phys. stat. sol. (c) 4, No. 12, 4384–4387 (2007) / DOI 10.1002/pssc.200777325

© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Giant parallel and perpendicular exchange biases

in MnPd/Co bilayers

Nguyen Thanh Nam*1

, Nguyen Phu Thuy2,3

, Nguyen Anh Tuan2, Nguyen Nguyen Phuoc

1,

and Takao Suzuki1

1 Information Storage Materials Laboratory, Toyota Technological Institute, Nagoya, Japan 2 International Training Institute for Materials Science, Hanoi University of Technology, Vietnam 3 Vietnam College of Technology, Vietnam National University, Hanoi, Vietnam

Received 7 May 2007, revised 5 September 2007, accepted 19 October 2007

Published online 18 December 2007

PACS 75.25.+z, 75.30.Gw, 75.70.Cn

A systematic study of exchange bias in MnPd/Co bilayers has been carried out in both parallel and per-

pendicular directions, where the dependences of blocking temperature, exchange bias and unidirectional

anisotropy constant on the thicknesses of MnPd and Co layers were investigated. One of the particular in-

terests is that the blocking temperature of parallel exchange bias is higher than that of perpendicular ex-

change bias which can be interpreted as the difference of the ordering in parallel and perpendicular direc-

tion. The other is that a huge unidirectional anisotropy constant, JK = 5.5 erg/cm2 was observed, which is

in reasonable agreement with the theoretical prediction based on the model by Meiklejohn and Bean.

© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

1 Introduction

Exchange bias, discovered by Meiklejonh and Bean in 1957 [1], is the phenomenon associated with the

exchange anisotropy created at the interface between antiferromagnet (AF) and ferromagnetic (FM)

layers when these layers are cooled in a magnetic field through the Néel temperature of the AF layer.

Although exchange bias was discovered more than fifty years old, it still has attracted much interest. One

of the reasons is that it can be used in magnetic sensors and high-density magnetic recording system. The

other is that until now the mechanism of this phenomenon is not clear.

It is known that the theoretically predicted exchange bias field is larger than the experimental value by

two orders of magnitude. There are several theoretical works, such as the domain wall models [3, 4] or

spin flopping model [5] proposed to account for this discrepancy. However, there is still a lack of ex-

perimental confirmation for these models. Recently several groups have found a great enhancement of

the exchange bias field in MnIr/CoFe [6, 7] and in MnPd/Co [8, 9] bilayer system, but its physical origin

is still in controversy. The fact that this large exchange bias shows contradicted results to that of the

“normal” exchange bias thus requires a modification of the theoretical works for the quantitative under-

standing of exchange bias coupling. Therefore, from the fundamental viewpoint, studies of materials

with giant unidirectional anisotropy play a vital role on the way to get a better understanding of the

mechanism of exchange bias phenomenon. From the application point of view, the quest for materials

exhibiting giant unidirectional anisotropy is indispensable for the realization of very thin read head sen-

sors used for ultra-high density magnetic recording.

* Corresponding author: e-mail: sd06508@toyota-ti.ac.jp

phys. stat. sol. (c) 4, No. 12 (2007) 4385

www.pss-c.com © 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

The present work reports about the observation of a very large unidirectional anisotropy constant in both

parallel and perpendicular directions of MnPd/Co bilayers. The dependences of blocking temperature,

exchange bias and unidirectional anisotropy constant on thicknesses are investigated and discussed.

2 Experimental

Samples with the structure of Co/MnPd/Si(111) were grown at room temperature using the RF sputter-

deposition system with the base pressure of 10–6 mbar. MnPd layers were fabricated from Pd target with

bonded Mn chips. No external field was applied during deposition. The argon pressure during deposition

process is 10-3 mbar. The composition of MnPd layer identified by the energy dispersion X-ray spectro-

scope is Mn30Pd70. The structural properties of the samples were investigated by X-ray diffraction. For

field cooling procedure, all the samples were heated in a vacuum oven (10-5 mbar) to temperature of 570

K and cooled down to 80 K in the field applied in both the parallel and perpendicular directions. The

magnetic properties of samples were characterized by the vibrating sample magnetometer (VSM) at

various temperatures from 80 K to 300 K with the field applied in both the parallel and perpendicular

directions

3 Results and discussion

Figure 1 shows the X-ray diffraction patterns of

MnPd(540 nm) single layer and MnPd(36 nm)/

Co(80 nm) bilayer. It is observed that in the sample

with very thick MnPd single layer, MnPd is fcc

phase of with several peaks. It indicates that the

sample is polycrystalline without texture. However,

for the bilayer of MnPd(36 nm)/Co(80 nm), only the

peak of MnPd(321) appears. This may suggest that

for the bilayer MnPd has the weak texture in (321)

orientation. Another point worthwhile noting is that

there is no peak for Co layers. It may therefore be

concluded that Co is in amorphous state.

