Influence of the annealing temperature of the Bi₄Ti₃O₁₂ seeding layer on the structural and electrical properties of Bi_3.15Nd_0.85Ti_2.99Mn_0.01O₁₂ thin films

Abstract

Highly (117)-preferred Bi_3.15Nd_0.85Ti_2.99Mn_0.01O₁₂ (BNTM) thin films with a Bi₄Ti₃O₁₂ (BTO) seeding layer were fabricated on Pt(111)/Ti/SiO₂/Si(100) substrates by using a sol–gel method. The effects of the Bi₄Ti₃O₁₂ (BTO) seeding layer under different annealing temperatures ranging from 550 to 700 °C on the structural and electrical properties of BNTM were investigated. X-ray diffraction results indicated that the BNTM thin films with a BTO layer processed with the annealing temperature of 600 °C exhibited the highest (117) orientation at a degree of 97.33%. This typical BNTM film also had the largest remanent polarization (2P_r = 114.5 μC cm⁻²), dielectric constant (ε_r = 614.9) and dielectric tunability (16.9%) as compared to the BNTM thin films without a seeding layer or with the BTO layer processed at a different temperature. It is also found that the significant enhancement of the piezoelectric properties was achieved in these typical BNTM thin films. Additionally, the BNTM thin films with BTO seeding layers displayed better fatigue properties, degraded by only 1.1% after 10⁹ pulse cycles as compared to 30.2% for those without seeding layers. The mechanism of the temperature dependence of BTO seeding layer on the properties of BNTM will be discussed.

---

Introduction

Bi_3.15Nd_0.85Ti₃O₁₂ (BNT) thin films have always been the most promising materials for their potential applications in next-generation information technology devices, such as ferroelectric field effect transistors (FeFET), thin-film transistors (TFT), ferroelectric tunnel junction (FTJ), and magnetoelectric effect (ME) devices.^1–6 Bi₄Ti₃O₁₂ (BTO) is a monoclinic crystal with a lattice constant along the c-axis (c = 3.284 nm) that is considerably larger than those along the other two axes (a = 0.544 nm, b = 0.541 nm) at room temperature. The spontaneous polarizations of BTO are about 50 and 4 μC cm⁻² along its a and c axes, respectively.⁷ Chon et al. reported that the c-axis oriented BNT capacitor exhibited a very high value of remanent polarization (2P_r > 100 μC cm⁻²).⁸ On the contrary, Garg et al. suggested that the polarization increased on moving away from the BNT c-axis on the underlying SrRuO₃/SrTiO₃ substrates.⁹ Lu et al. also found that BNT thin film with a-axis preferential orientation showed the largest remanent polarization in comparison with that on c-axis preferential orientation.¹⁰ It was further confirmed by Hu that the remanent polarization of (100)-predominant BNT film is substantially higher than those of (117)-preferred and (001)-oriented films.¹¹

It has been shown that many factors such as film composition, layer thickness, annealing condition, the precursor solution and the nature of substrates have great impacts on the orientation of BNT films.^4,9,12–14 Hu found that the film layer thicknesses of 30, 50, and 100 nm favor the formation of (001), (100), and (117)-preferred BNT films, respectively.¹¹ Zhong et al. reported that Bi_3.15Nd_0.85Ti_2.99Mn_0.01O₁₂ (BNTM) thin film exhibited enhanced dielectric constant and tunability.¹⁴ But the best annealing temperature of BNTM thin film was found to be 750 °C, higher than that of the BNT thin film (700 °C).¹⁵ Recently, a BTO seeding layer was applied to provide nuclei for crystallization resulting in substantial improvement of the electrical properties of Bi_3.25La_0.75Ti₃O₁₂ (BLT) thin films whose structure are similar to that of the BNT thin film.^16,17 However, very few reports are found on the studies of whether the BTO seeding layer could improve the performance of BNT thin films.

In order to acquire the good performance of BNTM film at low temperatures, highly (117)-preferred BNTM films with a BTO seeding layer were prepared by a sol–gel method in this work. Effects of BTO seeding layer under different annealing temperature between 550 to 700 °C on the structural and electrical properties of BNTM films were investigated. We will also discuss the mechanism of the temperature dependence of BTO seeding layer on the properties of BNTM.

