Design, fabrication and dielectric properties in core–double shell BaTiO₃-based ceramics for MLCC application

Abstract

BaTiO₃-based ceramics with a core–double shell structure were fabricated by precipitation and the sol–gel method. Bi(Zn_1/2Ti_1/2)O₃-BaTiO₃ (BZT-BT) or Nb oxide was chosen to be the shell-I (inner layer) or shell-II (outer layer) composition. The structure and dielectric properties were investigated by X-ray diffraction (XRD), HRTEM (High Resolution Transmission Electron Microscopy) and electron probe micro analysis (EPMA), with different core to shell ratio (n_c/n_s). Compared with the designed composition BT–Nb-(0.2BZT-0.8BT), BT-(0.2BZT-0.8BT)–Nb was found to possess improved dielectric temperature stability, where the capacitance variation ΔC/C ≦ ±15% was achieved over a temperature range of −60–155 °C, with the dielectric constant and dielectric loss being in the order of 1860 and 0.011 at room temperature.

---

1. Introduction

BaTiO₃ (BT) is the most actively studied ferroelectric material, the mainstay for multilayer ceramic capacitor (MLCC) dielectrics in the last few decades.^1–3 Forming a “core–shell” structure in BT dielectric ceramics is an effective approach to improve the properties of XnR MLCC (X gives the minimum temperature of −55 °C, n means the maximum temperature, such as 7 for 125 °C and 8 for 150 °C, while R symbolizes percentage of the capacitance variation limit ±15% in the whole temperature range, ΔC/C_25°C ≤ ±15%).^4–7 The existence of double dielectric anomalies over the studied temperature range, which is the characteristic phenomenon of “core–shell” structure in BT ceramics, can help to improve the temperature stability of the dielectric properties. In general, BT is the ferroelectric composition for “core” component while paraelectric oxides are the main compositions for the “shell” component.^8–10 For example, in Nb₂O₅ doped BT ceramics, the Nb⁵⁺ will diffuse into the crystal lattice, replacing Ti⁴⁺ as donor dopant and forming the chemically inhomogeneous “core–shell” structure,^11,12 to improve the temperature stability of the dielectric behavior. However, a chemically inhomogeneous “core–shell” structure is hard to form in the BT-based ceramics by a traditional solid state method, due to the fact that long soaking time at the sintering temperature will facilitate the oxide diffusion in the shell region.^13,14 On the contrary, a wet chemical coating method, such as sol–gel or precipitation coating, is expected to increase the core–shell structure formation rate. In addition, the accumulated dielectric properties of the gradient compositions in the “core–shell” structure are dominated by both core and shell parts.^15,16 Based on the above discussion, a core–double shell structure was designed to control the dielectric properties by tuning the double shell compositions and core–shell ratio (n_c/n_s).

In our previous work, BT based ceramics with a core–double shell structure were designed, where the shell-II (outer layer) composition was bismuth-based perovskite material Bi(Mg_1/2Ti_1/2)O₃-BaTiO₃ (BMT-BT),¹⁷ fabricated by two different approaches and a two-step sintering method. However, the solubility limit of the BMT composition in xBi(Mg_1/2Ti_1/2)O₃-(1 − x)BaTiO₃ perovskite structure is low, being only x = 0.07,¹⁸ making the property tuning difficult. In this work, a core–double shell structure was fabricated based on the traditional Nb-doped core–shell structure, where the Bi(Zn_1/2Ti_1/2)O₃-BaTiO₃ (BZT-BT)¹⁹ was selected as shell-I (inner layer) or shell-II (outer layer) composition. The designed core–double shell BaTiO₃-based powders were prepared by sol–gel and precipitation coating methods for different shells (Fig. 1), and the dielectric properties of ceramics with different core–shell ratios and shell compositions were studied.

Fig. 1 Schematic core–double shell structure materials and designed compositions for S-1 and S-2.

