Buschow K.H.J. (Ed.) Concise Encyclopedia of Magnetic and Superconducting Materials

Подождите немного. Документ загружается.

13. Three-terminal Spin Electronic

Device Variants

Ounadjela and Tiusan (1999) have proposed a neat

and versatile metallic spin-tunnel transistor with

particular relevance to three-terminal spin-tunnel

MRAM application. The device has two tunnel bar-

riers between emitter/base and base/collector respec-

tively. It operates on the SPICE principle and has

numerous variants which differ in respect of the pre-

cise location of the encoder and decoder electrodes

relative to the collector structure.

Mizushima et al. (1998, 1999) have also produced a

variety of interesting tunnel transistor designs, again

operating on the SPICE principle but, unlike the

Ounadjela/Tiusan device they comprise semiconduc-

tor collector elements.

14. Spin Electronics in Carbon Nanotubes

An unusual but prophetic venture in spin electronics

is the use of single-wall carbon nanotubes as spin

conductors (Alphenaar et al. 2000). Surprisingly it is

discovered that the spin diffusion lengths in such

tubes are many tens of nanometers and their future

application looks promising.

15. Voltage-induced Switching

Interesting predictions have been made with respect to

the simple RKKY-coupled GMR trilayer structure to

the effect that the interlayer magnetic coupling could

be changed in sign by application of sufﬁcient voltage

across the structure (You and Bader 1999). The object

of such endeavor is to produce a magnetic memory

element or switch whose transport in one dimension

may be altered by magnetic switching which is induced

not by external magnetic ﬁelds but rather by an electric

potential in n orthogonal direction. This remains to be

demonstrated experimentally.

16. Current-induced Switching

Current-induced switching of a spin electronic struc-

ture has also been predicted and experimental results

are available which appear to conﬁrm it. The prin-

ciple harks back to an idea of Luc Berger (Myers

et al. 1999) in which he examines the relationship

between current density and the torque/force on a

ferromagnetic domain wall. If a sufﬁciently large

spin-polarized current is incident normally on a

domain wall, the spin angular momentum exerts a

torque which twists the spins and hence effectively

causes the wall to move. When implemented in ferro-

magnetic nanocontacts, the effect is attractive as an

operating principle for simple MRAM memory ele-

ments (Coey 2001).

17. Quantized Conductance

When a point contact with dimensions of atomic size

is made between two metallic nanowires, the lateral

conﬁnement causes mutual distancing of the k-states

in a plane perpendicular to the direction of current

ﬂow (Ott et al. 1998). This means that only a small

number of conductance channels are below the metal

Fermi surface. By varying the contact width, the

individual conductance quanta may be observed.

Simple theory suggests that in the case of spin asym-

metric contacts, the quanta should be halved, but so

far this effect has eluded observation.

18. Measuring Spin Asymmetry

Various techniques are used for estimating the spin

asymmetry of spin electronic materials. They include

spin-resolved photoemission (Park et al. 1998), And-

reev reﬂection (Soulen et al. 1998), measurement of

the Zeeman effect of chemical potential (Sirisathitkul

et al. 2001b), and by spin tunneling to a material with

100% polarization (de Teresa et al. 1999). The last

tunneling technique relies on the control electrode

genuinely behaving as if it is 100% polarized. Merely

because its bulk bandstructure is fully polarized is no

guarantee of this. As discussed above, the polariza-

tion which magnetic electrodes manifest for tunneling

purposes is found to be a function of the magnetic

metal/insulator combination, not just of the magnetic

metal itself. Thus, it is important that the control

electrode is always accompanied by the same insula-

tor. This casts a new light also on the other meas-

urement techniques for measuring spin polarization

and shows not only the importance of establishing

exactly what they are measuring, but it also begs the

question as to what we really should be trying to

measure for practical purposes.

19. Competing Magnetotransport Phe nomena

There exists a large number of magnetotransport ef-

fects which are not strictly spin electronic in origin

but which nonetheless can display similar character-

istics. They include anisotropic magnetoresistance,

Hall effect and anomalous Hall effect.

In this category, an interesting and potentially use-

ful giant magnetotransport effect has been observed

in a ferromagnet/semiconductor structure in which

arrays of submicron ferromagnets are fabricated on a

heterostructure containing a near surface two-dimen-

sional electron gas. The magnetoresistance arises

from electrons propagating in open orbits along lines

of zero magnetic ﬁeld (Nogaret et al. 1997).

20. Measurement Artefacts

In spin electronics the currents and voltages meas-

ured are frequently small and this heightens the need

620

Magnetic Reco rding Systems: Spin Electronics

for vigilance against interference from rival transport

effect and spurious artefacts. The most obvious

sources of such trouble are radiated electrical noise

and earth loops. However there is also a more subtle

artefact which may occur in a.c. transport measure-

ments in applied magnetic ﬁeld if the ﬂying leads to

the sample are free to move. The current carrying

wires are subject to a harmonic torque in a magnetic

ﬁeld and if they are capable of mechanical displace-

ment this leads to a harmonic ﬂux variation. The

result is a parasitic voltage and hence an apparent

resistance which is proportional to the square of the

magnetic ﬁeld.

21. Applications of Spin Electronics

Applications to date of spin electronics include hard

disk read heads, advanced high density tape heads,

and various position and motion sensors based on

GMR and high frequency properties of spin elec-

tronic materials. Although not yet commercialized,

spin tunnel MRAM is at an advanced stage of de-

velopment also. More speculative uses with future

potential include applications of spin FETS to quan-

tum information processing and the radiofrequency

ferromagnetic single electron transistor.

Bibliography

Alphenaar B W, Tsukagoshi K, Ago H 2000 Spin electronics

using carbon nanotubes. Physica E 6, 848–51

Baibich M N, Broto J M, Fert A, Nguyen Van Dau F, Petroff

F, Etienne P, Creuzet G, Friederich A, Chazelas J 1988 Giant

magnetoresistance of (001)Fe/(001)Cr magnetic superlattices.

