter reaction times, e.g., 1, 2, and 5 h,
were also collected for each sample.
7) (a) Chen, J. Y.; McLellan, J. M.; Siekkinen, A.; Xiong, Y. J.; Li, Z. Y.;
Xia, Y. N. J. Am. Chem. Soc. 2006, 128, 14776–14777. (b) Lu, X. M.;
Au, L.; McLellan, J.; Li, Z. Y.; Marquez, M.; Xia, Y. N. Nano Lett.
2
007, 7, 1764–1769.
(
(
8) Wang, W. Z.; Poudel, B.; Wang, D. Z.; Ren, Z. F. AdV. Mater. 2005,
1
7, 2110–2114.
9) Yang, X.; Williams, I. D.; Chen, J.; Wang, J.; Xu, H.; Konishi, H.;
Characterization. The powder X-ray diffraction (XRD) patterns
of the products were recorded with a Rigaku D/MAX 2200 VPC
diffractometer using Cu KR radiation (λ ) 0.15406 nm) and a
graphite monochromator. Scanning electron microscopy (SEM)
images were taken using an FEI Quanta 400 Thermal FE scanning
electron microscope. The samples for transmission electron mi-
croscopy (TEM) studies were prepared by dispersing the powder
specimen on a holey carbon film supported on a copper grid, and
the images were recorded on a JEOL JEM-2010HR electron
microscope operated at 200 kV and equipped with a Gatan GIF
Tridiem system.
Pan, Y.; Liang, C.; Wu, M. J. Mater. Chem. 2008, 18, 3543–3546.
(
(
10) Teo, J. J.; Chang, Y.; Zeng, H. C. Langmuir 2006, 22, 7369–7377.
11) Hu, M. S.; Wenk, H. R.; Sinitsyna, D. Am. Mineral. 1992, 77, 359–
3
73.
12) Mather, G. C.; Islam, M. S.; Figueiredo, F. M. AdV. Funct. Mater.
007, 17, 905–912.
(
(
2
13) (a) Lee, W. T.; Salje, E. K. H.; Goncalves-Ferreira, L.; Daraktchiev,
M.; Bismayer, U. Phys. ReV. B 2006, 73, 214110. (b) Huang, Y. J.;
Chiu, H. T.; Lee, C. Y. CrystEngComm 2009, 11, 1904–1909. (c)
Huang, Y. J.; Tsai, M. C.; Chiu, H. T.; Sheu, H. S.; Lee, C. Y. Cryst.
Growth Des. 2010, 10, 1221–1225. (d) Wang, D. A.; Guo, Z. G.; Chen,
Y. M.; Hao, J.; Liu, W. M. Inorg. Chem. 2007, 46, 7707–7709. (e)
Croker, D.; Loan, M.; Hodnett, B. K. Cryst. Growth Des. 2009, 9,
2
207–2213.
Results and Discussion
(
(
14) Wang, X. S.; Xu, C. N.; Yamada, H.; Nishikubo, K.; Zheng, X. G.
AdV. Mater. 2005, 17, 1254–1261.
The three samples produced with different contents of water
in the synthetic solutions gave very similar XRD patterns as
shown in Figure 1. All the diffraction peaks can be indexed to
15) Inoue, M.; Rodriguez, A. P.; Takagi, T.; Katase, N.; Kubota, M.; Nagai,
N.; Nagatsuka, H.; Nagaoka, N.; Takagi, S.; Suzuki, K. J. Biomater.
Appl. 2010, 24, 657–672.
1
4280 J. AM. CHEM. SOC. 9 VOL. 132, NO. 40, 2010

Formation Mechanism of CaTiO
3
Hollow Crystals
A R T I C L E S
Figure 2. (a) Low-magnification SEM image of the as-synthesized CaTiO
3
product prepared in the water-free system. The inset in the middle shows a
corresponding cross-section TEM image of a particle showing a hollow structure. The inset on the right side is a higher magnification SEM image showing
the walnut-like surface. (b,c) SEM images of the cubic CaTiO particles from samples produced in the 1.25 and 5 vol % water systems, respectively. The
insets are corresponding SEM images of broken cubes, revealing the hollow structures.