Experimental

Synthesis and characterization

All chemicals and reagents were supplied by Sinopharm Chemical Regent, Co., Ltd. The Bi_3.15Nd_0.85Ti_2.99Mn_0.01O₁₂ and Bi₃Ti₄O₁₂ films were prepared by a sol–gel method. The precursor solution for the coating was prepared by first dissolving 3.396 g of bismuth nitrate [Bi(NO₃)₃·5H₂O, purity ≥ 99.0%] and 0.753 g of neodymium nitrate [Nd(NO₃)₃·6H₂O, purity ≥ 99.0%] in the mixed solution of acetic acid (purity ≥ 99.5%) and 2-methoxyethanol (purity ≥ 99.0%). About 2.076 g of titanium isopropoxide [Ti(OC₃H₇)₄, purity ≥ 99.0%] was slowly added with continuous stirring into the 2-methoxyethanol solution (6–8 mL) containing 2–3 drops of acetylacetone [C₅H₈O₂, purity ≥ 99.0%]. 0.005 g of manganese acetate [Mn(CH₃COO)₂·4H₂O, purity ≥ 99.0%] was also dissolved into the Ti-solution after adding a little amount of acetic acid. The Ti-solution was then added to the mixed solution containing Bi and Nd ions with continuous stirring. Finally, 2-methoxyethanol was added to adjust the concentration to obtain 20 mL 0.1 mol L⁻¹ BNTM precursor solution. The concentration of BTO solution prepared by the same method was 0.02 mol L⁻¹. These solutions were stirred for 10 hours at room temperature. The proportion of acetic acid and 2-methoxyethanol was set to 1 : 3. A 10 mol% excess of bismuth nitrate was added to compensate for possible bismuth loss during the high temperature process. The precursor solution was aged for 3–7 days before its use for spin coating.

The BNTM solution was spin coated on Pt/Ti/SiO₂/Si substrates at a rate of 4000 rpm (5000 rpm for BTO solution) for 40 s, followed by a drying process at 180 °C for 5 min, a pyrolysis process at 400 °C for 5 min, and an annealing process at 700 °C for 5 min under an O₂ pressure of 1.5 atm. The rapid thermal annealing (RTP) method was used for annealing process at a ramping rate of 15 °C s⁻¹. These processes were repeated four times to obtain the desired thickness. The bottom BTO seeding layers under the same drying and pyrolysis process were annealed at 550, 600, 650 and 700 °C for 2.5 min under an O₂ pressure of 1.5 atm. Films prepared with the BTO seeding layer processed under the annealing temperature of 550, 600, 650 and 700 °C are denoted by BTO-550, BTO-600, BTO-650 and BTO-700, respectively. All samples were cut into slices with a diamond knife and a part with good parallelism and clean cross-section was chosen to take scanning electron microscopy (SEM) pictures. The mean thicknesses of BNTM films were observed to be ∼421 nm according to the cross-sectional SEM images (as shown in Fig. 2). The thickness of the BTO seeding layer was estimated to be ∼10 nm. Pt top electrodes with a diameter of 200 μm were deposited on BNTM films by dc sputtering.

The crystallographic structure and the texturing state of these thin films were studied by X-ray diffraction (XRD, Rigaku Ultima IV, Japan) with Cu-Kα radiation. Cross-sectional and surface morphologies of these films were characterized by Scanning Electron Microscopy (SEM, Hitachi S4800, Japan), which has a high resolution below 1 nm with a cold field emission gun operating at 5 kV. The surface microstructures and roughness of BNTM films were determined by an atomic force microscope (AFM, Bruker Multimode 8, USA) working in a contact mode in ambient conditions. The dielectric and leakage properties of these films were measured using a semiconductor device analyzer (Agilent B1500A, USA). The ferroelectric property was analyzed by a ferroelectric test systems (Radiant Technologies Precisions workstations, USA). Piezoelectric responses of these films were characterized by using a piezoresponse force microscope (PFM, MFP-3D, Asylum Research, USA). Thermogravimetry analysis (TGA) tests were conducted to study the thermal properties of BTO and BNTM xerogels. These tests have been done using a thermogravimetric analyzer (STA-409 PC, NETZSCH, 6 Germany) under an air atmosphere with a gas flux of 40 mL min⁻¹. Samples were heated from 30 °C to 850 °C at a heating rate of 10 °C min⁻¹.

