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

Подождите немного. Документ загружается.

technologies in general can be found elsewhere, for example in the monograph

edited by W. Gerhartz [12] and in recent overviews [13,14].

During the few decades prior to 1980 the square magnetic hysteresis loop

ferrite ‘cores’ dominated computer memory technology. The binary information

was determined by the sense of the magetization around the core. They were

non-volatile and intrinsically ‘radiation hardened’ and still used in 1990 for

specialized applications, for example military, where those attributes were

essential. They have been totally displaced by the semiconductor memory

oﬀering much higher storage density and speed. It is interesting that FeRAMS

(see Section 5.7.5), the ferroelectric analogue oﬀering high storage density, high

access speed and non-volatility, are now manufactured commercially.

In the conventional tape-recorder the recording medium consists of small

needle-shaped particles dispersed in an organic binder and supported on a

ﬂexible polymer ﬁlm. The traditional recording technology is ‘longitudinal’, that

is the induced changes in magnetization occur in the place of the ﬁlm. Most tapes

incorporate particles of g-Fe

(maghemite, which has a defect spinel structure)

that are synthesised chemically in the form of needles typically 0.2 to 0.5 mm long

with aspect ratios in the range 5 to 10. The needles are single domain and the

magnetic anisotropy arises as a result of the diﬀerent demagnetizing ﬁelds along

the length and perpendicular to it (‘shape anisotropy’). The spontaneous

magnetization lies along the length of the needle and the coercive force is

typically in the range 50 to 100 kA m

1

The record/read head, positioned close to or in contact with the recording

surface, is essentially a gapped soft electromagnet (e.g. Mn-Zn ferrite) positioned

so that the fringing ﬁeld in the gap intersects the magnetic particles in the tape.

The signal fed to the head coil modulates the ﬁeld which, in turn, switches the

magnetization in the particles. In play-back mode the modulated magnetization

in the tape induces voltages in the coil as the tape moves relative to the head. The

‘bit’ length, the minimum distance between magnetization reversals, is about

1 mm, corresponding to one half wavelength in analogue recording.

Perpendicular recording, which involves magnetization changes normal to the

plane of the tape or disc, has attracted considerable interest because of the higher

storage density oﬀered than with longitudinal recording. This is because the

magnetic state of a particular bit is not subjected to the demagnetizing ﬁeld from

a neighbouring bit of opposite polarity, whereas in the case of longitudinal

recording it is, placing a lower limit on bit spacing. Small (typically

0.080.030.1 mm

) barium hexaferrite particles are arranged to lie with

their c-axes (the preferred direction of magnetization, or ‘easy’ direction; see

Section 9.2.2) perpendicular to the plane of the tape [12]. Although the

technology has not been adopted the interest stimulated the development of

routes to preparing suitable powders. These included both ‘mixed oxide’

followed by milling, and ‘chemical’, including ‘hydrothermal’ [15]. A rather novel

route involved fusing the barium and iron oxides with boric oxide to form a glass

532 MAGNETIC CERAMICS

TEAMFLY

Team-Fly

from which the barium hexaferrite particles were precipitated, the glass being

removed by acid dissolution.

Pure iron particles, coated to inhibit oxidation, are now used to an increasing

extent for data storage and may ultimately totally displace the ferrites.

The hard disk of a computer exploits thin magnetic metal ﬁlms with a present-

day data storage capacity of about 10 Gb in

2

whereas in the case of ﬂexible tape

memory systems incorporating magnetic particles the capacity is about 100 times

smaller.

Applications exploiting magneto-optical e¡ects

Magneto-optical recording exploits the Kerr eﬀect which describes rotation of

the plane of polarization on reﬂection from a magnetic material. Both the Kerr

and Faraday eﬀects (the latter describing rotation of the polarization plane on

passing through a transparent magnetic material) are exploited in optical signal

processing.

If a thin transparent layer of anisotropic magnetic garnet is viewed through a

crossed polarizer and analyser system in a microscope, a pattern of the type

shown in Fig. 9.50(a) is seen. This pattern is due to the Faraday rotation eﬀect

which was discussed in Section 9.3.4. Domains magnetized in a direction normal

to the ﬁlm rotate the plane of polarization of incident light in one sense while

those magnetized in the reverse direction rotate it in the opposite sense, so that

the domain pattern is observable in the light emerging from the analyser. If a

ﬁeld is now applied normal to the ﬁlm, one set of domains expands while the

other shrinks as magnetization in a single direction becomes established

(Fig. 9.50(b)). There is a stage before magnetization is complete when the

remaining domains, which have their magnetizations in the opposite direction to

APPLICATIONS 533

Fig. 9.50 The various domain structures observable in epitaxial iron garnet ﬁlms: (a) maze,

(b) ‘bubbles’, and (c) stripe.

the ﬁeld, are in the form of short circular cylinders with their axes normal to the

ﬁlm and appear under the microscope as circular ‘bubbles’ (Fig. 9.51).

