Nuclear Instruments and Methods in Physics Research B 216 (2004) 245–250 www.elsevier.com/locate/nimb
Structural and magnetic properties of Fe–Al silica composites prepared by sequential ion implantation C. de Juli an Fern andez a,*, M.A. Tagliente b, G. Mattei a, C. Sada a, V. Bello a, C. Maurizio c, G. Battaglin d, C. Sangregorio e, D. Gatteschi e, L. Tapfer b, P. Mazzoldi a a
e
Dip. Fisica, INFM Univ. Padova, via Marzolo 8, 54124 Padova, Italy b ENEA-CR Brindisi, SS.7 Appia km. 714, 72100 Brindisi, Italy c INFM – GILDA, ESRF, Rue J. Horowitz, BP 200 Grenoble, France d INFM Dip. Chimica-Fisica, Univ. Venezia, Dorsoduoro 2137, 30170 Mestre, Italy LAMM – Dip. Chimica, Univ. Firenze, via della Lastruccia 3, 50019 Sesto Fiorentino, Italy
Abstract The nanostructural and magnetic properties of Fe–Al/SiO2 granular solids prepared by ion implantation have been investigated. A strong effect of the implantation order of the Fe and Al ions has been evidenced. By implanting first the Al ions and later Fe ions, 5–40 nm core–shell nanoparticles are formed with a magnetic behavior similar to that of Fe. The lattice parameter of the nanoparticles is consistent with that of the a-Fe. By changing the implantation order, 10–15 nm core–shell nanoparticles of a bcc Fe-based phase with a lattice 2.5% smaller than that of a-Fe are formed. The temperature dependence of the magnetization indicates a superparamagnetic behavior. 2003 Elsevier B.V. All rights reserved. PACS: 36.40.C; 75.50; 61.46; 68.55.L Keywords: Magnetic nanoparticles; Silica composites; Fe–Al alloys; Ion implantation; Superparamagnetism
1. Introduction Composite materials made of metal particles with nanometric dimensions dispersed in dielectric or metallic matrices have received an increasing interest due to their peculiar properties which can
* Corresponding author. Tel.: +39-049-827-7040; fax: +39049-827-7003. E-mail address:
[email protected] (C. de Julian Fern andez).
be exploited for magnetic, transport, optoelectronic, photonic and catalytical applications. In particular, bimetallic nanoparticles are very promising candidates for technological purposes because the composition is a further freedom degree to tailor their properties for specific purposes. Ion-implantation is a suitable technique to obtain bimetallic nanoparticles in dielectric matrix [1–3]. In this way, it has been obtained composite systems consisting of dielectric matrices with Au–Cu, Au–Ag, Au–Pd clusters with interesting optical properties [1,3,4], with Co–Ni, Co–Cu, Fe–Pt,
0168-583X/$ - see front matter 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.nimb.2003.11.041
246
C. de Julian Fernandez et al. / Nucl. Instr. and Meth. in Phys. Res. B 216 (2004) 245–250
Co–Pt clusters with peculiar magnetic properties [5–7] and with GaN, GaAs, ZnSn clusters for luminescence applications [2,3,8]. In this work, we present a study of the nanostructural and magnetic properties of composites prepared by sequentially implanting Fe and Al ions in silica matrix in order to form Fe–Al nanoparticles. In the bulk Fe–Al alloys it was found that starting from pure Fe, the progressive substitution of Fe sites by Al atoms leads to a gradual decrease of the magnetization and for an Al concentration of about 35% the alloy becomes non-magnetic [9,10]. In this range of compositions two ordered compounds are present: the ferromagnetic Fe3 Al and the non-magnetic FeAl. However, alloys prepared by non-equilibrium techniques, like rapid quenching or mechanical milling, are magnetic up to 50 at.% Al [11,12]. This was explained considering that the composition and the structural disorder produce the competition of ferromagnetic and antiferromagnetic interactions between the Fe atoms [12,13]. Ion implantation technique is an out-of-equilibrium technique, so it is expected that Fe–Al nanoparticles show this effect even if it modified by the size [13]. Moreover, in this paper the magnetic properties of these composites and the relationship with the nanostructure and composition have been investigated.
