Kinetic Study of the Static Hydrothermal Synthesis of BaTiO3 Using

Apr 5, 2011 - ... Engineering, University of Aveiro, CICECO, 3810-193 Aveiro, Portugal ... Materials, University of Sheffield, Sheffield S1 3JD, Unite...
0 downloads 0 Views 3MB Size
ARTICLE pubs.acs.org/crystal

Kinetic Study of the Static Hydrothermal Synthesis of BaTiO3 Using Titanate Nanotubes Precursors Florentina Maxim,† Paula M. Vilarinho,*,† Paula Ferreira,† Ian M. Reaney,‡ and Igor Levin# †

Department of Ceramic and Glass Engineering, University of Aveiro, CICECO, 3810-193 Aveiro, Portugal Department of Engineering Materials, University of Sheffield, Sheffield S1 3JD, United Kingdom # Ceramics Division, NIST, Gaithersburg, Maryland 20899, United States ‡

ABSTRACT: A kinetic study of the growth of BaTiO3 under hydrothermal conditions using layered titanate nanotubes as precursors has been performed at 110 and 200 C. At the early stages of crystallization, irrespective of temperature, pseudocubic BaTiO3 nucleates on the surfaces of the nanotubes. Growth subsequently proceeds by a phase-boundary mechanism with exponent m ≈ 1. During this first regime, “wild type” dendritic particles form with a rough surface and a microstructure which contains planar defects. As the reaction proceeds beyond 70% completion, the nanotubes fully dissolve, and nucleation and growth of tetragonal rather than pseudocubic BaTiO3 take place. During this second regime, the morphology of the BaTiO3 particles remains predominantly dendritic with “seaweed” morphology and a smooth surface, but smaller spherical precipitates and particles with a small number of dendritic arms are also observed. For this stage in the hydrothermal process, the exponent, m ≈ 0.2, does not correspond to any known kinetic mechanisms.

’ INTRODUCTION BaTiO3 (BT) is cubic and paraelectric (PE) above ≈125 C but on cooling transforms to a tetragonal ferroelectric (FE) phase. The phase transition occurs due to the cooperative off-center displacement of Ti ions within O octahedra along the [001] direction. On further cooling, two more ferroelectric phases are observed: orthorhombic between 5 and 90 C and rhombohedral below 90 C. Because of its high relative permittivity (εr) and low dielectric loss, BT is utilized in many electronic and microelectronic applications.1 Its primary use, in modified form, is as the dielectric in multilayer capacitors (MLCs), but it has also found commercial exploitation in positive temperature coefficient resistors, ultrasonic transducers, piezoelectric devices,2 and tunable phase array antennas.3 In MLCs, thinner dielectric layers give rise to higher capacitance per unit volume and therefore enable miniaturization of the devices. MLCs are fabricated by a highly controlled tape casting process, and the dielectric layer thickness in part depends on the starting particle size of the BT powder. Therefore, there has been considerable recent interest in the development of techniques for the fabrication of sub-micrometer BT powders.4 The degree of tetragonality (c/a ratio) and phase transition temperature (TC) of BT decrease with particle size, but the magnitude of this effect varies with the method of powder fabrication.5,6 Consequently, the electrical properties of compacted bodies also degrade.5 An existence of a critical size below which ferroelectricity can no longer exist in BT has been postulated, but experimentally this value varies with the method of fabrication of the powder from 120 to 8 nm.5,7,8 Despite the lack of consensus in the literature, r 2011 American Chemical Society

these observations point to the role of extrinsic effects such as point defects and local strain gradients in degrading ferroelectric properties.9 In 2007, a theoretical study predicted the enhancement of ferroelectricity in one-dimensional (1D) nanorods and nanowires, claiming that nanowires only ≈1.2 nm in radius could preserve long-range cooperative Ti displacements along their length.10 Such effects may have potential applications in high density nonvolatile memory storage.11 However, the fabrication of ferroelectric nanostructures is still in its infancy, and numerous challenges need to be overcome. The first step is to achieve a cost-effective fabrication of large numbers of highly oriented BT nanorods with well-defined dimensions. Recently, efforts have been devoted to the synthesis of 1D ferroelectric nanoparticles,1214 and the properties of some individual 1D ferroelectrics have been assessed.1517 Within this context, and aiming to develop simple, low-cost environmentally friendly synthesis, the present authors reported hydrothermal synthesis of anisotropic barium titanate particles by reaction of titanate layered nanotubes (NTs) and barium hydroxide in aqueous solutions.18 Although it has been claimed that anatase nanotubes can behave as growth directing agents to obtain BT NTs,13,19 the present authors18 and other recent studies2022 clearly showed that the titanate layered structures in general and nanotubes, in particular, do not easily act as templates for the 1D Received: November 6, 2010 Revised: March 21, 2011 Published: April 05, 2011 3358