Phys. Rev. Lett. 61, 2472–5

Barnard J A, Waknis A, Tan M, Haftek E, Parker M R,

Watson M L 1992 Giant magnetoresistance observed in

single layer Co–Ag alloy ﬁlms. J. Magn. Magn. Mater. 114,

L230–4

Barnas J, Fert A 1998 Effects of spin accumulation on single-

electron tunneling in a double ferromagnetic microjunction.

Europhys. Lett. 44, 85–90

Barthelemy A, Fert A 1991 Theory of the magnetoresistance in

magnetic multilayers analytical expressions from a semiclas-

sical approach. Phys. Rev. B 43, 3124–9

Bean C P, Livingstone J D 1959 J. Appl. Phys. 30, 120

Berger L 1991 Galvanomagnetic voltages in the vicinity of a

domain wall in ferromagnetic thin ﬁlms. J. Appl. Phys. 69

(3), 1550–5

Berkowitz A E, Mitchell J R, Carey M J, Young A P, Zhang S,

Spada F E, Parker F T, Hutten A, Thomas G 1992 Giant

magnetoresistance in heterogeneous Cu–Co alloys. Phys.

Rev. Lett. 68, 3745–8

Cabrera G G, Falicov L M 1974 Phys. Stat. Sol. (b) 61, 539

Chui S T 1997 Bias dependence in spin-polarised tunneling.

Phys. Rev. B 55 (9), 5600–3

Coey J M D 2001 Magnetic peanut memory device. Irish Patent

ﬁling

Daniel D, Gregg J F, Thompson S M, Coey J M D, Fagan A,

Ounadjela K, Fermon C, Saux G 1995 Giant magnetother-

mal resistance. J. Magn. Magn. Mater. 1404, 493–4

Datta S, Das B 1990 Electronic analog of the electrooptic

modulator. Appl. Phys. Lett. 56, 665–7

de Groot R A, Mueller F M, van Engen P G, Buschow K H J

1984 Half-metallic ferromagnets and their magneto-optical

properties. J. Appl. Phys. 55 (6), 2151–4

De Teresa J M, Barthelemy A, Fert A, Contour J P, Montaigne

F, Seneor P 1999 Role of metal-oxide interface in deter-

mining the spin polarisation of magnetic tunnel junctions.

Science 286, 507–9

Devoret M H, Grabert H 1992 Single Charge Tunneling. Ple-

num, New York

Dieny B, Speriosu V S, Parkin S S P, Gurney B A, Wilhoit D R,

Mauri D 1991 Giant magnetoresistance in soft ferromagnetic

multilayers. Phys. Rev. B 43, 1297–300

Dubois S, Piraux L, George J M, Ounadjela K, Duvail J L, Fert

A 1999 Evidence for a short spin diffusion length in perm-

alloy from the giant magnetoresistance of multilayered nano-

wires. Phys. Rev. B 60 (1), 477–83

Fert A, Duvail J-L, Valet T 1995 Spin relaxation effects in the

perpendicular magnetoresistance of magnetic multilayers.

Phys. Rev. B 52, 6513–21

Fiederling R, Keim M, Reuscher G, Ossau W, Schmidt G,

Waag A, Molenkamp L W 1999 Injection and detection of a

spin polarised current in a light emitting diode. Nature 402,

787–90

Flatte M E, Byers J M 2000 Spin diffusion in semiconductors.

Phys. Rev. Lett. 84 (18), 4220–3

Gehring G A, Gregg J F, Thompson S M, Watson M L 1995

Electron spin depolarisation in granular magnetic systems.

J. Magn. Magn. Mater. 1404, 501–2

Gerritsen A N 1959 Physica 25, 489–502

Giaever I, Megerle K 1961 Phys. Rev. 122 (4), 1101

Gregg J F, Allen W, Ounadjela K, Viret M, Hehn M, Thomp-

son S M, Coey J M D 1996 Giant magnetoresistive effects in

a single element magnetic thin ﬁlm. Phys. Rev. Lett. 77 (8),

1580–3

Gregg J F, Allen W D, Viart N, Kirschmann R, Sirisathitkul C,

Schille J-P, Gester M, Thompson S M, Sparks P, da Costa V,

Ounadjela K, Skvarla M 1997 J. Magn. Mater. 175, 1–9

Hagele D, Oestreich M, Ruhle W W, Nestle N, Ebert K 1998

Spin transport in GaAs. Appl. Phys. Lett. 73 (11), 1580–2

Hammar P R, Bennett B R, Yang M J, Johnson M 1999 Ob-

servation of spin injection at a ferromagnet–semiconductor

interface. Phys. Rev. Lett. 83 (1), 203–6

Jia Y Q, Shi R C, Chou S Y 1996 Spin-valve effects in nickel/

silicon/nickel junctions. IEEE Trans. Magn. 32 (5),

Johnson M 1993 Bipolar spin switch. Science 260 (5106), 320–3

Johnson M, Silsbee R H 1985 Interfacial charge-spin coupling:

injection and detection of spin magnetisation in metals. Phys.

Rev. Lett. 55 (17), 1790–3

Julliere M 1975 Phys. Lett. A 54, 225

Kikkawa J M, Awschalom D D 1999 Lateral drag of spin

coherence in gallium arsenide. Nature 397, 139–41

Lee S F, Pratt W P, Yang Q, Holody P, Lololee R, Schroeder

P A, Bass J 1993 2-Channel analysis of CPP-MR data for

Ag/Co and AgSn/Co multilayers. J. Magn. Magn. Mater.

