Zuo-Guang. Ye Advanced Dielectric Piezoelectric and Ferroelectric Materials: Synthesis, Characterisation and Applications

Подождите немного. Документ загружается.

Handbook of dielectric, piezoelectric and ferroelectric materials620

crystal structure and lattice parameter of the substrate to those of the ferroelectric

nano-islands, in this way minimizing the density of interfacial dislocations

and other interfacial defects at the nanostructure/substrate interface. Nanosized

ferroelectric crystals are thus another area of technology where interface

engineering is indispensable.

20.5 Acknowledgments

This work has been performed in close collaboration with M.-W. Chu, M.

Dawber, C. Harnagea, R. Scholz and J.F. Scott, to whom the authors are

grateful for their respective contributions. Special thanks are due to U. Gösele

for his continuous support.

20.6 References

1. Scott J F, Ferroelectric Memories, Berlin, Springer, 2000.

2. International FRAM roadmap; http://public.itrs.net.

3. Wang Y G, Zhong W L and Zhang P L, ‘Surface and size effects on ferroelectric

films with domain structures’, Phys Rev B, 1995 51 5311.

4. Waser R (ed.), Nanoelectronics and Information Technology: Advanced Electronic

Materials and Novel Devices, Weinheim, Wiley-VCH, 2003.

5. Ganpule C S, Stanishevsky A, Su Q, Aggarwal S, Melngailis J, Williams E and

Ramesh R, ‘Scaling of ferroelectric properties in thin films’, Appl Phys Lett, 1999

75 409.

6. Harnagea C, Alexe M, Schilling J, Choi J, Wehrspohn R B, Hesse D and Gösele U,

‘Mesoscopic ferroelectric cell arrays prepared by imprint lithography’, Appl Phys

Lett, 2003 83 1827.

7. Alexe M, Harnagea, C and Hesse D, ‘Non-conventional micro- and nanopatterning

techniques for electroceramics’, J Electroceramics, 2004 12 69.

8. Alexe M, Harnagea C, Hesse D and Gösele U, ‘Patterning and switching of nanosize

ferroelectric memory cells’, Appl Phys Lett, 1999 75 1793.

9. Fujisawa H, Shimizu M, Niu H, Honda K and Ohtani S, ‘Observations of domain

structure at initial growth stage of PbTiO

thin films growth by MOCVD’, Mat Res

Soc Symp Proc, 2000 596 321.

10. Fujisawa F, Morimoto K, Shimizu M, Niu H, Honda K and Ohtani S, ‘Piezoresponse

measurements of Pb(Zr,Ti)O

island structure using scanning probe microscopy’,

Mat Res Soc Symp Proc, 2001 655 CC10.4.1.

11. Alexe M, Harnagea C, Visinoiu A, Pignolet A, Hesse D and Gösele U, ‘Patterning

and switching of nano-size ferroelectric memory cells’, Scripta Materialia, 2001 44

1175.

12. Seifert A, Vojta A, Speck J S and Lange F F, ‘Microstructural instability in single-

crystal thin films’, J Mater Res, 1996 11 1470.

13. Zhao L, Chien A T, Lange F F and Speck J S, ‘Microstructural development of

BaTiO

powders synthesized by aqueous methods’, J Mater Res, 1996 11 1325.

14. Lee W T, Salje E K H and Dove M T, ‘Effect of surface relaxations on the equilibrium

growth morphology of crystals: platelet formation’, J Phys: Condens Matter, 1999

11 7385.

WPNL2204

Nanosized ferroelectric crystals 621

15. Waser R, Schneller T, Hoffmann-Eifert S and Ehrhart P, ‘Advanced chemical deposition

techniques – from research to production’, Integr Ferroelectr, 2001 36 3.

16. Roelofs A, Schneller T, Szot K and Waser R, ‘Piezoresponse force microscopy of

lead titanate nanograins possibly reaching the limit of ferroelectricity’, Appl Phys

Lett, 2002 81 5231.

17. Roelofs A, Schneller T, Szot K and Waser R, ‘Towards the limit of ferroelectric

nanosized grains’, Nanotechnology, 2003 14 250.

