630 DOI: 10.17344/acsi.2018.4286 Acta Chim. Slov. 2018, 65, 630-637 ©amnions Scientific paper Plate-Like Bi4Ti3O12 Particles and their Topochemical Conversion to SrTiO3 Under Hydrothermal Conditions Alja Čontala,1,2 Marjeta Maček Kržmanc1,* and Danilo Suvorov1 1 Advanced Materials Department, Jožef Stefan Institute Jamova cesta 39, 1000 Ljubljana, Slovenia 2 International Postgraduate School, Jamova cesta 39, 1000 Ljubljana, Slovenia * Corresponding author: E-mail: marjeta.macek@ijs.si Received: 25-05-2018 Abstract Plate-like Bi4Ti3O12 particles were synthesized using a one-step, molten-salt method from Bi2O3 and TiO2 nanopowders at 800 °C. The reaction parameters that affect the crystal structure and morphology were identified and systematically investigated. The differences between various Bi4Ti3O12 plate-like particles were examined in terms of the ferroelec-tric-to-paraelectric phase transition and the photocatalytic activity for the degradation of Rhodamine B under UV-A light irradiation. The results encouraged us to conduct further testing of the as-prepared Bi4Ti3O12 plate-like particles as templates for the preparation of plate-like SrTiO3 perovskite particles using a topochemical conversion under hydrothermal conditions. The characteristics of the Bi4Ti3O12 plates and the reaction parameters for which the SrTiO3 preserved the shape of the initial Bi4Ti3O12 template particles were determined. Keywords: Bi4Ti3O12; Aurivillius layered perovskites; Molten-salt synthesis; Plate-like Bi4Ti3O12; Template plates; To-pochemical conversion 1. Introduction Bi4Ti3O12 (BIT) is a member of the Aurivillius-type perovskites that have a layered structure in which pseu-do-perovskite (Bi2Ti3O10)2- units are alternating with (Bi2O2)2+ layers. As a piezoelectric material with a high Curie temperature of 675 °C, Bi4Ti3O12 has the potential for use in high-temperature piezoelectric applications.1,2 The most common synthesis approaches for the preparation of Bi4Ti3O12 ceramics and particles are solid-state, molten-salt, hydrothermal and some other methods with lower reaction temperatures, e.g., coprecipitation and sol-gel.3,4,5 During the preparation of Bi4Ti3O12 from Bi2O3 and TiO2 in molten salt or by solid-state reaction, secondary phases such as Bi12TiO20 or Bi2Ti2O7 are commonly formed and must be removed from the product prior to further use due to their detrimental effect on the piezoelectric properties of Bi4Ti3O12 ceramics.6-8 Platelike Bi4Ti3O12 particles with a side length of around 1 ^m and a thickness of 50 nm were formed in molten NaCl/KCl at 800 °C.9 In contrast to the thickness, the side length of the plates was not uniform, varying from 0.25 ^m to 3 ^m. He et al. systematically studied the influence of the salt/ precursor ratio on the morphology of plate-like Bi4Ti3O12 particles prepared in molten salt. 10 They observed that the average side length and thickness of the Bi4Ti3O12 plates decreased with an increase in the amount of salt; however, at high dilutions (molar ratios: NaCl:KCl:Bi4Ti3O12 >32:32:1) the influence of the salt content was no longer significant. These Bi4Ti3O12 (00/)-oriented plates were also proved to be more effective for the photocatalytic degradation of Rhodamine B (RhB) than irregularly shaped Bi-4Ti3O12 particles obtained using the solid-state method.10 Thickness, surface defects and faceting of the surface are all parameters that influence the photocatalytic activity (PA). Bi4Ti3O12 particles prepared with the hydrothermal method usually exhibited smaller dimensions and higher aggregation than those prepared in molten salt.11,12 Therefore, a decision has to be taken regarding the preparation method for the specific application of the material. In contrast to MTiO3 perovskites (M = Ba, Sr, Pb) with their cubic or tetragonal crystal structures, layered perovskites such as Aurivillius (Bi4Ti3O12 and MBi4Ti4O15, where M = Ba, Sr, Pb) and Ruddlesden-Popper (Sr3Ti2O7) Čontala et al.: Plate-Like Bi4Ti3O19 Particles Acta Chim. Slov. 2018, 65, 630-637 631 have a tendency to grow in an anisotropic shape due to their large crystal anisotropy. 