CHAPTER 6 RESULTS AND DISCUSSION: HYDROTHERMAL … · 6.2. Effects of Hydrothermal Synthesis Conditions on Particle Size and Morphology 6.2.1. Effects of Mineraliser Concentration

CHAPTER 6 RESULTS AND DISCUSSION: HYDROTHERMAL SYNTHESIS OF PEROVSKITE PZT POWDERS

95

CHAPTER 6

RESULTS AND DISCUSSION:

HYDROTHERMAL SYNTHESIS OF PEROVSKITE PZT POWDERS

In this chapter, firstly, the hydrothermal synthesis conditions for perovskite PZT phase formation

are investigated and effect of the synthesis conditions on the resulting particle sizes and their

morphologies are discussed. The nucleation and growth mechanisms operating under the present

synthesis conditions are then analysed in order to better control the particle size distribution and

morphology during hydrothermal synthesis. Finally, the hydrothermal synthesis conditions for

producing perovskite PZT powders with a submicron-size and narrow distribution are discussed.

The discussion presented here is based on two recent papers [Su et al., 1997b; 1998].

6.1. Hydrothermal Synthesis Conditions for Perovskite PZT Formation

6.1.1. Effect of Mineraliser Type and Concentration on PZT Synthesis

Fig. 6.1 shows the XRD patterns for the hydrothermal products synthesised at 200°C for 2 hours

from a one-step derived feedstock (see Fig. 5.1 (a)), using different bases as the mineraliser but

at the same molar concentration (2 M). As can be seen, the mineralisers each exhibit different

catalytic effects on the perovskite PZT formation. On using a strong inorganic base, i.e. NaOH or

KOH, phase-pure perovskite PZT was formed without any other impurities being present. By

contrast, only titania was observed to form, as both


96

20 25 30 35 40 45 50 55 60 65 702 θθ°

Arb

itar

y In

tens

ity

*****

** * *∆ ∇ ∇

♦♦♦♦♦

♦

♦♦

♦

♦♦

♦- Perovskite PZT

* - Perovskite PbTiO3

- PbO (Litharge)

∆∇ - TiO2 (Rutile)

ammonia (aq)

TMAH

KOH

NaOH

#

#

- TiO2 (Anatase)

#

Fig. 6.1. XRD patterns for the hydrothermal products synthesised at 200°C for 2 hours using

different bases, at a fixed 2 M concentration, as the mineraliser for the one-step derived

feedstock.

20 25 30 35 40 45 50 55 602θθ°

Arb

itar

y In

tens

ity

∇∇ ∇∇∆∆

##

Two-step derived feedstock

One-step derived feedstock

♦♦

♦♦♦♦♦♦

♦

♦

* *******

*

*

*

∆∆∇∇

#

- Perovskite PZT (Tetragonal)

- PbO (Litharge)- TiO2 (Rutile)

- TiO2 (Anatase)

♦

*

∆∆∇∇∇∇

##

- Perovskite PbTiO3 (Tetragonal)

Fig. 6.2. XRD patterns for the hydrothermal products synthesised at 200°C for 2 hours using 2

M TMAH as the mineraliser but with different methods of feedstock preparation. Note that the


97

perovskite PZT phase is formed when the two-step derived feedstock is used, however; titania

and lead oxide are still present as impurity phases.

anatase and rutile, when using the weak base NH3 (aq) as a mineraliser. When using the strong

organic base TMAH as a mineraliser in the hydrothermal synthesis, the perovskite PZT phase

was not formed unless a two-step derived feedstock was used (see Fig. 6.2), indicating that

perovskite PZT formation depends strongly on the hydrothermal environment. Using the one-step

derived feedstock, the major product is PbTiO 3, with the zirconium component not being

incorporated to form a PbTiO3-PbZrO3 solid-solution. The results confirm that both the

mineraliser type and the hydrothermal synthesis feedstock have a significant effect on the

perovskite PZT formation.

The effect of the KOH mineraliser molar concentration and the feedstock type used during

hydrothermal synthesis on perovskite PZT formation is shown in Fig. 6.3. The critical KOH

mineraliser concentration required for perovskite PZT formation is 1.5 M when using the one-

step derived feedstock at 200°C (see Fig. 6.3 (a)) compared with 0.6 M when using the two-

step derived feedstock at 200°C and 0.4 M at 300°C (see Fig. 6.3 (b)). This is attributed to the

latter feedstock species having less of a steric hindrance effect and/or neutralising action on the

base mineraliser as discussed later. When using the one-step derived feedstock and a mineraliser

concentration below the critical level for this feedstock (0.1 M KOH), the hydrothermal synthesis

product is amorphous. Pyrochlore PbTiO 3 is formed when the KOH concentration is 0.5 M. As

the KOH concentration is increased further to 1.0 M, the hydrothermal product is a mixture of

PbTiO3 and PbZrO3. The XRD peak intensities for the pyrochlore phase decrease, while those


98

for the perovskite phase increase. When the KOH concentration is increased to 1.5 M, phase-

pure perovskite PZT is formed, which has a tetragonal structure as is evident from the peak

splitting in the XRD

(a)

20 30 40 50 60

2θ°θ°

Arb

itary

Inte

nsity

0.4 M

0.6 M

0.2 M

0.4 M

0.6 M

0.8 M

200°C

300°C♦

♦♦♦♦♦♦♦♦

♦♦

♦

♦

♦♦♦♦♦

⊕⊕

⊕

⊕⊕

⊕⊕⊕⊕

⊕⊕ ⊕⊕⊕⊕ ⊕

∇ ∇** #∆

♦

♦ ∇

*⊕#

- P e r o v s k i t e P Z T ( T e t r a g o n a l )

- P e r o v s k i t e P T ( T e t r a g o n a l )

- P X - P h a s e P T ( T e t r a g o n a l )

- P b O ( L i t h a r g e )

-TiO 2 ( R u t i l e )

∆ -TiO2 (Anatase)

Two-step derived feedstock

20 25 30 35 40 45 50 55 60 65

2θθοο

Arb

itary

Inte

nsity

0.1 M0.5 M

1.0 M

1.5 M

4.0 M

♦ ♦♦

♦

♦♦ ♦ ♦ ♦ ♦

◊

◊◊ ◊ ◊◊◊ ◊

♥

♥♥♥◊ ◊ ◊ ◊◊ ◊◊ *

** * * * *

⊗ ⊗⊗⊗⊗

⊗

⊗♦

◊

-Perovskite PZT (Rhombohedral)

-Perovskite PZT (Tetragonal)

* -Perovskite PbTiO3 (Tetragonal)

♥ -Perovskite PbZrO3 (Orthorhombic)

-Pyrochlore Pb2Ti2O6 (Cubic)

One-step derived feedstock

200°C


99

(b)

Fig. 6.3. XRD patterns for the hydrothermal products synthesised at 200°C and 300°C for 2

hours using KOH as the mineraliser, from (a) one-step (200°C) and (b) two-step (200 &

300°C) derived feedstock.

pattern. Finally, the tetragonal perovskite PZT is transformed to the rhombohedral form when the

KOH concentration is increased to 4 M. The results indicate that the PZT composition shifts

gradually towards the PbZrO3-rich side of the PZT phase diagram (see Fig. 2.5) as the KOH

concentration increases. When using a two-step derived feedstock at 200°C below the critical

KOH concentration for perovskite PZT formation, however, it can be seen that the hydrothermal

product is basically composed of amorphous or nanocrystalline material, together with small

amounts of TiO2, PbO and tetragonal PbTiO 3 (PT) Phase. These results suggest that the PZT

formation mechanism operating with the one-step and the two-step derived feedstock is different.

At a higher synthesis temperature of 300°C, a PX-phase PbTiO3 together with PZT are formed

below the critical KOH concentration for this temperature, which is in agreement with the results

of Suzuki et al. [1987]. The PX-phase PbTiO3 is reported to be stable at medium pH and high

synthesis temperatures, and has a tetragonal body-centred-type structure [Cheng et al., 1992].

6.1.2. Effects of Hydrothermal Temperature and Time on PZT Synthesis

The effect of the synthesis temperature on perovskite PZT formation is shown in Fig. 6.4. It can

be seen that the minimum temperature for perovskite PZT formation depends strongly on the

mineraliser used and its molar concentration. At a relatively low mineraliser concentration (e.g.


