Materials Research Bulletin 45 (2010) 1762–1767

Large-scale synthesis of nanocrystals of barium titanate and other titanatesthrough solution-phase processes

Chen Chen a, Chao Li b, Qiao Su a, Qing Peng a,*a Department of Chemistry, Tsinghua University, Beijing 100084, PR Chinab Department of Materials Science and Engineering, Tsinghua University, Beijing 100084, PR China

A R T I C L E I N F O

Article history:

Received 25 November 2009

Received in revised form 26 March 2010

Accepted 8 June 2010

Available online 17 June 2010

Keywords:

Large-scale synthesis

Nanocrystal

Barium titanate

Strontium titanate

Solution-phase processes

A B S T R A C T

We reported a large-scale synthesis of nanocrystals of BaTiO3, SrTiO3, PbTiO3, SrxBa1�xTiO3 through low-

temperature and solution-phase processes without any surfactant. The series of samples were

characterized by X-ray diffraction (XRD), transmission electron microscopy (TEM) and scanning electron

microscopy (SEM). Samples obtained were of high purity, consisting of nanoparticles with fine

crystallinity and uniformity as well as good dispersibility in ethanol. This method might also offer an

effectively new way to synthesis other titanate nanocrystals with perovskite structure in the future.

� 2010 Elsevier Ltd. All rights reserved.

Contents lists available at ScienceDirect

Materials Research Bulletin

journal homepage: www.e lsev ier .com/ locate /mat resbu

1. Introduction

Owing to its excellent dielectric and ferroelectric properties,barium titanate (BaTiO3) with the perovskite structure has utilitiesin a wide variety of applications, such as multilayer ceramic chipcapacitors (MLCC), and so on [1–3]. Thus, research on large-scalesynthesis of barium titanate at low expense is of great importancein manufacture. In the past few years, following the trend ofminiaturization of electronic components, relevant researcheshave also been focused on the preparation of barium titanatenanoparticles to meet the ever-growing requirements [4].

Conventional industrial synthesis of barium titanate powdersadopts the method of solid-state reaction [5]. The preparation isperformed by high temperature treatment (above 1273 K) of solidmixture of barium carbonate and titanium dioxide powders, or bycalcinations of chemically derived intermediates. This process issimple but the powders obtained are often of low purity, withrelatively large size (0.5–1.5 mm) and poor uniformity. Thus, in orderto obtain products with higher purity and smaller particle size, albeitat lower temperature, many other synthetic methods have beenproposed, such as sol–gel method [6,7], hydrothermal method [8–13], microwave heating [14], co-precipitation method [15], etc. Ofcourse, each of them has certain advantages as well as disadvan-tages, while among them sol–gel method and hydrothermal method

* Corresponding author. Tel.: +86 10 62792798; fax: +86 10 62792798.

E-mail address: pengqing@mail.tsinghua.edu.cn (Q. Peng).

0025-5408/$ – see front matter � 2010 Elsevier Ltd. All rights reserved.

doi:10.1016/j.materresbull.2010.06.032

are widely employed owing to the relative simplicity of syntheticprocess and the high quality of their respective products. Sol–gelsynthesis is performed in low-temperature solutions, and both thesize of particles and the purity of products are fine. However, thestarting materials here are too expensive and the synthetic process istoo complicated for large-scale synthesis. Also, hydrothermalmethod has many advantages: low temperature and low cost forsynthesis, yet high purity and well morphology of the products. Still,there are also some limitations: the necessity of some pretreatmentsand post-treatments in most cases, and the relatively largeconsumption of time (from several hours to days).

In this paper, we propose a solvothermal method [16–18], andmore importantly, a refluence process subsequently developedfrom the solvothermal method for the preparation of bariumtitanate nanocrystals. The syntheses in both ways were performedin water/ethanol mixed solvent without any surfactant, onlyrelatively inexpensive reagents needed. The synthetic process issimple and at low temperature (85 8C, water bath). Compared withlow temperature direct synthesis (LTDS) proposed by Wada et al.[19,20], samples obtained are of higher purity, consisting ofnanoparticles with better crystallinity and finer uniformity.Though prepared without any surfactant, the powders have shownwell dispersibility in ethanol (no precipitate observed after severalweeks) and no significant difference has been observed in bothsamples. Hence, the processes we propose here, especially therefluence process, is appropriate to large-scale synthesis.

We have also attempted to extend this synthetic process to be ageneral strategy for other titanate nanocrystals with perovskite

[(Fig._1)TD$FIG]

Fig. 1. XRD pattern of BaTiO3: (a) prepared with solvothermal method (* can be

indexed as the peak of anatase TiO2 (1 0 1) crystal plane); (b) prepared with

refluence method.

