Piezoelectric response of epitaxial Pb(Zr0.2Ti0.8)O3 films measured by scanning tunneling microscopy

(1)

Article

Reference

Piezoelectric response of epitaxial Pb(Zr

_0.2

Ti

_0.8

)O

₃

films measured by scanning tunneling microscopy

KUFFER, Olivier, et al.

Abstract

We report on scanning tunneling microscopymeasurements of the piezoelectric response in ferroelectricheterostructuresgrown by off-axis rf magnetronsputtering. The samples are composed of a single-crystalline ferroelectricfilm of Pb(Zr0.2Ti0.8)O3deposited on a conducting substrate and covered with an ultrathin metallic film of gold. The high quality of the c-axis oriented ferroelectric layer is evidenced by sharp polarization hysteresis loops. By applying a voltage to the bilayer and recording the inverse piezoelectric effect with the scanning tunneling microscope, we demonstrate the ability to measure the phase response as well as the ferroelectric switching. We obtained strain-field plots with a butterfly loop shape, and a quantitative measurement of the longitudinal piezoelectric coefficient (d33).

KUFFER, Olivier, et al . Piezoelectric response of epitaxial Pb(Zr

_0.2

Ti

_0.8

)O

₃

films measured by scanning tunneling microscopy. Applied Physics Letters , 2000, vol. 77, no. 11, p. 1701

DOI : 10.1063/1.1309017

Available at:

http://archive-ouverte.unige.ch/unige:35911

Disclaimer: layout of this document may differ from the published version.

1 / 1

(2)

Piezoelectric response of epitaxial Pb(Zr 0.2 Ti 0.8 )O 3 films measured by scanning tunneling microscopy

O. Kuffer, I. Maggio-Aprile, J.-M. Triscone, O/. Fischer, and Ch. Renner

Citation: Applied Physics Letters 77, 1701 (2000); doi: 10.1063/1.1309017 View online: http://dx.doi.org/10.1063/1.1309017

View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/77/11?ver=pdfcov Published by the AIP Publishing

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 129.194.8.73 On: Wed, 05 Mar 2014 09:24:57

(3)

Piezoelectric response of epitaxial Pb „ Zr

_0.2

Ti

_0.8

… O

₃

films measured by scanning tunneling microscopy

O. Kuffer,^a)I. Maggio-Aprile, J.-M. Triscone, and Ø. Fischer

DPMC, Universite´ de Gene`ve, Quai Ernest-Ansermet 24, 1211 Gene`ve 4, Switzerland Ch. Renner

NEC Research Institute, 4 Independence Way, Princeton, New Jersey 08540

共

Received 12 May 2000; accepted for publication 13 July 2000

兲

We report on scanning tunneling microscopy measurements of the piezoelectric response in ferroelectric heterostructures grown by off-axis rf magnetron sputtering. The samples are composed of a single-crystalline ferroelectric film of Pb

共

Zr_0.2Ti_0.8

兲

O₃deposited on a conducting substrate and covered with an ultrathin metallic film of gold. The high quality of the c-axis oriented ferroelectric layer is evidenced by sharp polarization hysteresis loops. By applying a voltage to the bilayer and recording the inverse piezoelectric effect with the scanning tunneling microscope, we demonstrate the ability to measure the phase response as well as the ferroelectric switching. We obtained strain-field plots with a butterfly loop shape, and a quantitative measurement of the longitudinal piezoelectric coefficient (d₃₃). © 2000 American Institute of Physics.

关

S0003-6951

共

00

兲

02037-4

兴

In the past several years, ferroelectric thin films have been broadly investigated for several applications such as nonvolatile memory elements,¹pyroelectric detectors,²elec- tromechanical actuators,³ and field-effect devices.⁴ Local probe techniques, such as atomic-force microscopy

共

AFM

兲

, have been widely and successfully used to investigate and help understand their physical properties at the nanometer scale.^5–9 In particular, AFM is a powerful tool to study the local inverse piezoelectric response and can be employed to image and study ferroelectric domains structures.^5–9 These studies are directly related to the longitudinal piezoelectric coefficient (d₃₃), whose small amplitude requires very high- resolution measurements. Most AFM piezoelectric measurements are based on the phase detection of the response rather than its magnitude,^5–9and a quantitative measurement of the piezoelectric response coefficient is limited by the difficulty in estimating accurately the electric field between the tip and the sample.¹⁰A possible way to overcome these problems is to use a top flat electrode in order to have a well-defined field.^11,12

Nonlocal studies, such as double-beam laser interferometry, pneumatic pressure rig, or the wafer flexure technique, have been performed by other groups,^13–15with the advantage of taking into account the strain due to the substrate, which complicates the determination of the piezoelectric coefficients, since it can add a component due to the simulta- neous bending of the substrate.

