3pdf PDF
3pdf PDF
3pdf PDF
Exchange bias (EB) is known as a magnetic coupling that gives rise to conventional EB [15,18,19]. Also, in these
phenomenon at ferromagnetic (FM)-antiferromagnetic EB systems, proportionality between the moment at the
(AFM) interfaces with strong implications for magnetic interface and the AFM bulk magnetic moment is a faintly
field sensor applications, read heads, and modern ap- motivated assumption. The latter is far more reasonable in
proaches to spintronics [1–6]. It is initialized by field the case of a thin FM pinning layer with a homogeneous
cooling the heterosystem to below the blocking tempera- spin structure along the normal of the film. We also present
ture, TB , at which AFM order is established, at least on a a transparent theory of the TE adapted to all FM bilayers,
mesoscopic scale [7]. The most striking and widely studied which shows excellent agreement with our experimental
feature in EB systems is the shift of the FM hysteresis loop data.
along the magnetic field axis by the amount 0 HEB . Figure 1 (left frame) displays a schematic of our sample
Furthermore, a gradual degradation of the EB field can structure and its overall magnetic hysteresis m vs 0 H
be observed upon cycling the heterostructure through con- (dashed line) where m is the magnetic moment and H is the
secutive hysteresis loops [8–12]. This aging phenomenon applied magnetic field. The measurements are done at
is know as the training effect (TE) and is quantified by
0 HEB vs n, where n is the number of loops cycled after HB
0
0.1
first setting the EB via field cooling. EB and its training
have been observed in various magnetic systems ranging 5
m[ Am]
2
m [n Am ]
2
room temperature by means of alternating gradient force 0.081 25
m [nAm ]
2
Hset=-0.34T
magnetometry. The shape of the overall loop reflects the 0
1.0
0.080 24
m [nAm ]
well separated switching field distributions of the HL and
2
SL. The thick solid lines are low field minor loops after 0.079 23
0.078 0.079 30
Hset=-0.36T
positive and negative saturation, respectively. The low field 0
HB[T]
HB [T]
0.078
response of these minor loops is dominated by the magne- -0.078
0.8
0 HB to distinguish it from the EB field, 0 HEB , in AFM- 0 5 10 15 20 0.00 0.04 0.08 0.12
FM systems. Figure 1 (right frame) shows an example of loop# n HAmp[T]
HB(n=1)- 0HB [ mT ]
ing place during a 10 loop training sequence in the field
interval 0 0 H 0 HAmp can be used as a measure of
e
the change of the bias field HB HB n 1 HB n 2
0
multaneous with the onset of magnetization reversal in the 1
n
X j [2] Ch. Binek and B. Doudin, J. Phys. Condens. Matter 17,
1
Sn K 1n1 S1 KSe K 1n : (4) L39 (2005).
j2 K 1 [3] J. Nogués and I. K. Schuller, J. Magn. Magn. Mater. 192,
203 (1999).
Taking into account that S / mr for our thin FM biasing [4] A. Berkowitz and K. Takano, J. Magn. Magn. Mater. 200,
films [17,20] and the already demonstrated proportionality 552 (1999).
mr / 0 HB one obtains an explicit expression for the cycle [5] R. Stamps, J. Phys. D 33, R247 (2000).
dependent bias field [6] M. Kiwi, J. Magn. Magn. Mater. 234, 584 (2001).
[7] I. Roshchin, O. Petracic, R. Morales, Li Zhi-Pan,
X. Batlle, and I. K. Schuller, Europhys. Lett. 71, 297
0 HB n K 1n1 0 HB 1 (2005).
[8] D. Paccard, C. Schlenker, O. Massenet, R. Montmory, and
K 1n 1 1 A. Yelon, Phys. Status Solidi 16, 301 (1966).
K0 HBe K 2 : (5)
KK 1n1 [9] C. Schlenker, S. S. P. Parkin, J. C. Scott, and K. Howard,
J. Magn. Magn. Mater. 54, 801 (1986).
Figure 2 (left frame, lines) shows the results of least- [10] K. Zhang, T. Zhao, and M. Fujiwara, J. Appl. Phys. 89,
squares fits of Eq. (5) to our experimental data sets. 6910 (2001).