18. Harnagea C, Pignolet A, Alexe M and Hesse D, ‘Piezoresponse scanning force

microscopy: what quantitative information can we really get out of piezoresponse

measurements on ferroelectric thin films? Integr Ferroelectr, 2002 44 113.

19. Szafraniak I, Harnagea C, Scholz R, Bhattacharyya S, Hesse D and Alexe M,

‘Ferroelectric epitaxial nanocrystals obtained by a self-patterning method’, Appl

Phys Lett, 2003 83 2211.

20. Shchukin V A and Bimberg D, ‘Spontaneous ordering of nanostructures on crystal

surfaces’, Rev Mod Phys, 1999 71 1125.

21. Williams R S, Medeiros-Ribeiro G, Kamins T I and Ohlberg D A A, ‘Thermodynamics

of the size and shape of nanocrystals: epitaxial Ge on Si(001)’, Annu Rev Phys

Chem, 2000 51 527.

22. Dawber M, Szafraniak I, Alexe M and Scott J F, ‘Self patterning of arrays of ferroelectric

capacitors: description by theory of substrate mediated strain interactions’, J Phys:

Condens Matter, 2003 15 L667.

23. Alexe M and Gruverman A (eds.), Nanoscale Characterisation of Ferroelectric

Materials – Scanning Probe Microscopy Approach, Berlin, Springer, 2004.

24. Chu M-W, Szafraniak I, Harnagea C, Scholz R, Hesse D, Alexe M and Gösele U,

‘Impact of misfit dislocations on polarization instability of epitaxial nanostructured

ferroelectric perovskites’, Nature Materials, 2004 3 87.

25. Streiffer S K, Eastman J A, Fong D D, Thompson C, Munkholm A, Ramana Murty

M V, Auciello O, Bai G R and Stephenson G B, ‘Observation of nanoscale 180°

stripe domains in ferroelectric PbTiO

thin films’, Phys Rev Lett, 2002 89 067601.

26. Fong D D, Stephenson G B, Streiffer S K, Eastman J A, Auciello O, Fuoss P H and

Thompson C, ‘Ferroelectricity in ultrathin perovskite films’, Science, 2004 304

1650.

27. Chu M-W, Szafraniak I, Hesse D, Alexe M and Gösele U, ‘Elastic coupling between

90° twin walls and interfacial dislocations in epitaxial ferroelectric perovskites’,

Phys Rev B, 2005 72 174112.

WPNL2204

622

21.1 Introduction

In recent decades, domain engineering has become one of the most important

areas of ferroelectric science and technology. The main target of domain

engineering is the manufacturing of stable tailored domain patterns in

commercially available ferroelectrics possessing characteristics important

for applications. The ability to introduce the spatial modulation of the electro-

optic, photorefractive, piezoelectric, and nonlinear optical properties by the

creation of periodic domain structure opens wide possibilities in the

manufacturing of devices with upgraded performance. The relatively low

price of these devices can be achieved under mass production using lithographic

electrode patterning for periodical poling. The potential for this technology

is ready to expand into a wider market.

The optimization of the poling process resulting from the fundamental

studies of the domain kinetics is expected to enable the fabrication of sub-

micrometer-pitch gratings and engineered 1D and 2D structures, which could

meet the demanding specifications for the photonic applications. The ultimate

interest is the exploitation of fine (nanoscale) domain structures with capabilities

for optical devices. It is necessary to keep in mind that for achieving high

conversion efficiency the structure has to be precisely reproducible with the

period dispersion about 10nm.

The idea to optimize the ferroelectric properties by the creation of stable

tailored domain structures was advanced 30 years ago by Newnham, Cross

and coworkers (Newnham et al., 1975), who discussed the ability to modify

the crystal properties for piezoelectric devices by producing ‘domains which

are not to be switched during device operation.’ In parallel the independent

branch of the domain engineering aimed at creating the periodically poled

nonlinear optical crystals for light frequency conversion and production of

quasi-phase-matched devices has been developed. The original method of

phase-matching by reversing the sign of the nonlinear coefficient every

coherence length, when a phase difference over π is accumulated between

Nano- and micro-domain engineering in

normal and relaxor ferroelectrics

V Y S H U R, Ural State University, Russia

WPNL2204

Nano- and micro-domain engineering 623

the interacting waves, was proposed by Armstrong et al. (1962) and was

applied experimentally to an array of ferroelectric domains by Miller (1964).

