Article pubs.acs.org/crystal
Tuning the Microstructure of γ‑Ba4Nb2O9 Polymorph Prepared from Single-Molecular Precursor Jasminka Popović,* Martina Vrankić, and Marijana Jurić Ruđer Bošković Institute, Bijenička cesta 54, 10000 Zagreb, Croatia ABSTRACT: Although γ-Ba4Nb2O9 was considered to be metastable at room temperature (RT) and, therefore, prepared only by quenching from high temperatures, a new procedure for preparation of pure γ-Ba4Nb2O9 from a singlemolecular precursor, the oxalate-based complex {Ba2(H2O)5[NbO(C2O4)3]HC2O4}·H2O, is described. This study focuses on controlling the phase composition and the crystallite domain lengths by altering preparation conditions, namely, (i) the time for which samples were held at the given temperature and (ii) the cooling rate. The high-temperature γ-Ba4Nb2O9 polymorph has been successfully retained and stabilized at RT; the desired crystallite size in the nanoscale regime, ranging from ∼5 to 20 nm, can easily be tuned. The crystallite domain length and lattice strain were calculated from X-ray diffraction line-broadening analysis performed during Rietveld structure refinement.
■
due to their catalytic behavior.14,15 Recent investigation emphasized that Ba4Nb2O9 exhibits mixed electronic, oxide ion, and proton conductivity, which makes it especially attractive in the field of fuel cells, steam electrolyzers, and humidity sensors.15 The Ba4Nb2O9 exists in various polymorphic forms: low-temperature α-modification, high-temperature γ-modification, and closely related β-modification.16−20 Most recently, a new hexagonal δ-polymorph was reported, having a 6H-perovskite structure. The γ-Ba4Nb2O9 exhibits several orders higher conductivity than α-Ba4Nb2O9 due to a faster protonic and oxide ionic transport.15 The enhanced proton conduction in the hydrated γ-Ba4Nb2O9 phase originates from the structural peculiarities: 2D layers containing Nb5+ cations with the low oxygen coordination number (4 or 5) separated by discrete OH groups that facilitate ionic transport.15,21 The γ-Ba4Nb2O9 has been prepared mostly by quenching from high temperature, followed by multiple reheating steps, typical for conventional ceramics.15−20 On the other hand, we recently reported a study on the oxalatebased three-dimensional (3D) network, {Ba2(H2O)5[NbO(C2O4)3]HC2O4}·H2O (1), which can act as a single-source precursor for the formation of bimetallic BaII−NbV oxides.21 Up to then, the γ-polymorph was considered to be metastable at RT and so could be isolated only by quenching the sample from high temperatures, because slow cooling led to a transformation to α-Ba4Nb2O9.15−20 Interestingly, we managed to obtain a mixture of γ- and δ-polymorphs that was stable at RT, simply by cooling oxalate precursor 1 at 1175 °C. Under specific conditions (at 1175 °C held for 3h, heating rate of 3 °C
INTRODUCTION The synthesis of nanoparticles has been studied extensively in order to understand the physics and chemistry at a “small” scale and because of promising applications due to their sizedependent properties.1,2 There are two major reasons why nanocrystals and their bulk counterparts differ from each other: (i) in nanocrystals, the number of surface atoms is a large fraction of the total, and (ii) the size of the nanocrystal interior causes a systematic transformation in the density of electronic energy levels.1−3 Namely, in any material, the surface atoms make a distinct contribution to the free energy, so the changes in thermodynamic properties of the nanocrystals compared to those of the crystalline bulk material of the same composition can appear.3 Also, changes in thermodynamic stability associated with size can induce modification of cell parameters and/or structural transformations.4−8 To display mechanical or structural stability, a nanoparticle must have a low surface free energy. As a consequence of this requirement, phases that have a low stability in bulk materials can become very stable in nanostructures. This structural phenomenon has been detected in TiO2, VOx, Al2O3, or MoOx oxides.5−8 For (semi)conductors, as the particle gets smaller, an additional effect, the so-called quantum-size effect or confinement effect, which essentially arises from the presence of discrete, atom-like electronic states, influences the energy shift of exciton levels and optical band gap.3 Therefore, synthesis of nanoparticles with a controlled size has been found to be demanding for tailoring the desired material properties since the thermal, electric, optical, catalytic, and magnetic properties are strongly composition-, structure-, but also, size- and shape-dependent.9−13 Interesting applications of multicomponent Nb-containing oxides, known ion conductors, can be found in the literature due to their ferroelectric and piezoelectric properties, as well as © 2013 American Chemical Society