2. Experimental procedure

xBZT-(1 − x)BT ceramics were prepared by the traditional solid state method. Bi₂O₃ (99.99%), ZnO (99%), BaTiO₃ (99.0%) and TiO₂ (99.0%) powders were batched stoichiometrically and ball milled in alcohol for 24 hours and dried, followed by calcination at 1000 °C for 2 hours, 5 mol% excess Bi₂O₃ was added to compensate the bismuth loss. The calcined powders were then re-milled and pressed into pellets, sintered at temperatures between 1150–1400 °C for 2 hours. Phase purity of the sintered samples was determined using X-ray powder diffraction (XRD) (PANalytical X'Pert PRO).

A core–double shell structure was designed as S-1 and S-2 (Fig. 1). S-1: the composition of shell-I layer is Nb oxide, coating the BT powder (with average diameter of 400 nm) by precipitation method (denoted as BT/shell-I), and the composition of shell-II layer is 0.2BZT-0.8BT, which is coated BT/shell-I powder by sol–gel method (denoted as BT/shell-I/shell-II). S-2: the composition of shell-I is 0.2BZT-0.8BT and the composition of shell-II layer is Nb oxide, respectively.^20,21

The fabrication method of 0.2BZT-0.8BT sol was as following: first, Ti(C₄H₉O)₄ was dissolved in a citric acid solution with a pH value of 6. After stirring at 70 °C for 3 h, a transparent yellow aqueous solution was obtained (named Ti sol). Then, Zn(CH₃COO)₂·2H₂O, Ba(CH₃COO)₂ and Bi(NO₃)₃·5H₂O were stoichiometrically dissolved in acetic acid according to the nominal composition of 0.2BZT-0.8BT, and then added to the citric acid solution to help the complexation reaction complete. The pH value of the obtained mixture solution was adjusted to 6–7 with ammonia solution, and a transparent aqueous solution was achieved (refer to Bi–Zn–Ba sol). Finally, stable sol of 0.2BZT-0.8BT was obtained after stirring the mixture of Ti sol and Bi–Zn–Ba sol. On the other hand, Nb(OH)₅ was dissolved in 0.3 mol L⁻¹ oxalic acid solution to produce oxalic acid of Nb.

The procedure for S-1 was as following: the BT powder was dispersed by ultrasonic treatment about 30 min in isopropanol to obtain a BT slurry, and then the oxalic acid of the Nb compound was added to the slurry. The pH value of the slurry was adjusted to above 5 with ammonia solution, so the Nb compound was able to precipitate completely and coat on the BT particles. The dried powders were calcined at 500 °C for 2 h to burn out the organics and obtain the BT/shell-I powders. Then, the BT/shell-I powder was dispersed by ultrasonic treatment for about 30 min in deionized water and added to the prefabricated 0.2BZT-0.8BT sol. The obtained mixture was stirred to produce a gel at 85 °C, and then dried at 120 °C to obtain xerogel. Finally, the BT/shell-I/shell-II powders were obtained by calcining the xerogel at 650 °C for 5 h. The molar ratio (n_c/n_s) of the BaTiO₃ (core) to shell layers (including shell-I or shell-II) will affect the dielectric properties of the obtained ceramics, thus, compositions with n_c/n_s = 1 : 1, 1 : 2, 1 : 3 and 1 : 4 were investigated. Compared with S-1, the procedure sequence for S-2 was adjusted according to the composition, the precipitation method for the Nb oxide and the sol–gel method for the 0.2BZT-0.8BT.

Microstructure and composition analysis of the powder and sintered samples were performed by transmission electron microscopy (TEM) and energy dispersive spectroscopy (EDS) using a JEOL JEM-2100F operated at 200 kV. The element distributions of ceramics were analyzed by Electron Probe Micro-Analyzer (EPMA, JXA-8230). The samples were polished to get parallel surfaces, and then electroded using fire-on silver paste. Dielectric measurements were carried out on samples using a multi-frequency precision LCRF meter (4184A, Agilent) in the temperature range of −60–200 °C.

