Exchange bias in Co Õ Fe Õ Fe
xMn
1ÀxÕ Cu „ 100 … ultrathin films
C. C. Kuo, W. Pan, Y. C. Chen, and Minn-Tsong Lina) Department of Physics, National Taiwan University, Taipei 106, Taiwan 共Presented on November 15, 2002兲
Stable and well-grown face-centered-cubic Fe films were prepared on buffer layers with varying lattice constants by depositing FexMn1⫺x alloy film on Cu共100兲 single crystal. No ferromagnetic ordering was observed at the stage of 30 ML Fe on the FexMn1⫺x/Cu(100) systems in the temperature range from 100 to 350 K. Furthermore, capping of Co on Fe/FexMn1⫺x/Cu(100) was employed as the probe of antiferromagnetic ordering by study of exchange bias coupling in these films. The exchange bias of the hysteresis loops can be observed after field cooling of the films.
Further analyses by varying the measurement temperature and Fe coverage of the films were also carried out to clarify the origin of the exchange bias coupling observed. The exchange bias field found here is attributed to the interlayer coupling between the Co and Fe–Mn films through the spacing layer Fe. © 2003 American Institute of Physics. 关DOI: 10.1063/1.1540136兴
Much attention has been directed to ultrathin fcc iron films, not only because of the artificial structure not found in nature, but also because of the interesting magnetic behav- iors that are extremely sensitive to crystalline structures.1In theoretical aspects, the antiferromagnetic 共AF兲 phase of fcc Fe on the Cu共100兲 substrates was proposed by ab initio cal- culation with both the local spin-density approximation and generalized gradient approximation.2 In experimental as- pects, the fcc-like phase of iron can be epitaxially grown on the fcc Cu共100兲 substrate.3,4 The AF phase of fcc iron with ferromagnetic 共FM兲 top layers was observed for room- temperature-grown films, while the coverage of iron ranged from 6 to 11 ML.1,5 On the other hand, the exchange bias, another fascinating phenomenon for AF materials can be ob- served for the FM/AF bilayer systems.6,7There is a shift of the magnetic hysteresis loops from the origin of the applied field due to the interlayer coupling of AF and FM layers.
However, no exchange bias was observed for fcc Fe/Cu共100兲 films up to now. This can be attributed to the following rea- sons: First, there are two structures, fcc and fct, mixed in Fe/Cu共100兲 films for coverage from 6 to 11 ML. It could affect exchange bias coupling, which is sensitive to the FM/AF interface. Second, the blocking temperature (TB), at which the exchange bias coupling vanishes, for Fe/Cu共100兲 films could be too low due to the finite size effect in low coverage films.8,9Furthermore, the instability of Fe/Cu共100兲 upon thickness results in low Ne´el temperature (TN), or probably in weak exchange bias coupling.
In order to study the inquiries mentioned above, we tried to deposit more stable and thicker iron films with only one fcc crystalline structure for investigation of the exchange bias induced by fcc iron films. This can be done by deposit- ing fcc FexMn1⫺x films on Cu共100兲 single crystal as a new synthetic substrate. The Cu共100兲 single crystal was prepared by sequential sputtering and annealing in an ultra-high- vacuum chamber.10 The coverage and alloy composition of
the FexMn1⫺xalloy films in our experiments can be precisely controlled with the benefit of a codeposition technique.11,12 The interlayer distance of FexMn1⫺x/Cu(100) can be ma- nipulated by varying the alloy composition x.13,14 In this study, we fixed the alloy composition of the FexMn1⫺x alloy at x⬃50%, where the Fe–Mn alloy can be grown on Cu共100兲 substrates in layer-by-layer mode and Fe can be deposited on top with a stable fcc structure, as shown in Fig.
1. Figure 1 complies with the medium energy electron dif- fraction共MEED兲 关Fig. 1共a兲兴 as well as intensities of the 共00兲 beam of low energy electron diffraction 共LEED兲 关Fig. 1共b兲兴 for Fe/FeMn/Cu共100兲 and Fe/Cu共100兲 films. It is obvious that the fcc structure, which corresponds to the plateau of MEED intensity, in Fe/FeMn/Cu共100兲 is more stable than that in Fe/Cu共100兲. The fcc iron can be grown on an FeMn/
a兲Author to whom correspondence should be addressed; electronic mail:
mtlin@phys.ntu.edu.tw
FIG. 1. 共a兲 MEED intensity oscillations of Fe/FeMn/Cu共100兲 and Fe/
Cu共100兲 films grown at 300 K. 共b兲 LEED 共00兲-beam intensity as a function of beam energy for Fe/FeMn/Cu共100兲 and Fe/Cu共100兲 films. Up and down arrows indicate the characteristics of fct and fcc structures, respectively.
JOURNAL OF APPLIED PHYSICS VOLUME 93, NUMBER 10 15 MAY 2003
8743
0021-8979/2003/93(10)/8743/3/$20.00 © 2003 American Institute of Physics
Cu共100兲 substrate up to 30 ML, while it can be observed on the Cu共100兲 substrate up to only 12 ML, as depicted in Fig.
