View Article Online / Journal Homepage / Table of Contents for this issue
PAPER www.rsc.org/crystengcomm | CrystEngComm
(111) Twinned BaTiO3 microcrystallites†
Shubin Qin, Duo Liu,* Feifei Zheng, Zhiyuan Zuo, Hong Liu and Xiangang Xu
Received 8th December 2009, Accepted 14th May 2010
DOI: 10.1039/b925863a
We report here the controlled synthesis of (111) twinned BaTiO3 microcrystallites, which are widely
known for their (110) ferroelectric twins. This (111) twinned structure was obtained through
a composite hydroxide mediated process by using amorphous TiO2 powders. A twin assisted re-entrant
growth model is proposed to account for the microstructural aspects determined. It is considered that
the (111) twins originate from the Ti–O octahedra of the amorphous TiO2 powders. The role of oriented
aggregation is also discussed.
Published on 02 June 2010 on http://pubs.rsc.org | doi:10.1039/B925863A
Downloaded by Linkopings universitetsbibliotek on 02 March 2013
Introduction
Twinned crystals have been known by mankind for a long time
owing to their unique morphologies Besides, twinned structures
are also technologically important due to their unique mechan-
ical, electrical, and optical properties. In particular, twins in
ferroelectric perovskites are especially of interests due to their
close relationships linked to ferroelectric domain switching,
which enables a variety of device applications.1 As a model
ferroelectric perovskite, barium titanate (BaTiO3) is best known
for its room temperature tetragonal structures characterized by
(110) twins. Besides the ferroelectric domain twins, structural
twins such as (111) twin have also been reported for single crystal
BaTiO32,3 and bulk ceramics.4 However, (111) twins of BaTiO3
can only be formed at high temperatures (1300  C) and under
the conditions of compositional non-stoichiometry.5,6 Moreover,
penetrated BaTiO3 structures similar to the (111) twinned CaF2
and diamond7,8 have not been reported.
Recently, we reported that BaTiO3 powders exhibited photo-
chromic and ferromagnetic properties.9 To better understand the
unusual observations, in the present work, we studied the
microstructural characteristics and the formation mechanism of
the BaTiO3 (111) twins in detail. The (111) twinned BaTiO3
microcrystallites can be formed at low temperatures (<200  C) by
using amorphous TiO2 as the starting material. The micro-
crystallites exhibit multiple penetrated structures as evidenced by
scanning electron microscopy (SEM) and atomic force micros-
copy (AFM). A growth model is proposed and the technical
potentials of this unique structure are discussed.
Experimental
The BaTiO3 microcrystallites were synthesized through a two-
step composite hydroxide mediated (CHM) approach by taking
advantage of the low eutectic point of mixed NaOH and KOH
solids.10 Anatase TiO2 and BaCl2 were employed as the raw
materials. First, the anatase TiO2 powders were pretreated in
molten hydroxides to form an amorphous phase. During this
State Key Laboratory of Crystal Materials, Shandong University, 27 South
Shanda Road, Jinan, Shandong, 250100, P. R. China. E-mail: liuduo@sdu.
edu.cn; Fax: +86 531-88362807; Tel: +86 531-88363901
† Electronic supplementary information (ESI) available: Fig. S1 and S2.
See DOI: 10.1039/b925863a
This journal is ª The Royal Society of Chemistry 2010
process, 0.5 mmol of TiO2 and 9 g of composite hydroxides
(NaOH : KOH ¼ 0.515 : 0.485) were added into a 25 ml Teflon-
lined autoclave. The autoclave was sealed and heated at 180  C
for 5 h. After the autoclave was cooled down to room temper-
ature, 2 ml of 0.3M BaCl2 solution was added. The sealed
autoclave was heat treated again at 180  C for a range of time
from 4 h to 20 days. The synthesized powders were collected and
washed thoroughly with deionized water to remove the
impurities.
The microstructures of the as-synthesized products were
investigated by using a powder X-ray diffractometer (XRD,
Bruker D8 Advance, CuKa, Ni filter, l ¼ 1.540598 A,  40 kV,
40 mA). Raman analysis was carried out at room temperature on
a NEXUS 670 spectrometer equipped with a Nd:YVO4 laser
(1064 nm) and an InGaAs detector. The morphologies were
determined by using a field emission scanning electron micro-
scope (FESEM, Hitachi S-4800) and an atomic force microscope
(AFM, D3100, Nanoscope IIIa, Digital Instruments).
Results and discussion
General
Fig. 1a shows XRD patterns of the final products obtained after
CHM treatment of various periods. All the diffraction peaks can
be indexed to BaTiO3 (P4/mmm, JCPD 79-2264) without any
evidence of a hexagonal phase. BaTiO3 formed after reaction for
4 h and the diffraction intensities increased as the reaction pro-
ceeded. The composition of the product is also characterized by
energy dispersive spectroscopy (EDS), which indicates that there
are only barium, titanium and oxygen in the sample (Fig. S1,
ESI†). As Raman spectroscopy is very sensitive to structural
phase transitions, it is used to verify that only the tetragonal
phase is present in the final product after reaction for 20 days. As
shown in Fig. 1b, all the Raman modes can be attributed to
tetragonal BaTiO3.11,12 The Raman bands located at 265 cm1
and 518 cm1 are assigned to TO vibrations of A1 symmetry. The
band at 715 cm1 can be attributed to the highest frequency LO
mode of A1 symmetry. The sharp peak located at 305 cm1 is
associated with the B1 mode. The modes at 305 cm1 and
715 cm1 indicate that the samples are tetragonal BaTiO3.12 No
Raman bands can be attributed to hexagonal BaTiO3, which
exhibits characteristic Raman peaks at 153 cm1 and 640 cm1.13
CrystEngComm, 2010, 12, 3003–3007 | 3003