118, L1–5

Loraine D R, Pugh D I, Jenniches H, Kirschman R, Thompson

S M, Allen W, Sirisathitkul C, Gregg J F 2000 Effect of

silicon crystal structure on spin transmission through spin

electronic devices. J. Appl. Phys. 87 (9), 5161–3

Mizushima K et al. 1998 US Patent 5747859

Mizushima K et al. 1999 US Patent 5973334

Monsma D J, Lodder J C, Popma J A, Dieny B 1995 Perpen-

dicular hot-electron spin-valve effect in a new magnetic ﬁeld

621

Magnetic Reco rding Systems: Spin Electronics

sensor the spin-valve transistor. Phys. Rev. Lett. 74 (26),

5260–3

Mott N F 1936 Proc. R. Soc. 153,69

Myers E B, Ralph D C, Katine J A, Louie R N, Buhrman R A

Current-induced switching of domains in magnetic multi-

layer devices. Science 285 (5429), 86770

Nogaret A, Carlton S, Gallagher B L, Main P C, Henini M,

Wirtz R, Newbury R, Howson M A, Beaumont S P 1997

Observation of giant magnetoresistance due to open orbits in

hybrid semiconductor/ferromagnet devices. Phys. Rev. B 55

(24), R16037–40

Ohno Y, Young D K, Beschoten B, Matsukura F, Ohno H,

Awschalom D D 1999 Electrical spin injection in a ferro-

magnetic semiconductor heterostructure. Nature 402, 790–2

Ott F, Barberan S, Lunney J G, Coey J M D, Berthet P, de

Leon-Guevara A M, Revcolevschi A 1998 Quantised con-

ductance in a contact between metallic oxide crystals. Phys.

Rev. B 58 (8), 4656–9

Ounadjela K, Tiusan C 1999 Dispositif microelectronique a

jonctions tunnel et reseau de memoires et capteur compren-

ant de tels dispositifs. Demande de brevet Francais 9904227

Park J-H, Vescovo E, Kim H -J, Kwon C, Ramesh R, Venk-

atesan T 1998 Direct evidence for a half-metallic ferromag-

net. Nature 392, 7946

Parkin S S P, More N, Roche K P 1990 Oscillations in exchange

coupling and magnetoresistance in metallic superlattice

structures: Co/Ru, Co/Cr and Fe/Cr. Phys. Rev. Lett. 64,

23047

Piraux L, George J M, Despres J F, Leroy C, Ferain E, Legras

R, Ounadjela K, Fert A 1994 Giant magnetoresistance in

magnetic multilayered nanowires. Appl. Phys. Lett. 65 (19),

1–3

Prins M W J, Jansen R, van Kempen H 1996 Spin-polarised

tunneling with GaAs tips in scanning tunneling microscopy.

Phys. Rev. B 53 (12), 8105–13

Renard J P, Bruno P, Megy R, Bartenlian B, Beauvillain P,

Chappert C, Dupas C, Kolb E, Mulloy M, Prieur J, Veillet P,

Velu E 1996 J. Appl. Phys. 79 (8), 5270–5

Sirisathitkul C, Allen W D, Gregg J F, Sparks P D, Coey J M

D, Kirschman R, Thompson S M 2001a Spin blockaded

silicon Schottky barrier. (In preparation)

Sirisathitkul C, Allen W D, Gregg J F, Sparks P D, Coey J M

D, Kirschman R, Thompson S M 2001b Measuring spin

asymmetry via the chemical potential Zeeman shift. (In prep-

aration)

Soulen R J, Byers J M, Osofsky M S, Nadgorny B, Ambrose T,

Cheng S F, Broussard P R, Tanaka C T, Nowak J, Moodera

J S, Barry A, Coey J M D 1998 Measuring the spin polar-

isation of a metal with a superconducting point contact.

Science 282, 85–8

Tang H X, Monzon F G, Lifshitz R, Cross M C, Roukes M L

2000 Ballistic spin transport in a two dimensional electron

gas. Phys. Rev. B 61 (7), 4437–40

Tedrov P M, Merservey R 1971 Phys. Rev. Lett. 26, 192

Tiusan C, Hehn M, Ounadjela K, Henry Y, Hommet J, Meny

C, van den Berg H, Baer L, Kinder R 1999 Artiﬁcial antif-

erromagnetic tunnel junction sensors based on Co/Ru/Co

sandwiches. J. Appl. Phys. 85,

Valet T, Fert A 1993 Theory of the perpendicular magnetore-

sistance in magnetic multilayers. Phys. Rev. B 48, 7099–113

van Wees B J, et al. 2000 Comment on observation of spin

injection at a ferromagnet–semiconductor interface. Phys.

Rev. Lett. 84 (21), 5022

Versluijs J J, Bari M A, Coey J M D 2001 Magnetoresistance of

half-metallic oxide nanocontacts. in press

Yang Q, Lee S F, Henry L L, Lololee R, Schroeder P A, Pratt

W P, Bass J 1994 Diffusion length and giant magnetoresist-

ance at low temperatures. Phys. Rev. Lett. 72, 3274–7

You C-Y, Bader S D 1999 Prediction of switching/rotation of

the magnetisation direction with applied voltage in a con-

trollable interlayer exchange coupled system. J. Magn. Magn.

Mater. 195, 488–500

J. Gregg

University of Oxford, Oxford, UK

Magnetic Recording Systems: Spin Valves

The rapid increase in information storage density in

magnetic recording systems over recent years has

been made possible in part by the development of

spin-valve (SV) read heads. These rely on the phe-

nomenon of giant magnetoresistance (GMR), in

which a change in resistance is observed on applica-

tion of a magnetic ﬁeld (see Giant Magnetoresist-

ance). This article discusses the GMR effect on which

the devices are based, and also the parameters that

are important in consideration of the design of spin-

valve structures, such as the choice of materials for

the various layers in the device, and thermal stability.

The effect of microstructure and of magnetization

reversal on device operation are also discussed.

1. Introduction

A change in resistance on application of a magnetic

ﬁeld in magnetic materials gives rise to a tremendous

number of applications in the electronics industry.

One of the most spectacular examples is the applica-

tion of giant magnetoresistance in ultrahigh-density

magnetic recording technology, and one of the de-

vices based on this phenomenon is the SV structure

that has applications as read heads (Brug et al. 1996)

in hard disk storage systems or as magnetoresistive

random access memory (MRAM) elements (Daugh-

ton 1997). The rapid increase in interest in these de-

vices has been motivated by the high demand for

improvement in information storage density and per-

formance, mainly for graphics-intensive applications.

