Dielectric and micromechanical<b style=""> </b>studies of barium titanate substituted (1-y)Pb (Zn_1/3Nb_2/3)O₃-yPT ferroelectric ceramics

Dielectric and micromechanical studies of barium titanate substituted (1-y)Pb (Zn

Nb

)O

-yPT ferroelectric ceramics

A K Himanshu^1*, S K Bandyopadhyay¹, Pintu Sen¹, D C Gupta², Riya Chakraborty³, Anoop K Mukhopadhyay³, B K Choudary⁴ & T P Sinha⁵

1Advanced Materials Laboratory, VECC, 1/AF, Bidhan Nagar, Kolkata 700 064, India

2School of Studies in Studies in Physics, Jiwaji University, Gwalior 474 011, India

3Central Glass & Ceramic Research Institute, Jadavpur, Kolkata 700 032, India

4University Department of Physics, Ranchi University, Jharkhand 834 001, India

5Department of Physics, Bose Institute, 93/1, APC Road, Kolkata 700 009, India

*E-mail: himanshu_ak@yahoo.co.in, akh@veccal.ernet.in Received 20 October; revised 15 January 2010; accepted 4 March 2010

Dielectric studies have been carried out on barium titanate substituted lead zinc niobate ferroelectrics ceramics, (Pb_0.88Ba_0.12)[(Zn_1/3Nb_2/3)_0.88Ti_0.12)]O₃ and (Pb_0.8Ba_0.2)[(Zn_1/3Nb_2/3)_0.8Ti_0.2)]O₃. The dielectric dispersion of the solid solutions has been studied as a function of temperature in the frequency range 100 Hz-1MHz. The temperature variation of the real components of dielectric permittivity (ε′) at different frequencies shows a diffuse phase transition. Micromechanical studies have been carried on (Pb0.88Ba0.12)[(Zn1/3Nb2/3)0.88Ti0.12)]O3 and (Pb0.8Ba0.2)[(Zn1/3Nb2/3)0.8Ti0.2)]O3 by nanoindentation at different loads. Hardness, Young’s modulus, elastic energy and total energy have been extracted. It is seen that in case of (Pb0.8Ba0.2)[(Zn1/3Nb2/3)0.8Ti0.2)]O3, the plastic deformation energy is lower than the other sample, which is due to extra Ti content in B-site. This is also reflected in polarization studies, which confirms that the presence of Ti⁴⁺ in B-site brings forth rigidity in Ti-O bond.

Keywords: Polarization, Ferroelectric, Ceramics, Nanoindentation

1 Introduction

Lead based relaxor ferroelectric compositions are used in a variety of applications¹. High dielectric constant, large electric-field-induced electrosrtictive strains, and good optical transparencies of these materials are utilized in devices such as high-K capacitors, actuators and optical shutters. Among the relaxor ferroelectric materials, lead zinc niobate (Pb(Zn_1/3Nb_2/3O₃-PZN) is one of the few compositions with a high dielectric transition temperature of about 140°C at 1 kHz measuring frequency. The rhombohedral PZN forms a complete solid solution with tetragonal lead titanate (PbTiO₃-PT) and the morphotropic phase boundary² (MPB) falls between 7 and 12% of PT. The ceramic preparation of these compositions remains a challenge as it always results in the formation of non-ferroelectric phases with a pyrochlore structure. In the past 40 years, several additives have been successfully used to stabilize the perovskite phase are discussed in our paper³. Among the different additives, barium titanate (BaTiO3-BT) stabilizes the perovskite phase with only 7-8 mol % addition. The dielectric properties of (1-x-y)PZN

.xBT.yPT compositions were investigated extensively^4,5 and some of the compositions are being manufactured as multiplayer capacitors. By varying the concentration of BT and PT, it is possible to tailor the composition⁶ in the temperature range from −100°

to 220°C. Low temperature dependence of the dielectric constant is important when high temperature stability of component properties is an issue for a particular application, such stability is not possible in high dielectric constant materials due to the high temperature dependence of dielectric constant (ε′). There is a need for high dielectric constant capacitor materials, which can be sintered at temperatures of <1000°C, with low temperature dependence of capacitance. Hence, modification of the dielectric response by optimizing a material through proper choice of the substituent is of utmost importance.