100

0.3 M NaOH or 0.4 M KOH), the perovskite PZT formation onset-temperature, as indicated by

a main peak intensity of 150 counts above the baseline in the XRD pattern for perovskite PZT, is

at about 230°C to 250°C, whereas it is decreased to between 190°C and 210°C when using the

same mineraliser at a higher concentration (e.g. 0.5 M NaOH or 0.8 M KOH). Also, NaOH

seems more efficient in promoting PZT formation as the onset-temperature is about 200°C for

0.5 M NaOH versus about 210°C for 0.8 M KOH.

Fig. 6.4. PZT perovskite phase evolution (obtained by XRD) as a function of sampling

temperature during hydrothermal synthesis, showing the effect of the inorganic alkali mineraliser

species and concentration used on the formation temperature of PZT. The results show that for

KOH, the minimum temperature for perovskite PZT phase appearance is lower at 0.8 M that at

0.4 M. An even more striking result is shown for NaOH.

150

650

1150

1650

2150

2650

3150

150 170 190 210 230 250 270 290Sampling Temperature, °C

Inte

nsity

abo

ve b

ase

line

of m

ain

PZT

peak

at 2

θθ =

31° (

coun

ts)

KOH 0.4 MKOH 0.8 MNaOH 0.3 MNaOH 0.5 M


101

The time required for the complete formation of perovskite PZT is a strong function of the

synthesis temperature, mineraliser type and concentration. Fig. 6.5 shows a tentative temperature-

time formation diagram for the hydrothermal synthesis of PZT powder. At a low temperature and

for short times, i.e. up to 150°C and 2 hours, respectively, the hydrothermal product is

amorphous. This is because the diffusion rates are not very high and so insufficient time is

available for crystallisation. Increased temperatures (> 200°C) and/or longer reaction times at the

given temperature cause the transformation from an amorphous phase to the perovskite PZT

phase. At intermediate temperatures, phases such as PbTiO 3, PbO, and TiO 2 were found to co-

exist with the perovskite PZT phase, as were amorphous phases having non-stoichiometric PZT

compositions (as indicated by TEM EDX analysis). The presence or otherwise of such phases

was found to depend markedly on the mineraliser type and molar concentration used (see Fig.

6.1 and 6.3, respectively). When the concentration of the given mineraliser is much higher than its

critical concentration for perovskite PZT phase formation at the chosen synthesis temperature, the

multi-phase region becomes fairly narrow and the perovskite PZT phase is formed rapidly from

the amorphous phase.

100

150

200

250

300

350

0 5 10 15 20 25 30 35

Synthesis Time, hours

Synt

hesi

s Te

mpe

ratu

re, °

C

Amorphous Phase Region

Perovskite PZT Region

Multi-phase (perovskite PZT, PT, PbO, TiO2, amorphous) Region


102

Fig. 6.5. A tentative temperature-time phase formation diagram for powders synthesised

hydrothermally using 0.3 M NaOH as a mineraliser, together with the two-step derived

feedstock. The dashed lines represent nominal temperature-time boundaries for the perovskite

PZT phase (n), multi-phase (l), and amorphous phase (¨) regions.

6.2. Effects of Hydrothermal Synthesis Conditions on Particle Size and Morphology

6.2.1. Effects of Mineraliser Concentration

Fig. 6.6 shows SEM micrographs of the PZT powders synthesised at 200°C for 2 hours using

KOH as a mineraliser at different molar concentrations together with the one-step derived

feedstock. As the KOH concentration increases, both the particle size and the extent of particle

agglomeration increase significantly. The morphology of the PZT powder changes from

comprising definitely cubic particles at lower mineraliser concentrations to comprising increasingly

large spherical “cubic PZT particle” agglomerates (of larger and less obviously cubic particles) at

higher mineraliser concentrations. In contrast, when the PZT powder is synthesised from the two-

step derived feedstock, the morphology of the PZT powder is essentially cubic, as shown in

Fig.6.7, since a lower mineraliser concentration is necessary. However, the particle size increases

markedly as the mineraliser concentration increases. It can be seen from Fig. 6.8 (a) and (b) that

increasing the KOH concentration from 0.4 to 0.6 M results in the particle size distribution

becoming narrower but shifting to a larger particle size, with the modal particle size increasing

from 0.37 µm to 0.71 µm. Increasing the mineraliser molar concentration from 0.6 M to 1.0 M in


103

0.2 M steps increases the distribution modal particle size in increments of about 0.25 µm to a

modal particle size of 1.22 µm at a 1.0 M KOH concentration, although the distribution widths

remain essentially the same. A similar trend was shown by the volume % particle size distribution

curves. The modal sizes for both the “elementary particles” as represented by number % and the

“particle agglomerates” as represented by volume %, increase with the mineraliser concentration

as shown in Fig. 6.9.


104

Fig. 6.6. SEM micrographs of the PZT powders synthesised using different mineraliser

concentrations (KOH, 200°C/2 hours): (a) 2 M; (b) 5 M; and (c) 8 M, together with the one-

step derived feedstock, showing that both the particle size and extent of particle agglomeration

increases with mineraliser concentration.

a

c

b


105

Fig. 6.7. SEM micrographs of the PZT powders synthesised at 300°C for 2 hours using (a) 0.6

M and (b) 1.0 M KOH as the mineraliser, together with the two-step derived feedstock, showing

that the morphology PZT particle is cubic and the particle size increases as the mineraliser

concentration increases.


106

(a)

(b)

Fig. 6.8. Effect of mineraliser molar concentration on the particle size distribution of the

hydrothermal PZT powders synthesised at 300°C for 2 hours using KOH as the mineraliser,

together with the two-step derived feedstock. (a) Differential number %; (b) cumulative number

% <.

0

2

4

6

8

10

12

14

0.1 1 10Particle Size, µµm

Diff

eren

tial N

umbe

r %

0.4 M0.6 M0.8 M1.0 M

0

10

20

30

40

50

60

70

80

90

100

0.1 1 10Particle Size, µ µm

Cum

ulat

ive

Num

ber %

<

0.4 M0.6 M0.8 M1.0 M


107

Fig. 6.9. The relationship between the modal particle size in Fig. 6.8 and the mineraliser

concentration for “elementary particles” and “particle agglomerates”, respectively. The synthesis

conditions are the same as in Fig. 6.8. The “elementary particles” sizes are represented by the

number % data because they are less biased by the presence of a small number of large “particle

volume % data.

6.2.2. Effects of Synthesis Temperature and Time

The effect of synthesis temperature on the particle size distribution is related closely to the

mineraliser concentrations employed. As can be seen from Fig. 6.4, the onset-temperature for

perovskite PZT phase powder formation is decreased significantly as the mineraliser

concentration increases; for example, it is reduced by about 50°C from 250°C to 200°C when

0

0.5

1

1.5

2

2.5

0.4 0.5 0.6 0.7 0.8 0.9 1KOH Mineraliser Concentration, M

Mod

al P

artic

le S

ize,

µµm

"Elementary particles"(from number % data)

"Particle agglomerates"(from volume % data)


108

the NaOH concentration is increased from 0.3 M to 0.5 M. For a given synthesis temperature

and time, there is a corresponding minimum mineraliser concentration (specific to the mineraliser

type) for the complete formation of perovskite PZT powders. Yet, the width of the particle size

distribution and the modal particle size are both determined to a large extent by this minimum

mineraliser concentration because the former narrows and the latter increases as the mineraliser

concentration increases, as shown in Fig. 6.8.

The effects of the synthesis temperature and time at a given NaOH mineraliser molar

concentration on the particle size distribution are shown in Fig. 6.10 and Fig. 6.11. First,

comparing the particle size data in Fig. 6.10 for a synthesis temperature of 250°C and synthesis

time of 2, 4 and 6 hours, in terms of the “elementary” particle size (see number % data in Fig.

6.10 (a) and (b)) and the “particle agglomerate” size (as represented by the corresponding

volume % data in Fig. 6.10 (c) and (d)), it can be seen that the differential volume % curve (Fig.

