Moulson A.J., Herbert J.M. Electroceramics: Materials, Properties, Applications

Подождите немного. Документ загружается.

way the birefringence of the plate is modiﬁed across its area according to the

incident light intensity distribution. The stored image can be read non-

destructively with monochromatic light used in conjunction with the usual

polarizer and analyser arrangement.

An improved display exploits scattering in a PLZT of similar composition to

that described above but with the grain size increased to about 4.5 mm. This

device has the advantage that it eliminates the need for the polarizer and analyser

and the straining arrangement; also, white light can be used for both the ‘write’

and ‘read’ operations.

The ceramic plate, ITO electrodes and photoconductive layer are arranged as

in the previous example. In its unpoled state the PLZT scatters light, probably at

the 718 and 1098 domain boundaries associated with the rhombohedral structure.

When a voltage pulse is applied simultaneously with a light pattern incident on

the plate, the polarization is switched to a direction normal to the plane of the

plate and the scattering is reduced.

The photoconductive ﬁlm is ﬁrstly uniformly illuminated so that the PLZT

plate can be poled to saturation remanence. The image to be stored, for

example from a transparent photographic ‘positive’, is then focused on the

photoconductive ﬁlm while a voltage of opposite polarity to that of the poling

voltage is applied. This pulse is suﬃcient to bring the ferroelectric to the

electrically depoled and optimum scattering state. Erasure of the stored image

is achieved by uniformly illuminating the area and repoling the plate to

saturation remanence.

462 ELECTRO-OPTIC CERAMICS

Fig. 8.17 Image storage device: (a) ferroelastic polarization; (b) ﬁeld-induced polarization.

TEAMFLY

Team-Fly

The stored image can be read by passing collimated white light through the

plate and then focusing the unscattered light to pass through a 2 mm aperture

and onto a ﬁlm. The PLZT plate is positioned in one focal plane of the focusing

lens and the aperture in the other. Images can be formed with tones of grey and a

resolution of approximately 80 lines mm

The techniques are capable of storing sequential information by using a raster-

scanned laser beam with appropriately timed and sized voltage pulses.

8.3.5 PLZT ¢lms

PLZT in thin and thick ﬁlm form can be exploited for a variety of applications

including optical signal switching and modulation. An attraction of PLZT is its

high electro-optic coeﬃcient and in consequence low drive voltages to perform a

particular function (c.f Table 8.1). The PLZT thin ﬁlm switch described below is

included to illustrate principles rather than because of very signiﬁcant

commercial importance.

There are various thin-ﬁlm optical switch principles based on interference,

diﬀraction or total internal reﬂection (TIR). The TIR switch, the essentials of which

are shown in Fig. 8.18, depends on TIR of light passing from one medium to

another of lower refractive index. TIR occurs for small diﬀerences in refractive

indices when the angle of incidence approaches 908, i.e. for light almost parallel to

the surface of the material with the lower refractive index.

A suitable substrate is sapphire with the PLZT ﬁlm deposited on an (0001)

face. Because of the smallness of the mismatch (about 2%) between the O–O

distances (287 pm in an (0001) face for sapphire and approximately 280 pm in a

(111) face for PLZT) the PLZT grows epitaxially, with its (111) face registering

with the sapphire (0001) face (c plane).

Typical dimensions of the sputtered PLZT ﬁlm are as follows: thickness t

about 0.35 mm; light guide step height t

about 0.05 mm; light guide step width t

APPLICATIONS 463

Fig. 8.18 Total internal reﬂection thin-ﬁlm optical switch (after K. Wasa et al. (1984) J.

Lightwave Techn. IEEE., LT-2, 710–14).

about 20 mm; angle f between guides about 28. The metal electrodes are of length

l about 1.7 mm and interelectrode gap g about 4 mm. There is a buﬀer layer of

between the electrodes and the PLZT to prevent the metal aﬀecting the

propagation of light in the guide.