Shown in Fig. 2 is the temperature dependence of

exchange bias field HE for some of representative

samples of series MnPd(x nm)/Co(20 nm), with x =

7, 14, 18 and 22 nm. It is observed that HE in paral-

lel direction are always larger than those in perpen-

dicular direction. Moreover the blocking tempera-

tures, defined as the temperature beyond which

exchange bias vanishes, in parallel direction are also

larger than those in perpendicular direction. This

effect has been observed by Phuoc and Suzuki in the

FePt/FeMn multilayer [10, 11] and can be explained

by assuming that the large shape anisotropy arising

from demagnetization field at interface of AF and

FM layers forces the spins to lie in the plane of the

films. Therefore the ordering of the antiferromag-

netic spins at the interface in the parallel direction is

larger than that in the perpendicular direction and

consequently the parallel TB and HE are higher than the perpendicular TB and HE. The dependences of

blocking temperature TB, exchange bias HE and the unidirectional anisotropy constant JK on MnPd thick-

ness (tMnPd) are shown in Fig. 3. Here, JK is the unidirectional anisotropy which is defined from the equa-

Fig. 1 (a) XRD pattern of Si(111)/MnPd (540

nm) single layer. (b) XRD pattern of Si(111)/

MnPd(36 nm)/Co(80 nm) bilayer.

20 40 60 80 1000

50

100

150

200

(b)

(a)

(321)

(220)

(200)

Si(111)

(111)

Inte

nsi

ty (

a.u.)

2θ (deg.)

Fig. 2 Temperature dependence of parallel and

perpendicular exchange bias fields for MnPd(x

nm)/Co(20 nm) bilayer (x = 7, 14, 18 and 22 nm.)

80 120 160 200 240 280

0

200

400

600

800

tMnPd

= 14 nm

HE (

Oe)

T (K)

Parallel

Perpendicular

80 120 160 200 240

0

300

600

900

1200

tMnPd

= 7 nm

Parallel

Perpendicular

HE (

Oe)

T (K)

80 120 160 200 240 280

0

200

400

600

800

1000

tMnPd

= 18 nm

HE (

Oe)

T (K)

Parallel

Perpendicular

80 120 160 200 240 280

0

200

400

600

800

tMnPd

= 22 nm

Parallel

Perpendicular

HE (

Oe)

T (K)

4386 N. T. Nam et al.: Exchange biases in MnPd/Co bilayers

© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.pss-c.com

tion JK = HE*MS*tFM, where HE is the exchange bias field, MS the saturation magnetization of the FM

layer and tFM the thickness of FM layer. It can be seen that TB, HE and JK are independent on tMnPd in both

directions. Normally, in exchange bias systems, the dependence of HE and JK on antiferromagnetic

thickness tAF is divided into two regions. When the antiferromagnetic thickness is larger than a critical

thickness, HE and JK are independent on tAF. When tAF is smaller than the critical thickness, as tAF is re-

duced, HE and JK decrease abruptly. But in this case there is only one region, HE and JK are independent

on JK. Therefore the critical thickness in this MnPd/Co system is smaller than 4 nm.

The dependences of blocking temperature TB, exchange bias field HE and the unidirectional anisotropy

constant JK on the thickness of Co layer are shown in Fig. 4. With tCo is larger than 40 nm both parallel

and perpendicular HE decrease, indicating that this is an interfacial effect. This is confirmed by the fact

that JK with the value of about 5.5 erg/cm2 and 3.5 erg/cm2 for parallel and perpendicular respectively, is

nearly constant over this Co thickness range. But when Co thickness is less than 40 nm, JK and HE de-

crease as the thickness of Co decreases. The trend is observed in several systems which were attributed

to the less well crystalline of the FM layer when its thickness is small [12]. However, in this study, this

argument may not be correct since the XRD pattern show that Co layer is amorphous.

It is of great interest to see that the maximum parallel and perpendicular JK is 5.5 erg/cm2 and 4.2

erg/cm2. These JK are several orders of magnitude larger than those observed on most of the metallic

exchanged bias thin films [2]. For example, in MnPd/Fe bilayers, the JK value is only 0.032 erg/cm2 [13]

and 0.017 erg/cm2 [14], in MnIr/Co bilayers, the JK value is 0.14 erg/cm2 [15] and MnFe/Co bilayers, the

JK value is 0.059 erg/cm2 [15]. In the literature, there are few systems, which exhibit such large unidirec-

tional anisotropy constant. The largest JK ever reported in the literature is 2.11 erg/cm2 in the system of

Fe3O4/CoO bilayers measured at T = 5 K [16]. Therefore, it may be concluded that the present obtained

JK of 5.5 erg/cm2 is the largest value ever found in the literature.

Fig. 3 Thickness (MnPd) dependences of

blocking temperature TB, exchange bias field

HE and unidirectional anisotropy constant JK

for MnPd(x nm)/Co(20 nm).

0

200

400

600

800

1000

1200 P a ra lle l

P e rp en d icu la r

HE (

Oe)

0 10 20 30 40 50 600 .0

0 .5

1 .0

1 .5

2 .0

2 .5

3 .0

3 .5 P ara lle l

P erp en d icu la r

J K(erg/cm

2

)

M nP d th ickness (nm )

0

50

100

150

200

250

300

P ara lle l

P erpend icu lar

TB (

K)

Fig. 4 Thickness (Co) dependences of block-

ing temperature TB, exchange bias field HE and

unidirectional anisotropy constant JK for

MnPd(36 nm)/Co(x nm).