dx.doi.org/10.1021/cg101466u | Cryst. Growth Des. 2011, 11, 3358–3365

Crystal Growth & Design

ARTICLE

growth of BT. Instead, anisotropic growth of barium titanate particles was observed when the synthesis temperature was 90 C, resulting in the formation of pseudocubic BT particles with a “wild”-type dendritic shape. At 200 C, barium titanate “seaweed”-type dendritic particles with a predominantly tetragonal structure were obtained. To better understand the role of the titanate nanotubes in the crystallization of BT under hydrothermal conditions, a detailed kinetic study is required. The kinetics of hydrothermal (and aqueous) synthesis of roundshaped barium titanate have been thoroughly studied and two concurrent mechanisms have been proposed.2329 Hertl suggested an in situ transformation mechanism that assumes that the dissolved Ba2þ and Ba(OH)þ ions react topochemically with the titanium precursor and BT is formed by heterogeneous nucleation on its surface.23 The second mechanism, dissolutionprecipitation,2427 assumes that Ba and Ti sources are rapidly dissolved and homogeneous nucleation of BT takes place. However, none of these studies reported the kinetics of the crystallization of anisotropic BT particles. Moreover, the kinetics of BT crystallization starting from titanate NTs has never been described. In this work, a detailed kinetic investigation (at 110 and 200 C) of the morphogenesis of anisotropic BT under hydrothermal conditions using layered titanate nanotubes as a precursor source is presented and a kinetic model is proposed.

’ EXPERIMENTAL SECTION Ba(OH)2 3 8H2O (Merck, 98% purity; the use of brand or trade name does not imply endorsement of the product by NIST) and a layered titanate precursor with the general formula Na2xHxTi2O5 3 1.8H2O were used as barium and titanium sources, respectively. The titanate NTs were obtained starting from anatase by hydrothermal treatment in alkaline medium. All the barium titanate syntheses were performed in Teflon-lined autoclaves. The detailed preparation of the barium titanate samples is reported elsewhere.18 Briefly, the barium titanate preparation was initiated by mixing barium and titanium precursors in a 1.01 Ba/Ti molar ratio. The pH of the starting mixture was above 13. After being stirred for 30 min, the mixtures were transferred into Teflon-lined autoclaves, and the hydrothermal syntheses were performed at 110 and 200 C for several periods of time. The samples are identified with BT and two numbers, indicating the temperature and time of synthesis, respectively. The formed phases were identified by powder X-ray diffraction (XRD) using a Philips X’Pert diffractometer and CuKR radiation. The diffraction patterns were acquired from 20 to 50 (2θ) with a step length of 0.04 and a fixed counting time of 20 s/step. In order to obtain reliable quantitative data for the kinetic analysis, the scan time and the samples weight were always the same.26 Raman spectroscopy was utilized as a complementary tool for the phase identification. Raman spectra were recorded on a Bruker RFS 100/S FT Raman spectrometer using a 1064 nm excitation of the Nd/YAG laser. Variable-temperature X-ray diffraction measurements was performed using Panalytical X’pertPro powder diffractometer equipped with an incident beam monochromator (Cu KR1 radiation), pixel position sensitive detector, and Anton Paar TTK-450 temperature stage having an operating range from 190 C to þ450 C. The measurements were conducted in a vacuum. The sample powder was mixed with a small amount of vacuum grease to improve thermal conductivity. The microstructures were analyzed by scanning electron microscopy (SEM), with a Hitachi SU-70 (S-4100) microscope, and by transmission electron microscopy (HRTEM) with a 300 kV Hitachi H9000-NA instrument.

Figure 1. X-ray diffraction patterns of the samples prepared at (a) 110 C and (b) 200 C for various periods of time; as a comparison, the X-ray diffraction pattern of standard BaTiO3 obtained by solid-state reaction at 1100 C is also presented. The kinetic analysis was carried out based on the method described by Hancock and Sharp30 and using the JohnsonMehlAvrami equation.31,32 The fraction of crystallized barium titanate as a function of reaction time was extracted from the XRD patterns presented in Figure 1. The degree of crystallization of a material can be represented by the gross intensity (Igross) of a diffraction peak determined by the area of the peak and defined as26 Igross ¼ ðintegrated phase counts=scan timeÞ  ðΔ2θÞ