With no voltage applied to the electrodes a signal entering at 1 will exit at 3;

with a suﬃciently large voltage applied TIR occurs in the region between the

electrodes, causing the signal to be switched to exit at 4. TIR occurs because the

electro-optic eﬀect gives rise to a reduction in refractive index for the guided

light, whose E vector is parallel to the short dimension of the guide.

It is readily shown that the critical angle y

for TIR is given by

¼ sin

1

ð1 

Þð8:50Þ

where the change in refractive index causing TIR is regarded as being due to the

Kerr eﬀect.

In the case of LiNbO

(n ¼ 2.29), which exhibits the linear Pockels eﬀect

¼ 17.5 pm V

), with an assumed y

of 888 (on the basis of the assumed 28 angle

between the guides), E is calculated to be 3.3 MV m

which, for a gap of 4 mm, is an

applied voltage of 13 V. If a 9/65/35 quadratic PLZT is assumed, then reasonable

corresponding values are R ¼ 400 nm

, n ¼ 2.5 and a voltage of 1.4 V.

Thin PLZT ﬁlms deposited by, for example, sputtering are too thin (51 mm) to

achieve the necessary retardation for transverse mode optical devices. There is,

however, potential for ﬁlms having thicknesses in the range typically 2–25 mm. If

these can be successfully and economically produced then they oﬀer potential for

a variety of devices including optical shutters, modulators and displays.

The substrate on which thick ﬁlms are grown also has to be chosen with care

since it inﬂuences grain orientation in the ﬁlm, and particularly mechanical

stresses arising because of thermal expansion mismatch. To avoid ﬁlm cracking

the ﬁlm has to be maintained in compression requiring the substrate thermal

expansivity to be greater than that of the PLZT.

The ﬁlms can be grown by various methods (see Section 3.6.9). G.H. Haertling

and G.C. Robinson [10] describe the production of 6 mm thick (9/65/35) PLZT

ﬁlms on sapphire substrates by an automated repeated dipping/ﬁring process.

The dipping solution consisted of the acetates of the metal ions, La, Zr, Ti and

Pb. The ﬁlms were the basis of a successful shutter and display based on the

principles described above (see Sections 8.21 and 8.33).

8.4 Optical Non-linearity in Glass and

Glass-ceramics

Glass in the form of the optical ﬁbre plays an unrivalled passive role in

communications’ technology. Many messages, in the form of a string of light

pulses generated by solid state (e.g. GaAs- or InP-based) lasers, are multiplexed

464 ELECTRO-OPTIC CERAMICS

and sent along a single ﬁbre; at the receiving end there has to be an optical-to-

electronic signal conversion using photodetectors and the messages demulti-

plexed by electronic switching.

At the present time the fastest electronic devices operate in the picosecond

range (10

12

s), slow in comparison to the duration of optical pulses which can be

as short as a femtosecond (10

15

s). The communications systems engineer

describes the use of electronics for signal processing as the ‘electronics

bottleneck’.

Optically non-linear crystals (e.g. LiNbO

) provide a means of controlling

optical signals but they have some disadvantages, including relatively high cost

resulting from crystal growth, cutting and polishing operations, signal losses at

interconnections with ﬁbres caused by thermal expansion mismatch, and the high

voltages needed to drive them. The possibility and attractiveness of exploiting

glass ﬁbres as ‘active’ devices for the control of signals, as well as for carrying

them, has stimulated research interest.

Glasses have the advantage of high transparency over a wide frequency range,

excellent ‘forming’ characteristics and, of course, total joining compatibility with

the ﬁbre network. However, the apparently macroscopically isotropic nature of

glass would be expected to rule out the possibility of second harmonic

generation, as discussed above. Third harmonic generation is expected, and

second harmonic generation does occur in optical ﬁbres irradiated with intense

laser light. Non-linearity in glasses and the potential it has for optical modulators

and switches is reviewed up to about 1990 by E.M. Vogel [11] and by E.M. Vogel

et al. [12].