3
the orthorhombic CaTiO
3
(JCPDS card No. 82-0229, unit cell
The shape of the particles is basically spherical. Figure 3 shows
more detailed microstructure of these hollow particles. It can
be seen that the walnut-like shells are piled up by nanocubes
of about 40 nm in average size. The selected area electron
diffraction (SAED) pattern from the whole hollow particle in
Figure 3b shows that all the nanocubes are perfectly oriented
according to the cubic subunit cell; i.e., all the nanocubes are
a ) 5.4086, b ) 5.4553, and c ) 7.6782 Å). Regarding the
peak intensities, all the strong peaks can be indexed to a
pseudocubic perovskite subcell with the cell dimension of a ≈
3
.84 Å. The orthorhombic unit cell can be regarded as a ꢀ2 ×
ꢀ2 × 2 superunit cell based on the cubic perovskite subcell,
leading to a cubic-to-orthorhombic transformation of the Miller
indices, e.g., (100) to (110), (110) to (020), (200) to (220),
c
c
c
c
connected with the {100} faces. However, if the real orthor-
etc., where the subscript c denotes the cubic subcell. It is well
known that the formation of the superstructure is due to regular
hombic unit cell is considered, the crystal orientations of the
nanocubes are along three zone axes, [110], [110], and [001],
j
tilting of the TiO
0.817) of CaTiO
6
octohedra because the lattice tolerance factor
is much smaller than 1 (Figure S1, Supporting
as described in Figure 4. The experimentally observed SAED
pattern (inset of Figure 3b), although it looks like it results from
a 2×2 superstructure based on the perovskite subunit cell, can
be regarded as a combination of the three patterns, and only
the three patterns, of the orthorhombic unit cell, corresponding
(
3
Information). It is interesting to notice that the relative intensities
of the (100) and (200) peaks of the sample obtained with 5
c
c
vol % water are significantly increased compared to those of
the other two samples with less water in the synthetic systems.
SEM micrographs of the product shown in Figures 2a depict
that the uniformly dispersed particles yielded from the water-
free system have a walnut-like morphology with an average
size of about 600 nm. The TEM micrograph of a thin slice of
an individual particle clearly reveals a hollow interior with a
shell thickness of about 200 nm (middle inset in Figure 2a).
c c c
to the orientations of [010] , [100] , and [001] of the pseudocu-
bic subunit cell.
To confirm the above conclusion, HRTEM images of several
individual nanocubes were examined. For example, the HRTEM
images and the corresponding FFT patterns of the marked areas
in Figure 3b reveal that each cubic building block is a single
3
crystalline orthorhombic CaTiO with one of the above-
J. AM. CHEM. SOC. 9 VOL. 132, NO. 40, 2010 14281

A R T I C L E S
Yang et al.
Figure 3. (a) SEM image and (b) cross-section TEM image of individual walnut-like CaTiO
3
particles produced in the water-free system. The inset of (b)
is the corresponding SAED pattern from the whole hollow particle consisting of many nanocubes. (c,d) HRTEM images of the marked areas in (b) with the
FFT patterns from individual nanocubes. The [001] axes of these nanocubes are indicated by arrows.
tolerant factor has little effect on the morphology of the
monocrystalline nanocubes and the orientations of these nanocubes
in aggregates. The formation of nanocubes with six {100} facets
c
implies that the minimum surface energies achieved by these
facets along the three different axes are almost uniform. The
particles behave like cubic crystals. The principal reason for
such a phenomenon is that the difference in the adsorption of
the PEG-200 molecules on these {100}
c
surfaces is not
significant. Consequently, the adsorption of the PEG-200
molecules concealed the difference between these {100}
surfaces.
c
Compared to the walnut-like hollow particles obtained from
the water-free system as described above, particles with a more
regular cubic shape and smoother surface were fabricated by
adding a small amount of water (e.g., 1.25 vol % for sample II
and 5 vol % for sample III) into the system, with other reaction
conditions unchanged (Figure 2b,c). Both the particle sizes and
the uniformities are similar to those of the walnut-like particles
in sample I. The broken particles clearly show they are also
hollow, with larger empty cores.
Figure 4. Schematic drawing to explain a combination of three SAED
patterns (a-c) along the different zone axes of the orthorhombic unit cell
of CaTiO
b.
3
to result in (d) the SAED pattern shown in the inset of Figure
3
mentioned different orientations (Figure 3c,d). Such a micro-
structure and the formation of the nanocubes indicate that the
orthorhombic distortion of CaTiO due to the small lattice
3
The HRTEM images and SAED patterns show that the
microstructures of these two samples are different. Figure 5a
1
4282 J. AM. CHEM. SOC. 9 VOL. 132, NO. 40, 2010

Formation Mechanism of CaTiO
3
Hollow Crystals
A R T I C L E S
Figure 5. (a) TEM image of a broken hollow box from the specimen produced from the 1.25 vol % water system. (b) TEM image of specimen produced
from the 5 vol % water system. The insets in (a) and (b) are the corresponding SAED patterns from the whole particles. (c,d) HRTEM images from the shells
of the particles of the samples produced from 1.25 and 5 vol % water systems, respectively.
shows a TEM image of a broken hollow cube of sample II.