Published on 02 June 2010 on http://pubs.rsc.org | doi:10.1039/B925863A
Downloaded by Linkopings universitetsbibliotek on 02 March 2013
Fig. 1 (a) XRD patterns of BaTiO3 synthesized at 180  C for 4 h, 8 h,
12 h, and 20 days,respectively. (b) Raman spectrum of BaTiO3 micro-
crystallites synthesized at 180  C for 20 days. (c) Raman spectrum of the
pretreated TiO2.
Fig. 2a shows a typical SEM image of BaTiO3 microcrystallites
grown at 180  C for 20 days. The crystallites are 30–40 mm in
size and cubic in shape with penetrated structures. The cubic
elements could overlap and form a chained structure (Fig. 2b).
Small cubes could also be embedded into large ones (Fig. 2c)
where Ostwald coarsening could be observed. A statistical
analysis reveals that (1) all the exposed surfaces are (100) planes
due to their relatively low surface energy,14 (2) that the cubes
intersect their neighbors in the [221] directions and form a re-
entrant angle of 109.5 . As a result, it can be tentatively
concluded that the growth process involves (111) twins.
The role of pretreated TiO2
The utilization of pretreated anatase TiO2 is found to be crucial
for obtaining (111) twinned BaTiO3. The synthesis route solely
3004 | CrystEngComm, 2010, 12, 3003–3007
View Article Online
Fig. 2 SEM images of (111) twinned BaTiO3 microcrystallites synthe-
sized at 180  C for 20 days.
using as-purchased anatase TiO2 results in separated single
crystal BaTiO3 nanocubes, as demonstrated by previous inves-
tigations.10,15 As a result, microstructural features of the pre-
treated TiO2 need to be clarified. Our XRD analysis reveals that
the pretreated product is amorphous. It is well known that
anatase TiO2 can be easily dissolved in alkaline solution at
elevated temperatures.16,17 In alkaline solutions, Ti–O octahedra
may combine with hydroxyl groups or other Ti–O octahedra to
form clusters, which yields short range ordered Ti–O octahedra
units. This result has been confirmed by previous studies, which
showed that amorphous TiO2 contain large numbers of defects,
i.e. oxygen vacancies and distorted Ti–O octahedra.18,19 Bursill
et al. showed that face-sharing octahedra can exist in non-stoi-
chiometric rutile (TiO2x).20 Recent studies by Penn et al. also
revealed twinned structures in aged TiO2 nanoparticles.21,22 This
short range ordered structure is also confirmed by Raman
analysis shown in Fig. 1c, where Ti–O vibrations of various
forms can be easily identified. These vibrations are characterized
by broadened bands with several peaks associated with either
rutile or anatase TiO2. Previous studies proved that the BaTiO3
(111) twin plane is composed of oxygen-deficient face-sharing
Ti–O octahedra,23 so that the pretreated TiO2 could play an
important role in the formation of (111) twins in BaTiO3.
Formation mechanism
As mentioned previously, (111) twins have been reported in
BaTiO3 single crystals and ceramics.24 It was considered that the
(111) twins originate from the oxygen-deficient hexagonal
BaTiO3 and could be synthesized through control of oxygen
This journal is ª The Royal Society of Chemistry 2010
 View Article Online