6.10 (c)) for the synthesis time of 2 hours exhibits a tri-modal distribution, with the first-peak

modal particle size of 0.17 µm (see Fig. 6.10(c)) being comparable to the “elementary” particle

size (≤ 0.1 µm) in the differential number % curve (see Fig. 6.10 (a)), and the other two peaks

representing “particle agglomerate” sizes of about 1.4 µm and 4.5 µm. The TEM and XRD

results show that at this stage the hydrothermal product does not comprise fully-crystallised

perovskite PZT particles. About 5~10 area % of the hydrothermal product in the TEM

micrograph is still in the amorphous state, and may be more prone, therefore, to gel formation or

agglomeration. As the synthesis time increases to 4 hours, the degree of crystallisation increases

with further particle growth. The modal “elementary” particle size increases to 0.26 µm, and the


109

particle morphology is perhaps becoming more equiaxed, as indicated by the increasing symmetry

of the distribution, as shown in Fig. 6.10 (a) and (b). The “particle agglomerate” size distribution

is still tri-modal due to agglomerate formation, and the first-peak modal particle size has increased

from 0.17 µm to 0.70 µm; likewise, the two “particle agglomerate” peaks have shifted to larger

modal sizes though the volume percentages are dramatically decreased (see Fig. 6.10(c) and (d)).

Finally, when the synthesis time increases to 6 hours, the degree of crystallisation and the extent of

particle growth are significant. The “elementary” particle size has increased to 0.37 µm and the

particle size distribution become narrower (see Fig. 6.10 (a) and (b)). Consequently, the

tendency to agglomerate decreases, and so, the modal “particle agglomerate” size becomes only

slightly larger (0.75 µm), although the size distribution has become mono-modal, narrower, and

reasonably symmetrical (see Fig. 6.10 (c) and (d)).

The situation at a synthesis temperature of 300°C is even clearer, since the crystallisation of the

PZT powders is more complete. As can be seen from Fig. 6.11 (a) and (b), the modal

“elementary” particle size increases as the synthesis time increases, while the distribution width

decreases and the distribution symmetry increases. Interestingly, however, the modal “particle

agglomerate” size remains almost constant, or perhaps decreases slightly, as the synthesis time

increases, while the distribution width decreases and the distribution symmetry increases. Careful

examination of the particle size data in Fig. 6.11 (a) to (d) demonstrates, in the case of the 4 and

6 hours synthesis time, that although the “particle agglomerate” size distributions shift to larger

particle sizes compared with the “elementary” particle size distributions, there is essentially no

change in the distribution width. The striking difference is for the 2 hours synthesis time, where the


110

PZT powder is not fully crystallised and/or the crystallite size very small, making the powder very

prone to agglomeration. The “elementary” particle and “particle agglomerate” size distributions

for the 2 hours synthesis time are much wider than those for the 4 and 6 hours synthesis time, and

are asymmetric; especially the number % curve.

(a)

(b)

Fig. 6.10 (a) and (b). Captions overleaf.

(c)

0

5

10

15

20

25

0.1 1 10 100Particle Size, µµm

Diff

eren

tial N

umbe

r %

2 hours

4 hours6 hours

250°C

0

10

20

30

40

50

60

70

80

90

100

0.1 1 10 100

Cum

ulat

ive

Num

ber %

<

2 hours4 hours6 hours

250°C

0

1

2

3

4

5

6

7

8

9

0.1 1 10 100Particle Size, µµm

Diff

eren

tial V

olum

e %


250°C

0

10

20

30

40

50

60

70

80

90

100

0.1 1 10 100

Particle Size, µµm

Cum

ulat

ive

Vol

ume

% <


250°C


111

(d)

Fig. 6.10. Effect of synthesis temperature and time on the particle size distribution of the

hydrothermal PZT powders synthesised at 250°C for 2, 4, and 6 hours, respectively, using 0.3

M NaOH as the mineraliser, together with the two-step derived feedstock.

(a)

0

2

4

6

8

10

12

14

16

18


Diff

eren

tial N

umbe

r %


300°C


112

(b)

Fig. 6.11 (a) and (b). Captions overleaf.

(c)

0

10

20

30

40

50

60

70

80

90

100

0.1 1 10

Cum

ulat

ive

Volu

me

% <


300°C0

2

4

6

8

10

12

14

16

18


Diff

eren

tial V

olum

e %


300°C

0

10

20

30

40

50

60

70

80

90

100


Cum

ulat

ive

Num

ber

% <


300°C


113

(d)

Fig. 6.11. Effect of hydrothermal synthesis temperature and time on the particle size distribution of

the hydrothermal PZT powders synthesised at 300°C for different holding times using 0.3 M

NaOH as a mineraliser, together the with two-step derived feedstock. Note the similar trends to

those at 250°C (in Fig. 6.10 (a) to (d)), although the “particle agglomerate” size distributions are

narrower than at 250°C.

Finally, when comparing for a given synthesis time, the particle size data at 250°C and 300°C, it

can be seen that the modal “elementary” particle size increases and the distribution becomes

narrower (Fig. 6.10 and 6.11; (a) and (b)), while the “particle agglomerate” size distribution

becomes narrower and more symmetric (Fig. 6.10 and 6.11; (c) and (d)), as the synthesis

temperature increases from 250°C to 300°C. This behaviour is most clearly shown by the data

for the 4 and 6 hours synthesis times. Disregarding the “elementary” particle size data for the 2

hours synthesis time, which are at the lower size detection limit for the Coulter LS130 particle size

analyser, the modal “elementary” particle sizes increase from 0.26 µm and 0.37 µm at 250°C to

0.41 µm and 0.49 µm at 300°C for a synthesis time of 4 and 6 hours, respectively. The width of

the volume % distribution at 300°C is narrower than that at 250°C.

Therefore, a number of conclusions can be drawn from the above particle size data: (i) smaller

“elementary” particles are more prone to agglomeration than larger ones; likewise, smaller

“particle agglomerates” have a greater tendency to agglomerate than do larger “particle

agglomerates”, (ii) longer synthesis times enable further particle growth with a narrowing of the

size distribution and/or a tendency towards a more equiaxed particle morphology, and (iii) higher


114

synthesis temperatures reduce the extent of agglomeration but the “elementary” particles grow

larger. These conclusions are supported by the TEM observations, as shown in Fig. 6.12. The

“elementary” particle morphology for a synthesis time of 2 hours at 300°C is essentially cubic but

the particles are bonded together in a dendrite-like manner, with an “elementary” particle size of

about 0.1 µm to 0.15 µm (see Fig. 6.12 (a)). After a synthesis time of 4 hours, the morphology

has become more noticeably cubic with an increased size of about 0.3 µm to 0.4 µm and a

somewhat narrower size distribution (see Fig. 6.12 (b)). Finally, after 6 hours synthesis time, the

particle size distribution has narrowed even further and the particle size has increased to between

0.4 µm and 0.5 µm, while the cubic morphology is more fully developed (see Fig. 6.12 (c)).

These observations concerning the “elementary” particles; namely, that they become more

equiaxed and cubic in morphology, whilst increasing in modal size (but with a narrowing of their

size distribution) at higher synthesis temperatures and longer synthesis times, are consistent with a

process of particle dissolution/recrystallisation becoming predominant, at least under the present

synthesis conditions, i.e. using 0.3 M NaOH as the mineraliser, together with the two-step

derived feedstock.


115

a

c

b


116

Fig. 6.12. TEM micrographs of the hydrothermal PZT powders synthesised at 300°C for (a) 2;

(b) 4; and (c) 6 hours using 0.3 M NaOH as the mineraliser, together with two-step derived

feedstock. Note that the morphology of the PZT particles becomes more equiaxed, i.e. “cubic”

as the hydrothermal synthesis time increases.6.3. Hydrothermal Formation Mechanism of

Perovskite PZT Powders

In this section, the formation mechanism of the perovskite PZT powders synthesised

hydrothermally using the two-step derived feedstock is discussed.

6.3.1. Structural Evolution of Hydrothermal PZT Powders

6.3.1.1. FT-IR data

Fig. 6.13 gives the FT-IR spectra for the reaction products extracted during the hydrothermal

synthesis process. The absorption band for the stretching vibration of the lattice hydroxyl group at

about 3500 cm-1 and the surface-adsorbed hydroxyl group at about 3400 cm-1 [Wada et al.,

1996] decreases as the reaction temperature increases from 250°C to 300°C, i.e. in the

sequence (a) to (c). The wave numbers of the double bands at around 1445 cm-1 (νsymCOO)

and 1570 cm-1 (νasymCOO) which are assigned to the acetate (OAc) group, suggest that OAc

ligands are strongly coordinated to the zirconium in a typical bidentate mode [Nakamoto, 1970]

in the starting precursors. Meanwhile, the two bands at 1590 cm-1 and 1530 cm-1, which are

overlapped with the OAc absorbance in the starting precursors, are assigned to AcAc groups

bonded to titanium. They correspond to the ν(C-O) and ν(C-C) stretching vibrations [Sanchez


117

et al., 1988]. When the temperature is raised to 280°C or above, the bending vibration of Zr/Ti

O6 metal-oxygen octahedra at 500-700 cm-1 appears, which is an indication of the structural

development of the PZT perovskite phase [Zhu et al., 1995]. However, the bands at 1020 and

1720 cm-1, corresponding to the ν(C-O) and ν(C=O) stretching vibration respectively, may be

attributed to free acetic acid [Merle-Mejean et al., 1990] or free acetylacetone [Sanchez et al.,

1988], which are still visible in the final product. In addition, the weak bands at 1560 cm-1 and

1350 cm-1 remaining in the final product may be attributed to coordinated carbonate groups

[Merle-Mejean et al., 1990], which is probably responsible for almost certain presence of the

lattice defects in the hydrothermal PZT powders.

