J Mater Sci: Mater Electron DOI 10.1007/s10854-016-5610-2

Structural, ferromagnetic and optical properties of pure bismuth A-site polar perovskite Bi(Mg3/8Fe2/8Ti3/8)O3 synthesized at ambient pressure Tiantian Wang1 • Hongmei Deng2 • Wenliang Zhou1 • Pingxiong Yang1 Junhao Chu1

•

Received: 15 June 2016 / Accepted: 22 August 2016 Ó Springer Science+Business Media New York 2016

Abstract Bi(Mg3/8Fe2/8Ti3/8)O3 (BMFT) perovskite oxide ceramic has been synthesized by a conventional solid-state reaction method at ambient pressure. Structural characterization is detected by X-ray diffraction (XRD) and Raman spectroscopy, reflecting a polycrystalline perovskite structure. The peaks in XRD pattern shift toward lower angle, which indicates larger lattice constant than BiFeO3 (BFO). High frequency modes blue shift and the peak broadening can be observed clearly in micro-Raman spectroscopy compared with BFO patterns. Morphology analysis through scanning tunneling microscope shows dense and well-interlinked grains. Through the UV–Vis–NIR spectra, BMFT displays a narrow band gap of 2.23 eV, smaller than the 2.8 eV band gap of BFO. The room-temperature ferromagnetism of BMFT ceramic is measured for the first time, which can be attributed to the nonmagnetic ions of Mg2? and Ti4? ions replacing Fe3? ions that causing the remanent magnetic moments. These results will open an avenue to further design multiferroic data storage and photovoltaic devices.

& Pingxiong Yang [email protected] 1

Key Laboratory of Polar Materials and Devices, Ministry of Education, Department of Electronic Engineering, East China Normal University, Shanghai 200241, China

2

Instrumental Analysis and Research Center, Institute of Materials, Shanghai University, 99 Shangda Road, Shanghai 200444, China

1 Introduction Multiferroics which possess two or three of the so-called ‘ferroic’ properties: ferroelectricity, ferromagnetism and ferroelasticity have attracted great attention on fundamental study and photovoltaic applications [1–3]. As we know, the most studied room-temperature (RT) multiferroic material is BiFeO3 (BFO) because of its high polarization. At the same time, BFO has the lowest band gap (Eg) of 2.8 eV compared with most of ferroelectrics oxide having band gap over 3.0 eV. However, it’s still not capable for solar photovoltaic devices. Meanwhile at bulk state, by reason of the spiral spin structure, oxygen vacancies and multi valence of Fe (Fe2? and Fe3?), BFO is antiferromagnetic and has high leakage current. Due to the small size of Bi ions, it usually needs high pressure to synthesize the perovskite, there are only four pure bismuth A-site perovskites that can be formed under ambient-pressure until now, they are BFO [4], Bi(Mn4/3Ni2/3)O6 [5], Bi(Mg3/8 Fe2/8Ti3/8)O3 (BMFT) [6] and Bi(Ni3/8Fe2/8Ti3/8)O3 [4, 5]. BMFT as a RT multiferroic material has excellent optical property, extensively applied on photovoltaic devices. But because of the multi composition of B-site ions and complex calcinations process, synthesized difficultly within a short time, there are few articles about BMFT until now. For further studying, we decided to form BMFT ceramics and test the properties systematically. In this work, the BMFT ceramics were synthesized successfully though simplified experimental procedure, to study the structure, morphology, optical and magnetic properties. To our excitement, the as-prepared BMFT sample has a relatively small Eg of 2.23 eV compared with BFO (2.8 eV) [7]. Meanwhile, it is noted that the RT ferromagnetic was investigated firstly. These properties provide us a new thinking on fundamental study and practical application.