Improvements in signal processing and disk media,

plus the development of SV read heads, have allowed

the density to be increased to 50 Gb/in

with a pro-

jected obtainable density of 80–100 Gbits/in

. The

limits on memory density that can be achieved using

solid-state memory (for example DRAM) means that

there has been renewed interest in the development of

magnetic memory based on the GMR effect, namely

MRAM. In addition, MRAM would be nonvolatile.

2. The GMR Effect

GMR is observed in ﬁlms consisting of alternate

ferromagnetic (FM) and nonmagnetic layers, as a

622

Magnetic Reco rding Systems: Spin Valves

large change in resistance with applied magnetic ﬁeld.

The resistance is a function of the relative alignment

of the magnetization in the magnetic layers, and thus

of spin-dependent scattering of the conduction elec-

trons. It was ﬁrst observed for Fe/Cr multilayers with

thicknesses of a few nanometers.

(For a general description of GMR the reader is

referred to a number of review papers, for example

Dieny (1994).) The GMR ratio is deﬁned as:

DR=R ¼ðR  R

Þ=R

ð1Þ

where R

is the resistance for parallel alignment of

the FM layer magnetization directions. The maxi-

mum GMR ratio is observed when

R ¼ R

ð2Þ

where R

is the resistance of the antiparallel mag-

netization conﬁguration.

3. The Spin Valve Structure

SV structures (see, for example, review of SV struc-

tures in Portier and Petford-Long (1999), however,

offer greater possibilities for development for real

applications than multilayer structures. In their sim-

plest form, a SV consists of two ferromagnetic layers

separated by a nonmagnetic spacer layer. The mag-

netization direction (easy axis) of one FM layer is

ﬁxed by an adjacent antiferromagnetic (AF) layer

(pinning layer) through exchange coupling (see Mul-

tilayers: Interlayer Coupling). The magnetization di-

rection of the other FM layer (sense layer) can be

rotated by applying a small external magnetic ﬁeld

(of the order of a few oersteds) in the plane of the

layers. A schematic illustrating the structure of a

simple SV is shown in Fig. 1(a).

In a spin valve, unlike in a GMR exchange cou-

pling of the two FM layers across it is small. For a

spin valve, the GMR is linearly dependent on (cos y)

where y is the angle between the sense and pinned

layer magnetizations:

RðyÞ¼R

þðR

 R

Þð1  cosyÞ=2 ð3Þ

When applying the external ﬁeld perpendicular

to the easy axis, the change in resistance is linear

with low coercivity and a sense layer magnetiza-

tion reversal characterized by coherent rotation (see

Fig. 1(b)). This conﬁguration is used for sensors. The

second commonly used conﬁguration is to apply the

external ﬁeld along the easy axis direction. In this

case, two remanent states can be stabilized with either

parallel (P) or antiparallel (AP) alignment of the

magnetization of the pinned and sense layers. The

magnetization reversal is then a hysteretic reversal

process, with the P and AP states corresponding

to the minimum and maximum resistance values,

respectively. This is suitable for MRAM devices (high

and low resistance could correspond to ‘‘1’’ and ‘‘0’’

bit states). A GMR ratio of up to 23.4% has been

obtained for a SV structure at room temperature.

Despite the fact that SV materials have been in-

troduced commercially as read heads in the latest

generations of hard disk drives, a number of pro-

blems remain to be solved. These include magneto-

static coupling of the pinned and sense layers, which

leads to an offset of the GMR curve with respect to

remanence, and thus to a need for a higher operating

magnetic ﬁeld, low exchange coupling between the

pinned and pinning layer for some conﬁgurations,

and a reduction in GMR after lithographic process-

ing. Solving these problems requires understanding

the microstructure and local magnetic structure of the

SV ﬁlms.

3.1 Microstructure

To enhance the GMR ratio, several different SV

structures have been proposed, for example the ‘‘top’’

SV (see Fig. 1(a)) and the ‘‘bottom’’ SV in which the

layer order is reversed. Changing the deposition order

changes both the crystallographic texture of the ﬁlm

Figure 1

(a) schematic cross-section through a simple SV

structure, (b) schematic GMR curve for a SV with the

external ﬁeld applied perpendicular to the easy axis of

magnetization in the pinned layer (- ----) and with the

external ﬁeld applied parallel to the easy axis of

magnetization in the pinned layer (––––).

623

Magnetic Reco rding Systems: Spin Valves

and the interfacial roughness, which are of crucial

importance in controlling the GMR. Figure 2 shows

a schematic of the dual, or symmetric, SV, which

contains one central FM sense layer and two biased

FM layers. This structure has been found to increase

the GMR ratio up to 23.4% compared with a top SV

made with the same materials which gives only a 10%

GMR ratio. In all cases, the FM material used for the

pinned and sense layers can be either the same or

different, depending on the properties required. The

structures are usually protected from oxidation by a

cap layer.

One of the important requirements in operating SV

structures is to be able to reverse the magnetization of

the sense layer in a low external magnetic ﬁeld, hence

the use of soft magnetic materials such as Permalloy

(Ni

). Cobalt is also widely used and has a

higher saturation magnetization, coercivity, and mag-

netocrystalline anisotropy, particularly for the h.c.p.

structure, which tends to form for layer thicknesses

greater than 4 nm. Sense layer thickness is also im-

portant, with an increase in thickness leading to a

larger grain size and decrease in coercivity. A typical

grain size is about 20 nm for an 8 nm-thick NiFe

sense layer, which gives an optimal GMR ratio.

Another alternative for the pinned layer is to use

an artiﬁcial antiferromagnet (AAF) such as Co/Ru/

Co, in which the cobalt layers are antiferromagnet-

ically exchange-coupled across the ruthenium, and

one of the cobalt layers is pinned by an adjacent

AFM material. The net magnetization of the pinned

layer is then close to zero. SVs containing an AAF

show good sensitivity to an external magnetic ﬁeld

because stray-ﬁeld coupling between the pinned and

sense layers in lithographically deﬁned elements is

very much reduced.

