eScholarship provides open access, scholarly publishing
services to the University of California and delivers a dynamic
research platform to scholars worldwide.
Lawrence Berkeley National Laboratory
Lawrence Berkeley National Laboratory
A direct measurement of rotatable and frozen CoO spins in exchange bias system of CoO/Fe/
Lawrence Berkeley National Laboratory
LBNL Paper LBNL-3533E
Physical Review Letters
A direct measurement of rotatable and frozen CoO spins in exchange bias
system of CoO/Fe/Ag(001)
J. Wu1, J. S. Park,1 W. Kim1, 2, E. Arenholz,3 M. Liberati,3 A. Scholl,3 Y. Z. Wu,4 Chanyong
Hwang2, and Z. Q. Qiu1
1 Department of Physics, University of California at Berkeley, Berkeley,
California 94720, USA
2 Korea Research Institute of Standards and Science,Yuseong, Daejeon 305-340, Koera
3 Advanced Light Source, Lawrence Berkeley National Laboratory, Berkeley, California
4 Dept. of Physics, Fudan University, Shanghai, P. R. China
The exchange bias of epitaxially grown CoO/Fe/Ag(001) was investigated using X-
ray Magnetic Circular Dichroism (XMCD) and X-ray Magnetic Linear Dichroism (XMLD)
techniques. A direct XMLD measurement on the CoO layer during the Fe magnetization
reversal shows that the CoO compensated spins are rotatable at thinner thickness and frozen
at larger thickness. By a quantitative determination of the rotatable and frozen CoO spins as
a function of the CoO film thickness, we find the remarkable result that the exchange bias
is well established before frozen spins are detectable in the CoO film. We further show that
the rotatable and frozen CoO spins are uniformly distributed in the CoO film.
PACS numbers: 75.70.Ak
As a ferromagnetic(FM)/antiferromagnetic(AFM) system is cooled down within a
magnetic field to below the Néel temperature (TN) of the AFM material, the shift of the FM
hysteresis loop in the magnetic field is referred to as exchange bias . Investigation of
exchange bias has been one of the most active research areas because of its importance to
spintronics technology . While it is well accepted that the AFM order is responsible for
the exchange bias [3,4], it remains a mystery on how the AFM spins behave during the FM
magnetization reversal. Consequently, different AFM spin structures have been proposed
to explain the exchange bias [5,6,7,8]. Most measurements are based on the FM layer
hysteresis loops which explore only
[9,10,11,12,13,14,15,16,17]. Recently, the development of X-ray Magnetic Circular
Dichroism (XMCD) and X-ray Magnetic Linear Dichroism (XMLD)  allows an
element-specific study of the FM/AFM systems. The result shows clearly a correlation
between the FM and the AFM domains, and the existence of a small amount of
uncompensated spins in the AFM layer . It was further shown that only a small
percentage of the uncompensated spins is pinned to account for the exchange bias
[20,21,22,23] and that these pinned uncompensated AFM spins actually extend into the
AFM layer , suggesting a bulk-like effect of the AFM spins in the exchange bias
[25,26,27,28,29]. Despite the above progress, the compensated AFM spin behavior
remains unclear during the FM layer reversal. It is usually assumed that the AFM
compensated spins should be frozen to generate an exchange bias. However, one direct
measurement on Co/bulk NiO(001) shows that the NiO spins at the Co/NiO interface may
exhibit a spring-like winding structure during the Co magnetization alignment . This
result raises a critical issue, i.e. whether it is necessary to freeze the majority of the AFM
compensated spins to generate an exchange bias in a FM/AFM thin film system. In this
Letter, we report an experimental study of CoO/Fe/Ag(001) single crystalline thin films.
We find the remarkable result that as the CoO thickness increases, the exchange bias is well
established before frozen spins are detectable in CoO film.