The LiNbO

(LN) and LiTaO

(LT) single crystals are the most popular

materials because of their record electro-optic and nonlinear optical properties.

LN and LT are favorable for the production of precise periodic domain

structures, because 180° domain walls in both crystals are strictly oriented in

proper crystallographic directions.

Feng et al. (1980) and Ming et al. (1982) were the first to produce

periodically poled LN (PPLN) by growing doped LN crystals in periodically

variable electric field conditions, giving rise to the artificially controlled

growth striations, which led to the formation of the periodic domain structure.

Many other methods were proposed later for the creation of PPLN, such as

diffusion, proton exchange and, electron beam, but not all of them satisfy the

requirements of industrial technology.

The most important step in the development of domain engineering in LN

and LT occurred in 1993, when the poling of congruent composition LN

(CLN) by electric field application using the photolithographic electrode

pattern was realized by Yamada et al. (1993), thus opening the way to the

mass production of quasi-phase-matched devices. It was pointed out by Byer

(1997) that ‘the leverage of mass production made possible by lithographic

patterning with subsequent domain inversion has led to a rapid transition

from nonlinear crystals that cost thousands of dollars to fabricate to nonlinear

chips that cost less than one dollar each to fabricate’. Further activities were

directed to the improvement of the electrical poling method, the decrease of

threshold fields, and the struggle with optical damage. The development of

domain engineering requires the understanding of the main physical

mechanisms governing the formation of the domain structure from micro- to

nanoscale.

This chapter gives a review of the most important recent achievements in

the experimental and theoretical study of the domain wall kinetics in single

crystalline LN and LT. Our experimental study of the polarization reversal

reveals that the stable tailored domain patterns are of a kinetic nature. The

crucial result is that almost any domain pattern can be stabilized by the

screening of a depolarization field (Shur, 1996, 2005). The developed approach

based on the critical review of experimental facts allows us to achieve the

essential improvement of the poling processes both for micro- and nanoscale

domain patterning (Shur, 1996, 2005).

The chapter is organized in the following way. In Section 21.2 we introduce

and discuss the main experimental stages of the domain structure evolution

during polarization reversal. Section 21.3 reviews briefly the properties of

the materials under investigation and the experimental methods for domain

observation: from optics to scanning force microscopy. The general approach

to the explanation of the experimentally observed variety of the domain

WPNL2204

Handbook of dielectric, piezoelectric and ferroelectric materials624

structure evolution scenarios is presented in Section 21.4. In Section 21.5 we

discuss the evolution of the domain growth scenario from continuous motion

of the plane domain wall to discrete switching due to the controlled change

of the switching conditions from quasi-equilibrium to highly non-equilibrium.

We demonstrate also how the effectiveness of screening and the domain

merging can drastically change the shape of individual domains. In Section

21.6 the process of the formation of self-assembled nanoscale domain structures

is systematically studied and discussed. Some modern tricks of nanoscale

domain engineering are introduced in Section 21.7. The original kinetics of

nanoscale domains in the heterophase media of relaxor ferroelectrics is

discussed in Section 21.8.

21.2 Main experimental stages of domain structure

evolution during polarization reversal in normal

ferroelectrics

The classical experimental investigations of domain dynamics during

polarization reversal (switching) were realized by in situ optical methods

(Merz, 1954; Little, 1955; Cameron, 1957; Miller and Savage, 1959). Later,

systematic study of domain structure evolution was carried out in several

uniaxial ferroelectric single crystals with optically distinguished domains,

such as lead germanate Pb

(PGO) (Newnham and Cross, 1974; Shur

et al., 1989), gadolinium molybdate Gd

(MoO

)

(Kumada, 1969; Flippen,

1975; Shur et al., 1990), LN (Shur et al., 1999c; Gopalan et al., 1999) and

LT (Gopalan and Mitchell, 1998,1999), which have been considered as model

systems.

