ISSN 10637850, Technical Physics Letters, 2011, Vol. 37, No. 10, pp. 952–955. © Pleiades Publishing, Ltd., 2011. Original Russian Text © E.N. Ubushaeva, E.V. Likhushina, K.G. Abdulvakhidov, M.A. Vitchenko, B.K. Abdulvakhidov, V.B. Shirokov, N.V. Lyanguzov, Yu.I. Yuzyuk, E.M. Kaidashev, I.V. Mardasova, 2011, published in Pis’ma v Zhurnal Tekhnicheskoi Fiziki, 2011, Vol. 37, No. 20, pp. 23–31.
Effect of Mechanical Activation on Physical Properties of Relaxor Ferroelectric Pb2ScNbO6 Ceramics E. N. Ubushaeva*, E. V. Likhushina, K. G. Abdulvakhidov, M. A. Vitchenko, B. K. Abdulvakhidov, V. B. Shirokov, N. V. Lyanguzov, Yu. I. Yuzyuk, E. M. Kaidashev, and I. V. Mardasova Southern State University, RostovonDon, 344090 Russia Moscow Aviation Institute (State University of Aerospace Technology), Moscow, 125080 Russia RostovonDon State Technical University, RostovonDon, 344010 Russia *email:
[email protected] Revised manuscript received June 2, 2011
Abstract—Relaxor ferroelectric lead scandium niobate Pb2ScNbO6 (PSN) ceramics was obtained by solid state synthesis and a modified ceramic technology, whereby the sintering stage was preceded by roomtem perature compression and shear straining of the synthesized PSN powder in Bridgman anvils. It is established that this mechanical activation leads to the development of dynamic recrystallization processes in PSN grains, which significantly influence the physical properties of the final ferroelectric ceramic material. DOI: 10.1134/S1063785011100282
The process of mechanical activation of a solid consists in relaxation of the supplied mechanical energy via different channels, in particular, plastic deformation and fracture as a result of the accumula tion of stressinduced point defects and dislocations [1–3]. The related processing method provides a pow erful tool that allows material particles to be obtained with linear dimensions ranging from several dozen microns to several nanometers and the physical prop erties of these particles to be controlled during the treatment [4]. As is known, the dependence of dielectric permit tivity ε, spontaneous polarization Ps, and a number of other physical properties of ferroelectric ceramics on the dimensions of grains is already manifested when the average grain size decreases below 1 μm. It is com monly accepted that crystalline grains with dimen sions below 10 nm do not contain dislocations, whereas greater crystals almost always contain both point defects and dislocations. Therefore, by inten tionally changing the dimensions of grains, concen tration of defects, and their types by mechanical load ing of the initial material, it is possible to influence the properties of ceramics. This Letter presents the results of a complex investigation of the effect of mechanical activation on the physical properties of lead scandium niobate Pb2ScNbO6 (PSN) ceramics, which is a promising relaxor ferroelectric material. The initial stage of sample preparation was solid state synthesis of PSN from the corresponding oxides (PbO, Sc2O3, Nb2O5) of analytical grade taken in a stoichiometric ratio. Synthesis was carried out in a
hermetically closed platinum crucible for 4 h at a tem perature of 1273 K. Then, weighted amounts of the synthesized PSN powder were processed by compres sion in a mechanical press with a load of up 3 tons and shear straining in Bridgman anvils, the bottom of which rotated at ω = 0.3 rad/min. Seven aliquots of PSN powder compacted at the same pressure and shear strain were triturated for 10 min in an agate mor tar with ethyl alcohol. Thus, eight batches of PSN powder were prepared, which were mechanically acti vated at various pressures P ranging from 80 to 360 MPa. For comparison of physical properties, one aliquot in every batch of the same initial composition was not mechanically activated and served as an untreated control. The structures of samples were studied by Xray diffraction (XRD) on a DRON3 diffractometer using CuKα radiation and by electron microscopy on a Supra25 instrument. The integral intensities of dif fraction reflections were measured for two reflections (111 and 220) by point scanning at an angular step of Δν = 0.01° and an exposure time of τ = 8 s at each point. The XRD measurements were performed at two temperatures, 473 and 573 K, because polar clusters could probably be retained in the nonpolar PSN matrix at lower temperatures. On the other hand, these temperatures are yet insufficient for the onset of the order–disorder transition (To–d ≈ 1480 K). Note also that the temperature Tm of attaining the maximum dielectric permittivity in disordered PSN is about 370 K. The sample temperature during these measure ments was controlled to within ±1 K by a VRT3 device.