In 1991 R.A. Myers et al. [13] reported SHG in an electrically poled glass. A

strong (2.5 MV m

1

) electric poling ﬁeld was applied across a 1.6 mm silica

glass plate at 300 8C for 15 min and maintained during cooling to room

temperature. Second harmonic generation was observed with an eﬃciency of the

same order as that achievable from LiNbO

, the incident infrared laser light

emerging with a green component. The second harmonic component was found

to originate in a thin layer of the silica glass close to the anode. The observation

led to a rapid growth of interest in poling glasses, particularly the chalcogenide*

glasses, to induce non-linearity.

K. Tanaka et al. [14] found SHG in tellurium-containing glasses conﬁrming

that in this case the light originated throughout the volume of the glass and

suggesting that the SHG is probably caused by local polarizations involving

TeO

and TeO

structural units.

A further signiﬁcant development is the observation by T. Fujiwara et al. [15]

of enhanced SHG achieved through a combination of ultraviolet laser irradiation

and electrical poling. A glass of composition 0.157GeO

–0.843SiO

, poled at a

ﬁeld strength of 30 MV m

1

and simultaneously irradiated with ultraviolet laser

OPTICAL NON-LINEARITY IN GLASS AND GLASS-CERAMICS 465

* Glasses containing the chalcogenide elements, sulphur, selenium and tellurium.

light, developed strong non-linearity, the measured d-coeﬃcient (3.4 pm V

1

)

being larger than the corresponding value for LiNbO

(2.6 pm V

1

). J. Khaled

et al. [16] observed similar eﬀects in sputtered thin (1 mm) glass

(GeO

:SiO

¼1:1) ﬁlms. The origins of the non-linearities are not clear but are

thought to involve defects and associated local electric ﬁelds. The eﬀects decay

with time but it seems likely that as understanding improves the decay rates will

be reduced to acceptable values.

H.G. Kim et al. [17] observed SHG in a transparent tellurite-based glass-

ceramic (see Section 3.7.4) and Y. Ding et al. [18] in thin (0.35 mm) surface layers

of LiNbO

having partially aligned crystallites and grown in a base glass of

composition 0.35Li

O–0.3Nb

–0.35SiO

. In addition to low voltage operation

(for optical switches, for example) a further advantage oﬀered by thin ﬁlms is that

the phase-matching requirement discussed above is relaxed.

The roles glass-based devices can play in photonics is well outside the scope of

the present text and are dealt with comprehensively elsewhere [e.g. 11,12].

However, in addition to the eﬀects mentioned above, the following two examples

give some indication of the potential.

Second harmonic radiation has been generated along chemically modiﬁed (e.g.

Ge- and P-doped) photo-irradiated silica ﬁbres [12]. The high second harmonic

conversions (5%) along ﬁbres as long as 30 cm require the SHG light to be

phase-matched. The mechanism(s) by which phase-matched SHG can be

generated in such long ﬁbres is not fully understood but involves a periodic

distribution of photo-induced defects (for example trapped electrons associated

with dopant-induced oxygen vacancies) along the ﬁbre length at just the

periodicity demanded for phase-matching.

Very recently J.D. Mills et al. [19] have ‘written’ a nano-structured Bragg

reﬂection grating within the volume of silica glass using focused infrared laser

light. Such photo-induced structures have potential for a variety of photonics

applications, including information storage.

There seems little doubt that the future will see glass playing an increasingly

important ‘active’ role in communications and related technologies, in addition

to its dominant role for carrying information along ﬁbres.

Problems

1. A light wave of E vector amplitude E

is incident at an angle y

to the interface

between two transparent media with refractive indices n

and n

. For E vectors in the

plane of incidence the reﬂection coeﬃcient r

cos y

 n

cos y

þ n

cos y

466 ELECTRO-OPTIC CERAMICS

where y

is the angle of refraction. The corresponding coeﬃcient when the E vector is

perpendicular to the plane of incidence is

cos y

 n

cos y

þ n

cos y

These are the Fresnel equations.

