Flash sintering of highly insulating nanostructured phase-pure BiFeO3

1 Flash Sintering of highly insulating nanostructured phase-pure BiFeO3 Luis A. Perez-Maqueda,a,b Eva Gil-Gonzalez,a Antonio Perejon,a,c Jean-Marie Lebrun,b,d Pedro E. Sanchez-Jimeneza, Rishi Rajb aInstituto de Ciencia de Materiales de Sevilla, Consejo Superior de Investigaciones Científicas−Universidad de Sevilla. Calle Américo Vespucio 49, Sevilla 41092, Spain bDepartment of Mechanical Engineering, University of Colorado at Boulder, Boulder, CO 80309-0427, USA cDepartamento de Química Inorgánica, Facultad de Química, Universidad de Sevilla, Sevilla 41071, Spain dPresent address: Saint-Gobain Innovative Materials – NRDC, Process Technology CRL - Forming & 3D Printing Group, 9 Goddard Road - Northboro, MA 01532, USA Submitted as a Communication to the Journal of the American Ceramic Society May 2017 Abstract We show that BiFeO3 that is electrically homogeneous, is a good insulator, and has a low dielectric constant (the properties desired in its applications) can be produced by flash sintering, which is nominally difficult to achieve by conventional and spark plasma sintering processes. The flash sintered specimens had a uniform microstructure with a nanometric grain size of ~20 nm. Keywords: Flash Sintering, BiFeO3, high-purity, insulating, electrically homogeneous, nanostructured. 2 1. Introduction Bismuth iron oxide is a multiferroic material, having applications in data storage and in magneto-electric sensors.1-3 Nevertheless, most conventionally prepared materials have high leakage current densities that limit their applications. This high leakage current has been associated to secondary phases4-7, the generation of oxygen vacancies, the loss of bismuth, or the reduction of Fe3+.8-10 Obtaining stoichiometric high-purity and dense pellets of BiFeO3 is quite challenging as this compound decomposes at relatively low temperatures producing undesirable phases.11-14 Thus, conventional sintering, which requires high temperatures and long treatments, often produces non-stoichiometric compounds with high leakage currents. Spark Plasma Sintering (SPS)15-22 yields high-density BiFeO3, but the highly reducing conditions produced by the graphite dies cause reduction of the sample and an increase in the electrical conductivity.20-21 Flash sintering was first applied to 3 mol% yttria-stabilized tetragonal zirconia.23 Different compounds have been sintered in this way.24-27 In the present work we explore the application of flash sintering and its effect on the electrical properties of BiFeO3. We are able to sinter BiFeO3, which not only has high density but also is single phase with a nanoscale grain size, and has desirable electrical properties, including insulating behavior. 2. Experimental Methods In the present work, nanoscale powders of BiFeO3 were synthesized by milling the stoichiometric amounts of the single commercial oxides, that is, Fe2O3 (Sigma −Aldrich, St. Louis, Missouri, USA, 310050, ≥99% purity) and Bi2O3 (Sigma−Aldrich, St. Louis, Missouri, 3 USA, 223891, 99.9% purity), using a high energy planetary ball mill (Fritsch Pulverisette 7, Fritsch GmbH, Idar-Oberstein, Germany) under 7 bar of oxygen pressure.28 Flash sintering experiments were performed using the standard procedure.23 The powders were uniaxially pressed at 500 MPa into dog-bones shaped pellets. The samples were suspended into a tubular furnace by means of two platinum wires attached to the handles of the dog bone. The furnace was heated from room temperature at a constant rate of 10 ºC min-1 up to flash event. A DC electric field was applied through the two platinum wires using a DC power supply. The field was held constant up to the point of the flash event, which was signaled by a non-linear rise in conductivity. The voltage control at the power supply was switched to current control when the current reached a preset limit. Constant current was held for 15 seconds before the power supply to the specimen was switched off. The furnace was then allowed to cool down to room temperature. The change in the sample dimensions was monitored with a CCD camera. These data were used to measure the linear shrinkage arising from sintering. X-ray diffraction patterns were recorded with a PanalyticalX’Pert Pro (Panalytical B.V., Almelo, The Netherlands) diffractometer. Sample microstructure was analyzed by scanning