1共a兲. In addition, the 共00兲 beam of the LEED I(E) spectrum in Fig. 1共b兲 reveals the mixing of fcc and fct structures cor- responding to the magnetic structures of the mainly antifer- romagnetic phase共fcc兲 with the top-most ferromagnetic lay- ers共fct兲.15However, there is only one fcc phase observed in the Fe/FeMn/Cu共100兲 films, which is compatible with our magnetic measurements by magneto-optical Kerr effect 共MOKE兲14 and the results shown in Fig. 2.
In order to clarify the magnetic structure of Fe/FeMn/
Cu共100兲 and exclude the unexpected effect, such as the con- tamination, from our investigation, comparison of the mag- netic properties between the Fe/Cu共100兲 and Fe/FeMn/
Cu共100兲 films was carried out, as shown in the schematic diagram of the Fig. 2. The Fe50Mn50 film was deposited on half of the Cu共100兲 substrate before the growth of the iron film. Thus, the magnetic hysteresis loops of Fe/Cu共100兲 and Fe/FeMn/Cu共100兲 films can be taken quasisimultaneously in this way with variation of the measurement temperature. Fig- ure 2 shows the hysteresis loops for Fe/Cu共100兲 共right half兲 and Fe/FeMn/Cu共100兲 共left half兲 in the polar configuration of MOKE at 180 K. The out-of-plane magnetization can be observed for Fe/Cu共100兲 up to 220 K. However, no ferro- magnetic phase was observed for Fe/FeMn/Cu共100兲 at all temperatures investigated 共100–350 K兲 and magnetic field 共up to 350 Oe兲 in our experiments in both polar and longi- tudinal configurations.
The investigation of exchange bias in the Fe/FeMn/
Cu共100兲 films was carried out by depositing 3 ML cobalt films on top of it at room temperature. There is no significant exchange bias for the as-deposited Co films. After deposi- tion, the films were cooled down to 100 K with a 350 Oe applied field in in-plane orientation共field cooling兲. The ex- change bias as well as the coercive field of the hysteresis loops were enhanced after the field-cooling procedure, as shown in Fig. 3共140 K兲. The magnetic hysteresis loops of 3 ML Co on the Fe/14 ML FeMn/Cu共100兲 film for various temperatures are compiled in Fig. 3. The hysteresis loop at lower temperature reveals both larger coercive field 共⬃50 Oe兲 and larger bias field 共⬃7 Oe兲. The bias field (Hex) de- creases rapidly as the temperature increases, as shown for the loop at 160 K 共⬃2 Oe兲. Hex vanishes for the temperature above 180 K. It means that the blocking temperature, the
characteristic temperature of the exchange bias coupling be- tween the ferromagnetic and antiferromagnetic layers, is around 180 K. This can be easily observed in the bias field versus temperature diagrams, as indicated in Fig. 4. In addi- tion, Auger electron spectroscopy was utilized before and after all of the magnetic property measurements. It confirms that there is no CoO antiferromagnetic layer formed in our systems.
As expected, there exists an exchange bias coupling in the Co/Fe/FeMn/Cu共100兲 films. However, these films are more complicated than a simple FM/AF bilayer system. The exchange bias field found in these films can be attributed to either the interlayer coupling between Co and Fe films or that between Co and Fe–Mn alloy films. In order to distin- guish the interlayer coupling between Co and Fe from that between Co and Fe–Mn, further studies of Co/Fe/FeMn/
Cu共100兲 films were carried out. There is an easy way to
FIG. 2. MOKE measurements of Fe/Cu共100兲 and Fe/FeMn/Cu共100兲 films in the polar configuration at 180 K. The films were deposited on one Cu共100兲 single crystal, as indicated by the top view of the schematic geometries.
FIG. 3. Magnetic hysteresis loops of the 3 ML Co/8 ML Fe/14 ML FeMn/
Cu共100兲 film at various measurement temperatures after the field-cooling procedure.
FIG. 4. Bias field of the hysteresis loops as a function of temperature for Co/Fe/FeMn/Cu共100兲 films with variation of Fe coverage. Inset: bias field vs Fe coverage at 140, 160, and 180 K.