A parallel beam of light is incident normally on a thin slab of PLZT (n ¼ 2.5).

Ignoring absorption, what fraction of the incident intensity is transmitted through the

slab? [Answer: 67%]

2. (a) The slab referred to in Question 1 has a thickness of 300 mm and an absorption

coeﬃcient at the particular wavelength (800 nm) of 65 m

. Calculate the total

transmitted intensity taking absorption into account. (b) Calculate the thickness and

refractive index of suitable antireﬂection coatings that might be applied to the PLZT.

[Answer: 65%; 200 nm; 1.58]

3. A parallel-sided PLZT plate is positioned between two crossed Polaroid plates and the

major faces of all three elements are normal to a parallel beam of monochromatic light

(l ¼ 750 nm). The thickness of the PLZT plate is 1 mm and the electrodes, spaced

1 mm apart, are arranged such that a uniform electric ﬁeld can be applied through the

volume of the PLZT plate parallel to its major faces and at 458 to the transmission axis

of each Polaroid plate.

Calculate the voltage that must be applied between the electrodes to achieve

maximum light transmittance through the system. Estimate the transmittance

assuming 5% loss at each Polaroid surface and assuming that the PLZT element

carries antireﬂection coatings and the material has an absorption coeﬃcient of

65 m

. The PLZT is quadratic with R ¼ 4.0  10

16

1

and n ¼ 2.5. [Answer:

346 V; 38% transmitted]

4. If in Problem 3 V

is the voltage necessary to produce maximum transmittance, draw

the time variation of the average transmitted intensity when a voltage of V

sin(10pt)is

applied.

5. Discuss the requirements for optical transparency in polycrystalline ceramics and

outline a route designed to achieve transparency.

6. Explain why it is that a small addition of La

to BaTiO

leads to semiconductivity

whereas a similar addition to PbTiO

increases electrical resistivity.

7. Compile a list of the various methods for producing thin ﬁlms of PLZT on a suitable

substrate pointing out the merits or otherwise of each.

PROBLEMS 467

Bibliography

1. Haertling, G.H. (1988) Electro-optic ceramics and devices. In Electronic Ceramics,

L.M. Levinson (ed.), Marcel Dekker, New York, pp. 371–92.

2. Herbert, J.M. (1982) Ferroelectric Transducers and Sensors, Gordon and Breach,

London.

3. Aguillo

-Lo

pez, F. et al. (1994) Electrooptics, Academic Press, London.

4. Lines, M.E. and Glass, A.M. (1977) Principles and Applications of Ferroelectrics and

Related Materials, Clarendon Press, Oxford.

5. Burfoot, J.C. and Taylor, G.W. (1979) Polar Dielectrics and their Applications,

Macmillan, London.

6. Wood, R.M. (1993) Optical materials, The Institute of Materials, London.

7. Margulis, W. et al. (1998) Poled glasses, Mater. Res. Soc. Bull., November, 30–5.

8. Yariv, A. and Yeh, P. (1984) Optical Waves in Crystals, John Wiley and Sons, Inc.,

New York.

9. Lengyel, B.A. (1966) Introduction to Laser Physics, John Wiley and Sons, Inc., New

York.

10. Haertling, G.H. and Robinson, G.C. (1998) PZT and PLZT thick ﬁlms on silver,

sapphire and silicon. In Integrated thin ﬁlms and applications. Pandey, R.K. et al.

(eds.) Ceramic Transactions 86, pp. 15–30.

11. Vogel, E.M. (1989) Glasses as non-linear photonic materials, J. Am. Ceram. Soc., 72,

719–24.

12. Vogel, E.M., Weber, M.J. and Krol, D.M. (1991) Nonlinear optical phenomena in

glass, Phys. Chem. Glasses, 32, 231–54.

