Magnetic properties of Ni-NiO (Ferromagnetic-Antiferromagnetic) nanocomposites obtained from a partial mechanochemical reduction of NiO

Copyright © 2008 American Scientific Publishers All rights reserved Printed in the United States of America

Journal of Nanoscience and Nanotechnology Vol. 8, 2923–2928, 2008

Magnetic Properties of Ni-NiO (Ferromagnetic–Antiferromagnetic) Nanocomposites Obtained from a Partial Mechanochemical Reduction of NiO J. Nogués1
∗ , V. Langlais2 , J. Sort3 , S. Doppiu4 , S. Suriñach2 , and M. D. Baró2 1

Institució Catalana de Recerca i Estidus Avan¸cats (ICREA) and Institut Catalá de Nanotecnologia, Edifici CM7, Campus Universitat Autónoma de Barcelona, 08193 Bellaterra, Barcelona, Spain 2 Depratament de Física, Universitat Autónoma de Barcelona, 08193 Bellaterra, Barcelona, Spain 3 Institució Catalana de Recerca i Estidus Avan¸ cats (ICREA) Delivered by Ingenta to:and Depratament de Física, Universitat Autónoma de Barcelona, 08193 Bellaterra, Barcelona, Spain Argonne National Laboratory 4 IFW Dresden, Institute of Metallic P.O. Box 270116, D-01171 Dresden, Germany IPMaterials, : 146.139.245.76

Thu, 19 Jun 2008 21:18:44

The magnetic properties of ferromagnetic (FM)–antiferromagnetic (AFM), Ni-NiO, nanocomposites obtained from a reactive ball milling reduction of NiO in H2 atmosphere have been studied. The formation of ferromagnetic Ni from antiferromagnetic NiO can be accurately followed by the increase of the saturation magnetization. The microstructure of the nanocomposite, consisting of FM Ni nanoparticles embedded in an AFM NiO matrix leads to exchange bias effects, i.e., loop shifts and coercivity enhancement, after field cooling from above the Néel temperature of NiO. Bias.

1. INTRODUCTION The partial reduction of transition metal (TM) oxides has been extensively studied from the applied and fundamental points of view.1–11 In particular, at the early stages of the reduction, composites consisting of TM nanoparticles embedded in a TM oxide matrix can be obtained. This type of microstructure has found applications in different fields such as catalysts, fuel cell electrodes, gas sensors, supercapacitors/battery hybrids or magnetic memories.12–16 In the specific case of ferromagnetic TM (i.e., Fe, Co and Ni), the evolution of the TM-oxide reduction can be accurately monitored, either in-situ or ex-situ, by the progress of the magnetic properties as the TM oxide (antiferromagnetic, AFM, or ferrimagnetic) transforms into strongly ferromagnetic (FM) metal.17–22 Moreover, since some of the TM oxides are AFM (e.g., FeO, CoO or NiO), the arrangement consisting of FM metallic nanoparticles embedded in a AFM oxide matrix leads to interesting magnetic properties, e.g., exchange bias.23–25 The exchange bias coupling between FM-AFM materials is usually obtained by field cooling the FM-AFM system from above the Néel ∗

Author to whom correspondence should be addressed.

J. Nanosci. Nanotechnol. 2008, Vol. 8, No. 6

temperature of the AFM. This procedure leads to a range of induced properties on the FM, such as a shift of the hysteresis loop in the field axis or an increase of the coercivity.23
26 Although traditionally oxide reduction is carried out by high temperature annealing in a H2 atmosphere,1–11 we have recently demonstrated that reduction of oxides can be carried out controllably at room temperature by mechanochemical reactions, i.e., reactive ball milling in a H2 atmosphere.27 At the early stages of milling this process leads to a nanocomposite consisting of metallic nanoparticles (Ni-FM) embedded in a oxide matrix (NiO-AFM). The magnetic properties of this novel nanocomposite have not been reported yet. It is noteworthy that mechanical milling is a widespread technique for the synthesis of a large variety of equilibrium and non-equilibrium phases and phase mixtures. Due to its low cost and straightforward scalability, it has been implemented as a basic processing tool in many industrial procedures, such as the production of permanent magnets.28
29 In this article we present the study of the magnetic properties of ferromagnetic (FM)–antiferromagnetic (AFM) Ni-NiO nanocomposites produced by the reactive ballmilling of pure NiO. It was found that the Ni percentage obtained by magnetic measurements agrees with the one

1533-4880/2008/8/2923/006

doi:10.1166/jnn.2008.010

2923

RESEARCH ARTICLE

Keywords: Nickel Oxide, Nickel, Reactive Ball Milling, Nanocomposites, Coercivity, Exchange

Magnetic Properties of Ni-NiO (Ferromagnetic–Antiferromagnetic) Nanocomposites

Nogués et al.