In the following we will restrict ourselves to the simplest case of the

polarization reversal process in the 180° domain structures. This is done not

only for the clarity of presentation, but also because the ferroelectrics LN

and LT, which are discussed throughout this chapter, belong to uniaxial

crystals. It must be understood that 180° switching is realized also in multiaxial

ferroelectrics with c-domain structure.

The domain evolution occurring during the polarization reversal of a

single-domain state by application of an external electric field pulse can be

divided into the following main stages (Fatuzzo and Merz, 1967; Shur, 2005):

(1) ‘nucleation of new domains’, (2) ‘forward growth’, (3) ‘sideways domain

growth’, (4) ‘domain coalescence’ and (5) ‘spontaneous backswitching’

(Fig. 21.1). The division into these stages is based on the experimental

observation of the domain dynamics by in situ optical methods (Merz, 1954;

Little, 1955; Cameron, 1957; Miller and Savage, 1959).

‘Nucleation of new domains’ is the most mysterious stage, because it is

practically impossible to visualize the appearance of individual nanoscale

domains. It is a matter of discussion whether the initial domain state could

WPNL2204

Nano- and micro-domain engineering 625

be completely single domain or it contains a large quantity of nanoscale

irremovable residual domains (Shur, 1996). In the latter case the nucleation

stage represents a transition of the invisible residual nano-domains into visible

ones as a result of the field-induced domain growth.

In high-quality crystals the appearance of new domains is observed as a

rule at the surface (Fig. 21.1a). The intrinsic surface dielectric gap existing

in any ferroelectric capacitor provokes incomplete compensation of the

depolarization field in the vicinity of the electrodes (Janovec, 1959; Shur,

1996). This residual field can provide the existence of residual ‘shallow’

domains and also facilitates the ‘nucleation’ process at the surface.

The second stage, ‘forward growth’ (domain tip propagation), represents

a fast expansion of the formed ‘nuclei’ in the polar direction. The tip of the

needle-like domain with charged domain walls rapidly moves through the

sample (Fig. 21.1b). The direct observation of this short stage encounters

experimental difficulties. That is the reason why this process is poorly

explored. The method used to study this stage consists of partial switching

by application of short pulses. It allows us to obtain and to study in detail the

stable structure with charged domain walls consisting of needle-like domains

which have not reached the opposite polar surface (Shur et al., 2000b). The

depolarization fields produced by bound charges result in pronounced optical

contrast due to electro-optical effect. The formation of the quasi-regular

cogged ‘charged domain wall’ and its expansion have been visualized in the

(a) (b)

(d)(c)

(e)

21.1

The main stages of the polarization reversal in ferroelectrics.

WPNL2204

Handbook of dielectric, piezoelectric and ferroelectric materials626

bulk of LN under continuous switching using the uniform metal electrodes

(Shur et al., 2000b).

The third stage, ‘sideways domain growth’ (spreading), is best studied

experimentally by direct in situ methods of domain observation with high

time resolution due to the remarkable optical contrast of the regions in the

vicinity of the domain walls. At this stage the domain growth is achieved

through the domain expansion in the direction transversal to the polar one

(Fig. 21.1c). It has been shown experimentally that under ‘equilibrium’

switching conditions (see Section 21.6) the essential anisotropy of the sideways

domain wall motion results in the growth of regular-shaped polygon domains

with sides oriented along preferred crystallographic directions (Fatuzzo and

Merz, 1967). The detailed study of this stage reveals that the domain shape

depends crucially on the switching conditions (Shur, 2005a, 2006). The

variation of the field pulse parameters and modification of the dielectric gap

properties allow the domain shape to be governed and various regular

and irregular shapes to be obtained (Shur et al., 1984b, 1985, 2002b; Shur,

2006).

The fourth stage, ‘coalescence of residual domains’ (merging), is observed

when the switching process is close to the completion (Fig 21.1d). At this

stage the pronounced deceleration of the approaching walls is observed. The

walls’ motion stops and after definite rest-time the residual domains between

these walls disappear very rapidly. This process leads to the jump-like switching

behavior, which can be one of the sources of the switching current

noise (Barkhausen noise) (Miller, 1960; Lines and Glass, 1977; Shur et al.,

2002a).