13,14 These layered perovskites are often used as the structural templates for the preparation of anisotropic MTiO3 perovskites via topochemical conversion.15-17 The first use of topochemical conversion for the preparation of (100)-oriented SrTiO3 tabular particles was demonstrated by Watari et. al. 18 This conversion was performed by the reaction of a Ruddlesden-Popper Sr3Ti2O7 template and TiO2 in molten KCl between 1000°C and 1200 °C. Saito and Takao were the first to use an Aurivillius SrBi4Ti4O15 template for the preparation of SrTiO3 (100)-oriented plates with smaller dimensions and a higher aspect ratio (side length 5-10 ^m, thickness 0.5 ^m).19 The mechanism of the reaction between SrCO3 and SrBi4Ti4O15 in the molten salt was further studied by Chang et al., who optimized the reaction conditions to minimize the Bi remains.15 Under hydrothermal conditions SrTiO3 plate-like particles were produced using other templates such as HL07TiL73O4 ■ nH2O and sheet-like TiO2 mesocrystals.20,21 Bi4Ti3O12 particles were already used for a topochemical conversion to BaTiO3 or SrTiO3 in molten salt.9,22,23 Based on our previous work Bi4Ti3O12 is very appropriate template for preparation of BaTiO3 plate-like particles in the molten salt. 9 To the best of our knowledge, the topochemical conversion from Bi4Ti3O12 to BaTiO3, Sr-TiO3 or CaTiO3 particles under hydrothermal conditions has not yet been reported. It is well known that the quality and morphology of the initial precursor particles, in addition to the reaction conditions, greatly influence the particle morphology. In this study we report on the preparation of Bi4Ti3O12 particles with two morphologies and their topochemical transformation to SrTiO3 under hydrothermal conditions. 2. Methods 2. 1. Experimental Bi4Ti3O12 plates were synthesized using the molten-salt method. The salts KCl (Sigma-Aldrich, > 99.0%) and NaCl (Merck, > 99.7%) were weighed in a 1:1 molar ratio, ground and mixed well in a mortar to achieve a homogeneous mixture. Next, an appropriate amount of Bi2O3 nanopowder (99.9% Alfa Aesar) and TiO2 nanopowder (P25, Degussa) were weighed, added to the salt mixture and mixed well again. A homogeneous mixture of powder was then transferred to the Al2O3 crucible, covered and placed in the furnace. The heating rate was 10 °C/min until the temperature reached 800 °C. The morphology and crystal structure for the Bi4Ti3O12 obtained under different reaction conditions were investigated in order to determine the best reaction conditions for the preparation of well-developed, plate-like Bi4Ti3O12 particles with a narrow size distribution. The influence of various reaction conditions such as time (20 minutes and 2 hours), Bi:Ti molar ratio (1.33, 2.0 and 2.67), NaCl:KCl:Bi4Ti3O12 molar ratio (50:50:1 or 25:25:1) and cooling rate (5 °C/min, 10 °C/min and natural cooling) were examined. In the case of natural cooling, the heating system was turned off after the reaction was performed, and the product was left in the furnace until the furnace reached room temperature. Natural cooling means uncontrolled cooling in a furnace by natural convection, conduction and radiation to room temperature. It was demonstrated that the cooling time from 800 °C to the eutectic temperature of 650 °C in the case of natural cooling was 8 minutes, while for cooling rates of 10 °C/min and 5 °C/min, the cooling times were 15 and 30 minutes, respectively. In this temperature range, where the particle growth still took place, the natural cooling was the fastest. Below 650 °C, the diffusion is expected to be too slow and thus the growth of the particles is negligible. For comparison, Bi4Ti3O12 was also synthesized under selected conditions (Bi:Ti = 1.33, 800 °C 2 h, natural cooling) using anatase TiO2 (Sachtleben Pigments Oy, HOMBITAN AFDC 001517011, 99%) ^m-sized powder instead of the P25 nanopowder. After the reaction, the product powders were washed with deionized water by suction filtration in order to remove the salt. Afterwards, they were also washed with 2-M HNO3 (soaking time 10 minutes) in order to remove the secondary phases and finally with deionized water again (until pH = 7). The product powders were freeze-dried. In the second part of the study, Bi4Ti3O12 template plates were tested for the topochemical conversion to Sr-TiO3 under hydrothermal conditions. For this reaction SrCl2 x 6H2O (Sigma Aldrich, > 99%) was added to the Bi4Ti3O12 in the molar ratio Sr:Ti = 3. The reaction was performed under hydrothermal conditions in 4-M NaOH (Merck, 99%) with stirring at 200 °C for 12 hours. The product was cooled naturally and washed with deionized H2O, 1-M HNO3 and once again with deionized H2O. The product was freeze-dried. 