13. Myers, R.A., Mukherjee, N. and Brueck, S.R.J. (1991) Large second order

nonlinearity in poled fused silica, Opt. Lett., 16, 1732–4.

14. Tanaka, K. et al. (1996) Optical second harmonic generation in poled MgO–ZnO–

TeO

and B

–TeO

glasses, J. Non-Cryst. Solids, 203, 49–54.

15. Fujiwara, T. et al. (1997) Second-harmonic generation in germanosilicate glass poled

with ArF laser irradiation, Appl. Phys. Lett., 71, 1032–4.

16. Khaled, J. et al. (2000) Generation of second harmonics in Ge-doped SiO

thin ﬁlms

by ultraviolet irradiation under poling electric ﬁeld, J. Appl. Phys., 87, 2137–41.

17. Kim, H.G. et al. (1996) Transparent tellurite-based glass-ceramics with second

harmonic generation, J. Non-Cryst. Solids, 208, 303–7.

18. Ding, Y. et al. (1995) Second order optical non-linearity of surface crystallised glass

with lithium niobate, J. Appl. Phys., 77, 2208–10.

19. Mills, J.D. et al. (2002) Embedded anisotropic microreﬂectors by femtosecond-laser

nanomachining, Appl. Phys. Lett., 81, 196–8.

468 ELECTRO-OPTIC CERAMICS

MAGNETIC CERAMICS

Over the past 60 years ceramic magnets have become ﬁrmly established as

electrical and electronic engineering materials; most contain iron as a major

constituent and are known collectively as ‘ferrites’.

The science and technology of ferrites have their roots in the eﬀorts of many

in the early part of the last century. Signiﬁcant contributors are S. Hilpert [1] in

Germany, H. Forestier [2] in France and Y. Kato and T. Takei [3] in Japan.

The remarkable discovery that the addition of non-magnetic zinc ferrite

(ZnFe

) to a magnetic ferrite caused an increase in magnetism was explained

in 1948 by L. Ne

el’s [4] theory of uncompensated antiferromagnetism, or

‘ferrimagnetism’.

During the 1930s the Philips researchers in The Netherlands made important

contributions to the science and technology of ferrites, especially the pioneer,

J.L. Snoek who, in the West, saw the industrial signiﬁcance of ferrites as

telecommunications was rapidly advancing. The development of the magneti-

cally ‘hard’ hexagonal ferrite (‘Ferroxdure’) is also due to Philips [5].

The classic text by J. Smit and H.P.J. Wijn [6] reviews the subject up to the late

1950s and the authoritative paper by M. Sugimoto [7] surveys the past, present

and future of ferrites from the perspective at the end of the 20th century. From

the point of view of electrical properties they are semiconductors or insulators, in

contrast to metallic magnetic materials which are electrical conductors. One

consequence of this is that the eddy currents produced by the alternating

magnetic ﬁelds which many devices generate are limited in ferrites by their high

intrinsic resistivities. To keep eddy currents to a minimum becomes of

paramount importance as the operating frequency increases and this has led to

the widespread introduction of ferrites for high-frequency inductor and

transformer cores for example. Laminated metal cores are most widely used

for low-frequency transformers. This single example illustrates how metal and

ceramic magnetic materials complement each other; often it is physical properties

Electroceramics: Materials, Properties, Applications. 2nd Edition. Edited by A. J. Moulson and J. M. Herbert.

& 2003 John Wiley & Sons, Ltd: ISBN 0 471 49747 9 (hardback) 0 471 49748 7 (paperback)

which determine choice, but sometimes it is cost. Ferrites dominate the scene for

microwave applications, and the transparency required for magneto-optical

applications is oﬀered only by them. Ferrites have also become ﬁrmly established

as the ‘hard’ (or permanent) magnet materials used for high-ﬁdelity speakers and

small electric motors, to mention just two mass-produced components.