Received: February 12, 2013 Revised: March 28, 2013 Published: April 10, 2013 2161
dx.doi.org/10.1021/cg400242h | Cryst. Growth Des. 2013, 13, 2161−2165
Crystal Growth & Design
Article
Table 1. Sample Notation, Synthesis Conditions, and the Results of X-ray Powder Diffraction (XRD) Qualitative and Quantitative Analysis Performed by the Rietveld Refinement32 sample
holding time (h)
cooling rate (°C min−1)
sample composition (wt %)
BN2-3 BN2-7 BN2-12 BN1-12 BN0.5-12
2 2 2 1 0.5
3 7 12 12 12
82.3(4) γ-Ba4Nb2O9 + 17.7(2) δ-Ba4Nb2O9 90.3(4) γ-Ba4Nb2O9 + 9.7(1) δ-Ba4Nb2O9 pure γ-Ba4Nb2O9 pure γ-Ba4Nb2O9 pure γ-Ba4Nb2O9
min−1, cooling rate of 3 °C min−1), the final product contained 81.5 wt % of γ-Ba4Nb2O9 and 18.5 wt % of δ-Ba4Nb2O9.21 The study aims at exploring the full potential of the proposed “molecular-to-materials” pathway related to greater control of the final phase composition and the resulting crystallite domain sizes. A systematic study of preparation conditions was undertaken, namely, tuning the final structural and microstructural parameters by a simple means, such as (i) the holding time at 1175 °C and (ii) the cooling rate for the cooling process from 1175 °C back to RT. This new pathway for preparation of γ-Ba4Nb2O9, rather than quenching, was highly desirable, since considerable crystal lattice defects are otherwise introduced, resulting in large average maximum lattice strains. Generally, the preparation method from single-molecular precursor,14,22−31 as compared to conventional methods, has several advantages. Further efforts should be directed into extending this procedure we applied on BaII−NbV oxide to targeted nanoscale tailoring of different mixed oxide materials.
■
equals 2:1, as it is in the starting precursor, the heterometallic oxalate-based complex 1, confirming the efficiency of the proposed synthetic route. No additional phases were found, indicating the possibility that the existence of a chelating ligand in the precursor prevented the metal separation during the oxide formation. Sample BN2-3 was characterized as a polymorphic mixture of both γ-Ba4Nb2O9 and δ-Ba4Nb2O9 phases (Table 1). For a starting structural model of γ-Ba4Nb2O9, we used the “average” orthorhombic structure (with a subcell of a well-ordered monoclinic supercell containing several hundred independent atoms) reported by Ling et al.15 All unit-cell parameters together with the profile function parameters were refined, but, due to the complexity of the structure, some atomic coordinates were constrained. The structure of δ-Ba4Nb2O9 was refined as a 6H-perovskite structure type, which, unlike the ideal BaTiO3 (P63/mmm), crystallizes in the reduced-symmetry space group P63/m.21 From our previously performed DCS measurements,21 it is known that the cooling process of 1, in the range from 1200 °C to RT, exhibits two exothermic maxima. The first exothermic effect at ∼1150 °C is associated with the γ → α transition and is in agreement with one noted by Ling15 and Bezjak.16,17 However, we have observed an additional exothermic maximum at 610 °C caused by the formation of the δ phase. The appearance of both γ and δ polymorphs in sample BN2-3 at RT indicated that only a partial γ → α transformation occurred, followed by an additional transition of the obtained amount of α phase to δ phase at ∼600 °C. Sample BN2-7 also contained both polymorphs; however, the δ-Ba 4 Nb 2 O 9 polymorph was present in a smaller amount (Table 1). The difference in phase compositions between samples BN2-3 and BN2-7 is a result of faster cooling in the case of sample BN2-7, which caused a less successful conversion of γ- to α-Ba4Nb2O9 (phase transition at 1150 °C), and consequently a smaller amount of δ-Ba4Nb2O9 present in the final sample at RT. On the other hand, samples BN2-12, BN1-12, and BN0.5-12 contained only the γ-Ba4Nb2O9 polymorph, with no additional phases present (Table 1). Since all samples prepared by a cooling rate of 12 °C min−1, BN2-12, BN1-12, and BN0.5-12, contained only the γ-Ba4Nb2O9 polymorph, it is evident that faster cooling prevented the γ → α phase transition, which resulted in the retention of the high-temperature (HT) γBa4Nb2O9 phase at RT. This result is noteworthy since, according to previous studies,15−20 the γ-polymorph is metastable at RT and, therefore, could only be isolated by quenching the sample from high temperatures, leading to unavoidable lattice strain. The reason for retention of the HT phase at RT will be the subject of our further investigations, although a plausible explanation might be that particles of the HT phase with a size smaller than the critical particle size do not transform to low-temperature forms, whereas those particles with a size above this critical size are subjected to transformation into low-temperature phases. Similar effects of
