Soft magnetic FeCo alloys

Soft magnetic FeCo alloys: alloy development, processing, and properties R. S. Sundar{ and S. C. Deevi* FeCo intermetallic alloys are well known to possess a unique combination of high saturation magnetisation, high Curie temperature, low magnetocrystalline anisotropy and good strength. They are ideally suited for applications requiring high flux densities. Intense research and development efforts over the years resulted in significant improvements in mechanical properties and better understanding of phase transformations in FeCo alloys. However, the high cost of the alloys has prevented their widespread applications. Emerging needs for soft magnetic materials to operate at high temperatures have caused renewed interest in FeCo alloys. The earlier studies are summarised in this review with special emphasis on phase transformations, processing, mechanical, magnetic and electrical properties of binary, ternary and more complex FeCo alloys. Current interest in developing nanocrystalline and thin films based on FeCo compositions is also summarised. The research areas that need to be addressed are also emphasised to clearly elucidate the effects of alloying additions and processing on the electrical and magnetic properties and on the high temperature mechanical properties of these alloys. Keywords: FeCo soft magnetic alloys, Magnetic properties, Mechanical properties, Nanocrystalline material, Thin films, Intermetallic compound, Metallic glasses, Processing

Introduction Soft magnetic materials constitute an important class of engineering materials1,2 since they can be easily magnetised and demagnetised under a small external field. In general, soft magnetic materials are characterised by high initial permeability in the range of 1.1 to 100 000 and low coercivity in the range of 0.4–1000 A m21. Their magnetic properties are critical to applications involving power generation and distribution, actuator, magnetic shielding, data storage, and microwave communication. Soft magnetic alloys can broadly be divided into three categories, namely metallic alloys, intermetallics and ceramics. Some of the commercially available soft magnetic materials are iron, low carbon steels, iron– silicon alloys, ferritic steels, iron–nickel alloys, iron– cobalt alloys, soft ferrites and soft magnetic amorphous alloys.2 Selection of soft magnetic material for a given application is based on attainment of the optimum combination of properties such as saturation magnetisation, permeability, coercivity, electrical resistivity, mechanical strength, and more importantly, cost. Continuous improvement in magnetic, mechanical and electrical properties of the existing alloys, and discovery of new materials led to severe competition between the

Chrysalis Technologies Incorporated Richmond, VA 23237, USA { Present address: Research Center, Philip Morris USA, Richmond, VA 23234, USA. *Corresponding author. Present address: RD&E Center, Philip Morris USA, 4201 Commerce Road, Richmond, VA 23234, USA, email [email protected]

ß 2005 Institute of Materials, Minerals and Mining and ASM International Published by Maney for the Institute and ASM International DOI 10.1179/174328005X14339

soft magnetic materials for a given application. Among the soft magnetic materials, electric steels lead in terms of market share (Table 1).2 However, alloys based on intermetallic compounds are well known for their superior soft and hard magnetic properties.3,4 Alloys based on the intermetallic compound FeCo exhibit the highest saturation magnetisation5–7 among the commercial magnetic materials (Table 2). In addition, they exhibit high Curie temperatures, good permeability, good strengths, and are ideally suited for applications requiring high flux density.8 However, their widespread applications are restricted as a result of the high cost of Co. Currently, FeCo alloys are used in applications where their high saturation values provide an advantage in reducing weight or volume of the components. For example, a careful design of components with FeCo alloys over standard Fe–Si alloys may result in weight savings of 20–25%.9 Some of the current applications of FeCo alloys are listed in Table 3. A worldwide initiative is under way to develop a new generation of aircraft called ‘More Electric Aircraft’ (MEA), wherein electric power will be utilised to drive aircraft subsystems which are currently run by a combination of hydraulic, pneumatic, electric, and mechanical power transfer systems.10,11 Utilisation of electric power to drive various subsystems will greatly increase the aircraft reliability, maintainability, and supportability and greatly reduce the need for ground support systems. The key technologies such as the integrated power generating unit, an internal starter/ generator for the main propulsion engines and magnetic bearings of MEA require a soft magnetic material which

International Materials Reviews

2005

VOL

50

NO

3

157

Sundar and Deevi

Development, processing and properties of FeCo alloys

can meet the demands of high temperature service conditions. Some of the property requirements are higher induction (.2 T) in the temperature range of 500–600uC, good thermal stability for a minimum period of 5000 h, and a core loss less than 480 W kg21 at 5 kHz. FeCo alloys are the automatic choice for such applications owing to their high Curie temperature and high saturation induction. However, the mechanical properties of FeCo-based alloys are inadequate to meet the stringent design requirements.11 In the past three decades, many efforts were directed at improving the physical, magnetic and mechanical properties through various alloy design principles. This review summarises these efforts and presents the authors’ current understanding on the physical metallurgy of FeCo alloys. The initial part of this review (the next section on ‘Binary FeCo’) deals with the properties of binary FeCo alloys with an emphasis on the effect of composition on the mechanical and magnetic properties. Then the section ‘Ternary FeCo–V alloys’ deals with the commercially important ternary alloys based on Fe–Co–V. Here, the recent advances in understanding the phase transformations and structure–property correlations of FeCo–V alloys are highlighted. The section ‘Beyond FeCo–V alloys’ describes the development of new alloys with superior mechanical and magnetic properties over the standard Fe–49Co–2V alloy. Recent efforts directed at developing nanocrystalline and thin film based on FeCo compositions for a variety of emerging applications are summarised in the section on ‘Nanocrystalline alloys and thin films’. Despite the progress made during the past several decades, there are several aspects of FeCo alloys that are not clearly understood. The final section highlights the research areas that should be focused in future research.

1 Phase diagram of binary Fe–Co system (reprinted from Ref. 12, ß1990, with permission from ASM International)

sub-lattices in which the Fe atom occupies one sublattice and the Co atom occupies the other sub-lattice (Fig. 2a). The a9 phase undergoes an order–disorder transformation when heated to high temperatures. The variation in the degree of long-range order with temperature of FeCo and FeCo–V alloys has been studied by specific heat measurements,14,15 theoretical calculations16,17 and by X-ray18 and neutron diffraction techniques.19 The change in degree of order (S) of FeCo with temperature is shown in Fig. 2b. The continuous variation of S from 0 to 1 when the temperature is lowered below the critical ordering temperature (Tc) Table 3 Applications of FeCo alloys

Binary FeCo Crystal structure and transformations The binary Fe–Co phase diagram12 is shown in Fig. 1. At ambient temperatures, the intermetallic compound FeCo (a9) is stable in the range of 29–70 at.-%Co.13 The B2 (CsCl) structure of FeCo is an ordered bcc structure and can be viewed as two interpenetrating simple cubic

High performance transformers Pole tips for high field magnets Magnetically driven actuators in impact printers Diaphragm in telephone handsets Solenoid valves Magnetostrictive transducers Flux guiding parts in inductive speed counters Electromagnetically controlled intake and exhaust nozzles Internal starter/generator in aircraft

Table 1 US markets for soft magnetic materials (millions of dollars) (reprinted from Ref. 2, ß1995 TMS) US market share (in millions of dollars) Material

1992

1997

Average annual growth rate, %

Steel Ceramic Other metallic alloys Total

1620 195 66 1881

2265 293 84 2642

7 8.5 5 7

Table 2 Saturation magnetisation and Curie temperatures of commercial soft magnetic alloys

158

Material

Saturation magnetisation (T)

Curie temperature (uC)

Iron Fe–(30–50)Co Low carbon steels (AISI C1010) Fe–(1–3%)Si alloys Amorphous Fe–Co–B Ni–Fe alloys Fe3Al alloys

2.15 2.45 2.0–2.15 1.9–2.0 1.9 0.8–1.6 1.14

770 920–985 770 730–760 370–420 250–500 540


2005

VOL

50

NO

3

Sundar and Deevi


3 Variation of lattice parameter of Fe–Co alloys as a function of Co content (reprinted from Ref. 8, ß1951 IEEE)

slightly larger lattice parameter compared to the disordered state. In the 1960s, FeCo alloys have been studied as a model system to bring out the effect of order on the mechanical and magnetic properties of intermetallic alloys. As discussed later, the state of order, and the lattice strains owing to the differences in lattice parameters of the ordered and disordered phases exert a profound influence on both the mechanical and structure-sensitive magnetic properties. 2 a B2 crystal structure of FeCo and b variation of order parameter of stoichiometric FeCo and FeCo–V alloys with temperature (reprinted from Ref. 18, ß1964, with permission from Elsevier)

indicates the occurrence of a second order order– disorder transformation. The critical ordering temperature Tc depends on the Co content and exhibits a maximum of 730uC near the stoichiometric composition. Ordering takes place rapidly when the alloy is cooled from the disordered state (a). In thin specimens, it is possible to retain the alloy in the completely disordered condition by rapidly quenching from temperatures greater than Tc.20 Reordering of disordered alloys occurs by homogeneous ordering followed by coalescence of antiphase domains at temperatures greater than 500uC.21–26 On the other hand, the diffusion of atoms is slow at low temperatures and the ordering occurs by nucleation and growth of ordered regions at the grain boundaries.20–22 Growth of ordered regions involves simple rearrangement of atomic positions across the interface between the ordered nuclei and the surrounding disordered matrix and does not require any longrange diffusion of atoms. At higher temperatures, the disordered a undergoes a polymorphous transformation into fcc phase (c). The Curie temperature of FeCo coincides with the a to c transition temperature. Like the critical ordering temperature, the Curie temperature also depends on the alloy composition. The alloy with 46 at.-%Co shows a maximum Curie temperature of 985uC. The high Curie temperature of the alloy allows it to retain good magnetic properties to higher temperatures, thereby making FeCo an attractive candidate material for high temperature soft magnetic applications. The lattice parameter of FeCo decreases with an increase in Co content (Fig. 3).8,27 Alloys in the ordered state have a

Atomic defects Knowledge of the defect behaviour is fundamentally important to gain an understanding of phase transformation, plastic deformation processes and diffusion and transport properties of an alloy. Point defects

In B2 alloys, deviation from stoichiometric composition is accommodated through formation of two types of point defects, namely anti-structural defects and triple point defects.28 Neumann29 predicted the formation of anti-structural defects in compounds with low enthalpy of defect formation, DHf (,30 kJ mol21), wherein lattice defects of A atoms on the b sub-lattice (in alloys rich in A) and B atoms on the a sub-lattice (in alloys rich in B) are created. A triple point defect is formed in a compound when its DHf is greater than 30 kJ mol21. The deviation of composition from stoichiometry is accommodated through the formation of A atoms on the b sub-lattice and creation of vacancies on the a sublattice. The DHf of FeCo is 10 kJ mol21.30 In accordance with Neumann’s prediction, deviation from either side of the stoichiometric composition is accommodated through the formation of anti-structural defects in FeCo.31 Slip systems and dislocations

Von Mises32 proposed that the plasticity of a polycrystalline alloy depends on the operation of five independent slip systems in each grain, to maintain compatibility with the neighbouring grains during the plastic deformation. Surface slip trace analysis and TEM studies have indicated that plastic deformation of FeCo occurs on {110}n111m slip systems.33–36 They provide more than five independent slip systems, and thereby, satisfy Von Mises criteria for plasticity in polycrystalline


2005

VOL

50

NO

3

159

Sundar and Deevi


a disordered alloy exhibiting wavy slip; b partially ordered condition revealing planar slip 4 Optical micrographs of FeCo alloys revealing nature of surface slip line markings in deformed samples (reprinted from Ref. 33 with permission from Taylor & Francis Ltd, http://www.tandf.co.uk/journals/titles/14786435.html)

materials. However, FeCo alloys are brittle in the ordered condition. Baker and Schulson37 analysed the ductility of various intermetallic alloys and pointed out that the Von Mises criterion is not a sufficient condition to ensure ductility in ordered intermetallic alloys. Additional factors like disordering at grain boundaries and suppression of environmental embrittlement are necessary to overcome the brittleness in the ordered condition. The state of order in the material determines the nature of dislocations observed during the deformation of FeCo. Deformation in FeCo at temperatures greater than Tc or in the alloys quenched from the disordered state occurs by movement of 1/2n111m unit dislocations. At temperatures less than Tc, anti phase boundary (APB) coupled 1/2n111m super partials contribute to the deformation of ordered alloys. The magnitude of the APB energy dictates the separation distance between the partials and determines the ease of cross slip of dislocations in the ordered alloys. The separation distance between the partials is low in alloys with high APB energies and cross slip occurs easily in these alloys. Similarly, in alloys with low APB energies, the partial dislocations are widely separated and each partial dislocation can independently cross-slip. However, the separation distance between the partials is intermediate in alloys with moderate APB energy and requires higher energy to cross slip. Hence, partial dislocations are restricted to slip on their slip plane. FeCo has a moderate APB energy (0.12–0.16 J m22).18,33,38 Hence, the APB coupled 1/2n111m dislocations have an intermediate partial separation in the ordered condition and planar slip is promoted. As a result, the nature of slip mode varies with the state of order.33,34,36,39 Alloys in the disordered state exhibit wavy slip behaviour

160


2005

VOL

50

NO

3

(Fig. 4a), indicating the occurrence of a cross slip. On the other hand, a planar slip mode is observed during the initial stages of deformation of alloys in the ordered condition (Fig. 4b).33 With the progress of deformation or straining, superdislocations tend to uncouple into single dislocations followed by APB trails. During this stage, the slip behaviour of the ordered state is predominantly wavy in nature.33,39 The deformation and fracture behaviour of FeCo alloy is significantly affected by the restriction of cross slip and promotion of planar slip behaviour in the ordered condition.

Processing One of the critical factors determining the selection of alloy for a given application is the overall cost of the alloy. Successful commercialisation requires processing of the alloys to the required shape and size without increasing the overall cost significantly. This is especially true in the case of soft magnetic materials as a result of the intense competition between the various alloys for a given application.2 Binary FeCo alloy can be hot worked at temperatures greater than 900uC.21,40,41 Generally, hot forging or hot rolling is employed to break down the cast microstructure. Zhao and Baker41 examined the effect of composition on the extrusion characteristics of FeCo alloys. During hot extrusion, FeCo alloys undergo dynamic recrystallisation (DRX). During hot working, DRX has been widely recognised as one of the safest mechanisms to process the materials and to reconstitute the microstructure.42 The extent of DRX depends on the Co to Fe ratio.41 DRX occurs easily when the Co/Fe ratio is greater than or equal to 1. On the other hand, recovery occurs easily in Fe rich alloys and the extent of DRX is limited. Recently, Hosoda et al.43 characterised the cold rolling behaviour

Sundar and Deevi


5 Variation of hardness with % reduction of thickness during cold rolling in ordered and disordered FeCo alloys (reprinted from Ref. 43, ß2000, with permission from Elsevier)

of different B2 intermetallic compounds such as FeAl, NiTi and FeCo. They noted that FeCo alloys (with 0.6%V) are difficult to cold roll in the ordered state, as indicated in Fig. 5. The hardness of the ordered alloy increases rapidly with the amount of cold working and cracks formed even at a reduction as low as 5% owing to rapid work hardening. The difficulties associated with cold rolling of ordered FeCo are attributed to the rapid work hardening associated with the ordered state. In contrast, disordered alloy exhibits sufficient room temperature ductility and relatively moderate work hardening rates and hence could be cold worked up to 40% at room temperature.43,44 Mao et al.44 characterised the recrystallisation behaviour of cold worked FeCo alloy in the ordered and disordered region. It was found that the cold worked alloy readily recrystallises when annealed at temperatures greater than 600uC. The recrystallisation process of FeCo in the disordered region is similar to that of bcc metals.45 It occurs through a mechanism involving normal nucleation and migration of high angle grain boundaries. During recrystallisation in the disordered region, the {111} rolling texture is replaced by a {111} recrystallisation texture.44 When the cold worked alloy is recrystallised in the ordered region, the ordering process interferes with the recrystallisation kinetics.44,46 During annealing, the alloy undergoes ordering and reaches an equilibrium ordered state well before the start of recrystallisation.44 The massive atomic movement during ordering decreases the driving force for normal recrystallisation and leads to strong recovery. Consequently, recrystallisation occurs via a continuous mechanism involving strong recovery of the deformed matrix with no drastic changes in the orientation of deformed grains. Texture is unaltered during recrystallisation of the cold worked alloy in the ordered region,44 which further supports the continuous recrystallisation mechanism. Moreover, recrystallisation kinetics is retarded as a result of the sluggish migration rate of grain boundaries in the ordered matrix. Davies and Stoloff47 also reported the effect of ordering on the recrystallisation and grain growth of FeCo–2V. Similar to binary FeCo, recrystallisation is observed when annealed below Tc. Grain growth kinetics in the ternary FeCo–V alloy is lowered below Tc. The lower rate of

6 Stress–strain curves for a disordered and b ordered alloys at different temperatures (reprinted from Ref. 39, ß1974 TMS)

grain growth with the onset of atomic order is attributed to slower diffusion kinetics and to an increase in the grain boundary energy in the ordered condition.

Mechanical properties It is feasible to vary the degree of order in FeCo from zero to one by suitable heat treatments. Hence, FeCo alloy has been selected as a model intermetallic system to bring out the effects of long-range order on the mechanical properties. The state of order, temperature and grain size influence the deformation and fracture behaviour of FeCo alloy and this is discussed below. Effect of temperature and state of order

Marcinkowski and Chessin33 reported the flow behaviour of stoichiometric FeCo alloy as a function of temperature (2196 to 364uC) and degree of long-range order. The flow behaviour of the disordered alloy exhibits single stage deformation with profuse cross slip and a strong temperature-dependent flow stress at all strains (Fig. 6a). On the other hand, the ordered alloy exhibits three distinct stages of deformation similar to


2005

VOL

50

NO

3

161

Sundar and Deevi


8 Effect of cobalt content on tensile properties of FeCo alloys (reprinted from Ref. 36, ß1994, with permission from Elsevier)

FeCo–V alloys.18,48 Factors responsible for the flow peak will be discussed later. Effect of grain size

7 Variation of Hall–Petch parameters (a) s0 and (b) k with strain for both disordered and ordered binary FeCo alloy (reprinted from Ref. 49, ß1965 TMS)

fcc metals (Fig. 6b). The flow stress of ordered alloys in stage I is independent of strain and mildly dependent on the temperature. In stage II, the flow stress is nearly independent of temperature and exhibits rapid work hardening rates. Slip line and TEM observations reveal planar slip behaviour of superdislocations and no evidence of cross slip. On the other hand, slip lines become stepped in stage III and TEM observations reveal dissociation of superdislocation into unit dislocations and extensive cross slip of unit dislocations.33,39 Temperature-dependent flow behaviour of an ordered alloy in stage III is similar to that of disordered FeCo.33,39 Hence, it was proposed that the deformation mechanism in Stage III of an ordered alloy is similar to that in a disordered alloy. The deformation mechanism was later identified as overcoming the Peierls barrier by unit 1/2n111m dislocations.39 Marcinkowski and Chessin33 observed a peak in the flow stress near Tc, when the flow stress is studied as a function of quench-in temperature or test temperature. They attributed the flow stress peak to the variation of strengthening due to short range order with quench or test temperature. A similar flow stress peak was later reported in ternary

162


2005

VOL

50

NO

3

Marcinkowski and Fisher49 and Marcinkowski50 reported the effect of atomic order and grain size on the flow behaviour of stoichiometric FeCo alloy. The variation in compressive flow stress with grain size in ordered and disordered conditions follows the classical Hall–Petch equation of the form sys ~so zkd {1=2

(1)

where sys is the flow stress, d is grain size, so is the intrinsic lattice resistance to deformation and k is the Hall–Petch parameter, a measure of resistance to slip transmission across grain boundaries. The variation of Hall–Petch parameters with strain is shown in Fig. 7. so at low strains is greater in disordered alloys and is attributed to strengthening owing to short-range order. However, so increases with strain and reaches a maximum value and is higher for an ordered material than for a disordered material. Overall, the Hall–Petch slope k is higher for the ordered condition. The k value increases with strain to a maximum and then decreases for both ordered and disordered states. The higher k value in the ordered condition is attributed to a limited number of slip systems and to the higher stress necessary to activate dislocation loops in the ordered condition.49,50 This is consistent with the theories of grain boundary strengthening by Cottrell51 and Petch.52 In a more recent work, Zhao and Baker36 systematically investigated the effect of grain size and the ordered state on the room temperature deformation behaviour of stoichiometric and two off-stoichiometric FeCo alloys. The yield strength of the alloy exhibits a