RESEARCH ARTICLE

obtained from X-ray diffraction after correcting for the Fe contamination from the milling medium. Moreover, at room temperature the system exhibits loop shifts and coercivity enhancements typical of FM-AFM coupling, vanishing around T = 360 K.

crystallite sizes exhibit a similar trend. It is noteworthy that although selective high temperature H2 reduction of oxides produces roughly spherical nanoparticles with rather narrow size distributions,5
25 H2 reduction of oxides carried out in less controlled conditions36
37 or other chemical routes,22 can also result in irregular shapes and broad size distributions. However, reactive ball milling allows for a 2. EXPERIMENTAL DETAILS more controlled partial reduction of oxides and hence a more accurate control of the amount of Ni produced. Sample preparation has been described in detail As can be seen in Figure 2(a), the overall saturation 27 elsewhere. In short, NiO powders were milled at room magnetization, MS , of the composites steadily increases temperature in a planetary mill using stainless steel vials for larger total dose. Since NiO is AFM and has no net and balls, under H2 (6 bar) atmosphere at a frequency of magnetic moment the increase of the MS is ascribed to the 300 rpm. Two different types of milling experiments were formation of ferromagnetic Ni during the reactive milling. carried out: (i) at the same milling intensity (21.3 W—ball of Ni (M (Ni) = 57 5 emu/g) the percentage From M S S to powder ratio 12:1) varying the milling time (2, 4, 8, or of Ni in the composite can be estimated. In Figure 2(b), 12 h) and (ii) at different milling intensities varying the the evolution of the percentage of reduced Ni with the ball to powder ratio (5.3, 10.6, 21.3, and 31.9 W—ball total dose obtained from magnetic measurements and from to powder ratio 3:1, 6:1, 12:1 and 18:1) for a constant quantitative analysis of the XRD are compared. It can be milling time of 8 h. The total milling energy per unit mass by Ingenta to: Delivered observed that the amount of Ni appears to be larger when (i.e., total dose) transferred during the milling is givenNational by Argonne Laboratory measured magnetically. A closer inspection of the XRD the ET = It/mp , where I is the milling intensity,30
31IPt :is146.139.245.76 data indicates that a small amount of Fe is present in the 32
33 milling time and mp is the mass of the powder. Thu, 19 Jun 2008 21:18:44 samples, which increases for higher total dose, reaching X-ray diffraction (XRD) was performed in the 10–120 about 2% of the total sample for the largest dose. This range on the as-milled samples with a Philips 3050 diffracsmall Fe content is debris from the milling medium, vials tometer using CuK radiation. The analysis of the diffracand balls, concomitant with the synthesis method. The tion patterns was carried out with a fitting program based contribution of the Fe can be easily corrected by using on the Rietveld full pattern fitting method.34
35 TransMS (Fe) = 221 9 emu/g and assuming a linear relationship mission electron microscopy imaging was obtained at an between the amount Fe and the total milling dose. After accelerating voltage of 200 kV using a JEOL JEM-2011. the correction the agreement between the structural and Magnetic measurements were carried out at room tempermagnetic results is quite reasonable (see Fig. 2(c)). It is ature using a vibrating sample magnetometer with a maxinoteworthy that magnetization measurements are far more mum applied field of 12 kOe. To induce exchange coupling precise than quantitative analysis of XRD spectra. Thus, between the AFM NiO and the FM Ni, the as-milled powin principle, MS would be a valuable tool for the study ders were field cooled in HFC = 6 kOe, from 600 K, i.e., of the first stages of the magnetic transition metal oxides, above TN (NiO) = 525 K, to room temperature. although in our case due to the presence of Fe contamination the accuracy is somewhat reduced specially for large milling doses. For example, for a dose of 30 kJ/g, 3. RESULTS AND DISCUSSION obtained for a ball-to-powder ratio of 3:1, although XRD The structural evolution of the NiO reduction was measurements show no appreciable amount of Ni, magdescribed in detail in Ref. [27]. Briefly, as the milling netic measurements indicate the presence of about 1.5% energy increases NiO transforms progressively into Ni, of Ni. This is consistent with the error in determining although there is a threshold energy necessary to initiate phase percentages using the Rietveld method which should the transformation. The maximum amount of Ni obtained be around 2%. in the present conditions was about 20%. The morphology Note that although the milling conditions of the two consisting of metallic Ni nanoparticles embedded in an experiments were chosen to lead roughly to the same total oxide NiO matrix can be observed in Figures 1(a and c). doses, there is some scattering in MS . This indicates that The polycrystalline nature of the sample is evidenced by although, in principle, both types of processes should be the presence of several Ni and NiO rings in the selected equivalent, it seems that perhaps higher milling intensity area diffraction rings (see Fig. 1(b)). As can be seen in for shorter times is more efficient than lower intensity Figure 1(c), the Ni nanoparticles exhibit irregular shapes for longer times to reduce NiO into Ni. This could arise with a rather broad size distribution. The NiO crystallite because the high energy milling conditions could be more size, obtained from XRD, decreases from >100 m in the effective in producing defects in NiO, which are known to unmilled powder to ∼25 nm for the highest milling energy. act as seeds for the formation of Ni.6
7
10 Contrarily, the Ni XRD crystallite size initially increases, Shown in Figure 3 is the coercivity, HC , of the as milled powders as a function of the total dose. There levelling off at DNi ∼ 14 nm (see Fig. 1(c)).27 The TEM 2924