It is interesting to note that when the same intermediate product extracted from the autoclave at

250°C is calcined at different temperatures (from 300 to 600°C) in air for 2 hours, the bending

vibration of the Zr/Ti O6 octahedra only appears after calcination at 600°C, double the

temperature of 300°C required for its appearance under hydrothermal conditions. This indicates

that the hydrothermal environment is kinetically more favourable to the formation of the perovskite

PZT structure.


118

Fig. 6.13. FT-IR spectra for the hydrothermal PZT products synthesised using 0.3 M NaOH as a

mineraliser, extracted at: (a) 250°C; (b) 280°C; and (c) 300°C.

Fig. 6.14. TGA-DTA curve (obtained in argon atmosphere at a heating rate of 10°C/min) for the

hydrothermal product synthesised and extracted at 250°C using 0.4 M KOH as the mineraliser,

together with the two-step derived feedstock.

93

94

95

96

97

98

99

100

50 100 150 200 250 300 350 400 450 500

Temperature, °C

Per

cent

age

of W

eigh

t, %

-9

-8

-7

-6

-5

-4

-3

-2

-1

0

DTA

Sig

nal,

µµV/m

g,

End

othe

rm ⇔⇔

Exo

ther

m

⇐

⇒


119

6.3.1.2. TGA-DTA and evolved gas analysis

A typical TGA-DTA trace for the intermediate hydrothermal product is illustrated in Fig. 6.14.

The analysis was conducted in flowing argon. It shows three endothermic DTA peaks associated

with three TGA weight loss regions at about 100°C, 250°C and 330°C. Above 450°C, the

weight loss is complete. The first weight loss is attributed to the desorption of surface-adsorbed

hydroxyl groups. In general, the surface-adsorbed hydroxyl groups possess various bonding

energies with the surface owing to their adsorption on several surface sites with different

coordination numbers, which although mostly desorbing at about 100°C, will continue desorbing

over the wide temperature range from room temperature to 600°C [Wada et al., 1996]. The

second and third weight loss regions at about 250°C and 330°C are due to the dissociation of

organic groups; in this work mainly OAc groups and AcAc groups bonded to the zirconium and

titanium atoms as indicated by the FT-IR (Fig. 6.13) and evolved gas mass spectroscopy (MS)

results. The MS results show that the dissociation is initiated by an endothermic departure of

gaseous acetic acid AcOH (mass/charge ratio, i.e. m/z = 45) at about 200°C followed by a

dissociation of the acetate and acetylacetonate groups bonded to Zr or Ti atoms with the release

of CO2 (m/z = 44) and acetone (m/z = 58) at about 300°C for the intermediate product

extracted at 250°C from autoclave. The exothermic oxidation peak will only become visible when

TGA-DTA is conducted in air because the inert argon atmosphere suppresses oxidation. The MS

results also show that only AcOH and CO2 were detected in the final PZT powders synthesised

at 300°C for 2 hours, indicating that the OAc group is strongly bonded to the Zr atom, and

therefore, difficult to eliminate completely, as shown by the FT-IR results (Fig. 6.13). The nature


120

of the double peaks at 250°C and 330°C is due to the removal of the free acetic acid in the

crystallised structure and the dissociation of the OAc groups bonded to Zr atom, respectively

[Merle-Mejean et al. 1990].

Fig. 6.15 shows the weight loss of the hydrothermal products as a function of the synthesis

temperature for different KOH mineraliser concentrations. The rate of weight loss increased

markedly when the mineraliser concentration was doubled. The weight loss of the PZT powder

synthesised using 0.8 M KOH as a mineraliser is almost complete at 250°C, whereas for powder

synthesised in the presence of 0.4 M KOH, the weight loss end-point is about 300°C.

Obviously, the higher mineraliser concentration facilitates the elimination of the OAc functional

groups, and thereby the nucleation process of the PZT powder particles, as will be discussed

later.

-0.2

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

120 140 160 180 200 220 240 260 280 300 320 340

Sampling Temperature, °C

Wei

ght L

oss

Bet

wee

n 20

0 an

d 45

0°C

at

TGA

Curv

e

0.4 M KOH0.8 M KOH


121

Fig. 6.15. Weight loss (%) between 200°C and 450°C according to TGA curve data versus

sampling temperature, for the hydrothermal PZT products synthesised using 0.4 M and 0.8 M

KOH as a mineraliser.

6.3.1.3. XRD data

Fig. 6.16 (a) and (b) show the powder XRD patterns for the extracted hydrothermal products

synthesised at various stages in presence of 0.4 M and 0.8 M KOH, respectively. When the

KOH concentration is 0.4 M, the hydrothermal product is a completely amorphous gel at 200°C.

As the temperature is increased to 250°C, the perovskite PZT phase, as a mixture of the PZT (T)

and the PZT (R) forms, begins to form. Note that PZT (T) and PZT (R) denote a Ti-rich

tetragonal PZT phase and a Zr-rich rhombohedral PZT phase, respectively [Cheng et al., 1994].

Co-existing with the PZT phase, are PbO (tetragonal) and TiO2 (anatase and rutile) as undesired

phases. When the temperature is increased further to 280°C, the PbO transforms to PT(ss), a

PbO-TiO2 solid-solution (TiO2 < 10 mol %) [Takai et al., 1983], accompanied by an increase in

the amount of PZT (T), whereas the level of PZT (R) remains low. Finally, when the temperature

is increased to 300°C, the peak intensities of both the PZT (T) and PZT (R) increase significantly.

However, when the KOH concentration is increased to 0.8 M, the onset-temperature for

perovskite PZT phase formation is decreased to 200°C in comparison with 250°C when using

0.4 M KOH as the mineraliser (see Fig. 6.16 (b)). It can be seen that the amorphous product at

150°C transforms to a mixture of PZT (T) and PZT (R) phases plus TiO2 (anatase and rutile) and

tetragonal PbO (litharge) phases at 200°C. As the temperature is increased to 250°C and above,


122

the peak intensities of the perovskite PZT phase increases markedly and the peaks become

noticeably narrower and sharper in comparison with those using 0.4 M KOH as the mineraliser,

which indicates that PZT powder with a larger crystal size and a higher degree of homogeneity is

formed. Both the PZT (T) and PZT (R) shift towards the morphotropic phase boundary as the

mineraliser concentration increases.

It is interesting to note that, under the same synthesis condition as for Fig. 6.16 (a), only PbO (ss)

is formed as an intermediate phase at temperatures up to 280°C when 10 wt. % excess PbO is

used in the feedstock. When comparing the XRD patterns in Fig. 6.16 (c) with those in Fig. 6.16

(a) at 280°C and 300°C for 2 hours, it can be seen that, the peak of PZT phase at 2θ ~ 31° in

Fig. 6.16 (c) is narrower and sharper than that in Fig. 6.16 (a), indicating excess Pb facilitating in

the perovskite PZT phase formation and particle growth. In the final product, however, extra

peak due to PbO is found (see Fig. 6.16 (c)).

Fig. 6.16. XRD patterns for the hydrothermal reaction products extracted at different stages,

synthesised using KOH as the mineraliser, together with the two-step derived feedstock.