3. Results and discussions

3.1 Dielectric properties of xBi(Zn_1/2Ti_1/2)O₃-(1 − x)BaTiO₃ ceramics

Pure perovskite xBZT-(1 − x)BT solid solution can be formed in the range of x = 0.05–0.35, above which, a second phase of Bi₄Ti₃O₁₂ appears, revealing the solid solution limit (Fig. 2). According to the peak splitting of the {200} diffraction, as shown in Fig. 2b, the sample exhibits tetragonal symmetry between x = 0.05–0.10 and transforms to a pseudo-cubic phase when x > 0.1.^22,23

Fig. 2 XRD diffraction patterns for xBZT-(1 − x)BT ceramics.

The dielectric constant and ΔC/C_25°C as a function of temperature for xBZT-(1 − x)BT ceramics at a 1 kHz frequency are shown in Fig. 3. With the BZT content increasing, the value of the dielectric constant at room temperature is gradually decreased, accompanied by a diffused dielectric peak, giving a more smeared behavior. Meanwhile, the temperature variation of the capacitance (ΔC/C_25°C) is found to maintain the same trend with increasing BZT content, but with values gradually increased in the high temperature region (about 25–200 °C). To meet the requirement of the capacitance variation limit ±15% (dash line in Fig. 3b), BZT-BT with composition of x = 0.2 shows improved stability in the high temperature region, demonstrating that a certain amount of BZT will benefit the capacitance stability of BaTiO₃ at high temperature.

Fig. 3 Dielectric constant (a) and ΔC/C_25°C (b) as a function of temperature for xBZT-(1 − x)BT ceramics at 1 kHz frequency.

Temperature variations of the capacitance for 0.2BZT-0.8BT and 0.25BZT-0.75BT ceramics were found to meet ΔC/C_25°C ≦ ±15% in the high temperature range of up to 200 °C, but fail at the low temperature end (−55 °C). It should be noted that both 0.2BZT-0.8BT and 0.25BZT-0.75BT ceramics exhibit a wider temperature range (ΔC/C_25°C ≤ ±15%) when compared to other compositions (as compared in Table 1). However, considering the values of the dielectric constant and dielectric loss, the 0.2BZT-0.8BT ceramic was selected as “shell” composition for core–double shell structure.

Table 1 Dielectric properties for xBZT-(1 − x)BT at 1 kHz

BZT content	Tm	Dielectric constant	Dielectric loss	ΔC/C25°C ≤ ±15% temperature range
0.10	32	2755	0.05	−7–114
0.15	45	2010	0.06	−3–165
0.20	48	1750	0.08	2–200
0.25	74	1435	0.10	6–200
0.30	92	1240	0.11	7–48
0.35	104	1100	0.11	6–45

---

3.2 Structure and dielectric properties of BT-based ceramics with core–double shell structure

3.2.1 S-1 specimens. TEM images of Nb-oxide coated BT powder (BT/shell-I) (n_c/n_s = 1 : 1) by the precipitation method are given in Fig. 4a and b. The domain structure of BT can be examined in the powder, as shown in Fig. 4a, around which there is an amorphous coating layer (Fig. 4b), due to the low sintering temperature (500 °C). The powder of 0.2BZT-0.8BT coated BT/shell-I was also analyzed by TEM, as given in Fig. 4c and d, showing an obviously thicker coating layer around the core when compared with BT/shell-I.

Fig. 4 TEM micrograph of Nb-oxide coated BT powder (BT/shell-I) by precipitation method (n_c/n_s = 1 : 1) (a and b) and 0.2BZT-0.8BT coated BT/shell-I powder by sol–gel method (n_c/n_s = 1 : 1) (c and d).

The EDS analysis together with TEM was performed on S-1 ceramic samples and given in Fig. 5, in which the “core–shell” structure can be observed. To confirm the main composition of the core and shell, EDS analysis was employed on different locations (spots 1, 2, 3 and 4) of the grain, including the core and shell parts. It is obvious that the concentrations of Nb and Bi in the shell area (spot 1) are much higher than those on other spots (spots 2 and 3), corresponding to the shell-I composition, while the main composition of the core (spots 2 and 3) is BaTiO_3. In addition, the Ba/Ti relative concentration on spot 4 is found to be lower than those on spots 2 and 3, corresponding to shell-II composition.