The spacer layer needs to be very conductive to

reduce spin-independent scattering of the conduc-

tion electrons and is commonly copper, silver, or

gold. Copper is most widely used as its structure

matches well (1.85% mismatch) with those of cobalt

or Ni

. Intermixing between NiFe and copper

can lead to the formation of a paramagnetic interfa-

cial alloy layer, resulting in spin-ﬂip scattering and

loss of GMR, but this can largely be avoided by in-

serting a thin cobalt layer at the NiFe/Cu interface.

The GMR properties of SVs are very dependent on

the nature of the interfaces between the layers, and

three different types of interlayer coupling can arise

across the FM/NM/FM interfaces:

FM exchange coupling between the FM layers

through pinholes in the spacer, although this is

largely avoidable for spacer layers thicker than about

1.5 nm.

Magnetostatic ‘‘orange-peel’’ coupling which

arises because of the waviness of the FM layer/spac-

er layer interface leading to free magnetic poles of

opposite sign across the spacer layer, for which Ne

proposed a model relating the period and the ampli-

tude of the waviness to the coupling strength.

An oscillatory Ruderman-Kittel-Kasuya-

Yosida (RKKY-like) coupling which is dependent

on the spacer layer thickness and decays with in-

creasing thickness in an oscillatory manner between

FM and AFM coupling. The spacer layer thickness is

usually chosen to be high enough that the RKKY

coupling is negligible, but if it is too high, increased

spin-independent scattering of the conduction elec-

trons and increasing shunting of the current through

the spacer layer lead to a decrease in GMR. The op-

timum thickness has been found to be B2.5–3 nm.

One of the key parameters for a SV device is the

strength of the exchange coupling (H

) between the

antiferromagnetic (pinning) and pinned layers (see

Nogue

s and Schuller 1999) for a review of exchange-

biasing effects). The exchange ﬁeld is set by applying

a magnetic ﬁeld during deposition or postdeposition

anealing in order to align the magnetization of the

ﬁrst monolayer of the AFM material with that of the

adjacent FM layer. The coupling strength depends on

the AFM material used and on its microstructure,

with the most commonly used materials being FeMn,

NiMn, and IrMn, although other materials such as

CoO, NiO, TbCo, a-Fe

, and other manganese al-

loys are being considered. An increase in H

and in

the pinned-layer coercivity are observed when reduc-

ing the thickness of the FM material, and an increase

of H

and of the blocking temperature (the temper-

ature above which no exchange ﬁeld is observed)

have been linked to an increase in grain size.

Figure 2

Schematic of dual SV structure.

624

Magnetic Reco rding Systems: Spin Valves

The surface topology is also important since a

smoother interface between the pinning and pinned

layers tends to improve H

The ﬁrst systems to be studied used FeMn because

it is structurally very similar to the FM and spacer

layers. However, H

(B150 Oe) and the blocking

temperature (B 150 1C) are relatively low and its cor-

rosion resistance is not very high (this can be im-

proved by adding chromium, iridium, platinum,

rhodium, or ruthenium during deposition). The most

promising AFM material to date is MnNi with H

up to 500 Oe after annealing, a blocking temperature

of B380 1C, a reasonable corrosion resistance, and

very good thermal stability. The as-grown MnNi is

cubic and nonmagnetic, but an anneal at about

220 1C for 1 hour leads to a phase transformation to

the antiferromagnetic face-centred-tetragonal (f.c.t.)

phase. A difﬁcult task is to determine the real pro-

portion of MnNi grains with the f.c.t. structure com-

pared to those still with the f.c.c. structure.

The most common deposition techniques used for

fabricating SVs are ion-beam deposition (IBD), DC

magnetron sputtering, and MBE. MBE-grown ﬁlms

tend to be epitaxial and single crystal, whereas the

sputtered ﬁlms are polycrystalline. Films grown by

IBD usually have higher resistivities and lower GMR

than DC magnetron-sputtered ﬁlms because of

greater intermixing between the layers. Polycrystal-

line ﬁlms tend to show better GMR properties than

MBE-grown ﬁlms, but a good /111S texture be-

tween the pinning and the pinned layers is preferable

for a strong pinning effect. In addition, the coercivity

of the sense layer is lower for a well-textured ﬁlm,

which enables a rapid transition between the P and

AP states for MRAM applications.

To increase the grain size and degree of texture

in the ﬁlm, a seed layer such as niobium, tantalum,

titanium, or zirconium is commonly used. Tantalum

is the most widely used, as it induces the /111S tex-

ture for most of the various FM materials used as

the sense layer resulting in a columnar growth from

bottom to top in the SV structure. In addition, the

tantalum/FM layer interface usually has a good ther-

mal stability. However, an intermixed, magnetically

‘‘dead’’ layer of a few monolayers is often observed

between the seed and the FM layer. Note that for

bottom SV structures it is sometimes necessary to

deposit a thin FM layer below the AFM material to

obtain a good crystallographic texture.

4. Thermal Stability

The GMR amplitude decreases linearly with temper-

ature by a factor of about three from 4 K to room

temperature. The main effect is magnon scattering of

the conduction electrons in the bulk of the FM layers

which has been correlated to the Curie temperature of

the FM material, but another contribution will be

phonon scattering in the NM layer preventing the

exchange of conduction electrons between the sense

and pinned layers. Operating a SV structure at ele-

vated temperatures thus leads to a decrease of the

GMR effect, but the critical temperature above which

the linear dependence is lost depends strongly on the

structure used.

The critical temperatures are much lower than in

bulk materials because of the numerous grain bound-

aries whose diffusion coefﬁcients are much higher than

the bulk values, favoring layer interdiffusion at lower

temperatures. In addition, changes to the interface

between the seed layer and the sense layer can degrade

the GMR effect following annealing, because of the

formation of a high-resistivity alloy at the seed layer/

FM interface. The cobalt/copper system is more

thermally stable than other systems such as NiFe/

copper, for example, because cobalt/copper is a phase-

separating system and, because the Curie temperature

of cobalt is high in comparison with NiFe, although

the GMR properties are not as good.

