High-temperature dielectric properties of TbFeO3 ceramics

Available online at www.sciencedirect.com

CERAMICS INTERNATIONAL

Ceramics International 42 (2016) 657–660 www.elsevier.com/locate/ceramint

High-temperature dielectric properties of TbFeO3 ceramics Da Zhang, Qiuju Lin, Chunchang Wang, Nan Zhang, Haibo Li Laboratory of Dielectric Functional Materials, School of Physics and Material Science, Anhui University, Hefei 230039, China Received 22 July 2015; received in revised form 22 July 2015; accepted 28 August 2015 Available online 10 September 2015

Abstract TbFeO3 ceramics were prepared via conventional solid-state reaction route. By means of the complex dielectric permittivity, electric modulus, and impedance analysis, the high-temperature dielectric properties of TbFeO3 were investigated in the temperature range from 300 to 1073 K and frequency range of 102–107 Hz. Rich dielectric phenomena, including three dielectric responses (R1–R3), a relaxor-like dielectric anomaly and negative capacitance were found in the order of ascending temperature. The low-temperature relaxation (R1) near room temperature is caused by the surface-layer effect. The high-temperature relaxations R2 and R3 were argued to be related to the dipolar and Maxwell–Wagner relaxation, The anomaly is confirmed to be caused by the negative capacitance effect. & 2015 Elsevier Ltd and Techna Group S.r.l. All rights reserved.

Keywords: Ceramics; Dielectric anomaly; Maxwell–Wagner relaxation; Negative capacitance

1. Introduction In the trend towards miniaturization, colossal dielectric constant (CDC) materials are going to play an important role. As a typical representative of CDC materials, CaCu3Ti4O12 (CCTO) had been widely investigated during the past 15 years [1–5]. Apart form the stunning CDC behavior in the temperature range below room temperature (RT), CCTO also shows a relaxor-like behavior in the temperature range above RT. Numerous experimental work indicated that the high-temperature relaxor-like dielectric anomaly is unrelated to ferroelectric transition and is almost a common dielectric phenomenon for oxides [6–10]. This anomaly is now called pseudo-relaxor behavior [11]. Li and coauthors attributed the behavior to be an artificial phenomenon resulting from negative capacitance as the investigated sample is highly leaky at high temperatures [12]. Our recent work pointed out that the anomaly is composed of two close relaxations with the low-temperature one being the dipolar relaxation and the high-temperature one being the MW relaxation. Both relaxations are related to the hopping motion of oxygen vacancies [6,7]. n

Corresponding author. Tel.: þ86 551 63861902; fax: þ 86 551 65846849. E-mail address: [email protected] (Q. Li).

http://dx.doi.org/10.1016/j.ceramint.2015.08.161 0272-8842/& 2015 Elsevier Ltd and Techna Group S.r.l. All rights reserved.

Recently, investigation on the low-temperature (100–300 K) dielectric properties of TbFeO3 (TFO) revealed that the sample shows CDC behavior similar to that found in CCTO [13]. It is expected that TFO might exhibit relaxor-like behavior at high temperatures. In the present work, the dielectric properties of TFO in the temperature range from 300 to 1073 K were investigated. Three dielectric relaxations and an dielectric anomaly along with negative capacitance effect were observed. The mechanisms of these relaxation and anomaly were discussed.

2. Experimental TFO ceramic samples were prepared by the conventional solid state reaction method using high purity (4 N grade) Tb4O7 and Fe2O3 powders as described in our previous paper [13]. Dielectric properties were measured using a Wayne Kerr 6500B precise impedance analyzer. Annealing treatments were performed in flowing (200 ml/min) oxygen and nitrogen (both with purity 4 99.999%). Electrodes were made by printing silver paste on both sides of the disk-type samples and then fired at 1073 K for 2 h in order to remove the polymeric component.