952
EFFECT OF MECHANICAL ACTIVATION
953 Δa/a × 10–3 3.6
D, E 600
3.4
500
3.2 400 3.0 300 2.8 200 2 μm
(a)
0
2.6 40
80 120 160 200 240 280 320 360 P, MPa
Fig. 2. Plots of CSD size D and microdeformation Δa/a versus pressure P applied during mechanical activation of PSN powders.
1 μm
(b)
Fig. 1. Microphotograps of (a) initial and (b) mechanically activated (P = 200 MPa) PSN powders.
The samples for electrical measurements were pre pared by pressing into disks with a thickness of 1 mm and a diameter of 10 mm. The ceramic samples of ini tial PSN and powders mechanically activated at vari ous pressures P were obtained by sintering under iden tical temperature–time regime in a closed platinum crucible with PbZrO3. The electrodes were formed by applying aquadag. The permittivity was measured by E720 immittance meter at a frequency of 1 kHz in a temperature interval from room temperature to 450°C. Figure 1 shows typical microphotograps of the ini tial and mechanically activated PSN powders. A com parative analysis of the dimensions and shapes of par ticles in the powder before and after mechanical acti vation showed that all particles in the latter case possessed irregular shapes and their dimensions were spread within 30–800 nm depending on the applied pressure. Investigation in an atomic force microscope (Ntegra, NT MDT) showed that the particle size dis TECHNICAL PHYSICS LETTERS
Vol. 37
No. 10
tribution in some powders exhibited a multimodal character. A comparison of the XRD and electronmicro scopic data allowed factors responsible for the observed particle size distribution to be elucidated. As is known, intense loading by combination of shear straining with certain pressures leads to grain refine ment and a decrease in the activation energy of diffu sion processes, so that the effective diffusion coeffi cient of generated point defects and their concentra tion increase by several orders of magnitude [5]. These factors provide favorable conditions for the socalled lowtemperature dynamic recrystallization [6]. The recrystallized grains contain a lower number of struc tural defects. As a result, the XRD patterns of PSN powders mechanically activated at some pressures exhibited growth in both the integral intensity I and coherent scattering domain (CSD) size D. However, the mechanical activation for a longer period of time or at higher pressures for the same time can lead to the refinement of recrystallized grains, which would result in a significant change of the physical properties and thermodynamic parameters of PSN. Previously, we have reported [4] that the mechanical activation can be used to control even the degree of longrange order ing in relaxor ferroelectric PbSc0.5Ta0.5O3, which is analogous to PSN. Thus, we can also expect that the mechanical activation would also change the electrical properties of PSN ceramics sintered from processed powders. Let us consider the behavior of some parameters that characterize the structural state of mechanically activated PSN powder and ceramics. Figure 2 shows the dependences of the CSD size D and microdefor mation Δa/a on pressure P applied during the mechanical activation of PSN powders. The process 2011
954
UBUSHAEVA et al.