Using a scanning tunneling microscope

共

STM

兲

for local and quantitative piezoelectric response measurements ap- pears to be an interesting alternative approach, because of the extremely high vertical resolution of the STM. Previous studies have reported on such work, but all of them were done on nonperovskite materials.^16–18 Applying this technique to high-quality bilayer heterostructures, composed of a perovskite ferroelectric film covered with an ultrathin elec-

trode, has the advantage of giving clear and reproducible piezoelectric response measurements.

In this letter, we present piezoelectric response measurements made by STM on lead zirconate titanate thin films.

Most of the quantitative local probe measurements of the piezoelectric coefficients of these compounds, using AFM or a tunneling acoustic microscope,^10,19 were performed near the morphotropic phase boundary, because of their high piezoelectric coefficient values.²⁰

A key parameter to reduce strain effects is to choose a composition of PZT having a low lattice mismatch with the substrate. Epitaxial films of Pb

共

Zr_0.2Ti_0.8

兲

O₃

共

PZT

兲

were deposited by off-axis radio-frequency magnetron sputtering on conducting

共

001

兲

Nb-doped SrTiO₃

共

NSTO

兲

single-crystal substrates. The PZT growth is achieved in an argon and oxy- gen discharge at a pressure of 200 mTorr

共

75% Ar/25% O₂

兲

, with a substrate temperature of

⬃

500 °C.²¹ X-ray measurements performed on 300-Å-thick samples displayed finite- size effects on the PZT 001 and 002 reflections, which demonstrate the high-crystalline quality⁸ and allow us to precisely calibrate the deposition rate

共

250 Å/h

兲

.

For the measurements discussed here, we typically grew thicknesses of PZT of the order of 3000–4000 Å in order to reduce leakage current effects. A 60-Å-thick layer of gold is deposited in situ at room temperature. Figure 1

共

a

兲

shows the x-ray diffraction analysis of one of the bilayers with a␪^–2␪ scan revealing epitaxial,²²single-crystalline growth of c-axis oriented PZT with a c-axis lattice parameter of 4.16 Å. The in-plane lattice parameter value is 3.95 Å, giving rise to a mismatch with NSTO of about 1%. The full width at half maximum of the rocking curve taken around the PZT 001 reflection has a value of 0.087°, indicating the high degree of crystallinity. Low-angle measurements allow us to precisely measure the thickness of the gold electrode

共

60 Å

兲

and demonstrate the high quality of the interface between the ultrathin metallic film and the PZT surface. We then performed STM topography images on the patterned gold elec-

a兲Electronic mail: olivier.kuffer@physics.unige.ch

APPLIED PHYSICS LETTERS VOLUME 77, NUMBER 11 11 SEPTEMBER 2000

1701

(4)

trode

共

typical size 0.04– 0.09 mm²

兲

revealing smooth sur- faces with a roughness lower than 6 Å rms

共⬍

15 Å peak to valley

兲

over 1␮^m⫻1␮m areas, as seen in Fig. 1

共

b

兲

. To further characterize the bilayer, polarization hysteresis loops were taken through standard Sawyer–Tower measurements,²³as shown in Fig. 1

共

c

兲

. We obtained a remanent polarization of 35␮^C/cm²and a coercive voltage of about 4 V corresponding to a coercive field of 100 kV/cm for a 4000- Å-thick PZT film.

The experimental setup for measuring the local piezoelectric activity using a STM is shown in Fig. 2. We use a

homemade STM head in ambient conditions²⁴ with electro- chemically etched Ir tips. The tip is brought in constant cur- rent mode (I_t⫽0.5 nA) and with a bias voltage (V_t⫽0.5 V) on top of the ultrathin gold electrode. The film is prepoled before any piezoelectric response measurement. A low- frequency sawtooth voltage

共

10 Hz

兲

is applied to the substrate

共

NSTO

兲

with the top electrode

共

Au

兲

as the ground reference. The frequency is chosen slow enough to ensure constant current operation mode of the STM, since in this configuration, the tip will follow accurately the motion of the PZT layer. Recording this vertical movement directly mea- sures the induced strain

共

z response

兲

, that is the inverse piezoelectric response of the bilayer. A topographic image is taken to measure the slope of the sample, and thus, to verify that the vertical movement of the tip is perpendicular to the surface. The misalignment is always found to be less than 1°.