2. 2. Characterization The crystal structure was characterized with X-ray powder diffraction (Bruker AXS D4 Endeavor) using Cu-Ka radiation (1.5406 A). For an estimation of the preferential orientation, a few drops of suspension of the particles in iso-propanol were deposited on a Si single crystal and left for the alcohol to evaporate. The morphology and size of the prepared particles were studied by field-emission scanning electron microscopy (FE-SEM, JSM 7600 F, JEOL). The specific surface area was measured using the BET method with nitrogen adsorption (Gemini 2370 V5.00). For the PA of the Bi4Ti3O12 particles, the decomposition of RhB was measured in UV-A light. The concentration of the sample, mixed with RhB and exposed to UV-A light, was 0.2 mg/ml, and the concentration of the RhB was 10 mg/l. Prior to irradiation, the solution was sonicated for Contala et al.: Plate-Like Bi4Ti3O12 Particles 632 Acta Chim. Slov. 2018, 65, 630-637 632 1 minute (pulse:pause was 2:1 seconds) at 80% and afterwards it was stirred in the dark at 500 rpm for 30 minutes. Samples were taken before irradiation and after 1 h, 2 h, 3 h, 4 h and 24 hours of irradiation with UV-A light. The powder was removed after centrifugation. The absorbance was measured at X = 554 nm using a Synergy Micro Plate Reader (BIOTEK). Control reactions in the dark for RhB with the sample and a parallel reaction for pure RhB were also performed. Differential scanning calorimetry (DSC) measurements were made on a Jupiter 449 simultaneous thermal analysis (STA) instrument (Netzsch, Selb, Germany). The measurements were made with a heating rate of 20 °C/min in an Ar/O2 (40/20) atmosphere using a TG/DSC-cp sample holder and platinum crucibles. The temperature and enthalpy calibrations of the STA instrument were made with BaCO3, CsCl, K2CrO4, KClO4 and RbNO3 standards. 3. Results and Discussion 3. 1. Influence of the Reaction Conditions on the Morphology and Crystal Structure Bi4Ti3O12 was formed in the molten salt by dissolution-precipitation. This means that Bi2O3 and TiO2 firstly dissolved in the molten salt and then Bi4Ti3O12 plates precipitated at a high degree of supersaturation. Further growth of the plates occurred by Ostwald ripening. Therefore, the particle size could be tailored with the duration of the reaction. Longer and shorter times were selected for the preparation of larger (^m-sized) and smaller (sub-^m-sized) Bi4Ti3O12 plates, respectively.9 In particular, sub-^m- and nano-sized anisotropic particles of ferroelectrics (Bi4Ti3O12 and BaTiO3) have recently become of great scientific and technological interest due to their unique shape- and size-dependent functional properties with reduced sub-^m-size dimensions. At first, the phase composition and morphology of the Bi4Ti3O12 particles obtained at 800 °C after 2 hours in the first experiment (BIT1) and 20 minutes in the second experiment (BIT2) were examined and compared (Fig. 1, Table 1). In both of these cases the reaction was performed with surplus Bi2O3 (Bi:Ti = 2.67, stoichiometric Bi:Ti = 1.33). The Bi2O3 was added in excess in order to provide a high concentration of Bi3+ for the formation of Bi4Ti3O12 plates. In the case of a shorter synthesis time (20 minutes), the Bi4Ti3O12 plates exhibited a smaller average side length as well as a reduced thickness compared to the plates obtained after a longer synthesis time (2 hours), when the particles had more time available for Ostwald-ripening growth. In both cases the XRD analyses showed that a significant amount of Bi12TiO20 secondary phase (PDF #034-0097) was present (Fig. 2, XRD pattern A) in addition to the Bi4Ti3O12 (PDF #035-0795). The formation of Bi12TiO20 was a consequence of excess Bi2O3 in the reaction mixture. During washing with 2-M HNO3 the Bi12TiO20 was dissolved and monoclinic Bi-4Ti3O12 became the only phase in both cases. Reaction conditions for the described Bi4Ti3O12 (BIT1 and BIT2) and for the Bi4Ti3O12 prepared under other conditions are presented in Table 1. Bi:Ti and NaCl:KCl:Bi4Ti3O12 are molar ratios. In the next step the molar ratio of bismuth to