Effect of bottom electrodes on polarization switching and energy storage properties in Pb

(1)

Effect of bottom electrodes on polarization switching and energy storage properties in Pb

0.97

La

0.02

(Zr

0.95

Ti

0.05

)O

3

antiferroelectric thin ﬁ lms

X.G. Fang

^a

, S.X. Lin

^a

, A.H. Zhang

^a

, X.B. Lu

^a

, X.S. Gao

^a

, M. Zeng

^a,n

, J.-M. Liu

^b,n

aInstitute for Advanced Materials and Guangdong Provincial Key Laboratory of Quantum Engineering and Quantum Materials, South China Normal University, Guangzhou 510006, China

bLaboratory of Solid State Microstructures and Innovation Center of Advanced Microstructures, Nanjing University, Nanjing 210093, China

a r t i c l e i n f o

Article history:

Received 8 May 2015 Received in revised form 16 June 2015

Accepted 18 June 2015 Communicated by Y.E. Lozovik Available online 25 June 2015 Keywords:

A. Antiferroelectricﬁlms B. Sol–gel process D. Energy storage D. Polarization switching

a b s t r a c t

Polarization switching and energy storage properties of a series of Pb0.97La0.02Zr0.95Ti0.05O3(PLZT) thin films deposited on (100)-textured LaNiO3(LNO)-buffered Si substrates and (111)-textured Pt/Ti/SiO2/Si substrates were investigated. It was revealed that the PLZTfilms deposited on the (100)-textured LNO- buffered Si substrates prefer the (100) textured structure, while the orientation of thefilms deposited on the (111)-textured Pt-coated Si substrates is random. With respect to the films on the Pt-coated Si substrates, the (100) textured PLZT films have bigger compressive residual stress, larger electrical polarization, better dielectric properties, and better energy storage performances. For the (100)- orientated PLZTfilms, the energy density (Ws) and efficiency (η) measured at room temperature are about 15.3 J/cm³and 56% respectively. Moreover, the better frequency stability in the range from 20 Hz to 10 kHz, and temperature stability in the range from 25 to 2701C are demonstrated in the (100)- orientated PLZTfilms. These results indicate that the PLZTfilms with LNO bottom electrode could be potential candidate for applications in high energy storage density capacitors.

1. Introduction

Antiferroelectric (AFE) materials with the electricﬁeld induced antiferroelectric-to-ferroelectric phase switching have received considerable attentions due to the promising candidates for future high-energy and fast-speed storage capacitors[1–3]. Among this family, the lead lanthanum zirconate titanate (PLZT) materials exhibiting high dielectric constant and low energy loss have been extensively studied. The AFE PLZT is the solid solution with La substitution at Pb site of lead zirconate titinate (PZT) ferroelectric (FE) materials. Depending on the compositions, various phases, typically including orthorhombic antiferroelectric (AFE), rhombohe- dral or tetragonal FE, and paraelectric (PE), can be obtained in PLZT solid solutions[4,5]. Currently, many research groups have reported the energy storage performances of PLZT dielectrics by modifying compositions,[6,7]and using different substrates, buffered layers, [8]and top/bottom electrodes[9,10]. For examples, Hao et al.[11]

reported a high recoverable energy density of28.7 J/cm³(under maximal electricﬁeldEmax¼2 MV/cm) and an energy efﬁciency of

60% in (Pb0.92La0.09)(Zr0.65Ti0.35)O3films deposited on Pt coated Si substrates. Tong et al.[8]studied (Pb0.92La0.08)(Zr0.52Ti0.48)O3cera- micfilms grown on LaNiO3buffered Ni substrates, and reported a high recoverable electric energy (22 J/cm³atEmax2 MV/cm) and an energy efficiency of 77%. In these works, one key factor affecting the polarization and energy storage performances is the bottom electrode, which chemically interacts with the AFE PLZT [12]. Generally, Pt bottom electrode is adopted in most of the previous studies on the antiferroelectricfilms. However, Pt layer as the bottom electrode could deteriorate ferroelectric properties by the interdiffusion with the interposed adhesion layer of Ti on the substrate [13]. LaNiO3conductive oxide layers as electrodes were the excellent alternatives for solving this problem because it has an acceptable resistivity and the same perovskite structure as many ferroelectric and antiferroelectric materials with small lattice mismatch so that it can be suitable for an effective interface layer for the growth of highly textured and smooth thinfilms[9]. Further- more, temperature and frequency responses of polarization switching are also important for the device applications. Currently, study on the temperature and frequency dependent energy storage properties is still limited.