These eﬀects are shown by all magnetic materials that have uniaxial

anisotropy, and the size of the stable bubbles is inversely proportional to the

saturation magnetization; thus barium ferrite and metallic cobalt give bubbles of

diameter about 0.05 mm and 0.01 mm respectively. Garnet is a cubic structure and

therefore does not have uniaxial anisotropy in the massive form. When it is

grown as a single-crystal epitaxial layer on a substrate of diﬀering composition,

strain is induced by the mismatch between the lattice dimensions of the magnetic

garnet and the substrate. This is enhanced on cooling to room temperature by

the diﬀerence in thermal expansion between the two garnets and results in an

anisotropic axis normal to the ﬁlm. The anisotropy and magnetization of the

ﬁlm, and consequently the size of the bubbles, can be adjusted by replacing

yttrium by other rare earths and iron by other trivalent cations such as Ga

and Al. Y

2.6

0.4

1.2

3.8

is a typical formulation giving bubbles of

diameter 4–8 mm in ﬁlms of similar thickness to the bubble diameter.

In 1990 there was strong interest in data storage devices which exploited trains

of small ‘bubbles’ injected into a garnet ﬁlm; the presence or absence of a bubble

representing the binary-coded information. Bubble memories oﬀered the

advantage of non-volatility but are now commercially unimportant because of

inadequate access speed.

It is also possible to develop ‘stripe’ domains as shown in Fig. 9.50(c), the

domain geometry being determined by ﬁelds applied in the plane and

perpendicular to it. The pattern as a whole can be rotated by rotating the in-

plane ﬁeld. The stripe pattern, arranged to have an appropriate spacing by the

applied magnetic ﬁeld, can act as a diﬀraction grating, the paths of light passing

through it being controlled through corresponding changes in the applied

magnetic ﬁeld. The eﬀect has the potential for switching signals in optical

communications systems (cf. Sections 8.3.5 and 8.4).

534 MAGNETIC CERAMICS

Fig. 9.51 Schematic diagram of bubbles showing directions of magnetization.

An important application for garnets is for isolators in optical communica-

tions systems. Reﬂections from end-faces of components and interconnections in

an optical transmission line have a destabilizing eﬀect on the operation of the

laser sources and have to be eliminated. This is achieved using an optical

‘isolator’, the optical analogue (l typically in the range 1.30–1.55 mm) of the

microwave isolator described in Section 9.5.5.

An optical isolator comprises a magnetic garnet crystal plate sandwiched

between crossed polarizing elements, and a permanent magnet for applying the

necessary ﬁeld. The device allows the transmission of forward light whilst

blocking any reﬂected light. Such isolators are incorporated in the semiconduc-

tor laser modules and so are miniaturized with overall dimensions typically

363 mm. Bi-substituted iron garnet is used for the active element because of its

low saturation magnetization and large rotating power, and rare-earth magnets

are employed to apply the appropriate magnetic ﬁeld.

For one particular commercial isolator ((YBi)

(FeAl)

; precise details not

disclosed) an optical rotation of 1060 deg cm

is quoted for a wavelength of

1.55 mm which compares with 175 deg cm

for YIG (Y

). An additional

advantage oﬀered by the substituted garnet is a smaller saturation magnetization

of 60 mT compared with 178 mT for YIG.

The substituted garnet plates can be produced by slicing from a single crystal

rod grown by the ﬂoating zone method or, more economically, from thick ﬁlms

grown onto a host substrate by liquid phase epitaxy (see Section 3.11). The plates

are lapped, polished, coated with antireﬂection layers and then diced to produce

the ﬁnal elements (2 mm side by 100 mm thickness) ready for assembly into

the isolator.

There are many potential applications for thin ﬁlm magnetic garnets, including

wave-guiding in the plane of ﬁlms, and image storage and display technologies.

The monograph by A.K. Zvezdin and V.A. Kotov [16] is recommended for a

comprehensive treatment of the materials and magneto-optical applications.

9.5.5 Microwave devices

The elementary description of microwave propagation in magnetically saturated

ferrites given in Section 9.3.4 provides a basis for an understanding of the

principles of operation of devices such as circulators, isolators and phase shifters,

all of which ﬁnd extensive use in microwave engineering. For example, a

circulator might be used in a radar system where a single antenna is employed for

both transmitting microwave power and receiving a reﬂected signal, as shown

schematically in Fig. 9.52(a).