J. Nanosci. Nanotechnol. 8, 2923–2928, 2008

Nogués et al.

Magnetic Properties of Ni-NiO (Ferromagnetic–Antiferromagnetic) Nanocomposites

Ni (nm)

Delivered by Ingenta to: 20 Argonne National Laboratory IP : 146.139.245.76 Ball-to-powder ratio Milling time Thu, 19 Jun 2008 21:18:44

18:1; 8 h

12:1 - 8 h

15

12:1; 8 h 12:1; 6 h

12:1; 12 h

10 6:1; 8 h

(d)

5 0

50

100

150

200

Total dose (kJ/g)

Fig. 1. (a) Bright field transmission electron micrograph of the Ni-NiO nanocomposite with 184 kJ/g dose (obtained for a ball-to-powder ratio of 12:1 and a milling time of 12 h); (b) electron diffraction indexed for Ni and NiO; (c) dark field corresponding to the Ni(200) peak of the highlighted area of Figure 1(a); (d) dependence of the Ni XRD crystallite size, DNi , on the total dose obtained during the milling. Indicated in the plot are milling conditions. Note that the half filled symbols correspond to the sample with a ball-to-powder ratio of 12:1 milled for 8 hours, which belongs to both series of experiments.

is a clear tendency for HC to increase as the total dose becomes larger. Hence, when comparing Figure 1(d) and Figure 3, it is clear that HC increases with increasing DNi . This fact may appear unusual at first sight, since usually the contrary trend is held.38 However, it has to be taken into account that the critical size for single domain behaviour at room temperature for spherical Ni nanoparticles is about DCrit ∼ 55 nm. Thus, DNi for all studied doses is clearly below DCrit , consequently the nanoparticles will be single domain but prone to thermal fluctuations. As the nanoparticles become larger they become less sensitive to thermal fluctuations and thus HC increases. Interestingly, from the definition of blocking temperature, TB = KV /25kB (where K is the anisotropy, V the volume and kB the Boltzman constant) and given the low anisotropy of Ni (KNi ∼ 5 0 × 104 erg/cm3 (Ref. [38])) one J. Nanosci. Nanotechnol. 8, 2923–2928, 2008

would expect that in order for Ni nanoparticles to remain ferromagnetic at room temperature (i.e., TB = 300 K) they should be larger than about D = 34 nm. However, our Ni nanoparticles are clearly ferromagnetic, i.e., HC  0, despite DNi < 34 nm. This effect has several origins. First, to carry out the calculations we have used bulk values of K, however in nanoparticles surface anisotropy may play an important role.38 Due to the irregular morphology of the Ni nanoparticles, shape anisotropy38 could also play a key role in the observed effects. Moreover, other anisotropies such as strain anisotropy could also be a significant contribution to the total anisotropy of the nanoparticles.39 Finally, it has been shown that when FM nanoparticles are embedded in an AFM matrix an additional uniaxial anisotropy is also introduced to the nanoparticles.16
23 The larger effective anisotropy increases 2925

RESEARCH ARTICLE

3:1; 8 h

Magnetic Properties of Ni-NiO (Ferromagnetic–Antiferromagnetic) Nanocomposites 25