20 25 30 35 40

2θ°θ°

Arb

itary

Inte

nsity

200°C

250°C

280°C

300°C

300°C/2h

∆

∆

∆

∇

∇

∇

∇

∇

#

##

#

+

++

⊗

⊗

⊗

⊗

⊗

⊗

⊗

⊗

⊗

♦

♦

♦

♦

♦

♦ ♦

♦

♦

♦

♦

♦

♦

♦

♦

♦

♦

- PZT(T)

- PZT(R)

- PbO (ss)

- PbO (Litharge)

- TiO2 (Rutile)

- TiO2 (Anatase)

(a) 0.4 M KOH

20 25 30 35 40

2θ°θ°

Arb

itary

Inte

nsity

(b) 0.8 M KOH

150°C

200°C

250°C

280°C♦ ♦

♦

♦

♦

♦

♦

♦

♦

♦

∇

∇∇∆

∆

##

#

#

⊗

⊗

⊗

⊗

⊗

⊗

⊗

⊗♦

♦

- PZT (R)

- PZT (T)

- PbO (Litharge)- TiO 2 (Rutile)

- TiO 2 (Anatase)

20 25 30 35 40

2 θ°θ°

Arb

itar

y In

tens

ity

200°C

250°C

280°C

300°C/2h

♦

♦

♦

♦

♦♦

♦

♦

♦♦

♦

(c) 0.4 M KOH10 wt. % excess Pb

#

#

+

++

+ +

⊗⊗

⊗

⊗

⊗

⊗

⊗

- PZT (T)

- PZT (R)

- PbO (Litharge)

- PbO (ss)


123

6.3.1.4. TEM observations

Fig. 6.17 shows the typical characteristics of the reaction products present at various stages

during the hydrothermal synthesis process. The initial hydrothermal products treated at lower

temperatures appear amorphous and nearly identical to the precursor gel, exhibiting some

clustered structures (Fig. 6.17 (a)). As the temperature is increased to 250°C or above,

crystallisation seems to proceed via collapse of the amorphous gel, resulting in grains with a

mosaic structure (Fig. 6.17 (b)). A SAD pattern taken from the particles is inserted in Fig. 6.17

(b) and shows that the particles are well crystallised. However, EDX analysis on the particles

showed that they are non-stoichiometry in comparison with the pre-designed PZT composition in

the feedstock. Both Ti-deficiency and Zr-deficiency particles were detected. These results are in

agreement with the XRD results as shown in Fig. 6.16 (a) that the initial PZT phases formed are

PZT (T) and PZT (R). At the same time, some minor intermediate phases are also formed. As

can be seen from the TEM micrograph (Fig. 6.17 (c)), acicular (A) and tabular (B) particles are

present, together with an amorphous gel (C) and perovskite PZT particle (D). These features are

also observed during the hydrothermal synthesis of PbTiO3 [Watson et al., 1987] and PLZT

[Cheng et al., 1996]. EDX analysis shows that these tabular and acicular particles have various

Pb and Ti compositions in majority and only minor Zr composition. The Pb/Ti ratio is about 1 for

tabular particles, and < 1 for acicular particles, which indicates that they are PbTiO3 or PZT (T),

and PbO or PbO (ss), respectively, as is also evident from the XRD pattern. After synthesis at

300°C for 2 hours, cubic morphology perovskite PZT powders are formed (Fig. 6.17 (d)). The

coexistence of various particle morphologies, i.e. equiaxed cubic, tabular and acicular, suggests


124

simultaneous formation mechanisms. The particles formed by in-situ transformation are typically

equiaxed in nature, while the dissolution /recrystallisation

Fig. 6.17. TEM micrographs of the hydrothermal reaction products extracted at (a) 200°C; (b)

250°C; (c) 280°C; and (d) after 2 hours at 300°C, using 0.4 M KOH as a mineraliser. Note in

A

B

C

D


125

(c) that acicular (A) and tabular (B) particles co-exist with an amorphous gel (C) and cubic

morphology perovskite PZT particle (D) during the synthesis process.

process accounts for the presence of acicular and tabular faceted particles [Watson et al., 1987].

Judging from the number of equiaxed particles formed, the former mechanism is dominant

throughout the hydrothermal synthesis process, at least under the conditions investigated in the

present work.

6.3.1.5. Electrophoretic mobility data

The electrophoretic mobility of the reaction products in aqueous solution has been measured to

evaluate the electrostatic surface charge characteristics of the sol particles. The mobility of the

reaction products as a function of solution pH is shown in Fig. 6.18. For comparison, the

mobility versus pH curves for the commercial PbO (masicot) and the hydrothermally synthesised

zirconia-titania (ZTO) powders are also plotted in the same figure. The isoelectric point (IEP) of

the hydrothermal PZT powder decreases as the hydrothermal synthesis temperature increases,

from pH9.8 to pH6.3. Note that the IEP of the PbO is pH11.2, whereas that of the ZTO is

pH5.4. These data show that as the hydrothermal synthesis temperature increases, the IEP of the

reaction products moves away from that of the PbO and becomes closer to that of the ZTO. This

implies that in the early stages of synthesis, the ZTO gel is heterocoagulated with adsorbed PbO

sol particles on the gel surface, and that as the synthesis temperature increases and greater inter-

diffusion occurs, the surface charge characteristics of the reaction product are modified, bringing

the IEP closer to that of the ZTO, i.e. effectively the IEP of doped ZTO. From Fig. 6.18, it can

also be concluded that the PZT powder is likely to be formed more readily within a certain pH


126

range. Over the pH range 7 to 11, the PbO sol particles are positively charged, whereas the ZTO

powder particles are negatively charged. The electrostatic repulsion barrier between the PbO and

the ZTO particles is thus absent; so, a strong attractive interaction between them can be

expected, resulting in particle heterocoagulation. Even above pH11, PbO and ZTO powder

particles with the same charge polarity but with different surface potential magnitudes may still be

mutually attracted and are thus likely to undergo rapid heterocoagulation [Biscan et al., 1993]. If,

however, the pH is high enough for the particles to develop a similar surface potential magnitude

they will tend to repel each other, making heterocoagulation unlikely. This is a further possible

reason why PZT cannot be formed at too high a pH, as predicted thermodynamically [Lencka et

al., 1995] and observed experimentally [Ohba et al., 1996].

Fig. 6.18. Electrophoretic mobility versus solution pH for the hydrothermal reaction products

extracted at 250°C; 280°C; and after 2 hours at 300°C. Electrophoretic mobility versus solution

-5

-4

-3

-2

-1

0

1

2

3

4

4 5 6 7 8 9 10 11 12Solution pH

Ele

ctro

phor

etic

Mob

ility

, (µµm

/sec

)/(V

/cm

)

250°C280°C300°C/2hPbOZTO


127

pH curves of commercial PbO (masicot) and hydrothermal zirconia titania (ZTO) powders are

also plotted for comparison.

6.3.1.6. ICP data

ICP measurements have been conducted on the filtrate solutions of the samples extracted at

different stages during the hydrothermal synthesis process on heating up from room temperature

to the maximum synthesis holding temperature of 350°C to identify the residual ions in the

hydrothermal solution. The levels of both titanium and zirconium were found to be negligible within

experimental error. The lead concentration in the extracted filtrate increased slightly as the

hydrothermal synthesis temperature increased but is strongly dependent on the mineraliser

concentration in the hydrothermal solutions, as shown in Fig. 6.19. Below 280°C, no lead solute

was found in the hydrothermal solution when employing 0.4 M KOH as a mineraliser. However,

when the KOH concentration was increased to 0.8 M, the lead content in solution was found

even at a temperature of 200°C, which is almost doubled that measured for the 0.4 M KOH at a

temperature of 300°C. This may be a further indication that different PZT particle formation

mechanisms operate at different mineraliser concentrations. Nevertheless, the total lead content

remaining in the supernatant solution is still relatively low (< 2 wt. % of the lead acetate precursor)

under the current synthesis conditions. A similar result is found when using 0.3 M NaOH as the

mineraliser, as shown in Fig. 6.19. Below 280°C, no lead is detected in solution. Above 280°C,

it increases steadily with temperature (see Fig. 6.19). Furthermore, as the holding time at 300°C

was increased from 2 hours to 6 hours, the lead concentration in the supernatant solution

increased markedly (see Fig. 6. 20). All these ICP results, in conjunction with the XRD and TEM

results, suggest that a minor dissolution process occurs after the perovskite PZT powder particles


128

are formed. It is also interesting to note from Fig. 6.19 that, when the KOH concentration was

0.4 M, the mineraliser cation concentration in the filtrate solution increased rapidly as the

hydrothermal synthesis temperature increased. This tends to suggest that the potassium ions

initially adsorbed on the precursor gel are gradually desorbed into solution during the

hydrothermal synthesis process. Although when the KOH concentration was increased to 0.8 M,

this trend was much less obvious.