Ferroelectric properties like polarization depend on the extent of lattice distortion, which also affects mechanical properties like elastic constants, elastic energy, plastic energy, etc. Thus, study of mechanical properties with respect to plastic and elastic energy is

modified columbite in three steps with sintering temperature < 1000°C. In order to understand the abnormal decrease of the average phase transition temperature with increase of BT content in PZN and PZN-PT ceramics, and the manner in which the dielectric properties are influenced by Ba²⁺ and Ti⁴⁺

substitution in ABO₃ provides insight into the changes in the dynamical behaviour of Polar Regions in these materials, which is useful for understanding the relaxor behaviour of these materials. In this context, nano-indentation opens up a new vista where one can evaluate hardness, Young’s modulus as well as elastic and plastic energies from applied load-displacement plots^7-10. The mechanical properties of two ferroelectric materials (Pb_0.88Ba_0.12)[(Zn_1/3Nb_2/3)_0.88Ti_0.12)]O₃ (PZN- BT) and (Pb_0.8Ba_0.2)[(Zn_1/3Nb_2/3)_0.8Ti_0.2)]O₃ (PZN-BT- PT) have been studied by nano-indentation and correlated with electrical properties.

2 Experimental Details

PZN-BT and PZN-BT-PT compounds were synthesized by the columbite precursor method in three stages to minimize the secondary phase formation. High purity of raw materials PbO, BaCO₃, ZnO, TiO₅ and Nb₂O₅ was used for the preparation.

The columbite precursor (ZnNb₂O₆) was first prepared by mixing predetermined amounts of ZnO and Nb2O5. It was wet milled in acetone for a day and calcined at 1150°C for 3 h. In the second stage, ZnNb₂O₆, TiO₂ and BaCO₃ were mixed in stoichiometric ratio and calcined at 1150°C for 3 h for formation of ZN-BT. In the third stage, PbO and ZN-BT were mixed in stoichiometric ratio, calcined at 850°C for 2 h and brought to room temperature under controlled cooling. The pellets of the calcined materials were finally sintered at 950°C for 1 h. At each stage of calcinations and final stage sintering, the phases were confirmed with XRD taken at room temperature. Each sintered disc of the prepared ceramic at room temperature was polished on different grades of emery paper to obtain parallel flat surfaces. Silver paste was used on both sides of the disc for electrical contacts. The temperature and frequency dependence of the electrical property was studied by impedance spectroscopy using an LCR meter in our laboratory in the frequency range 50 Hz-1 MHz and in a temperature range from 133 K to room temperature. A programmable controller controlled the temperature. All the data were collected

micrograph (SEM). Depth sensing nano-indentation experiment was done on two samples (PZN-BT and PZN-BT-PT) with a triangular pyramidal Berkovich diamond tip using a commercially available nano- indentation machine (FISCHERSCOPE H100 XYp, Fischer, Switzerland) in a load range between 0.4 and 1000 mN. The force sensing resolution was 0.2 µN.

Similarly, the depth sensing resolution was 0.1 nm.

The experiments were done at five different loads (50, 100, 300, 500 and 1000 mN) with a constant loading as well as unloading time of 30 s. At least 25 data were taken for each case. Polarization (P) versus electric field (E) hysteresis loops were recorded with computer interfaced loop tracer based on modified Sawyer Tower Circuit (Radiant Technologies). The measurements of the saturation polarization (P_s), remnant polarization (P_r) and coercive field (E_c) were carried out on samples of PZN-BT and PZN-BT-PT ceramics.

3 Results and Discussion

XRD profiles at room temperature of developed structures of two compositions PZN-BT and PZN- BT-PT are shown in Figs 1 and 2. The sintered pellets at 950°C show perovskite phase. All the reflections peaks of the X-ray profiles were indexed in a pseudo- rhombohedral (for PZN-BT) and pseudo-tetragonal (for PZN-BT-PT) cell and lattice parameters were determined using a least-squares method with the help of a standard computer program (POWD). Good

Fig. 1 — XRD pattern of PZN-BT and scanning electron micrograph of PZN-BT sintered at 950°C in the inset

agreement between the observed and calculated interplanar spacing (d-values) suggests that the PZN-BT and PZN-BT-PT compounds are having pseudo-cubic structures with a = 4.05932 (106) and tetragonal structures with α = β = γ = 90° (a = 4.0498, b = 4.0498 and c = 4.073), respectively.