9.1 Magnetic Ceramics: Basic Concepts

9.1.1 Origins of magnetism in materials

Ampe

re, Biot, Savart and Oersted were among the ﬁrst to demonstrate that

conductors carrying currents produced magnetic ﬁelds and exerted ‘Lorentz’

forces on each other. They were also responsible for determining the laws

governing the magnetic ﬁelds set up by currents. It was established that a small

coil carrying a current behaved like a bar magnet, i.e. as a magnetic dipole with

magnetic moment l (Fig. 9.1), and this led Ampe

re to suggest that the origin of

the magnetic eﬀect in materials lies in small circulating currents associated with

each atom. These so-called amperian currents each possess a magnetic moment

ðl ¼ IA), and the total moment of the material is the vector sum of all individual

moments. The amperian currents are now identiﬁed with the motion of electrons

in the atom.

Both the orbital and the eﬀective spinning motions of the electron have

associated angular moments quantized in units of h

¼ 1:055  10

34

Js.Itisan

elementary exercise in physics to show that the relationship between the magnetic

dipole moment l and the angular momentum L for a moving particle of mass m

and charge Q is

l ¼ gL ð9:1Þ

in which g ¼ Q=2m is the gyromagnetic ratio. Therefore there is a unit of

magnetic moment, corresponding to the quantum unit of angular momentum, in

terms of which the moment of the spinning or orbiting electron is measured. This

unit is the Bohr magneton (m

¼ðe=2m

Þh

) which has the value 9.274610

724

470 MAGNETIC CERAMICS

Fig. 9.1 The magnetic moment of a current loop.

It follows that once the total angular momentum of an ion, atom or molecule

is known, so too is its magnetic moment. Most free atoms possess net angular

momentum and therefore have magnetic moments, but when atoms combine to

form molecules or solids, the electrons interact so that the resultant angular

momentum is nearly always zero. Exceptions are atoms of the elements of the

three transition series which, because of their incomplete inner electron shells,

have a resultant magnetic moment.

There are two major contributions to the magnetic moment of an atom – from

the orbiting electrons and from their spin. There is also a nuclear magnetic

moment, but because it is of the order of 10

of that of the Bohr magneton, it is

insigniﬁcant in the present context and will be disregarded.

The description of the orbital and spin states of an electron in terms of the

quantum numbers n, l, m

and s, and the calculation of the total angular

momentum of a number of electrons in an atom in terms of J are outlined in

Chapter 2. From that discussion, together with the above, it follows that an

electron with orbital quantum number l possesses a total magnetic moment

flðl þ 1Þg

1=2

, and this can be oriented with respect to a magnetic ﬁeld in

directions such that the component of magnetic moment in the ﬁeld direction is

also quantized with values m

The electron also behaves as if it had a component of magnetic moment of 1m

in the direction of an applied ﬁeld, resulting from its spin angular momentum s,

and the total magnetic moment is l ¼ 2fs(s þ 1Þg

1=2

. A consequence of the

factor 2 is that the gyromagnetic ratio g for a system of electrons possessing both

orbital and spin angular momenta must be written g ¼ gðe=2m

Þ, in which the

Lande

g factor is given by

g ¼

SðS þ 1ÞLðL þ 1Þ

2JðJ þ 1Þ

ð9:2Þ

and can take values from 1 to 2. For many ferrites g is close to 2, indicating that

it is the total spin magnetic moment which makes the dominating contribution to

magnetization; the orbital magnetization is said to be ‘quenched’. The later

discussions are concerned with the net magnetic moment of the partly ﬁlled 3d

electron shell of atoms of the ﬁrst transition series elements where, because of

quenching, only the electron spin moments need be considered (see Chapter 2,

p. 10).

9.1.2 Magnetization in matter from the

macroscopic viewpoint

The ﬁeld vector determining the Lorentz force on a current is the magnetic

induction B, which is measured in teslas. In principle, B at a point P can be

MAGNETIC CERAMICS: BASIC CONCEPTS 471