ELSEVIER Physica B 208&209 (1995) 781-783
IWIIlU
Orientation of the Ni magnetic moments in N i - Z n ferrites
W.F. Pong a'*, Y.K. Chang a, M.H. Su a, H.J. Lin b, G.H. Ho b, P.K. Tseng c, T.M. Uen a,
G. Meigs e, C.T. Chen e
a Department of Physics, Tamkang University, Tamsui, Taiwan 251, Republic of China b Synchrotron Radiation Research Center, Hsinchu, Taiwan 300, Republic of China CDepartment of Physics, National Taiwan University, TaipeL Taiwan 107, Republic of China d Department of Electrophysics, National Chaio Tung University, Hsinchu, Taiwan 300, Republic of China
eAT&T Bell Laboratories, Murray Hill, NJ 07974, USA
Abstract
We report the Ni L2.3-edges magnetic circular dichroism measurements of the Zn~Nil _ xFe204 (x = 0.0, 0.26, 0.50, and 0.75) ferrimagnet. The Ni average magnetic moments were found to decrease significantly with increasing Zn content. This observation is interpreted in terms of the Yafet-Kittel type canting of the Ni magnetic moment and shows values greater than those of the Fe magnetic moment determined from M6ssbauer and neutron diffraction measurements. This work provides new information on the effect of Zn doping of N i - O - N i and F e - O - F e superexchange couplings.
Z n ~ N i l - x F e 2 0 4 is a ferrimagnetic compound which exhibits interesting doping-dependent magnetic proper- ties. The saturation magnetization and the N6el temper- ature were found to change with Zn content; and a Yafet-Kittel (Y-K) type canting of the local Fe mag- netic moments has been proposed to account for these observations I-1, 2]. Due to the lack of Ni-specific mag- netic measurements, however, the role of Ni 2 ÷ ions in the magnetic properties of the N i - Z n ferrites is still not understood. F o r example, whether the Ni magnetic mo- ments have a Y - K type canting similar to that of Fe magnetic moments, and how the F e - O - F e and N i - O - N i superexchange couplings are modified upon Zn doping remain as interesting open questions. To help reveal the role of the Ni 2 ÷ ions in this magnetic system, soft-X-ray magnetic circular dichroism (MCD) measure- ments at the Ni L2.a-edges were conducted.
The soft-X-ray M C D measurements were performed at the AT&T Bell Laboratories Dragon beamline at the * Corresponding author.
National Synchrotron Light Source [3] using the fluo-. rescence yield X-ray absorption spectroscopic (XAS) technique. The sample preparation, X-ray diffraction, and the chemical analysis of the ZnxNil -xFe204 samples used in this study have been described elsewhere [4]. To facilitate M C D measurements, the samples were ground and sieved to fine powders, and then rubbed onto an adhesive type which was kept at room temperature.
Fig. 1 shows the photon-flux-normalized Ni L2,a-edges XAS and M C D spectra of Zno.sNio.sFe20 4. The I+ (I_) is the absorption spectrum taken with the projection of the spin of the incident photons parallel (anti-parallel) to the spin direction of the Ni 3d majority electrons. The two white line regions, labeled as L3 and L2 in the top panel, are the electron transitions from the Ni 2pa/2 and 2pl/2 core levels to the Ni 3d unoccupied states, respec- tively [5, 6]. The MCD spectrum, e.g. I + - I_, is shown in the bottom panel. Unlike the case of solid element nickel [5], rather complicated XAS and M C D spectra were observed, exhibiting strong multiplet and crystal field effects [6]. The lineshapes of the XAS and M C D
0921-4526/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved
782 W..F. Pong et al./Physica B 208&209 (1995) 781 783 120 100 80 60 40 20 0 i L a i t i Zno.5Nio.sFez0 4 - - I + ... I L z > . . e. CY~ z z r ~ 20 10 - 1 0 - 2 0 / - 3 0 ~ i I 840 850 I I I I 860 870 880 890 PHOTON ENERGY (eV)
Fig. 1. Ni L2.3-edges fluorescence yield X-ray absorption (XAS) and magnetic circular dichroism (MCD) spectra of Zno.sNio.sFe204.