electron microscopy (SEM) using a Hitachi S-4800 (Hitachi, Ltd., Tokyo, Japan) instrument equipped with energy dispersive X-ray spectrometer (EDX). Impedance spectroscopy measurements of Au sputter-coated samples were performed, taking into account the blank capacitance of the sample holder and the overall pellet geometry, in a Newtons4th Ltd (Leicester, United Kingdom) impedance analyzer over the frequency range from 100 Hz to 1 MHz, with AC measuring voltage of 0.1 V. 4 3. Results Plots of the power dissipation, which is equal to the product of the electric field and the current density, are shown in Fig. 1a. The rise in current under voltage control signals the onset of the flash. The peak in the power density occurs when the power supply is switched to current control. The plateau thereafter corresponds to the steady state of flash. The sample sinters near the power peak. The specimen temperature at the power peak can be calculated from the blackbody radiation model;29 these values, which have been derived in the Supporting Information, are shown as TBBR in the figure. The shrinkage data are shown in Fig. 1b. The 0 V data represent conventional sintering. At 15 V cm–1, the sintering behavior remains the same as in conventional sintering. At 50 V cm–1 and above, the sample sinters by the flash mechanism. As the field is increased, sintering occurs at a lower temperature, for example, at 150 V cm–1, the sample sinters at a furnace temperature of ~400 oC. The diffraction patterns of sintered specimens are given in Fig. 1c. The retention of pure BiFeO3 perovskite phase under flash sintering is indeed noteworthy. Conventionally sintered BiFeO3 decomposes into secondary phases. Figure 1. a) Power dissipation and (b) linear shrinkage as a function of furnace temperature at constant heating rate of 10°C min-1 for different applied fields. Specimen temperatures, TBBR, at the power peak have been calculated from the blackbody radiation model. c) X-ray diffraction patterns of the resulting pellets obtained from experiments at different applied fields. 0 and 15 V cm-1 patterns resulted very similar, so only the 15 V cm-1 is shown. 5 The microstructure of the specimen flashed at 100 V cm –1 and 20 mA mm –2 is shown in Fig. 2. The specimen is well sintered with low porosity. A histogram of the grain size distribution showing an average grain size of 20±6 nm is also shown in Fig. 2. The chemical composition of the sample, as determined by EDX is given in Table S1. There is no loss of bismuth under flash sintering conditions. Figure 2. SEM micrograph and histogram plot of the grain-size distribution for BiFeO 3 (100 V cm -1 , 20 mA mm -2 ). The electrical properties of the specimens were measured by impedance spectroscopy for the sample that was flashed at 100 V cm -1 , with a current limit of 20 mA mm –2 . The measurements were done with a frequency range of 100 Hz to 1 MHz at 0.1 V. The impedance plots, giving the real and imaginary parts, are given in Fig. 3a, which includes SPS data for comparison. The graph highlights the much higher resistance of the flash-sintered sample. The inset in Fig. 3a, giving the data for the SPS sample, shows two arcs, one from grain boundaries and the other from grain matrix. 20-21 In contrast, the flashed sample shows a single arc, suggesting no difference between grain boundary and grain matrix resistivity. Arrhenius plots of the conductivity for flash-sintered and SPS specimens are shown in Fig. 3b. 20-21 The flash sintered specimens are more insulating. 6 The activation energies of flash sintered samples was 1.01 eV; the value for high quality single crystals has been reported as 1.3 eV.30 The measurements of conductivity by impedance spectroscopy are compared with the change in conductivity of the powder specimen during the constant heating rate flash-sintering experiment at a field of 100 V cm-1 in Fig. S1. Surprisingly, the conductivity of the porous specimens is higher than the impedance measurements by about an order of magnitude, which it is contrary to the expectation that pores and interparticle interfaces would lower the conductivity.31 Perhaps the difference is related to nonlinear effects under the much higher fields used in the flash experiments than in impedance spectroscopy. Figure 3. a) Impedance complex plane plots, showing the imaginary part (Z”) vs the real part (Z’) of the impedance at 300ºC for FS