Fig. 5 TEM-EDS micrograph of S-1 ceramics (n_c/n_s = 1 : 1).

3.2.2 S-2 specimens. TEM images of 0.2BZT-0.8BT coated BT powder (BT/shell-I) (n_c/n_s = 1 : 1) are shown in Fig. 6a and b. As observed for S-1, the BT powder has small average particles size (400 nm) and large specific surface area, leading to particle agglomeration. However, domain structure of the BT (shown by the arrow in Fig. 6a) can still be observed. The spacing of the lattice fringe at 0.2821 nm observed by HRTEM (Fig. 6b) corresponds to the (110) plane of the tetragonal BT crystal, whereas the lattice fringe at 0.4644 nm corresponds to the double (111) plane of the shell-I pseudo-cubic 0.2BZT-0.8BT, being in good agreement with the designed composition. In addition, the powder of Nb-oxide coated BT/shell-I was analyzed by TEM, as given in Fig. 6c and d. Similarly, particles are still in agglomeration, around which there are amorphous coating layers, with thickness being larger than that of the BT/shell-I, revealing the optimized coating behavior.

Fig. 6 TEM micrographs of 0.2BZT-0.8BT coated BT powder by sol–gel method (n_c/n_s = 1 : 1) (a and b HRTEM) and Nb-oxide coated BT/shell-I powder by precipitation method (n_c/n_s = 1 : 1) (c and d).

S-2 ceramic samples were analyzed by TEM, as shown in Fig. 7. The “core–shell” structure can be observed, similar to S-1, the core is ferroelectric BaTiO₃ with a typical domain structure, while the shell is the solid solution 0.2BZT-0.8BT or Nb oxide, which is in the paraelectric phase. To confirm the main compositions of the “core–shell” structure, three representative spots (corresponding to spots 1, 2 and 3) were analyzed by EDS. It is obvious that the concentrations of Bi, Zn and Nb in the core area (spot 2) are much lower than those in the shell (spot 1 and spot 3), indicating the main composition of the core is BaTiO₃. Meanwhile, it can be confirmed that the main compositions of the shell material follow the expectation. The Nb oxide-coating layer is the outer layer, corresponding to spot 1, where niobium and small amount of bismuth were detected, while Ba/Ti relative contents were found to decrease on spot 3 with increased bismuth concentration, revealing that the BT-BZT composition dominate the shell-I layer. Compared with S-1 “core–shell” structure (Fig. 5), the core component of S-2 was found to reduce, with increased shell thickness, which was expected to affect the dielectric properties.

Fig. 7 TEM-EDS micrograph of S -2 ceramic (n_c/n_s = 1 : 1).

S-2 ceramic samples were analyzed by Electron Probe Micro-Analyzer (EPMA), as given in Fig. 8, showing the distribution of the Ba element. High Ba concentration was observed at some spots, as marked by the black arrows, corresponding to the core portion, whose main composition is BaTiO₃. The distributions of Bi and Zn were found to be higher in the area where Ba concentration was low, as shown by the white arrows in the graph.

Fig. 8 EPMA micrograph of ceramic surface (n_c/n_s = 1 : 1).

3.3 Dielectric properties of S-1 and S-2 ceramics

The dielectric constant and temperature variation of capacitance as a function of temperature for the S-1 and S-2 samples with different n_c/n_s ratios are shown in Fig. 9 and 10, respectively. It was observed that the dielectric–temperature curves for both fabrication methods showed a “double peaks” effect, which is the typical feature of the “core–shell” structure.^11,24 For S-1 (Fig. 9), the dielectric peak was found to be suppressed at the low temperature region, but enhanced gradually as the shell volume fraction decreased at the high temperature region. It should be noted that there is only minimal dielectric constant variation with increasing n_c/n_s ratio. Therefore, the capacitance–temperature stability doesn't show significant improvement, regardless of the variation of n_c/n_s ratios for the S-1 samples (Fig. 9b). On the contrary, the dielectric constant for S-2 is sensitive to the n_c/n_s ratios, where the dielectric constant was found to increase in both low temperature and high temperature regions with improved capacitance–temperature stability, as a function of n_c/n_s ratio (Fig. 10).