220 200 180

15

12:1; 12 h

HC (Oe)

MS (emu/g)

18:1; 8 h

Ball-to-powder ratio Milling time 12:1 - 8 h

20

160 140

10

12:1; 6 h

12:1; 8 h

6:1; 8 h

100

12:1; 2 h

(a)

Ni% - XRD (ball-to powder ratio) Ni% - XRD (milling time) Ni% -MS (ball-to powder ratio) Ni% -MS (milling time) , 12:1 - 8 h

Ni %

10

(b) 0 Ni% - XRD (balls) Ni% - XRD (time) Ni% -MS (balls) corr Ni% -MS (time) corr , 12:1 - 8 h

Ni %

20

10

(c) 0

20

40

60

80

100

120

140

160

180

200

Total dose (kJ/g) Fig. 2. (a) Dependence of the saturation magnetization, MS of the NiNiO composites on the total dose (indicated in the plot are milling conditions); (b) dependence of Ni percentage on the total dose as obtained from MS and X-ray diffraction (XRD); (c) dependence of Ni percentage in the Ni-NiO composites on the total dose as obtained from MS after correction for Fe contamination. The data for Ni% obtained from XRD is also shown for comparison. The lines are guides for the eye. Note that the half filled symbols correspond to the sample with a ball-topowder ratio of 12:1 milled for 8 hours, which belongs to both series of experiments.

2926

100

150

200

Delivered by Ingenta to: temperature, thus leading to the thermal stathe blocking Argonne Nationalbility Laboratory of the nanoparticles.16
23
25
40 IP : 146.139.245.76 Figure 4 shows the evolution of the loop shift (HE ) Thu, 19 Jun 2008 and21:18:44 coercivity enhancement (HC , i.e., the difference of

20

0

50

Fig. 3. Dependence of the coercivity, HC , of the as milled powders on the total dose. Note that the half filled symbols correspond to the sample with a ball-to-powder ratio of 12:1 milled for 8 hours, which belongs to both series of experiments.

0

30

0

Total dose (kJ/g)

3:1; 8 h

30

Ball-to-powder ratio Milling time 12:1 - 8 h

120

5

RESEARCH ARTICLE

Nogués et al.

coercivity between after and before field cooling) for different total dose. The observed exchange bias effects are somewhat small, HE = 10 Oe and HC = 30 Oe. Nevertheless, these effects are expected to be small for systems containing a low anisotropy AFM such as NiO (KNiO ∼ 2 8 × 106 erg/cm3 (Ref. [41])).23
26 Interestingly, opposite trends are observed for HE and HC . Namely, while HE decreases with milling time, HC increases. This effect is probably related to the progressive creation of Ni and the development of defects in NiO during the reactive milling. For short milling times only a reduced number of small FM nanoparticles is created. As the milling proceeds, the Ni nanoparticles grow in size and new nanoparticles are formed. Since in nanoparticle systems HE is assumed to be inversely proportional to the FM particle size,23
26 the reduction of the loop shift is, to a certain extent, expected for larger doses since DNi increases. The increase of HC probably arises from a combination of several effects. For example, as the milling proceeds more defects are induced in NiO, e.g., crystallite size reduction or nonstoichiometry and this probably favours partial dragging of the AFM spins during magnetization reversal of the FM, hence increasing HC .23
26 Simultaneously, the Ni nanoparticles increase their intrinsic coercivity (i.e., HC before the field cooling process) for larger milling doses. This fact could also influence the coercivity enhancement, if, for example, the coercivity enhancement would depend on the FM anisotropy.42 In this system, although the loop shift and coercivity enhancement change markedly with the milling energy, the induced remanence enhancement is rather small, in contrast with what has been observed in other FM nanostructures embedded in, or in contact with, an AFM.16
23
24 J. Nanosci. Nanotechnol. 8, 2923–2928, 2008

Nogués et al.