titanium (denoted as Bi:Ti in Table 1) was studied. We selected three values of Bi:Ti: 2.67, 2.0 and 1.33 (stoichiometric) for BIT3, BIT4 and BIT5, respectively. In the experiment using a ratio of 1.33 the plates had a more defined shape and a narrower size distribution (Fig. 1C) than in the other two cases and Bi4Ti3O12 was the main phase, observed by XRD after the synthesis. In the other two cases, washing with HNO3 was necessary to remove the secondary Bi12TiO20 phase. The influence of the amount of salt on the morphology and crystal structure was also investigated. In accordance with the literature10 we used a 50:50:1 molar ratio (BIT6) of KCl:NaCl:Bi4Ti3O12. However, when the ratio was reduced to 25:25:1 (BIT5) the size distribution of the product particles became more uniform. With a smaller amount of salt, the diffusion distance of the reactant particles is smaller; therefore, the reaction occurs faster and more uniformly. Powder XRD patterns confirmed the dominance of the monoclinic Bi4Ti3O12 phase (PDF #0350795) for both products. Table 1 : Reaction conditions of BIT1-BIT9 for the investigation of the influence of the reaction parameters on the morphology and crystal structure of Bi4Ti3O12 Sample TiO2 Bi:Ti NaCl:KCl:Bi4Ti3O12 T (°C) Time Heating Cooling BIT1 P25 2.67 50:50:1 800 2 h 10 °/min 10 °/min BIT2 P25 2.67 50:50:1 800 20 min 10 °/min 10 °/min BIT3 P25 2.67 25:25:1 800 2 h 10 °/min 10 °/min BIT4 P25 2.0 25:25:1 800 2 h 10 °/min 10 °/min BIT5 P25 1.33 25:25:1 800 2 h 10 °/min 10 °/min BIT6 P25 1.33 50:50:1 800 2 h 10 °/min 10 °/min BIT7 P25 1.33 25:25:1 800 2 h 10 °/min 5 °/min BIT8 P25 1.33 25:25:1 800 2 h 10 °/min natural BIT9 anatase 1.33 25:25:1 800 2 h 10 °/min natural Contala et al.: Plate-Like Bi4Ti3O12 Particles Acta Chim. Slov. 2018, 65, 630-637 633 Cooling rates of 10 °C/min (BIT6), 5 °C/min (BIT7) and natural cooling (BIT8-Fig. 2D) were also compared. Based on SEM observations we concluded that there is no significant difference in the morphology when the cooling rates were 10 °C/min, 5 °C/min or natural cooling. The most differing BIT products (BIT2 and BIT8) were selected for further analysis. SEM micrographs and XRD patterns for those plate-like Bi4Ti3O12 samples are shown in Figures 1 and 2, respectively. The higher degree of aggregation of BIT2 compared to BIT8 is evident from the SEM micrographs as well as from the XRD patterns of the Bi4Ti3O12 particles deposited on the Si single crystal (Figure 2, XRD patterns C and D). Due to the random orientation of the plate-like Bi4Ti3O12 particles in the aggre- Figure 1: SEM micrographs of the Bi4Ti3O12 plates that were synthesized at 800 °C for 20 min (B-BIT2) and for 2h (A-BIT1, C-BIT5, D-BIT8). Other synthesis details are shown in Table 1. Figure 2: XRD patterns of Bi4Ti3O12 plates prepared at 800 °C for 20 min ("patterns A and C for BIT2) and for 2 h (patterns B and D for BIT8). The A and B patterns represent the XRD patterns of Bi4Ti3O12 powders washed only by water in order to determine possible secondary phases. The XRD patterns C (for BIT2) and D (for BIT8) were obtained from Bi4Ti3O12 plates that were cast on the Si single crystal. The Bi4Ti3O12 plates prepared for examination of their average preferential orientation in these two patterns were washed with 2-M HNO3 to ensure a single-phase product. In the XRD pattern A, ♦ and * denote the Bi4Ti3O12 and Bi12TiO20 phases, respectively. The hkl indexation is given for Bi4Ti3O12. Contala et al.: Plate-Like Bi4Ti3O12 Particles 634 Acta Chim. Slov. 2018, 65, 630-637 634 gates of BIT2 sample, the relative intensities of the (00/)/ (117) planes (/ = even number) were considerably lower in comparison to those of the larger Bi4Ti3O12 (BIT8) plates, which oriented during deposition on a Si single crystal in a way that the (00/) planes were parallel to the substrate. Based on this we can infer that 1-2-^m-sized Bi4Ti3O12 plates exhibited a high (00/) preferential orientation (Figure 2, XRD pattern D). Finally, Bi4Ti3O12 was also prepared from ^m-sized anatase TiO2 particles (BIT9) instead of P25 TiO2 nano-powder, but under the same reaction conditions as for the BIT8. As expected, the plates were less uniform in size. The majority of the plates were much smaller (approx. 