ð1Þ

The gross intensity of the (110) reflection in the 2θ range from 30.2 to 32.5 for each sample at the two reaction temperatures 110 and 200 C was determined. Crystallization fractions (f) were calculated taking into account that at time 0 (pure titanium precursor) the BT crystallization fraction is 0. The maximum extent of the reaction (f = 100%) is obtained from a BT sample fabricated by solid-state reaction at 1100 C and is denominated as BT standard. The BT crystallization fraction at time t was calculated by the ratio between the gross intensity (Igross) at the specific time of each sample and the gross intensity (Igross) of the standard sample. 3359

dx.doi.org/10.1021/cg101466u |Cryst. Growth Des. 2011, 11, 3358–3365

Crystal Growth & Design

ARTICLE

Figure 3. Comparison of traces of the 103 reflection (2θ ≈ 75) measured at 20 C (red) and 180 C (blue) for the sample prepared at 110 C for 18 h (BT11018).

Figure 4. Temperature dependence of the fwhm for the 103 peak for the sample prepared at 110 C for 18 h (BT11018). Figure 2. Raman spectra of the samples prepared at (a) 110 C and (b) 200 C for various periods of time.

’ RESULTS Structural Evolution. Figure 1 presents the X-ray diffraction patterns from samples prepared at 110 C (Figure 1a) and 200 C (Figure 1b). Crystalline BT is formed after ≈6 h and after 30 min at 110 and 200 C, respectively. For comparison, the XRD pattern of BT ((200)/(002) peaks) fabricated by solid-state reaction at 1100 C is also presented. For the hydrothermal process, a gradual transformation from a cubic to a predominantly tetragonal symmetry is observed at both reaction temperatures as synthesis time increases. At short reaction times, BaCO3 is present, for example, at 200 C for 30 min and 1 h (Figure 1b), but the volume fraction decreases as the reaction time increases with single phase BT present after 2.5 h. The Raman spectra of the samples prepared at 110 and 200 C are shown in Figure 2, panels a and b, respectively. Eight Ramanactive modes are expected for tetragonal BaTiO3, 3A1gþ B1gþ 4Eg.33 The peaks at around 517, 250, and 185 cm1 are assigned to the fundamental TO modes (transverse component of the optical mode) of A1g symmetry, and the peak at 307 cm1 is assigned to the B1g mode, indicating asymmetry within the [TiO6] octahedra. The broad band at around 715 cm1 is related to the highest frequency longitudinal optical mode (LO) of A1

symmetry. If the peak at 307 cm1 broadens and becomes indistinct, then the tetragonal phase is not considered dominant and the BT is pseudocubic.34 In Raman spectra from samples collected after 18 h at 110 C (Figure 2a) and 30 min at 200 C (Figure 2b), the band at 307 cm1 is broad and indistinct, but with increasing time it sharpens, suggesting a transition from a pseudocubic to tetragonal structure. The transition occurs more rapidly at 200 C than at 110 C. The transition from pseudocubic to tetragonal structure was studied using variable-temperature XRD for the samples prepared at 110 C for 18 h and 24 h, respectively. Figure 3 shows traces of the 103 reflection (2θ ≈ 75) recorded at 20 C (red) and 180 C (blue), respectively, for the 18 h sample. The peak is relatively narrow but asymmetric and exhibits a broadened base likely due to heterogeneities within the sample. The line broadening increases significantly with decreasing temperature suggesting a lattice distortion. Figure 4 displays the temperature dependence of the FWHM for the 103 peak of the same sample (BT11018). A clear onset of the strong line broadening is observed below 80 C; the broadening increases progressively with decreasing temperature. This trend is attributed to a distortive phase transition, most likely from cubic to tetragonal. The error bars reflect statistical uncertainties associated with profile fitting. 3360

dx.doi.org/10.1021/cg101466u |Cryst. Growth Des. 2011, 11, 3358–3365

Crystal Growth & Design

ARTICLE

Figure 5. Trace of the 103 reflection recorded at 25 C (red), 50 C (blue), and 120 C (green) for the sample prepared at 110 C for 24 h (BT11024). Figure 7. JohnsonMehl Avrami plots for the two different reaction temperatures (a) 110 C and (b) 200 C.

and 200 C, respectively. It is evident that f increases with the reaction time but that the reaction is faster at 200 C than at 110 C. After 24 h at 110 and 200 C, f = 0.75 and 0.85, respectively. The maximum extent of the reaction (100% of BT) is reached after ≈24 and 96 h at 200 and 110 C, respectively. The time dependence of the crystallization fraction has a sigmoidal shape.31,32 From the Avrami-Erofe’ev equation, Hancock and Sharp derived eq 2 from which the kinetic parameters of the crystallization in a solid-state reaction can be extracted.30 ln½  lnð1  f Þ ¼ lnðkÞ þ m lnðtÞ

Figure 6. Fraction of BaTiO3 crystalline phase represented as a function of the synthesis time for the samples prepared at (a) 110 C and (b) 200 C.