To demonstrate the technique, we show in Fig. 3 a record of the piezoelectric response of the bilayer, as a function of time, when the STM tip is regulating at a fixed loca- tion. If the driving voltage is smaller than the coercive voltage of the PZT film, the strain response is linear, and in or out of phase with respect to the excitation, depending on the polarization direction. In this case, part I of Fig. 3 shows the sawtooth in-phase response as a function of time. At time t₁, the applied voltage is increased above the coercive voltage of the PZT film, and the ferroelectric switching is observed by a doubling of the frequency response

共

part II

兲

. Finally, at time t₂

共

part III

兲

, the applied voltage is again reduced below the coercive voltage, and the response is out of phase, as seen by the 180° jump of the response (

⌬⌽

) with respect to the signal obtained in part I.

When the PZT film is poled, we found the piezoelectric response to be homogeneous at various positions of the tip on the sample. We emphasize that the Au/PZT interface and the thickness of the gold electrode play an important role in the reproducibility. The ultrathin gold electrode allows one to minimize any artifact such as intergrain or elastic defor- mation effects in the measurement. Furthermore, the surface of the electrode (0.04– 0.09 mm²) is large enough to neglect clamping effects present at the edges.²⁵

We next investigated the longitudinal piezoelectric coef-

FIG. 1. 共a兲Cu K␣ ^x-ray␪^–2␪diffractogram of 60 Å Au/ 4000 Å PZT/

NSTO showing the first two sets of reflections. The size effect at low angle is due to the 60-Å-thick gold electrode.共b兲STM topography.共c兲Sawyer–

Tower measurement of the polarization hysteresis loop of the bilayer. Pr

⫽35␮C/cm²and Ec⫽100 kV/cm.

FIG. 2. Setup for the piezoelectric measurement. A sawtooth voltage is applied to the ferroelectric layer, while the STM tip, in constant current mode on the gold contact共grounded兲, can directly sense the inverse piezoelectric effect at a fixed position.

FIG. 3. Three-dimensional STM image of the piezoelectric response共verti- cal scale兲of the bilayer, measured at one point on the sample, as a function of time共horizontal scales兲. The image was taken by stacking consecutive segments of constant duration共time axis tx兲next to one another共time axis ty兲. Parts I and III of the image show the in- and out-of-phase piezoelectric response for a driving voltage (⫾3 V) lower than the coercive voltage共4 V兲. Part II shows the switching response with a doubling of the frequency for a driving voltage of⫾10 V.

1702 Appl. Phys. Lett., Vol. 77, No. 11, 11 September 2000 Kufferet al.

(5)

ficient (d₃₃) of the bilayer. When the driving voltage is larger than the coercive voltage, we obtain clear strain-field plots in Fig. 4

共

a

兲

, by representing the switching response

共

see part II in Fig. 3

兲

as a function of the applied voltage. This butterfly loop shows both the inverse piezoelectric effect and the polarization switching of the ferroelectric PZT films. Its numerical derivative

关

Fig. 4

共

b

兲兴

gives the d₃₃hysteresis loop.

This kind of measurement yields a d₃₃ value of 50

⫾2 pm/V.²⁶ This value is in good agreement with a recent local probe measurement,²⁷ and can also be compared with double-beam interferometry measurements.²⁸ A striking re- sult is that the coercive voltage values and the shape of the local d₃₃ hysteresis loop

关

Fig. 4

共

b

兲兴

, including the slight asymmetry, perfectly match the macroscopic polarization hysteresis loop measurement obtained in Fig. 1

共

c

兲

, although the nature of the two measurements is different.

In conclusion, we have shown that a STM can be used to probe locally the piezoelectric response of a ferroelectric film covered with an ultrathin electrode. The sample quality is a key point in this technique, since the interfaces and the surface corrugations have to be optimized in order to perform tunneling in good and reproducible conditions. For epitaxial single-crystalline thin films, the high resolution of the instru- ment allows one to obtain clear strain-field plots and gives a direct and quick access to the measurement of the piezoelec- tric d₃₃ coefficient. This technique opens up promising po- tentials for local investigations of ferroelectric memory devices.

The authors would like to thank C. H. Ahn and T. Tybell for fruitful discussions, and J.-G. Bosch and A. Stettler for

valuable technical help. This work was supported in part by the Swiss National Science Foundation.

1J. F. Scott and C. A. Paz de Araujo, Science 246, 1400共1989兲.

2D. L. Polla, C. Ye, and T. Tamagawa, Appl. Phys. Lett. 59, 3539共1991兲.