5. Micromagnetic Considerations

5.1 Sense-layer Reversal

One of the important parameters for the operation of

a SV is the magnetization-reversal mechanism of the

sense layer, because of its inﬂuence on the GMR re-

sponse of the SV. If the applied ﬁeld is along the easy

axis direction, the reversal of the sense layer magnet-

ization occurs via domain wall motion for NiFe, and

by magnetization rotation mechanisms for cobalt,

with a higher reversal ﬁeld for the cobalt-based SVs.

This is expected because of the higher saturation

magnetization and magnetocrystalline anisotropy of

cobalt. However, some other features can be seen,

which are directly attributable to the FM interlayer

coupling. First, the domain structure for a NiFe-

based SV is more complex than for a single NiFe ﬁlm.

In addition, domain walls are observed to lie at an

angle with respect to the easy axis direction because

of the residual torque associated with coupling to the

pinned layer.

Third, the P–AP and AP–P reversal are not sym-

metric, with more magnetization rotation observed

for the AP–P reversal. This is a result of the combi-

nation of the unidirectional anisotropy imposed by

the exchange ﬁeld of the AFM layer on the pinned

FM layer and the FM ‘‘orange-peel’’ magnetostatic

coupling between the pinned and sense layers.

A further important feature is that the hysteresis

loop is shifted with respect to zero ﬁeld, and there are

a number of terms that contribute to this effect. The

ﬁrst of these is the orange-peel coupling (H

), which

is FM in nature. Second, there is a magnetic ﬁeld

induced by the current passing through the device

), whose effect depends on the direction of current

ﬂow and its magnitude.

625

Magnetic Reco rding Systems: Spin Valves

A simple model based on Ampere’s law:

¼ 0:75Ið4p  10

3

Þ=2h ð4Þ

where I is the current and h is the height (width) of

the element, provides good agreement with experi-

mental observations. Third, magnetostatic stray-ﬁeld

coupling occurs at the edges of a lithographically

deﬁned SV element because of the demagnetizing

ﬁeld (H

) between the pinned and sense layers. This

phenomenon becomes of primary importance as the

element size is reduced and is strongly dependent

on element size and shape. The stray-ﬁeld coupling

is AFM in nature, and H

can be simply estimated

using:

¼ð4pMt=hÞu

ð5Þ

where M and t are the saturation magnetization and

thickness of the sense layer, respectively, h is the

width of the element, and u

is a unit vector in the

easy axis direction. A possible expression for the off-

set ﬁeld (H

off

), by which the GMR curve is displaced

from zero, can then be given as:

off

¼ H

þ H

ð6Þ

As mentioned previously, a decrease in sense-layer

thickness and/or the use of an AAF as the pinned

layer lead to a substantial reduction of the stray ﬁeld

coupling. Flattening the FM/NM/FM layer inter-

faces, for example by the use of a surfactant such as

indium, can reduce the orange peel coupling. H

off

should be kept as low as possible to reduce the ex-

ternal ﬁeld that must be applied to reverse the sense

layer magnetization.

Finally, 3601 domain walls are often observed,

mainly at the end of the P–AP transition, which

consist of a pair of very close domain walls separated

by a very narrow region of ﬁlm in which the mag-

netization in the two FM layers is still in the P

conﬁguration.

These defects remain up to a much higher applied

ﬁeld value and act as domain nucleation sites for the

opposite reversal, which can result in differences in

reversal ﬁeld for different devices with nominally ex-

actly the same structure. Applying an external ﬁeld

perpendicular to the easy axis results in a magneti-

zation reversal of the sense layer via coherent mag-

netization rotation with a single domain behavior.

Assuming that the magnetization of the pinned

layer remains parallel to the easy axis during the

reversal, the change in resistance is only due to the

change in magnetization direction in the sense layer.

This assumption is based on the fact that the ex-

change bias ﬁeld is usually much higher than the

reversal ﬁeld of the sense layer. An important issue is

thus the uniformity of the pinned layer magnetiza-

tion. Nonuniformity can result in different responses

from nominally identical devices, since the GMR is

determined by the misorientation between the mag-

netization directions of the FM layers, and can be

caused by variations in the exchange coupling

imposed by the adjacent AFM layer, for example

because of incomplete transformation to the f.c.t.

phase for MnNi.

The operating temperature of devices such as read

heads in hard-disk systems can be of the order of

150 1C, so the thermal behavior of SV ﬁlms is an

important issue. An increase in temperature results

in a reduction in both the ﬁeld value and the ﬁeld

range over which the sense layer magnetization

reverses. This is believed to be a result of the in-

crease in thermal energy which compensates the effect

of the FM interlayer coupling. If the temperature

becomes too close to the blocking temperature, then

becomes zero, leading to simultaneous magnet-

ization reversal of the two FM layers and loss of the

GMR effect.

Current density through the SV is an important

parameter in operating SV elements since it imposes

the amplitude of the signal to be read for distin-

guishing the P and AP remanent states. A typical

value is of the multilayer, the spacer layer thickness is

such that the order of 10

Acm

2

and a decrease in

GMR and in coercivity are usually observed with

increasing current density. The current can also have

a more subtle effect on the sense layer magnetization

reversal because the increase in thermal energy sup-

plied to the device can counteract the coupling be-

tween the sense and pinned layers. There will also be

a current density limit above which the ﬁlm resistance

increases dramatically and, consequently, above

which the element begins to be damaged by heating

effects. The critical current density depends on the

size and shape of the element.

See also: Magnetic Recording Heads: Historical Per-

spective and Background; Magnetoresistive Heads:

Physical Phenomena; Thin Films: Domain Forma-

tion

Bibliography

Brug J A, Anthony T C, Nickel J H 1996 Magnetic recording

head materials. MRS Bull. 21 (9), 23–8

Daughton J M 1997 Magnetic tunneling applied to memory.