Sundar and Deevi

maximum at the stoichiometric composition (Fig. 8). All three alloys obey the classical Hall–Petch relationship. The variation of so with composition is similar to the variation of yield strength with composition. In the stoichiometric and Fe rich alloys, so is higher in the disordered condition than in the ordered condition, whereas the reverse is true for the Co rich alloy. Zhao and Baker36 attributed the difference in magnitude of so between the ordered and disordered state in a given alloy to the extent of short-range order (SRO) strengthening in the disordered condition. The Hall–Petch parameter k is maximum at the stoichiometric composition in both the ordered and disordered conditions and decreases on either side of stoichiometry. Similar to an earlier study,33 the k value is higher in the ordered condition than in the disordered condition in all three alloys. The above observations indicate that both compositional and thermal disordering lead to a lower k value. In B2 intermetallics such as NiAl and FeAl, the decrease in k correlates well with a decrease in so.53 However, this observation does not apply to FeCo. As pointed out by Zhao and Baker36 in the case of FeCo alloys, the difference in the behaviour is due to the additional contribution of SRO to so, whereas SRO does not influence k. Ductility and fracture

Normally, a sufficient number of slip systems and moderate ordering energy are associated with the ductile behaviour in intermetallics. However, this is not true in the case of ordered FeCo. As a result of brittleness of the binary FeCo alloys in the ordered state, many studies relied on compression testing to study the deformation behaviour.33,39 Limited tensile data on binary Fe–Co alloys indicate that the ductility of the stoichiometric and Fe rich compositions in the ordered state is virtually zero (Fig. 8b).36 On the other hand, the ductility increases on the Co rich side of the stoichiometry.36 Both the ordered and disordered Co rich alloy fail by dimple, ductile rupture. In contrast, the stoichiometric and Fe rich alloys fail predominantly by intergranular fracture in both the ordered and disordered conditions. Careful SEM examination of the brittle intergranular fracture features by Zhao et al.54 suggests limited plasticity in the ordered condition. The brittleness of FeCo is mainly as a result of the planar slip behaviour of the alloy in the ordered condition.33,34,36 Johnston et al.34 explicitly linked the brittleness of ordered FeCo–V to slip character changes with ordering. As discussed earlier, the moderate APB energy of FeCo prompts 1/2n111m super-partials to move as pairs connected through APB. As a result, cross slip of screw dislocations is suppressed and planar slip is favoured in the ordered state.33,36 Recently, Zhao and Baker36 proposed that deformation-induced disordering of the slip plane could be another reason for planar slip in ordered FeCo. Their suggestion is based on a substructure analysis of deformed Ni3Al, which revealed that the disordering of the slip plane owing to passage of numerous superdislocations lowers the APB energy in this slip plane.55 As a result, dislocations prefer to move in the same slip plane and promote planar slip behaviour. Similar behaviour is expected during the deformation of ordered FeCo. As a result of planar slip, slip transfer across the grain boundaries is difficult and the fracture at grain


9 Effect of alloying elements on saturation magnetisation of iron (reprinted from Ref. 8, ß1951 IEEE)

boundaries is initiated at lower strain levels. Thus, the ordered alloy fails in a brittle manner. Based on the dramatic improvement in ductility of Ni3Al owing to boron addition, Baker and Schulson37 proposed that disordering of the grain boundaries may facilitate the slip transfer across the grain boundaries and result in ductility improvement. Zhao and Baker36 reported that the ductility of binary FeCo alloys is more in the disordered state than in the ordered state, and deviation from stoichiometry resulted in ductility improvement. The above results are consistent with Baker and Schulson’s37 suggestion that promoting the slip transfer through thermal or constitutional disordering improves the ductility.27,33,36,49 Nevertheless, the stoichiometric and Fe rich alloys still fail by intergranular fracture,36 indicating the intrinsic weakness of grain boundaries.

Magnetic properties Pure iron is one of the best known soft magnetic materials. However, in the pure elemental form, it does not have enough strength and resistivity to meet application requirements. Significant efforts were made to improve the strength and resistivity of iron through alloying additions.8 However, almost all alloying additions invariably decrease the saturation magnetisation of iron (Fig. 9). Cobalt and manganese increase the saturation magnetisation of iron.8 The saturation magnetisation of iron increases with Co additions and exhibits a maximum saturation of 2.45 T (Tesla) at 35 wt-%Co5,6,8,56 (Fig. 10a). The saturation magnetisation of Fe–35 wt-%Co is about 13% higher than that of pure iron. The increase in magnetisation with Co content is attributed to an increase in the polarisation of Fe atoms with Co addition.57–60 High saturation in Fe–35Co was first identified by Preuss5 and Weiss6 in 1912. Later in 1927, Ellis61 and Elmen62 reported that stoichiometric FeCo has better permeability (Fig. 10b) and a saturation value equal to 2.4 T. The stoichiometric FeCo alloy was commercialised as ‘Permendur’.62 MacLaren et al.60 have carried out band structure studies (based on Layer Korringa-Kohn-Rostoker method) to bring out the effect of composition on the magnetic moments of iron and cobalt in FeCo alloys. The variation of iron and cobalt moments in the ordered a9 and disordered a phases are presented in Fig. 11. Iron moment exhibited a maximum near the equiatomic composition, while the cobalt moment remained almost


2005

VOL

50

NO

3

163

Sundar and Deevi


10 Variation of a saturation magnetisation and b permeability with Co content in Fe–Co alloys (reprinted from Ref. 8, ß1951 IEEE); maximum permeability (mm) of alloys depends on heat treatment temperature (shown in parentheses in b)

constant in the composition range studied. These results are in good agreement with the experimental Fe and Co moment values obtained by neutron diffraction studies.63 The combined effect of constant cobalt moment and varying iron moment with the alloy composition is attributed to the observed average moment/atom peak

11 Effect of cobalt content on iron and cobalt moments of ordered (a9) and disordered (a) phases in FeCo (reprinted with permission from Ref. 60, ß1999, American Institute of Physics)

164


2005

VOL

50

NO

3

12 a Relative change in induction ((Iordered–Idisordered)/ Iordered) of ordered stoichiometric FeCo alloy as a function of applied field and b variation of coercivity with test temperature in stoichiometric FeCo alloy64

at 30 at.-%Co. First principles calculations on an equiatomic composition reveal that the interatomic exchange coupling for ordered phase is slightly larger than that of disordered phase and are significantly larger than those found for pure iron and cobalt. Bogma64 characterised the magnetic properties of stoichiometric FeCo as a function of temperature and degree of order. Saturation induction of ordered FeCo is 2–3% higher than that of the disordered state. At low applied fields, the difference in magnetisation between the ordered and disordered state increases with increasing field and reaches a maximum at 12 kA m21 (151 Oe) (Fig. 12a). A further increase in field strength reduces the difference between the magnetisation of the ordered and disordered states. At ambient temperatures, the coercivity of FeCo in the disordered condition is less than that of the ordered condition (Fig. 12b). With increasing temperature, the coercivity of ordered alloys decreases with increasing temperature, whereas the coercivity of disordered alloys begins to increase with

Sundar and Deevi


That is, the value of K1 is zero for the stoichiometric alloy in the ordered condition. Similarly, the easy direction of magnetisation changes from n100m to n111m when the Co content is increased over 41 wt-%.56 Saturation magnetostriction (ls) represents the change in dimension due to stress under the magnetic field and is substantial for FeCo alloys (Fig. 13).66 Low magnetic anisotropy (K150) combined with high magnetostriction makes the stoichiometric ordered FeCo a good candidate material for transducer applications. In addition, the high saturation magnetisation of FeCo alloys could be utilised to induce large attractive forces to operate moving parts. FeCo alloys are considered to be ideal materials for relay armature applications. Some of the important magnetic properties of the stoichiometric FeCo alloy are summarised in Table 4.

Ternary FeCo–V alloys 13 Variation of magnetic anisotropic constant (K1) and saturation magnetostriction (ls) with Co content; dotted line shows variation of K1 in ordered condition (reprinted with permission from Ref. 66, ß1960 TMS)

the ordering of the alloy. The increase in coercivity is due to internal stresses created in the material as a result of the co-existence of ordered and disordered regions. Ordering occurs rapidly at temperatures greater than 500uC, and the coercivity of disordered material begins to fall and approaches a value close to that of an ordered alloy. Tailoring the magnetic properties of a soft magnetic alloy requires an understanding of the magnetic anisotropy of the material, and the ability to control it through the thermo-mechanical treatment. The magnetocrystalline or anisotropy energy is the energy that directs the magnetisation of the ferromagnetic crystal along certain crystallographic axes called the easy axes of magnetisation. This arises mainly because of the interaction of the magnetisation with the crystal lattice, which occurs via the orbital overlap of electrons: spinorbit coupling. Recently, Chu et al.65 prepared good quality single crystals of FeCo to characterise the anisotropy in magnetic properties. Pristine FeCo single crystals were prepared by annealing the binary FeCo ingot at 930uC/20 h in vacuum followed by slow cooling. Crystals parallel to the (110) plane were sectioned from the annealed ingot. Back reflection Laue and synchrotron X-ray diffraction techniques were used to confirm the single crystalline nature and ordered state of the crystals, respectively. They also reported preliminary results of their investigation on the magnetic properties of the single crystals along various crystallographic directions. However, further studies are needed to gain valuable anisotropic magnetic properties of single crystalline FeCo alloys. Apart from having high saturation magnetisation, FeCo alloys have an attractive magnetic property, namely low magnetocrystalline anisotropy.56 The anisotropy constant K1, which expresses the magnitude of anisotropy, depends on the state of order as well as on the Co content (Fig. 13).66 In the disordered condition, K1 decreases with increasing Co content and changes its sign from positive to negative at a composition close to 40%Co. On the other hand, in the ordered condition, the crossover occurs at the stoichiometric composition.66

Despite the discovery of high induction and high permeability stoichiometric FeCo alloys in 1927, they were unattractive for commercial applications as a result of their high cost, difficulties associated with processing into strips, and inherent low electrical resistivity. Processing the alloy into thin strips is essential to minimise the eddy current losses generated in AC applications. Similarly, increasing the resistivity of the alloy helps to further minimise the eddy current losses. In 1932, White and Wahl67 discovered that the addition of vanadium alleviated the poor ductility and low resistivity of stoichiometric FeCo. Ternary FeCo–V alloys were found to be cold workable up to 90% in the disordered state. Several commercially important soft magnetic alloys were developed based on FeCo– 2 wt-%V. The amount of vanadium added to the alloy is limited to below 3 wt-% to retain good soft magnetic properties. FeCo alloys with higher vanadium contents68–71 were developed for semi-hard (Remendur 22 to 5%V) or hard magnetic (Vicalloys 28 to 15%V) applications. The following discussions on FeCo–V alloys are restricted to soft magnetic alloys with vanadium additions up to 3.0 wt-%.

Physical metallurgy Site occupancy

At low temperatures, the solid solubility of vanadium in FeCo is less than 2%.72 Despite the extensive research done on FeCo–V alloys, the site occupancy of vanadium atoms in FeCo is not clearly understood.73 Mo¨ssbauer spectroscopic studies, size and electronegativity considerations suggest that vanadium atoms occupy the Fe Table 4 Magnetic properties of stoichiometric ordered FeCo (reprinted from Ref. 8, ß1951 IEEE) Property

Value

Curie temperature (uC) Saturation magnetisation (T) Coercivity (A m21) Initial permeability Maximum permeability Saturation magnetostriction (ordered condition) l100 l111 lpolycrystal Magnetocrystalline anisotropy constant (J m23)

980 2.4 150 800 5000–8000


2005

15061026 2561026 6061026 0

VOL

50

NO

3

165

Sundar and Deevi


sub-lattice.74 However, by combining X-ray and neutron diffraction methods, Williams et al.75 suggested that vanadium atoms occupy the Co sub-lattice. Further, the authors augmented their claim by noting that substitution of vanadium atoms with Co atoms is energetically favourable from a magnetic interaction point of view. Knowledge of the site occupancy of vanadium is important in understanding and controlling the properties of FeCo alloys and is critical for alloy design. Hence, further studies are necessary to clarify this issue. Phase diagrams and phase transformations

Raynor and Rivlin13 compiled and critically reviewed the constitution relationships in the Fe–Co–V system. They corrected inconsistencies in the published phase diagrams to comply with the latest and more accurate information available at the time. The phase equilibria near the FeCo (with low vanadium content) portion of the ternary Fe–Co–V systems are discussed below. Ko¨ster and colleagues76–78 pioneered the investigation of phase relationships in the Fe–Co–V system. Information related to liquidus and solidus boundaries in the ternary Fe–Co–V system can be found in Ref. 77. The present discussion on phase equilibria in the solid state is limited to two temperature ranges, namely 900– 950uC where hot working is carried out, and around 600–700uC where the final heat treatment is done to improve the mechanical and magnetic properties. Bennett and Pinnel79 utilised the polythermal sections provided by Ko¨ster and Schmid78 to construct isothermal phase diagrams at temperatures between 600 and 1000uC. Precise knowledge of the phase boundaries is vital for processing of alloys since quenching from single phase c poses problems during further processing at room temperature if temperature–phase relations are not well established. Bennet and Pinnel79 experimentally determined five tielines at temperatures between 900 and 950uC and noted that vanadium stabilises the c phase and expands the c and cza phase fields towards lower temperatures. They suggested that the sensitivity of the microstructure to annealing temperature and the corresponding effect on properties of FeCo–V is due to the rapid changes in the shapes of azc and c phase fields with temperature and vanadium content. Considerable ambiguity exists between the various published phase diagrams at temperatures less than 985uC (the maximum temperature for the c=a transformation in the Fe–Co system).13,72,80,81 The differences may arise owing to the sluggishness of various transformations,13,72 and the sensitivity of phase fields to a small change in alloy composition.77–81 For example, Ko¨ster and Schmid,77 and later Martin and Geisler80 reported that the c«a transformation exhibits considerable hysteresis between heating and cooling cycles and increases with increasing vanadium content. As a result of hysteresis and the sluggishness of the c=a reaction, Ko¨ster and Schmid77,78 noted considerable differences between the constitution of the alloys after short and long annealing treatments. Comprehensive knowledge of the transformations in FeCo–V is essential to understand the variation in mechanical properties with heat treatment. The transformations involved during the heat treatments of FeCo–V alloys are schematically shown in Fig. 14. It is well established that the addition of vanadium slows

166


2005

VOL

50

NO

3

14 Schematic illustration of phase transformations during heat treatment of FeCo–V alloys; c, high temperature fcc phase; c2, (Co,Fe)3V precipitates (fcc); a, equilibrium disordered bcc phase; a9, equilibrium ordered bcc phase; a2, metastable bcc phase

down the order–disorder transformation in FeCo.21 Rapid quenching retains the disordered state of FeCo–V alloys.19 When a vanadium-containing FeCo alloy is quenched from the single phase c or (cza) phase field, c transforms into the metastable vanadium rich bcc phase (a2).80,82,83 The above metastable transformation takes place via martensitic as well as massive transformation mechanisms.82 During quenching, the high temperature c phase undergoes a massive transformation at progressively decreasing temperatures. If the cooling rate is rapid enough to prevent completion of the massive transformation before the temperature drops below the martensitic start temperature Ms, the remaining c phase is converted into a2 through martensitic transformation. In addition, the Ms temperature decreases with increasing vanadium content.72,82,83 Thus, the relative contributions of the above two mechanisms depend on the cooling rate and vanadium content. By rapid cooling, it is possible to suppress ordering in the metastable a2 phase. When aged at temperatures below 700uC, the disordered a2 transforms to equilibrium a9 and precipitates as c2 phase.83–88 The kinetics of c2 precipitation are reported to follow classical ‘C curve’ behaviour (Fig. 15), indicating that precipitation is controlled by nucleation and growth events. For FeCo–2V alloys, the nose of the ‘C curve’ is close to 550uC.83 During the initial stages of aging, vanadium segregates to APB and the local composition around the APB is close to c2 composition. With further aging, c2 precipitation occurs homogeneously as well as on APBs. The activation energy for precipitation is equal to the activation energy for diffusion in the disordered state,89 suggesting that growth of precipitates is controlled by diffusion of atoms along the disordered layers associated with APB. Cold working the alloy before aging accelerates the precipitation kinetics (Fig. 15),79,83 and precipitation takes place extensively on the dislocations. In this case, the activation energy for precipitation is less than that of diffusion in ordered FeCo implying that the precipitation process is controlled by the diffusion of atoms (pipe diffusion) through the core of dislocations. Though the c2 phase is fcc, its composition is quite different from the high temperature c phase. Chemical

Sundar and Deevi

15 TTT diagram for precipitation of c2 during aging of quenched FeCo–V alloy (reprinted from Refs 83 and 165, ß1977 & 1981, with permission from Institute of Materials, Minerals and Mining

analysis of extracted precipitates as well as analytical TEM studies suggest that c2 is rich in vanadium,77–80,83,85 and its composition is close to (Co,Fe)3V. The mechanism of ordering strongly depends on the disordering temperature.19,20,24 In specimens quenched from temperatures less than Tc, ordering occurs homogeneously within the large domains, whereas in specimens quenched from above Tc, ordering occurs by a nucleation and growth process.24 The degree of order achieved after a given annealing treatment depends on the amount of prior cold working and the initial disordering temperature. Cold working up to 20% accelerated ordering.24 On the other hand, when the cold working was greater than 20%, ordering was retarded and was independent of the amount of prior cold working.19 Aging studies on disordered alloys revealed that the activation energy for ordering is not affected by prior cold working19 and is equal to the activation energy89 for diffusion in the disordered alloy. The retardation of ordering in cold worked material is due to the destruction of ordering nuclei or short-range order. The initial ordering process depends on the initial disordering temperature. Except for a small discontinuity near the a to azc phase boundary, the degree of long range order achieved after annealing treatment increases with increasing disordering temperature and it is attributed to the presence of excess vacancies in aiding the ordering reaction.

Processing FeCo–V alloys are melted and cast in vacuum to avoid contamination. The presence of elements like S, P, N and O are detrimental to the workability and to the desired mechanical and magnetic properties of FeCo alloys. Hence, careful control of impurities is required.90 Generally, initial processing is done in the c phase field (.940uC) to break up the as cast microstructure of the ingots. Subsequently, the billets are hot rolled in c or azc phase field (850–1000uC). Final processing is done through a cold rolling step to obtain thin sheets of FeCo–V alloy. As in the case of FeCo alloys, the cold workability of FeCo–V alloys in the ordered condition is limited by low room temperature ductility and rapid work hardening behaviour. Hence, successful cold


rolling requires that the alloy is in the disordered condition. The addition of vanadium retards the ordering reaction in FeCo, and the alloy can be quenched rapidly to retain the disordered condition. Hence, the alloy should be annealed at temperatures greater than Tc (.720uC) before cold rolling and subsequently quenched in water or ice brine solution to retain the disordered condition. Grain growth during high temperature annealing results in embrittlement of the alloy and inhibits cold workability.90 Small amounts (,0.5%) of niobium or zirconium are added to restrict grain growth during high temperature annealing.91 The final cold rolling step of FeCo–V alloys involves 80% or more reduction to form thin sheets of 4 to 50 mil (100– 1300 mm) thickness. Kawahara92 investigated the effect of microstructure on the cold workability of the FeCo–2V alloy and correlated the ease of cold rolling to the extent of homogeneity in the microstructure after cold rolling. The duplex structure (aza2) formed by quenching from the azc phase field exhibits a more homogenous microstructure after cold rolling and is readily cold workable up to 90% without any difficulties. On the other hand, disordered a or metastable a2 structures exhibit heterogeneous structures after cold working. These heterogeneities in the structure act as nucleation sites for cracks during the early stages of cold rolling and limit the cold workability. Texture measurements indicated the presence of duplex texture {001} n110m and {112} n110m in the cold rolled alloy.93–95 The nature of the texture was independent of the starting microstructure. Though {001} n110m was the dominant texture component after conventional cold rolling, special rolling such as cross rolling or clock rolling sequences intensify the {112} n110m texture component in the rolled sheets.96 Unlike Fe–Si alloys,56 the easy axis of magnetisation n111m of FeCo is not present in the plane of the rolled sheet. Hence, anisotropy owing to texture could not be used to enhance the magnetic properties of FeCo alloys. As discussed below, the presence of a texture results in distinct anisotropy in the mechanical and magnetic properties of the FeCo–V alloys.