partial pressure at high temperatures (>1300  C)23,25 A twin

assisted model has been proposed to account for the high growth
rate at the re-entrant edges where the nucleation barrier is
considerably lower.26–28 However, we believe that the mechanism
governing the formation of (111) twinned BaTiO3 in a CHM
environment is fundamentally different and should be closely
linked to the microstructural aspects of the pretreated anatase
TiO2.
A schematic representation of the growth model is shown in
Fig. 3. Fig. 3a shows a pair of face-sharing Ti–O octahedra. A
BaTiO3 nucleus containing a (111) twin plane is formed when
Ba2+ cations were captured by the face-sharing Ti–O octahedra
(Fig. 3b). The re-entrant angle between the neighboring BaTiO3
lattices reduces the nucleation barrier (to 5% of the value on
Published on 02 June 2010 on http://pubs.rsc.org | doi:10.1039/B925863A

a flat interface)25 and accelerates the growth rate significantly. As

the twin plane proceeds, a penetrated structure is formed. The
Downloaded by Linkopings universitetsbibliotek on 02 March 2013

ideal morphology of a (111) penetration twin is shown in Fig. 3c,

where two identical cubes penetrate coordinatively. However,
local thermodynamic instabilities also enable an Ostwald coars-
ening process, which may lead to the growth of larger cubes in
compensation of the smaller ones. In addition, multiple (111)
twins promote growth rate in different directions. As the super-
saturation decreases, the morphologies of the BaTiO3 particles
could resemble the patterns shown in Fig. 3d,e.
Fig. 4 shows re-entrant angles between two cubes, based on
which the crystallographic directions of the twin planes can be
determined. Fig. 4a shows an exposed corner of BaTiO3, similar
to the schematic representation shown in Fig. 3e. The re-entrant
edge at the (111) twin plane in Fig. 4a provides nucleation sites
where favorable growths are expected. The (100) faces grow
through the conventional 2D growth model. Grooves are formed
at the interfaces between the two adjacent cubes due to lattice
mismatch between the (221) plane and the (100) plane. Growth
steps can be clearly visualized in Fig. 4b. Nuclei can be easily
formed at re-entrant edges (indicated by arrows) and provide
terrace sites for 2D growth. It is also found that in areas close to
the re-entrant edges the surface steps are of small height with
higher density. Both the height and the lateral separations
of these surface steps become greater at areas far from the
Fig. 3 Schematic representation of the evolutionary process of the (111)
twinned BaTiO3 crystallites. (b,c) are the ideal models and (d,e) are the
practical models of the as-grown BaTiO3 microcrystallites.
Fig. 4 SEM images of (111) twinned BaTiO3 microcrystallites show the
details of re-entrant edge (a) and layered growth traces (b).
re-entrant edges. It is known that impurities can block the
motion of surface steps on crystal surfaces, i.e., on KH2PO4 and
K2Cr2O7.29,30 As impurities such as metastable nuclei and Ti–O
clusters can not be avoided during the growth of BaTiO3
microcrystallites, their accumulations on the (100) surfaces create
pinning sites and hinder monolayer growth. It appears that
macrosteps may overcome the pinning centers due to the high
degree of supersaturation.29 Consequently, bunches or macro-
steps can be clearly visualized.
Contact mode AFM measurements were performed on the
(100) surface of the as-grown BaTiO3 microcrystallites, as shown
in Fig. 5a. It can be clearly verified that foreign particles block
the proceedings of the growth layers where a convex front can be
found. Fig. 5b shows a cross-section plot of some surface steps.
Typically, they are 2–3 nm in height, containing 5–8 unit
 c ¼ 4.018 A
lattices (a ¼ 3.9998 A,  for BaTiO3). Fig. 5c shows
a hole of 70 nm in depth. The possible reason for the formation
of holes lies in the fact that the surface layers can not cover the
inclusions completely. Another surface feature of the crystallites
is numerous nuclei randomly distributed on the (100) surfaces
(Fig. 5a). All the nuclei exhibit sizes greater than 16 nm with an
average value of 23.2 nm (based on a statistical analysis of more
than 100 nuclei). As nuclei smaller than a critical value will
spontaneously diminish, our results suggest that BaTiO3 nuclei
have a critical size greater than 16 nm. This finding agrees with
many previous studies which revealed that it is almost impossible
to synthesize BaTiO3 nanoparticles smaller than 20 nm.31–33
Theoretically, the (100) surface of BaTiO3 is characterized by
a high value of surface energy (1.237 eV per surface unit cell),34
which may help to explain the large size of the critical nuclei.
To further clarify the growth process of (111) twinned BaTiO3
microcrystallites, the initial growth stages were studied. After the