C. Chen et al. / Materials Research Bulletin 45 (2010) 1762–1767 1763

structure. A series of other titanates: BaTiO3, SrTiO3, PbTiO3,SrxBa1�xTiO3 and rare-earth-doped BaTiO3 nanocrystals have beensuccessfully synthesized.

2. Experimental

2.1. Synthesis of BaTiO3 nanocrystals through solvothermal process

In a typical synthesis, NaOH (1.0 g) and Ba(NO3)2 (1.0 mmol)were, respectively dissolved in deionized water (10 mL) to form atransparent solution at room temperature. Ti(OC4H9)4 (1.0 mmol)was added in ethanol (30 mL) under magnetic stirring. The NaOHand Ba(NO3)2 solution formed was subsequently added into thestirring ethanol. The mixture was then transferred to an autoclave,which was sealed and heated to 140 8C for 12 h.

After the reaction, the autoclave was cooled to room tempera-ture naturally, and the products were precipitated by centrifuga-tion. The dried white powder was finally redispersed in ethanol.

2.2. Synthesis of BaTiO3 nanocrystals through refluence process

In a typical synthesis, TiCl4 (0.2 mol) dissolved in ethanol (1.0 L)was added into a three-necked flask with reflux condenser. BaCl2

(0.2 mol) and NaOH (40 g) were orderly added into the three-necked flask. The mixture was stirred for 3 h, and heated by waterbath (85 8C) in whole process.

The post-treatment in refluence process is similar to solvother-mal process as described in Section 2.1.

2.3. Characterization

The samples were characterized by a Bruker D8 Advance X-raydiffractometer (XRD) with Cu Ka radiation (l = 1.5418 A), the[(Fig._2)TD$FIG]

Fig. 2. TEM image of BaTiO3 nanocrystals: (a) prepared with solvothermal method; (b

temperature of 200 8C; (d) prepared with solvothermal method for the reaction time o

operation voltage and current keeping at 40 kV and 40 mA,respectively. The 2u range used was from 108 to 708 in steps of0.028 with a count time of 0.2 s. The size and morphology of all thenanocrystals were obtained by a JEOL JEM-1200EX transmissionelectron microscope (TEM).

3. Results and discussions

3.1. Characterization of samples prepared with both methods

Fig. 1 shows the XRD patterns of the samples prepared withboth solvothermal and refluence methods, which confirm that theproducts are barium titanate of pure phase. Compared to that of

) prepared with refluence method; (c) prepared with solvothermal method at the

f 72 h.

[(Fig._3)TD$FIG]

Fig. 3. A typical batch size of BaTiO3 nanocrystals prepared with refluence method

in a large scale.

C. Chen et al. / Materials Research Bulletin 45 (2010) 1762–17671764

Fig. 1(b), peaks in Fig. 1(a) are of higher intensities, indicatingsamples prepared by solvothermal reaction are of better crystal-linity. Other than this, no significant difference is observed.

The TEM image presented in Fig. 2(a) shows that the typicalsolvothermally synthesized sample consists of well-morphologynanocrystals, whose diameters are between 20 and 30 nm. Fig. 2(b)is an image of typical sample prepared by refluence means. Thenanocrystals here are of similar size, yet slightly worse morphology.

Both the XRD and TEM characterizations suggest that thesample prepared solvothermally are of slightly better quality thanthat by refluence means. However, the typical batch size in theformer way is merely at the order of 0.1 g, which limits itsapplication in large-scale synthesis. As for the refluence method,tens of grams of samples can be obtained as shown in Fig. 3,furthermore, even larger amount of product is easily accessible aslong as more starting materials are introduced and larger flask isadopted. In addition, the simplicity of the apparatus setup in therefluence way further facilitates this method extended toindustrial manufacture.

3.2. Influences of various factors

3.2.1. Influence of temperature

In the solvothermal process, synthetic temperature varied from200 8C (as the TEM image shown in Fig. 2(c)) down to 120 8C, and

[(Fig._4)TD$FIG]

Fig. 4. XRD pattern of BaTiO3: (a) prepared with half amount of NaOH in solvothermal pr

indexed as the peak of anatase TiO2 (1 0 1) crystal plane, *2 can be indexed as (0 0 4)

solvothermal process; (d) prepared with double amount of NaOH in solvothermal proc

barely any difference was observed. As a matter of fact, this is thevery observation that enlightened us to try the refluence synthesis,since at such a low temperature (120 8C), the conditions which thereagents undergo within the solvothermal process are quitesimilar to that in the refluence process.

ocess; (c) prepared with double amount of NaOH in solvothermal process (*1 can be

). TEM image of BaTiO3 nanocrystals: (b) prepared with half amount of NaOH in

ess.