Evidently, there is excellent agreement between the here [11] S. G. te Velthuis, A. Berger, G. P. Felcher, B. Hill, and
derived theoretical expression and our experiments, vali- E. Dahlberg, J. Appl. Phys. 87, 5046 (2000).
dating our theoretical approach. In our analysis, K and [12] A. Hoffmann, Phys. Rev. Lett. 93, 097203 (2004).
the equilibrium bias field 0 HBe 0 HB n ! 1 enter [13] W. H. Meiklejohn and C. P. Bean, Phys. Rev. 105, 904
Eq. (5) as fitting parameters while 0 HB n 1 is fixed (1957).
[14] V. Skumryev, S. Stoyanov, Y. Zhang, G. Hadjipanayis,
as being the bias field of the first loop. Note that the line has
D. Givord, and J. Nogués, Nature (London) 423, 850
physical meaning only at integer values. Furthermore, it is (2003).
remarkable that the parameter K varies only by about 25% [15] A. Hochstrat, Ch. Binek, and W. Kleemann, Phys. Rev. B
around K 0:2 throughout all fits indicating that K is 66, 092409 (2002).
virtually independent from the field initialization. In con- [16] D. Niebieskikwiat and M. B. Salamon, Phys. Rev. B 72,
trast to the here discussed FM-FM systems, the analysis of 174422 (2005).
AFM-FM systems gives rise to a free energy expression [17] A. Berger, D. T. Margulies, and H. Do, Appl. Phys. Lett.
with a leading term of 4th order in S Sn Se , result- 85, 1571 (2004).
ing in a slower power law approach to equilibrium, a fact [18] H. Ohldag, A. Scholl, F. Nolting, E. Arenholz, S. Maat,
that was previously confirmed experimentally [28,29]. A. T. Young, M. Carey, and J. Stöhr, Phys. Rev. Lett. 91,
In summary, we have shown that FM bilayers posses 017203 (2003).
[19] P. Kappenberger, S. Martin, Y. Pellmont, H. J. Hug, J. B.
prototypical properties for the fundamental understanding
Kortright, O. Hellwig, and Eric E. Fullerton, Phys. Rev.
of exchange bias and demonstrated for the first time its Lett. 91, 267202 (2003).
corresponding training phenomenon. From an experimen- [20] A. Berger, Ch. Binek, D. T. Margulies, A. Moser, and E. E.
tal standpoint, these systems are vastly superior to conven- Fullerton, Physica (Amsterdam) 372B, 168 (2006).
tional exchange bias systems due to the FM nature of the [21] M. Toney, J. Borchers, K. O’Donovan, C. Majkrzak,
pinning layer. The latter allows a characterization of its D. Margulies, and Eric E. Fullerton, Appl. Phys. Lett.
magnetization state by means of simple magnetometry, 86, 162506 (2005).
enabling us to unambiguously demonstrate that the devia- [22] L. Thomas, J. Luning, A. Scholl, F. Nolting, S. Anders,
tions from equilibrium in the pinning layer are the driving J. Stohr, and S. Parkin, Phys. Rev. Lett. 84, 3462 (2000).
force behind the exchange bias training effect. We further- [23] U. Nowak, K. D. Usadel, J. Keller, P. Miltényi,
more derive a theoretical description of the training effect B. Beschoten, and G. Güntherodt, Phys. Rev. B 66,
014430 (2002).
for such all ferromagnetic bilayer systems based upon the
[24] D. Suess, M. Kirschner, T. Schrefl, J. Fidler, R. L. Stamps,
discretized dynamical LK equation. The resulting equation and J.-V. Kim, Phys. Rev. B 67, 054419 (2003).
shows excellent quantitative agreement with our experi- [25] M. D. Stiles and R. D. McMichael, Phys. Rev. B 60, 12 950
mental data, corroborating the underlying physical picture (1999).
of the training effect as a triggered relaxation mechanism [26] H. Xi and R. M. White, Phys. Rev. B 64, 184416 (2001).
towards the equilibrium state of the pinning layer. [27] S. Brems, D. Buntinx, K. Temst, C. Van Haesendonck,
We thank NSF-MRSEC for financial support. F. Radu, and H. Zabel, Phys. Rev. Lett. 95, 157202 (2005).
[28] Ch. Binek, Phys. Rev. B 70, 014421 (2004).
[29] Ch. Binek, Xi He, and S. Polisetty, Phys. Rev. B 72,
054408 (2005).
[1] G. Prinz and K. Hathaway, Phys. Today 48 No. 4, 24 [30] G. Vizdrik, S. Ducharme, V. M. Fridkin, and G. Yudin,
(1995). Phys. Rev. B 68, 094113 (2003).
067201-4