The SEM photographs of the fired PZN-BT and PZN-BT-PT ceramics sintered at 950°C for one hour are shown in the inset of Figs 1 and 2. The average grain size is found to be around ∼0.71 and 1.55 µm for PZN-BT and PZN-BT-PT, respectively. Figures 3 and 4 show the real part (ε′) and imaginary part (ε") of dielectric permittivity of two compositions PZN- BT and PZN-BT-PT as a function of temperature at various frequencies. The peak in the dielectric constant ε′ is shifted towards higher temperatures as the measuring frequency is increased, and a dispersion of ε′ exists in the low temperature region [Figs 3(a) and 4(a)]. The peak value (ε′_max) of ε′

decreases with the increase of frequency. These indicate that the relaxor nature is retained in PZN-BT and PZN-BT-PT. It is observed that addition of Ba concentration decreases not only the maximum permittivity but also shifts the T_m to lower temperature. The variation in T_m at 1 kHz frequencies is listed in Table 1 for PZN-BT and PZN-BT-PT.

Figures 3(b) and 4(b) show the temperature dependence of ε″ at various frequencies for PZN-BT and PZN-BT-PT, respectively. It is seen that the peak value (ε"max) of ε″ increases with the increase in frequency which is typical characteristic of relaxor ferroelectric materials. The above described features of (ε′−T) and (ε"−T) variation shown in Figs 3 and 4

are very much similar to the observation by Smolensky et al.¹¹ for various lead-based materials.

It is well known that dielectric permittivity of a normal ferroelectric above Curie temperature follows the Curie-Weiss law described by:

1 ( )

' T

C

  − θ

 =

ε  ^{(T >}θ ) … (1)

(where θ is the Curie-Weiss temperature and C is the Curie-Wiess constant), is observed in the paraelectric phase only at temperatures much higher than T_m i.e.

above Burns temperature, T_B. Figure 5 (a and b) shows the inverse of ε′ as a function of temperature at 1 kHz for the PZN-BT and PZN-BT-PT samples, respectively and it’s fit to the experimental data by

Fig. 2 — XRD pattern of PZN-BT-PT and scanning electron micrograph of PZN-BT-PT sintered at 950°C in the inset

Fig. 3 — Real part (ε′) (a) and imaginary part (ε″) (b) of dielectric permittivity of dielectric permittivity as function of temperature at various frequencies for PBZN15

Curie-Wiess law. A deviation from Curie-Wiess law starting at T_B can be clearly seen. The parameter ∆Tm, which is often used to show the degree of deviation from the Curie-Wiess law, is defined as:

∆Tm = TB − Tm … (2)

The values of TB as determined from the Curie- Wiess law fit and other characteristic parameters calculated, are listed in Table 1. It is observed that the generalized empirical relation in Eq. (1) is obeyed in the paraelectric phase only at temperature much higher than Tm, i.e. above Burns temperature,TB. In the temperature range between Tm and TB, the temperature dependence of permittivity is known to deviate from the typical Curie-Weiss behaviour and obeys the empirical expression by a modified Curie-Weiss relationship¹¹ to describe the diffuseness of the phase transition as follows:

max

max 1

( )

1 1

' '

T T C

  − γ

− =

 

ε ε

  …(3)

where γ = 1 and C1 are modified constants, with 1 = γ

= 2. The parameter gives the information on the character of the phase transition. Its limiting values are γ = 1 and γ = 2 in the expression given in Eq. (3) of the Curie-Weiss law, γ = 1 is for the case of a normal ferroelectric and the quadratic dependence is valid for an ideal ferroelectric relaxor, respectively.

Thus, the value of γ can also characterize the relaxor behaviour. The plot of log{(1/ε′(T)−1/ε′max)} versus

Fig. 4 — Real part (ε′) (a) and imaginary part (ε″) (b) of dielectric permittivity of dielectric permittivity as function of temperature at various frequencies for PBZN25

Fig. 5 — Inverse dielectric constant (1/ε′) as function of temperature at 1 kHz for ceramic PZN-BT (a) and PZN-BT-PT (b). The symbol represents experimental data points and solid line shows fitting to the Curie-Wiess law

Table 1 — Characteristic parameters determined and calculated for ε′(T) measurements at 1 kHz

Composition Frequency T_m (K) θ (K) T_B (K) ∆Tm

PZN-BT 1 kHz 315.78 307.61 371.944 56.164 PZN-BT-PT 1kHz 326.431 314.425 356.951 30.52

log(T−Tmax) is shown in Fig. 6 by fitting with Eq. (3).