EXPERIMENTAL SECTION
Sample Preparation. The mixed BaII−NbV oxides were prepared using {Ba2(H2O)5[NbO(C2O4)3]HC2O4}·H2O (1) as a singlemolecular precursor. Details on preparation of 1 are given in our previous publication.21 All samples were prepared by heating finely ground crystalline powders of 1 in a Mettler-Toledo TGA/DSC Star 1 System analyzer to 1175 °C, in the stream of the synthetic air (20.5% O2 and 79.5% N2), but the time period during which the samples were held at that temperature and the cooling rate to RT were systematically altered. Preparation conditions and notation of the samples are given in Table 1. Samples are denoted by the letters BN, followed by two numbers, the first one being the holding time (at 1175 °C) in hours and the second one being the cooling rate in °C min−1. The products obtained by pyrolysis of precursor 1 were explored by X-ray powder diffraction at RT. Methods. X-ray powder diffraction (XRD) patterns were measured in reflection mode with monochromated Cu Kα radiation on a Philips diffractometer PW1830 in steps of 0.02° (2θ) in the 2θ range from 10 to 100° with a fixed counting time of 5 s per step. Rietveld refinement32 was performed by a Panalytical HighScore X́ pert Plus ver. 2.1.33 A polynomial model was used to describe the background. During the refinement, zero shift, scale factor, half-width parameters (U, V, W), asymmetry parameters, and peak shape parameters were simultaneously refined. Microstructural information, namely, volumeweighted crystallite domain lengths and average maximum lattice strain, were obtained in the course of Rietveld refinement.32 The Si powder was used as an instrumental broadening standard. Diffraction profiles for both, sample and standard, were described by the pseudoVoigt function.
■
RESULTS AND DISCUSSION Different nanocrystalline products obtained by thermal decomposition of metal-complex precursor 1 were investigated by X-ray powder diffraction at RT. The results of XRD qualitative and quantitative analysis performed on prepared samples by the Rietveld refinement32 are given in Table 1. The molar ratio of Ba/Nb of prepared bimetallic oxide Ba4Nb2O9 2162
dx.doi.org/10.1021/cg400242h | Cryst. Growth Des. 2013, 13, 2161−2165
Crystal Growth & Design
Article
Figure 1. Graphical result of the final Rietveld refinement32 of RT data for (a) BN2-12 and (b) BN0.5-12. Experimental data are shown in blue, the calculated pattern is red, and the background and the difference curves are gray. The blue vertical marks represent the positions of γ-Ba4Nb2O9.
data and a summary of the structure refinements for the samples BN2-12, BN1-12, and BN0.5-12 are given in Table 2. It was noticed that all samples can be characterized as strainfree samples. Calculated lattice strain for all samples was e = 0% or negligible, so it was omitted from the table. However, a dependence of crystallite size (volume-weighted crystallite domain lengths ⟨L⟩V) for γ-Ba4Nb2O9 on holding time was established. The graphic representation given in Figure 2 shows a correlation of the cooling rate and the holding time to crystallite size of prepared samples. Crystallite sizes were ∼20 nm for γ-Ba4Nb2O9 prepared with a holding time of 2 h (sample BN2-12) while it decreased to only ∼5 nm with shortening of the holding time to 1/2 h (sample BN0.5-12). The synthesis of samples BN2-12, BN1-12, and BN0.5-12 was repeated four times independently; sample composition was reproducible within standard deviations while crystallite sizes were reproducible within 2%. Also, XRD
size-dependent phase transformation are evidenced in the case of stabilization of the HT tetragonal ZrO2 phase at RT.34,35 Once the conditions for preparation of the pure γ-Ba4Nb2O9 phase were established, namely, heating of 1 to 1175 °C, followed by cooling with 12 °C per minute, the effect of holding time (at 1175 °C) was further investigated. Shortening of the time period for which samples were held at 1175 °C had no impact on the phase composition of samples BN2-12, BN112, and BN0.5-12. However, the observed difference in the width of the X-ray diffraction lines indicated the changes of crystallite sizes and/or lattice strain. A detailed line-broadening analysis was undertaken in order to establish the correlation of the time period for which samples were held at 1175 °C and microstructural parameters of γ-Ba4Nb2O9. Figure 1 shows the graphical result of the Rietveld refinement32 on powder X-ray diffraction data for samples BN2-12 and BN0.5-12. The crystal 2163
dx.doi.org/10.1021/cg400242h | Cryst. Growth Des. 2013, 13, 2161−2165
Crystal Growth & Design
Article
Table 2. Crystal Data and Summary of Structure Refinements for the Samples BN2-12, BN1-12, and BN0.5-12 phase
γ-Ba4Nb2O9
formula sum formula mass/g mol−1 density (calculated)/g cm−3 space group (No.)