A Ag(001) substrate was prepared in an ultra-high vacuum system by cycles of Ar
ion sputtering at ~2keV and annealing at 600oC. A 15 monolayer (ML) Fe film was
grown on top of the Ag(001) substrate. Then a CoO wedge (0-8 nm) was grown on top of
the Fe film by a reactive deposition of Co under an oxygen pressure of 1×10-6 Torr. Low
Energy Electron Diffraction (LEED) shows well-defined diffraction spots, indicating the
formation of single crystalline CoO film . The sample is covered by a 2nm Ag
indirectly the AFM spin behavior
protection layer and then measured at beamlines 4.0.2 and 11.0.1 of the Advanced Light
Source (ALS). Fe film on Ag(001) has a bcc structure with the Fe  axis parallel to
the Ag  axis and CoO film on Fe(001) has an fcc structure with the CoO  axis
parallel to the Fe  axis.
XMLD effect is clearly seen by measuring the X-ray absorption spectrum (XAS) at
the CoO L3 edge (Fig. 1) at the normal incidence. The L3 ratio (RL3), defined as the ratio
of the XAS intensity at 778.1 eV and at 778.9 eV, is used to quantify the XMLD effect .
The sample of CoO(6 nm)/Fe(15 ML)/Ag(001) was cooled to 90 K and measured by
PEEM-3 with the incident x-ray at 60o incident angle and circularly polarized for Fe
domain imaging, and linearly polarized for CoO domain imaging. The CoO domains
follow exactly the Fe domains (Fig. 1b). Noting that Fe  axis is parallel to CoO
 axis, then based on the L3 line shape and ratio analysis [32,33,34,35], we conclude
that the in-plane CoO AFM spins have a 90o-coupling to the Fe spins in the CoO(6
After the sample was cooled down to 90 K within a 4 kOe magnetic field along the
Fe  crystal axis, Fe and CoO hysteresis loops are measured at 90 K with the applied
field in the field cooling direction. A small transverse in-plane field was applied during
the hysteresis loop measurement to ensure a rotational Fe magnetization reversal. In the
XMLD measurement of the CoO L3 edge, the x-ray polarization direction is also parallel to
the field cooling direction. Since the exchange bias and the amount of rotatable/frozen
spins could depend on the field frequency , we performed all our measurements in the
DC regime. The most interesting observation is the appearance of the CoO XMLD
hysteresis loop (Fig. 2) in dCoO=2.0 nm sample, showing clearly that the CoO compensated
spins rotate during the Fe magnetization reversal. In contrast, the absence of the CoO
hysteresis loop in dCoO=6.0 nm sample shows that the CoO spins are totally frozen during
the Fe magnetization reversal.
Fig. 3 shows the Fe film coercivity (HC) and exchange bias (HE) as a function of the
CoO thickness. While the HC starts to increase at dCoO~0.2 nm, the HE develops only
above a critical thickness of dCoO=0.8 nm. The increase of the HC is due to the
establishment of the AFM order of the CoO layer . Then the onset of HE at dCoO=0.8
nm shows that the exchange bias does not develop right after the CoO establishes its AFM
order. We further carried out the following measurement to separate the rotatable and
frozen spins in the CoO layer. We performed XMLD measurement as a function of the
polarization angle () to obtain the -dependence of the L3 ratio RL3, RL3cos2with
the coefficient A proportional to the amount of the AFM compensated spins. Therefore a
RL3- measurement allows the determination of the amount of AFM spins under specific
We first measured the RL3- dependence right after the field cooling. Under this
condition, both rotatable and frozen CoO spins should be aligned to the same direction so
that the RL3 difference at =90o and =0o [e.g., RL3=RL3(90o)-RL3(0o)] is proportional to
the total CoO spins (top row of Fig. 4). We then rotate the in-plane magnetic field by 90o
to rotate the Fe magnetization by 90o in the film plane. Under this condition, the rotatable
CoO spins should follow the Fe magnetization to rotate by 90o while the frozen CoO spins
should remain in their original direction. Then RL3 in this case should correspond to the
difference between the frozen spins and the rotatable spins inside the CoO film. The
result (lower row of Fig. 4) indeed shows a thickness dependent RL3. At dCoO=6.0 nm,
the RL3- dependence remains unchanged after the field rotates by 90o, showing that there
is no rotatable spins at this thickness. As the CoO thickness decreases to dCoO=2.5 nm,
RL3 decreases, showing that some CoO spins rotate away from the field cooling direction
in the CoO film. Thinner than dCoO=2.5nm, RL3 reverses its sign, showing that there are
more rotatable spins than frozen spins in the CoO film. The difference of RL3 for field
parallel and perpendicular to the field cooling direction allows us to determine
quantitatively the percentage of the frozen spins in the CoO film (Fig. 3). The CoO film
has no detectable frozen spins below 2.2 nm (with an error bar of ~0.16 nm), becomes
partially frozen for 2.2 nm<dCoO<4.5 nm, and has all spins frozen for dCoO>4.5 nm. We
then find the remarkable result that the exchange bias develops even when no frozen CoO
spins are detectable at dCoO<2.2 nm, reaches ~2/3 of its saturation value at the onset of the
frozen spins at dCoO=2.2 nm, and becomes saturated at dCoO=3 nm where 80% of the CoO
spins are frozen. We estimate an upper limit of no more than ~5% of frozen spins below
dCoO=2.2 nm. Therefore we conclude that ~5% frozen CoO spins should be enough to
generate an exchange bias in CoO/Fe/Ag(001) system. This result may explain why only
a small percentage of pinned uncompensated spins would be enough to account for the
exchange bias [20-24].