123

J Mater Sci: Mater Electron

2 Experimental details

3 Results and discussion

BMFT ceramics were prepared by a conventional solid-state reaction method using Bi2O3 (99 %), Fe2O3 (99 %), TiO2 (99 %) and MgO (98.5 %) as starting raw materials. The starting materials were mixed according to the stoichiometric ratio and ground in an agate mortar for at least 1 h. The resulting powder were presintering at 750 °C for 1 h, 800 °C for 2 h and 850 °C for 2 h. Then the sample were reground for another 1 h and finally sintered at 890 °C for 8 h. BFO ceramics were grounded for 1 h and sintering at 800 °C using Bi2O3 (99 %), Fe2O3 (99 %) as raw materials. The crystal structure of the as-prepared ceramics were detected by X-ray diffraction (XRD, Bruker D8 Advance, ˚ ). Raman scattering with Cu Ka radiation, k = 1.54056 A analysis for the samples was performed with a microRaman spectrometer (Jobin–Yvon LabRAM HR 800UV). The morphologies were characterized by scanning tunneling microscope (SEM, Philips XL30FEG). The optical absorption of the samples were studied by ultraviolet– visible-near infrared (UV–Vis–NIR) spectrophotometer (cary500, USA Varian) equipped with integration sphere. The magnetic properties were analyzed by physical property measurement system (PPMS-9, Quantum Design).

Figure 1 shows the XRD patterns of the BFO and BMFT samples. The BFO pattern matches well with the ICDD data (JPDF 86-1518). The BMFT pattern matches closely the data from previous articles [5, 8], which implies the material was synthesized successfully. The Bragg reflections for both BFO and BMFT ceramics are indexed to the rhombohedral structure of R3c space group [9]. The BMFT perovskite structure was stabilized by complex B-site ions. But as marked by ‘‘*’’, an impurity phase of Bi25FeO40 is detected. Figure 1b is the magnified patterns of (012) peaks around 22.4°, it can be obviously observed that BMFT peaks shift towards lower angle, which suggested a bigger lattice constant compared with BFO XRD pattern. This may be due to the bigger ion radius of Mg2?, making an expanded lattice. Figure 1c is the magnified patterns of (116) and (122) peaks around 51.5°, (116) and (122) peaks overlap with each other in BMFT, indicating the phase evolution of the rhombohedral structure [8]. This phase evolution can also be observed at (104) and (110) reflections at about 31°. According to the XRD data, basic parameters of BFO and BMFT are calculated and displayed in Fig. 1d. The

Fig. 1 a XRD patterns of BFO and BMFT samples. b The magnified patterns of (012) peaks around 22.4°. c The magnified patterns of (116) and (122) peaks around 51.5°. d Basic parameters of BFO and BMFT samples

123

J Mater Sci: Mater Electron

Fig. 2 Measured Raman scattering spectra (open circle) for BFO and BMFT samples, together with fitted spectra (red solid line) and the decomposed active modes (green solid line) for BFO sample (Color figure online)

crystallite size is calculated by Scherrer Equation D = kk/ (b*cos(h)), k = 0.89, b is the full width at half maximum, meanwhile, the tolerance factor is counted by t = (\rA[ ? rO)/H2(\rB[ ? rO), where \rA[ and \rB[ are average ion radius of A-site and B-site respectively [10]. In Fig. 1d, BMFT has a bigger lattice constant, consistent with Fig. 2b. The c/a ratio of BMFT is 2.459, smaller than 2.488 of BFO, indicating a smaller lattice distortion. Additionally, cell volume of BMFT is slightly bigger than BFO due to the larger ion radius of Mg2?, but the calculated crystallite size of BFO is great larger than BMFT. Besides X-ray diffraction, Raman spectroscopy is also a famous and useful way to investigate the lattice structure, as shown in Fig. 2. According to group theory, Raman active modes of BFO with space group R3c at RT can be regarded as C = 4A1 ? 9E in which A1 and E modes represent Raman-active and IR-active modes [11]. In