2. Experimental The same doses of Fe and Al ions (15 · 1016 / cm2 ) were sequentially implanted at room temperature into fused silica (Heraeus Herasil), at energies of 110 keV for Fe and 50 keV for Al, such that the respective projected ranges were similar. Also, single-implanted with Fe and Al samples were prepared with the same implantation dose and energies. Fe–Al implants were performed by implanting first Fe ions and later Al ions whereas in Al–Fe samples the ion implantation order was reversed. Implants were performed at the ENEA Ion Implantation laboratory in Brindisi (Italy). Samples were not annealed after implantation. GIXRD experiments were carried out employing a X-ray diffractometer Philips MPD PW1880
in Parallel Beam geometry, equipped with a X-ray tube emitting CuKa radiation ðkCuKa ¼ 0:154186 nm) and operated at 40 kV, 40 mA. The incident X-ray beam was fixed at 0.5 (which corresponds to a penetration depth of 280 nm in silica) and the detector was moved along the goniometer circle in the 2h range between 10 and 70. The samples for transmission electron microscopy (TEM) were prepared and examined at CNR-LAMEL Institute in Bologna with a FEI TECNAI F20 Supertwin field emission microscope, operating at 200 kV. Rutherford backscattering spectrometry (RBS) measurements were performed using 4 Heþ ions of energy 2.2 MeV at the Van der Graaff Accelerator of INFN Legnaro Laboratory. The magnetic characterization was carried out using a Cryogenic S600 SQUID magnetometer. Zero-field-cooled (ZFC) and field-cooled (FC) measurements were performed applying a magnetic field of 5 mT in the plane of the glass slide.
3. Results and discussion From RBS measurements, we observed that the implanted ion doses in the Fe–Al and Al–Fe samples differed from the nominal ones. In particular, in the Fe–Al sample the measured doses were 15 · 1016 Feþ /cm2 and 10 · 1016 Alþ /cm2 so that the average composition was Fe60 Al40 . In the Al–Fe sample, the respective doses were 17 · 1016 Feþ /cm2 and 8 · 1016 Alþ /cm2 and the average composition was about Fe70 Al30 . Figs. 1(a) and (b) show the bright-field crosssectional micrographs of the Al–Fe and Fe–Al samples, respectively. In both samples, nanoparticles dispersed in the matrix and concentrated in a layer with a width of about 100 nm are observed. The morphology of the nanoparticles appears to depend on the implantation order: in the Al–Fe sample the nanoparticles have diameters between 3 and 40 nm whereas in the Fe–Al sample the particle size is between 10 and 15 nm. In both cases, the smaller nanoparticles are dispersed in the deeper side of the implanted region and a core– shell structure is observed in the larger nanoparticles. However, in the Al–Fe sample the core–shell nanoparticles had several cores while in the Fe–Al
C. de Julian Fernandez et al. / Nucl. Instr. and Meth. in Phys. Res. B 216 (2004) 245–250
247
Al(220) Fe(111)
Al(200) Fe(110)
Al(111)
one there were single-core clusters. Further studies on the morphology are in progress. Fig. 2 shows the GIXRD scans of the Al, Fe, Fe–Al and Al–Fe samples. The patterns corresponding to the single-implanted samples Fe and Al exhibit the diffraction peaks of the bulk bcc aFe and bulk fcc Al phases, respectively. The pattern of the sample Al–Fe shows a broad diffraction peak which was ascribed to the (1 1 0) diffraction of the bcc a-Fe phase. No peak is observed that indicates the presence of metallic aluminium. In the Fe–Al sample pattern the Al (1 1 0) peak is observed, indicating the presence of Al crystalline domains. The pattern of sample Fe–Al shows two others diffraction peaks at about 46 and 67 from an other crystalline phase not clearly identifiable. A detailed peak shape analysis was accomplished to these patterns. In particular, the diffraction line profiles were fitted by a pseudo-Voigt function and it was assumed that the background in proximity of the reflection is linear. According
X-ray diffracted intensity (a.u.)
Fig. 1. Cross-sectional bright-field TEM micrographs of the (a) Al–Fe and (b) Fe–Al samples.
Fe-Al Fe Al Al-Fe 10
20
30
40
50
60
70
2θ (degrees) Fig. 2. GI-XRD spectra of the Fe, Al, Fe–Al and Al–Fe samples.