J. Appl. Phys. 81 (8), 3758–63

Dieny B 1994 Giant magnetoresistance in spin-valve multi-

layers. J. Magn. Magn. Mater. 136, 335–59

Nogue

s J, Schuller I K 1999 Exchange bias. J. Magn. Magn.

Mater. 136, 203–32

Portier X, Petford-Long A K 1999 Electron microscopy studies

of spin-valve structures. J. Phys. D: Appl. Phys. 32, R89–108

A. K. Petford-Long

University of Oxford, Oxford, UK

626

Magnetic Reco rding Systems: Spin Valves

Magnetic Recording Technologies:

Overview

Magnetic recording technology is the key element of

large-scale information storage and retrieval. In ad-

dition to its obvious importance in science and en-

gineering, it has become indispensable to our daily

lives. When we make a bank transaction, reserve an

airplane ticket, use a credit card, or watch a movie

from a videotape, we are using the technology of

magnetic recording. The general requirements for

data storage are information integrity, fast access,

and low cost. In its simplest form, a magnetic re-

cording system consists of a magnetic head and a

magnetic medium, as schematically shown in Fig. 1.

The head is made of a piece of magnetic material in a

ring shape, with a small gap facing the medium and a

coil away from the medium.

The head records and reproduces information,

while the medium stores the information. The re-

cording process is based on the phenomenon that an

electric current i generates a magnetic ﬂux j as de-

scribed by Ampere’s law. The ﬂux j leaks out of the

head core at the gap, and magnetizes the magnetic

medium. The orientation of the magnetization is de-

termined by the current direction. As the medium

moves from left to right with a velocity V under the

head gap, magnetization is retained in the medium at

different locations.

There are two types of head used in reproducing.

One senses magnetic ﬂux change rate, and the other

senses the magnetic ﬂux. During the reproduction

process, a magnetized medium is placed under the

head gap. A part of the ﬂux coming out of the me-

dium surface enters the head through the gap. As

the medium moves, the prerecorded information is

reproduced by the head.

The distance between two closest transitions in the

medium is the ﬂux change length B, and the distance

between two adjacent data tracks is the track pitch

W, which is wider than the data track width w. The

ﬂux change length can be directly converted into bit

length with the proper coding scheme. The recipro-

cals of the bit length and the track pitch are termed

the linear density and the track density, respectively.

Their product yields the areal density of the infor-

mation stored in the medium. This areal density is a

ﬁgure of merit for a recording system.

The research and development efforts of the mag-

netic recording industry have been focused on in-

creasing the areal density. During the 1990s, the areal

density was increasing at 60% annual rate. More re-

cently, the rate has been accelerated to more than

100%. At present, hard disk drives used in desktop

computers feature areal densities of more than

30 Mb mm

2

(Bo0.06 mm and Wo0.6 mm). This

yields a total storage capacity of more than 30 Gb

for a disk of 95 mm diameter.

1. Magnetic Media and Magnetic Heads

Magnetism is the result of uncompensated electron

spin motions in an atom. Only transition elements

exhibit this property, and nearly all practical interest

in magnetism centers on the ﬁrst transition group of

elements (manganese, iron, nickel, and cobalt) and

the rare earth elements (gadolinium, and terbium).

While the ﬁrst transition group of elements and their

alloys are the materials used in magnetic recording,

amorphous rare earth-transition metal alloys are the

only commercially used magneto-optical storage me-

dium. The strength of magnetism is represented by

magnetization M, and is related to magnetic ﬁeld H

and magnetic ﬂux density B by

B ¼ m

ðH þ MÞð1Þ

where m

is the permeability of vacuum. The ratio of

B with and without a magnetic material is the relative

permeability m of that magnetic material.

When a magnetic ﬁeld H is applied to a piece of

demagnetized magnetic material, the magnetization

M starts increasing with H from zero. The rate of

increase gradually slows down and M asymptotically

approaches a value M

.IfH is reduced to zero, M

follows to a lower value M

. Further reduction of H

to a very high negative value will magnetize the ma-

terial to M

. In order to bring the material to the

demagnetized state, a positive ﬁeld H

is required.

Continuous increase in H ﬁeld will bring the trace of

M to a closed loop. This loop is the major hysteresis

loop, as illustrated in Fig. 2.

The hysteresis loop shows that magnetic material

has memory. It is this memory that is used in the

medium for storing information. H

is the coercivity,

indicating the strength of magnetic ﬁeld required to

erase the memory of a magnetic material. Magnetic

materials with high H

are ‘‘hard’’ magnets, and are

Figure 1

Conceptual diagram of a magnetic medium and an

inductive head.

627

Magnetic Reco rding Technologies: Overview

suitable for recording media applications if they also

have high M

. On the other hand, magnetic materials

with low H

are ‘‘soft’’ magnets, and are candidates

for head core materials if they also have high M

and

high m. M

and M

are the remanent and saturation

magnetization, respectively, and their ratio is the

remanent squareness.

There are two groups of magnetic media. The ﬁrst

group is called particulate media because the mag-

netic materials are in the form of particles. This group

includes oxide (g-Fe

), cobalt-modiﬁed iron oxide

(g-Fe

þCo), metal particles, and barium ferrite

(BaFe

). Some of these have been used in mag-

netic recording for several decades. The other group

of media is called the thin-ﬁlm media, because the

magnetic layer can be deposited as a continuous thin

ﬁlm. Most materials in this group are cobalt-based

alloys. Compared with particulate media, the thin-

ﬁlm media usually have a higher coercivity H

and a

higher remanence M

, and are suitable for high areal

density applications.

Magnetic media can be classiﬁed into three general

forms of applications. Tape is the oldest form and

remains an important medium today. It is central to

most audio, video, and instrumentation recording,

although it is also used in the computer industry for

archival storage. It is economical and has a large ca-

pacity, but suffers slow access time. Hard disk is tra-

ditionally the primary storage unit inside a computer,

and is just being introduced in the consumer market

(audio/video box). It provides the user with fast data

access, but has poor transportability. Flexible disk is

designed for easy data removability, but is limited in

capacity. Besides these three general forms of appli-

cations, a hybrid of ﬂexible medium and hard disk

has appeared recently. It is a removable rigid disk

capable of holding up to several gigabytes of digital

data. In addition, magnetic stripes are getting widely

used in various forms of cards.