Results and discussion

Microstructure and morphology

X-ray diffractions presented a highly (117)-preferred growth of BNTM, BTO-550, BTO-600, BTO-650 and BTO-700 are shown in Fig. 1(a). XRD profiles focusing on peak value of (117) are shown in Fig. 1(b). To quantify the texturing state, we define the degree of orientation as α₁₁₇ = I₍₁₁₇₎/(I₍₀₀₆₎ + I₍₁₁₇₎ + I₍₂₀₀₎), where I_(hkl) is the XRD peak intensity of (hkl) crystal plane. The degrees of (117) orientation of BNTM, BTO-550, BTO-600, BTO-650 and BTO-700 were found to be 66.50%, 92.33%, 97.33%, 96.30% and 85.38%, respectively. It can be concluded from Fig. 1(b) that the BNTM films with BTO seeding layer possess a much higher (117) orientation than those without seeding layer. This implies that the BTO seeding layer provides nuclei for (117)-preferred grains growth. What's more, BTO-600 shows the highest peak value of (117) orientation than what BTO-550, BTO-650 and BTO-700 do. It is also clear that the peak value of (117) decreases with the annealing temperature increasing from 600 to 700 °C. As the annealing temperature of seeding layer increases, the competition between the nucleation and grain growth may determine the intensity of the diffraction peaks. A more detailed explanation can be found in the discussion part.

Fig. 1 XRD patterns: (a) BNTM thin films prepared without and with BTO seeding layer; (b) XRD profiles of (117) peaks.

The SEM micrographs of BNTM thin films' surface and cross-section without or with BTO seeding layer are shown in Fig. 2(a)–(j). It can be seen that all the BNTM films deposited on Pt/Ti/SiO₂/Si substrates are dense and crack-free. The BNTM thin films with BTO seeding layer show the larger grains than the BNTM thin films (without BTO seeding layer), which is further confirmed by AFM micrographs as shown in Fig. 3(a)–(e). It is worth noting that all the films are mainly composed of the rod-like grains of (117)-oriented with different length (as shown in Fig. 2(a)–(e)) while the (200)-oriented grains and (001)-oriented grains are more or less equiaxed and plate-like, respectively.^11,18 The average grain sizes of BNTM, BTO-550, BTO-600, BTO-650 and BTO-700 were observed to be 419.2 nm, 435.5 nm, 463.3 nm, 548.3 nm and 589.4 nm, respectively. It is obvious that the grain size increases with the annealing temperature of the BTO seeding layer (as seen in Fig. 2(a)–(e)). It is likely that the BTO seeding layer processed at 700 °C preserves the largest (117)-preferred grain size corresponding to the lowest surface energy for (117) plane.¹⁹ Film thicknesses of BNTM, BTO-550, BTO-600, BTO-650, and BTO-700 were observed to be 421 nm, 435 nm, 426 nm, 426 nm and 433 nm according to the cross-sectional SEM images (as shown in Fig. 2(f)–(j)), respectively, which was in favour of the (117)-oriented BNTM thin films with a thickness of 105 nm for each layer, in agreement with what has been founded by Hu.¹¹ What's more, a better crystallization of BTO-600, BTO-650 and BTO-700 was found from the bottom part of BNTM thin films compared with BNTM and BTO-550. The corresponding root mean square (Rms) roughness of such films was found to be 10.3 nm, 12.3 nm, 12.2 nm, 12.7 nm and 12.9 nm, respectively (as shown in Fig. 3(f)–(j)). The BNTM films with BTO seeding layer likely have a rougher surface than the BNTM thin films without BTO seeding layers, probably due to their much larger grains as reported by Pei et al.¹⁷

Fig. 2 SEM surface and cross-section images: (a) and (f) for BNTM; (b) and (g) for BTO-550; (c) and (h) for BTO-600; (d) and (i) for BTO-650; (e) and (j) for BTO-700.

Fig. 3 AFM planar and three dimensional images: (a) and (f) for BNTM; (b) and (g) for BTO-550; (c) and (h) for BTO-600; (d) and (i) for BTO-650; (e) and (j) for BTO-700.