to the results, the lattice parameter of Fe in sample Al–Fe is 0.287(2) nm which is consistent with that of the bulk value and the coherent domains size is 1.6(2) nm. In the sample Fe–Al the not-identified bcc phase has a lattice parameter a ¼ 0:279ð2Þ nm and the average size of the coherent domains is
248
C. de Julian Fernandez et al. / Nucl. Instr. and Meth. in Phys. Res. B 216 (2004) 245–250
Fe Al-Fe Fe-Al
Fe
2
Al-Fe Fe-Al 1
0
2 1
-1
0 -1 -2 -0.10
-2 -6
attribute the reduced magnetic moment to a feature of the bcc Fe-based nanoparticles. Indeed, the presence of oxides or free atoms dispersed in the matrix should give, at 3 K and with high magnetic fields, as in our measurements, a much larger increase of the magnetization with the field. In the inset of Fig. 3 the low field region of the hysteresis loops has been represented. It is found that the coercive fields of the Fe–Al and Al–Fe samples are 3 and 1.4 mT, respectively. In first instance, the nanoparticles of all the samples must be single domains due to their nanometric size (this approximation will be discussed further). Then, considering a mechanism of coherent rotation of the magnetization, the coercive force is proportional to the magnetic anisotropy of the nanoparticles and inversely proportional to their magnetization saturation [14,15]. Considering the difference of magnetic moment per Fe atom in the Fe–Al and Al–Fe samples we conclude that the magnetic anisotropy of the bcc phase, presented in the Fe–Al sample, is similar or smaller than that of the Fe–Al. Fig. 4 shows the zero-field cooled (ZFC) and field cooled (FC) magnetization measurements of the Fe, Al–Fe and Fe–Al samples. This type of measurements is used to evidence the superparamagnetic behavior, that is, in nanograins, due to the size, the thermal effects help the magnetization in overcoming the energy barrier that fixes this in a
Magnetization (arb. units)
Magnetic moment per Fe atom (µB)
6.5(4) nm. Such a phase could be the Fe phase hardly stressed considering that it is contracted of 2.5% with respect to the lattice parameter measured in the Fe case. The observed pattern does not correspond with any Fe, Al oxides or silicate. We observe that in both Fe–Al and Al–Fe samples the TEM particle size is 5–10 times larger than the XRD coherent domains dimensions, which indicates that the clusters are polycrystalline. Finally, we point out that Fe–Al bulk bcc alloys have a larger lattice parameter than that of Fe. Therefore the formation of the Fe–Al alloy in the nanoparticles of the Al–Fe and Fe–Al samples was ruled out. The magnetic behavior of the samples was investigated by measuring the hysteresis loops at 3 K and applying a magnetic field up to 6 T. The diamagnetic contribution of the silica was subtracted from the measured magnetization. Then the magnetic moment per Fe atom was calculated considering the number of atoms in the sample from the RBS measurements. The results are presented in Fig. 3. The magnetic moments per Fe atom in the Fe implanted and Al–Fe samples at 6 T are similar, about 2.1 ± 0.2 lB , and slightly smaller than the bulk value of a-Fe, 2.2 lB . These results confirm the absence of Fe–Al intermetallic compound and indicate that a little oxidation occurred. However the magnetic moment in the Fe–Al sample is much smaller: 1.0 ± 0.2 lB . We
-5
-4
-3
-2
-1
0
1
-0.05
0.00
0.05
Applied Field (T)
2
3
4
0.10
5
6
Applied Field (T)
0
100
200
300
Temperature (K) Fig. 3. Hysteresis loops measured at 3 K of the Fe, Fe–Al and Al–Fe samples. Inset: detail at low field of the hysteresis loops of the Fe–Al and Al–Fe samples.
Fig. 4. Temperature dependence of the ZFC–FC magnetizations of the Fe, Fe–Al and Al–Fe samples.
C. de Julian Fernandez et al. / Nucl. Instr. and Meth. in Phys. Res. B 216 (2004) 245–250
direction of the grain and then the magnetization tilts freely. This energy barrier is responsible of the hysteresis that the nanoparticles show and it is proportional to the magnetic anisotropy and to the particle volume. The thermal effects produce the demagnetization of the particle: it seems as a paramagnet with high moment. The temperature threshold, called blocking temperature TB , above which a grain of volume V and with a magnetic anisotropy, Kef , shows the superparamagnetic behavior is [14,15] TB ¼
Kef V ; a kB
ð1Þ
where kB is the Boltzman constant and a is a constant equal to 25 in the case of the SQUID measurements. The ZFC–FC curves presented in Fig. 4 are representative of a distribution of TB , mainly associated to the broad size distribution, and being the temperature of the maximum of the ZFC curve linked to the average TB . In addition the magnetization does not follow a Curie law at high temperature in these films. The temperature at which ZFC–FC branches joints is the temperature above which all the nanoparticles are superparamagnetic. In the Fe and Fe–Al samples the blocking temperature is below the room temperature while for the Al–Fe sample TB is above the room temperature. Considering the much larger particle size of the Al–Fe than the Fe–Al sample, the differences in the ZFC–FC curves can be discussed in terms of superparamagnetism considering the measured particle size. However we point out that the nanoparticles of these samples have a core–shell morphology. The magnetic anisotropy will include the contribution of the different magnetic surfaces (that of the core and those of the shell) as occurs in many nanostructures [15]. The core–shell structure promotes incoherent rotation modes in which the magnetic anisotropy is not related to the intrinsic properties of the material. Moreover, as can be observed in Figs. 1(a) and (b) nanoparticles are very densely packed so that interparticle interactions must be dominant giving rise to different demagnetization processes. Therefore, we think that the observed behavior appearing similar to superparamagnetism is the result of several magnetic effects and must be
249
investigated in order to analyse correctly the ZFC– FC curves. Finally we want to point out that, both from the point of view of the nanostructural and magnetic characterization, FeAl intermetallic alloy nanoparticles were not formed in silica by using sequential ion implantation. Several authors have studied the mechanism of nucleation and alloying of the nanoparticles concluding that it is driven by competitive physical and chemical reactions between the implanted elements and those of the matrix [16,17]. For Fe and Al the respective Gibbs energy of oxide formation are negative and very high [18], so the formation of oxides is expected. However, as it was shown in this work, metallic nanoparticles were formed suggesting that competitive chemical paths determine the nanostructural features as elsewhere suggested [16]. Several studies are in progress to investigate the mechanism of formation of the different nanostructures and their chemical state, the identification of the bcc phase observed in the Fe–Al sample and the characterization of the magnetic properties.