Mechanical properties The mechanical properties of FeCo–V alloys are very sensitive to alloy composition, processing conditions, test temperature, microstructure, and state of order.18,88,94,95,97–101 The effect of the above variables on the low temperature (,0.4 Tm) deformation and fracture behaviour of the alloy is discussed below: Strength

Effect of degree of order Stoloff and Davies18 reported the influence of degree of order on the room temperature flow strength of the FeCo–2 at.-%V alloy when quenched from high temperatures in the range 500–900uC (Fig. 16). The strength of the alloy in the disordered state is higher than that in the fully ordered state and reaches a peak at an intermediate degree of order (S50.2). The flow stress peak occurs at a temperature, defined as Tp (y710uC), just below the order–disorder temperature. In samples quenched from temperatures greater than Tp, the movement of unit dislocations in a short range ordered


2005

VOL

50

NO

3

167

Sundar and Deevi


17 Effect of state of order on temperature dependence of yielding in FeCo–V alloy (reprinted from Ref. 34, with permission from Taylor & Francis Ltd, http:// www.tandf.co.uk/journals/titles/14786435.html)

16 Variation of tensile properties of FeCo–2V alloy as a function of quenching temperature (reprinted from Ref. 18, ß1964, with permission from Elsevier)

matrix controls the deformation, while in samples quenched from temperatures below Tp, the movement of superdislocations in the long range ordered matrix controls the deformation. When quenched from temperatures greater than Tp, the degree of short-range order retained decreased with an increase in quenching temperature. As a result, the strength of the alloy decreased as the quenching temperature was increased over Tp. On the other hand, the increase in strength when Tp was approached from lower temperatures was related to the increase in deformation resistance to the movement of superdislocations in the partially disordered matrix. The spacing between super-partials is inversely proportional to S2. At low values of S, the spacing between partials is large and they tend to disassociate into unit dislocations. Marcinkowski and Chessin33 confirmed the above results for a binary FeCo alloy. Later, Moine et al.48 observed a similar peak in the compression strength as a function of S, when testing FeCo–2V alloy at a higher strain rate. According to Stoloff and Davies,18 Tp (S50.2) represents the dissociation of superdislocations into unit dislocations. However, TEM experiments by Moine et al.48 revealed the existence of paired dislocations even at temperatures greater than Tc, different from the prediction of the Stoloff and Davies mechanism.18 According to Moine et al.,48 conventional Frank-Read type sources operate in the partially ordered condition and generate single dislocations or paired dislocations with only the leading dislocation active. They further proposed that the superlattice Frank-Read sources operate in the fully ordered condition and generate

168


2005

VOL

50

NO

3

paired superdislocations. The variation in stress required to operate these sources with quenching temperature (i.e. S) is linked to the observed flow stress peak at intermediate S value. Even though a lot of attention has been paid recently to understanding flow stress anomalies in intermetallic alloys,102,103 no further work has been reported on the anomalous flow behaviour in FeCo alloys since the observations of Moine et al.48 in 1971. Effect of temperature and grain size Johnston et al.34 investigated the flow behaviour of FeCo–2V alloy in both the ordered and disordered conditions over a temperature range of 2196 to 200uC (Fig. 17). The flow stress increases with a decrease in temperature in both states. The above strong dependence of the flow stress on temperature resembles that of a bcc alloy. The magnitude of yield stress in the ordered condition is considerably lower than in the disordered condition. Koylu et al.99 studied the temperature dependence of flow stress and deformation mechanisms of FeCo–2.3%V as a function of the degree of order. Irrespective of the state of order, deformation is controlled by the Peierls mechanism at temperatures less than room temperature as in the case of bcc alloys at low temperatures.104 Above room temperature, the deformation process is controlled by a thermal activation mechanism, such as cross slip in the case of partially ordered or disordered alloys. However, the deformation becomes athermal above room temperature, the occurrence of which in the fully ordered alloy is attributed to the difficulty of cross slip. Recently, Shang et al.105 reported the grain size dependence of the hardness of three commercially available FeCo alloys, namely Hiperco 27, 50 and 50HS. Hiperco 50 and 50HS represent ordered alloys with composition close to stoichiometry, whereas Hiperco 27 represents a disordered alloy. Surprisingly, the hardness variation of all the three alloys with grain size could be fitted with a single Hall–Petch type equation. The authors concluded that yield strength (hardness) variation with hardness of the alloys is independent of alloy composition. It is possible that the heat treatment utilised to vary the grain size could

Sundar and Deevi

have varied the microstructural state of the alloys. Hence, a more careful evaluation, separating out variables such as the amount of second phase precipitate and degree of order present in the material, is needed before accepting the above broad conclusion. Ductility and fracture

Stoloff et al.34,106 reported the fracture behaviour of ordered and disordered FeCo–2V alloy in the temperature range of 2196 to 500uC and grain size range of 13– 75 mm. The ductile-to-brittle transition temperature (DBTT – corresponding to a 50% reduction in area), and Zero Ductility Temperature (ZDT – corresponding to a yield stress equal to fracture stress) are sensitive to the state of order and grain size (Table 5). In the disordered condition, the DBTT is less than or equal to 2110uC. Depending on the grain size, the DBTT of ordered condition shifts to temperatures as high as 425uC. Both ordered and disordered alloys fail by transgranular cleavage at temperatures less than the DBTT. Cleavage fracture in the alloy is controlled by crack nucleation.106 Above the DBTT, disordered alloys failed by dimple fracture. The ductility of ordered alloys above the DBTT coincides with the onset of thermally activated recovery processes. However, the alloy fails by mixed intergranular and transgranular fracture.106 Johnston et al.34 noted a correlation between the slip mode and ductility of the alloys. Wavy slip led to ductile failure, whereas planar slip caused cleavage failure. However, Jordan and Stoloff98 in a later study found no


clear correlation between the ductile-to-brittle transition and planar/wavy slip mode. Pitt and Rawlings107 compiled published data on the variation in ductility with grain size of FeCo–V alloys and showed that the ductility is proportional to d21/2, where d is the grain, subgrain, or deformation cell size. Many intermetallic alloys, especially aluminides, are prone to environmental embrittlement at room temperatures as a result of moisture present in the atmosphere.108 The ductility of aluminides is significantly higher when tested in vacuum or a moisture free atmosphere. Unlike aluminides, the room temperature ductility of FeCo–V alloy (Table 6) is independent of the test environment109 suggesting that FeCo–V alloy is not susceptible to moisture-induced embrittlement. However, limited information available in the literature suggests that hydrogen may embrittle the alloy in the disordered condition.110 The yield strength is not affected by hydrogen. However, the ductilities of partially ordered and disordered alloys are lower when tested in hydrogen compared to air (Table 7). The inherent brittleness of the ordered alloy does not allow us to understand the influence of hydrogen and the alloy fails by cleavage under both air and hydrogen atmospheres. Processing conditions As cast FeCo–V is brittle at room temperature.92 Ductility can be induced in the cast alloy by quenching from the high temperature disordered region. The

Table 5 Ductile–brittle transition temperatures in FeCo–V alloy (reprinted from Ref. 106, ß1970, with permission from Claitor’s Law and Publishing) Ordered

Disordered

Grain size, mm

ZDT, uC

DBTT (50%RA), uC

ZDT, uC

DBTT (50%RA), uC

13 18 23 60 75

2170 2160 2160 to 2196 2170 Below 2196

350 425 420 420* 500*

Below 2196 Below 2196 Below 2196 2196 2196

2175 2175 2150 2110 290

ZDT, zero ductility temperature; DBTT, ductile to brittle transition temperature; RA, reduction in area. *Estimated, if intergranular cracking could be suppressed. Table 6 Effect of environment on room temperature tensile properties of FeCo–2V alloy (reprinted from Ref. 109, ß2002, with permission from Elsevier) Condition

Environment

Yield strength, MPa

UTS, MPa

Elongation, %

Ordered

Air Vacuum Water Oxygen

284 288 380 394

608 590 654 634

9.2 8.6 17.7 17.6

Disordered

UTS, ultimate tensile strength. Table 7 Effect of hydrogen on room temperature tensile properties of FeCo–2V alloy110 Condition

Environment

Yield strength, MPa

UTS, MPa

% Elongation

Disordered

Air Hydrogen Air Hydrogen Air Hydrogen

385 380 303 303 227 227

842 615 980 537 552 434

18 3 19 2 4 2.5

Partially ordered (S50.4) Fully ordered UTS, ultimate tensile strength.


2005

VOL

50

NO

3

169

Sundar and Deevi


18 Anisotropy in yield strength of cold rolled FeCo–V alloys at room temperature and at 480uC (reprinted from Ref. 11, ß1997, with permission from ASME)

quench temperature determines the ductility of the alloy. Alloys quenched from a single-phase a region exhibit higher ductility than those quenched from a two-phase (azc) or single-phase c region. However, improvements in ductility are lost if ordering is introduced in the material by aging below Tc. As discussed in the previous section, the final steps in commercial processing of the alloys involve cold rolling. In the cold rolled material, the presence of texture leads to anisotropic mechanical properties.11,111 Strength in the transverse direction is greater than that in the rolling direction, whereas the strength is lowest along 45u to the rolling direction (Fig. 18). Cold-rolled material is generally annealed at high temperatures to improve the mechanical and magnetic properties. Thornburg,100 and Stoloff and Dillamore106 reported that the room temperature yield strength of the cold worked alloy decreases with increasing annealing temperature, whereas the ductility peaks at an intermediate annealing temperature94,100 when the vanadium content is in the range of 1.5–2.0%. On the other hand, Pinnel et al.94 reported that the yield strength and ductility peak at an intermediate annealing temperature in an alloy containing higher vanadium (3 wt-%) content. In contrast to 1.5–2.0% vanadiumcontaining alloys, the ductility peak in a 3.0 wt-% vanadium-containing alloy occurs in the disordered region. Factors such as high dislocation density, fine precipitation, and fine grain size increase yield strength while recrystallisation, grain growth, dissolution of precipitates, and change in texture lead to a loss in strength. The influence of these factors on the overall strength of the alloy may vary with annealing temperature and alloy composition. For example, very little or no precipitation strengthening occurs in 1.5– 2.0% vanadium alloys, and the optimum conditions for ductility occur when the alloy is in the partially recrystallised condition. On the other hand, in high vanadium containing alloys (as in the case of 3.0 wt-%V), the presence of c2 precipitates leads to strengthening at intermediate temperatures, and as a result, the optimum condition for ductility is reached at higher temperatures where the alloy is in the fully recrystallised condition. Hence, it may not be possible to correlate the observed peak in ductility to a single mechanism.

170


2005

VOL

50

NO

3

19 Effect of rolling reduction on room temperature a tensile strength and b elongation of ordered FeCo–V alloy (reprinted from Ref. 95, ß1983, with permission from Kluwer Academic Publishers)

Ductility in ordered FeCo–V alloys As discussed earlier, a vanadium addition is more effective in improving the ductility of FeCo alloys in the disordered state than in the ordered state. Introduction of order in the binary and vanadiumcontaining FeCo alloys results in ductility loss at room temperature. However, two reports indicate enhanced ductility in the ordered condition.95,109 Kawahara95 has shown that ductility of the FeCo–V alloys in the ordered condition can be improved by cold rolling the alloys over a critical reduction of around 70% (Fig. 19). The beneficial effect of cold rolling is thought to be due to the refined Co3V clusters.112–115 The Co concentration in the matrix near the clusters is reduced resulting in the formation of a local concentrated disordered (LCD) region around the clusters (Fig. 20). In the as cast condition, the LCD regions are coarse and are non-uniformly distributed in the matrix and, thus, are not effective in enhancing the ductility. As a result of extensive cold rolling, LCD regions are refined and uniformly distributed in the form of elongated fibres in the matrix. The presence of these ductile regions in the matrix improves the ductility of the alloy through fibreinduced toughening. George et al.109 investigated the effects of small amounts of B and C on the mechanical properties of

Sundar and Deevi

a LCD zone around precipitate; b variation of Co content; c probability of ordering around precipitate 20 Schematic diagram illustrating LCD model (reprinted from Ref. 166, ß1983, with permission from Kluwer Academic Publishers)

binary FeCo and FeCo–V alloys. Alloying with B or C did not improve the ductility of binary FeCo in the ordered condition. However, B and C were effective in improving the ductility of FeCo–V alloy in the ordered state (Fig. 21). The maximum improvement was noted when the vanadium was in the range of 2.1–2.5 wt-%. In the case of B, the improvement in ductility was accompanied by a decrease in yield strength. The improvement in ductility was attributed to the presence of fine precipitates rich in B or C, and precipitates were believed to disperse the slip and reduce the stress concentrations due to planar slip in the ordered condition.

High temperature strength and creep behaviour Successful applications of new alloys at high temperatures require a thorough knowledge of the creep behaviour of the alloy at the operating temperatures. Designers of high temperature components need an understanding of the relationship between the steady state creep rate and rupture time with variation of stress and temperature. The variation of steady state creep rate (˙e) with stress (s) and temperature (T) can be described by the standard Dorn creep equation : e~Ksn exp({Q=RT) (2) where n is the stress exponent and Q is the activation energy for creep. Similarly, the rupture time (tr) is related to stress and temperature through the Larson– Miller parameter (P), which is given by P~(TR =1000)(log tr zC)

(3)

where TR is absolute temperature in Rankine and C is a constant. Earlier creep studies on FeCo and FeCo–V indicate that ordering enhances the creep resistance.97 As compared to an ordered condition, the disordered state exhibits a large increase in the creep rate and a decrease in activation energy. Similar effects of ordering on creep are observed in many ordered alloys such as b-brass, Fe3Al, Mg3Cd, MgCd and Ni3Fe.97 The creep activation


21 Room temperature ductility of FeCo–2V as a function of boron and carbon content (reprinted from Ref. 109, ß2002, with permission from Elsevier)

energies of various metals, alloys and intermetallic compounds are correlated with the activation energies for diffusion.116 It is interesting to note that the selfdiffusion coefficients of Fe and Co exhibit marked deviation from linearity because of the onset of B2 order and the diffusion kinetics is related to long-range order parameter in the ordered region.117 However, Stoloff and Davies97 pointed out that the creep activation energy is independent of degree of order over the range of S50–0.7 in ordered alloys such as FeCo, b-brass and Fe3Al. At present, there is no clear theory available to address this discrepancy. Material requirements for MEA applications have motivated researchers to characterise the creep behaviour of commercially available FeCo alloys. Design requirement of components in MEA require a soft magnetic material with a yield strength of 600 MPa or greater at 600uC and a creep strength of 5610210 s21 for a period of 5000 h at 550uC. Fingers and colleagues11,118,119 reported creep and high-temperature tensile behaviour of a high strength FeCo alloy (50 HS of Carpenter Technologies). The typical room temperature yield strength of this alloy is about 450–650 MPa.120 The strength of the alloy is derived from the combined effect of fine Nb rich precipitates and fine grain size.105 Tensile test results118 indicate the occurrence of a sharp yield point followed by Lu¨ders elongation at temperatures less than 320uC. Between 320 and 600uC, serrated flow curves are observed and the average yield stress is found to be independent of temperature. Both the factors indicate the occurrence of dynamic strain aging. The above findings are in broad agreement with an earlier report on dynamic strain aging in FeCo alloys.121 The creep behaviour of the alloy was reported over a temperature range of 375–600uC, and over a wide stress range of 200–800 MPa.118 Depending on the stress level, two types of creep curve were recorded as may be noted from Fig. 22. At stress levels close to the tensile yield strength of the alloy, the alloy exhibited abnormal creep behaviour wherein the creep rate increased abruptly in the steady state creep region. At stress levels well below or above the yield stress, normal creep curves were


2005

VOL

50

NO

3

171

Sundar and Deevi


22 Typical creep curves for Hiperco50 HS alloy at 510uC; open circles, 450 MPa (normal creep curve); open squares, 550 MPa (abnormal creep curve)118

24 S–N curves (stress controlled) of ordered and disordered FeCo–2V alloy [reprinted from Ref. 123, courtesy of AIME (www.aimehq.org)]

recorded. The abrupt increase in creep rate during the steady state deformation was linked to the occurrence of dynamic strain aging, where the mobile dislocation density increases with increasing creep rate by sudden unpinning of dislocations. The creep results of 50 HS at stress levels below its yield strength is expressed (Fig. 23a) as : : e~9:37x10{15 s3 83 exp({415,000=RT) (4)

(Fig. 23b) by plotting stress versus the Larson–Miller parameter. The fine grain size (,5 mm) of the commercial FeCo–V alloys could be one of the reasons for the observed low creep resistance. However, a special heat treatment devised to produce large grain size in a commercial FeCo–V alloy resulted in only a marginal improvement in the creep properties.122 The available creep data on commercial FeCo alloys indicate that they do not meet the design requirements of the emerging MEA applications. Clearly, development of new FeCo alloys with better creep resistance is essential to meet the proposed design requirements.

where s is in MPa, and Q is in J mol–1. The stress rupture properties of FeCo alloy are represented

Fatigue behaviour As in the case of creep, only very limited information is available on the fatigue behaviour of FeCo. Stoloff and colleagues123–125 characterised the fatigue behaviour of FeCo–2V in both the ordered and disordered conditions. Under stress-controlled cycling (Fig. 24), the ordered state exhibited better fatigue life than the disordered state.123 On the other hand, the disordered condition exhibited better fatigue life under strain-controlled cyclic conditions125 (Fig. 25). Irrespective of the state of order, the fatigue life of FeCo–V alloys under strain-controlled mode was better in vacuum than in ambient atmosphere (Fig. 25). The difference in fatigue resistance was explained based on the strain hardening behaviour of the materials under cyclic deformation. The ordered

23 a stress dependent on steady state creep rate and b stress rupture behaviour (Larson-Miller plot) of Hiperco alloy (50 HS)118

172


2005

VOL

50

NO

3

25 Low cycle fatigue behaviour of FeCo–2V alloy under strain controlled mode (reprinted from Ref. 125, ß1992, with permission from Elsevier)

Sundar and Deevi


Magnetic properties

26 Effect of atomic order and environment on fatigue crack growth behaviour of FeCo–2V alloy110

alloy reaches critical stress for fracture sooner compared to a disordered alloy owing to rapid strain hardening. On the other hand, rapid cyclic hardening of the ordered alloy under stress-controlled cyclic deformation gave a progressive decrease in the amount of plastic strain/ cycle. Consequently, crack initiation and fracture during stress controlled fatigue conditions were delayed in the ordered condition. As discussed earlier, the onset of order in many intermetallic compounds with moderate APB energies leads to restriction of cross slip and promotion of planar slip behaviour. During cyclic deformation, planar slip retards crack propagation by promoting the ease of slip reversibility. Consequently, ordering improves fatigue resistance through retardation of crack propagation.123–127 However, Kuruvilla110 reported that the fatigue crack propagation rate increases with increasing degree of order in FeCo–V (Fig. 26). The brittleness of ordered FeCo–V leads to premature crack initiation and enhancement of the crack growth rate. Thus, the lack of ductility of FeCo in the ordered condition overshadows the beneficial effect of ordering on the fatigue resistance. The presence of a hydrogen environment further accelerates the fatigue crack growth rate of the disordered and partially ordered alloys (Fig. 26). Hydrogen suppresses cross slip and enhances stress concentration at the grain boundaries, thereby promoting intergranular failure. For example, the fracture mode of the disordered alloy changes from transgranular with ductile features in vacuum to brittle intergranular failure in hydrogen. At present, there are no reports available on the elevated temperature fatigue behaviour of FeCo alloys. Emerging applications in MEA warrant further studies on high cycle fatigue and notched fatigue tests up to the maximum expected temperature of operation (550uC).11