Ba12.00Nb6.00O28.00H2.00 2655.39 5.59 Pmn21 (31)
sample
BN2-12
BN1-12
BN0.5-12
a/Å b/Å c/Å V/106 pm3 fitting mode profile function U V W asymmetry parameter 1 peak shape parameter 1 peak shape parameter 2 R (weighted profile)/% R (profile)/% GOF
6.0297(3) 12.4243(7) 10.551(1) 790.3 structure fit pseudo-Voigt −0.012(3) 0.42(1) 0.157(1) 0.00012(1) 0.22(3) 0.002(1) 3.70683 2.69910 1.07176
6.0215(9) 12.426(1) 10.530(3) 787.9 structure fit pseudo-Voigt 0.021(2) 0.43(5) 0.55(1) 0.00019(1) 0.15(1) 0.000 5.06023 3.52584 2.15013
6.016(1) 12.421(2) 10.551(4) 788.3 structure fit pseudo-Voigt −0.002(1) 0.38(1) 1.66(1) 0.00022(3) 0.651(7) 0.000 3.91673 2.84584 1.49880
■
Table 3. Volume-Weighted Crystallite Domain Lengths of γBa4Nb2O9 Phase in All Samples sample
volume-weighted crystallite domain lengths (nm)
BN2-3 BN2-7 BN2-12 BN1-12 BN0.5-12
20.8(9) 21.2(5) 20.3(8) 11.6(9) 5.9(6)
ACKNOWLEDGMENTS This research was supported by the Ministry of Science, Education and Sports of the Republic of Croatia (Grant Nos. 098-0982886-2893 and 098-0982904-2946).
■
(1) Fernández-García, M.; Martínez-Arias, A.; Hanson, J. C.; Rodriguez, J. A. Chem. Rev. 2004, 104, 4063−4104. (2) Burda, C.; Chen, X.; Narayanan, R.; El-Sayed, M. A. Chem. Rev. 2005, 105, 1025−1102. (3) Ayyub, P.; Palkar, V. R.; Chattopadhyay, S.; Multani, M. Phys. Rev. B 1995, 51, 6135−6138. (4) Millot, N.; Aymes, D.; Bernard, F.; Niepce, J. C.; Traverse, A.; Bouree, F.; Cheng, B. L.; Perriat, P. J. Phys. Chem. B 2003, 107, 5740− 5750. (5) Schoiswohl, J.; Kresse, G.; Surnev, S.; Sock, M.; Ramsey, M. G.; Netzer, F. P. Phys. Rev. Lett. 2004, 92, 206103 1. (6) McHale, J. M.; Auroux, A.; Perrotta, A. J.; Navrotsky, A. Science 1997, 277, 788−791. (7) Zhang, H.; Banfield, J. F. J. Mater. Chem. 1998, 8, 2073−2076. (8) Song, Z.; Cai, T.; Chang, Z.; Liu, G.; Rodriguez, J. A.; Hrbek, J. J. Am. Chem. Soc. 2003, 125, 8059−8066. (9) Murray, C. B.; Kagan, C. R.; Bawendi, M. G. Annu. Rev. Mater. Sci. 2000, 30, 545−610. (10) Alivisatos, A. P. Science 1996, 271, 933−937. (11) Heath, J. R. Acc. Chem. Res. 1999, 32, 388−388. (12) Puntes, V. F.; Krishnan, K. M.; Alivisatos, A. P. Science 2001, 291, 2115−2117. (13) Pan, Z. W.; Dai, Z. R.; Wang, Z. L. Science 2001, 291, 1947− 1949. (14) Bayot, D.; Tinant, B.; Devillers, M. Inorg. Chem. 2005, 44, 1554−1562. (15) Ling, C. D.; Avdeev, M.; Kutteh, R.; Kharton, V. V.; Yaremchenko, A. A.; Fialkova, S.; Sharma, N.; Macquart, R. B.; Hoelzel, M.; Gutmann, M. Chem. Mater. 2009, 21, 3853−3864. (16) Bezjak, J.; Jančar, B.; Rečnik, A.; Suvorov, D. J. Eur. Ceram. Soc. 2008, 28, 2771−2776. (17) Bezjak, J.; Rečnik, A.; Jančar, B.; Boullay, P.; Evans, I. R.; Suvorov, D. J. Am. Ceram. Soc. 2009, 92, 1806−1812. (18) Bezjak, J.; Abakumov, A. M.; Rečnik, A.; Kržmanc, M. M; Jančar, B.; Suvorov, D. J. Solid State Chem. 2010, 183, 1823−1828.