Dependence of the characteristic Debye temperature Θ, temperaturedependent Debye–Waller factors B, and mean square displacements 〈U2〉 for PSN powders mechanically activated at various pressures 2
P, MPa 0 80 120 160 200 240 280 320 360
Θ, K 112 172 92 208 132 187 118 92 97
B473, Å2 1.28 0.56 1.97 0.39 0.96 0.48 1.20 1.97 1.77
B573, Å2 1.61 0.69 2.39 0.47 1.16 0.58 1.45 2.39 2.15
ing of PSN powders at pressures increasing to 200 MPa is accompanied by a growth in the concen tration of dislocations and, hence, the degree of grain mosaicity and the integral diffraction intensity I. Starting with this pressure, the concentration of point defects generated in the course of mechanical activa tion and their role in diffusion processes begin to increase. This interval of pressures also shows a decrease in the CSD size D, in correlation with the results of electronmicroscopic measurements. The increase in D at 240 MPa (Fig. 2) corresponds to the aforementioned dynamic recrystallization process, whereas the subsequent decrease in D with increasing pressure is due to refinement of recrystallized grains and a growth in the concentration of point defects. We have calculated the characteristic Debye tem perature Θ, Debye–Waller thermal factors B, and mean square displacements 〈U2〉 for the control and mechanically activated samples in the framework of the Debye approximation, assuming that the oscilla tions of all ions in PSN paraphase (Pm3m) are inde pendent and isotropic. This assumption is permissible [7] for the choice of diffraction profiles with neglect of the diffusion background that arises between the Bragg peaks due to thermal displacement of ions. Accord ingly, the results obtained in this study for the parame ters of lattice dynamics possess somewhat rough esti mation character. In order to take into account the influence of tex ture, each sample was characterized by the Debye temperature Θ that was determined by the classical twotemperature method [8] using the following for mula for the ratio of intensities I of the XRD profiles measured at T1 = 473 K and T2 = 573 K: 2 IT 12h F ( x 1 ) F ( x 2 ) 2 ln ⎛ 1⎞ = 2 – S , (1) ⎝ IT ⎠ x1 x2 m a kΘ 2 where x1 = Θ/T1, x2 = Θ/T2, h is the Planck constant, k is the Boltzmann constant, F(x) is the Debye tem
〈 U d〉 , Å
2
〈 Us 〉 , Å 1.050 1.040 1.038 1.041 1.052 1.057 1.037 1.040 1.030
473 K
573 K
0.062 0.057 0.060 0.057 0.056 0.068 0.061 0.061 0.060
0.136 0.080 0.139 0.105 0.058 0.153 0.152 0.154 0.134
perature, ma = Σnnmi/Σmi is the mean atomic mass, ni is the number of atoms in the PSN unit cell, and S = sinν/λ. Using these Θ values, the isotropic Debye– Waller factors B for all samples at the two temperatures indicated above were calculated as follows: 2
6h T F ( x ) + x . B ( T ) = 2 4 m a kΘ
(2)
The obtained values of Θ and B(T) are presented in the table. The values of total mean square displacements 〈U2〉 were calculated according to [9] with allowance for the fact that PSN occurs in a disordered state. As is 2 known, 〈U2〉 represents a sum of the static (〈 U s 〉) and 2
dynamic (〈 U d 〉) mean square displacements of atoms from their ideal crystallographic positions: 2
2
2
〈 U 〉 = 〈 U s 〉 + 〈 U d〉 .
(3)
The lefthand side can also be determined using the following simplified expression: 2
2 ( ν 2 )P 1 f ( ν 1 ) F 1 2 3λ ln I 〈 U 〉 = , (4) 2 2 2 2 16π ( sin ν 1 – sin ν 2 ) I ( ν 1 )P 2 f ( ν 2 ) F 2
where f(ν1) and f(ν2) are the polarization factors, P1 and P2 are the corresponding repeatability factors, F1 and F2 are the structural factors, and I(ν1) and I(ν2) are the integral intensities. In order to separate the dynamic and static displacements, the temperature dependences of total displacements were plotted as 2 h 〈U2〉 – and extrapolated to T = 0 K. Then, the m a kT intercept with the abscissa axis at T = 0 K determines the static displacement, while the difference 〈U2〉 – 2 〈 U s 〉 gives the dynamic component. The results of these calculations are also given in the table.