The saturation magnetisation of FeCo decreases when alloying elements are added to improve its strength.8,128,129 The effect of various alloying elements on the saturation magnetisation of FeCo is summarised in Table 8. Alloying addition decreases the saturation magnetisation of FeCo alloy. Only limited information is available on the nature of magnetic interaction of solute or impurities in FeCo alloys.12 Recently, Reddy et al.130 have investigated the effect of substitution alloying elements Al, V, Mn and Ru on the magnetic properties of FeCo alloys by the cluster based ‘linear combination of atomic orbitals molecular orbital approach’. Calculated magnetic moments of Fe and Co using a 67 atom cluster are in good agreement with the band structure calculations and experimental values. All the solute atoms studied occupied iron sites with a finite magnetic moment. Aluminium and vanadium antiferromagnetically coupled to the host atoms and ruthenium coupled ferromagnetically to the host atoms. On the other hand, manganese coupled ferromagnetically in Fe–Co alloys and the moment varied with the relative concentration of Fe and Co. It is interesting to note that the authors utilised a small size cluster calculation to obtain the location of impurities and their magnetic coupling, and the information was used in large cluster calculation to obtain accurate local moments. Further studies based on similar lines are necessary to bring out the effect of other solute additions on the magnetic properties of FeCo. It is necessary to carry out the band structure calculations to determine the density of states and the exchange interactions of Fe and Co in the presence of potential alloying elements. From these calculations, the atomic moments and Curie temperature can be estimated within the framework of mean field theory.60,131–133 Like other alloying elements, vanadium additions are detrimental to the saturation magnetisation of FeCo.134,135 The effect of vanadium on the saturation magnetisation of FeCo is shown in Fig. 27.135 Vanadium couples anti-ferromagnetically136,137 with both Fe and Co atoms and decreases the saturation magnetisation of the alloy. On the other hand, the decrease in saturation magnetisation when the vanadium content is beyond its solid solubility limit is because of precipitation of paramagnetic c2 phase. The paramagnetic c2 phase reduces the saturation magnetisation as a result of a dilution effect.134 Gould and Wenney138 Table 8 Effect of ternary additions on magnetisation values of FeCo alloys from Ref. 8, ß1951 IEEE and from ß1984, with permission from Kluwer Publishers)

saturation (reprinted Ref. 141, Academic

Alloy (X)

Saturation magnetisation of FeCo–2 X (Ms), T

% Loss in Ms

V Mo Nb W Ni Ti Mn C Cr

2.30 2.34 2.32 2.37 2.35 2.33 2.35 2.35 2.29

4.2 2.5 3.3 1.2 2.1 2.9 2.1 2.1 4.6


2005

VOL

50

NO

3

173

Sundar and Deevi


27 Effect of vanadium content on saturation magnetisation of stoichiometric FeCo alloy135

reported that minimising the contamination during melting by using high purity starting materials and controlled melting atmospheres (through dry and wet hydrogen treatment) leads to an improvement in the structure-sensitive magnetic properties. Further, they reported that final annealing under a magnetic field results in further improvement in magnetic properties of FeCo–V alloy. The alloys prepared using the above procedures are called Supermendur.138 Table 9 compares the properties of regular 2V-Permendur with that of Supermendur. Structure-sensitive magnetic properties such as permeability and coercivity are greatly improved in Supermendur. Structure-sensitive magnetic properties such as the coercivity of ternary FeCo–V alloys depend on grain size,88,139,140 precipitate size and its distribution,88 internal stresses,88,140 state of order,88,100,141–146 texture,111 imposed external stress147 and test temperature.144 In general, vanadium refines the grain size of FeCo, and increases the coercivity.87,91 By carefully varying the temperature and annealing time, Yu et al.148 studied the effect of grain size on the coercivity of FeCo– V in both the disordered (S50) and ordered (S50.88) conditions. Both the DC and AC coercivity are found to be inversely proportional to the grain size (Fig. 28). The linear relationship between the coercivity and 1/D indicates that the grain boundaries act as pinning sites for magnetic domain walls. The higher coercivity of the disordered samples is attributed to the smaller grain size and higher magnetocrystalline anisotropy associated with the disordered condition. Internal strain created in the material due to c to a2 phase transition contributes to an increase in coercivity.139,140 The two-phase (a2za) structure, produced by quenching from the (azc) region, exhibits higher coercivity compared to the single phase a2 or a structure.139 In the two-phase structure,

28 Change in coercivity (DC and AC) with grain size in Fe–49Co–2V alloy (reprinted from Ref. 148, ß1999 IEEE)

transformation strains because of c to a2 and internal strains due to the lattice parameter difference between a2 and a contribute to coercivity. Aging or annealing at high temperatures can relieve transformation strains through recovery, thereby decreasing the coercivity. However, annealing or aging in the (a9zc2) region may actually increase the coercivity of the material due to paramagnetic c2 precipitates (which impede the domain wall movements).139,140 Figure 29 presents the effect of aging time at 600uC on the magnetic properties of commercial FeCo–V alloys.149 The magnetic properties significantly degrade with aging. The increase in coercivity and the decrease in magnetic induction are due to the increase in volume fraction of paramagnetic c2 precipitate with aging time.149 Pinnel and Bennett139 reported the variation of coercivity as a function of annealing temperature for FeCo–3%V alloy under two microstructural states namely as-quenched (from the

Table 9 Properties of commercial and high purity FeCo– 2V alloys56

174

Property

Permendur

Supermendur

Electrical resistivity (mV cm) Saturation magnetisation (T) Remanent induction (T) Maximum permeability Coercivity (A m21)

25 2.4 1.5 4000–8000 393

25 2.4 2.22 92 500 16


2005

VOL

50

29 Variation of magnetic properties of FeCo–V alloys as a function of aging at 600uC (reprinted from Ref. 149, ß2000 IEEE)

NO

3

2.21 2.23 2.12 2.28 2.23 2.22 1.54 1.60 1.69 1.72 1.96 2.21 1.46 1.52 1.60 1.63 1.90 2.15 1.37 1.42 1.48 1.50 1.83 2.09 1.22 1.28 1.32 1.30 1.74 1.98 0.98 1.00 0.76 0.84 1.36 1.57 1.3 1.7 0.92 1.5 0.6 0.76

1.06 1.12 1.04 1.04 1.54 1.75

2.15 2.24 2.27 2.29 2.28 2.31 1.12 1.72 1.18 2.2 2.24 2.22 0.98 1.7 1.03 2.17 2.2 2.18 0.81 1.66 0.81 2.13 2.16 2.14 0.59 1.6 0.54 2.05 2.07 2.07 0.12 1.4 0.09 1.78 1.62 1.64

2005

VOL

IBQ, ice brine quenched; CR, cold rolled; WQ, water quenched.

7 8 9 10 11 12 Hot rolled condition


1100uC WQ 1100uC WQ 950uC WQ 950uC WQ 800uC WQ 800uC WQ

1100uC IBQ 90%CR 1100uC IBQ 90%CR 950uC IBQ 90%CR 950uC IBQ 90%CR 950uC IBQ 90%CR 950uC IBQ 90%CR 1 2 3 4 5 6 Cold rolled condition

As rolled 800uC 5 min WQ As rolled 800uC 2 min WQ 800uC 5 min WQ 800uC 5 min WQ & aged 500uC/1 h As-quenched 500uC 1 h As-quenched 500uC 1 h As-quenched 500uC 1 h

2.6 0.37 1.8 0.27 0.29 0.28

0.26 1.48 0.23 1.9 1.69 1.91

Saturation magnetisation (T) 12 9.6 7.2 4.8 1.2 Initial

Final

Coercivity (kA m21)

2.4


Processing condition/Experiment no.

disordered region) and cold worked conditions (Fig. 30). The variation in coercivity with temperature is well correlated with the change in microstructural parameters such as grain size, extent of recovery, formation of c2 precipitate with annealing temperature and their effect on the domain wall movement. The coercivity of the quenched as well as the cold worked alloy exhibits a maximum near 900uC where equal amounts of a2 and a phase are present. In the cold worked alloy, an additional peak in coercivity is observed near 600uC, where c2 precipitation occurs. The presence of texture in the cold rolled condition results in magnetic anisotropy, and the anisotropy can be reduced by annealing at high temperatures.111 In addition to the annealing temperature, the cooling rate from the selected annealing temperature may affect the coercivity.150 For instance, rapid cooling results in high coercivity as a result of both the fine grain size and the disordered state of the material. On the other hand, c2 precipitates pin the magnetic domains and increase the coercivity in a slow cooled alloy. Cooling of an alloy at an intermediate cooling rate gives rise to low coercivity owing to lack of c2 precipitates and the ordered state of the alloy. The best magnetic properties are achieved at higher annealing temperatures120 and at intermediate cooling rates.138,150 The magnetisation behaviour of the alloy under a low applied field is important for applications like rotors and motors. Kawahara and Uehara141 characterised the magnetisation behaviour of FeCo–2V under different microstructural conditions. Their results are presented in Table 10. The early increase in magnetisation is sensitive to processing and heat treatment conditions. The early rise in magnetisation is rather low in the ascold rolled alloys. The microstructure before rolling had an influence on the magnetisation behaviour at low applied fields. In the cold rolled alloys, a duplex structure (a2za) exhibits a better early rise in magnetisation than the metastable a2 structure. On the other hand, a disordered a structure, produced by quenching of hot rolled alloys from single phase a, exhibits a better response to low applied fields. Irrespective of rolling and quenching conditions, additional annealing or aging treatment further improves the magnetisation under a low applied field. Recently, Hug et al.151 reported the

Magnetisation (T) at different magnetic fields (kA m21)

on coercivity of permission from

Heat treatment condition

30 Effect of annealing temperature FeCo–V alloy (reprinted with Ref. 139, ß1975 IEEE)

Table 10 Effect of microstructure on magnetisation of FeCo–2V alloy under low applied fields (reprinted from Ref. 141, ß1984, with permission from Kluwer Academic Publishers)

Sundar and Deevi

50

NO

3

175

Sundar and Deevi


32 Power loss characteristics of 0.010 thick Hiperco and Fe–Si alloys at 60 Hz and 600 Hz (reprinted from Ref. 9, ß1947 IEEE)

B–H hysteresis loop and hence Wh ~kh Hc Bm f

(6)

where Bm is the maximum flux density swing, Hc is the coercivity f is the frequency and kh is a constant. On the other hand, eddy current loss is related to eddy currents induced owing to the large change in magnetic induction. The amount of energy loss due to eddy currents is significant during AC applications, and 31 a effect of plastic strain on initial magnetisation and b power loss characteristics (at 50 Hz) of FeCo–2V alloy (reprinted from Ref. 151, ß2000, with permission from Elsevier)

effect of plastic straining (e51.5–3.5% plastic strain) on the initial magnetisation and power loss of FeCo–2V. Their findings are summarised in Fig. 31. The initial magnetisation of the alloy deteriorates when the alloy is subjected to plastic deformation (Fig. 31a). The deterioration of magnetic properties along the rolling direction is greater than in the transverse direction. Similarly, the power loss of the alloy increases with plastic strain; however, it is less sensitive to directional effects (Fig. 31b). The degradation of magnetic properties with plastic deformation is also observed in other soft magnetic materials like non-grain oriented Fe–Si alloy.152 Nevertheless, compared to an Fe–Si alloy, the magnetic properties of FeCo–V alloys are more sensitive to mechanical strengthening and are attributed to its high saturation magnetostriction.151 Power loss (W) in a soft magnetic material during AC application can be separated into hysteresis losses (Wh), eddy current losses (We) and anomalous losses (Wa)56,151 W ~Wh zWe zWa

(5)

Hysteresis loss is due to magnetic hysteresis of the material and can be minimised by reducing the impurities as well as removing the internal stresses. Hysteresis losses are proportional to the area under the

176


2005

VOL

50

NO

3

We ~ke

B2m f 2 Lt rp

(7)

where t is the thickness of the sample, r is resistivity, L is domain width and ke is a constant. The anomalous loss is given by the difference between the total power loss and the sum of hysteresis and eddy current losses. The current interest in MEA applications requires operation of FeCo alloys at frequencies close to 4000 Hz and at temperatures close to 550uC. Under these AC operating conditions, losses because of eddy currents become important. However, very little attention has been paid to characterising AC magnetic behaviour of FeCo alloys. Power loss characteristics of FeCo under different frequencies are shown in Fig. 32.9 For comparison, the power loss data on two Fe–Si alloys are also included. Among the alloys, Fe–(4–5)%Si alloys exhibit low power loss. However, it cannot be used at flux densities above 2.0 T. On the other hand, FeCo alloys exhibit comparable power loss to Fe–1%Si alloy at low frequencies (60 Hz) and perform better at higher frequencies. Wieserman et al.153 studied the core loss characteristics of FeCo–2V and grain oriented Fe–3%Si alloy over a temperature range of 25–300uC and frequency range of 0.1–10 kHz. Their results are summarised in Table 11. The core loss of FeCo–2V is lower only at high magnetic induction and low frequency (highlighted numbers). The low core loss observed in Fe–3%Si alloys at high frequencies could be due to the high electrical resistivity (see equation (7)).

Sundar and Deevi


34 Power losses in commercial FeCo alloys (after annealing at 720uC for 1 h) as a function of temperature (Bmax51.8 T and f51 kHz) (reprinted with permission from Ref. 156, ß2002, American Institute of Physics)

33 Effect of sheet thickness on power loss characteristics of Fe–Co–V alloys at different frequencies (reprinted from Ref. 154, ß1996, with permission from Elsevier)

report, similar information in FeCo alloys is not available. Fingers et al.156 reported the effect of long term aging on the magnetic properties of three commercial FeCo alloys (Hiperco 50, Hiperco 50HS and Hiperco 27) in the annealed condition (720uC) and after aging at 500uC. They characterised the AC magnetic property, namely power loss characteristics, as a function of aging time of up to 2000 h, temperature up to 500uC and frequencies up to 2000 Hz. Figure 34 presents the temperature dependence of the power losses of the three alloys in the 720uC annealed condition. Amongst the three alloys, magnetically soft Hiperco 50 exhibits a lower power loss than the other two alloys. The temperature-dependent power loss behaviour of Hiperco 27 is due to the stronger temperature dependence of resistivity compared to Hiperco 50 and Hiperco 50HS alloys. The above results again indicate the

For a given f and Bmax, it is possible to control the losses by varying the thickness and electrical resistivity. However, it is easier and more practical to control power losses by decreasing the thickness of the component. The effect of thickness on losses is shown in Fig. 33.154 At an industrial frequency range of 60 Hz, the power loss exhibits a minimum near 0.3 mm. The minimum in power loss shifts to lower thicknesses at higher frequencies. A similar trend in power loss with thickness was also observed in conventional Fe–Si alloys.155 The effects of temperature, microstructure and heat treatments on the power loss characteristics of Fe–Si alloys have been well documented. Except for one recent

Table 11 Comparison of core loss behaviour of Fe–3%Si (Magnesil) and FeCo–2V (Supermendur) alloys153 Core losses, W kg–1 Fe–3%Si alloy

FeCo–2V alloy

f, kHz

Bm, T

23uC*

150uC

300uC

23uC{

23uC*

150uC

300uC

23uC{

0.1

2.0 1.4 1.0 0.6 2.0 1.4 1.0 0.6 2.0 1.4 1.0 0.6 1.0 0.6 0.2 0.6 0.4 0.2

–

–

– –

–

6.4 3.5 2.2 1.1 34.0 19.6 12.6 6.2 92.2 54.5 34.8 18.1 206.4 105.4 22.9 – 129.0 49.4

5.3 3.1 1.8 0.9 31.7 17.9 11.0 5.5 91.3 53.4 34.0 17.2 222.7 115.3 27.1 – 144.8 58.4

0.0 2.4 1.5 0.8 24.9 13.4 7.7 3.7 77.2 43.2 25.8 12.3 224.9 113.1 21.2 – 145.5 44.5

7.1 3.1 1.8 0.8 24.0 13.4 7.7 3.7 62.4 37.3 21.4 9.9 143.7 68.1 10.6 – 80.5 24.0

0.4

1.0

5.0 10.0

6.8 2.8 1.2 –

6.1 2.3 0.9 –

27.7 10.8 4.5

6.8 2.8 1.2

1.9 0.8 –

24.2 9.6 4.0

– 19.3 7.7 3.1

28.9 11.7 4.7

–

–

–

–

73.3 28.6 12.2 174.4 72.1 9.2 170.7 96.5 28.4

64.6 25.1 10.3 153.8 63.6 8.5 141.6 77.5 23.7

49.8 20.4 7.7 123.5 50.0 6.1 119.3 58.7 16.7

78.2 30.8 12.2 187.4 74.0 11.5 193.7 100.5 27.2

*Initial measurement at 23uC. {Measurement made at 23uC after 300uC exposure.


2005

VOL

50

NO

3

177

Sundar and Deevi


35 Effect of applied compressive stress on magnetic properties of commercial FeCo alloys (reprinted with permission from Ref. 147, ß2003, American Institute of Physics)

importance of high resistivity in minimising the power losses at high frequencies. Magnetic components such as stator/rotor of aircraft power generators and starters are formed by stacking thin laminates of FeCo alloy. An axial load is imposed on the stack to improve its rigidity. Stoichiometric FeCo has positive magnetostriction and application of an applied compressive stress may degrade the magnetic properties and increase core losses.157 Turgut et al.147 reported the effect of compressive axial stress on the DC and AC magnetic properties of a stack of laminates formed from three commercial FeCo alloys: Hiperco 50, Hiperco 50HS and Hiperco 27. The effects of a compressive stress on the magnetic properties, namely coercivity, maximum permeability and core loss of the alloys are shown in Fig. 35. Coercivity of Hiperco 27 alloy increases with increasing applied stress. On the other hand, coercivities of Hiperco 50 and Hiperco 50HS alloys decreased initially with the applied stress, and then increased gradually. The maximum permeability of all three alloys decreases with the increasing applied compressive stress. Similarly, core losses increased with increasing applied stress. The increase in total core losses is attributed to the increase in anomalous losses with the application of external stress. All the above measurements were made by insulating the laminates from one another. Both the DC and AC magnetic properties of the alloys degraded further

178


2005

VOL

50

NO

3

when tested without the insulating layer between the laminates. Extensive future studies are needed to examine the effect of heat treatment conditions, which are necessary to attain a balance between the mechanical and magnetic properties and control the power loss characteristics of the FeCo alloys.

Resistivity We have shown in the previous section that the eddy current losses can be minimised by using a soft magnetic material with high resistivity (see equation (7)). Unfortunately, binary FeCo has a low resistivity ((7 mV cm).8,56 The effects of various ternary additions on the resistivity of FeCo alloy are shown in Fig. 36.8 The addition of ternary alloying elements with a high valence is effective in increasing the resistivity of FeCo.8 Vanadium is a particularly effective alloying element, in improving the resistivity of FeCo alloys. The electrical resistivities of commercial FeCo–V alloys range between 28 and 49 mV cm, and depend on the vanadium content, presence of impurities or other alloying elements, amount of cold work and aging treatment.8,88,135,158,159 For example, Ashby et al.158 investigated the effect of cold working and aging treatment on the resistivity of FeCo–2 wt-%V alloy, which was subjected to an initial disordered treatment.

Sundar and Deevi


38 Resistivity versus temperature of commercial FeCo alloys (heat-treated at 720uC/90 min) (reprinted with permission from Ref. 160, ß2003, American Institute of Physics) 36 Effect of alloying additions on resistivity of stoichiometric FeCo (reprinted from Ref. 8, ß1951 IEEE)

resistivity of FeCo–V alloys without degrading their magnetic properties.

The resistivity variation with aging time was correlated to the evaluation of microstructure with aging time. For example, the schematic variation of normalised resistance (with respect to the initial disordered condition) with aging time is shown in Fig. 37. The initial rise in resistivity of the disordered alloy is associated with the formation of ordered domains in a disordered matrix. The rise in resistivity with initial ordering is attributed to the creation of interfaces between the ordered and disordered phases. During the intermediate aging stage, segregation of vanadium to the APB before the formation of c2 phase is postulated to increase the resistivity. However, this trend is reversed with the nucleation of c2 phase formation and coarsening of ordered domains during longer aging times. Formation of the c2 phase leads to a depletion of solute in the matrix and leads to a decrease in resistivity. Geist et al.160 reported the effect of aging at 500uC on the resistivity of Hiperco 50, Hiperco 50HS and Hiperco 27 alloys. Figure 38 shows that the electrical resistivity is dependent on the composition of the alloy and increases with increase in temperature. A comparison of Hiperco 50 and Hiperco 50HS reveals the influence of precipitates and grain size on the electrical resistivity.105,120 Further alloy design efforts are needed to enhance the electrical

Beyond FeCo–V alloys

37 Schematic variation of normalised resistivity with aging time of FeCo–2V alloy (in initially disordered alloy) (reprinted from Ref. 158, ß1978, with permission from Wiley-VCH Verlag GmbH & Co)

During the last two decades, many researchers have developed new compositions based on FeCo with enhanced mechanical properties.115,161–167 The driving force behind these endeavours is the recognition that improvement in mechanical properties of FeCo-based alloys is imperative for their prospective applications at elevated temperatures. For example, materials used for rotors in advanced power generating units in aerospace should exhibit a minimum yield strength of about 700 MPa at room temperature, and 500 MPa at 600uC.10,11 More importantly, the intended applications require good creep resistance for the candidate materials. However, the current ternary FeCo–2V alloy is far from these requirements. In most cases, the attendant negative effect associated with the improvement in mechanical properties of FeCo alloys is the deterioration of their soft magnetic properties.8 However, several reports point out that through careful alloy design, it is possible to attain significant improvement in the mechanical properties without losing soft magnetic properties. The research efforts directed towards the development of new compositions are discussed next. The addition of alloying elements such as tungsten, niobium, tantalum, molybdenum, carbon and nickel up to 2% is effective in improving the cold workability of FeCo. In contrast, the addition of gold, silver, aluminium, beryllium, copper, manganese, silicon, titanium and zirconium did not improve the cold workability.115,141,166 Kawahara and colleagues115,141,166 correlated the effectiveness of a given alloying element (X) in improving the cold workability to its ability to form ordered clusters whose composition is close to Co3X. Cold rolling of the alloy improves the ductility by uniformly distributing the ductile LCD regions in the matrix. Table 12 summarises the solid solubility of ternary alloying elements in FeCo alloys. The mechanical and magnetic properties of the promising ternary alloys are presented in Table 13. As may be noted from Table 13, the improvements in strength and ductility in these ternary alloys are attained with very little degradation in magnetic properties.