Fig. 9 Dielectric constant (a) and ΔC/C_25°C (b) as a function of temperature for S-1 ceramics with different n_c/n_s ratio.

Fig. 10 Dielectric constant and ΔC/C_25°C as a function of temperature for S-2 ceramics with different n_c/n_s ratio.

The dielectric properties of the materials with a core–shell structure can be regarded as the accumulated properties of the ferroelectric core, paraelectric shell and the gradient compositions, using Lichteneckers equation log ε = V_c log ε_c + V_s log ε_s, where V_c, and V_s are the volume ratio of the core and shell, respectively, ε_c, and ε_s are the dielectric constants for the core and shell, respectively. Thus, it is expected that the dielectric properties of the ceramics could be modified by tuning the core–shell volume ratios.^25,26 Considering that the diffusion between multilayers will affect the dielectric properties, 0.2BZT-0.8BT/BT and Nb-oxide/BT were fabricated, and the EPMA elemental distribution testing was employed to analyze the ion diffusion at the interfaces (Fig. 11). There is no obvious diffusion between the BT and Nb-oxide interface, while there is obvious Bi diffusion from 0.2BZT-0.8BT to BT. Thus, in the S-1 samples, the major component of the core is BT according to the n_c/n_s ratio, leading to a relatively high dielectric constant peak at the Curie temperature (about 4500). However, in S-2, the core composition on the proximity of the 0.2BZT-0.8BT shell layer will fluctuate, leading to the reduced ferroelectric core (BT), being confirmed by TEM, as shown in Fig. 5 and 7, accounting for the decreased dielectric constant peak at the Curie temperature (from 2000 to 3500, depends on the n_c/n_s ratio, Fig. 11a). The dielectric constant corresponding to shell composition increases with n_c/n_s ratio increasing, due to the fact that the Bi diffusion make BZT component decrease correspondingly in the shell composition, leading to the increased dielectric constant. Therefore, the dielectric–temperature stability is improved for S-2 ceramics. Compared with the properties of the reported Nb-coated BT, double shells (0.2BZT-0.8BT and Nb oxides) of the S-2 sample with n_c/n_s = 1 : 1 helped to extend the temperature range, meeting the requirement of the ΔC/C ≦ ±15% (as summarized in Table 2).

Fig. 11 Element distribution EPMA micrograph for Nb–BT and BZT-BT/BT ceramics.

Table 2 Dielectric properties (25 °C) for S-2 ceramics with different n_c/n_s and selected Nb-doped BT at 1 kHz

Composition (nc/ns)	Dielectric constant	Dielectric loss	ΔC/C ≦ ±15% temperature range (°C)
1 : 1	1860	0.011	−60–155
2 : 1	2095	0.016	−60–115
3 : 1	2375	0.016	−60–119
4 : 1	2260	0.022	−32–106
Nb (3 mol%) coated BT	1983	0.006	−55–125 (ref. 27)
Nb (2 mol%) mixed BT	1601	0.015	−55–125 (ref. 27)
Nb (0.05 mol%) coated BT	1100	0.032	0–100 (ref. 24)

---

4. Conclusions

A core–double shell structure was designed using BZT-BT and Nb oxide as different coating compositions and BT as core composition, which were named as S-1 and S-2. Different structure design gives rise to divergences in microstructure of ceramic materials, playing an important role on the dielectric properties. The ceramic samples designed as S-2 exhibit improved dielectric properties when compared to S-1, with dielectric constant and loss being in the order of 1860 and 0.011 at room temperature, respectively, while the ΔC/C ≦ ±15% is maintained over the temperature range of −60–155 °C, meeting the X8R MLCC specification.

Acknowledgements

This work was supported by Natural Science Foundation of China (no. 51102189, no. 51372191), the program for New Century Excellent Talents in University (no. NCET-11-0685), International Science and Technology Cooperation Program of China (2011DFA52680) and the National Key Basic Research Program of China (973 Program) (no. 2015CB654601).