The fifth stage, ‘spontaneous backswitching’ (flip-back), represents partial

revival of the initial domain state after the electric field switch off. This

stage is poorly studied experimentally. Nevertheless, the role of this stage

during cyclic switching is crucial. The backswitching represents the enlargement

of the residual domains by the backward wall motion and the nucleation of

the domains with the initial orientation of spontaneous polarization after

abrupt decrease of the switching field (Fig. 21.1e). It is necessary to point

out that the backswitching stage occurs under the action of the abnormally

high field. That is the reason why various self-assembled nano-domain

structures can be produced (Shur et al., 2000a).

It must be understood that the different stages of the domain structure

evolution can occur simultaneously in different parts of a real ferroelectric

sample even during switching in uniform electric field. Moreover, during the

given switching experiment it can happen that not all the stages discussed

can be realized. For example, such stages as nucleation and forward growth

(Fig. 21.1a,b) are not realized for switching from the multidomain initial

state.

WPNL2204

Nano- and micro-domain engineering 627

21.3 Materials and experimental conditions

21.3.1 LN and LT family: from congruent to

stoichiometric

In this chapter the main attention will be paid to experimental study of the

single crystals of LN and LT family owing, first of all, to practical reasons.

LN and LT are the favourite materials of domain engineering. The production

of periodically poled nonlinear chips for light conversion based on these

materials opens wide possibilities for further upgrading of quasi-phase matching

devices.

Moreover, LN and LT have recently become the recognized leaders in the

study of the domain structure stimulated by practical reasons. It is possible

to claim that now LN and LT can be used also as model crystals for the study

of the domain kinetics as they possess the simplest domain structure and

permit in situ optical domain visualization.

LN and LT are uniaxial crystals with C

) symmetry in ferroelectric

phase, thus possessing the domain structure with 180° domain walls only.

The direct study of the domain kinetics in LN and LT is due to optical

contrast of the domain walls. The origin of the contrast is the essential

change of the refractive index in the vicinity of the domain walls induced by

incompletely compensated depolarization field in materials with a pronounced

electro-optical effect.

In the most popular congruent compositions of LN (CLN) and LT (CLT)

the coercive field is enormously high and attains 210kV/cm. Thus, the

application of the necessary threshold field to the sample is coupled with

great experimental difficulty. For this reason both materials were named as

‘frozen ferroelectrics’ for many years. Recently, new representatives of LN

and LT family closer to stoichiometric composition have become available

for research and application (Furukawa et al., 1998; Kitamura et al., 1998;

Niwa et al., 2000; Huang et al., 2001). The stoichiometric LN (SLN) and LT

(SLT) demonstrate essentially lower values of the coercive field and

substantially different domain kinetics as compared with the congruent ones

(Kitamura et al., 1998; Niwa et al., 2000). In addition, we have studied

domain kinetics in LN and LT doped by MgO in order to increase essentially

the optical damage threshold, which is crucial for production of the periodical

poled elements for high-energy light conversion.

The LN and LT single crystals of congruent composition were grown

from the melt by the Czochralski method. The double-crucible modification

of this method and top-seeded solution growth method allows the production

of the near-stoichiometric LN and LT (Kitamura et al., 1998; Niwa et al.,

2000; Huang et al., 2001). Switching experiments were carried out in optical-

grade single-domain wafers cut perpendicular to the polar axis. The typical

thickness of the studied wafers ranges from 0.2 to 0.5mm for CLN and CLT

WPNL2204

Handbook of dielectric, piezoelectric and ferroelectric materials628

and from 1 to 2 mm for SLN and SLT. The polar surfaces of all studied

samples were carefully polished.

For in situ study of the domain kinetics in a uniform field we prepared

two types of circular transparent electrodes: (1) liquid electrolyte (water

solution of LiCl or pipe water) in a special fixture and (2) In

:Sn (ITO)

films deposited by magnetron sputtering, with diameter down to 1 mm. The

visualization by polarizing microscopy in transmitted and reflected light

with simultaneous TV-recording and subsequent image processing was used

for detailed study of the domain kinetics.