The exponent γ, determining the degree of the diffuseness of the phase transition, is obtained from the slope of log{(1/ε′(T)−1/ε′max)}versus−log(T−Tmax) plot. We obtain the value of the parameter γ = 1.286 and 1.259 at 1 kHz for PZN-BT and PZN-BT-PT, respectively. A clear linearity of this relation was observed for both the investigated samples from which γ is calculated. The value of γ deviates from 2 with the increase in Ba which indicates that there is a decrease in relaxor nature within the solid solution where the relaxor nature is retained in PZN-BT. We observe that the addition of lead titanate (PT) to the PZN-BT system has shown a weak frequency dispersion of the dielectric constant, offering an interesting material (PZN-BT-PT) for practical application. Yet another parameter, which is used to characterize the degree of relaxation behaviour in the frequency range 100 Hz-919.28 kHz, is described as follows¹²:

∆Trelax = Tε′m(919.28 kHz) −Tε′m(100 Hz) …(4) The value of ∆Trelax was determined to be 18.22 and 4.99 K for the sample PZN-BT and PZN-BT-PT, respectively. The above characterization done on the basis of Curie-Weiss law and the value of empirical parameters like ∆Tm, γ, and ∆Trelax suggest that the permittivity of PZN-BT and PZN-BT-PT ceramics follows the Curie-Weiss law only at temperatures much higher than Tm. Thus, large deviation from the Curie-Weiss behaviour, large relaxation temperature Trelax and γ, suggests that (PZN-BT and PZN-BT-PT are RFE.

The detailed results of nano-indentation experiment are presented in Table 2. The average load-depth plots of different loads are shown in Fig. 7(a-e) for sample PZN-BT and in Fig. 8(a-e) for sample PZN-BT-PT.

The photomicrographs of the impression of Berkovich indentation array at 1000 mN are shown in Fig. 9 (a-b) for both PZN-BT and sample PZN-BT-PT, respectively. From Table 2 and Figs 7 and 8, it is

Table 2 — Details of nano-indentation experimental results

Sample 1 (0.88 PZN-0.12BT) (PZN-BT) Load

(mN)

Hardness (GPa)

Standard Deviation (GPa)

Young's Modulus (GPa)

Standard Deviation (GPa)

Total Energy, Wt

(mJ)

1000 1.63 0.39 32.09 8.81 1740

500 2.02 0.92 40.84 16.07 610

300 2.11 0.73 46.09 14.63 280

100 3.06 1.13 65.46 17.93 40

50 3.68 1.54 79.94 30.44 10

Load (mN)

Standard Deviation (mJ)

Elastic EnergyWe

(mJ)

Standard Deviation (mJ )

Projected area, Ap (µm²)

Standard Deviation (µm²)

1000 240 590 110 408.11 108.59

500 130 190 40 199.07 104.57

300 70 80 20 106.77 42.3

100 10 10 0 25.62 14.42

50 0 0 0 11.14 8.18

Sample 2 (0.88PZN-0.08BT-0.12PT) (PZN-BT-PT) Load

(mN)

Hardness (GPa)

Standard Deviation (GPa)

Young's Modulus (GPa)

Standard Deviation (GPa)

Total Energy, Wt

(mJ)

1000 2.09 0.43 36.6 7.46 1580

500 1.83 0.32 27.89 3.18 590

300 2.44 0.68 40.1 7.35 240

100 2.86 1.31 69.54 19.74 40

50 3.6 1.51 79.85 22.81 10

Load (mN)

Standard Deviation (mJ)

Elastic Energy, We

(mJ)

Standard Deviation (mJ )

Projected area, Ap (µm²)

Standard Deviation (µm²)

1000 130 600 70 314.35 61.09

500 50 250 10 162.62 34.02

300 30 100 10 80.94 25.85

100 10 10 0 30.03 14.04

50 0 0 0 11.4 5.96

Fig. 6 — log (1/ε′(T) -1/ ε′m) versus log(T−Tm) for (a) PZN-BT and (b) PZN-BT-PT at 1 kHz

Fig. 7—Load-depth plots at loads (a) 50 mN, (b) 100 mN, (c) 300 mN, (d) 500 mN, (e) 1000 mN for PZN-BT