An isolator is a device which allows microwave power to be transmitted in one

direction with little attenuation while power ﬂowing in the reverse direction is,

ideally, completely absorbed. It is used to avoid interaction between the various

APPLICATIONS 535

parts of a microwave system. For example, in high-microwave-power industrial

processing systems it is essential to prevent the high-power source – a magnetron

for instance – from instability and destruction due to reﬂected power from the

load. This function is illustrated diagrammatically in Fig. 9.52(b).

An example of the use of ferrite phase shifters is in a phased-array antenna.

This comprises an array of coherently radiating dipoles, in which the microwaves

from each are controlled in phase and amplitude. The main beam emitted is

produced by interference between the radiation from the individual elements; the

phase diﬀerence between the radiation from the individual dipoles determines the

attitude of the beam in relation to the array. Stepwise changes in the phase

diﬀerences thus make it possible to control the direction of the beam

electronically. In fact the individual phase shifters can be controlled by a

computer, thus enabling a single immobile array to perform inertialess search,

acquisition and tracking functions simultaneously on a large number of targets.

These devices are described in the following section, but ﬁrst the Faraday

rotation isolator is brieﬂy discussed because it illustrates the basic ideas; at one

time it was an important device but it has now been superseded, principally by

the junction circulator.

Faraday rotation isolator

Figure 9.53 shows the essential parts of a Faraday rotation isolator. The electric

ﬁeld of the forward wave is perpendicular to the plane of the resistive card (a thin

536 MAGNETIC CERAMICS

Fig. 9.52 (a) Schematic diagram of a circulator for a radar system; (b) an isolator.

(a)

(b)

vane of absorbing material such as a coating of nichrome deposited onto a thin

glass strip) and so it passes it with little attenuation. The length and

magnetization of the ferrite and the applied ﬁeld H

are chosen so that the

plane of polarization of the radiation is rotated by 458 in the correct sense for it

to pass into the right-hand guide without reﬂection. Radiation propagated in the

reverse direction will also be rotated by 458 so that the plane of polarization is

now at 90 8 to that of the forward wave. Therefore it will not propagate in the

left-hand guide, and also the resistive card will be in the same plane as the electric

vector so that the radiation is absorbed.

Although this type of isolator has largely been superseded by other types, it

has advantages at very high frequencies over devices which necessitate the ferrite

operating at resonance: the very high applied magnetic ﬁelds required for

millimetre wave resonance isolators are avoided.

Y-junction circulator

The most frequently used ferrite device is a three-port circulator in waveguide or

stripline conﬁguration, as illustrated in Fig. 9.54. In the waveguide conﬁguration

a ferrite cylindrical disc is located symmetrically with respect to the three ports. In

the stripline conﬁguration two ferrite discs are placed on each side of the strips. In

both cases a magnetic ﬁeld is applied perpendicular to the planes of the discs.

Unlike the case of the Faraday rotation isolator, even an elementary

description of the operational principle of the circulator is not easy, involving

as it does the dimensional resonance of the microwave ﬁeld within the ferrite

cylinder. In this context the word ‘resonance’ does not signify ‘gyromagnetic

resonance’ but a standing-wave resonance determined by the dimensions of the

APPLICATIONS 537

Fig. 9.53 Faraday rotation isolator.

ferrite (cf. Section 9.3.1). In the particular resonance modes excited, the electric

ﬁeld vectors are perpendicular to the plane of the disc. The ﬁeld conﬁguration

generated can be regarded as two counter-rotating patterns which, because of the

applied ﬁeld H, do not resonate at the same frequency. This is because of the

diﬀerent permeability values depending on whether the magnetic component of

the microwave ﬁeld is rotating in the same or the opposite sense to that of the

electron precession. By appropriate choice of dimensions and biasing ﬁeld,

depending on the operating frequency, the positions of the maximum and

minimum electric ﬁeld can be made to coincide with the port positions. For

example, for power entering port 1 (Fig. 9.54(a)) the electric vector E can be

arranged to be zero opposite port 3, so that no voltage exists there and in

consequence no power leaves; because E is a maximum opposite port 2, power

passes through it. The device therefore acts as a transmission cavity with power

entering at port 1 and leaving at port 2. Because the device is symmetrical, power

entering port 2 will set up similar ﬁeld conﬁgurations resulting in power leaving

538 MAGNETIC CERAMICS

Fig. 9.54 Y circulators in (a) waveguide and (b) stripline conﬁguration.

(a)

(b)

port 3 and port 1 will be isolated. Similarly, power entering port 3 is transmitted

to port 1 and port 2 is isolated.

In addition to its use as a circulator, the device can be operated as an isolator,

by terminating one of the ports with a matched load, or as a microwave power

switch, by reversing the sense of the biasing ﬁeld.