Acknowledgements This work was financially supported by ENEA in the framework of the PROTEMA project of ‘‘Intesa ENEA-MIUR’’ under project 4335/04 and by the Italian MURST (National Projects).
References [1] F. Gonella, P. Mazzoldi, in: H.S. Nalwa (Ed.), Handbook of Nanostructured Materials and Nanotechnology, Vol. 4, Academic Press, San Diego, 2000, p. 81. [2] A. Meldrum, L.A. Boatner, C.W. White, Nucl. Instr. and Meth. B 178 (2001) 7. [3] G. Mattei, Nucl. Instr. and Meth. B 191 (2002) 323. [4] G. Battaglin, E. Cattaruzza, G. De Marchi, F. Gonella, G. Mattei, C. Maurizio, P. Mazzoldi, M. Parolin, C. Sada, I. Calliari, Nucl. Instr. and Meth. B 191 (2002) 392. [5] C. de Julian Fernandez, C. Sangregorio, G. Mattei, C. Maurizio, G. Battaglin, F. Gonella, S.A. Lascialfari, S. Lo Russo, D. Gatteschi, P. Mazzoldi, J.M. Gonzalez, F. DÕAcapito, Nucl. Instr. and Meth. B 175–177 (2001) 468. [6] E. Cattaruzza, F. dÕAcapito, C. de Julian Fernandez, A. de Lorenzo, F. Gonella, G. Mattei, C. Maurizio, P. Mazzoldi,
250
[7]
[8]
[9]
[10]
C. de Julian Fernandez et al. / Nucl. Instr. and Meth. in Phys. Res. B 216 (2004) 245–250 S. Padovani, B.F. Scremin, F. Zontone, Nucl. Instr. and Meth. B 191 (2002) 40. C.W. White, S.P. Withrow, K.D. Sorge, A. Meldrum, J.D. Budai, J.R. Thompson, L.A. Boatner, J. Appl. Phys. 93 (2003) 5656. E. Borsella, C. de Julian Fernandez, M.A. Garcıa, G. Mattei, C. Maurizio, P. Mazzoldi, S. Padovani, C. Sada, G. Battaglin, E. Cattaruzza, F. Gonella, A. Quaranta, F. DÕAcapito, M.A. Tagliente, L. Tapfer, Nucl. Instr. and Meth. B 191 (2002) 447. H.P.J. Wijn (Ed.), Physical Properties and Other Data, Landolt-B€ ornstein, Vol. III, Part 19b, Springer, Berlin, 1987, p. 295. J.S. Kouvel, J.H. Westbrook, R.L. Fleischer (Eds.), Intermetallic Compounds: Principles, Vol. 1, Wiley, New York, 1994, p. 935.
[11] G.R. Caskey, J.M. Franz, D.J. Sellmyer, J. Phys. Chem. Solids 34 (1973) 1179. [12] A. Hernando, X. Amils, J. Nogues, S. Suri~ nach, M.D. Bar o, M.K. Ibarra, Phys. Rev. B 58 (1998) R11864. [13] B.V. Reddy, S.C. Deevi, F.A. Reuse, S.N. Khanna, Phys. Rev. B 64 (2001) 132408. [14] J.L. Dormann, D. Fiorani, E. Tronc, Adv. Chem. Phys. 98 (1997) 283. [15] X. Batlle, A. Labarta, J. Phys. D: Appl. Phys. 35 (2002) R15. [16] E. Cattaruzza, Nucl. Instr. and Meth. B 169 (2000) 141. [17] T. Ito, O. Kitakami, Y. Shimada, Y. Kamo, S. Kikuchi, J. Magn. Magn. Mater. 235 (2000) 165. [18] O. Knacke, O. Kubaschewski, K. Hesselmann (Eds.), Thermo-Chemical Properties of Inorganic Substances, Springer, Berlin, 1991.
