Comparative study of the d -band filling effect on the magnetic behavior of Co

(1)

Comparative study of the d -band filling effect on the magnetic behavior of Co

_x

Ni

₁_Àx

and Fe

_x

Ni

₁_Àx

ultrathin films on Cu „ 100 …

C. C. Kuo, S. F. Chuang, W. Pan, W. C. Lin, and Minn-Tsong Lin^a) Department of Physics, National Taiwan University, Taipei 106, Taiwan

The ultrathin Fe_xNi₁_⫺x/Cu(100) and Co_xNi₁_⫺x/Cu(100) alloy films with x⭐10% were prepared for the study of spin-reorientation transition with precise variation of composition and coverage of the films. The spin-reorientation transition with the variation of the film coverage was observed in Co_xNi₁_⫺x/Cu(100) with x⬍10% as well as FexNi₁_⫺x/Cu(100) with x⬍6%. This transition was found to be strongly affected by the alloy composition of the alloy films. The critical thickness for the spin-reorientation transition changes from 7.5 ML with x⫽0 to 17.5 ML with x⫽8% for Co_xNi₁_⫺x/Cu(100). A more sensitive composition-driven effect has been found in Fe_xNi₁_⫺x/Cu(100) by varying the Fe concentration for the critical thickness to 16 ML with x

⫽5%. The more drastic influence of the concentration of Fe than Co on the critical thickness of the spin-reorientation transition can be attributed to the greater deviation of the difference in 3d electron number of majority and minority bands between Fe and Ni than that between Co and Ni. © 2002 American Institute of Physics. 关DOI: 10.1063/1.1456401兴

Numerous attempts have been made to reconfigure the easy axis, or so-called the spin-reorientation transition 共SRT兲, of the magnetic ultrathin films. It has been shown that the SRT can be manipulated not only by the coverage,^1–3but also by the temperature^{4 – 6}of the films. In addition, the composition of alloy ultrathin films is another factor determining the SRT if two composite elements prefer different easy axes. The alloy composition-driven SRT has been successfully demonstrated for the Fe_xCo₁_⫺x/Cu(100)⁷ and Co_xNi₁_⫺x/Cu(100)⁸ systems. In both of these two alloy ultrathin films, the morphology, crystalline structure, and interlayer distance keep invariant in the range of the alloy composition where the SRT is observed. The alloy composition can thus be treated as a variable that modifies the 3d-band filling for the films. As a consequence, the spin-reorientation transition found in the Fe–Co and Co–Ni alloy ultrathin films is attributed to the evolution of the 3d-band filling in these systems.^7,8It provides a venue to investigate the effect of 3d-band filling on the magnetic characteristics such as the magnetic anisotropy and, in turn, the spin-reorientation transition of the films.

In particular, there is good evidence to show a critical evolution of inverse SRT similar to Ni/Cu共100兲 for Co_xNi₁_⫺x/Cu(100) alloy films driven by the subtle variation of alloy composition x.⁸ The critical thickness of the SRT was changed drastically from 7.5 to 17.5 ML for x varying from 0% to 8%. This system can be identified as a Ni/Cu共100兲-like system since the deviation of the crystalline structure between Ni/Cu共100兲 and these alloy films is insignificant. The alloy composition in this system can therefore be treated as a variation of the effective filling of the 3d-band electrons for Ni/Cu共100兲 films since Co is the ele- ment adjoining Ni with the 3d-electron number decreasing by one. In addition, the difference in number of 3d-electrons

between Fe and Ni is twice as large as that between Co and Ni. An interesting question then arises: Will a more drastic variation of critical thickness for SRT be found than Co_xNi₁_⫺x/Cu(100) films if Fe is substituted for Co? In addition, the complicated transitions of the structural and magnetic anisotropy in Fe/Cu共100兲 films⁹may be involved in the alloy films during the range of the composition in which the SRT is observed. Therefore, it is imperative to conduct a comparative study of the correlation between structure and magnetic properties for Co_xNi₁_⫺x/Cu(100) and Fe_xNi₁_⫺x/Cu(100) films.

In this work, a comparative study, including the structural and magnetic properties, of the magnetic ultrathin Co_xNi₁_⫺x/Cu(100) and Fe_xNi₁_⫺x/Cu(100) alloy films were performed. The alloy composition-driven SRT were found in both of these two alloy films. In addition, the contribution of the concentration of Fe and Co on the critical thickness of the SRT was also estimated and discussed.

The magnetic alloy ultrathin films were prepared and investigated in situ in an UHV chamber.¹⁰The coverage and alloy composition of the alloy films in our experiments can be precisely controlled to an accuracy of 0.05 ML⫾0.5%, respectively.¹¹The crystalline structure as well as interlayer distance of the alloy films were performed via the low energy electron diffraction共LEED兲 and LEED I(E) in the kinematic approximation.³Auger electron spectroscopy was also em- ployed as the tool for studying the chemical structure of the alloy films. In addition, the study of the magnetic hysteresis loops was carried out by means of magneto-optical Kerr effect in polar and longitudinal configurations quasi- simultaneously with benefit of the modulation and lock-in technique to investigate the evolution of magnetic structures, such as spin-reorientation transition.