100 nm) and the rest were between 1 and 2 microns. This non-uniform particle size distribution was most probably a consequence of the larger anatase particles, which dissolved more slowly and unevenly than the P25 nanoparticles. Also in this case, the XRD pattern confirmed that mono-clinic Bi4Ti3O12 was the main phase. For smaller and larger Bi4Ti3O12 plates (BIT2 and BIT8) the specific surface area (BET), the photocatalytic activity (PA) and the DSC measurements were performed in order to confirm the significant difference in size and the specific surface area. The BET results were in agreement with the SEM observations. The specific surface area was larger for BIT2 than for BIT8 (Table 2), which is in accordance with the SEM observation and confirmed that the BIT2 particles were smaller and consequently the specific surface area was larger than for BIT8. Since it is well known that PA is influenced by many factors such as surface defects, particle size, morphology, crystallinity and band gap, we decided to test both types of plates in terms of their capability to degrade organic dye (RhB) under UV-A radiation. A high concentration of defects is not beneficial, neither for the photo catalysis nor for the epitaxial growth of a new phase on the template. Nevertheless, the acceptable level of defects is different for both processes. For photocatalysis, some defects are advantageous when they introduce intermediate surface states that narrow the band gap. In general, a high density of defects, which act as recombination centers for photo-induced electrons and holes, lowered the PA. The photodegradation of RhB in BIT2 was 88% and in BIT8 it was 65% after 4 hours of irradiation with UV-A light. This confirmed our expectations that the PA of BIT2 is larger than that of BIT8. The larger PA of BIT2 can also be ascribed to smaller particles with a higher specific surface area, which provide more active sites for the photocatalyt-ic reaction. For the DSC measurements both samples (BIT2 and BIT8) were first washed with 2-M HNO3 to ensure single-phase Bi4Ti3O12. The DSC measurements revealed a very similar DSC peak temperature (Tc) for the ferroelec-tric-to-paraelectric phase transition for both types of Bi-4Ti3O12 plates (643.6 °C for BIT2 and 642.9 °C for BIT8) (Fig. 3). However, the absolute values of the phase-transi- tion enthalpies (|AHFET|) of the two samples differ significantly. The measured|AHFEF| for BIT2 and BIT8 were 1.457 J/g and 4.555 J/g, respectively (Fig. 3, Table 2). The decrease of the phase-transition enthalpy with a decrease in the particle size was already observed for other ferroelectric particles (BaTiO3).24 Taking into account that the enthalpy of the phase transition is proportional to the polarization (P) (AHFET = 2 nP2 Tc /C (Eq. 1), where C is the Curie-Weiss constant) the decrease of |AHFEF| could be explained by the decrease of P. It is also known that the phase-transition enthalpy is related to the domain struc-ture.24,25 The larger |AHFET| of BIT8 could correlate with larger particles and be expected to exhibit a multi-domain structure. Due to the domain clamping the enthalpy of the phase transition was higher for larger particles compared to the single-domain smaller particles. In addition, with a decrease of the particle size, the ratio between the disordered surface and the ordered bulk is increasing, which additionally causes a destruction of the polar state.24 3. 2. Bi4Ti3O12 Plates as a Template for Topochemical Conversion to SrTiO3 Due to the pseudo-perovskite units the Bi4Ti3O12 plates are regarded as a suitable template for the preparation of MTiO3 perovskite plates via topochemical conversion. It was already shown that the transformation from Bi4Ti3O12 to BaTiO3 is possible in the molten salt.9 There is great interest in whether this kind of conversion is also -0,04 -0,08 - O "O.