The SAED pattern from a large area of the shell is very
similar to that in the inset of Figure 3b, indicating a
combination of three different diffraction patterns, as de-
scribed in Figure 4. This conclusion is confirmed by a
textured structure revealed by the HRTEM images of the
shells as shown in Figure 5c. Consequently, the shells consist
perfectly indexed to the [110] zone axis of the orthorhombic
unit cell of CaTiO , as shown in Figure 4a,b. The single
crystalline property of sample III has been confirmed by
HRTEM imaging, as shown in Figure 5d. Such well-oriented
3
c
single crystalline hollow cubes with six {100} facets, or four
{110} and two {001} facets according to the orthorhombic unit
cell, can explain why the intensities of the (110,002) and
(220,004) peaks in its XRD pattern are higher than for the other
two samples (samples I and II) (Figure 1).
of nanodomains of orthorhombic CaTiO
orientations, as mentioned in Figure 4.
3
with three different
In contrast, the shells of hollow cubes in sample III, prepared
in the 5 vol % water system, are truly single crystalline, as seen
from the SAED pattern of the inset of Figure 5b, recorded from
a whole broken cubic particle. This SAED pattern can be
To further investigate the formation mechanism of these
hollow crystals, series specimens with shorter reaction times
were collected in each synthesis. It is obvious that, in the water-
3
free system for sample I, nanocubes of CaTiO formed at an
J. AM. CHEM. SOC. 9 VOL. 132, NO. 40, 2010 14283

A R T I C L E S
Yang et al.
Figure 6. TEM images of particles in sample I obtained from the water-free system with different crystal growth times: (a) 1, (b) 2, and (c) 5 h. The
corresponding SAED patterns are shown below (d-f).
early stage of the solvothermal process. The PEG-200 polymer
molecules adsorbed on the {100} surfaces of the nanocubes
3
Since cubic CaTiO is a thermodynamically stable phase at
c
temperatures over 1250 °C for bulk materials, it is interesting
to note that it can be synthesized at much lower temperatures
(180 °C) and exists at ambient temperature when the material
is in nanoscale. Similar size-induced phase transformations have
can enhance the oriented assembly of these nanocubes to form
spherical aggregates, as observed in many similar processes,
although the detailed mechanism is yet to be studied. Figure 6
shows TEM images of typical particles from specimens after
reaction for 1, 2, and 5 h in the water-free system. It can be
seen that the particle size increases with the reaction time from
1
6
3
been previously observed in BaTiO nanocrystals; i.e., the
smaller crystal size favors the cubic phase of BaTiO
than the tetragonal phase.
3
rather
9
0 nm (1 h) to 150 nm (2 h) and then 600 nm (5 h) in diameter.
All the particles are aggregates of nanocubes which are about
0 nm or less in size. These polycrystalline particles are not
When a small amount of water (1.25 vol %) was added into
the reaction system, the primarily formed nanocubes also
underwent an oriented assembly to form spherical particles, as
shown in Figure 7a. The SAED pattern of the particle in Figure
7a confirms the oriented aggregation regarding the pseudocubic
subunit cell but random arrangement of three principal orienta-
tions according to the orthorhombic unit cell (Figure 7d). As
the particle sizes increased with longer thermal treatment, the
surface of the particles recrystallized, leading to morphology
change from spheres to cubes. At this stage, the cubes are
nonhollow (Figure 7b). As compared to sample I (Figure 6b),
surface recrystallization was enhanced by an addition of water,
which may reduce the number of adsorbed PEG-200 molecules
on crystal surface and increase the solubility of the crystals,
which helps recrystallization. The recrystallized surface of such
a particle is believed to have a higher crystallinity or to contain
some larger crystallites in comparison with the inner crystallites.
Therefore, a following Ostwald ripening process will result in
an extension of recrystallization from the surface to the core.
Eventually, hollow cubes can form, as shown in Figure 7c. The
cage volume further increased until the shell density approached
the maximum value and the Ostwald ripening stopped (Figures
2b and 5a). The epitaxial intergrowth of the orthorhombic
4
hollow at first (Figure 6a,b), although many voids are presented
randomly in the whole particles. A small hole was developed
after reaction for 5 h (Figure 6c), which became much larger in
the final product (Figures 2a and 3b). Another phenomenon we
discovered is that the size of nanocubes in small aggregates is
also small in comparison with that in larger aggregates. A natural
consequence of the crystal growth in the solvothermal synthesis
is that the potential for aggregation is so high that the growth
of the nanocubes and their aggregation occur simultaneously.