C. Chen et al. / Materials Research Bulletin 45 (2010) 1762–1767 1765

3.2.2. Influence of reaction time

Various periods of reaction time were adopted in thesolvothermal process, from 3 to 96 h. As shown in Fig. 2(d),samples obtained after 72 h heating are with similar morpholo-gy to that in Fig. 2(a), only of larger size (around 50 nm indiameter). We have observed that after 3 h heating, the particlesize increases slowly, and the growth velocity decreases asreaction time is prolonged. As a result, particle growth almoststops after 72 h.

Similar experiments were also carried out with refluencemethod, with reaction time varying from 1 to 16 h, and similarresults were obtained, which indicate that reaction time longerthan 3 h has little influence on the quality of products. For the sakeof high efficiency and low consumption, 3 h may be appropriate tolarge-scale synthesis.

3.2.3. Influence of concentrations of reagents

Other than shorter reaction time, higher efficiency results fromhigher concentrations of reagents. When the concentrations ofboth reagents were doubled, product quality was barely influ-enced. However, when we further doubled the concentrations, thecrystallinity of product sharply decreased, mostly amorphouspowders obtained.

3.2.4. Influence of pH

The addition of NaOH can significantly influence the crystallin-ity and morphology of products in the solvothermal process. When[(Fig._5)TD$FIG]

Fig. 5. TEM images of BaTiO3 nanocrystals post-treated by hydrochlo

the NaOH amount was 50% less than that in the typicalexperiment, samples of poor crystallinity were obtained as theXRD pattern (Fig. 4(a)) and TEM image (Fig. 4(b)) shows. When theNaOH amount was doubled, cubic nanocrystals were obtainedwith larger size and better crystallinity, as shown in Fig. 4(c)and (d).

3.2.5. Influence of acid post-treatment on dispersibility

As shown in Fig. 1(a), barium titanate nanocrystals aggregate tosome extent. As a result, when these nanocrystals are dispersed inethanol, visible precipitate is observed after several days. Thedispersibility in ethanol can be improved by addition of acid [21].We introduced hydrochloric acid to adjust the pH. The results areas shown in Fig. 5.

From Fig. 5(a)–(d), dispersibility of the nanocrystals enhancesas the acidity increases. Fig. 5(d) shows that excessive additionof acid has great impact on both morphology and size of thesenanocrystals. The optimized pH value is 2–3; and correspondingTEM image (Fig. 5(c)) shows no aggregation of the particlesand no degradation of the morphology. After the optimizedpost-treatment, barium titanate nanocrystals can be stablydispersed in ethanol (no precipitate observed at least within 2weeks).

The improvement of dispersibility by addition of acid isprobably due to the electrostatic repulsion between the positivelycharged nanoparticles, which result from the adsorption of H+ onthe surface of neutral particles.

ric acid with different pH value: (a) 7, (b) 5–6, (c) 2–3 and (d) 1.

C. Chen et al. / Materials Research Bulletin 45 (2010) 1762–17671766

3.3. Probable mechanism of the synthetic process

Herein we propose a presumed synthetic mechanism. Before theintroduction of NaOH, small amount of white precipitate is alreadypresent in the system. When NaOH is introduced, the nucleation ofbarium titanate starts immediately, numerous crystal nucleiformed. Within the following heating process, the subsequentlyformed amorphous barium titanate dissolves and recrystallizes ontothe surface of the crystal nuclei formed beforehand. The bariumtitanate crystals formed are of cubic structure, with six identicalcrystal planes. As a consequence, the equal growth velocities on eachcrystal surface give rise to the ultimate formation of near sphericalnanocrystals. When the particle size reaches a certain threshold, thecrystallization–dissolution equilibrium is finally achieved, whichterminates the growth of crystals. Consequently, the crystal size iscontrolled at the order of nanometer. This mechanism is furtherconfirmed by the observation in reaction time varying experiments(see Section 3.2.1).

3.4. Extension to some other titanates

Some other titanates with excellent electric properties havealso been successfully synthesized following the same process,with corresponding starting materials.[(Fig._6)TD$FIG]

Fig. 6. (a) XRD pattern of SrTiO3, Sr0.5Ba0.5TiO3 and PbTiO3 prepared with solvotherma

prepared with solvothermal method.