Phase shifters

The phase of a microwave passing along a guide can be shifted by inserting pieces

of ferrite at appropriate positions. For example, in a rectangular guide along

which the simplest microwave ﬁeld conﬁguration (TE

mode) is propagating, the

magnetic ﬁeld component is circularly polarized in the plane of the broad face of

the waveguide at a distance approximately a quarter of the way across the guide;

on the opposite side of the guide the magnetic ﬁeld component is circularly

polarized in the opposite sense. This can be seen from Fig. 9.55 which shows the

instantaneous magnetic ﬁeld conﬁguration looking down on the broad face of

the guide. As the pattern propagates from left to right the ﬁeld directions at A

and B change as indicated, as the ﬁeld conﬁgurations at points 1, 2, 3 and 4 arrive

at A and B. Therefore if two slabs of ferrite are positioned in the guide and

magnetic ﬁelds, well below resonance strength, are applied as shown in Fig. 9.56,

the eﬀective path length of the wave is altered and a phase shift is introduced.

APPLICATIONS 539

Fig. 9.55 Magnetic ﬁeld conﬁguration looking down on the broad face of a waveguide

propagating the TE

mode.

It is worth noting that if the ferrites are magnetized to resonance and a

forward wave is absorbed by the ferrite, it will be relatively unaltered in the

reverse propagation direction. The device can therefore function as an isolator.

A development of the conﬁguration of Fig. 9.56(a) is to complete the magnetic

circuit using permanently magnetized ferrite pieces, as shown in Fig. 9.56(b); this

is the basis of the ‘latching ferrite phase shifter’. The static magnetization may

now correspond to positive or negative remanence. Thus the phase can be shifted

by a magnetizing pulse through wires threading the core. By using a series of

such phase shifters in the feeder to a dipole of an antenna, the overall phase shift

can be controlled over a range of steps as small as desired. The control is digital

and can therefore be computer directed, forming one element of a ‘phased-array

antenna’.

The way in which the ferrite parts are arranged in the guide is illustrated in

Fig. 9.57. Because considerable power can be handled by radar phase shifters the

540 MAGNETIC CERAMICS

Fig. 9.56 Phase shifters: (a) non-reciprocal isolating arrangement; (b) latching.

Fig. 9.57 Schematic diagram of a four-bit latching phase shifter; the driver terminals pass

current pulses along the central conductor to switch ferrite bits (cf. Fig. 9.56(b)).

inevitable heat developed has to be removed. This can be achieved by ensuring

good thermal contact between the ferrite and the metal parts. BeO and BN

ceramics can be designed in as heat sink materials.

Microwave ferrite types

Perhaps more than in most other examples that can be cited of materials

exploitation, designing a ferrite for a particular microwave application is a

compromise: a desired value for a particular parameter is generally achieved at

the expense of the optimum values of other parameters. Nevertheless, a select

number of spinel and garnet compositions have become favoured for microwave

applications.

As is evident from the discussion in Section 9.3.4, saturation magnetization is

an important parameter. In the nickel spinel family M

values ranging from 400–

40 kA m

(4pM

¼ 5000–500 G) can be tailored by substituting Zn

for Ni

to raise M

or by introducing Al

to the B sites to lower M

. Lowering M

also

increases the power-handling capability of the ferrite since it partly determines

the power threshold P

(see Section 9.3.4) of the microwave at which spin-wave

resonance sets in. Nickel ferrites are used to best eﬀect for applications at X-band

frequencies and above and, because of their large spin-wave linewidth, for

handling high peak power levels. Undesirable features of the nickel spinels are

their magnetic and dielectric loss factors which are higher than acceptable for

some applications.

The magnesium ferrites have M

values ranging from 200 to 40 kA m

(4pM

¼ 2500–500 G). Small additions of Mn

are found to suppress the

formation of Fe

which would increase dielectric losses due to electron

hopping. As in the case of nickel spinels, the addition of Al

, which goes mainly

into the B sites, leads to a reduction in M

The lithium spinels Fe

(Li

1þ

0:5

3þ

2:5

have M

values ranging from 380 to

32 kA m

(4pM

¼ 4800–400 G). They have exceptionally high Curie tempera-

tures – up to about 670 8C – and so the M

values are relatively insensitive to

temperature under normal operating conditions. M

values can be reduced by the

substitution of Ti

into B sites, with additional Li

for charge compensation.

Coercivity can be tailored by the addition of Zn onto A sites which reduces

anisotropy; zinc addition also increases M

. Addition of Co

increases the

power-handling capability (see Sections 9.1.7 and 9.3.1). Care has to be exercised

in the sintering to prevent the formation of Fe

and consequent poor dielectric

properties; small quantities (about 1%) of Bi

are added to reduce sintering

temperatures and so alleviate the problem. Except for the low insertion loss,

where the garnets have a clear advantage, the lithium spinels are regarded as

having superior properties for most situations.

APPLICATIONS 541