The growth rate of the nanocubes in the aggregates decreased
significantly, leading to a gradual size increase of the nanocubes
from the center to the surface of the walnut-like particles. The
particles then underwent an Ostwald ripening process within a
further thermal treatment when the relatively smaller nanocubes
disappear to form hollow centers. The cage size increases with
the time of the thermal treatment. At the same time, the density
of the shells becomes higher when the size of the nanocubes
increases, as estimated from the image contrast (see Figures 3b
and 6c).
The SAED patterns from these particles are also displayed
in Figure 6. It is obvious all the particles have almost perfect
oriented aggregation. Moreover, smaller particles at earlier stages
show less orthorhombic distortion from the cubic symmetry.
(
16) (a) Dutta, P. K.; Gregg, J. R. Chem. Mater. 1992, 4, 843–846. (b)
Suzuki, K.; Kijima, K. J. Alloys Compd. 2006, 419, 234–242.
1
4284 J. AM. CHEM. SOC. 9 VOL. 132, NO. 40, 2010

Formation Mechanism of CaTiO
3
Hollow Crystals
A R T I C L E S
Figure 7. TEM images of particles in sample II obtained from the system containing 1.25 vol % water with different crystal growth times: (a) 1, (b) 2, and
c) 5 h. The corresponding SAED patterns are shown below (d-f).
(
Figure 8. TEM images of particles in sample III obtained from the system containing 5 vol % water with different crystal growth times: (a) 1, (b) 2, and
c) 5 h. The corresponding SAED patterns are shown below (d-f).
(
CaTiO
3
domains or the textured structure (Figure 5c) was
line-like property, regarding the orthorhombic unit cell of
CaTiO , was detected by SAED (Figure 8d-f) for all the
retained during the whole process of crystal growth.
3
In the 5 vol % water system, the development of the crystals
in sample III was similar to that of sample II. The crystallinity
of the former was much higher than the latter. A monocrystal-
particles, even when the cubic morphology was not developed
(Figure 8a). This indicates that water helped the nanocubes to
aggregate with a perfect orientation according to the orthor-
J. AM. CHEM. SOC. 9 VOL. 132, NO. 40, 2010 14285

A R T I C L E S
Yang et al.
Scheme 1. Schematic Illustration Showing Three Different Growth Mechanisms in (a) Water-Free, (b) 1.25 Vol % Water, and (c) 5 Vol %
a
Water Systems
a
j
3 3
. The light blue and dark blue cubes are orthorhombic CaTiO particles with random orientations of [110], [110],
The yellow cubes are cubic CaTiO
and [001].
hombic unit cell. Increasing the water content greatly suppressed
the adsorption of PEG-200 molecules on the crystal surface and
enhanced a better oriented aggregation of the nanocubes when
nanocubes, resulting in oriented aggregation according to both
the pseudocubic subunit cell and the real orthorhombic unit cell.
the difference of the {100}
c
became obvious. Surface recrys-
Conclusions
tallization followed by Ostwald ripening resulted in single
crystalline hollow cubes, which was also confirmed by HRTEM
studies (Figure 5d). It was noticed that, similar to sample II,
the outer surface of the cubic shells is smooth, while the inner
surface is quite rough.
The phenomenon of surface recrystallization followed by an
extension of recrystallization from the surface to the core of a
polycrystalline aggregate is similar to the reversed crystal growth
of some zeolites previously reported by one of the authors, Zhou
1
7
and co-workers. In Zhou’s recent publication, the general
The experimental results obtained in the present work clearly
18
mechanism of the crystal growth is discussed. At a very early
demonstrate the formation mechanisms of the CaTiO
3
hollow
stage of crystal growth, there is a competition between the
aggregation and growth processes. If crystal growth is fast,
individual crystallites quickly approach a certain size, at which
aggregation becomes difficult. This is the classic growth route.