Fig. 2, two peaks at about 171, 225 cm-1 belong to A1symmetry longitudinal-optical phonons [A1(LO)] rather than transverse-optical phonons [A1(TO)] [12]. At the same time the peaks at around 135, 277, 292, 345, 370, 475, 525, 614 cm-1 are written transverse-optical phonons [E(TO)] [11, 13–15]. From other studies, it is noted that low frequency Raman modes are related to A–O (Bi–O) bonds, however higher frequency modes are consistent with B–O (Fe–O, Mg–O and Ti–O) bonds [16]. As a result, mode 1 in BMFT basically isn’t offset compared with BFO, but at frequency over 200 cm-1, modes of BMFT obviously blue shift, this phenomenon may be due to lighter Mg2? and Ti4? ions replaced Fe3? ions, influencing lattice vibration of BO6 octahedron. At the whole view of Raman pattern of BFO and BFMT, it is clearly that peaks of BMFT are broadened compared with BFO. Combining with the c/a ratio from Fig. 1d which is 2.4878 and 2.4590 for BFO and BMFT respectively, the degree of lattice distortion of BMFT is smaller than BFO, in other words, the BMFT lattice is more relaxation than BFO. It is known that lattice distortion is related to Raman peaks, so this may be a reason that can explain the broadening of peaks in BMFT. The SEM micrographs of all samples are displayed in the Fig. 3. Comparing Fig. 3a, b, the BFO sample has an irregular and well-interlinked grains, the size of BFO grains are obviously bigger than BMFT grains, and the intergranular porosity is looser in BFO than BMFT. It can be explained by Kirkendall effect [17]. The replacement of Fe3? by Mg2? and Ti4?, leading different diffusion rates, these doping particles have obstructive effect on growing process of Bi, Fe and O, lowering the growth rate of particles, leading a smaller grain size and decreasing the porosity. Therefore, in Fig. 2b, high density compact clusters of BMFT can be seen. Figure 4 shows the optical absorption spectra of the BMFT samples. The absorption cut-off wavelength is about 700 nm, which means BMFT has a good ability to observe

Fig. 3 a, b SEM micrographs of BFO and BMFT respectively

123

J Mater Sci: Mater Electron

Fig. 4 UV–Vis–NIR absorption spectra of BMFT ceramic sample. The inset is (ahm)2 versus hm for the absorption spectra

visible light and be used on solar devices [18]. In the inset, according to Kubelka–Munk function, (ahm)2 versus hm for the direct band gap materials where a is absorbance and hm is photon energy [19, 20], the optical Eg of BMFT can be obtained from the tangent line in Fig. 4b, suggesting that the Eg is about 2.23 eV for BMFT ceramics. It presents an obvious decrease compared to the BFO with the Eg of 2.8 eV [8]. The lower Eg of BMFT may be due to an increase in density of states in the valence band with the appearance of localized states in the band gap [21]. In the fundamental absorption regions, the band-to-band transition from the top of valence band to the bottom of conduction band directly leading to the absorption. For BFO, the bottom of conduction band is the Fe3? 3d orbital, and the top of valence band is set by O 2p orbital. Increasing the varieties of B-site ions conduce the lower state of conduction band, decreasing the difference between valence band and conduction band, as a result lowering the band gap. Figure 5a shows the magnetization versus magnetic field (M–H) curves of BFO and BMFT measured with the magnetic field from -10 to 10 kOe at RT. In can be seen that BFO curve is almost linear, indicating antiferromagnetic nature. When replacing Fe3? by Ti4? and Mg2?, the M–H line transits into S-type hysteresis, appearing relatively obvious ferromagnetism. This phenomenon can be explained using the model displayed in Fig. 5b. BFO is antiferromagnetic, it has antiparallel magnetic moment in nearest-neighbor site, and parallel magnetic moment in next-nearest-neighbor sites [22], so at macro level, BFO doesn’t show ferromagnetic behavior. Ti4? and Mg2? are nonmagnetic ions, doped at Fe3? site, spin-up and spindown pair at adjacent cannot be compensate, leading to the appearance of remanent magnetization [23]. At the same time, the doping of Ti4? introduces 3d0 orbital, adjacent oxygen have tendency to donate it’s two electrons to the

123

Fig. 5 a M–H hysteresis curves of BFO and BMFT ceramic samples. b Scheme of the spins arrangement on BFO and BMFT, light blue octahedrons represent Fe–O octahedrons, dark blue and purple octahedrons represent Ti–O and Mg–O octahedrons respectively (Color figure online)

empty 3d orbital causing oxygen vacancies, causing remanent magnetization. Generally, the occurrence of microscopic ferromagnetic needs a certain concentration of unbalanced spin moments and localized electrons to maintain strong effective coupling [24].

4 Conclusions In summary, BMFT ceramics were synthesized successfully and the structure, morphology, optical and magnetic properties were studied. From XRD results, the main diffraction peak of BMFT shifts towards lower angle, together with the overlap of peaks suggesting evolution of lattice structure. Micro-Raman spectroscopy displays that high frequency modes of BMFT show blue shift compared with BFO patterns, which is mainly due to lighter Mg2? and Ti4? ions replaced Fe3? ions influencing lattice

J Mater Sci: Mater Electron

vibration. And the more lattice relaxation of BMFT causes peaks broadening. SEM micrographs display smaller grains and decreased porosity in BMFT. The optical absorption spectra reveal that BMFT has a relatively narrow band gap of 2.23 eV. Doping nonmagnetic Ti4? and Mg2? ions break the offset of antiparallel magnetic moments between two neighbouring Fe3? ions, leading to ferromagnetic in BMFT. Acknowledgments This work was supported by the National Natural Science Foundation of China (61474045) and the State Key Basic Research Program of China (2013CB922300).