Under this motivation, we adopted Pb0.97La0.02(Zr0.95Ti0.05)O3

with 2 mol% La-doping since La concentration of 2 mol%

Contents lists available atScienceDirect

journal homepage:www.elsevier.com/locate/ssc

Solid State Communications

nCorresponding authors.

E-mail addresses:[email protected](M. Zeng),[email protected](J.-M. Liu).

Solid State Communications 219 (2015) 39–42

(2)

destabilizes the original FE state of PZT and favors an incommen- surate AFE state [14]. The conductive LaNiO3 and Pt films were chosen as the bottom electrodes and coated on Si substrates. In this study, the effects of the bottom electrodes on orientation, microstructure, polarization, and dielectric properties of the PLZT thinfilms are reported. Finally, we place special emphasis on the frequency- and temperature-dependent polarization, dielectric, and energy storage properties of the PLZT thinfilms.

2. Experimental procedure

The Pb0.97La0.02(Zr0.95Ti0.05)O3 (PLZT) AFE thin films were deposited on (100)-textured LaNiO3(LNO)-buffered Si substrates and (111)-textured Pt/Ti/SiO2/Si substrates via a chemical solution deposition method. The conductive LNO-buffered layers with thickness of 150 nm were coated on Si substrate by a simple sol– gel method. The detailed deposition conditions can be found elsewhere [15]. PLZT precursor solutions with excess Pb-acetate 10 mol% were prepared by dissolving appropriate amount of lead acetate trihydrate [Pb(CH3COO)23H2O], lanthanum acetate [La (CH3COO)3H2O], zirconium propoxide [Zr(OC3H7)4], titanium iso- propoxide [Ti(OCH(CH3)₂)₄] into 2-methoxythanol (all agents from Aladdin). After that the solution was passed through a springe filter and spin-coated on the LNO- and Pt-buffered Si substrates at 4000 rpm for 30 s to ensure thefilm uniformity. After each coating cycle, the sample was pyrolyzed at 4501C for 10 min in air. This process was repeated several times to reach the expectedfilm thickness of200 nm. Finally, the as-prepared thin films were introduced to a rapid thermal annealing process at 7001C for 10 min at oxygen atmosphere to crystallize into perovskite structure.

For electric measurements, top Au electrodes with a 200

μ

^{m in}

diameter were deposited by a rf-sputtering (Model: SBC-12, China) through a shadow mask. The crystallization characteristics of the PLZTfilms were characterized by an X-ray diffractometer (XRD) with CuKa radiation (PANalytical X'Pert PRO). Surface microstructures of thefilms were observed using an atomic force microscope (AFM) (Cypher, Asylum Research). Temperature-dependent dielectric properties of thefilms were examined by an Agilent 4980A LCR analyzer with the ac drive amplitude of 100 mV at 1 kHz. To evaluate energy storage performances of thefilms, temperature- and frequency-dependent electricfield-induced polarization (P–E) hysteresis loops of the films were measured using a Radiant ferroelectric testing system (Multiferroics, Radiant Company).

The temperature was changed from room temperature to 463 K with the ac applied electric ﬁeld of 1 MV/cm at 1 kHz, and the frequency was altered from 1 Hz to 100 kHz at 1 MV/cm.

3. Results and discussion

Fig. 1shows the XRD patterns of the two representative PLZT films deposited on LaNiO3 (LNO)-buffered and Pt-coated Si substrates respectively. All the peaks of the two films are indexed according to the pseudo-cubic structure, indicating that the PLZT films are pure perovskite phase without any second peaks such as parasitic pyrochlore phase. It is noted that the PLZT/LNO films exhibit a (100)-textured structure because the two strongest peaks are (100) and (200). Actually, the textured property can be determined by comparing the relative intensities of the family of planes {100} with the cumulative intensities of all the peaks in the PLZTfilms. The ratio in the PLZT/LNOfilms is up to 83.3%, while it is only 62.1% in PLZT/Ptfilms. The (100) texture in the PLZT/LNO films can be explained by two reasons. One is that the LNO- buffered Si substrate itself presents a preferential (100) texture to

some extent [not shown here]. Other is that the oxide LNO bottom electrode exhibiting a perovskite structure provides good compat- ibility in structure and lattice matching with the PLZTﬁlms than the metal Pt bottom electrode[16,17].