8744 J. Appl. Phys., Vol. 93, No. 10, Parts 2 & 3, 15 May 2003 Kuoet al.
make clear the ambiguity by preparing the Co/Fe/FeMn/
Cu共100兲 films with variation of Fe coverage and measure- ment temperature. The coverage of the Fe film adopted in our study increases up to 18 ML for avoiding the compli- cated mixture of the fcc and bcc structures near the phase transition of fcc to bcc. Figure 4 represents the results of our attempt on solving this problem. It complies with the bias field as functions of Fe coverage 共inset兲 and measurement temperature. The exchange bias field for all of the films in our experiments decrease rapidly and vanish at almost the same temperature around 180 K. In addition, the bias field at the same temperature decreases with a small oscillation as the coverage of Fe increases. The bias field should become larger as the Fe coverage increases if it is attributed to the interlayer coupling between Co and Fe films. On the con- trary, if the bias field is attributed to the interlayer coupling between Co and Fe–Mn alloy films, one can expect a de- crease of bias field while increasing the Fe coverage because the iron film plays the role of a spacing layer and makes no contribution to the exchange bias coupling. From the bias field versus Fe coverage diagram, one can conclude that the exchange bias found here should be attribute to the interlayer coupling between Co and Fe–Mn through the Fe spacing layer. The small oscillation observed here is due to the long- range interlayer exchange coupling overcoming the antifer- romagnetic coupling and dominating the exchange bias cou- pling of the film, which is similar to our previous experiments on the NiO/Cu/NiFe system.7Furthermore, this conclusion also can be verified by the behavior of the bias field as a function of temperature in Fig. 4. It indicates that the exchange bias of the hysteresis loops observed in our study is not contributed by Fe, but contributed by the Fe–Mn alloy films since all the bias field versus temperature curves for different Fe coverages vanish at the same temperature.
The reason responsible for the results that the Fe film in Co/Fe/FeMn/Cu共100兲 makes no contribution to the exchange bias coupling could be twofold. First, the magnetic structures of the fcc Fe/Cu共100兲 films were found to be out of plane.5 However, Co on the Fe/FeMn/Cu共100兲 reveals in-plane mag- netization. The exchange coupling may only provide very weak unidirectional anisotropy in this geometry. Further- more, the detailed spin structure in the fcc AF Fe layers was
recently reported to be a complicated spin-density-wave configuration,16 which may result in the absence of the sig- nificant exchange bias field, in contrast to the well-ordered AF spin configuration as usual.
In conclusion, exchange bias coupling was found for Co/
Fe/FeMn/Cu共100兲 films by varying Fe coverage and mea- surement temperature. The bias field for the hysteresis loops of Co decreases as the coverage of Fe increases. The block- ing temperature for films with different Fe coverages in our study remains the same, indicating that the bias field of the hysteresis loops is attributed to the interlayer exchange bias coupling between Co and Fe–Mn alloy films. The fcc Fe film in the Co/Fe/FeMn/Cu共100兲 system plays the role of a spac- ing layer to modify the interlayer coupling between Co and Fe–Mn films.
This research was supported by a grant from the Na- tional Science Council at Taiwan through Contract No. NSC- 91-2112-M-002-058 and the MOE Program for Promoting Academic Excellence of Universities.
1S. Mu¨ller, P. Bayer, C. Reischl, K. Heinz, B. Feldmann, H. Zillgen, and M.
Wuttig, Phys. Rev. Lett. 74, 765共1995兲.
2M. Fria´k, M. Sˇob, and V. Vitek, Phys. Rev. B 63, 052405共2001兲.
3W. A. A. Macedo and W. Keune, Phys. Rev. Lett. 61, 475共1988兲.
4J. Thomassen, F. May, B. Feldmann, M. Wuttig, and H. Ibach, Phys. Rev.
Lett. 69, 3831共1992兲.
5D. Li, M. Freitag, J. Pearson, Z. Q. Qiu, and S. D. Bader, Phys. Rev. Lett.
72, 3112共1994兲.
6J. Nogue´s and I. K. Schuller, J. Magn. Magn. Mater. 192, 203共1999兲.
7M.-T. Lin, C. H. Ho, C.-R. Chang, and Y. D. Yao, Phys. Rev. B 63, 100404共R兲 共2001兲.
8C. M. Schneider, P. Bressler, P. Schuster, and J. Kirschner, Phys. Rev. Lett.
64, 1059共1990兲.
9T. Ambrose and C. L. Chien, Phys. Rev. Lett. 76, 1743共1996兲.
10C. C. Kuo, W. C. Lin, C. L. Chiu, H. L. Huang, and M.-T. Lin, J. Appl.
Phys. 89, 7153共2001兲.
11M.-T. Lin, W. C. Lin, C. C. Kuo, and C. L. Chiu, Phys. Rev. B 62, 14268 共2000兲.
12W. C. Lin, C. C. Kuo, C. L. Chiu, and M.-T. Lin, Surf. Sci. 478, 9共2001兲.
13F. Offi, W. Kuch, and J. Kirschner, Phys. Rev. B 66, 064419共2002兲.
14W. Pan, C. C. Kuo, Y. C. Chen, and M.-T. Lin共unpublished兲.
15R. D. Ellerbrock, A. Fuest, A. Schatz, W. Keune, and R. A. Brand, Phys.
Rev. Lett. 74, 3053共1995兲.
16D. Qian, X. F. Jin, J. Barthel, M. Klaua, and J. Kirschner, Phys. Rev. Lett.
87, 227204共2001兲.
8745
J. Appl. Phys., Vol. 93, No. 10, Parts 2 & 3, 15 May 2003 Kuoet al.