If aggregation occurs before the crystals become too large due
to a strong van de Waals interaction between the chemical
building units, the crystal growth of individual crystallites is
suppressed. Surface recrystallization of the aggregated particles
will take place, and the reversed crystal growth route will be
crystals as illustrated in Scheme 1. The nanocubes of CaTiO
3
form at an early stage in all three synthetic systems. In the water-
free reaction system for sample I (Scheme 1a), the nanocubes
undergo an oriented aggregation according to the pseudocubic
subunit cells enhanced by the polymer PEG-200, which conceals
c
the difference of the surface energies of {100} . Therefore,
according to the orthorhombic unit cell, the intergrown faces
j
are randomly (110), (110), and (001). The sizes of the nanocubes
in the nonhollow spherical aggregates increase from the center
to the surface. An Ostwald ripening process then leads to the
disappearance of smaller nanocubes in the central area, forming
hollow walnut-like hollow particles. When 1.25 vol % water is
added for sample II (Scheme 1b), a surface recrystallization
takes place in the spherical aggregates of the nanocubes to form
a high-density cubic shell with a textured structure. Since the
surface of the walnut-like particles in sample I was not
recrystallized into a single-crystal shell, gaps between the
nanocubes were obvious, as seen in SEM and TEM images
3
followed. Zhou predicted that the hollow CaTiO cubes shall
form with a mechanism similar to the latter route, i.e., the
reversed crystal growth route, as for the formation of hollow
cubes of zeolite A. A time-dependent investigation would reveal
more details. The present work confirms this prediction and
demonstrates the first ceramic example of the reversed crystal
growth. It shows again that when aggregation dominates in the
early stages of crystal growth, the classic theory of crystal
growth may not be followed. A single crystal may not be
developed from a single nucleus, and its regular polyhedral
morphology may not be directly related to the different crystal
growth rates along different zone axes. Surface recrystallization
of polycrystalline aggregates will select a morphology to keep
a minimum surface free energy, as predicted by Curie and Wulff
(
Figures 2a and 3). A textured structure was developed during
surface recrystallization in sample II. Therefore, the grain
boundaries were very fine and could only be observed in
HRTEM images (Figure 5c). Sample III had a true single-crystal
shell without grain boundaries. Water plays an important role
in recrystallization that extends from the particle surface to the
cores via an Ostwald ripening process. A hole forms eventually
in the center of each particle. If 5 vol % water is added in the
reaction system for sample III (Scheme 1c), it not only enhances
the surface recrystallization to form a cubic morphology but
also facilitates self-adjustment of the orientations of the
(
17) (a) Chen, X. Y.; Qiao, M. H.; Xie, S. H.; Fan, K. N.; Zhou, W. Z.;
He, H. Y. J. Am. Chem. Soc. 2007, 129, 13305–13312. (b) Yao, J. F.;
Li, D.; Zhang, X. Y.; Kong, C. H.; Yue, W. B.; Zhou, W. Z.; Wang,
H. T. Angew. Chem., Int. Ed. 2008, 47, 8397–8399. (c) Greer, H.;
Wheatley, P. S.; Ashbrook, S. E.; Morris, R. E.; Zhou, W. Z. J. Am.
Chem. Soc. 2009, 131, 17986–17992.
(18) Zhou, W. Z. AdV. Mater. 2010, 22, 3086–3092.
1
4286 J. AM. CHEM. SOC. 9 VOL. 132, NO. 40, 2010

Formation Mechanism of CaTiO
3
Hollow Crystals
A R T I C L E S
1
9
1
00 years ago. This mechanism can be used to explain the
their fruitful discussions. This work was financially supported by
National Natural Science Foundation (NNSF) of China, the Govern-
ment of Guangdong Province and Guangzhou City (Nos. U0734002,
formation of many other hollow crystals, including some natural
minerals. For example, some precious agate stones have a hole
inside with or without liquid, the latter being often called water
agate. These agate stones may form via the reversed crystal
growth, similar to what we observed in the present work. The
better understanding of the formation of the hollow crystals will
enable us to control the syntheses of these materials for various
applications.
50872158, 10774195, 8251027501000010 and 10C22051347), and the
China Postdoctoral Science Foundation (No. 20080440117). W.Z.Z
thanks the Royal Society for financial support to an International
Collaboration project in this field.
Supporting Information Available: Structural description of
the orthorhombic unit cell and the pseudocubic subcell of
CaTiO . This material is available free of charge via the Internet
3
Acknowledgment. We gratefully thank Dr. Yanbin Wang at the
University of Chicago and Dr. Qisheng Huo at Jilin University for
at http://pubs.acs.org.
(
19) (a) Curie, P. Bull. Soc. Fr. Mineral. Cristallogr. 1885, 8, 145–150.
b) Wulff, G. Z. Kristallogr. 1901, 34, 449–480.
(
JA106461U
J. AM. CHEM. SOC. 9 VOL. 132, NO. 40, 2010 14287