The next question is where the rotatable and frozen CoO spins are located? To
answer this question, we inserted a 2 ML NiO probe layer at the CoO/Fe interface and at
the surface of CoO/Fe by growing two samples of CoO(wedge)/NiO(2 ML)/Fe(15
ML)/Ag(001) and NiO(2 ML)/CoO(wedge)/Fe(15 ML)/Ag(001), and measured the Ni
XMLD as a function of the CoO thickness. The result shows that the NiO XMLD follows
exactly the CoO thickness dependence (Fig. 5). This result shows that the rotatable and
frozen spins distribute uniformly inside the entire CoO film, supporting the doping and
FM/AFM/FM results [25-29] that the exchange bias depends on the bulk AFM spin
structure. It should be mentioned that the NiO spins are much softer than CoO spins due to
a much weaker NiO magnetic anisotropy . Consequently, the inserted 2 ML NiO won’t
produce a noticeable shift of CoO onset thickness for the frozen spins.
The last question is the role of FeO layer at the Fe-CoO interface. The interfacial
FeO layer was identified to be only in the monolayer regime [38,39,40] and won’t alter the
exchange bias of Fe/CoO system . In addition, the fact that we have the same FeO
interface at different CoO thicknesses allows us to single out the effect of the AFM
thickness on the exchange bias. We conclude that the presence of the FeO interfacial layer
does not alter our conclusion.
In summary, we investigated epitaxially grown CoO/Fe/Ag(001). Using element-
specific XMLD measurement, we find that the CoO spins are rotatable below 2.2 nm CoO
thickness, become partially frozen between 2.2 nm and 4.5 nm, and totally frozen above 4.5
nm. Contrary to the expectation, the exchange bias of the Fe film develops at dCo>0.8 nm
even when no frozen spins are detectable in CoO film, reaches ~2/3 of its saturation value
at the onset of frozen CoO spins at dCo=2.2 nm, and saturates at dCoO=3 nm where 80% of
the CoO spins are frozen. With the XMLD sensitivity estimation, we conclude that ~5%
of frozen CoO spins is enough to establish the exchange bias in CoO/Fe/Ag(001) system.
We further show that the rotatable/frozen spins distribute uniformly in the CoO film.
This work was supported by National Science Foundation DMR-0803305, U.S.
Department of Energy DE-AC02-05CH11231, KICOS through Global Research
Laboratory project, and Chinese Education Department.
Fig. 1: (color online) (a) X-ray Absorption Spectra (XAS) of Co L3 edge taken at two
orthogonal linear polarizations (=0o and 90o) for CoO(6.0 nm)/Fe(15
ML)/Ag(001). The asymmetry of two spectra represents XMLD signal. (b)
Magnetic domain images of ferromagnetic Fe and antiferromagnetic CoO taken
by XMCD and XMLD, respectively. Arrows indicate the orientation of Fe and
CoO spins. It is clear that the antiferromagnetic CoO spins are 90o coupled to Fe
Fig. 2: Hysteresis loops of ferromagnetic Fe and antiferromagnetic CoO for CoO/Fe(15
ML)/Ag(001) taken by XMCD and XMLD, respectively. Arrows indicate the
ramping direction of magnetic field. The presence and absence of the CoO
response to the magnetic field at dCoO=2.0 nm and at dCoO=6.0 nm show rotatable
and frozen compensated spins in the 2.0 nm and 6.0 nm thick CoO films
Fig. 3: (color online) Fe film coercivity (HC), exchange bias (HE) and the percentage of
CoO frozen spins in CoO/Fe(15ML)/Ag(001) as a function of CoO thickness.