Magnetic Properties of Ni-NiO (Ferromagnetic–Antiferromagnetic) Nanocomposites

30 Ball-to-powder ratio Milling time 12:1 - 8 h

200

(a) HC (Oe) (after field cooling)

25

∆HC (Oe)

20 15 10 5

(a)

180

160

140

120

0 100

–5

(b) (b)

8 10

Delivered by Ingenta to: 4 Argonne National Laboratory IP : 146.139.245.76 Thu, 19 Jun 2008 21:18:44 2 HE (Oe)

HE (Oe)

6

5

0

0

0

50

100

150

200

–2

300

320

340

360

Total dose (kJ/g)

Finally, when increasing the measuring temperatures, HE vanishes at about T = 360 K. HC , after field cooling, decreases progressively with temperature, exhibiting a slight kink around T = 400 K (see Fig. 5). The analysis of HC is more complex since the changes due to the interface coupling are superimposed to the strong decrease of the intrinsic coercivity of Ni nanoparticles, due to their small size. However, both the loss of HE and the kink in HC occur at temperatures much lower than TN (NiO), contrarily to systems where FM + NiO have been ball milled together, rather than obtained from mechanochemical reduction. This agrees with the argument that reactive milling may be inducing more structural damage to NiO than conventional AFM + FM milling.43
44 In conclusion the progressive formation of Ni during the mechanochemical reduction of NiO has been studied magnetically. The amount of Ni obtained from magnetization measurements is consistent with X-ray results after correction of the milling medium contamination. The microstructure consisting of ferromagnetic Ni nanoparticles embedded in an antiferromagnetic NiO matrix leads J. Nanosci. Nanotechnol. 8, 2923–2928, 2008

400

420

440

460

Fig. 5. Dependence of (a) the coercivity, HC , and (b) the exchange bias, HE , after field cooling, on measuring temperature, T , for the sample milled for 8 h in a 12:1 ball-to-powder ratio. The lines are guides to the eye.

to the existence of loop shifts and coercivity enhancements in the system. Acknowledgment: Financial support from the MAT2007-66302-C02, the MAT2007-61629 and the 2005SGR-00401 research projects is acknowledged.

References and Notes 1. H. Yan, C. F. Blanford, B. T. Holland, M. Parent, W. H. Smyrl, and A. Stein, Adv. Mater. 11, 1003 (1999). 2. J. A. Rodriguez, J. C. Hanson, A. I. Frenkel, J. Y. Kim, and M. Pérez, J. Am. Chem. Soc. 124, 346 (2002). 3. B. Delmon, Handbook of Heterogeneous Catalysis, edited by G. Ertl, H. Knozinger, and J. Weitkamp, Wiley-VCH, New York (1997). 4. R. P. Fusternau, G. McDougall, and M. A. Langell, Surf. Sci. 150, 55 (1985). 5. Ch. Laurent, Ch. Blaszczyk, M. Brieu, and A. Rousset, Nanostruct. Mater. 6, 317 (1995). 6. L. Soriano, M. Abbate, A. Fernández, A. R. González-Elipe, F. Sirotti, G. Rossi, and J. M. Sanz, Chem. Phys. Lett. 266, 184 (1997).

2927

RESEARCH ARTICLE

Fig. 4. Dependence of (a) the enhanced coercivity, HC , and (b) exchange bias, HE , after field cooling on the total dose. The lines are guides to the eye. Note that the half filled symbols correspond to the sample with a ball-to-powder ratio of 12:1 milled for 8 hours, which belongs to both series of experiments.

380

T (K)

RESEARCH ARTICLE

Magnetic Properties of Ni-NiO (Ferromagnetic–Antiferromagnetic) Nanocomposites

Nogués et al.