When comparing with the XRD patterns of the two ﬁlms, the peaks of (100)/(200) and (110) in PLZT/LNOﬁlms show slight shift towards lower 2

θ

angle related to that of PLZT/Ptﬁlms. Note that an expansion ofd-spacing in the out-of-plane direction appears, which is a result of compressive strain in-plane direction that is parallel to the substrate surface [18]. Consider that the two PLZT ﬁlms are deposited using the same deposited conditions, the intrinsic and transformation stress can be treated as the same for all samples[19].

Therefore, the compressive stress exhibited in PLZT/LNOﬁlms could be originated from the lattice mismatch between LNO (3.81 Å[20]) and PLZT (4.078 Å[5]in a pseudo-cubic).

The surface morphologies of the two PLZTfilms observed from AFM imaging are shown inFig. 2. Clearly, the PLZTfilm with the Pt bottom electrode exhibits homogeneous surface morphology, while the surface of PLZT film with the LNO bottom electrode represents microstructure consisting of small grains and coarse grains. Such a non-uniform texture in PLZT/LNO films can be ascribed to the non-uniform LNO layer, as reported in previous published papers[17]. The average grain size in PLZT/LNOfilms is 25 nm which is slightly smaller than that with the Pt bottom electrode (30 nm).

The polarization–electricﬁeld (P–E) curves of the PLZTﬁlms on the two substrates measured at 1 MV/cm and 1 kHz are given inFig. 3.

The double loops and nearly zero remnant polarization demonstrate the antiferroelectric nature of PLZTfilms. Moreover, the electricfield- induced switching between the AFE and FE states is clear for the two films. The forward (AFE-to-FE) phase switchingfield (EF) and backward (FE-to-AFE) phase switchingfield (E_A) can be determined by a drastic increase and decrease in polarization, respectively. The values ofEFandEA, respectively, are 566 kV/cm and 365 kV/cm for the PLZT with the LNO bottom electrode, which are larger than those with the Pt bottom electrode (410 kV/cm and 230 kV/cm, respectively). Inter- estingly, the saturation polarizationPsof PLZT/LNOfilms is44.8mC/

cm²(measured at 1 MV/cm), which is much larger than that of PLZT/

Ptfilms (31.6mC/cm²). Oppositely, the remnant polarizationProf the former is smaller than that of the latter. Lee et al. revealed that the saturation polarizations increase with an increase in the in-plane compressive stress, which is beneficial to dipole alignment along the out-of-plane direction that is parallel to the appliedfield[21]. There- fore, the significantly enhanced ferroelectric polarization in the PLZT/

LNOﬁlms can be attributed to the higher compressive stress, textured structure and better interface.

Fig. 1.Room temperature XRD patterns of PLZT thinﬁlms deposited on LNO- and Pt-buffered Si substrates. The arrows denote the peaks of LNO bottom electrode.

The insets show the ampliﬁed XRD patterns for (100) peak and (110) peak of the two PLZT thinﬁlms.

X.G. Fang et al. / Solid State Communications 219 (2015) 39–42 40

(3)

To clarify the effect of bottom electrode on the observed macroscopic dielectric behavior, we measured the dielectric properties at different temperatures. Fig. 4 displays the dielectric permittivity (

ε

r) and loss tangent (tan

δ

) of the two PLZT ﬁlms deposited on the LNO and Pt bottom electrodes from room temperature to 533 K. Similar to the polarization properties in Fig. 3, the PLZT/LNOﬁlms also show larger

ε

r. but lower tan

δ

^than

these of PLZT/Ptﬁlms in the whole temperature range, The reasons can be attributed to that the LNO oxide electrode can avoid a dead layer near the interface so that defects such as oxygen vacancies and space charges will not be accumulated on the interface, which in turn affects the dielectric properties. With increasing temperature,

ε

rin both PLZTfilmsfirst increases and then decreases, which can be attributed to the AFE-FE phase transition[22]. Moreover, the transition temperature of the PLZTfilms deposited on the LNO bottom electrode is lower (473 K) than that from the Pt bottom electrode (483 K). It has been pointed out that the transition temperature will shift to lower temperature if thefilm is placed under a substantial stress[23,24].

As to temperature-dependent energy storage performances, we measured P–Eloops of the two PLZT ﬁlms at 1 MV/cm and at different temperatures. The insets inFig. 4 show several representative P–E loops at selected temperatures. With increasing temperature, the measured EFand EA values gradually decrease, implying that the AFE domain energy recedes and its state becomes more and more unstable. At higher temperature, electric dipoles with a certain degree of freedom overcome the disorder state due to the thermal motion and turn to the uniform direction.

Thus, the double hysteresis behavior becomes less pronounced

with increasing temperature and eventually transforms into a similar ferroelectric hysteresis loop.