0

200

400

600

800

1000

1200

HE (

Oe)

P a ra lle l

P e rp en d icu la r

0 10 20 30 40 50 60 70 80 900

1

2

3

4

5

6

7

P aralle l

P erpend icu lar

C o th ickness (nm )

J K(erg/cm

2

)

0

50

100

150

200

250

300

TB (

K)

P ara lle l

P erp en d icu lar

phys. stat. sol. (c) 4, No. 12 (2007) 4387

www.pss-c.com © 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

It is interesting to find the physical origin of giant exchange bias energy. Recently, Tsunoda et al. [6, 7]

reported a very large value of the unidirectional anisotropy constant JK of MnIr/CoFe bilayers up to 1.3

erg/cm2 and found a strong correlation between exchange bias anisotropy and chemical ordering. In the

present study, it is difficult to find such a correlation since the giant exchange bias is strongly dependent

on the Co thickness. Based on the XRD patterns, it can only be concluded that Co is in amorphous state

and consequently, one cannot find any structural change when changing the thicknesses of Co layers.

In this paper, the simple model put forward by Meiklejohn and Bean [1] is employed to compare the

theoretical value with the obtained experimental result. According to this model, the AF spins at the

interface are uncompensated and fixed during the FM magnetization rotation. Hence, the unidirectional

anisotropy constant can be calculated as follows: JK = Jex SiSj/a2. Here, Jex is the exchange integral at the

interface, which is believed to be in the range of the exchange integrals of the AF and FM materials, Si

and Sj the spins of the interfacial atoms, and a the lattice parameter. If using Jex from the exchange inte-

gral of bulk materials, then JK is in the range from 1.8 erg/cm2 to 14.1 erg/cm2. Therefore, roughly speak-

ing, one may conclude that the present result lies in the range of theoretical prediction by the model of

Meiklejohn and Bean [1]. Even though the present result is consistent with the theoretical prediction, it

may not be considered as a firm support for the simple model by Meikljohn and Bean [1]. A more com-

plicated model which is able to cover for all the cases should therefore be developed.

4 Conclusion

In summary, the present study shows an interesting result in MnPd/Co bilayers that the parallel blocking

temperature is higher than the perpendicular one, which can be interpreted as the difference of the order-

ings in parallel and perpendicular directions. The present study also reported on the largest unidirectional

anisotropy constant ever found in the literature up to 5.5 erg/cm2 in MnPd/Co bilayers. This huge unidi-

rectional anisotropy constant is found to be in reasonable agreement with the simple model proposed by

Meiklejohn and Bean, which predicts that JK in MnPd/Co system is in the range from 1.8 erg/cm2 to 14

erg/cm2. Although the present system of MnPd/Co bilayer has a low blocking temperature, suggesting

that this system cannot be applied for spin-valve sensors, the finding of giant exchange bias may provide

some useful information for better understanding of the mechanism of exchange bias.

Acknowledgements This work is partially supported by the Vietnamese Fundamental Research Grant #4.049.06

(2006-2008). Also, the supports from the Academic Frontier Center for Future Storage Materials Research [MEXT

HAITEKU (2004-2008)] and from the Japanese Storage Research Consortium are gratefully acknowledged.

References

[1] W. H. Meiklejohn and C. P. Bean, Phys. Rev. 102, 1413 (1956).

[2] J. Nogués and I. K. Schuller, J. Magn. Magn. Mater. 192, 203 (1999).

[3] D. Mauri, H. C. Siegmann, P. S. Bagus, and E. Kay, J. Appl. Phys. 62, 3047 (1987).

[4] A. P. Malozemoff, Phys. Rev. B 37, 7673 (1988).

[5] N. C. Koon, Phys. Rev. Lett. 78, 4865 (1997).

[6] I. Imakita, M. Tsunoda, and M. Takahashi, Appl. Phys. Lett. 85, 3812 (2004).

[7] M. Tsunoda et al.., Appl. Phys. Lett. 89, 172501 (2006).

[8] N. P. Thuy et al, J. Magn. Magn. Mater. 192, 203 (2006).

[9] N. T. Nam et al., J. Magn. Magn. Mater. (accepted 2007).

[10] N. N. Phuoc and T. Suzuki, J. Appl. Phys. 99, 08C107 (2006).

[11] N. N. Phuoc and T. Suzuki, J. Appl. Phys. 101, 009E51 (2007).

[12] N. N. Phuoc and T. Suzuki, IEEE Trans. Magn. 42, 2996 (2006).

[13] R. F. C. Farrow et al., J. Appl. Phys. 81, 4986 (1997).

[14] Y. J. Tang et al., Phys. Rev. B 62, 8654 (2000).

[15] K. A. Seu et al., J. Appl. Phys. 93, 6611 (2003).

[16] P.J. van der Zaag et al., J. Appl. Phys. 79, 5103 (1996).