658

D. Zhang et al. / Ceramics International 42 (2016) 657–660

3. Results and discussion Fig. 1(a) and (b) shows, respectively, the typical temperature dependence of dielectric constant ε′(T ) and dielectric loss tangent, tan δ (T )( tan δ = ε″ / ε′, where ε″ is the imaginary part of the complex permittivity ε*) for TbFeO3. From Fig. 1(a), we can see that ε′(T ) exhibits an obvious stepwise increase around 500 K and a pronounced relaxor-like dielectric anomaly around 800 K followed by negative capacitance effect. In the tan δ (T ) curves [see Fig. 1(b)], a set of thermally activated relaxation peaks, humps, and sharp peaks correspond to the stepwise increase, anomaly, and negative capacitance, respectively, were observed. When the temperature rises higher than 500 K, the tan δ (T ) curve shows lager value and increases with increasing temperature causing remarkable background. This implies that the conductivity is notable in this temperature range. In order to get more information of the dielectric phenomena, we applied the dielectric function of electric modulus M*, which is defined as M * = 1/ε* and considered as “good dielectric function” that can provide information about relaxation in the absence of a well-defined peak in tan δ (T )[14,15]. Fig. 2 shows the temperature dependence of the imaginary part of electric modulus for TFO at various frequencies. We 300 Hz 800 Hz 10 kHz 500 kHz

500 Hz 2 kHz 100 kHz

ε

10

10

10 400

500

600

700

800

900

1000

T (K) Fig. 1. Temperature dependence of ε′ (a) and tan δ (b) of TFO measured at various frequencies.

R3 R2

(1 )

where f0 is the preexponential factor, Ea is the activation energy for relaxation, kB is the Boltzmann constant. The values of Ea and f0 were calculated to be 0.59 eV and 6.69  1013 Hz for R1, 1.08 eV and 2.81  1014 Hz for R2, and 1.19 eV and 1.1  1014 Hz for R3. We note the relaxation parameters ( Ea ¼ 0.59 eV and f0 ¼ 6.69  1013 Hz) of R1 are the same as those ( Ea ¼ 0.59 eV and f0 ¼ 6.78  1013 Hz) of the relaxation around room temperature in our previous report [13]. This relaxation was clarified to be a Maxwell–Wagner relaxation caused by the surface-layer effect and will not be further discussed in the following part. We now turn our attention to R2 and R3. It is well known that oxygen vacancies as native point defects are unavoidable in perovskite oxides. The activation energy around 1.0 eV is typical value for relaxation related to the hopping motion of oxygen vacancies in various perovskites [16–20]. To identify the relationship between oxygen vacancies and the observed relaxations, the sample of TFO was treated by the following processes: (1) annealing in O2 at 1073 K for 2 h to decrease the oxygen vacancies content; (2) annealing in N2 at 1073 K for 2 h to increase the oxygen vacancies content. Fig. 4 displays the comparison of the temperature dependence of electric modulus recorded at 10 kHz before and after being annealed in

R1 1

400

T (K)

480

Experimental data Fitting peak R2 Fitting peak R3

500 Hz

0.1

400

R1 R2 R3

5

560

Μ (·10−5)

Μ (×10−5)

10

f = f0 exp (Ea / k B Tp )

500

300 Hz 500 Hz 800 Hz 2 kHz 10 kH

logƒ (Hz)

tanδ

10

can clearly see that the M ″(T ) curves show two distinct sets of relaxation peaks and a set of humps between them. This finding indicates that there are three relaxations in TFO in the tested temperature range. For simplicity, these relaxations were designed as R1, R2, and R3 in the order of ascending temperature. To calculated the relaxation parameters, the accurate peak position is need. Since the position of the hump is strongly affected by the peak of R3, two Debye peaks were applied to fit the hump and the high-temperature peak in order to deduce the accurate positions of R2 and R3. As an illustration, the inset of Fig. 2 displays the fitting result to the experimental data recorded at 500 Hz. Based on the fitting result, the peak positions of R2 and R3 can be obtained. Fig. 3 presents the measuring frequency f versus the reciprocal of the peak temperature 1000/ Tp . The obtained data for the three relaxations fall on straight lines in the halflogarithmic scale indicating that these relaxations follow the Arrhenius law:

4 3

600

T (K) Fig. 2. Temperature dependence of M ″(T ) measured at various frequencies. The inset shows the fitting result of R2 and R3 peaks recorded at 500 Hz.