Electrical properties

Fig. 4 shows the plot of the leakage current density (J) of BNTM, BTO-550, BTO-600, BTO-650 and BTO-700 thin films at room temperature as the function of electric field. Current densities of BNTM, BTO-550 and BTO-600 are of the same order of 1 × 10⁻⁴ A cm⁻² at an electric field of 100 kV cm⁻¹, which is one order of magnitude lower than that of BTO-650 and BTO-700 when the positive electric field is applied. It can be inferred that the exacerbated evaporation of Bi from the surface of BTO seeding layer under annealing temperatures of 650 °C and 700 °C would increase the oxygen vacancies leading to the increase of current density. However, when the negative electric field is applied, the current density of BNTM is remarkably larger than the other films. This can be ascribed to the poor interface between BNTM and the bottom electrode of Pt, as the BTO seeding layer was proved to prevent the formation of a Bi deficient or Bi₂Pt layer.¹⁶ To better understand the behaviour of the leakage current of BNTM, an inset plot of log(J) versus log(E) for BNTM under positive electric field (as shown in Fig. 4) is divided into three distinct regions. One can easily see from Fig. 4 that the leakage current behaviour is dominated by space-charge-limited conduction (SCLC) as a relation of J_SCLC ∝ E^α.²⁰ In the SCLC mode, α is equal to 1 in the Ohm's law region and larger than 2 in the trap-filled limit region. In our case, as shown in Fig. 4, α is 0.9 (close to 1) under 20 kV cm⁻¹, and is 2.5 (>2) from 20 kV cm⁻¹ to 106 kV cm⁻¹. However, the realistic mechanism for the leakage current behaviour may be also influenced by ferroelectric polarization as α is significantly larger than 2, to 17.4 when E is above 106 kV cm⁻¹ (value of coercive field, E_c). The similar and detailed arguments can be found in some other works.^21–23

Fig. 4 Plot of log J (leakage current density) versus log E curve for BNTM, BTO-550, BTO-600, BTO-650 and BTO-700. The inset shows log J versus log E characteristic of BNTM and the linear fit for the slop.

Fig. 5(a) shows the variation of dielectric constant (ε_r) and dielectric loss tangent (tan δ) as the function of frequency for BNTM, BTO-550, BTO-600, BTO-650 and BTO-700. The dielectric constants and dissipation factors of such films at 10 kHz were 514.1, 353.6, 614.9, 435.1, 514.1 and 0.089, 0.045, 0.075, 0.08, 0.075, respectively. BTO-600 was found exhibits the largest dielectric constant, while BTO-550 has the smallest dielectric constant. Fig. 5(b) plots the C–V curves for different type of BNTMs at the frequency of 1 MHz. The hysteresis behaviours are confirmed by the butterfly shape for all the curves. The BTO-600 film displays the highest switching peaks (373.4 pF) and dielectric tunability (16.9%) compared with the others. Fig. 5(c) displays the polarization–electric field (P–E) hysteresis loops. At an applied electric field of 270 kV cm⁻¹, the remanent polarization 2P_r and the coercive field 2E_c of such films were found to be about 75.4, 59.2, 114.5, 65.8, 70.9 μC cm⁻² and 198.4, 186.8, 203.8, 197.3 and 192.8 kV cm⁻¹, for BNTM, BTO-550, BTO-600, BTO-650 and BTO-700, respectively. It is obvious that the BTO-600 has the highest remanent polarization, which is consistent with the results demonstrated in Fig. 5(a) and (b). To further study the piezoelectric properties of BNTM thin films, the piezoelectric tests were performed. Piezoelectric test results are shown in Fig. 5(d). It is found that the BTO-600 thin films demonstrate the largest piezoelectric. This enhancement can be interpreted by the larger grain in the BNTM thin films with the BTO seeding layer cause dipoles to polarize more sufficiently along the electric field. While for those films without BTO seeding layer, more crystal boundaries may constraint the piezoelectric deformation.

Fig. 5 The electrical properties of BNTM, BTO-550, BTO-600, BTO-650 and BTO-700 for (a) dielectric constant and dielectric loss curves; (b) C–V curves at 1 MHz; (c) P–E hysteresis loops at 1 kHz; (d) piezoelectric hysteresis loops.

