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

Подождите немного. Документ загружается.

xxx

WPNL2204

Part I

High-strain, high-performance piezo- and

ferroelectric single crystals

WPNL2204

Handbook of dielectric, piezoelectric and ferroelectric materials2

WPNL2204

1.1 Introduction

1.1.1 Background

Continual development in acoustic transduction devices has been made since

the end of World War II. This progress came about largely due to the availability

of improved materials. The example of PZT (lead zirconate titanate)

piezoelectric ceramics is well known. Recently, progress in the growth and

characterization of PMN–PT (lead magnesium niobate–lead titanate)-based

piezoelectric crystals has promoted the development of the next generation

of acoustic transduction devices.

The recent shift in focus from blue water to littoral operations for the US

Navy has placed additional requirements on sonar systems. New materials of

high energy density and improved properties are needed for enhanced sonar

transduction performance. Material property improvements include:

• increased strain to enhance acoustic source level;

• increased electromechanical coupling to broaden bandwidth;

• increased energy density to reduce transducer weight and give higher

efficiencies;

• reduced hysteresis to produce greater thermal stability;

• increased sensibility to improve signal/noise ratio.

A breakthrough was announced at the Piezoelectric Crystal Planning Workshop

sponsored by the Office of Naval Research in May 1997. Single crystals of

PZN–PT

(lead zinc niobate–lead titanate) and PMN–PT near MPB

(morphotropic phase boundary) compositions exhibit extraordinary

piezoelectric properties, namely, electrical field-induced strains exceeding

1%, and electromechanical coupling exceeding 90% (compared with 0.1%

and 70–75%, respectively, in the state-of-the-art PZT piezoceramics)

2,3

. These

perovskite relaxor ferroelectric crystals have opened up new opportunities

not only in current acoustic transduction devices, where traditional PZT

Bridgman growth and properties of PMN–PT-

based single crystals

P HAN, J TIAN and W YAN

H. C. Materials Corporation, USA

WPNL2204

Handbook of dielectric, piezoelectric and ferroelectric materials4

ceramics are used, but also in exploration of new applications, such as medical

ultrasonics, non-destructive detection, marine seismic exploration and energy

harvesting. These new crystals of giant-piezoelectric properties will enable

“revolutionary” developments for the next generation of acoustic transduction

devices. Thus, there is an urgent need to develop crystal-growth techniques

for the fabrication of large, high-quality piezoelectric crystals at industrial scale.

1.1.2 Challenges in the growth of large PMN–PT crystals

It is a great challenge to develop a cost-effective method for the growth of

the high-strain piezocrystals of large sizes (75–100mm (3–4 inches) in diameter

by 150–200mm (6–8 inches) in length, and high quality. Generally speaking,

the difficulties found when growing large-sized crystals of lead-containing

materials are their complex thermodynamic behavior and special physical

properties. For example, common problems include incongruent melting and

low thermal conductivity. The incongruent melting means that crystals cannot

be grown from stoichiometric melts. The low thermal conductivity affects

the transport of latent heat released during the crystallization process, thereby

causing interface instability, defects, inclusions and phase segregation, etc.

The difficulties here stemmed from a basic discrepancy between theoretical

predictions and experimental data on the congruent behavior and the perovskite

precipitation characteristics for the MPB solid solution systems associated

with PT.

In addition, solid-state phase transformations commonly occur on cooling

to room temperature that lead to twinning and possible cracking problems.

Furthermore, the growth of lead-containing crystals at high temperatures

encounters more special technical barriers, including:

• corrosion of container materials – platinum crucibles are attacked by

lead-containing melt at temperatures above 1300

C, leading to severe

leakage;

• high volatility of toxic PbO from the melt at high temperatures;

• difficulties in controlling compositional homogeneity for multi-component

systems due to the compositional segregation.

The critical problems enumerated above make the growth of large PMN–PT

crystals for commercialization a challenging task.