References 1. Z. Chen, W. Jin, J. Mater. Sci. Mater. Electron. 25, 4039–4045 (2014) 2. G. Dhir, P. Uniyal, N.K. Verma, J. Mater. Sci. Mater. Electron. 26, 3538–3544 (2015) 3. S. Sharma, P. Saravanan, O.P. Pandey, P. Sharma, J. Mater. Sci. Mater. Electron. 27, 5909–5915 (2016) 4. V.S. FilipNev, N.P. Smolyaninov, E.G. Fesenko, I.N. Belyaev, Kristallografiya 5, 958 (1960) 5. H. Hughes, M. Allix, C.A. Bridges, J.B. Claridge, X. Kuang, H. Niu et al., J. Am. Chem. Soc. 127, 13790–13791 (2005) 6. B.A. Craig, A. Mathieu, S.R. Matthew, X.J. Kuang, S. Iasmi, S.C. Derek et al., Angew. Chem. Int. Ed. 119, 8941–8945 (2007) 7. V. Kumar, A. Gaur, R.K. Kotnala, Superlattices Microstruct. 67, 233–241 (2014)

8. W.L. Zhou, H.M. Deng, P.X. Yang, J.H. Chu, Appl. Phys. Lett. 105, 111904 (2014) 9. J. Liu, H.M. Deng, X.Z. Zhai, T. Lin, X.J. Meng, Y.Y. Zhang et al., Mater. Lett. 133, 49–52 (2014) 10. M.K. Sharif, M.A. Khan, A. Hussain, F. Iqbal, I. Shakir, G. Murtaza et al., J. Alloys Compd. 667, 329–340 (2016) 11. Y. Yang, J.Y. Sun, K. Zhu, Y.L. Liu, L. Wan, J. Appl. Phys. 103, 093532 (2008) 12. M.K. Singh, H.M. Jang, S. Ryu, M.H. Jo, Appl. Phys. Lett. 88, 042907 (2006) 13. G.L. Yuan, S.W. Or, H.L.W. Chan, J. Phys. D 40, 1196 (2007) 14. S. Kamba, D. Nuzhnyy, M. Savinov, J. Sˇebek, J. Petzelt, J. Proklesˇka et al., Phys. Rev. B 75, 024403 (2007) 15. R. Palai, H. Schmid, J.F. Scott, R.S. Katiyar, Phys. Rev. B 81, 139903 (2010) 16. R.P.S.M. Lobo, R.L. Moreira, D. Lebeugle, D. Colson, Phys. Rev. B 76, 172105 (2007) 17. L. Klingera, E. Rabkin, Mater. Lett. 161, 508–510 (2015) 18. W.J. Zhang, J. Chen, X.X. An, Q. Wang, L.L. Fan, F.F. Wang et al., Dalton Trans. 43, 9255–9259 (2014) 19. J. Liu, H.M. Deng, X.Z. Zhai, H.Y. Cao, P.X. Yang, J.H. Chu, J. Mater. Sci. Mater. Electron. 26, 2977–2981 (2015) 20. L. Huang, H.M. Deng, J. He, X.K. Meng, L. Sun, P.X. Yang et al., J. Mater. Sci. Mater. Electron. 26, 3984–3988 (2015) 21. J. Liu, H.M. Deng, H.Y. Cao, X.Z. Zhai, J.H. Tao, L. Sun et al., Appl. Surf. Sci. 307, 543–547 (2014) 22. X.Z. Zhai, H.M. Deng, W.L. Zhou, P.X. Yang, J.H. Chu, Z. Zheng, Mater. Lett. 161, 423–426 (2015) 23. J. Wei, R. Haumont, R. Jarrier, P. Berhtet, B. Dkhil, Appl. Phys. Lett. 96, 102509 (2010) 24. W.L. Zhou, H.M. Deng, L. Yu, P.X. Yang, J.H. Chu, J. Appl. Phys. 117, 194102 (2015)

123