References

1. B. Tang, S. Zhang, X. Zhou, Y. Yuan and L. Yang, J. Electroceram., 2010, 25(1), 93.
2. S. Liu, H. Zhang, L. Sviridov, L. Huang, X. Liu, J. Samson, D. Akins, J. Li and S. O'Brien, J. Mater. Chem., 2012, 22, 21862.
3. Y. Mizuno, T. Hagiwara and H. Chazono, J. Eur. Ceram. Soc., 2001, 21, 1649.
4. M. T. Buscaglia, V. Massimo, Z. Zhao, V. Buscaglia and P. Nanni, Chem. Mater., 2006, 18(17), 4002.
5. H. Wen, X. Wang, Z. Gui and L. Li, J. Electroceram., 2008, 21, 545.
6. S. C. Jeon, C. S. Lee and S. J. L. Kang, J. Am. Ceram. Soc., 2012, 95(8), 2435.
7. H. Chazono and H. Kishi, J. Am. Ceram. Soc., 2000, 83(1), 101.
8. C. Huber, T. D. Mona, C. Elissalde, F. Weill and M. Maglione, J. Mater. Chem., 2003, 13, 650.
9. T. Schneller, S. Halder, R. Waser, C. Pithan, J. Dornseiffer, Y. Shiratori and L. Houben, J. Mater. Chem., 2011, 21, 7953.
10. R. Berthelot, B. Basly, S. Buffière and J. Majimel, J. Mater. Chem. C, 2014, 2, 683.
11. S. H. Yoon, J. H. Lee, D. Y. Kim and N. M. Hwang, J. Am. Ceram. Soc., 2003, 86(1), 88.
12. G. F. Yao, X. H. Wang, Y. C. Zhang, Z. B. Shen and L. T. Li, J. Am. Ceram. Soc., 2012, 95(11), 3525.
13. D. Hennings and G. Rosenstein, J. Am. Ceram. Soc., 1984, 67(4), 249.
14. B. Xiong, H. Hao, S. Zhang, H. Liu, M. Cao and Z. Yu, Ceram. Int., 2012, 38S, S45.
15. L. Xie, X. Huang, C. Wu and P. Jiang, J. Mater. Chem., 2011, 21, 5897.
16. Z. Tian, X. Wang, L. Shu, T. Wang, T. H. Song, Z. Gui and L. Li, J. Am. Ceram. Soc., 2009, 92(4), 830.
17. H. Hao, H. Liu, S. Zhang, B. Xiong, X. Shu, Z. Yao and M. Cao, Scr. Mater., 2012, 67(5), 451.
18. B. Xiong, H. Hao, S. Zhang, H. Liu, M. Cao and Z. Yu, J. Am. Ceram. Soc., 2011, 94(10), 3412.
19. C. C. Huang and D. P. Cann, J. Appl. Phys., 2008, 104, 24117.
20. Y. Zhang, X. Wang, J. Y. Kim, Z. Tian, J. Fang, K. H. Hur and L. Li, J. Am. Ceram. Soc., 2012, 95(5), 1628.
21. L. Xie, X. Huang, Y. Huang, K. Yang and P. Jiang, J. Phys. Chem. C, 2013, 117(44), 22525.
22. N. Triamnak, R. Yimnirun, J. Pokorny and D. P. Cann, J. Am. Ceram. Soc., 2013, 96(10), 3176.
23. C. C. Huang and D. P. Cann, J. Appl. Phys., 2008, 104(2), 024117.
24. M. Cernea, B. S. Vasile, A. Boni and A. Iuga, J. Alloys Compd., 2014, 587, 553.
25. J. S. Park and Y. H. Han, J. Eur. Ceram. Soc., 2007, 27, 1077.
26. Z. Li, L. A. Fredin, P. Tewari, S. A. DiBenedetto, M. T. Lanagan, M. A. Ratner and T. J. Marks, Chem. Mater., 2010, 22, 5154.
27. J. J. Gan and W. J. Wei, Int. J. Appl. Ceram. Technol., 2009, 6(6), 661.

---