1.1.3 Growth of PMN–PT crystals from stoichiometric

melt using the Bridgman method

PMN–PT is a binary solid solution of lead magnesium niobate (PMN) and lead

titanate (PT). It can be represented by the formula (1–x)[Pb(Mg

1/3

2/3

)

]–x[PbTiO

]. PMN–PT has the ABO

perovskite structure and an MPB

WPNL2204

Bridgman growth and properties 5

where the structure changes from rhombohedral to tetragonal at x ~0.34–

0.35. To obtain crystal samples, the high-temperature solution (flux) growth

as a ‘universal’ method has been widely used for a variety of complex oxide

compounds including the perovskite relaxor ferroelectric materials. In the

1990s millimeter-sized crystals of PMN–PT were grown from high-temperature

solution with PbO/B

as the flux

. Recently, inch-sized (~25 mm) PMN–

PT crystals with improved quality have been successfully grown using the

flux growth method

5–8

and a modified flux growth, the so-called ‘solution-

Bridgman’ method

. However, the growth rate and crystal size of the above

crystal growth methods were limited and not suitable for commercial

production. To suppress the evaporation of the volatile melt component PbO,

crystal growth of PMN–PT under high pressure (80 atm Ar with 1% oxygen)

was demonstrated using a vertical Bridgman furnace. It confirmed that lead

evaporation was significantly reduced even for the unsealed crucible, but the

crystal quality was degraded due to inclusions such as voids and Mg–Si–O-

rich impurities. A possible reason is that high pressure also influenced the

interface dynamic process, leading to the occurrence of constitutional

supercooling.

It has been well known that the most straightforward and economical way

of growing high-quality large crystals is the Bridgman

-Stockbarger

method,

which normally freezes stoichiometric melt without flux: a molten ingot is

gradually crystallized from one end to the other. However, the stoichiometric

melt growth of the single crystals of ABO

perovskite relaxor ferroelectric

materials is suitable only for systems that satisfy the following essential

criteria: (i) the system is congruent melting and/or (ii) in the compositional

phase diagram there must exist a window from which the perovskite as the

primary phase (instead of the pyrochlore phase of the same chemical

composition) can be directly crystallized from the melt. Unfortunately, most

of the known MPB systems associated with PT (PbTiO

) are incongruent

and thus no window exists in the phase diagrams for the perovskite phase to

crystallize first. As a result, these perovskite crystals cannot be grown from

stoichiometric melts. Fluxes or mineral agents must be used for the crystal

growth to avoid interference from unexpected nuclei of the non-perovskite

phases.

Perovskite PMN melts congruently at 1320°C

and perovskite PT melts

congruently at 1285°C. Thus, both end compounds of the PMN–PT binary

system are congruent-melting perovskite phases. This implies that PMN–PT

is more likely to form the perovskite phase, instead of the parasitic pyrochlore

phase, than the other MPB binary solid solutions such as PZN–PT (PZN

melts incongruently). Since the first experimental report in 1997

14,15

on the

melt growth of high-quality PMN–PT crystals from stoichiometric melt

(without flux) in sealed platinum crucibles using a modified Bridgman furnace,

more efforts

16–19

have been made to the melt growth of the PMN–PT-based

WPNL2204

Handbook of dielectric, piezoelectric and ferroelectric materials6

crystals; however, undesired compositional segregations were encountered

At the present time, <001>-seeded PMN–PT crystals of 75 mm (3 inch)

diameter and 200mm (8 inch) length (6kg each boule) have been commercially

manufactured using the multi-crucible vertical Bridgman method

In the following sections, the details of the melt growth of PMN–PT-

based crystals using the Bridgman method developed at the H. C. Materials

Corporation are described. The physical properties are systematically

characterized and discussed in the context of domain engineering and elasto-

piezo-dielectric tensor concepts, which are important for the appropriate

selection of crystal cut directions and vibration modes.

1.2 Crystal growth

1.2.1 Phase equilibrium

A prerequisite for growing PMN–PT crystals is the detailed knowledge of

the high-temperature phase equilibrium of the binary solid solution. There

are limited data about the PMN–PT phase diagram. As mentioned above,

only congruent melting at 1320°C for perovskite PMN and 1285°C for PT

was known. In 1999, we reported the results of the compositional segregation

(Fig. 1.1) based on ICP (induction coupled plasma spectroscopy, accuracy

better than 0.5%) analysis of a PMN–31%PT single crystal

. The growth

parameters are: seeding [210], growth rate 0.8mm/h, temperature gradient

20°C/cm, and maximum temperature 1365°C. The effective segregation

coefficient is estimated ~85% for PT. The result of the crystal growth from

stoichiometric melt indicated that the phase equilibrium diagram may be a

typical binary solid solution system. In 2003,

the first high-temperature phase

diagram of the PMN–PT binary system was reported

. A combination of all

available data, provides a phase equilibrium diagram, as proposed in Fig.

1.2. This phase diagram is accurate enough as the guidance for the thermal

process control during crystal growth and it has been successfully used in the

PMN–PT crystal growth.