1.5

2.0

2.5

1000/Tp(K-1) Fig. 3. The Arrhenius plots of the three relaxations.

D. Zhang et al. / Ceramics International 42 (2016) 657–660

O2 and N2. We can clearly see that both R2 and R3 can be affected by the annealing treatments indicating that they are closely related to the oxygen vacancies. In order to get a clear picture about how the peak intensity varies with the annealing treatments, nonlinear fittings were applied. Fig. 4(a) and (b) are the resulting peaks of R2 and R3, respectively. As expected, both R2 and R3 can be depressed by O2-annealing treatment and then increased by N2-annealing treatment. This finding further supports that R2 and R3 are related to oxygen vacancies. The hopping motions of oxygen vacancies between spatially fluctuating lattice potentials not only produce conductivity but also give rise to dipolar effects. The frequency dependent conductivity can be described by Jonscher's power law, i.e., [21,22] σ = σdc + a 0 f n

(2 )

where σdc is the direct conductivity, a0 and n (0 r nr 1) are the temperature-dependent adjusting constants. To clarify this inference, Fig. 5 shows the variation of ac conductivity with frequency at several temperatures in the temperature higher than 500 K. It is clearly seen that the ac conductivity shows a flat in the low frequency range followed by a rapid increase in the high frequency. This is the typical feature of Jonscher's power law. From Fig. 5, the dc conductivity σdc can be read out directly and was plotted as a function of the reciprocal of temperature as shown in the insert of Fig. 5. It follows the R2

10 kHz

R3

1.6

1

0.8

R1 0.0

R3 as-prepared O annealed

0.1

N annealed 400

500

600

3

Μ" (×10-5)

Μ"(×10-5)

R2

Μ" (×10-5)

10

0

700

500

600

T (K)

700

T (K)

Fig. 4. (a) Temperature dependence of M″(T) at 10 kHz for TFO sample before (as-prepared) and after O2 and N2 annealing treatments. (b) and (c) are the resultant M″ peaks for R2 and R3, respectively, in the as-prepared, O2-, and N2-annealed cases.

659

Arrhenius law: σdc = σ0 exp (

Edc / kB T )

(3 )

where Edc is the activation energy of conductivity, and σ0 is the pre-exponential factor. Linear fitting shown as a straight line in the inset yields Edc ¼ 1.10 eV. The dc conductivity value of activation energy is close to that for the relaxation of R2. This finding convincingly indicates that the mechanism for the R2 is related to the conductivity relaxation associated with the hopping motions of oxygen vacancies. The hopping motions of oxygen vacancies inside grains yield not only a dipolar-type relaxation, but also a Maxwell– Wagner (MW) relaxation when the carriers were blocked by the interfaces of grain boundaries. Therefore, R3 might be a MW relaxation. This inference is confirmed by impedance analysis. The inset (a) of Fig. 6 is the complex impedance spectrum (Z″ vs Z′ where Z′ and Z″ are the real and imaginary parts of the complex impedance Z*) at 953 K. A highfrequency semicircle and a low-frequency arc can be seen. This indicates that two relaxations are active at this measuring temperature. In terms of the result displayed in Fig. 1(b), R1 is accomplished at this high temperature. Thus, the two relaxations can be identified to be R2 and R3. It was found that the high-frequency semicircle terminates at the origin point. This fact indicates that the high-frequency semicircle corresponds to the bulk response. Consequently, the low-frequency arc is related to the grain boundary response. This result convinces that R2 and R3 stem form the grain and grain boundary responses, respectively. Since R2 and R3 appear in the temperature range very close to the dielectric anomaly, this seemly implies that the tworelaxation mechanism might account for the observed dielectric anomaly. However, as the negative capacitance effect can also give rise to the anomaly. This leads to the question that which model dominates the observed relaxor-like behavior? If the two-relaxation mechanism was true, the dielectric mismatch defined as the difference of the relaxation time between the grain and grain boundary (Δτ ¼ τgb  τg) should be less than 3–4 orders of magnitudes [6]. To clarify this point, the complex impedance spectra at different temperatures were modeled by two RC(R ¼ resistor, C ¼ capacitor) units [inset