The medium energy electron diffraction共MEED兲 intensities of Fe_xNi₁_⫺x/Cu(100) and Co_xNi₁_⫺x/Cu(100) alloy ultrathin films for all the alloy compositions (x⭐10%) exhibit the regular oscillations up to four layers, as shown in

a兲Author to whom correspondence should be addressed; electronic mail:

mtlin@phys.ntu.edu.tw

JOURNAL OF APPLIED PHYSICS VOLUME 91, NUMBER 10 15 MAY 2002

7185

Downloaded 14 May 2002 to 140.112.192.1. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/japo/japcr.jsp

(2)

Fig. 1. It is obvious that the characteristic in MEED oscillations of the alloy films look fairly similar to that of the pure Ni/Cu共100兲. Nevertheless, there are significant deviations of MEED characteristic between the alloy films and pure Fe/Cu共100兲 or Co/Cu共100兲 films. It implies both of these alloy films are the Ni/Cu共100兲-like systems within the composition range of our experiments. Similar results can also be found in the LEED patterns for Ni/Cu共100兲, Fe_xNi₁_⫺x/Cu(100), and Co_xNi₁_⫺x/Cu(100) films. The LEED patterns in the Fe_xNi₁_⫺x/Cu(100) and Co_xNi₁_⫺x/Cu(100) films both exhibit the crossing streaks on the LEED spots, just as the Ni/Cu共100兲 films do.¹¹This result agrees well with the previous scanning tunnel microscope study in Ni/Cu共100兲.¹² The 3D platelike islands from the additional periodic structure in the 关011兴 and 关01¯1兴 direc- tions account for the crossing streaks along these two direc- tions in the LEED patterns.

The average vertical interlayer spacing (a_⬜) for both the Fe_xNi₁_⫺x/Cu(100) and Co_xNi₁_⫺x/Cu(100) alloy ultrathin films determined by means of LEED I(E) measurement are compiled in Fig. 2. It is obvious that a_⬜ for all the alloy composition in both alloy films are almost the same within

an accuracy of ⫾0.02 Å. Different from the Fe/Cu共100兲 films, which exhibit a structure transition from fcc to bcc structure,⁹there is no structure variation or lattice relaxation for the alloy films for all the alloy compositions and thickness investigated. In addition, the vertical interlayer spacing for the alloy films are fairly close to the pure Ni/Cu共100兲 films, as shown in Fig. 2. It again confirms our assumption that both Fe_xNi₁_⫺x/Cu(100) and Co_xNi₁_⫺x/Cu(100) films are Ni-dominant systems behaving like the Ni/Cu共100兲 films.

As mentioned, the magnetic properties of the alloy films were characterized by the hysteresis loops in terms of Kerr intensities. The magnetic hysteresis loops of the alloy films Fe_xNi₁_⫺x/Cu(100) for x⫽3%, 4%, 5%, 6%, and 8% as well as Co_xNi₁_⫺x/Cu(100) for x⫽3%, 5.5%, 8%, 9%, and 10%

were investigated at 110 K with variation of the film thickness. The easy axis of these films is identified by both the polar and longitudinal geometries of hysteresis loops. For Co_xNi₁_⫺x/Cu(100) films, the critical thickness of the SRT varies from 7.5 ML for x⫽0 关pure Ni/Cu共100兲兴 to 17.5 ML for x⫽8%. There is no SRT found for x⭓10% with the coverage up to 20 ML. The similar behavior was also found in Fe_xNi₁_⫺x/Cu(100), but it was more sensitive to the Fe concentration. The critical thickness of SRT varies more drastically to 16 ML for x⫽5%, and no such transition can be observed for x⭓6%, as summarized in Fig. 3.

Comparing the electron configurations of the Fe, Co, and Ni atoms, the difference in number of 3d electrons between Fe and Ni is twice as large as that between Co and Ni. It is reasonable to expected that the influence of iron on the magnetic behavior is stronger than cobalt as these two elements are alloyed with Ni. However, the influence of Fe concentration in Fe_xNi₁_⫺x/Cu(100) alloy films was found to be less than expected. The effect of the alloy composition x for Fe on the critical thickness for the SRT is about 1.4 times as large as that for Co, with the estimation of the electron num- ber n_d in 3d minority band made by fitting the magnetostric-

FIG. 1. MEED intensity oscillations of Co/Cu共100兲, Ni/Cu共100兲, Fe/Cu共100兲 and alloy ultrathin films with different compositions. All the films were prepared at 300 K.