Fig. 8—Load-depth plots at loads (a) 50 mN, (b) 100 mN, (c) 300 mN, (d) 500 mN, (e) 1000 mN PZN-BT-PT

Fig. 9 — Optical photomicrograph of the impression of Berkovich indentation array at 1000 mN on (a) PZN-BT and (b) PZN-BT-PT

clear, that the mechanical parameters like hardness Young’s Modulus and elastic energy are, in general, more for the material PZN-BT-PT as compared to PZN-BT. This is due to more amount of Ti⁴⁺ in the former. Ti⁴⁺(d⁰) is in body centre position with octahedral coordination. Ti-O bond is quite rigid, whereby the permanent distortion of the bond leading to plastic deformation is less as compared to PZN-BT.

This is reflected in its large value of Young’s modulus and elastic energies at all load values, in general. But, the plastic deformation in BZN-PT causes its large contribution to plastic energy and hence total energy at all load values.

The above nature of PZN-BT and PZN-BT-PT is also reflected in the polarization behaviour. The ferroelectric loops of PZN-BT measured at room temperature are shown in Fig. 10(a). There is a typical saturation of ferroelectric polarization Ps=19.71 µC/cm², with remnant polarization P_r = 14.281 µC/cm². For PZN-BT-PT [Fig 10(b)] Ps = 12.41408 µC/cm²,

Pr = 4.97746 µC/cm². The polarization of PZN-BT- PT is significantly less than that of PZN-BT, which is due to the higher Ti⁴⁺ at B-site in the former and more contribution of rigid Ti-O bond making it less vulnerable to polarization.

4 Conclusions

The increase of BT content in PZN and PZN-PT completely suppresses the formation of pyrochlore phase but there is abnormal decrease of the average phase transition. The phase relations and dielectric properties of ceramics with PZN-BT and PZN-BT-PT reveal that both these compositions show a deviation from the Curie-Wiess law, which has only followed at temperatures much higher than the transition temperature (Tm). We observe that the addition of lead titanate (PT) to the PZN-BT system has shown a weak frequency dispersion of the dielectric constant. It is also seen that (PZN-BT-PT) has more elastic energy

Fig. 10 — Polarization hystersis curve of (a) PZN-BT and (b) PZN-BT-PT

energy. The deformation energy is, thus, much less in PZN-BT-PT than PZN-BT. In the former sample, Ti⁴⁺

content is more and it is in B-site of the perovskite ABO3. Ti⁴⁺(d⁰) is in body center position with octahedral coordination. Ti-O bond is quite rigid and restricts plastic deformation. Hence, the presence of more Ti in PZN-BT-PT has caused the increase in elastic energy and decrease in plastic deformation.

This is also evident from polarization (P versus E hysteric measurements) studies of these two systems.

Polarization value for PZN-BT-PT is lower than that of PZN-BT. The rigidity of Ti-O bond makes the former less vulnerable to deformation and the charge separation, which causes polarization in Ti-O bond and to the system.

Acknowledgement

The author (AKH) is thankful to Prof. Ratanmala Chaterjee, IIT Delhi, New Delhi 110 016, India, for providing P-E loop measurement facilities.

579.

3 Himanshu A K, Gupta D C, Dutta A, Sinha T P &

Bandyopadhayay S K, Indian J Pure & Appl Phys, 47 (2009) 212.

4 Halliyal A, Kumar U, Newnhan R E & Cross L E, Am Ceram Soc Bull, 66 (1986) 671.

5 Furukawa O, Yamashita Y, Harata M, Takahashi T &

Inagashi K, Japan J Appl Phys, 24 Suppl (1985) 96.

6 Halliyal A, Kumar U, Newnhan R E & Cross L E, J Am Ceram Soc, 70 (1987) 119.

7 Fischer-Cripps, A C Nanoindentation (Springer-Verag Inc.:

New York), 2002.

8 Oliver W C & Pharr G M, J Mater Res, 7 (1992) 1564.

9 Kanari M, Tanaka K, Baba S & Eto M, Carbon, 35 (1997) 1429.

10 Guicciardi S, Balbo A, Sciti D, Melandri C & Pezzotti G, J Eur Ceram Soc, 27 (2007) 1399.

11 Smolensky G A, Proc 2nd In. Meeting Ferroelctricity, Kyoto, (1969); J Phys Soc Japan Suppl, 28 (1970) 26.

12 Martirena H T & Burfoot J C, Ferroelectrics, 7 (1974) 151.