10

1000/T (K )

10

1.4

10 10

503K 633K 753K

543K 673K

10

593K 713K

logσ (Ω m )

σ (Ω-1m-1)

10

1.6

1.8

2.0

1 0 -1 -2

10

Ea=1.10eV 10

ƒ (Hz) Fig. 5. Frequency dependence of ac conductivity for TFO sample at various temperatures. The inset shows the Arrhenius plot for the dc conductivity.

Fig. 6. Temperature dependence of the relaxation time of the grain and grain boundary. The inset (a) shows the complex impedance plot for TFO at 953 K and (b) is the equivalent circuit.

660

D. Zhang et al. / Ceramics International 42 (2016) 657–660

(b) of Fig. 6] connected in series with one for the grain and the other for the grain boundary [23]. We can easily deduce the relaxation time for grain and grain boundary based on τi = Ri Ci (i¼ g and gb, representing the grain and grain boundary, respectively), The values of R and C are fitting by using Z-view. The dashed curves in the main panel of Fig. 6 are fitting results. It is seen that the relaxation times of both grain and grain boundary exponentially decrease with increasing temperature following the Arrhenius law. We found that the relaxation time of the grain and grain boundary tends to intersect. But their difference in the whole temperature range is larger than 3–4 orders of magnitudes. For example, the difference of τg and τgb at 953 K is 5 orders of magnitude, which is in the Debye-like region (see Ref. [6]). This finding excludes the possibility that the two-relaxation model underlies the observed relaxor-like behavior and confirms that the negative capacitance effect is the cause of the anomaly. When the temperature is high enough, the long-range transportation of charge carriers becomes significant. This makes the sample highly leaky and the inductive effect of the measuring system becomes evident [12]. This will contribute to negative capacitance effect giving rise to a dielectric anomaly. Therefore, we can conclude that the high-temperature anomaly is an artificial effect caused by negative capacitance.

4. Conclusions In summary, high-temperature dielectric properties of TbFeO3 ceramics were systematically investigated in the temperature range of 320–1073 K and the frequency range of 100 Hz to 10 MHz. Three relaxations and an anomaly were observed. The low-temperature relaxation (R1) near room temperature is due to surface-layer effect. The high-temperature relaxations R2 and R3 are related to the dipolar and Maxwell–Wagner relaxation, respectively. Both relaxations are induced by the hopping motions of oxygen vacancies. The high-temperature anomaly was found to be an artificial effect caused by negative capacitance.

Acknowledgments The authors thank financial support from National Natural Science Foundation of China (Grant no. 11404002) and Cooperative Innovation Research Center for Weak Signal-Detecting Materials and Devices Integration of Anhui University (Grant no. Y01008411). This work was supported in part by the China Postdoctoral Science Foundation (Grant no. 2014M561805) and Anhui Province Postdoctoral Science Foundation (Grant no. 2014B007).