2005

VOL

50

NO

3

179

Sundar and Deevi


Rawlings and colleagues134,167 showed that the addition of niobium or tantalum to FeCo results in balanced mechanical and magnetic properties. Unlike vanadium, addition of tantalum or niobium did not decrease the cRa transition temperature, thereby increasing the effective heat treatment temperature range over FeCo– V alloys. When niobium and tantalum are added beyond the solubility limit (Table 12), they result in precipitates rich in niobium or tantalum. As in FeCo–V alloys, the saturation magnetisation decreases with increasing amounts of the second phase. However, the loss in saturation for a given volume per cent of second phase is less in niobium-containing alloys (Fig. 39a). Similarly, for a given amount of second phase, the coercivity (Fig. 39b) of niobium-containing alloys is less than that of vanadium-containing alloys. However, it is important to note that the resistivity of niobium-containing FeCo alloy is low (Fig. 36), and hence niobium-containing alloys are restricted to DC applications.134,167 Fe–Co alloys with a low cobalt content (17– 35 wt-%Co) were developed168–173 in an effort to reduce the cost of the alloy while maintaining good magnetic properties. These alloys enjoy the distinct advantage of having the highest ductility and toughness among the commercially available Fe–Co alloys. When compared to stoichiometric FeCo alloys, these alloys exhibit equal or greater saturation induction, higher resistivity and lower magnetostriction. However, their use is restricted as a result of low permeability, high coercivity and more importantly high magnetocrystalline anisotropy. Typical properties of low cobalt alloys are summarised in Table 14. Low cobalt alloys, whose compositions are outside the ordering range can be processed and

Table 12 Room temperature solid solubility of ternary elements in stoichiometric FeCo (reprinted from Ref. 115, ß1983, with permission from Kluwer Academic Publishers) Element

Solubility limit (at.-%) in FeCo

V Ni Ti Mn Al Si Cu Zr Be Nb W Mo Ta Cr C B

(2 (2 (2 (2 (2 (2 (2 ,2 ,2 ,1 .0.5 .0.5 .0.5 .0.5 ,0.5 ,0.5

& & & &

,1 ,1 ,1 ,1

Although vanadium additions are beneficial in improving the room temperature ductility, it degrades the magnetic properties of FeCo. For example, addition of 2 wt-% vanadium to FeCo results in a 4% reduction in saturation magnetisation, 25% reduction in initial permeability, 50% decrease in maximum permeability and 10% increase in coercivity of FeCo.8,167 The decline in soft magnetic properties is because of the presence of paramagnetic c2 phase in the ternary FeCo–V alloy. Attempts were made to develop an alternative alloy composition with good mechanical properties without significantly losing the saturation magnetisation.

Table 13 Room temperature mechanical and magnetic properties of ternary FeCo alloys (reprinted from Refs 134 and 141, ß1983 & 1984, with permission from Kluwer Academic Publishers) Composition, at.-% FeCo–2V

Fe–40Co–2V FeCo–3.6V FeCo–5.6V FeCo–0.5C FeCo–2C FeCo–0.5Cr FeCo–2Cr FeCo–0.5Mo FeCo–0.5W FeCo–2W FeCo–2W FeCo–0.5Ta FeCo–2Ta FeCo–0.5Nb FeCo–2Nb FeCo–2Ni FeCo–1Nb FeCo–2Nb FeCo–3Nb

Thermomechanical treatment

YS, MPa

% Elongation

1100uC/IBQ/90%CR/as rolled 1100uC/IBQ/90%CR/500uC–1 h 1100uC–5 min/IBQ/90%CR/800uC/5 min 950uC–5 min/IBQ/90%CR 950uC–5 min/IBQ/90%CR/850uC/5 min/IBQ 800uC–5 min/IBQ/90%CR 800uC–5 min/IBQ/90%CR/800uC/5 min 1100uC–10 min/IBQ/90%CR/400uC/1 h

1167 1289 617 1334 573 1383 568 1246

5.6 4.4 9.0 2.2 8.5 2.8 10 5

800uC/10 min/IBQ/90%CR/ 800uC/5 min/IBQ/400uC/1 h 800uC/10 min/IBQ/90%CR/400uC/1 h 800uC/10 min/IBQ/90%CR/ 800uC/5 min/IBQ/400uC/1 h 1200uC/5 min/IBQ/90%CR/500uC/0.5 h 1200uC/5 min/IBQ/90%CR/500uC/0.5 h 1200uC/5 min/IBQ/90%CR/500uC/0.5 h 1200uC/5 min/IBQ/90%CR/500uC/0.5 h 1200uC/5 min/IBQ/90%CR/500uC/0.5 h 1200uC/5 min/IBQ/90%CR/500uC/1 h 1200uC/5 min/IBQ/90%CR/500uC/0.5 h 1200uC/5 min/IBQ/90%CR/500uC/0.5 h 1200uC/5 min/IBQ/90%CR/500uC/0.5 h 1200uC/5 min/IBQ/90%CR/500uC/0.5 h 1200uC/5 min/IBQ/90%CR/500uC/0.5 h

180


2005

VOL

50

NO

3

Coercive force, kA m–1

2.2 2.25 2.24 2.24

2.9 2.0 0.25 0.25

2.24

0.24

15

2.31 2.01 1.66 2.36

2.3 9.9 39.3 1.0

1050 838

9 15.5

2.37 2.34

2.5 1.0

1295 1117 1491 1538 1909 2092 1547 1961 1393 1765 1578

6 1.7 8.7 7 1.7 8.9 1 0.6 5.8 0 3.3

2.35 2.29 2.33 2.33 2.21 2.24 2.33 2.24 2.33 2.16 2.34 2.34 2.29 2.2

2.8 2.4 2.0 2.3 2.0 2.2 2.1 2.8 2.2 3.3 2.4 3.6 4.7 6.7

458

IBQ, ice brine quenched; CR, cold rolled.

Saturation magnetisation, T

Sundar and Deevi


40 Herzer diagram of grain size dependence of coercivity in nanocrystalline and microcrystalline magnetic alloys (reprinted from Ref. 174, ß1997, with permission from Elsevier)

39 Effect of niobium and vanadium on magnetic properties of FeCo alloy; a saturation magnetisation and b coercivity as a function of amount of paramagnetic precipitates in alloys (reprinted from Ref. 134, ß1991, with permission from Kluwer Academic Publishers)

machined without any difficulties.171 An alloy with cobalt content between 25% and 35% requires addition of either chromium or vanadium to improve its cold workability.168–173 Chromium or vanadium additions also improve the electrical resistivity. Low cobalt alloys are considered as low cost alternatives to FeCo in applications requiring saturation induction higher than conventional Fe–Si alloys and ferritic steels. Some of the current applications of low cobalt alloys are: solenoid valves in actuator, intake and exhaust nozzles and fuel injectors.

Nanocrystalline alloys and thin films Soft magnetic alloys with nanocrystalline and amorphous structure may allow us to develop novel soft magnetic materials with superior magnetic softness. As discussed in the section ‘Magnetic properties’ above, grain boundaries act as pinning points for magnetic domain wall movement and increase the coercivity of the conventional alloys. However, coercivity measurement of alloys with grain size less than 100 nm indicates that the coercivity of the nanocrystalline materials rapidly decreases with decreasing grain size. The reverse effect of grain size on the coercivity of the material is illustrated in Fig. 40. In nanocrystalline materials, a magnetic domain spans several grains and thus, domain wall movement is not hindered by grain boundaries. Averaging of the magnetocrystalline anisotropy over Table 14 Physical and magnetic properties of Fe–(15– 35)Co alloys Variable

Properties

Density Modulus Curie temperature Resistivity Saturation magnetisation Coercivity Maximum permeability Saturation magnetostriction Core losses (at 2 T and 50 Hz)

7.9–8.0 g cm23 160–200 GPa .800uC 20–30 mV cm 2.2–2.43 T 1.9 kA m21 2000–4000 25–4061026 8–10 W kg21

many grains coupled within an exchange length (1– 10 nm) is responsible for the observed low coercivity.174 Excellent reviews on the recent advances in the soft magnetic nanocrystalline and amorphous alloys can be found elsewhere.175–177 This section summarises the recent efforts in developing the nanocrystalline alloys and thin films based on Fe–Co compositions.

Powder synthesis of FeCo nanoparticles Unique magnetic softness found in magnetic nanoparticles with a grain size less than the exchange coupling length (,50 nm) has generated greater interest in exploring their use in a variety of applications.175–177 In this section, efforts directed towards the synthesis and characterisation of nanocrystalline FeCo powders are summarised. A variety of techniques such as plasma synthesis, gas condensation, chemical synthesis routes, electrodeposition, mechanical alloying and rapid solidification are employed to synthesise FeCo nanopowders.175,178–200 The magnetic properties of nanoparticles depend on their method of preparation and the size distribution, morphology and composition of the nanoparticles. McHenry et al.178–184 have employed KratschmerHuffman carbon arc and plasma torch methods to synthesise a variety of encapsulated (carbon and oxide coated) and un-encapsulated FexCo12x nanoparticles, with an average particle size less than 100 nm. Encapsulated nanoparticles may find applications in which fine particle magnets with protective coatings are important. In addition, the carbon coating could prevent particle coalescence during compaction and reduce eddy current losses. Energy filtered microscopy confirmed the chemical homogeneity of the nanoparticles, whereas XRD and TEM results revealed the presence of single phase bcc a-FeCo phase, along with graphite or oxide coating in all nanopowders with Fe contents ,80%. Thermomagnetic and neutron diffraction studies confirmed the rapid ordering of the initially disordered nanopowders when heated to 500uC.179,182 FeCo nanoparticles exhibit similar variation in the magnetisation with increase in cobalt content as in the case of bulk FeCo alloys. Hysteresis measurements (at room temperature) of oxide coated FeCo nanoparticles revealed an anti-ferromagnetic exchange biased coupling between the nanoparticle core and the oxide layer.183 Asprepared FeCo nanoparticles exhibit high coercivity (22.3 kA m21 at 5 K), with a strongly temperature dependent coercivity and saturation magnetisation


2005

VOL

50

NO

3

181

Sundar and Deevi


41 Temperature dependence of saturation magnetisation and coercivity of oxide coated FeCo nanoparticles produced by plasma torch technique (reprinted with permission from Ref. 183, ß1999, American Institute of Physics)

(Fig. 41), as with other oxide coated nanoparticles.185 Turgut et al.183 rationalised the observed high coercivity in the nanoparticles based on modified Neèl surface anisotropy model.201 While significant reduction in coercivity is observed after compaction of the powders by hot isostatic pressing,184 the coercivity values are still higher than the bulk FeCo alloys. Kim et al.191 characterised a nanocrystalline FeCo alloy produced by ball milling. The presence of high internal stress owing to mechanical alloying resulted in high coercivity in the as-milled powders. Annealing at temperatures as low as 400uC/1 h decreased the coercivity to below 1 kA m21. However, annealing led to grain growth (10 nm in the as milled condition to 42 nm after annealing for 1 h at 400uC). Baker et al.194 investigated the effect of various alloying additions on the grain refinement of FeCo powders during ball milling and subsequent annealing at 600uC. Alloying elements boron, carbon and (copper z boron) were effective in restricting the grain growth during annealing at 600uC/ 1 h (86 nm in the binary FeCo versus 21–61 nm in ternary and quaternary alloys). The coercivity of the nano-powders was found to decrease with increasing grain size. Consolidation of nanocrystalline powders into bulk nanocrystalline material without deterioration in their soft magnetic properties still remains a major challenge. Spark plasma sintering techniques may allow the consolidation of nanopowders without significantly altering the nanocrystalline state.191 Further studies are necessary to characterise the mechanical and magnetic properties of bulk nanocrystalline FeCo alloys.

Nanocrystalline alloys produced by primary crystallisation of amorphous precursors Successful development of nanocrystalline soft magnetic alloys like Finemet (FeSiBNbCu)202 and Nanoperm (FeBZrCu)203–205 has motivated researchers to develop nanocrystalline alloys based on FeCo for high temperature soft magnetic applications. These new generation soft magnetic alloys derive their superior magnetic properties from their unique two-phase microstructure consisting of ferromagnetic amorphous and nanocrystalline phases.174–176,202–207 The amorphous phase (present

182


2005

VOL

50

NO

3

42 Relationship between permeability (at 1 kHz) and saturation magnetisation of various soft magnetic materials (reprinted from Ref. 176, ß2000, with permission from Elsevier)

below its Curie temperature) promotes strong exchange coupling between the nanocrystalline grains and also increases the electrical resistivity of the alloy. Higher resistivity minimises eddy current losses and improves high frequency characteristics. The nanocrystalline ferromagnetic phase, with a grain size close to or less than the exchange coupling length provides lower coercivity and high magnetisation. These alloys are processed by conventional rapid solidification techniques (cooling rate .104 K s21) to produce an amorphous precursor. The nanocrystalline precipitates are produced within the amorphous matrix when annealed above the primary crystallisation temperature. Figure 42 compares the permeability and saturation magnetisation of the emerging nanocrystalline materials with conventional microcrystalline and amorphous alloys.173,205 Fe based nanocrystalline alloys such as Finemet and Nanoperm exhibit higher permeability and magnetic induction. These materials possess reduced hysteretic losses and improved high frequency range and response. Though Fe-based nanocrystalline materials like Finemet and Nanoperm alloys exhibit low coercivity at room temperature, they fail to maintain their properties at elevated temperatures as a result of the low Curie temperature (just above room temperature) of the amorphous phase.176 At temperatures greater than the Curie temperature of the amorphous phase, the nanocrystalline grains decouple from one another, resulting in an increase in coercivity. Therefore, the intergranular amorphous phase must also possess a high Curie temperature for the nanocrystalline alloys to operate at high temperatures. Amorphous precursors of FeCo alloys exhibit high Curie temperature (higher than the crystallisation temperature of nanocrystalline a9-FeCo phase) and are more suitable for high temperature applications over Fe based nanocrystalline alloys.207 By careful selection of alloy composition, it possible to further tailor the properties such as high permeability, low magnetostriction and high magnetisation. For example, by varying the ratio of Fe/Co, the magnetisation and magnetocrystalline anisotropy can be controlled. Similarly, near zero magnetostriction can be achieved by adjusting the composition such that the amorphous and nanocrystalline phases have near zero magnetostriction coefficient values.208,209

Sundar and Deevi

One of the promising compositions based on the FeCo system developed by researchers at Carnegie Mellon University is called HITPERM and has a composition of [(Fe12xCox)88M7B4Cu1 where M5Zr, Nb, Hf, Ta].207–213 Alloy compositions close to deep eutectics (normally hypo-eutectic compositions are used to maximise the volume fraction of high moment transition metal phase) are necessary to produce amorphous precursors by rapid solidification. Early transition metals (Zr, Hf and Nb) and metalloid elements (B, Si, P) are added to produce deep eutectics in ferromagnetic transition metal systems.202–207,210 Copper was added to HITPERM alloys to lower the primary crystallisation temperature and to promote crystallisation of the nanocrystalline a9-FeCo phase.210,212 However, recent atom probe and TEM observations revealed that copper does not form clusters before the primary crystallisation and remains in the amorphous matrix.213 Thus, copper in HITPERM alloys does not act as a heterogeneous nucleation site for nanocrystalline a9 phase and can be omitted in future alloys to improve the saturation magnetisation of the alloy. The addition of zirconium or niobium inhibits grain growth and retains good magnetic softness. Partial replacement of zirconium with hafnium is found to be beneficial in improving the stability of the amorphous phase.212 Phase transformation and crystallisation studies on HITPERM alloy have been carried out by differential scanning calorimetry (DSC),132,133,207,212,214–220 differential thermal analysis (DTA),208–213 X-ray diffraction (XRD) and synchrotron XRD,132,133,207–220 X-ray absorption fine structure,216 Mo¨ssbauer spectroscopy221–224 atom probe microscopy,213 atomic force microscopy (AFM) and magnetic force microscopy


(MFM),225 transmission electron microscopy (TEM)207–209,212–226 and thermomagnetic studies.132,210,215,216,218 Crystallisation and phase transformation temperatures and various magnetic properties of various Fe–Co based nanocrystalline alloys are summarised in Table 15. Crystallisation of HITPERM alloys occurs in two stages. Depending on the Co content, the primary crystallisation involves formation of nanocrystalline bcc Fe(Co) phase or B2 a9 FeCo phase embedded in the residual amorphous matrix rich in glass forming elements (B and early transition metals)132,210,215,216 During primary crystallisation, Co prefers to partition into the residual amorphous matrix.224 Cobalt enrichment has the beneficial effect of increasing the Curie temperature of the residual amorphous phase. During the secondary crystallisation, the amorphous matrix is converted into nonmagnetic or magnetically hard phases like (Fe,Co)3Zr, borides or niobides.220 Recent Mo¨ssbauer and XRD studies suggest that formation of the boride phase during secondary crystallisation is accompanied by dissolution of some of the FeCo grains.219,224 The effect of annealing temperature on the coercivity of Fe45CO43Cu1B3.6Zr7.4 is shown in Fig. 43.218 The local minimum in coercivity is observed after annealing in the temperature range corresponding to the primary crystallisation temperatures. The large increase in coercivity observed when the alloy is annealed above 650uC is due to the formation of borides during the secondary crystallisation. Therefore, the annealing temperature of the alloy should be selected such that the formation of magnetically hard phases can be avoided. Johnson et al.132 reported the effect of varying the ratio of Fe and Co from 50 : 50 (as in the original HITPERM alloy)

Table 15 Summary of crystallisation temperatures and magnetic properties of FeCo-based nanocrystalline alloys Composition

Tx1, uC

Tx2, uC

Bs, T

Fe44Co44Zr7B4Cu1 Fe44.5Co44.5Zr7B4 Fe57.2Co30.8Zr7B4Cu1 Fe44Co44Zr5.7B3.3Ta2Cu1 Fe44Co44Zr5.7B3.3Mo2Cu1 Fe45Co43Zr7.4B3.6Cu1 Fe45Co43Zr3.7B3.6Cu1Nb3.7 Fe45Co43B3.6Cu1Nb7.4 Fe45Co43Zr7.4B3.6Cu1Hf3.7 Fe45Co43Zr7.4B3.6Cu1Hf7.4 Fe49Co35Zr7B6Cu1 Fe43Co43Zr7B6Cu1 Fe56Co30Zr7B6Cu1 Fe66Co20Zr7B6Cu1 Fe61Co21Nb3B15 Fe59Co21Nb5B15 Fe57Co21Nb7B15 Fe62Co21Nb7B10 Fe64Co21Nb7B8 Fe67Co21Nb7B9 Fe63Co21Nb7B9 Fe59Co25Nb7B9 Fe55Co29Nb7B9 Fe51Co33Nb7B9 Fe4.5Co84.5Zr7B4 Fe4.4Co83.6Zr7B4Cu1 Fe42.7Co42.7Zr6.8B6.8Cu1 Fe42.7Co42.7Zr5.8B6.8Cu1Nb1

510 518 475 480 463 485 460 440 489 485 – – – – 421 462 503 456 436 458 449 443 439 439 481 475 440 457

700 689 – – – 701 750 740 708 701 – – – –

1.6–2.0

624 630

Hc, A m21

Tc, uC

l

Ta, uC

Reference

980

–

–

210 213 132

980 968 939

217

1.67 1.59 1.63 1.59 1.57 1.65 1.42 1.48 1.64 1.55 1.59 1.58 1.59 1.55

50 50 30 30 26 20 18 10 858 51.5 31.9 58.2 26.4 44.6 62 26.5 80 61.1 127.4 20.7

500 500 500 500 2.7361025 2.8561025 2.2461025 1.5361025 2.561025 2.2861025 2.1461025 2.0361025 2.2161025 1.7761025 2.0361025 2.5161025 3.261025 3.8761025 1.1561025 1.3561025

214

600 600 600 600 600 600 600 600 600 600 550 550

215

209 223

Tx1, primary crystallisation temperature; Tx2, secondary crystallisation temperature; Bs, saturation magnetisation; Hc, coercivity; l, magnetostriction; Ta, annealing temperature.