According to the temperature-dependentP–Eloops, the energy storage densityWsand the energy efﬁciency

η

are evaluated as a function of temperature, and the results are shown inFig. 5. Here, theWsvalue is calculated by the numerical integration of the area between the positive polarization axis and the backward switching curve of theP–Eloops, and the

η

value is calculated according to the formula: [10,25]

η

¼Ws/(WsþWloss). Where Wloss is the energy loss density calculated by the numerical integration of closed area of the hysteresis loops. Note thatWsof PLZTﬁlms in most reported works is greatly dependent upon the operating electricﬁeld[10]. Our results inFig. 5are calculated at 1 MV/cm, and the valuated value ofWsat room temperature is about 15.3 J/

cm³for PLZT/LNOfilms, which is larger than that of PLZT/Ptfilms (6.8 J/cm³). Our calculated W_s value for PLZT/LNO films is consistent with reported values for PLZT films with LNO top electrodes (14.8 J/cm³) [10]. With increasing temperature, the Wsvalue for the PLZT/LNOfilms decreases gradually to8.7 J/cm³ at 463 K, but is always larger than that of PLZT/Ptfilms over the whole temperature range. Moreover, the

η

in the PLZT/LNOﬁlms is almost constant (56%) from 298 K to 443 K. In contrast, the

η

ⁱⁿ

PLZT/Ptfilm is relatively low (42%). It is worth noting that some sudden jumps of energy storage performances in the higher temperature may be owing to the leakage affiliated with the switching current, as the gap between the start and end point of the P–Eloops increases with increasing temperature. It can be concluded from our results that the PLZT films deposited on the LNO bottom electrodes possess higher energy storage performances and better temperature stability in

η

^.

Fig. 2.AFM images of PLZTﬁlms on (a) LNO-, and (b) Pt-buffered Si substrates.

Fig. 3.Room temperature polarization–electricﬁeld (P–E) hysteresis loops measured at 1 kHz for PLZT thinﬁlms deposited on the LNO and Pt bottom electrodes.

300 350 400 450 500 550

500 1000 1500 2000 2500

-1000 0 1000 -60

0 60

-1000 0 1000 -60

0

60 -60-1000 0 1000 0 60

-1000 0 1000 -60

0 60

0.04 0.08 0.12

T(K)

Dielectricloss

Dielectricconstant

303K 403K

423K 453K

PLZT/Pt PLZT/LNO

Fig. 4.Dielectric constant (εr) and loss tangent (tanδ) as functions of temperature for PLZTﬁlms. Insets show the representativeP–Eloops at selected temperatures.

In theP-Eloop insets,x-axis units are in kV/cm, andy-axis units are inμC/cm².

(4)

Actually, the energy storage performance in the PLZTﬁlms is evaluated according to the P–Eloops. It is well known that the shape of P–E loop can be inﬂuenced greatly by the measuring frequency[26]. Particularly, the charge-releasing speed in energy storage capacitors is inseparably linked with the domain switching speed and an important parameter in practical applications. To correlate the effect of frequency on energy storage performance, we measured theP–Eloops at 1 MV/cm for different frequencies.

Fig. 6 depicts the dependence of Ws on frequency for the PLZT films deposited on the LNO and Pt bottom electrodes. TheW_sin PLZT/LNO films is always higher than that of PLZT/Pt films. It is also noticed that the Ws in the two types of PLZT films keeps nearly constant at lower frequency ofo10 kHz and then drops gradually at higher frequency of410 kHz. This result means that the electric field induced AFE-FE transition can only complete

within the time scale of 10⁴s, and the low domain switching speed may be attributed to the pinning effect of the electric dipoles in theﬁlms[27].

4. Conclusions

In summary, we have grown (100)-textured and nearly random oriented PLZT films on LaNiO3/Si and Pt/Ti/SiO2/Si substrates respectively by sol–gel processing. It was observed that the (100)-textured PLZTfilms have a large compressive residual stress and exhibit enhanced critical electricfieldsEFandEA, as well as electrical polarization, in comparison with those on Pt-coated Si substrates. An energy storage density of 15.3 J/cm³ and an efficiency of 56% were achieved in the (100)-textured PLZT films.

Moreover, the (100)-textured PLZT ﬁlms also showed better energy storage stability in the temperature range from 25 to 2701C and in the frequency range from 20 Hz to 10 kHz. These results suggest that the PLZT AFEﬁlms with LNO bottom electrode have a potential for application in pulsed power electronic systems.