The results of a similar analysis performed on the sample synthesized at 110 C for 24 h are presented in Figure 5; here, the 103 reflection was recorded at 25 C (red), 50 C (blue), and 120 C (green). The peak exhibits a clear splitting as expected for a tetragonal phase. The splitting decreases progressively upon heating consistent with occurrence of the tetragonal to cubic transition at higher temperatures. Kinetic Analysis. The crystalline fraction of BaTiO3 (f) as a function of the synthesis time, t, is shown in Figure 6a,b at 110

ð2Þ

where f is the fraction isothermally crystallized at the time t, k is the rate constant which partially depends on the nucleation frequency and growth rate and is very sensitive to the temperature, and m is an exponent which is independent of the temperature but responsive to the time of nucleation and growth rate and to the geometry of the particles. Hancock and Sharp30 showed that for reactions obeying a single theoretical rate equation JohnsonMehl Avrami plots, ln[ln(1  f)] versus ln(t), over the range f = 0.150.50 yield approximately straight lines, with slope (m) characteristic of three distinct reaction mechanisms;30 for m = 0.540.62, diffusion is the rate limiting step, for m = 1.01.24, a zero-order, first-order, or a phase-boundary controlled mechanism is indicated, and, when m = 2.03.0, the nucleation controls the rate. In the current study, the JohnsonMehl Avrami plots for two different reaction temperatures are depicted in Figure 7a,b. For 110 and 200 C synthesis, the kinetic analysis indicated that crystallization of BaTiO3 takes place in two stages, in a manner similar to that previously reported.24,26,27 In the first regime (f < 0.70) at both temperatures, m ≈ 1, which corresponds to a phase boundary mechanism.30 This value was also obtained for the in situ study reported by Walton et al. in which BaTiO3 was synthesized in a static hydrothermal cell starting from barium hydroxide and amorphous titanium oxide.25 For the synthesis of BaTiO3 via aqueous coprecipitation or in nonstatic hydrothermal cells,24,26,27 values as high as m = 2.63 have been reported and attributed to nucleation-dominated mechanisms. These data imply a fundamentally different rate-determining step in static hydrothermal synthesis of BaTiO3 compared to other methods. 3361

dx.doi.org/10.1021/cg101466u |Cryst. Growth Des. 2011, 11, 3358–3365

Crystal Growth & Design

ARTICLE

Figure 8. Timeline morphological evolutions studied by SEM for the samples obtained at 110 C after a) 6h (BT1106); b) 12h (BT11012); c) 18h (BT11018); d) 24h (BT11024); e) 96h (BT11096) and f) 144 h (BT110144). Insets of figures 8c) and 8f) present magnified views of the BT particles.

For the second regime, a value of m ≈ 0.2 is obtained for both temperatures. According to the table of Hancock and Sharp, this value does not fall into any of the three listed exponent ranges.30 The method of kinetic analysis proposed by Hancock and Sharp is only valid for f e 0.5, and no attempt has ever been made to rationalize kinetic data for f > 0.5. Therefore, a detailed morphological study was performed to understand further the crystallization process. Morphological Evolution. SEM images of the timeline of morphological evolution at 110 C are presented in Figure 8. At the early stages of the reaction (6 h, Figure 8a), dendrites with a rough surface are detected along with “wild-type” dendrites. These morphologies are similar to those reported for the BT synthesized at 90 C in the previous work of the present authors.18,35 As t increases more wild type dendrites are formed (Figure 8b for 12 h), but unreacted titanium precursor is still observed up to 18 h of reaction (Figure 8c). The unreacted titanium precursor cannot be detected by XRD due to its low crystallinity (see Figure 1a). Concomitantly, for this synthesis time (18 h) spherical barium titanate particles are formed (see inset Figure 8c). After 24 h of reaction, the presence of spherical particles is particularly noticeable (Figure 8d). After 96 h of reaction, when the maximum extent of crystallization is attained (≈75%) (Figure 6a), two types of barium titanate dendritic morphologies are observed (Figure 8e). “Wild-type” dendrites coexist with small and large seaweed-type dendritic aggregates characterized by a smooth surface.