The remarkable fact is that HE develops below 2.2 nm CoO thickness where no
frozen CoO spins are detectable. The solid lines are guides to eyes.
Fig. 4: (color online) Polarization angle dependence of the Co L3 ratio measured with a
0.4 Tesla in-plane magnetic field at different CoO thicknesses in CoO/Fe(15
ML)/Ag(001). Solid lines are fitting results of cos2-dependence. The L3 ratio
differenceRL3=RL3(90o)-RL3(0o) is proportional to the sum and subtraction of the
frozen and rotatable CoO spins for field parallel (top row) and perpendicular
(lower row) to the field cooling direction, respectively.
Fig. 5: (color online) A 2ML NiO layer is inserted on the top or bottom of CoO to detect
the depth-dependent distribution of the frozen CoO spins. The same thickness
dependences of the frozen CoO and NiO spins indicate a uniform distribution of
the frozen spins in the CoO film.
1. W. H. Meiklejohn and C. P. Bean, Phys. Rev. 102, 1413 (1956).
J. Nogues and I. K. Schuller, J. Magn. Magn. Mater. 192, 203 (1999). 2.
3. M. G. Blamire, M. Ali, C.-W. Leung, C. H. Marrows, and B. J. Hickey, Phys. Rev.
Lett. 98, 217202 (2007).
E. Shipton, K. Chan, T. Hauet T, O. Hellwig, and E. E. Fullerton, Appl. Phys. Lett. 95,
N.C. Koon, Phys. Rev. Lett. 78, 4865 (1997).
T. C. Schulthess and W. H. Butler, Phys. Rev. Lett. 81, 4516 (1998).
D. Mauri, H. C. Siegmann, P. S. Bagus, and E. Kay, J. Appl. Phys. 62, 3047 (1987).
A. P. Malozemoff, Phys. Rev. B 35, 3679 (1987).
T. Hauet, J. A. Borchers, Ph. Mangin, Y. Henry, and S. Mangin, Phys. Rev. Lett. 96,
Steven Brems, Kristiaan Temst, and Chris Van Haesendonck, Phys. Rev. Lett. 99,
S. K. Mishra, F. Radu, H. A. Dürr, and W. Eberhardt, Phys. Rev. Lett. 102, 177208
S. Maat, K. Takano, S. S. Parkin, and Eric E. Fullerton, Phys. Rev. Lett. 87, 087202
J. Olamit, Z. P. Li, I. K. Schuller, and K. Liu, Phys. Rev. B 73, 024413 (2006).
14. H. Ouyang, K.-W. Lin, C.-C. Liu, Shen-Chuan Lo, Y.-M. Tzeng, Z.-Y. Guo, and J.
van Lierop, Phys. Rev. Lett. 98, 097204 (2007)
15. X. P. Qiu, D. Z. Yang, S. M. Zhou, R. Chantrell, K. O’Grady, U. Nowak, J. Du, X. J.
Bai, and L. Sun, Phys. Rev. Lett. 101, 147207 (2008).
K. Liu, S.M. Baker, M. Tuominen, T. P. Russell, I. K. Schuller, Phys. Rev. B 63,
17. J. Sort, A. Hoffmann, S.-H. Chung, K. S. Buchanan, M. Grimsditch, M. D. Baró, B.
Dieny, and J. Nogués, Phys. Rev. Lett. 95, 067201 (2005).
18. J. Stöhr, A. Scholl, T. J. Regan, S. Anders, J. Lüning, M. R. Scheinfein, H. A.
Padmore, and R. L. White, Phys. Rev. Lett. 83, 1862 (1999).
19. H. Ohldag, T. J. Regan, J. Stöhr, A. Scholl, F. Nolting, J. Lüning, C. Stamm, S.
Anders, and R. L. White, Phys. Rev. Lett. 87, 247201 (2001).