1.2.2 Platinum crucible leakage and lead

oxide chemistry

A suitable container/crucible for the lead-containing melt is one of the essential

requirements for PMN–PT crystal growth at the maximum temperature up to

1370°C for a few hundred hours. Several materials such as iridium (Ir),

platinum (Pt)/rhodium (Rh) alloy, Pt/gold (Au) alloy and pure platinum have

been tried. Pure platinum has been proven to be the best crucible material. It

is chemically stable, so does not contaminate the melt. However, the operating

temperature is over the safety temperature limit of 1300°C in air for Pt

WPNL2204

Bridgman growth and properties 7

metal. Thus, it is probable that the platinum crucible could be attacked by

lead-containing melt, especially at temperatures above 1350°C. The crucible

leakages due to PbO attack pose a serious threat to crystal growth. After

systematic investigations, we have found that there are two main reasons

leading to the crucible leakage.

First the purity and the quality of the platinum crucible are not high

enough. During crystal growth at high temperatures platinum grains grow to

up to sub-millimeter sizes and some of the large grains can penetrate the wall

of the crucible. As the platinum grains grow, impurities in the platinum metal

1.8

1.6

1.4

1.2

1.0

0.8

0.6

01234567

Position along growth direction (cm)

Element content

(normalized to starting composition)

1.1

ICP analysis of a PMN–PT crystal boule.

WPNL2204

Handbook of dielectric, piezoelectric and ferroelectric materials8

can be segregated and expelled into the tri-junctions of platinum grains. The

lead-containing melt, acting as a strong flux, would ‘eat’ (dissolve) the

impurities and thus form micro-holes in the platinum crucible wall (Fig.

1.3).

Second, the chemical behavior of lead oxides before melting is complex

The commercial lead oxide (PbO) generally contains some Pb

. To

understand the influence of the chemical reaction on the platinum corrosion,

TGAs (thermal gravity analyses) on PbO and Pb

powders were performed

in conjunction with X-ray diffraction for phase identifications. Figure 1.4

clearly indicates that PbO absorbed oxygen in the temperature range from

430 to 520°C to form Pb

(or PbO

1.1/3

). Above 600 °C, the PbO

1.1/3

released

one-third of the oxygen and returned to PbO. So, one can imagine the following

scenario: if (i) some free PbO exists in the chemical batch loaded in a

platinum crucible and (ii) the temperature of the crucible in the furnace is

hotter at the upper segment, i.e. the heating rate of the upper crucible is

faster than the lower part, then oxygen deficiency inside the sealed crucible

will occur when the temperature at the lower segment of the crucible ramps

up to around 520°C. If the temperature of the upper segment increases over

the liquidus temperature of PbO 888°C, some PbO may be reduced to form

micrometer drops of lead metal. The reduced lead immediately forms alloy

with platinum, which has a much lower melting point than platinum or lead

and thus causes crucible leakage. To avoid the leakage of platinum crucible,

the following measures are suggested:

Liquidus

Starting 31% PT

∆

~ 12 °C

∆PT ~ 14%

solidus

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

PT content from 0–1.0

1325

1320

1315

1310

1305

1300

1295

1290

1285

1280

Temperature (°C)

1.2

Suggested phase equilibrium diagram of PMN–PT binary system.

WPNL2204

Bridgman growth and properties 9

• Use high purity platinum for the crucibles.

• Use Pb

instead of PbO as starting raw chemicals.

• Sinter the mixture of raw chemicals to eliminate the free PbO in the

precursor for crystal growth in any manner, such as a typical ceramic

process.

1.2.3 Apparatus for crystal growth

The Bridgman growth method is the most straightforward and inexpensive

way to grow large crystals of high quality. In order to fabricate PMN–PT

crystals of large size and high quality, cost effectively two major modifications

have been made to our growth system. First, a multi-crucible approach was

taken. Second, to suppress the effect of compositional segregation, a zone-

leveling

22–24

configuration was adopted by narrowing the hot zone of the

furnace chambers. A Bridgman crystal-growth system with a multi-crucible

configuration is given in Fig. 1.5(a). The growth furnace has five chambers;

each has three zones: upper, hot and lower. A platinum crucible is placed in

a ceramic buffer tube assembly on a levitation mechanism that controls the

moving rate of the crucible. A master computer system is used to control the

temperature, crucible moving rate and data acquisition. Five crystal boules

can be obtained from one run (Fig. 1.5(b). The typical growth parameters

1.3

SEM image of 10–40 µm holes at the tri-junctions of platinum

grains in the wall of platinum crucible after crystal growth at 1350°C

for 150 h.

WPNL2204