The LN and LT wafers, used for the periodical poling, were lithographically

patterned with periodic stripe NiCr or Al electrodes deposited on Z+-surface

only and oriented along one of Y-directions. The patterned surface was covered

by a photoresist layer about 1µm thick. A high voltage pulse producing an

electric field greater than the coercive one was applied to the structure through

the fixture containing the liquid electrolyte (Fig. 21.2). To observe the domain

patterns after partial poling both polar surfaces were etched with pure

hydrofluoric (HF) acid at room temperature. The surface relief obtained was

visualized by optical microscopy, scanning electron microscopy (SEM), and

various modes of scanning probe microscopy (SPM).

21.3.2 Relaxors

Relaxor ferroelectrics are among the most intensively studied materials due

to their unique properties which are useful for various important applications

(Cross, 1994; Blinc et al., 2006; Bokov and Ye, 2006a). They are included in

our investigations for the practical reasons resulting from prominent

electromechanical properties (Cross, 1994). Of course the domain kinetics in

these materials is far from the ‘model’ one because of their complicated

structure.

We have studied the hot-pressed high-density transparent lanthanum-doped

lead zirconate–titanate ceramics (Pb

1–x

)(Zr

0.65

0.35

PLZT (5-12)/65/

35, which exhibits a classic relaxor behavior within a broad temperature

range. The plate thickness varied from 90 to 300µm. Two types of electrodes

–4–3–2–1

21.2

Experimental setup used for periodical poling: 1 wafer, 2 liquid

electrolyte, 3 insulating layer, 4 periodic electrodes, 5 O-rings.

WPNL2204

Nano- and micro-domain engineering 629

were used: (1) transparent ITO deposited by magnetron sputtering and (2)

gold deposited by thermal evaporation.

The hysteresis loops were measured under the action of ac triangular field

pulses in the frequency range from 0.02 to 0.2Hz and the field amplitude up

to 10 kV/cm. The switching charge has been obtained by digital integration

of the recorded switching current. The frequency range was chosen to provide

the ‘quasi-static switching regime’. The switching under the action of the

series of rectangular unipolar pulses with rise time about 20µs, amplitude up

to 10kV/cm and duration ranging from 50µs to 5s was carried out. The

pulse frequency was chosen sufficiently low to provide the complete

backswitching between subsequent pulses which results in good reproducibility

of the current data. All experiments have been done in temperature range

from 20 to 200°C.

Piezo-response force microscopy (PFM) mode was realized on the basis

of EXPLORER™ AFM. Conductive Co-coated silicon cantilevers with 75kHz

resonance frequency were used for visualization of domain patterns. The ac

modulation voltage with amplitude 10V and frequency 17kHz was applied

between the bottom electrode and the conductive tip being in contact with

the surface. Local mechanical deformation at the frequency of the reference

signal was induced by applied ac voltage due to reversed piezoelectric effect.

The piezoelectric response was detected by analyzing the scanning probe

microscopy (SPM) photodetector signal using the lock-in amplifier. The

phase of recorded signal differed by 180° for domains with opposite polarization

directions. The main PFM advantage is the ability to obtain the domain

contrast even in the case of complete screening of the depolarization field at

the surface.

21.3.3 Experimental methods of domain observation in

LN and LT: from optics to SPM

The number of different techniques ranging from common optical microscopy

to more sophisticated SPM has been widely used for the investigation of the

static domain patterns and the kinetics of the domain structure in LN and LT

family crystals. We shall only briefly discuss the most important visualization

techniques. A review of modern experimental methods has been published

(Soergel, 2005).

Selective chemical etching is one of the oldest and most popular techniques

used for revealing the domain structure in ferroelectrics. The inherent difference

in the etch rate for domains with opposite directions of spontaneous polarization

leads to the formation of steps on the polished sample’s surface at the position

of the domain walls. It has been revealed in LN and LT that the etching rates

for the faces of anti-parallel domains are drastically different. The Z-face

etches hundreds of times faster than Z+ one. This method is commonly used

WPNL2204