20. H. Ohldag, A. Scholl, F. Nolting, E. Arenholz, S. Maat, A. T. Young, M. Carey, and J.
Stöhr, Phys. Rev. Lett. 91, 017203 (2003).
21. S. Roy, M. R. Fitzsimmons, S. Park, M. Dorn, O. Petracic, Igor V. Roshchin, Zhi-Pan
Li, X. Batlle, R. Morales, A. Misra, X. Zhang, K. Chesnel, J. B. Kortright, S. K.
Sinha, and Ivan K. Schuller, Phys. Rev. Lett. 95, 047201 (2005).
22. Hendrik Ohldag, Hongtao Shi, Elke Arenholz, Joachim Stöhr, and David Lederman,
Phys. Rev. Lett. 96, 027203 (2006).
23. Florin Radu, Alexei Nefedov, Johannes Grabis, Gregor Nowak, Andre Bergmann,
Hartmut Zabel, J. Magn. Magn. Mater. 300, 206 (2006).
24. Sebastian Brück, Gisela Schütz, Eberhard Goering, Xiaosong Ji, and Kannan M.
Krishnan, Phys. Rev. Lett. 101, 126402 (2008).
25. P. Miltényi, M. Gierlings, J. Keller, B. Beschoten, G. Güntherodt, U. Nowak, and K.
D. Usadel, Phys. Rev. Lett. 84, 4224 (2000).
Jung-Il Hong, Titus Leo, David J. Smith, and Ami E. Berkowitz, Phys. Rev. Lett. 96,
27. Marian Fecioru-Morariu, Syed Rizwan Ali, Cristian Papusoi, Martin Sperlich, and
Gernot Güntherodt, Phys. Rev. Lett. 99, 097206 (2007).
28. U. Nowak, K. D. Usadel, J. Keller, P. Miltényi, B. Beschoten, and G. Güntherodt,
Phys. Rev. B 66, 014430 (2002).
29. R. Morales, Zhi-Pan Li, J. Olamit, Kai Liu, J. M. Alameda, and Ivan K. Schuller,
Phys. Rev. Lett. 102, 097201 (2009).
A. Scholl, M. Liberati, E. Arenholz, H. Ohldag, and J. Stöhr, Phys. Rev. Lett. 92,
31. A. Brambilla, P. Sessi, M. Cantoni, L. Duò, M. Finazzi and F. Ciccacci, Thin Solid
Films 516, 7519 (2008).
32. Gerrit van der Laan, Elke Arenholz, Rajesh V. Chopdekar and Yuri Suzuki, Phys. Rev.
B 77, 064407 (2008).
33. Elke Arenholz, Gerrit van der Laan,, Rajesh V. Chopdekar, and Yuri Suzuki, Phys.
Rev. Lett. 98, 197201 (2007).
34. I. P. Krug, F. U. Hillebrecht, M. W. Haverkort, A. Tanaka, L. H. Tjeng, H. Gomonay,
A. Fraile-Rodríguez, F. Nolting, S. Cramm, and C. M. Schneider, Phys. Rev. B 78,
35. R. Abrudan, J. Miguel, M. Bernien, C. Tieg, M. Piantek, J. Kirschner, and W. Kuch,
Phys. Rev. B 77, 014411 (2008).
36. Jeffrey McCord, Roland Mattheis, and Dieter Elefant, Phys. Rev. B 70, 094420
37. Susumu Soeya, Shigeru Tadokoro, Takao Imagawa, Moriaki Fuyama, and Shinji
Narishige, J. Appl. Phys. 74, 6297 (1993).
38. T.J. Regan, H. Ohldag, C. Stamm, F. Nolting, J. Luning, J. Stohr and R.L. White, Phys.
11 Download full-text
Rev. B 64, 214422 (2001).
39. C. Tusche, H.L. Meyerheim, N. Jedrecy, G. Renaud, A. Ernst, J. Henk, P. Bruno, and J.
Kirschner, Phys. Rev. Lett. 95, 176101 (2005).
40. R. Abrudan, J. Miguel, M. Bernien, C. Tieg, M. Piantek, J. Kirschner, and W. Kuch,
Phys. Rev. B 77, 014411 (2008).
41. J. Gurgul et. al., Surf. Interface Anal. (2010). (online preprint, published online 29