The magnetic layer alone cannot be used as a me-

dium. It needs additional components to improve its

chemical and mechanical durability. Typical cross

sections of a particulate magnetic tape and a thin-ﬁlm

hard disk are shown in Fig. 3. In the case of tape

application, iron particles with a typical size of 0.5 mm

long and 0.1 mm wide are dispersed in a polymeric

binder, together with solvents, lubricants, and other

ﬁllers to improve magnetic and mechanical durability.

This dispersed material is then coated on a biaxially

oriented polyethylene terephthalate substrate. An op-

tional back coat may also be applied to the other side

of the substrate. The cross section of a hard disk is

more complex. A high-purity aluminum–magnesium

(5 wt.%) substrate is diamond-turned to a ﬁne surface

ﬁnish, and then electrolessly plated with a nonmag-

netic nickel–phosphorus (10 at.%) undercoat.

This layer is used to increase the hardness, reduce

the defects, and improve the ﬁnish of the aluminum–

magnesium alloy, and is polished to a super surface

ﬁnish. Next, an underlayer of chromium is sputtered

to control the properties of the magnetic ﬁlm, fol-

lowed by sputtering the magnetic ﬁlm. Finally, a layer

of hydrogenated or nitrogenated carbon is over-

coated on the magnetic ﬁlm, and an ultrathin layer of

perﬂuorinated hydrocarbon liquid lubricate is ap-

plied on top. The carbon and lubricant layers are

used to improve the corrosion and mechanical resist-

ance of the disk. To achieve high shock resistance,

glass disk is sometimes used, especially in mobile ap-

plications. For a 95 mm disk, the ﬁnished product

should have a surface ﬂatness better than 8 mm and an

arithmetic average roughness (R

) of less than 0.3 nm.

Magnetic heads serve three functions: recording,

reproducing, and erasing. Two groups of head are

employed to carry out these three functions. The

Figure 2

Hysteresis loop of a magnetic material.

Figure 3

Cross-sections of a particulate magnetic tape (top) and a

thin-ﬁlm hard disk (bottom).

628

Magnetic Reco rding Technologies: Overview

ﬁrst group is the inductive heads, as schematically

illustrated in Fig. 1. The evolution of the inductive

head follows the selection of core materials, as shown

in Fig. 4. Early heads used laminated molybdenum

Permalloy (Ni–Fe–Mo, 79–17–4 wt. %). The primary

drawbacks are frequency limitation, gap dimension

inaccuracy, mechanical softness, and frequency lim-

itation caused by eddy current loss. One way to re-

duce eddy current loss is to increase core material

electric resistivity. Two types of ferrite material have

high resistivity (four to nine orders higher than Perm-

alloy) and reasonable magnetic properties: nickel–

zinc and manganese–zinc.

These materials are also very hard, providing great-

er head durability during head/medium contacts. The

major deﬁciency of ferrite materials is their low B

values. In order to record in high H

media, high ﬂux

density B is needed in the head core. When the ﬂux

density in the core material reaches its saturation B

,it

will not increase despite the increase of recording cur-

rent or coil turns. This saturation starts from the cor-

ners of the gap due to its geometry. To remedy this

deﬁciency, a layer of metallic alloy material with much

higher B

is deposited on the gap faces. This type of

head is called the metal-in-gap (MIG) head.

Thin-ﬁlm heads capitalize on semiconductor-like

processing technology to reduce the customized fab-

rication steps for individual heads. The core, coil, gap,

and insulator layers are all fabricated by electroplat-

ing, sputtering, or evaporation. Due to the nature of

the semiconductor process, the fabrication is more

accurate for small dimensions. Small gap dimensions

are suitable for high linear and track density, and

small core dimensions allow the use of high B

Perm-

alloy material (nickel–iron, 80–20 wt. %) as a core

with low inductance for high data-rate applications.

The high cost of the semiconductor-like process is

offset by high throughput: a 150 mm 150 mm wafer

can produce 40 000 pico-slider heads (1.25 1.00 

0.30 mm). One disadvantage is the limited band-

recording capability because the small pole length

limits low-frequency response. A second disadvantage

is the Barkhausen noise, which is characterized as the

sudden jump of magnetic ﬂux ﬂowing around the

head, and is caused by the relatively small number of

magnetic domains in the core.

The recording and erasing functions are performed

by an inductive head in all magnetic recording devices.

The reproducing function, however, can be carried out

either by an inductive head or by a magnetoresistive

(MR) head. The reproducing performance of an in-

ductive head is described by Faraday’s equation:

v ¼nV

ð2Þ

where v is the output electric voltage from the head, n

is the number of the coil turns, V is the medium mov-

ing velocity, and dj/dx is the derivative of the mag-

netic ﬂux with respect to medium moving direction.

When an inductive head is used for reproducing,

this head is usually also used for recording and eras-

ing. The optimal performance cannot be achieved

because recording and reproducing have contradic-

tory requirements for head design. To solve this

problem, a new reproduction head was introduced in

the 1990s. This is the MR head. An MR head is al-

ways integrated together with an inductive head in

such a way that the MR head performs the repro-

ducing and the inductive head carries out the record-

ing and erasing.

The ﬁrst generation of MR head introduced in

1990 is the anisotropic type (AMR). As shown in

Fig. 5, conceptually, an AMR head has a magneto-

resistive element (MRE) and two electric leads. The

MRE is a Permalloy stripe (nickel–iron, 80–20 wt.%),

with thickness t, width w, and height h. An electric

current, with density J, passes through the MRE

through the leads. The electric resistivity of the

MRE is a function of the angle y between J and MRE

magnetization M:

¼ r 1 þ

cos



ð3Þ

Figure 4

Various inductive head structures: (a) laminated, (b)

MIG, and (c) thin ﬁlm.

629

Magnetic Reco rding Technologies: Overview