Fig. 6.19. Normalised concentration ratio of Pb or K content in supernatant to Pb or K content in

feedstock as a function of hydrothermal synthesis sampling temperature for different mineralisers

(KOH and NaOH) and concentrations (0.4 M and 0.8 M).

0

0.2

0.4

0.6

0.8

1

1.2

200 220 240 260 280 300 320 340

Sampling Temperature,°C

Pb

in H

ydro

ther

mal

S

olu

tio

n/F

eed

sto

ck, %

40

45

50

55

60

65

70

75

80

K in

Hyd

roth

erm

al

So

luti

on

/Fee

dst

ock

, %

- 0.4 M KOH- 0.8 M KOH- 0.3 M NaOH

⇒

⇒

⇐

⇐

⇐

o

u¡


129

Fig. 6.20. Pb content in the supernatant solution as a function of hydrothermal synthesis time at

300°C using 0.3 M NaOH as the mineraliser.

6.3.1.7. pH change of hydrothermal solution

The change in the filtrate solution pH as a function of the hydrothermal synthesis sampling

temperature for different mineraliser concentrations is shown in Fig. 6.21. When the hydrothermal

synthesis was conducted at a critical minimum mineraliser concentration for perovskite phase

formation (i.e. 0.3 M NaOH or 0.4 M KOH), the pH decreased non-linearly from about pH12

at room temperature to almost neutral at 300°C, while at higher mineraliser concentrations (i.e.

0.5 M NaOH and 0.8 M KOH), the pH decreased only slightly with increasing synthesis

temperature. The difference in the rate of change of pH with temperature for different mineraliser

concentrations may either be indicative of or account for the different PZT particle formation

processes envisaged on the basis of the present experimental results.

100

200

300

400

500

2 3 4 5 6

Reaction Time, hours

Pb

Con

cent

rati

on in

Sup

erna

tant

, g/l


130

Fig. 6.21. Filtration solution pH as a function of hydrothermal synthesis sampling temperature for

different mineraliser concentrations.

6.3.2. Formation Mechanism of PZT Powder Particles during Hydrothermal Synthesis

6.3.2.1. Structure of starting precursors

The Pb, Ti and Zr precursors each undergo different reactions during the hydrothermal feedstock

preparation process. It has been reported that the precipitated titania and zirconia precursor gels

tend to form polymeric chains such as [Ti-O-Ti]n and [Zr-O-Zr]n in preference to isolated Ti4+ or

Zr4+ ions except at strong acidic conditions [Baes and Mesmer, 1986]. The Ti4+ or Zr4+ ion does

not actually exist because of its high charge to ionic radius ratio [Kutty & Padmini, 1997]. In the

present work, the reaction of titanium alkoxide with acetylacetone (AcAc) leads to several

molecular compounds [Livage, 1994]. The basic reaction is as follows:

5

6

7

8

9

10

11

12

13

14

15

25 50 75 100 125 150 175 200 225 250 275 300Sampling Temperature, °C

Filtr

ate

Solu

tion

pH

0.8 M KOH0.4 M KOH0.5 M NaOH0.3 M NaOH


131

(6.1)

In the presence of excess AcAc, e.g. a TIPT to AcAc molar ratio of 1:2 as in this work, the

following molecular structure might be formed [Smith et al., 1972]:

(6.2)

During these stoichiometric reactions, the coordination number of the titanium atoms increases

from four to six with the low-molecular-weight oligometric species being formed during the

hydrolysis process. Upon hydrolysis, the alkoxide groups are removed first, whereas the

bidentate acetate ligands remain bonded to the titanium atoms. They prevent further condensation

during feedstock preparation, and so only small oligomers are formed at the stage. During

hydrothermal treatment, further condensation may occur, leading to the formation of fibrous titania

gels [Hennings et al., 1991] and polyhedral crystals in anatase and/or rutile structures depending

on the synthesis conditions [Waston et al., 1987]. Correspondingly, the solubility of the titania gel

will decrease with increasing hydrothermal temperature as reported by Ohba et al. [1996].

Zirconium acetate is a basic salt of indefinite composition and contains complex anions. The

compound is generally given the formula, ZrO(OAc)2.xH2O. In aqueous solutions, the compound


132

probably exists as polymeric chains of unknown lengths [Luebke, 1969]. Hydrolysis of the

acetate ligands, which leads to zirconium oxide formation can be expressed as [Vesteghem et al.,

1992]:

ZrO(OAc)2⋅xH2O (aq) + 2 H2O(l) → ZrO2⋅xH2O (s) + AcOH (aq) (6.3)

This reaction can be accelerated by the presence of bases as observed during feedstock

preparation, or by hydrothermal treatment at high temperatures [Vesteghem et al., 1992]. Owing

to the polymeric nature of the zirconium acetate, the hydrolysis product could be regarded as a

hydrous zirconia gel. According to the FT-IR and TGA-DTA results, this hydrolysis reaction is

incomplete with residual OAc groups remaining bonded to the zirconium atoms.

The precipitate from the lead acetate trihydrate dissolved in KOH solution will undergo following

reaction resulting in the formation of hydrous lead oxide:

Pb(OAc)2⋅3H2O(aq) + 2KOH(aq) → PbO⋅3H2O(s) + 2 K(OAc)(aq) + H2O(l) (6.4)

However, the form of PbO produced can vary depending on the feedstock pH according to the

following reaction equilibria [Margolis, 1966]:

PbO(s) + H2O(l) ⇔ Pb2+(aq)+ 2 OH-(aq) (6.5)

Pb2+(aq) + H2O(l) ⇔ PbOH+(aq)+ H+(aq) (6.6)

PbOH+(aq) + OH-(aq) ⇔ Pb(OH)2 (aq) (6.7)

Pb(OH)2(aq) + OH-(aq) ⇔ HPbO2-(aq) + H2O(l) (6.8)

Pb(OH)2(aq) ⇔ Pb(OH)2 (s) (6.9)


133

At a relatively low basic pH, reactions (6.5) and (6.6) are promoted, yielding Pb2+ and Pb(OH)+.

At a relatively high basic pH, reactions (6.7), (6.8) and (6.9) dominate. The neutral Pb(OH)2 (aq)

precipitates are in the form of Pb(OH)2 (s). Fergusson [1990] reported that, in the pH range 8

to10, the dominant species is PbOH+, which is replaced at about pH11 by Pb(OH)2 (aq). At

about pH11.5, negative species such as Pb(OH)3- occur, which is also supported by the

electrophoretic mobility results in this paper (see Fig. 6.18). In the present work, the pH of the

feedstock solution ranges from about 11 to 13.7; therefore, the structure of the as-coprecipitated

feedstock could be envisaged as being hydrated zirconia-titania gels bonded with unreacted

acetate (OAc) and AcAc groups, together with surface-adsorbed Pb(OH)3- or ultrafine

Pb(OH)2(s).

6.3.2.2. In-situ transformation

Based on the experimental observations in the present work, in the presence of 0.4 M KOH as a

mineraliser, the initial hydrothermal products are basically amorphous as shown by XRD, with a

clustered structure (as shown by TEM), that is less agglomerated than the starting precursor

owing to the breakdown of the Ti-O-Ti and/or Zr-O-Zr bridging bonds under the hydrothermal

conditions obtaining when the synthesis temperature is below 200°C. As the synthesis

temperature is increased, perovskite PZT phase begins to form, and it appears from the TEM

micrograph (Fig. 6.17) that the PZT particles form directly from the amorphous precipitates via

an in-situ structural rearrangement. Since no soluble Pb species were detected by ICP in the

filtrate solution of the hydrothermal products at this stage, and given that the solubility of zirconia


134

and titania gel is very low as reported by Ohba et al. (1996) and shown by the present ICP data,

the following reaction previously proposed for the ABO3 perovskite formation [Kiss et al., 1966]

can be ruled out:

A2+ (aq) + B(OH)62- (aq) → ABO3 (s) + 3H2O (l) (6.10)

where A2+ and B4+ are the alkaline earth and transition metal cations respectively.