The percentage of the round-shaped particles appears to be reduced compared to the 24 h sample . A coexistence of the “wild” and seaweed-type dendritic particles is also confirmed after 144 h of reaction, Figure 8f. A similar morphological evolution, from “wild-type” dendrites to spherical particles to seaweed dendrites, is also observed at 200 C (Figure 9). At this temperature, the wild-type dendrites are formed after only 30 min, and, at the same time, large quantities of unreacted titanium precursor are observed (Figure 9a). Spherical particles and small “seaweed” dendrites are formed after 1 h (Figure 9b). After 24 h at 200 C (f ≈ 85%), only large seaweed-type dendritic aggregates are apparent (Figure 9c). Figure 10a,b represents a sequence of transmission electron micrographs of BT particles synthesized at 110 and 200 C, respectively. Regardless of temperature, BaTiO3 nuclei form on the surface of the NTs during early stages of synthesis. The ensuing particle morphologies mirror those observed by SEM. However, diffraction patterns from dendrites oriented with the Æ001æ direction parallel with the electron beam confirmed that the dendritic arms occupied {001} growth habits.

’ DISCUSSION The data presented above lead to the four key points that need to be considered. First, at the early stages of crystallization, 3362

dx.doi.org/10.1021/cg101466u |Cryst. Growth Des. 2011, 11, 3358–3365

Crystal Growth & Design

ARTICLE

Figure 10. TEM micrographs at the early stages of BaTiO3 crystallization for the samples prepared at (a) 110 C and (b) 200 C.

phase boundary rather than a classic diffusion or nucleationcontrolled mechanism. The movement of the solidliquid interface within the reacting medium is physically unimpeded, so it is reasonable to assume that the slowest step occurs within the chemical reaction to form BaTiO3 at the phase boundary. The overall reaction for the formation of BaTiO3 is given below (eq 3): Ba2þ þ Ti4þ þ 6OH f BaTiO3 þ 3H2 O

Figure 9. Timeline morphological evolutions studied by SEM for the samples obtained at 200 C after (a) 30 min (BT20030); (b) 1 h (BT2001); (c) 24 h (BT20024).

BaTiO3 nucleates on the surface of the Ti-nanotubes. Second, there is a gradual transition with reaction time from pseudocubic to tetragonal barium titanate (Figures 1 and 2 and also Figures 35). Third, the transition from pseudocubic to tetragonal is accompanied by a change in the particle morphology from “wild”-type dendrites to spherical particles and seaweed dendrites (Figures 8 and 9). Finally, an exponent, m = 1, is obtained by fitting JohnsonMehl Avrami plots (f < 0.7) to the reaction, which suggests that the kinetics are dominated by a

ð3Þ

but as with all complex reactions there are multiple steps within the mechanism, one of which will be rate determining. On the basis of the thermodynamic study of Lencka and Riman,36 the aqueous chemistry of Ba2þ, Ti4þ cations and O2 is represented by BaðOHÞþ ¼ Ba2þ þ OH

ð4Þ

TiðH2 OÞ2 ðOHÞ4 þ 4Hþ ¼ Ti4þ þ 4H2 O

ð5Þ

2OH ¼ O2 þ H2 O

ð6Þ

and indicates that formation of stoichiometric BaTiO3 requires hydrated Ti(OH)4 (Ti(H2O)2(OH)4) species. As illustrated below, dissolution of Ti is accompanied by a sequence of reactions to form hydrous titanium cations (Ti(OH)x(4x)þ). 3363

dx.doi.org/10.1021/cg101466u |Cryst. Growth Des. 2011, 11, 3358–3365

Crystal Growth & Design

ARTICLE

Since each step generates H3Oþ ions, the isothermal stability of the Ti(OH)x(4x)þ species is related to pH which decreases with increasing titanium concentration.36 TiðH2 OÞ6 4þ þ H2 O S TiðH2 OÞ5 OH3þ þ H3 Oþ