TECHNICAL PHYSICS LETTERS
Vol. 37
No. 10
2011
EFFECT OF MECHANICAL ACTIVATION
As can be seen from the data in the table, the vari ation of Θ with the pressure of mechanical activation does not obey any clear law, which also implies a spread of the maΘ values that are proportional to the forces returning atoms to their equilibrium positions during thermal oscillations. The isotropic thermal fac tor B(T) increases with the temperature (see the third and fourth columns in the table). Figure 3 shows plots of the mean square displace ments 〈U2〉 at two temperatures versus mechanical activation pressure P. As can be seen, the two 〈U2〉 curves vary almost in phase up to P = 280 MPa. A sig nificant difference in their behavior is only observed at P = 320 Ma. Note that the 〈U2〉 values determined at T = 473 and 573 K for the sample treated at P = 200 MPa almost coincide. This behavior can be explained by more intense development of the dynamic recrystallization process due to local heating of the sample at higher pressures. Analogous coinci dence of 〈U2〉 values (intersection of the 〈U2〉 curves) takes place in the pressure intervals of 280–320 and 320–360 MPa. 2 〈 Ud 〉
Comparative analysis of the values deter mined at T = 473 and 573 K (see table) shows that this increase in the temperature leads to a more than two fold growth in the dynamic mean square displace ment, although the general character of the depen 2 2 dences of 〈 U d 〉 and 〈 U s 〉 on the pressure is essentially the same. The only exception is presented by the sam 2 ple treated at P = 200 MPa, for which the 〈 U d 〉 values determined at T = 473 and 573 K approximately coin cide. Now let us consider the behavior of the maximum dielectric permittivity (εm) for the ceramics sintered from PSN powders mechanically activated at various pressures. As can be seen from Fig. 3, the εm value exhibits nonmonotonic variation with increasing pres sure of mechanical activation. Of course, it would be not quite correct to seek a correlation between the total mean square displacements in powders upon the dynamic recrystallization process and the εm values in ceramic samples (where the sintering is also accompa nied by thermal recrystallization processes). Never theless, despite the fact that ceramics were prepared from different batches of powders mechanically acti vated at the same temperature, a decrease in the 〈U 2〉 value of powders corresponds to a growth in the εm of ceramics. In fact, there is no ground to believe that the mechanical activation of powders used to prepare the ceramics differs from that in the nonsintered pow ders. Thus, PSN powder mechanically activated at vari ous pressures occur in different metastable structural
TECHNICAL PHYSICS LETTERS
Vol. 37
No. 10
〈U2〉, Å 1.07
955 εm 8500
473 K 573 K
1.06
8000 7500
1.05
7000 1.04 6500 1.03
6000
1.02 1.01 0
5500 40
5000 80 120 160 200 240 280 320 360 P, MPa
Fig. 3. Plots of total mean square displacements 〈U2〉 in PSN powders and maximum dielectric permittivity εm of the corresponding ceramics versus pressure P of mechani cal activation.
states. Therefore, even for the same temperature–time regime, the initial conditions of sintering can differ, which allows the physical properties of obtained ceramics to be controlled in a broad range of parame ters. REFERENCES 1. V. V. Boldyrev, Soros. Obrazov. Zh., No. 5, 49 (1996). 2. R. Z. Valiev and I. V. Aleksandrov, Nanostructured Materials Fabricated by Severe Plastic Deformation (Moscow, 2000) [in Russian]. 3. N. F. Uvarov and V. V. Boldyrev, Usp. Khim. 70, 307 (2001) [Russ. Chem. Rev. 70, 265 (2001)]. 4. K. G. Abdulvakhidov, M. A. Vitchenko, I. V. Marda sova, and E. N. Oshaeva, Zh. Tekh. Fiz. 78 (5), 131 (2008) [Tech. Phys. 53, 661 (2008)]. 5. E. N. Ubushaeva, K. G. Abdulvakhidov, I. V. Mar dasova, B. K. Abdulvakhidov, M. A. Vitchenko, A. A. Amirov, A. B. Batdalov, and A. G. Gamzatov, Zh. Tekh. Fiz. 80 (11), 40 (2010) [Tech. Phys. 55, 1596 (2010)]. 6. M. A. Glezer and L. S. Metlov, Fiz. Tverd. Tela (St. Petersburg) 51, 52 (2009) [Phys. Solid State 51, 5 (2009)]. 7. Ya. S. Umanskii, Roentgenography of Metalls (Metal lurgiya, Moscow, 1967) [in Russian]. 8. R. James, The Optical Principles of the Diffraction of Xrays (London, 1950). 9. S. S. Gorelik, L. N. Rastorguev, and Yu. A. Skakov, Roentgenographic and ElectronOptical Analysis (Metal lurgiya, Moscow, 1970) [in Russian].
Translated by P. Pozdeev
2011