spectra of x = 0, 0.26, and 0.75 samples were found to be nearly identical to those of the x = 0.5 sample, except for their M C D asymmetry ratio, i.e. (I+ - I _ ) / ( I + + I_). Since the M C D asymmetry ratio is proportional to the magnetic m o m e n t averaged over different sites a n d mag- netic orientations, a quantitative evaluation of the asym- metry ratio can provide direct information on the Ni relative average magnetic m o m e n t as a function of Z n content. A m o n g the various spectral regions from which one can calculate the asymmetry ratio, the L2 white line region is the best choice. This is because the main peak at ~ 855 eV in the L 3 white line region suffers seriously from self-absorption and saturation artifacts due to its large absorption cross-section; it is also very difficult to isolate the positive M C D signal of the satellite peak at ~ 857 eV from the negative M C D signal of the main peak. After the removal of a linear background from the L2 white line (see the dashed straight line in Fig. 1), the XAS intensities between 870.9 and 875.6 eV were integ- rated separately for I + and I _ spectrum. The asymmetry ratios thus obtained are 0.2241, 0.2241, 0.1926, a n d 0.1413 for the x = 0, 0.26, 0.5, and 0.75 samples,
80 60 ,~ 40 20 0e 0.0 ZnxNil_xFe204 O MCD (Ni L2) [] Mossbauer <> M r s s b a u e r t /x N e u t r o n / / / / 0.2 0.4 0.6 0.8 1.0
Fig. 2. Yafet-Kittel canting angles, 0vK, as a function of Zn content, x, from different magnetic measurements of ZnxNiz xFe204. The dashed line serves only as a guide to the eyes for the MCD data.
respectively, showing a progressive reduction of the Ni average magnetic m o m e n t with Z n doping over a certain threshold.
This p h e n o m e n a can be interpreted as a non-collinear magnet m o m e n t arrangement proposed for N i - Z n fer- rites by Yafet and Kittel [7]. F o r simple ferrites such as M F e 2 0 4 (M = Mn, Co, a n d Ni), the M substitutes the Fe 3+ octahedral sites a n d leaves the Fe 3+ octahedral sites intact. Regardless of whether the intra-sublattice coupling (between magnetic moments in octahedral a n d tetrahedral sites) is ferromagnetic or antiferromagnetic, the magnetic moments of M 2 ÷ or Fe 3 ÷ in the octahedral sites are parallel to one another, i.e. collinear, just as the magnetic moments in the Fe 3 + tetrahedral sites are I-8]. U p o n Z n doping of the NiFe204, however, Yafet and Kittel proposed that a splitting of the octahedral-site into two subsites occurs, making the magnetic moments in the two subsites equal in magnitude but canted by an angle of 0vK with respect to the direction of their combined moment. The resultant combined magnetic m o m e n t of the octahedral subsites (Fe 3+ or Ni 2 +)is still collinear with, but antiparallel to, that of the tetrahedral Fe 3 ÷ sites. By adopting this interpretation, we can relate the observed reduction of the average Ni moments to an increase in 0VK- If we assume that 0VK = 0 for X = 0, then 0vK for any x can be calculated as 0w(X) = c o s - 1 (asy(x)/ asy(0)), where asy(x) is the aforementioned M C D asym- metry ratio. The 0vK of the Ni magnetic moments thus determined are 0 °, 0 °, 31 °, a n d 51 ° for x = 0, 0.26, 0.5, a n d 0.75 samples, respectively.
W.F. Pong et al./Physica B 208&209 (1995) 781-783 783
Fig. 2 compares the 0vK of the Ni moment determined from our M C D measurements with those obtained from neutron diffraction [1] and M~Sssbauer I-2] measure- ments. Since the M/Sssbauer technique measures only the Fe moments, the 0yK thus determined are attributed to the Y - K canting of the Fe moments. The 0we versus x plot of the Ni moment exhibits similar characteristics to that of the Fe moment, i.e. a monotonic increase in 0vK for Zn content above a threshold value, thus demonstrat- ing that the Y - K type canting also occurs for the Ni magnetic moments. More interestingly, the threshold Zn content for the Ni moments seems to be lower than that for the Fe moments, and the 0vj: are larger for Ni than for Fe. Considering that the Mfssbauer data were taken at 7 K, these differences are even more pronounced. This clearly suggests that the influence of Zn doping on the
superexchange coupling is different for N i - O - N i as com- pared to F e - O - F e .
References
[1] N.S. Satya Murthy et al., Phys. Rev. B 181 (1969) 969. [2] L.K. Leung et al., Phys. Rev. B 8 (1973) 29.
[3] C.T. Chen, Rev. Sci. Instrum. 63 (1992) 1229 and references therein.
[4] T.M. Uen and P.K. Tseng, Phys. Rev. B 25 (1982) 1848. [5] C.T. Chen et al., Phys. Rev. B 42 (1990) 7262; B 43 (1991)
6785.
I-6] G, van der Laan and B. T. Thole, Phys. Rev. B 43 (1991) 13401.
[7] Y. Yafet and C. Kittel, Phys. Rev. 87 (1952) 290. [8] L. N6el, Ann. Phys. 3 (1948) 137.