ð7Þ

TiðH2 OÞ5 OH3þ þ H2 O S TiðH2 OÞ4 ðOHÞ2 2þ þ H3 Oþ ð8Þ TiðH2 OÞ4 ðOHÞ2 2þ þ H2 O S TiðH2 OÞ3 ðOHÞ3 þ þ H3 Oþ ð9Þ TiðH2 OÞ3 ðOHÞ3 þ þ H2 O S TiðH2 OÞ2 ðOHÞ4 þ H3 Oþ ð10Þ Therefore, the protons released during Ti dissolution and hydrolysis result in a decrease in pH as the titanate NTs dissolve. It is postulated therefore that, under the current static synthesis conditions, the local concentration of Ti increases, and therefore local pH decreases at the surface of the NTs. The decrease in local pH shifts the equilibrium of reactions 710 to the left, thereby suppressing the formation of Ti(H2O)2(OH)4 which according to reaction 5 is critical for the formation of BaTiO3. The proposed rate-determining step (RDS) therefore relates not to nucleation or diffusion but to hydrolysis of the Ti precursor at the surface of the titanate NTs. This RDS corresponds directly to the phase boundary mechanism indicated by the exponent, m = 1, obtained from fitting the kinetic data, Figure 6a,b. Figure 11 schematically illustrates the proposed RDS. Walton et al.25 also reported a phase boundary mechanism (m = 1) for the crystallization of BaTiO3. The main similarity between the results presented here and the study by Walton et al.25 is that each was performed under static conditions. Therefore, a dissolution of the Ti precursor would locally decrease pH in a manner similar to that postulated, and reaction rate will be controlled by the formation of Ti(H2O)2(OH)4 at the solidliquid interface. Under nonstatic conditions, percolation of the aqueous solution eradicates the local decrease of pH at the precursor surface and much higher exponents are observed (m > 2.5) corresponding to nucleation as the RDS. However, only in the present study the dendritic particles are observed at early stages of the reaction.27,37 The apparent difference between the experimental conditions described by Walton et al.23 and in the present study is in the titanium precursor, which may be an indication that titanate NTs play a crucial role in the formation of anisotropic BaTiO3 at the early stages of the reaction. This effect may simply relate to the 1D geometry of the NTs and to multiple nucleation events on their surface. As the reaction progresses, the nanotubes fully dissolve (f > 0.7) and the aqueous solution becomes enriched in hydrated Ti ions. There are no longer any local conditions inhibiting the formation of Ti(H2O)2(OH)4 and the homogeneous precipitation of BaTiO3 is observed. In the early stages, BT nucleated on the surface of the NTs is pseudocubic, implying that defects are incorporated within the perovskite lattice. In contrast, BT nucleated in the latter stages of the reaction is tetragonal, implying a lower concentration of impurity ions or defects. Moreover, the BT formed on the surface NTs has a wild-type dendritic morphology with a rough surface with evidence of nanoscale planar defects in high resolution TEM images. The particles of homogeneously nucleated BT observed

Figure 11. Schematic representation of the proposed mechanism for the synthesis of BaTiO3 under static hydrothermal conditions starting from layered titanate NTs precursors.

in the latter stages of the reaction have a smooth surface and no evidence of planar defects within its microstructure. The change in structure, microstructure, and morphology of the BT is entirely consistent with a change in the kinetics and reaction mechanism at f = 0.7.

’ CONCLUSIONS The crystallization kinetics of anisotropic BaTiO3 particles prepared by a static hydrothermal synthesis starting from layered titanate nanotubes of general formula Na2xHxTi2O5 3 1.8H2O was studied at 110 and 200 C. At the early stages of crystallization, the reaction taking place at the solidliquid interface was considered to be the rate-determining step. The reaction is controlled by concentration of Ti(H2O)2OH4 which is limited in the vicinity of the NTs by a local decrease of pH. As the reaction proceeds, the NTs fully dissolve (f > 0.7), the bulk of the solution becomes enriched in hydrated Ti ions, and homogeneous nucleation of BaTiO3 occurs. The BaTiO3 formed when the kinetics are controlled by reaction at the phase boundary is pseudocubic with a rough surface morphology and a microstructure containing planar defects. In contrast, in the later stages of the reaction, the BaTiO3 is tetragonal with a smooth surface and a microstructure free from planar defects. Dendrites are formed throughout the reaction, but the morphology differs depending on f. In the early stages (f < 0.7), dendrites are “wild-type” but in the later stages “seaweed-type” are observed. The changes in structure and microstructure are consistent with a change in the kinetics at f = 0.7. ’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected].