References [1] P. Lunkenheimer, S. Krohns, S. Riegg, S.G. Ebbinghaus, A. Reller, A. Loidl, Colossal dielectric constants in transition-metal oxides, J. Eur. Phys 180 (2010) 61–89. [2] C.C. Wang, L.W. Zhang, Polaron relaxation related to localized charge carriers in CaCu3Ti4O12, Appl. Phys. Lett. 90 (2007) 142905. [3] M.A. Subramanian, D. Li, N. Duan, B.A. Reisner, A.W. Sleight, High dielectric constant in ACu3Ti4O12 and ACu3Ti3FeO12 phases, J. Solid State Chem. 151 (2000) 323–325. [4] A.P. Ramirez, M.A. Subramanian, M. Gardel, G. Blumberg, D. Li, T. Vogt, S.M. Shapiro, Giant dielectric constant response in a coppertitanate, Solid State Commun. 115 (2000) 217–220. [5] C.C. Homes, T. Vogt, S.M. Shapiro, S. Wakimoto, A.P. Ramirez, Optical response of high-dielectric-constant perovskite-related oxide, Science 293 (2001) 673–676. [6] C.C. Wang, M.N. Zhang, K.B. Xu, G.J. Wang, Origin of hightemperature relaxor-like behavior in CaCu3TiO4O12, J. Appl. Phys. 112 (2012) 034109. [7] C.C. Wang, M.N. Zhang, W. Xia, High-temperature dielectric relaxation in Pb(Mg1/3Nb2/3)O3–PbTiO3 single crystals, J. Am. Ceram. Soc. 96 (2013) 1521–1525. [8] J.L. Ye, C.C. Wang, W. Ni, X.H. Sun, Dielectric properties of ErFeO3 ceramics over a broad temperature range, J. Alloy. Compd. 617 (2014) 850–854. [9] C.C. Wang, Y.M. Cui, G.L. Xie, C.P. Chen, L.W. Zhang, Phase separation in La2Cu4 þ y ceramics probed by dielectric measurements, Phys. Rev. B 72 (2005) 064513. [10] J.L. Cohn, M. Peterca, J.J. Neumeier, Low-temperature permittivity of insulating perovskite manganites, Phys. Rev. B 70 (2004) 214433. [11] C.C. Wang, S.X. Dou, Pseudo-relaxor behaviour induced Maxwell– Wagner relaxation, Solid State Commun. 149 (2009) 2017–2020. [12] M. Li, D.C. Sinclair, A.R. West, Extrinsic origins of the apparent relaxorlike behavior in CaCu3Ti4O12 ceramics at high temperatures: a cautionary tale, J. Appl. Phys. 109 (2011) 084106. [13] W. Xia, C.C. Wang, P. Liu, J.L. Ye, W. Ni, Colossal dielectric behavior in TbFeO3 ceramics, Curr. Appl. Phys. 13 (2013) 1743–1745. [14] V. Gupta, A. Mansingh, Hopping conduction in insulating rf-sputtered zinc oxide films, Phys. Rev. B 49 (1994) 1989–1995. [15] C.T. Moynihan, L.P. Boesch, N.L. Laberage, Decay function for electricfield relaxation in-vitreous ionic conductors, Phys. Chem. Glasses 14 (1973) 122–125. [16] S.A. Prosandeyev, N.M. Teslenko, A.V. Fisenko, Breaking of symmetry of one-electron orbitalsat oxygen vacancies in perovskite-type oxides, J. Phys. Condens. Matter 5 (1993) 9327–9344. [17] B.S. Kang, S.K. Choi, C.H. Park, Diffuse dielectric anomaly in perovskite-type ferroelectric oxides in the temperature range of 400– 700 1C, J. Appl. Phys. 94 (2003) 1904–1911. [18] P. Moretti, F.M. Michel-Calendini, Impurity energy levels and stability of Cr and Mn ions in cubic BaTiO3, Phys. Rev. B 36 (1987) 3522–3527. [19] M.O. Selmer, P. Pecheur, The electronic structure of transition-metal impurities in SrTiO3, J. Phys. C Solid State Phys. 21 (1988) 1779–1790. [20] K.A. Muller, Paramagnetic point and pair defects in oxide perovskites, J. Phys. 42 (1981) 551–557. [21] S.R. Elliott, Ac conduction in amorphous chalcogenide and pnictide semiconductors, Adv. Phys. 36 (1987) 135–217. [22] A.K. Jonscher, Dielectric Relaxation in Solids, Chelsea Dielectrics, London, 1983. [23] D.C. Sinclair, A.R. West, Impedance and modulus spectroscopy of semiconducting BaTiO3 showing positive temperature coefficient of resistance, J. Appl. Phys. 66 (1989) 3850–3856.