FIG. 2. Interlayer distance of FexNi1⫺x/Cu(100) and CoxNi1⫺x/Cu(100) alloy ultrathin films with variation of alloy composition.

FIG. 3. Comparison of the magnetic phase diagrams for FexNi1⫺x/Cu(100) and CoxNi1⫺x/Cu(100) alloy films. The solid circles共䊉兲, hollow circles 共䊊兲, solid diamonds 共⽧兲, and hollow diamonds 共〫兲 represent the out-of- plane and in-plane easy axes for the FexNi1⫺x/Cu(100) films as well as the out-of-plane and in-plane easy axes for the CoxNi1⫺x/Cu(100) films, respectively. The phase boundary of SRT for Fe–Ni and Co–Ni alloy are illustrated by solid and dashed curves, respectively.

7186 J. Appl. Phys., Vol. 91, No. 10, 15 May 2002 Kuoet al.

(3)

tion coefficients ␭001 as a function of n_d.¹³ The deviation between the experimental result and expectation 共two times larger兲 should be attributed to fact that the distribution of electrons in metal is somewhat different from that in a free atom. The 3d-band filling for the solid metals Fe, Co, and Ni are given in Table I.¹⁴With these modified values, the effect of 3d-band filling for Fe caused by the alloy concentration on the magnetic anisotropy is about (2.2⫺0.6)/(1.7⫺0.6)

⫽1.45 times as large as Co, which is very close to our experimental results.

In conclusion, the critical thickness of spin-reorientation transition for alloy ultrathin films Fe_xNi₁_⫺x/Cu(100) and Co_xNi₁_⫺x/Cu(100) were investigated and compared in this work. In comparison to the Co_xNi₁_⫺x/Cu(100) films, Fe_xNi₁_⫺x/Cu(100) exhibit more drastic influence of the al- loy concentration x on the critical thickness of SRT. Based on the insignificant structure variation of the films in all the compositions, these composition-driven spin-reorientation transitions can be attributed to the effect of 3d-band filling on the magnetic anisotropy for the Ni/Cu共100兲-like films.

Based on our experimental results, the difference of the 3d electron number between the majority and minority bands for the Fe–Ni, and Co–Ni alloys was also successfully applied to estimate the effect of 3d-band filling of Fe_xNi₁_⫺x/Cu(100) films as compared with Co_xNi₁_⫺x/Cu(100) films.

This research was supported by the grant from the Na- tional Science Council at Taiwan through Contract No. NSC- 90-2112-M-002-036.

1D. P. Pappas, K.-P. Ka¨mper, and H. Hopster, Phys. Rev. Lett. 64, 3179 共1990兲.

2J. Thomassen, F. May, B. Feldmann, M. Wuttig, and H. Ibach, Phys. Rev.

Lett. 69, 3831共1992兲.

3M.-T. Lin, J. Shen, W. Kuch, H. Jenniches, M. Klaua, C. M. Schneider, and J. Kirschner, Phys. Rev. B 55, 5886共1997兲.

4M. Zheng, J. Shen, P. Ohresser, C. V. Mohan, M. Klaua, J. Barthel, and J.

Kirschner, J. Appl. Phys. 85, 5060共1999兲.

5F. Baudelet, M.-T. Lin, W. Kuch, K. Meinel, B. Choi, C. M. Schneider, and J. Kirschner, Phys. Rev. B 52, 12563共1995兲.

6M. Farle, W. Platow, A. N. Anisimov, P. Poulopoulos, and K. Baberschke, Phys. Rev. B 56, 5100共1997兲.

7A. Dittschar, M. Zharnikov, W. Kuch, M.-T. Lin, C. M. Schneider, and J.

Kirschner, Phys. Rev. B 57, R3209共1998兲.

8M.-T. Lin, W. C. Lin, C. C. Kuo, and C. L. Chiu, Phys. Rev. B 62, 14268 共2000兲.

9S. Mu¨ller, P. Bayer, C. Reischl, K. Heinz, B. Feldmann, H. Zillgen, and M.

Wuttig, Phys. Rev. Lett. 74, 765共1995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.

11W. C. Lin, C. C. Kuo, C. L. Chiu, and M.-T. Lin, Surf. Sci. 478, 9共2001兲.

12J. Shen, J. Giergiel, and J. Kirschner, Phys. Rev. B 52, 8454共1995兲.

13C. C. Kuo, S. F. Chuang, W. Pan, and M.-T. Lin共unpublished兲.

14R. M. Bozorth, Ferromagnetism共IEEE, New York, 1993兲.

TABLE I. Average number of electrons in 3d band for transition metals Fe, Co, and Ni.^a

3d↑ 3d↓ Asymmetry (3d↑ – 3d↓)

Fe 4.6 2.4 2.2

Co 5.0 3.3 1.7

Ni 5.0 4.4 0.6

aSee Ref. 14.

7187

J. Appl. Phys., Vol. 91, No. 10, 15 May 2002 Kuoet al.