Acknowledgment

This work was supported by the National 973 Project of China (Grant no.: 2015CB654602), and the National Natural Science Foundation of China (Grant nos.: 51101063, 51272078 and 51332007) and the Program for Changjiang Scholars and Innova- tive Research Team in University (Grant no. IRT1243) and the Program for International Innovation Cooperation Platform of Guangzhou (No. 2014J4500016).

References

[1]K. An, P. Li, D. Chen, et al., Solid State Commun. 180 (2014) 64–67.

[2]X.H. Hao, Z. Yue, J. Xu, et al., J. Appl. Phys. 110 (2011) 064109.

[3]X.H. Hao, J.W. Zhai, L.B. Kong, et al., Prog. Mater. Sci. 63 (2014) 1–57.

[4]S.S.N. Bharadwaja, S.B. Krupanidhi, Thin Solid Films 423 (2003) 88–96.

[5]X.H. Hao, J.W. Zhai, F. Zhou, et al., J. Sol–Gel Sci. Technol. 53 (2010) 366–371.

[6]X.H. Hao, J.W. Zhai, X. Yao, J. Am. Ceram. Soc. 92 (2009) 1133–1135.

[7]M. Sharifzadeh Mirshekarloo, K. Yao, T. Sritharan, Appl. Phys. Lett. 97 (2010) 142902.

[8]S. Tong, B. Ma, M. Narayanan, et al., ACS Appl. Mater. Interfaces 5 (2013) 1474–1480.

[9]X. Li, J. Zhai, H. Chen, J. Appl. Phys. 97 (2005) 024102.

[10]J. Ge, X. Dong, Y. Chen, et al., Appl. Phys. Lett. 102 (2013) 142905.

[11]X.H. Hao, Y. Wang, J. Yang, et al., J. Appl. Phys. 112 (2012) 114111.

[12]J. Ge, G. Pan, D. Remiens, et al., Appl. Phys. Lett. 101 (2012) 112905.

[13]D. Bao, N. Mizutani, X. Yao, et al., Appl. Phys. Lett. 77 (2000) 1041.

[14]X. Chen, K. Jiang, Z.G. Hu, et al., Appl. Phys. Lett. 101 (2012) 011914.

[15]J. Liu, W. Geng, X. Chou, et al., J. Sol–Gel Sci. Technol. 62 (2012) 414–418.

[16]X.H. Hao, J.W. Zhai, X. Chou, et al., Solid State Commun. 142 (2007) 498–503.

[17]F. Yan, T.J. Zhu, M.O. Lai, et al., J. Appl. Phys. 110 (2011) 084102.

[18]B.H. Ma, S. Liu, S. Tong, et al., J. Appl. Phys. 112 (2012) 114117.

[19]B.H. Ma, S. Liu, S. Tong, et al., Smart Mater. Struct. 22 (2013) 055019.

[20]H. Miyazaki, T. Goto, Y. Miwa, et al., J. Eur. Ceram. Soc. 24 (2004) 1005–1008.

[21]J.-W. Lee, C.-S. Park, M. Kim, et al., J. Am. Ceram. Soc. 90 (2007) 1077–1080.

[22]X. Chou, M. Guo, Y. Zhang, et al., J. Sol–Gel Sci. Technol. 61 (2011) 62–68.

[23]L. Lian, N.R. Sottos, J. Appl. Phys. 87 (2000) 3941.

[24]M. Narayanan, M. Pan, S. Liu, et al., RSC Adv. 2 (2012) 11901.

[25]X. Hao, J. Adv. Dielectr. 03 (2013) 1330001.

[26]W.Y. Pan, C.Q. Dam, Q.M. Zhang, et al., J. Appl. Phys. 66 (1989) 6014–6023.

[27]L. Jin, F. Li, S. Zhang, et al., J. Am. Ceram. Soc. 97 (2014) 1–27.

0 4 8 12 16 20

280 320 360 400 440 480 20

30 40 50 60 70 80

w

s

(J/c m

3

)

PLZT/Pt PLZT/LNO

T (K)

η (% )

Fig. 5.Dependences of energy densityWsand energy efficiencyηon temperatures in PLZTfilms deposited on LNO- and Pt-buffered Si substrates (The data were calculated at the appliedfield of 1 MV/cm).

10⁰ 10¹ 10² 10³ 10⁴ 10⁵ 0

4 8 12 16 20

W

s

(J/c m

3

)

Frequency (Hz)

PLZT/Pt PLZT/LNO

Fig. 6.Dependence of energy densityWson frequency in PLZTﬁlms deposited on LNO- and Pt-buffered Si substrates. The dashed line is a reference line.