2005

VOL

50

NO

3

183

Sundar and Deevi


43 Variation of coercivity of Fe45Co43Cu1B3.6Zr7.4 alloy with annealing temperature (reprinted from Ref. 217, ß2002 IEEE)

to 65 : 35 as well as partially substituting the glass forming elements (Zr and B) with transition metals (Ta and Mo) on the phase transition and crystallisation temperatures. The primary crystallisation temperature is not affected by the above compositional changes. Similarly, changing the Fe to Co ratio did affect the a– c phase transition. On the other hand, the a–c transition temperature is suppressed by Mo and Ta additions. Partial replacement of Zr with Nb increases the primary crystallisation temperature and decreases the Curie temperature (a–c) of the amorphous phase.132 Willard et al.216 analysed the kinetics of primary crystallisation of a Zr-based HITPERM alloy by a Kissinger analysis of DSC data. The activation energy for primary crystallisation was found to be 323 kJ mol21 and is comparable to those of other Fe-based nanocrystalline alloys. Johnson et al.132 reported that the activation energy for primary crystallisation is increased to 365 kJ mol21 in the copper free, Zr-based HITPERM alloy Fe44.5Co44.5Zr7B4. On the other hand, substitution of Zr with either Mo or Ta decreases the activation energy (,300 kJ mol21). Understanding the crystallisation kinetics will help in the design of new alloys with enhanced extrinsic magnetic properties by controlling the number and distribution of nucleation sites and by refining the grain size.216 Figure 44 compares the thermomagnetic properties of a Fe-based alloy, NANOPERM, and FeCo-based

44 Thermomagnetic properties of HITPERM (Fe44Co44Zr7B4Cu1) and NAANOPERM (Fe88Zr7B4Cu1) alloys (reprinted with permission from Ref. 210, ß1998, American Institute of Physics)

184


2005

VOL

50

NO

3

45 Variation of a AC permeability and b core loss as a function of frequency (at field amplitude of 2.5 Oe) of HITPERM (Fe44Co44Zr7B4Cu1) alloy after 600uC annealing (reprinted with permission from Ref. 211, ß1999, American Institute of Physics)

HITPERM alloy.210,216 The amorphous precursor of NANOPERM has a Tc just above room temperature. The low Curie temperature of the amorphous matrix restricts the maximum operating temperature of this alloy to close to room temperature. On the other hand, the Curie temperature of the amorphous precursor of HITPERM is estimated to be in the range of 800– 850uC.208 After the initial crystallisation treatment, the Curie temperature of the amorphous phase may change owing to segregation of the alloying elements. Thus, the Curie temperature of the amorphous matrix in HITPERM alloy is above the primary crystallisation temperature (y400–500uC). The magnetisation of the amorphous matrix is only partially suppressed before crystallisation. Hence, the amorphous matrix helps to maintain exchange coupling between the nanocrystalline phases at elevated temperatures. The Curie temperature of the nanocrystalline (FeCo) phase in the HITPERM alloy is close to 980uC. Willard et al.211,216 reported the room temperature AC permeability of HITPERM alloys heat treated at 600uC. The variation of the real (m9) and imaginary (m0) components of the permeability with frequency of the heat-treated HITPERM alloy is shown in Fig. 45a. The m0 value, which is related to the power loss because of eddy current and hysteresis losses, exhibits a maximum of 1800 at 20 kHz. The room temperature AC permeability characteristics of HITPERM alloy are comparable to the commercial Hiperco 50 alloy. Further, the alloy has a resistivity of 50 mV cm at room temperature. High resistivity may limit the eddy current losses at high frequencies. The core loss of the alloy as a function of frequency is shown

Sundar and Deevi


46 Magnetisation change in (Fe61.6Co26.4Hf7B4Cu1) sample with annealing time at 600uC (reprinted with permission from Ref. 212, ß1999, American Institute of Physics)

in Fig. 45b.212,217 Typical room temperature core loss of the alloy is 1 W g21 at an induction of 1 T and a frequency of 1 kHz, which is comparable to that of commercial FeCo alloys. The alloy should resist grain growth and suppress the formation of harmful second phases at high temperatures for ‘MEA’ application. Willard et al.207 studied the phase stability of HITPERM alloys by carrying out XRD and TEM studies on samples annealed for different times (up to 3072 h) at 600uC. The grain size of the alloy annealed for 3072 h at 600uC is about 61 nm whereas the grain size of the alloy annealed for a period of 10 h to obtain optimum magnetic properties is 49 nm. Further, TEM examination of samples annealed for 3072 h at 600uC revealed the formation of minor (Fe,Co)3Zr and ZrO phases. Iwanabe et al.212 reported changes in microstructure and magnetic properties of (Fe70Co30)88Hf7B4Cu1 alloy with annealing time (up to 1000 h) at 600uC. The grain size of the alloy annealed up to 20 h at 600uC is about 10 nm and grew to 30 nm when annealed for 100 h. Grain growth was accompanied by an increase in coercivity of the alloy. Samples annealed for 500 h or longer at 600uC exhibited Co2Hf and Co23Hf6 type precipitates. Precipitation of these Co rich precipitates depletes Co from the amorphous matrix and may lead to reduction in magnetisation at room temperature when annealed for longer times (Fig. 46). Kulik et al.217 reported that the crystallisation temperature strongly influences the thermal stability of the magnetic properties of HITPERM alloy. HITPERM alloys were developed by substituting Co to the original NANOPERM (Fe–Zr–Cu–B) alloy composition. Go´mez-Polo and colleagues227,228 reported the effect of Co addition on the structure and high temperature magnetic properties of FINMET (FeSiBNbCu) alloy. The addition of Co promoted the formation of Co rich-FeCoSi precipitates (L12) during primary crystallisation (at 550uC) of the amorphous precursor. The addition of Co improved the high temperature magnetic properties (.300uC) of FeSiBNbCu alloy and an alloy with a composition of Fe43.5Co30Si13.5B9Cu1Nb3 exhibited the best magnetic properties (Fig. 47). The improvement in high temperature magnetic properties is attributed to the enhanced coupling between the nanocrystalline and amorphous phases resulting from an increase in the Curie temperature of the ferromagnetic precipitate. With a further

47 Temperature dependence of a real (mr/mr) and imaginary (mi/mr) components of magnetic permeability of Fe73.52xCoxSi13.5B9Cu1Nb3 (where x50, 30, 45); Curie temperatures of residual amorphous phase (Tc2) are indicated with arrows in b (reprinted with permission from Ref. 227, ß2002, American Institute of Physics)

increase in Co content, the magnetic properties of the alloy deteriorated due to a decrease in the Curie temperature of the residual amorphous phase. The above observations confirm the contributions of both nanocrystalline and residual amorphous phases towards the improvement in the high temperature magnetic properties of bulk nanocrystalline alloys. In summary, FeCo-based nanocrystalline materials exhibit promising magnetic properties for high temperature applications. The superior soft magnetic properties rely on both the nanocrystalline FeCo and the residual amorphous phase at the grain boundaries. However, long annealing degrades the microstructure and the magnetic properties.212 Successful elevated temperature applications require alloy design to improve the thermal stability of both amorphous and nanocrystalline phases at the operating conditions. Only limited information is available on the mechanical properties of nanocrystalline FeCo alloys. Recently, Duckham et al.229 reported the tensile properties of ultra-fine grained FeCo–2V alloy over the temperature range of 25–500uC. Bulk nanocrystalline FeCo–V alloy, with a grain size in the range of 100–300 nm, was produced by heat treating a commercially available, cold-rolled FeCo–2V alloy between 440 and 650uC. The room temperature yield strength of these bulk nanocrystalline alloys is extremely high in the range of 1.0– 2.0 GPa. The high strength of these alloys is attributed to the significant contribution from the higher grain


2005

VOL

50

NO

3

185

186


2005

VOL

50

NO

3

Bs, saturation magnetisation; Hc-east, easy axis coercivity; f, resonance frequency; Hc-hard, hard axis coercivity; m, effective permeability; Hk, anisotropy of magnetic field; r, resistivity; l, magnetostriction.

244 249 4.0 2.8 2.0 >1.4 FeCoB (200 nm)/[Ti/glass (thermal annealed)] (Fe0.7Co0.3)88Zr7B4Cu1 (100 nm)/[SiO2]

– 2.4–10.4 2.1–2.5 2.2

Columnar grain (200 A˚) Nanocrystalline needles of FeCo in amorphous matrix Amorphous Amorphous

0.1–2.8 –

0.1–2.4 0.1

12–90 3.2–9.5 0.8–14.3

18-57

– .5061026

–

248 241

246 247 238 2000 1000 – 4761026 5–761025

115 94 40–200

3.1 3.2 1.0–3.2 2.4 1.6–2.0 1.8 2.1 1.7–2.3 2.3 2.4

FeCoZr FeCoB Fe–Co–Ta–N (250 nm)/[Si (100) substrate] Fe65Co35(50 nm)/[Cu-underlayer/glass] (Fe66Co34)1002x(Al2O3)x(50–3) (50–600 nm)/[NiFe/SiO2] (Fe54Co46)12xCrx (1000 A˚)/[Si/SiO2 substrate] (Fe65Co35)90B10 (1850–2250 A˚)/[NiFeCr]

Amorphous Amorphous fcc TaNza-Fe or a-FeCo (5–15 nm) Columnar grain (10 nm) –

2.4–8.0 –

Hc-soft, kA m21 Hc-hard, kA m21 Hk, kA m21 B s, T Microstructure (grain size) Material/film thickness

Table 16 Summary of magnetic properties of FeCo thin films

The data storage capacity of magnetic recording media has been increasing rapidly in recent years. Perpendicular magnetic recording is one of the promising technologies that could achieve areal densities as high as one terabit per square inch (Tb in–2), roughly 20 times the density of today’s state-of-the-art disc drive products. Perpendicular recording utilises recording media with a soft magnetic underlayer (SUL) which effectively increases the field during writing, lowers the demagnetisation fields in the recorded layer, promotes stronger playback signals and allows the use of media with higher anisotropies. A soft magnetic underlayer with a high moment allows the soft magnetic underlayer thickness to be reduced, which is critical in many potential applications. Efforts are under way to develop high moment thin films based on FeCo for SUL and write-head applications. The structure and magnetic properties of FeCo based thin films are summarised in Table 16. Amorphous or fine grained films with grain sizes smaller than the exchange length exhibit good soft magnetic properties. The as-deposited FeCo films typically show in-plane nearly isotropic square hysteresis loops with high coercivity (Hc.6.0 kA m21).231–234 It is possible to reduce the coercivity values to below 0.8 kA m21 by depositing the film on a suitable underlayer.232–239 Reduction in the grain size of the film owing to the underlayer is responsible for the observed reduction in coercivity. As in the case of Fe based thin films, FeCo based thin films exhibit large values of magnetostriction (5–761025).240 When using magnetic materials with non-zero magnetostriction, it is necessary to optimise the processing conditions to achieve positive magnetoelastic anisotropy to force the magnetisation to lie inplane. Minor et al.241 have shown that positive magnetoelastic anisotropy can be achieved in nanocrystalline FeCoB thin films (produced by depositing on NiFeCr) by creating tensile residual stresses in the film. The residual stress in the film changes from the compressive to the tensile state with increase in sputtering pressure, and the magnetic characteristics of the film change from magnetically isotropic to magnetically uniaxial with low coercivity (Fig. 48).Theoretical and experimental studies by Zou et al.242,243 showed that zero or tensile residual stress is beneficial in suppressing the undesirable strip domain (caused by perpendicular anisotropy, which leads to high coercivity and low permeability) formation in a Fe35Co65 film. Similarly, Yu et al.244 reported that strip domains and the associated noise levels in amorphous FeCoB thin films can be greatly reduced or eliminated through proper thermal annealing.

r, mV cm

FeCo thin films

0.5 0.4 – 0.6–1.0 –

l

m (f)

Reference

boundary strengthening observed in the ordered condition.36,98 Further, these nanocrystalline alloys exhibit at least 3% room temperature ductility in the ordered condition and only a moderate decrease in yield strength up to 400uC. Cheng et al.230 reported the AC core loss of these bulk nanocrystalline alloys at various frequencies up to 4500 kHz. The core loss of the alloy is found to decrease with decreasing frequency and increasing grain size. Based on the variation in total power loss with frequency, they identified that the major contribution of the core loss of the alloys is due to hysteresis loss.

245


0.2 0.3 0.2–1.9 0.1–0.2 0.1

Sundar and Deevi

Sundar and Deevi

48 Effect of process pressure on magnetic properties [easy axis coercivity (Hc), uniaxial anisotropy (Hk) and magnetostriction at 50 Oe (l50)] and film stress of as deposited FeCoB thin films (reprinted with permission from Ref. 241, ß2002, American Institute of Physics)

Noise due to domain instability of the soft magnetic layer is a serious problem and deteriorates the signal quality. Noise due to domain walls can be reduced or avoided by pinning the soft layer domain using an exchange coupled hard magnetic layer or by making the SUL nanocrystalline or granular. A soft underlayer with a granular structure using nonconductive matrix materials has the added advantage of improving the performance at high frequencies. Soo et al.250 explored the use of FeCoC granular and [FeCo/C]n laminated films for soft magnetic underlayers in perpendicular recording media. Both the films exhibit fine nanocrystalline granular microstructures. Laminated multilayer [FeCo/ C]n films showed good soft magnetic properties (Bs.1.4 T and a relative permeability of 300 up to 10 MHz) and the properties could be tailored by varying the carbon concentration and the number of layers. Read/write tests performed on [Co/Pd]n/NiP media with and without a FeCoC soft underlayer revealed that a higher recording density without any spike in noise is achieved when FeCoC is used as a SUL. Laughlin and colleagues249,251–253 have explored the possibilities of using thin films based on HITPERM composition as a soft magnetic underlayer for perpendicular recording media. They deposited single and multiple layers of HITPERM [(Fe0.7Co0.3)88Zr7B4Cu1] films on glass or SiO2 (100) substrates by rf sputtering. They investigated the effect of individual layer thickness, number of stacking layers, and sputtering power density on the microstructure and magnetic properties of the deposited films. An amorphous structure was observed in HITPERM films of 100 nm thickness when sputtered at room temperature, with a power density of 2.3 W cm22, while the films deposited at 250uC turned into nanocrystalline microstructure with ferromagnetic [a (FeCo)] and nonmagnetic (Fe2Zr) phases. On the other hand, the microstructures of the films deposited at sputter power density >4.5 W cm22 consist of nanocrystalline a (FeCo) particles in an amorphous phase. As discussed in the previous section, the primary and secondary crystallisation temperatures for bulk HITPERM alloys produced by rapid solidification are above 250uC whereas the crystallisation temperatures for sputtered films are between room temperature and 250uC (depending on the sputtering power density).


49 Magnetic properties of HITPREM films as a function of film thickness (reprinted with permission from Ref. 251, ß2002, American Institute of Physics)

Higher surface diffusion rates and lower nucleation energy barriers (due to the high energy of the sputtered atoms) are considered to be responsible for the lower crystallisation temperature in the sputtered films. Magnetic properties of HITPERM films as a function of film thickness are shown in Fig. 49. The degradation in soft magnetic properties at greater thicknesses is due to an increase in grain size and decrease in the exchange coupling as a result of increased separation between the grains. Magnetic properties of 100 nm thick films as a function of sputtering power density are shown in Fig. 50. The saturation magnetisation (1.3 T) of the film deposited at a sputtering density of 2.3 W cm22 did not change with the deposition temperature. On the other hand, the coercivity of the films deposited at 2.3 W cm22 and at 250uC is higher than that deposited at room temperature. The higher coercivity of the films deposited at 250uC is attributed to the presence of nonmagnetic Fe2Zr particles and to the higher average magnetic anisotropy of the nanocrystalline phase compared to the amorphous phase present in the films deposited at room temperature. Films produced at higher sputter densities (>4.5 W cm22) showed lower coercivity (0.3–0.4 kA m21) and higher saturation magnetisation (1.9–2.0 T). The presence of a higher volume fraction of ferromagnetic nanocrystalline a(FeCo) phase in the amorphous matrix is believed to be responsible for

50 Magnetic properties of HITPREM thin films as a function of sputtering power density (reprinted with permission from Ref. 253, ß2003, American Institute of Physics)


2005

VOL

50

NO

3

187

Sundar and Deevi


the observed lower coercivity and higher saturation magnetisation of the films deposited at higher power densities. The microstructure and magnetic properties of the films deposited at higher power densities did not change even after heat treating at 250uC. Further, a (00.2) textured CoCrPt magnetic recording layer with a thin Ti intermediate layer (to exchange couple the recording media and SML) could be grown over the HITPERM films by depositing the films at 250uC at power densities greater than 4.5 W cm22. It is important to note that the HITPERM films could retain their good soft magnetic properties and promote the growth of a strong (00.2) textured CoCrPt magnetic layer required to achieve thermally stable media with enhanced signalto-noise ratio (SNR). Okumura et al.252 characterised the microstructure and high frequency magnetic properties of HITPERM/SiO2 films by TEM and ferromagnetic resonance (FMR) methods, respectively. The as-deposited HITPERM film exhibits a fine nanocrystalline ferromagnetic a (FeCo) phase dispersed in an amorphous matrix. The volume fraction of nanocrystalline phase is less than 10%. The saturation magnetisation of the films determined from FRM measurement is about 1.45–1.5 T. The Landau-Lifshiftz-Gilbert damping parameters of the single and multilayered films are small (0.0055¡0.0004) and are comparable to that of a bulk alloy. Neither variation in thickness of each layer nor stacking number affected the damping characteristics. The soft magnetic properties of the films could further be improved by increasing the volume fraction of ferromagnetic phase through in situ heating during deposition and post-annealing treatment. Fe–Co–Ta–N thin films (250 nm) prepared by rf magnetron sputtering display good soft magnetic properties in the as deposited condition.246 Further, these films exhibit excellent high-frequency characteristics (with an effective permeability of 2000 up to 70 MHz) and high resistivity (100 mV cm). The corrosion resistance (in 0.5M NaCl) of the film is better than that of Co–Ni–Fe and Fe–Hf–N thin films. Vas’ko ˚) et al.248 proposed the use of FeCoCr films (y1000 A for magnetic recording head applications. They studied the effect of Cr additions and field annealing on the saturation magnetisation, resistivity and corrosion resistance of (Fe54Co46)12xCrx films. The saturation magnetisation decreases and the resistivity increases with Cr additions. As deposited films with and without Cr doping exhibited coercivities in the range of 6.4– 9.5 kA m21, which substantially decreased to 3.2 kA m21 after magnetic annealing at 220uC. The aqueous corrosion resistance of FeCoCr films (under the conditions encountered during fabrication) is comparable to that of the widely used Ni45Fe55 films.248 The magnetic properties of FeCo thin films (reported to date) show promise for many thin film applications (Table 17).245–249,251–255 However, further characterisation of the FeCo films is Table 17 Applications of FeCo thin films Recording head poles Soft magnetic underlayer in perpendicular recording media Giant magnetoresistance sensors Magnetoresistive memories (as multilayer) High frequency magnetic devices Specular GMR spin-valve Magnetic tunnel devices Magnetic microactuators

188


2005

VOL

50

NO

3

51 a effect of W fibre volume fraction on yield stress and saturation magnetisation of FeCo–W composite; b high temperature creep of FeCo–W composite in comparison with commercial FeCo–2V alloy (reprinted from Ref. 149, ß2000 IEEE)

necessary to elucidate the advantages of these films over other leading candidates like Fe–X–N films, where X is Ta, Ti, Al, etc. Thin films of FeCo alloys are also being considered for microelectro-mechanical applications such as magnetic micro-actuators. Various deposition methods were attempted to produce thin films required for microelectro-mechanical applications. For example, Shao et al.256 reported deposition of crack free FeCo films of up to several hundred micrometres from sulfamate-based solution. The composition of the film is found to be dependent on the deposition current density, deposition potential, ion concentration, additives, pH, deposition temperature, and agitation of the solution. They found that addition of vanadium to the solution greatly enhanced the film quantity. The magnetic properties of these thin films were found to be similar to the bulk FeCo material. Yu et al.149,257 reported the mechanical and magnetic properties of FeCo composites reinforced with tungsten fibres and Al2O3 particulates. These composites are produced by electrodeposition. The composite derives its unique soft magnetic properties from the pure FeCo matrix, and the reinforced fibres and particulates offer superior mechanical properties. Low coercivity and high permeability are achieved in the composites after high temperature annealing. Interestingly, the composites exhibited high strength (Fig. 51a) and superior creep resistance (Fig. 51b) compared to the commercial bulk FeCo alloys. It is possible to further improve the magnetic and mechanical properties of the composites by optimising the magnetic matrix and fibre network.