and SPS BiFeO3. b) Bulk Arrhenius plots for FS and SPS BiFeO3. The imaginary parts of the modulus, M’’, and the impedance, Z’’ as a function of frequency at 300ºC for flash sintered BiFeO3 are plotted in Fig. 4. They represent Debye peaks that reflect on the electrical homogeneity of the sample. The ideal RC element with the shape of a single Debye peak indicates a homogeneous material.32 In the present case, M’’ and Z’’ show 7 single peaks with a small displacement in their frequency maxima, as shown in Fig. 4a. These results show that the flash-sintered sample was electrically homogeneous. The capacitance values calculated from the imaginary part of the impedance are shown in Fig. 4b. The permittivity was calculated to be equal to 113; it was obtained by dividing the capacity by the permittivity of free space. Capacitance, equal approximately to ~10 pFcm-1, remains reasonably constant within the entire frequency range, again confirming that the sample was electrically homogeneous.32 Figure 4. a) Z”/M” spectroscopic plots at 300ºC. b) C’ vs frequency for BiFeO3 flash-sintered sample. 4. Discussion It is clear from above results that flash sintering helps to preserve the phase purity and compositional fidelity in (complex) ceramic oxides which are otherwise prone to decomposition and volatilization. Since sintering requires mass transport, its rate is controlled by the slowest diffusing species in the solid state, which may require temperatures where low melting and volatile species may be difficult to retain. However, it may be rather simplistic to imagine that 8 lower (furnace) sintering temperatures and shorter sintering time afforded by flash sintering can of themselves alleviate this issue. Flash sintering has complexity, and its role in obtaining unusual results may not be reduced to rather simple arguments. For example, while in the present case flash preserves the phase fidelity, there are examples where metastable phases can form during the flash-sintering experiments, for example, in the yttria-stabilized zirconia system,33 and more recently, in the alumina-magnesium aluminate system.34 A fuller understanding of how phase transitions and composition fidelity are influenced not only by time, temperature, and environment, but also by electrical field and current, during flash experiments, is far from complete. 5. Conclusions Flash sintering enables the processing of BiFeO3 into a pure perovskite, single phase structure, which is hard to obtain in conventional sintering as the material is unstable at temperatures required for processing. Moreover, high-density nanostructured ceramics with narrow grain-size distribution were obtained. The conductivity of flash sintering specimens was nearly one order of magnitude lower than samples made by SPS. Acknowledgments Financial support from Projects CTQ2014-52763-C2-1-R (MINECO-FEDER) and TEP- 7858 (JuntaAndalucía-FEDER), is acknowledged. AP thanks VPPI-US for his contract. LAPM acknowledge MECD and Fulbrigh-commission for a grant to support his visit to UCB. JML and 9 RR gratefully acknowledge support from the Office of Naval research under Grant Number, N00014-15-1-2505. References (1) Andrew, J. S.; Starr, J. D.; Budi, M. A. K. Prospects for nanostructured multiferroic composite materials. Scripta Materialia 2014, 74, 38-43. (2) Eerenstein, W.; Mathur, N. D.; Scott, J. F. Multiferroic and magnetoelectric materials. Nature 2006, 442, (7104), 759-765. (3) Roy, A.; Gupta, R.; Garg, A. Multiferroic Memories. Advances in Condensed Matter Physics 2012. 926290. (4) Muneeswaran, M.; Dhanalakshmi, R.; Giridharan, N. V. Structural, vibrational, electrical and magnetic properties of Bi1−xPrxFeO3. Ceramics International 2015, 41, (7), 8511- 8519. (5) Wang, Y. P.; Zhou, L.; Zhang, M. F.; Chen, X. Y.; Liu, J. M.; Liu, Z. G. Roomtemperature saturated ferroelectric polarization in BiFeO3 ceramics synthesized by rapid liquid phase sintering. Applied Physics Letters 2004, 84, (10), 1731-1733. (6) Das, S. R.; Choudhary, R. N. P.; Bhattacharya, P.; Katiyar, R. S.; Dutta, P.; Manivannan, A.; Seehra, M. S. Structural and multiferroic properties of La-modified BiFeO3 ceramics. Journal of Applied Physics 2007, 101, (3), 034104. (7) Yuan, G. L.; Or, S. W.; Wang, Y. P.; Liu, Z. G.; Liu, J. M. Preparation and multiproperties of insulated single-phase BiFeO 3 ceramics. Solid State Communications 2006, 138, (2), 76-81.