Equal amount of Sr(NO3)2 as Ba(NO3)2 used in the typicalsolvothermal synthesis was introduced. Strontium titanate nano-crystals were obtained, as the XRD pattern and TEM image(Fig. 6(a) and (b)) show, with size around 20 nm, similar to thebarium titanate nanocrystals previously prepared.

SrxBa1�xTiO3 nanocrystals were also successfully synthesizedsolvothermally, with x = 0.4, 0.5, 0.6, by adjusting the ratio ofstrontium and barium introduced. Fig. 6(a) and (c), respectivelyshow the XRD pattern and TEM image of typical Sr0.5Ba0.5TiO3

nanocrystals, which further confirm the similarity betweenSr0.5Ba0.5TiO3 and BaTiO3 and SrTiO3.

We also attempted to synthesize PbTiO3 nanocrystals with thesame method. Only at 200 8C and with doubled amount of NaOHcan the PbTiO3 crystallize. The XRD pattern in Fig. 6(a) confirms thefine crystallinity, while the TEM image in Fig. 6(d) demonstratesthe irregular needle-like morphology of the nanocrystals.

Corresponding refluence synthesis experiments of the titanatesmentioned above were also carried out, only SrTiO3 andSrxBa1�xTiO3 nanocrystals afforded as expected. The refluencesynthesis for crystalline PbTiO3 nanoparticles is barely accessible,which is ascribed to the crystallization temperature (200 8C) ofPbTiO3, while the highest temperature available in refluence(100 8C) is unable to provide the similar environment in thesolvothermal system.

l method. TEM image of (b) SrTiO3, (c) Sr0.5Ba0.5TiO3 and (d) PbTiO3 nanocrystals

C. Chen et al. / Materials Research Bulletin 45 (2010) 1762–1767 1767

4. Conclusion

In summary, nanocrystals of BaTiO3, SrTiO3, PbTiO3 andSrxBa1�xTiO3 have been successfully synthesized through low-temperature solution-phase processes (solvothermal process andrefluence process). The refluence process has the features of lowsynthetic temperature, short reaction time, easy apparatusaccessibility as well as well morphology, high dispersibility andfine crystallinity of the products, hence is especially appropriate tolarge-scale synthesis.

References

[1] N.A. Hill, J. Phys. Chem. B 104 (2000) 6694.[2] A.J. Millis, Nature 392 (1998) 147.[3] T.K. Song, J. Kim, S.I. Kuwn, Solid State Commun. 97 (1996) 143.[4] M. Nirmal, L.E. Brus, Acc. Chem. Res. 32 (1999) 407.

[5] E. Brzozowski, M.S. Castro, J. Eur. Ceram. Soc. 20 (2000) 2347.[6] H. Matsuda, N.J. Kobayashi, Non-Cryst. Solids 271 (2000) 162.[7] B. Lee, J. Zhang, Thin Solid Films 368 (2001) 107.[8] J. Ju, D. Wang, J. Lin, G. Li, J. Chen, L. You, F. Liao, N. Wu, H. Huang, G. Yao, Chem.

Mater. 15 (2003) 3530.[9] Y. Mao, S. Banerjee, S.S. Wong, Chem. Commun. (2003) 408.

[10] H.Jyh. Chen, Y.W. Chen, Ind. Eng. Chem. Res. 42 (2003) 473.[11] E. Ciftci, M.N. Rahaman, M. Shumsky, J. Mater. Sci. 36 (2001) 4875.[12] H. Xu, L. Gao, J. Guo, J. Eur. Ceram. Soc. 22 (2002) 1163.[13] Y. Hakuta, H. Ura, H. Hayashi, K. Arai, Ind. Eng. Chem. Res. 44 (2005) 840.[14] Y. Ma, E. Vileno, S.L. Suib, P.K. Dutta, Chem. Mater. 9 (1997) 3023.[15] M.Z.C. Hu, G.A. Miller, E.A. Payzant, C.J. Rawn, J. Mater. Sci. 35 (2000) 2927.[16] Y.D. Li, H. Liao, Y. Qian, et al. Inorg. Chem. 38 (1999) 1382.[17] G.R. Patzke, F. Krumeich, R. Nesper, Angew. Chem. Int. Ed. 41 (2002) 2446.[18] Y.D. Li, H. Liao, Y. Qian, et al. Chem. Mater. 10 (1998) 2301.[19] S. Wada, H. Chikamori, T. Noma, T. Suzuki, Key Eng. Mater. 181–182 (2000)

19.[20] S. Wada, T. Tsurumi, H. Chikamori, T. Noma, T. Suzuki, J. Cryst. Growth 229 (2001)

433.[21] S.S. Tripathy, A.M. Raichur, J. Dispers. Sci. Technol. 29 (2008) 230.