Future research Worldwide research efforts have enhanced the understanding of phase relations and structure–property

Sundar and Deevi

relations in FeCo alloys. However, further studies are clearly needed to bring out the effects of alloy composition on the creep, fatigue, and high temperature oxidation resistance of these alloys. For example, the limited information available indicates the sensitivity of fatigue behaviour to the environment. Additional studies are needed to fully understand the impact of environment on fatigue resistance. Similarly, it is necessary to characterise the high temperature creep behaviour of some of the promising alloy compositions. Earlier studies indicate that the mechanical properties of these alloys are notch-sensitive.94 However, systematic studies are necessary to bring out the notch sensitivity of the alloy under different microstructural conditions. The effects of alloying additions and processing conditions on the magnetic properties of alloys require more careful examination. The majority of the investigations are related to DC magnetic properties of the alloys. However, most of the emerging applications involve AC fields at higher frequencies. Hence, it is mandatory to characterise the magnetic behaviour of the alloy under AC fields. It is well known that the presence of impurities in soft magnetic materials is detrimental to their soft magnetic properties.8,56 However, there is very little or no information available on the effect of trace elements on the magnetic properties of alloys based on FeCo. Knowledge of the site occupancy of alloying elements is important in understanding and controlling the properties of FeCo alloys. Recently Bozzolo et al.258 reported the site occupancy of various alloying elements in several B2 intermetallic compounds including FeCo. Experimental verification of their results is warranted. Similarly, fundamental information such as the effects of alloying additions on the APB energies is lacking. Availability of this information will help in the design of better alloys with improved properties. Despite the current interest in FeCo–V alloys for high temperature applications, there is no information available on the oxidation behaviour of FeCo alloys. The oxidation resistance of commercial FeCo alloys is expected to be poor as the alloys do not have Cr or Al which form a protective oxide layer. In addition, for AC applications such as the rotor of a power generating unit, it is necessary to electrically insulate the laminates from each other to minimise the eddy current losses. Hence, a fundamental understanding of the oxidation of FeCo alloy is highly recommended. Current FeCo compositions are based on the presence of a second phase to improve their strength. However, it may be better to develop alloys by solid solution strengthening. There are many advantages to use solid solution strengthened alloys over precipitationstrengthened alloys for soft magnetic applications.56 For example, the degradation in magnetic properties owing to alloying additions can be minimised when the additions are restricted to below their solubility limit (Table 12). In addition, the increase in resistivity due to solute additions will be higher when the solutes are in solid solution. Solid solution alloys are expected to be more stable over the precipitation-strengthened alloys for long term high temperature applications. Unfortunately, solid solubility of several beneficial alloying elements in FeCo is less than 1.0 at.-%.115


Current efforts to develop nanocrystalline alloys for high temperature soft magnetic applications have resulted in alloys with promising soft magnetic properties. However, there is no information available on the mechanical properties of these alloys. Further research and development efforts are needed to consolidate the ribbons produced by rapid solidification processing to form bulk material and to characterise their mechanical and magnetic properties. Similarly, the long term stability of the alloys need to be evaluated to develop new alloys with improved stability. Finally, chemical methods may be employed to obtain nanopowders of FeCo alloys. Novel processing techniques based on powder processing approaches such as roll compaction, tape casting, plasma processing and self-assembly need to be investigated to obtain thin sheets for MEA applications. Powder processing approaches will enable incorporation of hard phases to enhance the strength and creep resistance, and insulating phases to increase the overall electrical resistivity of the alloy. Powder processing approaches combined with the chemical processing methods may pave the way for significantly superior soft magnetic products for the 21st century and beyond.

Concluding remarks FeCo–V alloys constitute an important class of intermetallic compounds and have been in commercial use for more than seven decades in high performance strategic applications. Previous research efforts significantly enhanced the understanding of phase relations and the effect of alloying additions/compositions on the magnetic, electrical, and mechanical properties of FeCo alloys. The mechanical and magnetic properties of FeCo alloys are highly susceptible to composition, processing and heat treatments. Often, an improvement in mechanical properties is achieved at the expense of magnetic properties. Hence, it is necessary to optimise the alloy composition and thermo-mechanical treatment to achieve balance between the mechanical and magnetic properties. Emerging applications for soft magnetic materials at high temperatures require further improvement in mechanical properties of FeCo alloys. The recent advances in amorphous and nanocrystalline materials suggest that these approaches may be combined with conventional alloys to balance the mechanical, electrical and magnetic properties of FeCo alloys.

References 1. G. E. Fish: Proc. IEEE, 1990, 78, 947–972. 2. T. Abraham: JOM, January 1995, 16–18. 3. J. S. Kouvel: in ‘Intermetallic compounds’, (ed. J. H. Westbrook), 529–568; 1967, Wiley. 4. H. H. Stadelmaier and B. Reinsch: in ‘Intermetallic compounds – Magnetic, electrical and optical properties and applications of intermetallic compounds’, (ed. J. H. Westbrook and R. L. Fleischer), 31–50; 2000, John Wiley & Sons. 5. A. Preuss: ‘Magnetic properties of Fe–Co alloys at different temperatures’, dissertation, University of Zurich, 1912. 6. P. Weiss: Trans. Faraday Soc., 1912, 8, 149–156. 7. P. Weiss and R. Forrer: Ann. Phys. (Paris), 1929, 12, 279–374. 8. R. M. Bozorth: ‘Ferromagnetism’; 1951, D. Van Nostrand Co. Inc. 9. J. K. Stanley and T. D. Yensen: Trans. Am. Inst. Elec. Eng., 1947, 66, 714–718. 10. R. E. Quigley: Proc. IEEE Applied Power Electronics Conf. ‘More electric aircraft’, 1993, 906–911.


2005

VOL

50

NO

3

189

Sundar and Deevi


11. R. T. Fingers, J. E. Coate and R. E. Dowling: Proc. Intersoc. Energy Conversion Eng. Conf., vol. 1, 1997, 563–568. 12. T. Nishizawa and K. Ishida: in ‘Binary alloy phase diagrams’, vol. 2 (ed. T. B. Massalski), 1186; 1990, Metals Park, OH, ASM. 13. G. V. Raynor and V. G. Rivlin: Int. Met. Rev., 1983, 28, 211–227. 14. M. L. Glazyrina, A. M. Glezer, B. V. Molotilov, S. M. Tret’yakova and V. I. Kleynerman: Fiz. Met. Metalloved., 1983, 56, 93–100. 15. S. Kaya and H. Sato: Proc. Phys. Math. Soc., Japan, 1943, 25, 261–273. 16. J. Cowley: Phys. Rev., 1950, 77, 669–675. 17. G. J. Dienes: Acta Metall., 1955, 3, 519–557. 18. N. S. Stoloff and R. G. Davies: Acta Metall., 1964, 12, 473–485. 19. A. W. Smith and R. D. Rawlings: Phys. Status Solidi (a), 1976, 34, 117–123. 20. M. Rajkovic and R. A. Buckley: Met. Sci., 1981, 15, 21–29. 21. D. W. Clegg and R. A. Buckley: Met. Sci. J., 1973, 7, 48–54. 22. R. A. Buckley: Met. Sci. J., 1975, 9, 243–247. 23. A. T. English: Trans. Met. Soc. AIME, 1966, 236, 14–20. 24. J. P. Eymery, P. Grosbras and P. Moine: Phys. Status Solidi (a), 1974, 21, 517–528. 25. P. Grosbras, J. P. Eymery and P. Moine: Acta Metall., 1976, 24, 189–196. 26. J. A. Rogers, H. M. Flower and R. W. Rawlings: Metall. Sci., 1975, 9, 32–35. 27. W. C. Ellis and E. S. Greiner: Trans. Am. Soc. Metal., 1941, 29, 415–432. 28. I. Baker: Mater. Sci. Eng., 1995, A 192/193, 1–13. 29. J. P. Neumann: Acta Metall., 1980, 28, 1165–1170. 30. J. P. Neumann, Y. A. Chang and C. M. Lee: Acta Metall., 1976, 24, 593–604. 31. I. M. Pike, Y. A. Chang and C. T. Liu: Mater. Res. Symp. Proc., 1995, 364, 65–70. 32. L. von Mises: Z. Angew. Math. Mech., 1928, 8, 161–164. 33. M. J. Marcinkowski and H. Chessin: Philos. Mag., 1964, 10, 837– 859. 34. T. L. Johnston, R. G. Davies and N. S. Stoloff: Philos. Mag., 1965, 12(116), 305–317. 35. M. Yamaguchi, Y. Umakoshi and T. Yamane: Scripta Metall., 1982, 16, 607–609. 36. L. Zhao and I. Baker: Acta Metall. Mater., 1994, 42, 1953–1958. 37. I. Baker and E. M. Schulson: Scripta Metall., 1989, 23, 345–348. 38. W. L. Bragg and E. J. Williams: Proc. R. Soc., 1935, A151, 540– 566. 39. S. T. Fong, K. Sadananda and M. J. Marcinkowski: Metal. Trans., 1974, 5, 1239–1247. 40. R. V. Major and V. Samadian: J. Mater. Eng., 1989, 11, 27–30. 41. L. Zhao and I. Baker: J. Mater. Sci., 1994, 29, 742–748. 42. Y. V. R. K. Prasad and T. Seshacharyulu: Int. Mater. Rev., 1998, 44, 243–258. 43. H. Hosoda, S. Miyazaki, K. Inoue, T. Fukui, K. Mizuuchi, Y. Mishima and T. Susuku: J. Alloys Comp., 2000, 302, 266–273. 44. W. Mao, G. Zhu and Y. Yu: Z. Metallkd., 2000, 91, 211–214. 45. D. Raabe and K. Lu¨cke: Mater. Sci. Forum, 1994, 157–162, 597– 610. 46. R. W. Cahn: in ‘High temperature aluminides and intermetallics’, (ed. S. H. Whang et al.), 245–270; 1990, Warrendale, PA, TMS. 47. R. G. Davies and N. S. Stoloff: Trans. Met. Soc. AIME, 1966, 236, 1605–1608. 48. P. Moine, J. P. Eymer and R. Grosbras: Phys. Status Solidi (b), 1971, 46, 177–185. 49. M. J. Marcinkowski and R. M. Fisher: Trans. Met. Soc. AIME, 1965, 233, 293–298. 50. M. J. Marcinkowski: Mem. Sci. Rev. Metall., 1972, 69, 579–590. 51. A. H. Cottrell: Trans. TMS-AIME: 1958, 212, 192–203. 52. N. J. Petch: Philos. Mag., 1958, 3, 1089–1097. 53. I. Baker, P. Nagpal, F. Liu and P. R. Munroe: Acta Metall., 1991, 39, 1637–1644. 54. L. Zhao, I. Baker and E. P. George: Mater. Res. Symp. Proc., 1993, 288, 501–506. 55. J. P. Horton, I. Baker and M. H. Yoo: Philos. Mag. A, 1991, 63, 319–335. 56. G. Y. Chin and J. H. Wernick: ‘Ferro magnetic material’, vol. 2, (ed. E. P. Wonlforth), 55–188; 1980, North-Holland Pub. Co. 57. M. S. Lataifeh: J. Phys. Soc. Jpn, 2000, 69, 248–250. 58. R. Richter and H. Eschring: J. Phys., F: Metal Phys., 1988, 18, 1813–1826. 59. N. Hamada: J. Phys. Soc. Jpn, 1979, 46, 1759–1766. 60. J. M. MacLaren, T. C. Schulthess, W. H. Butler, R. Sutton and M. McHenry: J. Appl. Phys., 1999, 85, 4833–4835.

190


2005

VOL

50

NO

3

61. W. C. Ellis: Renselaer Polytech. Inst. Bull. Eng. Sci. Ser., 1927, 16, 1–57. 62. G. W. Elmen: U.S. Patent 1 739 752, 1929. 63. M. F. Collins and J. B. Forsyth: Philos. Mag., 1963, 8, 401–410. 64. K. K. Bogma: Fiz. Met. Metalloved., 1966, 22, 148–151. 65. S. Y. Chu, C. Kline, M. Q. Huang, M. E. McHenry, J. Cross and V. G. Harris: J. Appl. Phys., 1999, 85, 6031–6033. 66. R. C. Hall: Trans. TMS-AIME, 1960, 218, 268–270. 67. J. H. White and C. V. Wahl: US Patent 1 862 559, 1932. 68. H. L. B. Gould and D. H. Wenny: AIEE Spec. Pub., 1957, T-97, 675–681. 69. E. A. Nesbitt and G. A. Kelsall: Phys. Rev, 1940, 58, 203. 70. E. A. Nesbitt: Trans. TMS-AIME, 1946, 166, 415–425. 71. D. H. Wenny: Bell Lab. Rec., 1965, 43, 257–261. 72. C. W. Chen: J. Appl. Phys., 1961, 32, 348S–355S. 73. Q. Zhu, L. Li, M. S. Masteller and G. J. Del Corso: Appl. Phys. Lett., 1996, 69, 3917–3919. 74. A. I. C. Persiano and R. D. Rawlings: Phys. Status Solidi (a), 1987, 103, 547–556. 75. A. Williams, G. H. Kwei, A. T. Ortiz, M. Karnowski and W. K. Warburton: J. Mater. Res., 1990, 5, 1197–1200. 76. W. Ko¨ster and K. Lang: Z. Metallkd., 1938, 30, 350–352. 77. W. Ko¨ster and H. Schmid: Arch. Eisenhu¨ttenwes., 1955, 26, 345– 354. 78. W. Ko¨ster and H. Schmid: Arch. Eisenhu¨ttenwes., 1955, 26, 421– 425. 79. J. E. Bennett and M. R. Pinnel: J. Mater. Sci., 1974, 9, 1083–1090. 80. D. L. Martin and A. H. Geisler: Trans. Am. Soc. Met., 1952, 44, 461–481. 81. E. Josso: IEEE Trans. Magn. Mag., 1974, 10, 161–165. 82. S. Mahajan, M. R. Pinnel and J. E. Bennet: Metall. Trans., 1974, 5, 1263–1272. 83. J. A. Ashby, H. M. Flower and R. D. Rawlings: Met. Sci., 1977, 11, 91–96. 84. C. Kuroda and T. Ogawa: Jpn J. Appl. Phys., 1966, 5, 733–734. 85. H. C. Fiedler and A. M. Davis: Metall. Trans., 1970, 1, 1036– 1037. 86. M. R. Pinnel and J. E. Bennett: Bell Syst. Tech. J., 1973, 52, 1325– 1340. 87. V. I. Gorbukov and B. G. Livshits: Fiz. Met. Metalloved., 1961, 12, 534–540. 88. J. E. Bennett and M. R. Pinnel: Metall. Trans., 1974, 5, 1273– 1283. 89. S. G. Fishman, D. Gupta and D. S. Lieberman: Phys. Rev., 1970, B2, 1451–1460. 90. M. S. Masteller and D. W. Dietrich: US Patent US05501747, 1996. 91. F. W. Ackerman, R. T. Casani, W. A. Klawitter and G. B. Heydt: US Patent US3634072, 1972. 92. K. Kawahara: J. Mater. Sci., 1983, 18, 3427–3436. 93. A. H. Geisler, J. P. Martin, E. Both and J. H. Crede: Trans. AIME, 1953, 197, 813–819. 94. R. M. Pinnel, S. Mahajan and J. E. Bennett: Acta Metall., 1976, 24, 1095–1107. 95. K. Kawahara: J. Mater. Sci., 1983, 18, 3437–3448. 96. M. L. Storch, A. D. Rollett and M. E. McHenry: J. Appl. Phys., 1999, 85, 6040–6042. 97. N. S. Stoloff and R. G. Davies: Prog. Mater. Sci. 1966, 13, 3–84. 98. K. R. Jordan and N. S. Stoloff: Trans. TMS–AIME, 1969, 245, 2027–2034. 99. Y. G. Koylu, G. F. Hancock and R. D. Rawling: Phys. Status Solidi (a), 1973, 16, 73–80. 100. D. R. Thornburg: J. Appl. Phys., 1969, 40, 1579–1580. 101. D. M. Pavlovic and F. G. Stone: INTERMAG Conf., Amsterdam, Netherlands, IEEE Trans. Magn., 1969, 390–391. 102. D. Caillard: Mater. Sci. Eng. A, 2001, 319–321, 74–83. 103. P. Veyssie`re: Mater. Sci. Eng. A, 2001, 309–310, 44–48. 104. R. E. Reed-Hill: ‘Physical metallurgy principles’; 1964, D. Van Nostrand Co., Inc. 105. C. H. Shang, R. C. Cammarata, T. P. Weihs and C. L. Chien: J. Mater. Res., 2000, 15, 835–837. 106. N. S. Stoloff and I. L. Dillamore: ‘Ordered alloys: structural applications and physical metallurgy’, (ed. B. H. Kear, C. T. Sims, N. S. Stoloff and J. H. Westbrook), 525–544; 1970, Baton Rouge, LA, Claitors Pub. Div. 107. C. D. Pitt and R. D. Rawlings: Met. Sci., 1983, 17, 261–266. 108. E. P. George and C. T. Liu: ‘High temperature ordered intermetallic alloys VI’, (ed. J. A. Horton, I. Baker, S. Hanada, R. D. Noebe and D. Schwartz), MRS Symp. Proc., 1995, 364, 1131–1145; Pittsburgh, PA, MRS.