Fig. 6 displays the fatigue characteristics of BNTM thin films in an applied electric field 190 kV cm⁻¹ at 100 kHz, while . Where P_N is the total polarization, N is the switching cycles, is the switched remanent polarization between the two opposite polarity pulses, and P^{^}_r is the non-switched remanent polarization between the same two polarity pulses. Values of dP_N/dP₀ of the five thin films were found to be 69.8, 87.1, 91.7, 89.7 and 98.9% after 10⁹ pulse cycles, respectively, indicating that the larger grain size can decrease density of oxygen vacancy in BNTM films and thus reduce the number of pinning domains as depicted by Zhong et al.²⁴ It is evident that the bottom BTO seeding layer has a quite positive effect on the fatigue properties.

Fig. 6 Fatigue characteristics of BNTM, BTO-550, BTO-600, BTO-650 and BTO-700.

Discussion

Fig. 7 shows the mass loss curves of BNTM and BTO xerogels. It can be obtained that perovskite structures of BNTM and BTO start to crystallize at 575 and 525 °C, respectively. Thus the BTO seeding layer has already started to crystallize at a temperature of 550 °C, which is further confirmed by the XRD pattern analysis and AFM planar images of BTO seeding layer (as shown in Fig. 8 and 9). The intensity of diffraction peak of (117) plane for the BTO layer increases with the annealing temperature as shown in Fig. 8. Possibly, it may be concluded that the seeding layers consisted by different size of (117)-oriented grains were formed as nuclei for BNTM thin films as shown in Fig. 9. Thus the larger grains for BTO-550, BTO-600, BTO-650 and BTO-700 should be due to the larger nuclei provided by the BTO seeding layer.

Fig. 7 Mass loss curves of BNTM and BTO xerogels.

Fig. 8 The X-ray diffraction patterns of BTO seeding layer with the annealing temperature ranging from 550 to 700 °C.

Fig. 9 AFM planar images for various annealed temperature BTO seeding layer: (a) 550 °C; (b) 600 °C; (c) 650 °C; (d) 700 °C.

It is generally known that grain growth and nucleation have decided the grain orientation distribution and ferroelectric property in a film.^25,26 As for the BTO seeding layer, the role of grain growth is more important when the BTO seeding layer crystallization occurs at low temperature from 550 to 600 °C, while in the opposite case the nucleation is the decisive role from 600 to 700 °C. The highest peak intensity and the largest remanent polarization (2P_r) for BTO-600 were ascribed to the optimal value between grain growth and nucleation as compared to those of BTO-550, BTO-650 and BTO-700. And the large grain of BTO layer under 650 °C and 700 °C annealing had suppressed the nucleation of the crystals, which was shown in Fig. 9. Thus the intensity of the diffraction peaks increase with increasing annealing temperature of seeding layer from 550 to 600 °C, and then decrease from 600 to 700 °C.

Conclusions

Highly (117)-oriented BNTM thin films were prepared with the BTO seeding layer processed under various annealing temperature. The orientation and morphology of BNTM thin films were strongly influenced by the annealing temperature of BTO seeding layer. It is found that 600 °C should be the most proper annealing temperature for BTO seeding layer. Enhanced dielectric and ferroelectric properties were achieved by BTO-600 thin films. Additionally, the improved piezoelectric properties were observed in BTO-600 thin films. Moreover, all the thin films with BTO seeding layer show better fatigue characteristics degraded by only 1.1% after 10⁹ pulse cycles as compared to 30.2% for those without seeding layers. The above results indicate that adding a BTO seeding layer at proper annealing temperature is a very effective way to improve the structural and the electrical properties of BNT thin films, which provides some insight into the application of BNT or BNTM thin films in ferroelectric devices.

Acknowledgements

This work was financially supported by the Project of National Natural Science Foundation of China (NSFC) (Grant No. 61274107, 51472210 and 61404113), Hunan Provincial Innovation Foundation for Postgraduate Grant No. CX2014B267.