’ ACKNOWLEDGMENT The authors acknowledge the Portuguese Foundation for Science and Technology (FCT), FEDER, the European Network of Excellence FAME, under the Contract FP6-500159-1 and COST 539, Electroceramics from Nanopowders Produced by Innovative Methods (ELENA). Florentina Maxim is thankful to FCT for the fellowship SFRH/BD/23375/2005. ’ REFERENCES (1) Haertling, G. H. Ferroelectric ceramics: History and technology. J. Am. Ceram. Soc. 1999, 82 (4), 797–818. 3364

dx.doi.org/10.1021/cg101466u |Cryst. Growth Des. 2011, 11, 3358–3365

Crystal Growth & Design (2) Vijatovic, M. M.; Bobic, J. D.; Stojanovic, B. D. History and Challenges of Barium Titanate: Part II. Sci. Sintering 2008, 40 (3), 235–244. (3) Tagantsev, A. K.; Sherman, V. O.; Astafiev, K. F.; Venkatesh, J.; Setter, N. Ferroelectric materials for microwave tunable applications. J. Electroceram. 2005, 14 (3), 199–203. (4) Pithan, C.; Hennings, D.; Waser, R. Progress in the synthesis of nanocrystalline BaTiO3 powders for MLCC. Int. J. Appl. Ceram. Technol. 2005, 2 (1), 1–14. (5) Zhao, Z.; Buscaglia, V.; Viviani, M.; Buscaglia, M. T.; Mitoseriu, L.; Testino, A.; Nygren, M.; Johnsson, M.; Nanni, P. Grain-size effects on the ferroelectric behavior of dense nanocrystalline BaTiO3 ceramics. Phys. Rev. B 2004, 70 (2) No. 024107. (6) Shiratori, Y.; Pithan, C.; Dornseiffer, J.; Waser, R. Raman scattering studies on nanocrystalline BaTiO3 - Part I - isolated particles and aggregates. J. Raman Spectrosc. 2007, 38 (10), 1288–1299. (7) Uchino, K.; Sadanaga, E.; Hirose, T. Dependence of the crystal structure on particle size in barium titanate. J. Am. Ceram. Soc. 1989, 72 (8), 1555–1558. (8) Wang, X. H.; Deng, X. Y.; Wen, H.; Li, L. T. Phase transition and high dielectric constant of bulk dense nanograin barium titanate ceramics. Appl. Phys. Lett. 2006, 89 (16). (9) Frey, M. H.; Payne, D. A. Grain-size effect on structure and phase transformations for barium titanate. Phys. Rev. B 1996, 54 (5), 3158–3168. (10) Morozovska, A. N.; Eliseev, E. A.; Glinchuk, M. D. Size effects and depolarization field influence on the phase diagrams of cylindrical ferroelectric nanoparticles. Phys. B-Condensed Matter 2007, 387 (12), 358–366. (11) Scott, J. F. Nanoferroelectrics: statics and dynamics. J. Phys.: Condens. Matter 2006, 18 (17), R361–R386. (12) Joshi, U. A.; Yoon, S.; Balk, S.; Lee, J. S. Surfactant-free hydrothermal synthesis of highly tetragonal barium titanate nanowires: A structural investigation. J. Phys. Chem. B 2006, 110 (25), 12249–12256. (13) Mao, Y.; Banerjee, S.; Wong, S. S. Hydrothermal synthesis of perovskite nanotubes. Chem. Commun. 2003, 408–409. (14) Wang, X. H.; Chen, L. L.; Zhao, J. L.; Jin, L.; Li, L. T. Hydrothermal synthesis of thin films of barium titanate nanotube arrays. Integr. Ferroelectr. 2008, 99, 125–131. (15) Wang, J.; Sandu, C. S.; Colla, E.; Wang, Y.; Ma, W.; Gysel, R.; Trodahl, H. J.; Setter, N.; Kuball, M. Ferroelectric domains and piezoelectricity in monocrystalline Pb(Zr,Ti)O-3 nanowires. Appl. Phys. Lett. 2007, 90, 13. (16) Spanier, J. E.; Kolpak, A. M.; Urban, J. J.; Grinberg, I.; Ouyang, L.; Yun, W. S.; Rappe, A. M.; Park, H. Ferroelectric phase transition in individual single-crystalline BaTiO3 nanowires. Nano Lett. 2006, 6 (4), 735–739. (17) Yun, W. S.; Urban, J. J.; Gu, Q.; Park, H. Ferroelectric properties of individual barium titanate nanowires investigated by scanned probe microscopy. Nano Lett. 2002, 2 (5), 447–450. (18) Maxim, F.; Ferreira, P.; Vilarinho, P. M.; Reaney, I. Hydrothermal synthesis and crystal growth studies of BaTiO3 using Ti nanotube precursors. Cryst. Growth Des. 2008, 8 (9), 3309–3315. (19) Mao, Y.; Park, T.-J.; Zhang, F.; Zhou, H.; Wong, S. S. Environmentally friendly methodologies of nanostructure synthesis. Small 2007, 3 (7), 1122–1139. (20) Bao, N. Z.; Shen, L. M.; Srinivasan, G.; Yanagisawa, K.; Gupta, A. Shape-controlled monocrystalline ferroelectric barium titanate nanostructures: From nanotubes and nanowires to ordered nanostructures. J. Phys. Chem. C 2008, 112 (23), 8634–8642. (21) Kang, S. O.; Park, B. H.; Kim, Y. I. Growth mechanism of shapecontrolled barium titanate nanostructures through soft chemical reaction. Cryst. Growth Des. 2008, 8 (9), 3180–3186. (22) Li, Y.; Gao, X. P.; Li, G. R.; Pan, G. L.; Yan, T. Y.; Zhu, H. Y. Titanate Nanofiber Reactivity: Fabrication of MTiO3 (M = Ca, Sr, and Ba) Perovskite Oxides. J. Phys. Chem. C 2009, 113 (11), 4386–4394. (23) Hertl, W. Kinetics of barium titanate synthesis. J. Am. Ceram. Soc. 1988, 71 (10), 879–883.