This journal is ª The Royal Society of Chemistry 2010 CrystEngComm, 2010, 12, 3003–3007 | 3005

Published on 02 June 2010 on http://pubs.rsc.org | doi:10.1039/B925863A
Downloaded by Linkopings universitetsbibliotek on 02 March 2013
Fig. 5 (a) Surface morphology measured by AFM of the (100) face of
BaTiO3 microcrystallites grown at 180  C for 20 days; (b) and (c) are the
height profiles along the lines indicated in (a).
solution was heated at 180  C for 1 day, two typical morphol-
ogies can be found, as shown in Fig. 6a,b. Fig. 6a shows that
these BaTiO3 particles are spherical and contain many pene-
trated cubes of 60 nm similar to pure BaTiO3 nanocubes
synthesized by the CHM approach.10,15 It is evident that these
spherical particles contain multiple (111) twins. This structure
Fig. 6 SEM images of BaTiO3 synthesized at 180  C for 1 day (a,b) and
4 day (c,d), respectively.
3006 | CrystEngComm, 2010, 12, 3003–3007
View Article Online
shows many re-entrant edges which facilitate 2D nucleation such
that the growth rate of some cubes is greatly enhanced. Because
re-entrant edges will finally disappear,26 Ostwald ripening would
become more significant, especially when supersaturation
decreases. As a result, small cubes dissolve and large cubes grow
preferentially (Fig. 6d, reaction for 4 days). Occasionally, we also
observe penetrated structures with exposed (111) faces, as shown
in Fig. 6b. It is clearly shown that these particles are (111)
twinned. A typical (111) twinned BaTiO3 particle is also shown in
Fig. S2a.† The HRTEM image (Fig. S2b†) obtained from small
BaTiO3 particles also supports the proposed twin assisted growth
mechanism, where a (111) twin can be clearly visualized. After
reaction for 4 days, the twinned structures will grow and form
aggregations (Fig. 6c and 6d). Similar to many other nano-
structured materials,21,22,35 (111) twins could also be formed
through oriented aggregations of adjacent grains with crystal-
lographic orientations close to the expected twinning relation.36
Conclusions
BaTiO3 microcrystallites were successfully synthesized through
a two-step CHM approach at low temperature. It is interesting to
note that these microcrystallites contain multiple (111) twins. A
twin assisted growth model is proposed to account for the
observations. The multiple (111) twins were formed through face
sharing of Ti–O octahedra, intentionally created through pre-
treating anatase TiO2 powders. These unique ferroelectric
nanostructures are potentially of importance, provided that the
(111) twin planes may serve as pinning sites for the motion of
(110) ferroelectric twin walls. A thorough exploration of the
ferroelectric properties of this twinned structure is currently
under way.
Acknowledgements
The authors thank National Science Foundation of China
(NSFC) (Grant No. 50702031, 60974117, 50872070), the Excel-
lent Young Investigators Award Foundation of Shandong
Province (Grant No. BS2009CL021), and National Basic
Research Program of China (973 Program) (Grant No.
2009CB930503) for financial support.
References
1 D. Dragan, Rep. Prog. Phys., 1998, 61, 1267.
2 J. P. Remeika and W. M. Jackson, J. Am. Chem. Soc., 1954, 76, 940.
3 E. A. D. White, Acta Crystallogr., 1955, 8, 845.
4 V. J. Tennery and F. R. Anderson, J. Appl. Phys., 1958, 29, 755.
5 V. Krasevec, M. Drofenik and D. Kolar, J. Am. Ceram. Soc., 1990,
73, 856.
6 B. K. Lee, S. Y. Chung and S. L. Kang, J. Am. Ceram. Soc., 2000, 83,
2858.
7 I. Sunagawa, Crystals Growth, Morphology and perfection, Cambridge
University Press, New York, 2005.
8 M. A. Tamor and M. P. Everson, J. Mater. Res., 1994, 9, 1839.
9 S. Qin, D. Liu, Z. Zuo, Y. Sang, X. Zhang, F. Zheng, H. Liu and
X. G. Xu, J. Phys. Chem. Lett., 2010, 1, 238.
10 H. Liu, C. Hu and Z. L. Wang, Nano Lett., 2006, 6, 1535.
11 R. Asiaie, W. Zhu, S. A. Akbar and P. K. Dutta, Chem. Mater., 1996,
8, 226.
12 C. H. Perry and D. B. Hall, Phys. Rev. Lett., 1965, 15, 700.
13 N. G. Eror, T. M. Loehr and B. C. Cornilsen, Ferroelectrics, 1980, 28,
321.
This journal is ª The Royal Society of Chemistry 2010