Sundar and Deevi

109. E. P. George, A. N. Gubbi, I. Baker and L. Robertson: Mater. Sci. Eng., 2002, A329–331, 325–333. 110. A. K. Kuruvilla: PhD thesis, Rensselaer Polytechnic Institute, Troy, NY, 1985. 111. C. H. Shang, T. P. Weihs, R. C. Cammarata, Y. Ji and C. L. Chien: J. Appl. Phys., 2000, 87, 6508–6510. 112. D. M. Dzhavadov and Ya. P. Selisskiy: Phys. Met. Metallogr., 1963, 15, 504–510. 113. D. M. Dzhavadov and Ya. P. Selisskiy: Phys. Met. Metallogr., 1964, 18, 790–793. 114. L. A. Alekseyev, D. M. Dzhavadov, Yu. D. Tyapkin and R. B. Levi: Phys. Met. Metallogr., 1977, 43, 99–104. 115. K. Kawahara: J. Mater. Sci., 1983, 18, 1709–1718. 116. D. H. Sastry and R. S. Sundar: in Proc. Int. Symp. on ‘Nickel and iron aluminides’, (ed. S. C. Deevi, V. K. Sikka, P. J. Massiasz and R. W. Chan), 123–144; 1997, ASM Int. 117. Y. Iijima and C. G. Lee: Acta Metall. Mater., 1995, 43, 1183– 1188. 118. R. T. Fingers: ‘Creep behavior of thin laminates of FeCo alloys for use in switched reluctance motors and generators’; 1999, AFRL-PR-WP-TR-1999–2053. 119. R. T. Fingers, J. E. Coate and N. E. Dowling: J. Appl. Phys., 1999, 85, 6037–6039. 120. Hiperco Alloy 50HS data sheet, Electronic Alloys #39, Carpenter Specialty Alloys, 1995. 121. J. F. Dinhut, T. Bonou and P. Moine: Acta Metall., 1976, 24, 445– 451. 122. S. Liu, S. Bauser, Z. Turgut, J. Coate and R. T. Fingers: J. Appl. Phys., 2003, 93, 7118–7120. 123. R. C. Boettner, N. S. Stoloff and R. G. Davies: Trans. TMSAIME, 1966, 236, 131–133. 124. N. S. Stoloff, G. E. Fuchs, A. K. Kuruvilla and S. J. Choe: in ‘High temperature ordered intermetallic alloys II’, (ed. N. S. Stoloff, C. C. Koch, C. T. Liu and O. Izumi), MRS Symp. Proc., 1987, 81, 247–261; Pittsburgh, PA, MRS. 125. N. S. Stoloff, S. J. Choe and K. Rajan: Scripta Metal. Mater., 1992, 26, 331–336. 126. H. D. Williams and G. C. Smith: Philos. Mag., 1966, 13, 835–854. 127. S. M. L. Sastry and H. A. Lipsitt: Metall. Trans., 1977, 8A, 299– 308. 128. C. W. Chen: ‘Magnetism and metallurgy of soft magnetic materials’; 1977, New York, North Holland. 129. R. S. Sundar and S. C. Deevi: Intermetallics, 2004, 12, 921–927. 130. B. V. Reddy, S. C. Deevi and S. N. Kanna: J. Appl. Phys., 2003, 93, 2823–2827. 131. V. A. Gubanov, A. I. Liechtenstein and A. V. Postnikov: ‘Magnetism and the electronic structure of crystals’, 149–151; 1992, New York, Springer-Verlag. 132. F. Johnson, P. Hughes, R. Gallagher, D. E. Laughlin, M. E. McHenry, M. A. Willard and V. G. Harris: IEEE Trans. Magn., 2001, 37, 2261–2263. 133. K. A. Gallagher, M. A. Willard, V. N. Zabenkin, D. E. Laughlin and M. E. McHenry: J. Appl. Phys., 1999, 85, 5130–5132. 134. A. I. C. Persiano and R. W. Rawlings: J. Mater. Sci., 1991, 26, 4026–4032. 135. M. Tanaka: US Patent # 05252940, 1991. 136. C. W. Chen: Philos. Mag., 1962, 7, 1753–1764. 137. C. W. Chen: Phys. Rev., 1962, 129, 121–124. 138. H. L. Gould and D. H. Wenny: Elect. Eng., 1957, 76, 208–211. 139. M. R. Pinnel and J. E. Bennett: IEEE Trans. Magn., 1975, Vol. Mag G-11(3), 901–907. 140. S. Mahajan and K. M. Olsen: Proc. Conf. on ‘Magnetism and magnetic materials’, San Francisco, CA, American Inst. Phys., 1974, 743S–744S. 141. K. Kawahara and M. Uehara: J. Mater. Sci., 1984, 19, 2575–2581. 142. C. W. Chen and G. W. Wiener: J. Appl. Phys., 1959, 30, 199S– 201S. 143. D. R. Thornburg and D. A. Colling: Metall. Trans., 1974, 5, 2241–2242. 144. L. Li: J. Appl. Phys. 1996, 79, 4578–4580. 145. R. K. Pandey, M. H. Weichold and D. W. Palmer: IEEE Trans. Magn., 1980, 16, 749–751. 146. P. E. Kueser, D. M. Pavlovic, D. H. Lane, J. J. Clark and M. Speweck: NASA Special Publication No. NASA SP-3043, 1967, 135–160. 147. Z. Turgut, J. C. Horwarth, M.-Q. Huang and R. T. Fingers: J. Appl. Phys., 2003, 93, 6692–6694. 148. R. H. Yu, S. Basu, Y. Zhang, A. Parvizi-Majidi and J. Q. Xiao: J. Appl. Phys., 1999, 85, 6655–6599.


149. R. H. Yu, S. Basu, L. Ren, Y. Zhang, A. P. Majidi, K. M. Unruh and J. Q. Xiao: IEEE Trans. Magn., 2000, 36, 3388–3393. 150. R. H. Yu, S. Basu, Y. Zhang and J. Q. Xiao: J. Appl. Phys., 1999, 85, 6034–6036. 151. E. Hug, O. Hubert and I. Guillot: J. Magn. Magn. Mater., 2000, 215–216, 197–200. 152. E. Hug, O. Hubert and M. Clavel: J. Appl. Phys., 1996, 79, 4571– 4573. 153. W. R. Wieserman, G. E. Schwarze and J. M. Niedra: Proc. 27th Intersociety Energy Conversion Engineering Conference, San Diego, P-259, vol. 1, Society of Automotive Engineers, Inc., Warrendale, PA, 1992, 1.275–1.282. 154. D. Gautard, G. Couderchon and L. Coutu: J. Magn. Magn. Mater., 1996, 160, 359–360. 155. M. F. Littman: J. Appl. Phys., 1967, 38, 1104–1108. 156. R. T. Fingers, R. P. Carr and Z. Turgut: J. Appl. Phys., 2002, 91, 7848–7850. 157. A. J. Moses: AIP Conf. Proc., 1975, 24, 741–742. 158. J. A. Ashby, H. M. Flower and R. W. Rawling: Phys. Status Solidi (a), 1978, 47, 407–414. 159. D. M. Dzhavadov and Ya. P. Selisski: Fiz. Met. Metalloved., 1963, 15, 504–510. 160. B. Geist, T. Peterson, J. C. Horwath, Z. Tugut, M.-Q. Huang, R. A. Snyder and R. T. Fingers: J. Appl. Phys., 2003, 93, 6686– 6688. 161. A. A. Couto and P. I. Ferreira: J. Mater. Eng., 1989, 11, 31–36. 162. H. C. Fiedler: Proc. Conf. on ‘Magnetism and magnetic materials’, San Francisco, CA, 1974, Am. Inst. Phys., 1975, 739–740. 163. R. V. Major, M. C. Martin and M. W. Branson: Proc. 2nd EPS Conf. on ‘Soft magnetic materials’, Wolfson Centre for Magnetics Technology, Wales, UK, 1975, 103. 164. M. W. Brason, R. V. Major, C. D. Pitt and R. D. Rawlings: J. Magn. Magn. Mater., 1980, 19, 222–224. 165. C. D. Pitt and R. D. Rawlings: Met. Sci., 1981, 15, 369–376. 166. K. Kawahara: J. Mater. Sci., 1983, 18, 2047–2055. 167. R. D. Rawlings: US Patent 04933026, 1990. 168. J. K. Stanley: US Patent 2 444 419, 1947. 169. J. K. Stanley and T. D. Yensen: Trans. AIEE, 1947, 66, 714–718. 170. Hiperco 27 Data Sheet, Electronic alloy 40, Carpenter Technologies, 1995. 171. G. Vaerst: Proc. of INTERTECH Conf. ‘SoftMag 2000’, Michigan, 2000, 30–34. 172. C. Radeloff: in ‘Werkstoffkunde der gebrauchlichen stahle’, Teil 2, Dusseldorf: Verlag Stahleisen, 1977, 296–312. 173. F. Pfeifer and C. Radeloff: J. Magn. Magn. Mater., 1980, 19, 190– 207. 174. G. Herzer: IEEE Trans. Magn., 1990, 26, 1397–1402. 175. M. E. McHenry, M. A. Willard and D. E. Laughlin: Prog. Mater. Sci., 1999, 44, 291–433. 176. M. E. McHenry and D. E. Laughlin: Acta Mater., 2000, 48, 223– 238. 177. D. C. Jiles: Acta Mater., 2003, 51, 5907–5939. 178. K. Gallagher, F. Johnson, E. M. Kirkpatrick, J. H. Scott, S. Majetich and M. E. McHenry: IEEE Trans. Magn. 1996, 32, 4842–4844. 179. Z. Turgut, M. Q. Huang, K. Gallagher, M. E. McHenry and S. A. Majetich: J. Appl. Phys., 81, 1997, 4039–4041. 180. Z. Turgut, J. H. Scott, M. Q. Huang, S. A. Majetich and M. E. McHenry: J. Appl. Phys., 1998, 83, 6468–8470. 181. J. H. Scott, Z. Turgut, K. Chowdary, M. E. McHenry and S. A. Majetich: Mater. Res. Soc. Symp. Proc., 1998, 501, 121–126. 182. J. H. Scott, K. Chowdary, Z. Turgut, S. A. Majetich and M. E. McHenry: J. Appl. Phys., 1999, 85, 4409–4411. 183. Z. Turgut, N. T. Nuhfer, H. P. Piehler and M. E. McHenry: J. Appl. Phys., 1999, 85, 4406–4408. 184. Z. Turgut, R. T. Fingers, H. R. Piehler, M. E. McHenry: IEEE Trans. Magn., 2000, 36, 3024–3025. 185. S. Gangopadhyay, G. C. Hadjipanayis, C. M. Sorensen and K. J. Klabunde: Mater. Res. Soc. Symp., 1991, 206, 55–60. 186. R. Elkalkouli, M. Grosbras and J. F. Dinhut: Nanostruct. Mater., 1995, 5, 733–743. 187. X. L. Dong, Z. D. Zhang, S. R. Jin and B. K. Kim: J. Magn. Magn. Mater. 2000, 210, 143–149. 188. G. H. Lee, S. H. Huh, J. W. Jeong, S. H. Kim, B. J. Choi, J. H. Jeong, H. I. Lee and H.-C. Ri: J. Korean Phys. Soc., 2003, 42, 367–370. 189. A. Arrott: Nanostruct Mater., 1999, 12, 1135–1140. 190. R. Bru¨ning, K. Samwer, C. Kuhrt and L. Schultz: J. Appl. Phys, 1992, 72, 2978–2983.


2005

VOL

50

NO

3

191

Sundar and Deevi


191. Y. D. Kim, J. Y. Chung, J. Kim and H. Jeon: Mater. Sci. Eng., 2000, A291, 17–21. 192. X. G. Li, T. Murai, T. Saito and S. Takahashi: J. Magn. Magn. Mater., 1998, 190, 277–288. 193. J. He, F. Zhou, G. Chang and E. J. Lavernia: J. Mater. Sci., 2001, 36, 2955–2964. 194. I. Baker, R. G. Quiller, M. Robson and D. Wu: in ‘High temperature ordered intermetallic alloys X’, Proceedings of the Material Research Society, 2003, 753, BB5.48. 195. D. Sudfeld, K. Wojczykowski, W. Hachmann, P. Jutzi, G. Reiss and A. Hutten: J. Appl. Phys., 2003, 93, 7328–7330. 196. A. K. Giri, K. M. Chowdary, K. D. Humfeld and S. A. Majetich: IEEE Trans. Magn., 2000, 36, 3026–3028. 197. K. Giri, K. M. Chowdary and S. A. Majetich: Mater. Phys. Mech., 2000, 1, 1–10. 198. S. N. Na, S. J. Suh, K. H. Shin, Y. S. Lee and S. H. Lim: J. Appl. Phys., 2003, 93, 8501–8503. 199. M. Sorescu and A. Grabias: Intermetallics, 2002, 10, 317–321. 200. P. S. Fodor, G. M. Tsoi and L. E. Wenger: J. Appl. Phys., 2003, 93, 7035–7037. 201. L. Neèl: J. Phys. Radium, 1954, 15, 225–239. 202. Y. Yoshizawa, S. Oguma and K. Yamaguchi: J. Appl. Phys., 1988, 64, 6044–6046. 203. K. Suzuki, A. Makino, N. Kataoka, A. Inoue and T. Masumoto: Mater. Trans. JIM, 1991, 32, 93–102. 204. A. Makino, T. Hatanai, Y. Naitoh, T. Bitoh, A. Inoue and T. Mosumoto: IEEE Trans. Magn., 1997, 33, 3793–3798. 205. A. Makino, K. Suzuki, A. Inoue and T. Masumoto: Mater. Trans. JIM, 1991, 32, 551–556. 206. M. E. McHenry, M. A. Willard, H. Iwanabe, R. A. Sutton, Z. Turgut, A. Hsiao and D. E. Laughlin: Bull. Mater. Sci., 1999, 22, 495–501. 207. M. A. Willard, D. E. Laughlin and M. E. McHenry: J. Appl. Phys., 2000, 87, 7091–7096. 208. M. A. Willard, F. Johnson, J. H. Claassen, R. M. Stroud, M. E. McHenry and V. G. Harris: Mater. Trans., Special issue on ‘Bulk amorphous, nano-crystalline and nano-quasicrystalline alloys IV’, Jpn Inst. Met., 2002, 43, 2000–2005. 209. M. A. Willard, J. H. Claassen, R. M. Stroud and V. G. Harris: J. Appl. Phys., 2002, 91, 8420–8422. 210. M. A. Willard, D. E. Laughlin, M. E. McHenry, D. Thoma, K. Sickafus, J. O. Cross and V. G. Harris: J. Appl. Phys., 1998, 84, 6773–6777. 211. M. A. Willard, M.-Q. Huang, D. E. Laughlin, M. E. McHenry, J. O. Cross, V. G. Harris and C. Franchetti: J. Appl. Phys., 1999, 85, 4421–4423. 212. H. Iwanabe, B. Lu, M. E. McHenry and D. E. Laughlin: J. Appl. Phys., 1999, 85, 4424–4426. 213. D. H. Ping, Y. Q. Wu, K. Hono, M. A. Willard, M. E. McHenry and D. E. Laughlin: Scripta Mater., 2001, 45, 781–786. 214. M. Mu¨ller, H. Grahl, N. Mattern and B. Schnell: J. Magn. Magn. Mater., 2000, 215–216, 437–439. 215. L. Kraus, V. Soyka, P. Duhaj and I. Mot’ko: J. Phys. IV France, 1998, 8, Pr 2–95–Pr2-98 z #2. 216. M. A. Willard, M. Gingras, M. J. Lee, V. G. Harris, D. E. Laughlin and M. E. McHenry: MRS Res. Symp. Proc., 1999, 577, 469–479. 217. T. Kulik, A. Wlazlowska, J. Ferenc and J. Latuch: IEEE Trans. Magn., 2002, 38, 3075–3077. 218. J. S. Bla´zquez, C. F. Conde and A. Conde: Appl. Phys. Lett., 2001, 79, 2898–2900. 219. J. S. Bla´zquez, C. F. Conde and A. Conde: J. Non-Crystalline Solids, 2001, 287, 187–192. 220. J. S. Bla´zquez, S. Lozano-Perez and A. Conde: Philos. Mag. Lett., 2002, 82, 409–417. 221. M. Kopcewicz, A. Grabias, M. A. Willard, D. E. Laughlin and M. E. McHenry: IEEE Trans. Magn., 2001, 37, 2226–2228. 222. T. Kemeńy, D. Kapta´s, L. F. Kiss, J. Balogh, L. Bujdoso´, J. Gubicza, T. Unga´r and I. Vincze: Appl. Phys. Lett., 2000, 76, 2110–2112. 223. J. Olszewski, H. Fukunaga, W. Ciurzynska, J. Zbroszczyk, A. Lukiewska, A. Mlynczyk, K. Perduta and J. Lelatko: J. Magn. Magn. Mater., 2003, 254–255, 516–518. 224. J. S. Bla´zquez, A. Conde and J. M. Greneche: Appl. Phys. Lett., 2002, 81, 1612–1614.

192


2005

VOL

50

NO

3

225. M. E. Hawley, G. W. Brown, D. J. Thoma, M. A. Willard, D. E. Laughlin and M. E. McHenry: Mater. Res. Soc., Symp. Proc., 1999, 577, 531–541; Pittsburgh, PA, MRS. 226. M. E. McHenry, F. Johnson, H. Okumura, T. Ohkubo, V. R. V. Ramanan and D. E. Laughlin: Scripta Mater., 2003, 48, 881–887. 227. A. Go´mez-Polo, J. I. Pe´rez-Landazabal, V. Recarte, J. Campo, P. Marıń, M. Lo´pez, A. Hernando and M. Va´zquez: Phys. Rev., 2002, 66, 012401–1–4. 228. A. Go´mez-Polo, P. Marıń, L. Pascual, A. Hernando and M. Va´zquez: Phys. Rev. B, 2002, 65, 024433. 229. A. Duckham, D. Z. Zhang, D. Liang, V. Luzin, R. C. Cammarata, R. L. Leheny, C. L. Chien and T. P. Weihs: Acta Mater., 2003, 51, 4083–4093. 230. X. M. Cheng, X. K. Zhang, D. Z. Zhang, S. H. Lee, A. Duckham, T. P. Weihs, R. C. Cammarata, J. Q. Xiao and C. L. Chien: J. Appl. Phys. 2003, 93, 7121–7123. 231. E. J. Yun, W. Win and R. M. Walser: IEEE Trans. Magn., 1996, 32, 4535–4537. 232. C. L. Platt, A. E. Berkowitz, D. J. Smith and M. R. McCartney: J. Appl. Phys., 2000, 88, 2058–2062. 233. S. X. Wang, N. X. Sun, M. Yamaguchi and S. Yabukami: Nature, 2000, 407, 150–151. 234. H. S. Jung and W. D. Doyle: IEEE Trans. Magn., 2002, 38, 2015– 2017. 235. H. S. Jung, W. D. Doyle, J. E. Wittg, J. F. Al-Sharab and J. Bentley: J. Appl. Phys. Lett., 2002, 81, 2415–2417. 236. H. Katada, T. Shimatsu, I. Watanabe, H. Muraoka and Y. Nakamura: IEEE Trans. Magn., 2002, 38, 2225–2227. 237. S.-H. Kong, T. Okamoto and S. Nakagawa: J. Appl. Phys., 2003, 93, 6778–6780. 238. K. Shintaku, K. Yamakawa and K. Ouchi: J. Appl. Phys., 2003, 93, 6474–6476. 239. J. Rantschler, Y. Ding, S. C. Byeon and C. Alexander: J. Appl. Phys., 2003, 93, 6671–6673. 240. C. L. Platt, M. K. Minor and T. J. Klemmer: IEEE Trans. Magn., 2001, 37, 2302–2304. 241. M. K. Minor, T. M. Crawford, T. J. Klemmer, Y. Peng and D. E. Laughlin: J. Appl. Phys., 2002, 91, 8453–8455. 242. P. Zou, W. Yu and J. A. Bain: J. Appl. Phys., 2002, 91, 7830–7832. 243. P. Zou, W. Yu and J. A. Bain: IEEE Trans. Mag., 2002, 38, 3501– 3520. 244. J. Yu, C. Chang, D. Karns, G. Ju, Y. Kubota, W. Eppler, C. Brucker and D. Weller: J. Appl. Phys., 2002, 91, 8357–8359. 245. M. K. Minor and T. J. Klemmer: J. Appl. Phys., 2003, 93, 6465– 6467. 246. J. M. Shin, Y. M. Kim, J. Kim, S. H. Han and H. J. Kim: J. Appl. Phys., 2003, 93, 6677–6679. 247. H. S. Jung, W. D. Doyle and S. Matsunuma: J. Appl. Phys., 2003, 93, 6462–6464. 248. V. A. Vas’ko, V. R. Inturi, S. C. Riemer, A. Morrone, D. Schouweiler, R. D. Knox and M. T. Kief: J. Appl. Phys., 2002, 91, 6818–6820. 249. M.-Q. Huang, Y.-N. Hsu, M. E. McHenry and D. E. Laughlin: IEEE Trans. Magn., 2001, 37, 2239–2241. 250. E. W. Soo, T. J. Zhou and J. P. Wang: J. Appl. Phys., 2002, 91, 8019–8021. 251. S. Kumar, T. Ohkubo and D. E. Laughlin: J. Appl. Phys., 2002, 91, 8360–8362. 252. H. Okumura, D. J. Twisselmann, R. D. McMichael, M. Q. Huang, Y. N. Hsu, D. E. Laughlin and M. E. McHenry: J. Appl. Phys., 2003, 92, 6528–6530. 253. S. Kumar and D. E. Laughlin: J. Appl. Phys., 2003, 93, 8158– 8160. 254. K. Bal, A. Keen, Th. Rasing and H. A. M. van den Berg: J. Appl. Phys., 2002, 91, 8584–8586. 255. J. O. Ventura, J. B. Sousa, M. A. Salgueiro da Silva, P. P. Freitas and V. Vloso: J. Appl. Phys., 2003, 93, 7690–7692. 256. I. Shao, P. M. Vereecken, C. L. Chien, R. C. Cammarata and P. C. Searson: J. Electrochem. Soc., 2003, 150, C184–C188. 257. R. H. Yu, L. Ren, S. Basu, K. M. Unruh, A. Parvizi-Majidi and J. Q. Xiao: J. Appl. Phys., 2000, 87, 5840–5842. 258. G. H. Bozzolo, R. D. Noebe and C. Amador: Intermetallics, 2002, 10, 149–159.