™ « BIT8 Temperature (°C) Figure 3: DSC curves of Bi4Ti3O12plates during heating for BIT2 and BIT8 sample. Table 2: Results for BET, DSC and PA measurements of BIT2 and BIT8 Sample BET (m2/g) Tc (°C) (J/g) PA4 hours BIT2 2.9075 643.6 1.457 88.6% BIT8 0.4855 642.9 4.555 65.0% Contala et al.: Plate-Like Bi4Ti3O12 Particles Acta Chim. Slov. 2018, 65, 630-637 635 possible under hydrothermal conditions. Kalyani et al. studied the hydrothermal crystallization of SrTiO3 on anatase nanowires and proved that the formation of Sr-TiO3 was driven by the topochemical reaction,26 since the formed SrTiO3 mesocrystals retain the wire-like shape of the initial anatase particles. To the best of our knowledge there are no literature reports about the topochemical conversion of Bi4Ti3O12 to MTiO3 perovskites under hydrothermal conditions. In the present study we firstly aimed to examine whether under these conditions Bi-4Ti3O12 plates are an appropriate template for the preparation of SrTiO3 plates and secondly we wanted to verify how differently sized Bi4Ti3O12 template plates (sub-^m-and above-^m-sized) influenced the final morphology of the SrTiO3 particles. From the standpoint of advanced development and the improvement of applications and electronic devices, controlling the shape and size of particles can lead to new or combined and improved properties of existing materials. The miniaturization of electronic devices is also of great interest in nanotechnology; however, it is already well known that materials lose some of their functional properties (i.e., ferroelectricity) below certain small dimensions. Additionally, it is easier to observe the growth of SrTiO3 on larger plates. For this reason we decided to study the formation of SrTiO3 from two types of Bi4Ti3O12: smaller, sub-^m-sized, Bi4Ti3O12, plate-like particles with a broad particle size distribution (BIT2) and larger, 1-2-^m-large and well-defined Bi4Ti3O12 plates (BIT8). The XRD examination of the reaction products in both cases confirmed the formation of SrTiO3 (PDF #0350734) from the Bi4Ti3O12 plates under alkaline (4-M NaOH) hydrothermal conditions at 200 °C for 12 hours (Fig. 4, XRD pattern A). The smaller Bi4Ti3O12 template particles (BIT2) that were very non-uniform in size, consequently resulted in a non-ideal morphology of the Sr-TiO3 particles. This means that the as-prepared SrTiO3 particles differed in their shape and size. The cube-like particles prevailed, but some of them also exhibited a preferable plate-like shape (Fig. 5A). In contrast, the SrTiO3 particles prepared from larger Bi4Ti3O12 template plates (BIT8) preserved the shape of the template. The XRD revealed that (00/)-oriented Bi4Ti3O12 plates transformed into (h00)-oriented SrTiO3 plates (Fig. 4, XRD pattern B). However, the morphology of the SrTiO3 plates was still not perfect. The particles varied in their size, which was most probably the consequence of the non-uniform size distribution of the initial Bi4Ti3O12 plates. Figure 5B demonstrates that there were small holes present in some of those SrTiO3 plates. These holes were also observed when the SrTiO3 plates were washed only with water after the synthesis. Therefore, the holes were not formed during HNO3-washing, although HNO3 is a strong acid and could cause the leaching of Sr and the remains of Bi from the plates. Additionally, the holes were not formed due to the HNO3 washing of the Bi4Ti3O12 template plates, because the holes were also present in the SrTiO3 plates, which were prepared from the water-washed Bi4Ti3O12 plates. Thus, the holes were not a consequence of the etching effect of the HNO3, although it is known that chemical etching can cause square-shaped holes in Bi4Ti3O12 grains.27 These holes were either the consequence of the defective surface of the Bi4Ti3O12 plates or originated from the lattice mismatch between the pseudo-perovskite layer of the template and the SrTiO3 plates. In addition, the removal of the (Bi2O2)2+ layer during the conversion could also cause exfoliation, the result of which could be those holes. An examination of the morphology of the SrTiO3 particles prepared from the different Bi4Ti3O12 template plates revealed that the completeness of the SrTiO3 plates strongly depended on both the quality of the Bi4Ti3O12 template plates and the reaction conditions. For example, the worst preservation of the initial Bi4Ti3O12 template shape during the conversion to SrTiO3 was observed for the Bi4Ti3O12 plates that were prepared from the large, ^m-sized, TiO2 anatase particles. With this study we confirmed that the ^m-sized Bi4Ti3O12 plate-like particles with a rather uniform particle size distribution (i.e. BIT8) can be used as a template for the topochemical transformation to plate-like SrTiO3 particles with a (h00) preferential orientation. 3 a a