3G. H. Haertling, J. Vac. Sci. Technol. A 9, 414共1991兲.

4C. H. Ahn, S. Gariglio, P. Paruch, T. Tybell, L. Antognazza, and J.-M.

Triscone, Science 284, 1152共1999兲.

5P. Gu¨thner and K. Dransfeld, Appl. Phys. Lett. 61, 1137共1992兲.

6A. Gruverman, O. Auciello, and H. Tokumoto, Appl. Phys. Lett. 69, 3191 共1996兲.

7C. H. Ahn, T. Tybell, L. Antognazza, K. Char, R. H. Hammond, M. R.

Beasley, Ø. Fischer, and J.-M. Triscone, Science 276, 1100共1997兲.

8T. Tybell, C. H. Ahn, and J.-M. Triscone, Appl. Phys. Lett. 72, 1454 共1998兲.

9O. Auciello, A. Gruverman, H. Tokumoto, S. A. Prakash, S. Aggarwal, and R. Ramesh, Mater. Res. Bull. 23, 33共1998兲.

10C. Durkan, D. P. Chu, P. Migliorato, and M. E. Welland, Appl. Phys. Lett.

76, 366共2000兲.

11G. Zavala, J. H. Fendler, and S. Trolier-McKinstry, J. Appl. Phys. 81, 7480共1997兲.

12J. A. Christman, R. R. Woolcott, Jr., A. I. Kingon, and R. J. Nemanich, Appl. Phys. Lett. 73, 3851共1998兲.

13A. L. Kholkin, Ch. Wu¨tchrich, D. V. Taylor, and N. Setter, Rev. Sci.

Instrum. 67, 1935共1996兲.

14F. Xu, F. Chu and S. Trolier-McKinstry, J. Appl. Phys. 86, 588共1999兲.

15J. F. Shepard, Jr., F. Chu, I. Kanno, and S. Trolier-McKinstry, J. Appl.

Phys. 85, 6711共1999兲.

16H. Birk, J. Glatz-Reichenbach, Li-Jie, E. Scheck, and K. Dransfeld, J.

Vac. Sci. Technol. B 9, 1162共1991兲.

17Li-Jie, C. Baur, B. Koslowski, and K. Dransfeld, Physica B 204, 318 共1995兲.

18R. Winters, M. Reinermann, C. Enss, G. Weiss, and S. Hunklinger, J. Vac.

Sci. Technol. B 13, 1316共1995兲.

19K. Takata, J. Appl. Phys. 79, 134共1996兲.

20B. Jaffe, W. R. Cook, and H. Jaffe, Piezoelectric Ceramics共Academic, London, 1971兲.

21J.-M. Triscone, L. Frauchiger, M. Decroux, L. Mie´ville, Ø. Fischer, C.

Beeli, P. Stadelmann, and G.-A. Racine, J. Appl. Phys. 79, 4298共1996兲.

22Off-axis measurements共␾^scans兲reveal typical fourfold symmetry with the PZT and NSTO peaks aligned, which clearly indicate cube-on-cube epitaxy.

23K. Aoki, T. Sakoda, and Y. Fukuda, Jpn. J. Appl. Phys., Part 2 37, L522 共1998兲.

24Ch. Renner, Ph. Niedermann, A. D. Kent, and Ø. Fischer, J. Vac. Sci.

Technol. A 8, 330共1990兲.

25R. Steinhausen, T. Hauke, W. Seifert, V. Mueller, H. Beige, S. Seifert, and P. Lo¨bmann, in Proc. of the 11th International Symposium on Applica- tions of Ferroelectrics, edited by E. L. Colla, D. Damjanovic, and N.

Setter共IEEE, Piscataway, NJ, 1998兲, p. 93.

26The standard deviation of⫾2 pm/V on d33is derived from the error on the strain response which is⫾5 pm.

27A. L. Roytburd, S. P. Alpay, V. Nagarajan, C. S. Ganpule, S. Aggarwal, E. D. Williams, and R. Ramesh, Phys. Rev. Lett. 85, 190共2000兲.

28S. Hiboux, P. Muralt, and T. Maeder, J. Mater. Res. 14, 4307共1999兲.

FIG. 4. 共a兲Strain-field plot obtained with STM. The vertical movement of the STM 共z response兲is plotted as a function of the driving voltage. 共b兲 Deduced d33hysteresis loop共numerical derivative兲.

1703

Appl. Phys. Lett., Vol. 77, No. 11, 11 September 2000 Kufferet al.