ARTICLE

(24) Eckert, J. O., Jr.; Hung-Houston, C. C.; Gersten, B. L.; Lencka, M. M.; Riman, R. E. Kinetics and mechanisms of hydrothermal synthesis of barium titanate. J. Am. Ceram. Soc. 1996, 79 (11), 2929–2939. (25) Walton, R. I.; Millange, F.; Smith, R. I.; Hansen, T. C.; O’Hare, D. Real time observation of the hydrothermal crystallization of barium titanate using in situ neutron powder diffraction. J. Am. Chem. Soc. 2001, 123, 12547–12555. (26) Moon, J.; Suvaci, E.; Morrone, A.; Costantino, S. A.; Adair, J. H. Formation mechanisms and morphological changes during the hydrothermal synthesis of BaTiO3 particles from a chemically modified, amorphous titanium (hydrous) oxide precursor. J. Eur. Ceram. Soc. 2003, 23 (12), 2153–2161. (27) Viviani, M.; Buscaglia, M. T.; Testino, A.; Buscaglia, V.; Bowen, P.; Nanni, P. The influence of concentration on the formation of BaTiO3 by directreaction of TiCl4 with Ba(OH)2 in aqueous solution. J. Eur. Ceram. Soc. 2003, 23, 1383–1390. (28) Testino, A.; Buscaglia, M. T.; Buscaglia, V.; Viviani, M.; Bottino, C.; Nanni, P. Kinetics and mechanism of aqueous chemical synthesis of BaTiO3 particles. Chem. Mater. 2004, 16 (8), 1536–1543. (29) Testino, A.; Buscaglia, V.; Buscaglia, M. T.; Viviani, M.; Nanni, P. Kinetic modeling of aqueous and hydrothermal synthesis of barium titanate (BaTiO3). Chem. Mater. 2005, 17, 5346–5356. (30) Hancock, J. D.; Sharp, J. H. Method of comparing solid-state kinetic data and its application to decomposition of kaolinite, brucite, and BaCO3. J. Am. Ceram. Soc. 1972, 55 (2), 74–77. (31) Avrami, M. Kinetics of phase change I - General theory. J. Chem. Phys. 1939, 7 (12), 1103–1112. (32) Avrami, M. Granulation, phase change, and microstructure kinetics of phase change. III. J. Chem. Phys. 1941, 9 (2), 177–184. (33) Asiaie, R.; Zhu, W.; Akbar, S. A.; Dutta, P. K. Characterization of submicron particles of tetragonal BaTiO3. Chem. Mater. 1996, 8 (1), 226–234. (34) Joshi, U. A.; Yoon, S. H.; Baik, S. G.; Lee, J. S. Surfactantfree hydrothermal synthesis of highly tetragonal barium titanate nanowires: A structural investigation. J. Phys. Chem. B 2006, 110 (25), 12249–12256. (35) Karma, A. Lectures on Dendritic Growth. In Branching in Nature; Fleury, V., Gouyet, J. F., Leonetti, M., Eds.; Springer-Verlag: New York, 2001; pp 365401. (36) Lencka, M. M.; Riman, R. E. Thermodynamic modeling of hydrothermal synthesis of ceramic powders. Chem. Mater. 1993, 5 (1), 61–70. (37) Vivekanandan, R.; Philip, S.; Kutty, T. R. N. Hydrothermal preparation of Ba(Ti,Zr)O-3 fine powders. Mater. Res. Bull. 1987, 22 (1), 99–108.

3365

dx.doi.org/10.1021/cg101466u |Cryst. Growth Des. 2011, 11, 3358–3365