Notes and references

1. D. Xie, Y. F. Luo, X. G. Han, T. L. Ren and L. T. Liu, J. Appl. Phys., 2009, 106, 4117.
2. H. J. Song, T. Ding, X. L. Zhong, J. B. Wang, B. Li, Y. Zhang and Y. C. Zhou, RSC Adv., 2014, 4, 60497.
3. S. G. Yuan, J. B. Wang, X. L. Zhong, F. Wang, B. Li and Y. C. Zhou, J. Mater. Chem. C, 2013, 1, 418.
4. Z. Y. Lu, C. H. Yang, G. D. Hu, J. C. Wang and X. Wang, J. Alloys Compd., 2010, 508, 106.
5. C. P. Cheng, M. H. Tang, Z. H. Tang and Y. C. Zhou, J. Sol-Gel Sci. Technol., 2013, 68, 346.
6. Z. H. Tang, Y. Xiong, M. H. Tang, Y. G. Xiao, W. Zhang, M. L. Yuan, J. Ouyang and Y. C. Zhou, J. Mater. Chem. C, 2014, 2, 1427.
7. A. D. Rae, J. G. Thompson, R. L. Withers and A. C. Willis, Acta Crystallogr., 1990, B46, 474.
8. U. Chon, H. M. Jang, M. G. Kim and C. H. Chang, Phys. Rev. Lett., 2002, 89, 087601.
9. A. Garg, Z. H. Barber, M. Dawber, J. F. Scott, A. Snedden and P. Lightfoot, Appl. Phys. Lett., 2003, 83, 2414.
10. C. J. Lu, Y. Qiao, Y. J. Qi, X. Q. Chen and J. S. Zhu, Appl. Phys. Lett., 2005, 87, 222901.
11. G. D. Hu, J. Appl. Phys., 2006, 100, 6109.
12. F. Hou, M. R. Shen and W. W. Cao, Thin Solid Films, 2005, 471, 35.
13. J. Yu, B. Yang, J. Li, X. M. Liu, C. D. Zheng, Y. Y. Wu and D. M. Zhang, J. Phys. D: Appl. Phys., 2008, 41, 035304.
14. X. L. Zhong, J. B. Wang, M. Liao, C. B. Tan, H. B. Shu and Y. C. Zhou, Thin Solid Films, 2008, 516, 8240.
15. X. L. Zhong, B. Li, J. B. Wang, M. Liao, H. Liao and Y. C. Zhou, Mater. Lett., 2008, 62, 2891.
16. K. T. Kim and C. I. Kim, Thin Solid Films, 2004, 447, 413.
17. L. Pei, N. Hu, G. Deng, Y. G. Bie, Y. W. Chen and M. Y. Li, J. Electron. Mater., 2015, 44, 2340.
18. H. N. Lee, D. Hesse, N. Zakharov and U. Gosele, Science, 2002, 296, 2006.
19. S. B. Ren, C. J. Lu, J. S. Liu, H. M. Shen and Y. N. Wang, Phys. Rev. B: Condens. Matter Mater. Phys., 1996, 54, R14337.
20. M. A. Lampert and P. Mark, Current Injection in Solids, Academic, New York, 1970.
21. J. C. Shin, C. S. Hwang, H. J. Kim and S. O. Park, Appl. Phys. Lett., 1999, 75, 3411.
22. F. Yang, F. W. Zhang, G. D. Hu, Z. H. Zong and M. H. Tang, Appl. Phys. Lett., 2015, 106, 172903.
23. W. Zhang, Y. Q. Gao, L. M. Kang, M. L. Yuan, Q. Yang, H. B. Cheng, W. Pan and J. Ouyang, Acta Mater., 2015, 85, 207.
24. N. Zhong, P. H. Xiang, Y. Y. Zhang, X. Wu, X. D. Tang, P. X. Yang, C. G. Chun and J. H. Chu, J. Appl. Phys., 2015, 118, 104102.
25. B. Yang, D. M. Zhang, B. Zhou, L. H. Huang, C. D. Zheng, Y. Y. Wu, D. Y. Guo and J. Yu, J. Cryst. Growth, 2008, 310, 4511.
26. W. Sun, J. C. Wang, G. D. Hu and J. Yan, J. Mater. Sci.: Mater. Electron., 2013, 24, 2853.

---