14 R. I. Eglitis, G. Borstel, E. Heifets, S. Piskunov and E. Kotomin,
J. Electroceram., 2006, 16, 289.
15 S. Qin, D. Liu, H. Liu and Z. Zuo, J. Phys. Chem. C, 2008, 112,
17171.
16 Z. Y. Yuan, J. F. Colomer and B. L. Su, Chem. Phys. Lett., 2002, 363,
362.
17 Z. Y. Yuan and B. L. Su, Colloids Surf., A, 2004, 241, 173.
18 V. V. Hoang, Phys. Status Solidi B, 2007, 244, 1280.
19 J. P. Rino and N. Studart, Phys. Rev. B: Condens. Matter Mater.
Phys., 1999, 59, 6643.
20 L. A. Bursill, M. G. Blanchin, A. Mebarek and D. J. Smith, Radiat.
Eff. Defects Solids, 1983, 74, 253.
21 R. L. Penn and J. F. Banfield, Am. Mineral., 1998, 83, 1077.
22 R. L. Penn, J. Phys. Chem. B, 2004, 108, 12707.
23 A. Recnik, J. Bruley, W. Mader, D. Kolar and M. R€ uhle, Philos. Mag.
B, 1994, 70, 1021.
24 R. C. DeVRIES, J. Am. Ceram. Soc., 1959, 42, 547.
25 A. Recnik and D. Kolar, J. Am. Ceram. Soc., 1996, 79, 1015.
Published on 02 June 2010 on http://pubs.rsc.org | doi:10.1039/B925863A
Downloaded by Linkopings universitetsbibliotek on 02 March 2013
This journal is ª The Royal Society of Chemistry 2010
View Article Online
26 K. A. Hu, B. V. Hiremath and R. E. Newnham, Phase Transitions,
1986, 6, 153–164.
27 M. K. Kang, Y. S. Yoo and D. Y. Kim, J. Am. Ceram. Soc., 2000, 83,
3202.
28 H. Y. Lee and J. S. Kim, J. Am. Ceram. Soc., 2002, 85, 977.
29 T. A. Land, T. L. Martin, S. Potapenko, G. T. Palmore and
J. J. D. Yoreo, Nature, 1999, 399, 442.
30 A. D. Derksen, W. J. P. van Enckevort and M. S. Couto, J. Phys. D:
Appl. Phys., 1994, 27, 2580.
31 H. Xu, L. Gao and J. Guo, J. Eur. Ceram. Soc., 2002, 22, 1163.
32 H. J. Chen and Y. W. Chen, Ind. Eng. Chem. Res., 2003, 42, 473.
33 T. Yan, Z. G. Shen, J. F. Chen, X. L. Liu, X. Tao and J. Yun, Chem.
Lett., 2005, 34, 1196.
34 J. Padilla and D. Vanderbilt, Phys. Rev. B: Condens. Matter, 1997, 56,
1625.
35 R. L. Penn and J. F. Banfield, Science, 1998, 281, 969.
36 G. Kastner, R. Wagner and V. Hilarius, Philos. Mag. A, 1994, 69,
1051.
CrystEngComm, 2010, 12, 3003–3007 | 3007