Sengupta et al. [1995] reported that molecular level heterogeneity, rather than homogeneity, is

the most significant feature of the amorphous PZT gels, since Ti-O-Ti and Zr-O-Zr linkages are

formed preferentially through the homo-condensation of M-OH and M-OR groups,

heterometallic Ti-O-Zr bonding comprises only a small fraction of the network system. The Pb

cations do not participate in the bonding with the Ti and Zr atoms because Pb atoms only attain

their regular coordination geometry, i.e. PbOn polyhedra, where n approaches 12, as required in

the perovskite structure, at a higher temperature. Instead, they occupy random positions within

the amorphous zirconia-titania gels. During hydrothermal synthesis, the PZT nuclei are formed

mostly via an in-situ transformation process, in which the relatively mobile Pb species become

incorporated within the zirconia-titania gel network, and eventually form the long-range ordered,

i.e. crystalline cubic morphology perovskite PZT structure. The role of the mineraliser in this

process can be envisaged as that of a template, rupturing the Ti-O-Ti and/or Zr-O-Zr bonds

during the hydrothermal treatment. Obviously, the smaller the mineraliser cation radius, the more

easily it will diffuse into the gel network, causing bond rupture; hence, the more effective it will be

in assisting the nucleation of PZT. Since the ICP data show that only a proportion of the

mineraliser cation, e.g. K+ content is found in the hydrothermal solution (see Fig. 6.19), it may be


135

postulated that the remaining mineraliser cations (and OH-) are adsorbed on the amorphous

zirconia-titania gel. During hydrothermal heating, these K+ or Na+ ions, on entering the gel, cause

chemical changes involving the rupture of the Ti-O-Ti and/or Zr-O-Zr bridging bonds via

dehydration as shown in (6.11). Note that Zr-O-Zr bridging bond can undergo a similar change

as in (6.11).

(6.11)+ KOH(aq) + H2O

At the same time, the residual organic groups (AcAc or OAc) continue to be replaced by

hydroxyl groups, leading finally to heterogeneous nucleation within the gel. The possible formation

process is shown schematically in Fig. 6.22.

PbO Ti or Zr

OH or H2O O (oxo or hydroxo bridges)

AcAc, OAc or OPr

Pb


136

Fig. 6.22. Schematic representation of the perovskite PZT particle formation process.

During hydrothermal synthesis, as the synthesis temperature is increased from 200°C to 250°C,

the hydrated lead oxide precursor Pb(OH)2 (s) and the titania precursor gel can dissolve partially

and transform to stable tetragonal PbO (litharge) [Cheng et al., 1994] and to TiO 2 in both the

anatase and rutile structures [Watson et al., 1987], respectively, as shown in Fig. 6.16 (a) and

(b). At the same time, non-stoichoimetric, equiaxed PZT (T) and PZT (R) particles are formed

in-situ as shown by the XRD and TEM results. As the synthesis temperature is increased further

to 280°C, PbO (ss) is formed as a result of partial replacement of Pb2+ by Ti4+ ions via

dissolution/recrystallisation. This feature is also reported by other authors during the hydrothermal

synthesis of PZT powders [Kutty et al., 1984]. Since PbO (ss) is a solid-solution of PbO-TiO2

with a TiO2 content < 10 mol % [Takai et al., 1983], it is formed preferentially when excess PbO

is used, as shown in Fig. 6.16 (c).

Since the tetragonal PbO phase (litharge) has a layer structure with a interlayer distance c of

0.5023 nm [Wells, 1975], the space between its layers could be large enough to intercalate the

small tetragonal TiO2 unit (rutile: a = 0.4594 nm; anatase: a = 0.3785 nm, c = 0.9514 nm) to

form PbTiO3, but not enough for the larger ZrO2 unit (cubic: a = 0.5070 nm; tetragonal: a =

0.5070 nm, c = 0.5150 nm; monoclinic: a = 0.5145 nm, c = 0.5311 nm, β = 99°14’) [Wyckoff,

1965]. In addition, the zirconium atom has a tendency to adopt coordination numbers higher than

six [Laaziz et al., 1992]; thus it is likely that the initial zirconia-titania gel contains titanium with a


137

coordination number of six and zirconium with a coordination number greater than six. The in-situ

transformation of this mixed gel to a perovskite structure requires a greater degree of ordering of

the lead and zirconium cations than of the lead and titanium cations because the coordination

number of the zirconium has to be reduced to six. This process is likely to be slow, especially

when the mineraliser concentration and the synthesis temperature are low. Increasing the

mineraliser concentration will increase not only the mineraliser cation number density, which

facilitates not only the breaking-down of the network together with the rearranging and re-

ordering of the zirconium and titanium atoms, but also the degree of the supersaturation of Pb.

Thus, diffusion of the Pb cation into the network is enhanced; consequently, both PZT (T) and

PZT (R) can form.

It must be pointed out that the homogeneous perovskite PZT structure may be formed only if

homogeneity of the zirconium and titanium atom distribution is achieved at a reasonably fine scale,

although not necessarily at the molecular level, through either the pre-mixing of the zirconia and

titania gel as in the two-step derived feedstock as shown in Fig. 6.3 (b), or the

dissolution/recrystallisation process occuring at higher mineraliser concentration as shown in Fig.

6.3 (b), or through a prolonged reaction time at a minimum temperature of 200°C as shown in

Fig. 6.5. Otherwise, single-cation oxide (e.g. TiO2 and PbO) or dual-cation oxide (e.g. PbTiO3)

will always form from the heterogeneous mixtures as soon as the diffusion distances (related to

temperature) are of the order of the heterogeneity scale. Upon increasing the homogeneity, the

diffusion-controlled reactions may convert to nucleation-controlled reactions because a


138

metastable multicomponent structure may crystallise only if homogeneity is achieved due to limited

diffusion crystallisation [Barboux et al., 1995].

6.4. Particle Size and Morphology Control of PZT Powders

6.4.1. Crystal Growth of PZT Particles

The crystal growth stage, which follows crystal nucleation, can be rate-controlled by either the

transport of polynuclear complexes to the growing particles or by their subsequent binding to the

particle surface depending on the pH and ionic strength of the hydrothermal solution [Segal,

1993].

It is well known that the surfaces of the oxide particles becomes hydrated in an aqueous solution,

resulting in a net surface electrostatic charge whose magnitude and sign depends on the solution

pH. The resultant electrostatic potential field will repel like-charged ions but attract unlike-

charged ions within the vicinity of the particles’ surface, increasing their local concentration

relative to that in the bulk solution in order to preserve electroneutrality. A diffuse electric double

layer is thus formed around each particle [Hunter, 1981]. If two particles approach each other,

their double layers interpenetrate, generating a repulsive force between them. However, the van

der Waals attractive potential is also operative, although over a much smaller distance from the

particles’ surface. The interaction of these long-range repulsive and short-range attractive

potentials is described approximately by the Derjaguin, Landau, Verwey, and Overbeek (DLVO)

theory [Israelachivili, 1991]. The attractive van der Waals potential is offset by the repulsive


139

double-layer potential. The resulting net interaction potential is shown in Fig. 6.23. The various

forms of the interaction potential that can occur between two colloidal particles under the

combined action of both the electrostatic repulsive and van der Waals attractive forces is also

shown schematically in Fig. 6.23 [After Israelachivili, 1991]. Coagulation can occur on increasing

the ionic strength of the solution and/or decreasing the surface electrostatic potential of the

particle.

Increasing saltconcentration,decreasingsurface potential

Double-layerrepulsion force Energy barrier

Secondary minimum(Vs)

Primary minimum(Vp)

Distance, nm

van der Waalsattraction force

Energybarrier

Secondaryminimum

Primary minimum

Resultant interaction energy

Fig. 6.23. Schematic particle interaction energy versus particle surface separation distance curve

according to DLVO theory [After Israelachvili, 1991]. (a) Surfaces repel strongly; small colloidal

particles remain ‘stable’. (b) Surfaces come into stable equilibrium at the secondary minimum if it

is deep enough; colloids remain ‘kinetically stable’. (c) Surfaces come into the secondary


140

minimum region; colloids coagulate slowly. (d) At the ‘critical coagulation concentration’, surfaces

may remain in the secondary minimum region or move closer together and adhere; colloids

coagulate rapidly. (e) Secondary minimum region absent; colloids coalesce immediately.

In the presence of a mineraliser at its critical concentration for perovskite PZT formation (e.g. 0.3

M NaOH or 0.4 M KOH), the hydrothermal solution pH, although measured at room

temperature, decreases markedly during the nucleation process at temperatures above 250°C,

owing to the consumption of KOH or NaOH as the residual OAc groups are removed by being

replaced by hydroxyl groups and the mineraliser cations enter the gel network. The pH values of

the filtrate solution extracted at 250, 280 and 300°C are 10.2, 8.8, and 7.8, respectively (Fig.