Figure 2. Correlation of the cooling rate and the holding time to the crystallite size of γ-Ba4Nb2O9.
patterns were retaken 1 h, 10 h, 48 h, 1 week, 1 month, and finally 2 months after cooling down to RT, and no changes with respect to phase composition nor morphology were detected.
■
REFERENCES
AUTHOR INFORMATION
Corresponding Author
*Tel: + 385 1 4561120. Mobile: +385 91 9574191. E-mail:
[email protected]. Notes
The authors declare no competing financial interest. 2164
dx.doi.org/10.1021/cg400242h | Cryst. Growth Des. 2013, 13, 2161−2165
Crystal Growth & Design
Article
(19) Leshchenko, P. P.; Lykova, L. N.; Kovba, L. M.; Stefanovich, S. Y.; Chechkin, V. V. Inorg. Mater. 1985, 21, 227−230. (20) Leshchenko, P. P.; Paromova, M. V.; Lykova, L. N.; Kovba, L. M. Vestn. Mosk. Univ., Ser. 2: Khim. 1979, 20, 148−151. (21) Jurić, M.; Popović, J.; Šantić, A.; Molčanov, K.; Brničević, N.; Planinić, P. Inorg. Chem. 2013, 52, 1832−1842. (22) Nunes, G. G.; Seisenbaeva, G. A.; Kessler, V. G. Cryst. Growth Des. 2011, 11, 1238−1243. (23) Thurston, J. H.; Whitmire, K. H. Inorg. Chem. 2003, 42, 2014− 2023. (24) Neo, K. E.; Ong, Y. Y.; Huynh, H. V.; Hor, T. S. A. J. Mater. Chem. 2007, 17, 1002−1006. (25) Louër, M.; Louër, D.; Gotor, F. J.; Criado, J. M. J. Solid State Chem. 1991, 92, 565−572. (26) Audebrand, N.; Auffrédic, J.-P.; Louër, D. Chem. Mater. 1998, 10, 2450−2461. (27) Boudaren, C.; Auffrédic, J. P.; Louër, M.; Louër, D. Chem. Mater. 2000, 12, 2324−2333. (28) Jeanneau, E.; Audebrand, N.; Louër, D. J. Mater. Chem. 2002, 12, 2383−2389. (29) Wada, S.; Narahara, M.; Hoshina, T.; Kakemoto, H.; Tsurumi, T. J. Mater. Sci. 2003, 38, 2655−2660. (30) Audebrand, N.; Raite, S.; Louër, D. Solid State Sci. 2003, 5, 783− 794. (31) Mahata, P.; Sarma, D.; Madhu, C.; Sundaresen, A.; Natarajan, S. Dalton. Trans. 2011, 40, 1952−1960. (32) Rietveld, H. M. J. Appl. Crystallogr. 1969, 2, 65−71. (33) X’Pert HighScore Plus Program, version 2.1; PANalytical: Almelo, The Netherlands, 2004. (34) Garvie, R. C. J. Phys. Chem. 1978, 82, 218−224. (35) Shukla, S.; Seal, S.; Vij, R.; Bandyopadhyay, S.; Rahman, Z. Nano Lett. 2002, 2, 989−993.
2165
dx.doi.org/10.1021/cg400242h | Cryst. Growth Des. 2013, 13, 2161−2165