7. J. T. Richardson, R. Scates, and M. V. Twigg, Appl. Catal. A: Gen. 24. J. Nogués, J. Sort, V. Langlais, S. Doppiu, B. Dieny, J. S. Muñoz, 246, 137 (2003). S. Suriñach, M. D. Baró, S. Stoyanov, and Y. Zhang, Int. J. Nano8. A. F. Benton and P. H. J. Emmett, J. Am. Chem. Soc. 74, 1194 technol. 2, 23 (2005). (1924). 25. J. Sort, V. Langlais, S. Doppiu, B. Dieny, S. Suriñach, J. S. Muñoz, 9. V. V. Boldyrev, M. Bulens, and B. Delmon, The Control of the ReacM. D. Baró, Ch. Laurent, and J. Nogués, Nanotechnology 15, S211 tivity of Solids, Elselvier, Amsterdam (1979). (2004). 10. J. M. McKay and V. E. Henrich, Phys. Rev. B 32, 6764 (1985). 26. J. Nogués and I. K. Schuller, J. Magn. Magn. Mater. 192, 203 (1999). 11. V. I. Budarin, V. Diyuk, L. Matzui, L. Vovchenko, T. Tsetkova, and 27. S. Doppiu, V. Langlais, J. Sort, S. Suriñach, M. D. Baró, Y. Zhang, M. Zakharenko, J. Therm. Anal. Calorim. 62, 345 (2000). G. Hadjipanayis, and J. Nogués, Chem. Mater. 16, 5664 (2004). 12. V. V. Belov, O. V. Isaev, L. G. Romanovskaya, E. V. Belyaeva, L. I. 28. B. S. Murty and S. Ranganathan, Internat. Mater. Rev. 43, 101 Sula, and P. N. Voronin, Neftekhimiya 31, 642 (1991). (1998). 13. M. Murai, K. Takizawa, K. Soejima, and H. Sotouchi, J. Elec29. C. Suryanarayana, Prog. Mater. Sci. 46, 1 (2001). trochem. Soc. 143, 2481 (1996). 30. M. Abdellaoui and E. Gaffet, J. Alloys Comp. 209, 351 (1994). 14. I. Hotovy, J. Huran, P. Siciliano, S. Capone, L. Spiess, and 31. M. Abdellaoui and E. Gaffet, Acta Metall. Mater. 43, 1087 (1995). V. Rehacek, Sens. Actuators, B 78, 126 (2001). 32. F. Delogu, L. Schiffini, and G. Cocco, Philos. Mag. A 81, 1917 15. P. A. Nelson and J. R. Owen, J. Electrochem. Soc. 150, A1313 (2001). (2003). 33. N. Burgio, A. Iasonna, M. Magini, S. Martelli, and F. Padella, Nuovo 16. V. Skumryev, S. Stoyanov, Y. Zhang, G. Hadjipanayis, D. Givord, Cimento D 13, 459 (1991). and J. Nogués, Nature 423, 850 (2003). 34. R. A. Young, The Rietveld Method, International Union of Crystal17. P. A. Chernavskii, G. V. Pankina, and A. S. Lermontov, Rus. J. Phys. lography, Oxford University Press, Oxford (1995). Chem. 79, 1014 (2005). 35. L. Lutterotti and S. Gialanella, Acta Mater. 46, 101 (1997). 18. D. L. Hou, X. F. Nie, and H. L. Luo, J. Magn. Magn. Mater. 188, 36. D. Potoczna-Petru and L. K¸epi´nski, Catal. Lett. 73, 41 (2001). Delivered by Ingenta to: H. K. Lin, and C. W. Tang, Catal. Lett. 94, 69 (2004). 169 (1998). 37. C. B. Wang, 19. U. Wiedwald, K. Fauth, M. Heßler, H. G. Boyen, F. Wigl, M. HilgenR. Skomski, J. Phys.: Condens. Mater. 15, R841 (2003). Argonne National38.Laboratory dorff, M. Giersig, G. Schütz, P. Ziemann, and M. Farle, Chem. Phys. 39. S. Chikazumi, Physics of Ferromagnetism, Oxford University Press, IP : 146.139.245.76 Chem. 6, 2522 (2005). New York (1997). Thu, 19 Jun 2008 21:18:44 20. G. Ennas, G. Marongiu, S. Marras, and G. Piccaluga, J. Nanopart. 40. D. Givord, V. Skumryev, and J. Nogués, J. Magn. Magn. Mater. 294, Res. 6, 99 (2004). 111 (2005). 21. A. A. Khassin, V. F. Anufrienko, V. N. Ikorskii, L. M. Plyasova, 41. M. T. Hutchings and E. J. Samuelsen, Phys. Rev. B 6, 3447 (1972). G. N. Kustova, T. V. Larina, I. Y. Molina, and V. N. Parmon, Phys. 42. S. M. Zhou and C. L. Chien, Phys. Rev. B 63, 104406 (2001). Chem. Chem. Phys. 4, 4236 (2002). 43. J. Sort, J. Nogués, X. Amils, S. Suriñach, J. S. Muñoz, and M. D. 22. C. Prada and E. Morán, Chem. Mat. 18, 2729 (2006). Baró, Appl. Phys. Lett. 75, 3177 (1999). 23. J. Nogués, J. Sort, V. Langlais, V. Skumryev, S. Suriñach, J. S. 44. J. Sort, J. Nogués, X. Amils, S. Suriñach, J. S. Muñoz, and M. D. Muñoz, and M. D. Baró, Phys. Rep. 422, 65 (2005). Baró, Mater. Sci. Forum 343–346, 812 (2000).

Received: 2 June 2006. Accepted: 17 November 2006.

2928

J. Nanosci. Nanotechnol. 8, 2923–2928, 2008