6.21). Interestingly, the isoelectric point (IEP) pH of the hydrothermal PZT products also

decreases as the hydrothermal synthesis temperature increases with the IEP pH values of the

hydrothermal PZT products extracted at 250, 280 and 300°C being 9.8, 8.2 and 6.3,

respectively (see Fig. 6.18). Thus, considering the ratios of the IEP pH to the room temperature

pH of the filtrate solution extracted at 250, 280 and 300°C of 0.96, 0.93 and 0.81, it is probable

that the PZT nuclei, as soon as they are formed, have a very strong tendency to coagulate

throughout the synthesis process. Therefore, crystalline PZT particles with a particle size of about

100 to 200 nm are formed almost instantaneously at the minimum mineraliser concentration

necessary for perovskite PZT formation as observed by TEM.

complete monolayer

surface nucleus monolayer growth


141

(a) (b)

Fig.6.24. Mononuclear (a) and polynuclear (b) growth models (After Pierre, 1991).

On the other hand, when the mineraliser concentration is low, the degree of supersaturation is

relatively low. Hence, mononuclear growth will dominate (see Fig. 6.24 (a)) [Pierre, 1991]. Most

of the particles, therefore, grow layer by layer, and the particle surfaces will be locally smooth at

the molecular level. They thus take on a well-defined cubic shape, which corresponds to the basic

crystal structure of the perovskite PZT, which is a cube with titanium or zirconium atoms lying at

the centre, eight lead atoms occupying the corners and six oxygen atoms being at the surface

centres (see Fig. 2.4). The minimal dissolution and recrystallisation occurs during the later stages

according to the ICP and TEM results. For a particle aggregate to be stable, the primary particles

will each have to adopt the same crystallographic orientation [Aksay et al., 1996]. During the

prolonged processing time, the aggregate surface begins to smooth out through the rearrangement

of surface species to form smoother particle surfaces and a more uniform particle size distribution,

as shown in Fig. 6.10 to Fig. 6.12. From a thermodynamic viewpoint, the rough primary particles

with a smaller radius of curvature have a higher rate of dissolution and so dissolve preferentially

and recrystallise in the neck regions [Iler, 1986].

However, in the presence of excess mineraliser (e.g. 0.5 M NaOH or 0.8 M KOH), the pH in

the hydrothermal solution decreased only slightly with increasing hydrothermal temperature (see


142

Fig. 6.21). This minimal decrease in pH may be the indication that there are enough OH- and Na+

or K+ ions present to remove the OAc groups and break-down ZTO network, respectively,

without the solution pH dropping significantly. Also, because of the minimal decrease in the

solution pH, the Pb concentration in solution is maintained at a relatively high level, increasing the

degree of supersaturation of Pb as the temperature increases (see Fig. 6.18). These two factors

together facilitates the rapid formation of perovskite PZT particles without the extensive formation

of undesired intermediate phases so long as the synthesis temperature is 250°C or above (see

Fig. 6.16 (b)). On the other hand, as the mineraliser concentration increases, the ionic strength of

the hydrothermal solution increases. The primary PZT nuclei become colloidally unstable, and

hence, prone to rapid coagulation, since the interaction potential barrier is reduced to a minimum

by the salt ions, i.e. “ion crowding” occurs, causing compression of the electric double layer. The

strong van der Waals attraction between a large number of the primary particles will thereby lead

to polynuclear growth (see Fig. 6.24 (b)). The primary particles appear to aggregate or

agglomerate together to form aggregates or agglomerates of cubic morphology PZT crystallites,

as shown in Fig. 6.6 (b) and (c).

6.4.2. Controlling Parameters during Hydrothermal Synthesis

From the above discussion, it can be seen that, in order to control the particle size and

morphology of hydrothermally synthesised PZT powders, the most important parameters, i.e.

mineraliser concentration, synthesis temperature and time, should be monitored during the

hydrothermal synthesis process. The mineraliser concentration used should be the minimum


143

necessary for the specified mineraliser to ensure solely phase-pure perovskite PZT particle

formation at the given synthesis temperature, where particle growth occurs via a mononuclear

growth mechanism; this produces PZT particles with a well-defined cubic morphology. Increasing

the mineraliser concentration will lead to polynuclear growth, with the result that the PZT

particles, which are in fact agglomerates of cubic morphology PZT crystallites, exhibit an

increased particle size and a morphology that is spherical rather than cubic. Increasing the

synthesis temperature can reduce the minimum mineraliser concentration required for perovskite

PZT formation, reducing the agglomerate particle size. Prolonging the synthesis time can lead to a

more symmetric particle size distribution due to the dissolution and recrystallisation process which

occurs during the latter stage of synthesis.

However, it should also be noted that the hydrothermal environment itself has a profound

influence on the nucleation and growth processes occuring during the hydrothermal synthesis of

PZT powders. The present work has shown that some ions such as acetate anions have a

deleterious effect on the perovskite PZT formation process, as they reduce the efficiency of the

mineraliser. A compromise has to be made between promoting perovskite PZT phase formation

and preventing excessive PZT particle growth. Yet, even at the critical mineraliser concentration

for perovskite PZT formation without excessive PZT particle growth, residual organic groups are

still found in the final PZT powders as shown by the FT-IR results (see Fig. 6.13). This is

because the hydrothermal solution pH has dropped close to neutral pH at the final stage (see Fig.

6.21), leaving few hydroxyl ions in solution, and hence, the residual acetate anions that are

strongly bound to the Zr and/or Ti atoms are not eliminated. They are likely to induce lattice


144

defects during nucleation and/or form micropores within the PZT particles during the subsequent

growth process. This probably accounts for the low density (about 6.5 g/cm3) of the

hydrothermal PZT powders synthesised at the critical mineraliser concentrations, in comparison

with that of the well-crystallised PZT powers (above 7.8 g/cm3). Finally, since there is very little

Pb in the hydrothermal solutions (see Fig. 6.19), the hydrothermally synthesised PZT powders

should exhibit negligible Pb-deficiency, as also supported by the TEM (EDX) analysis.

6.5. Summary

The base mineraliser (type and concentration), hydrothermal environment, temperature and time

are the most important parameters in determining the phase formation, particle size and

morphology of PZT powders during their hydrothermal synthesis. Nucleation appears proceed

mostly via in-situ transformation from the amorphous zirconia-titania gel and hydrated lead oxide

precursors in the presence of the critical mineraliser concentration (specific to the mineraliser

used) required for the formation of cubic morphology perovskite PZT particles. Only minor

dissolution/recrystallisation occurs, accounting for the formation of intermediate phases such as

PbO, TiO 2 and PbO-TiO2 solid solutions. This mechanism becomes more obvious when the

mineraliser concentration increases and/or during the later stages of crystal growth. The base

mineraliser type and its concentration play important roles in promoting the solubility and

rearrangement of the network by functioning as a template. The smaller the cation radius of the

base, the more efficient it is as a mineraliser. The mineraliser concentration should be above a

critical level in terms of providing both the necessary soluble Pb species and enough template ions


145

to cause disruption of the amorphous zirconia-titania gel network; thereby facilitating the diffusion

of the Pb species into the network, and hence, the in-situ nucleation of PZT. However, the

minimum mineraliser concentration is strongly dependent on both the hydrothermal temperature

and environment. Increasing the synthesis temperature and using the two-step derived feedstock,

which minimises the interference of the acetate ions, reduce the minimum mineraliser concentration

(of a specific mineraliser) required for perovskite PZT formation. Consequently, the tendency of

the primary PZT nuclei to coagulate during the particle growth stage is decreased because the

ionic strength of the hydrothermal solution is reduced. The particle size and morphology of the

PZT powders from hydrothermal synthesis can, therefore, be controlled by choosing carefully the

mineraliser type and concentration to be used with a given synthesis temperature and for a chosen

synthesis time. This will simultaneously ensure only perovskite PZT phase formation and prevent

overly rapid PZT particle growth. Perovskite PZT powders with a particle size of 0.2 to 0.3 µm

and a narrow particle size distribution have thus been synthesised hydrothermally at 300°C for 2

hours using either 0.4 M KOH or 0.3 M NaOH as the mineraliser, together with the two-step

derived feedstock.

Documents

CHAPTER 6 RESULTS AND DISCUSSION: HYDROTHERMAL … · 6.2. Effects of Hydrothermal Synthesis Conditions on Particle Size and Morphology 6.2.1. Effects of Mineraliser Concentration