ARTICLE IN PRESS Journal of Magnetism and Magnetic Materials 322 (2010) 1300–1303
Contents lists available at ScienceDirect
Journal of Magnetism and Magnetic Materials journal homepage: www.elsevier.com/locate/jmmm
Control of crystalline phases in magnetic Fe nanoparticles inserted inside a matrix of porous carbon M.P. Ferna´ndez a,, D.S. Schmool b,c, A.S. Silva c, M. Sevilla d, A.B. Fuertes d, P. Gorria a, J.A. Blanco a a
Dpto. de Fı´sica, Universidad de Oviedo, Calvo Sotelo, s/n, 33007 Oviedo, Spain IN-IFIMUP, Universidade do Porto, Rua do Campo Alegre 687, 4169-007 Porto, Portugal c Dpto. de Fı´sica, Universidade do Porto, Rua do Campo Alegre 687, 4440-661 Porto, Portugal d Instituto Nacional del Carbon (CSIC), Apartado 73, 33080 Oviedo, Spain b
a r t i c l e in f o
a b s t r a c t
Available online 3 May 2009
Two magnetic composites made up of Fe nanoparticles (Fe-NPs) embedded in a porous amorphous carbon matrix are presented. One of the samples, Fe-S-AC, was obtained with the aid of sucrose and the other, Fe-AC, in the absence of this substance. The XRD patterns show Bragg diffraction peaks associated with a-Fe and g-Fe crystalline phases in the Fe-AC sample, while only peaks corresponding to the a-Fe phase are observed for Fe-S-AC powders. The Fe-NPs exhibit broad particle-size distributions for both samples, 5–50 nm for Fe-AC, whereas two populations (2–8 and 10–70 nm) for the Fe-S-AC composite are found. This fact gives rise to poorly defined blocking temperatures, as it can be deduced from the broad maxima observed in MZFC(T) variations. In addition, M(H) curves for both Fe-AC and Fe-S-AC samples reveal the existence of exchange-bias effect for To60 K, probably due to a magnetic coupling within a core/shell structure of the Fe-NPs, although this effect was observed to be less significant for Fe-S-AC. & 2009 Elsevier B.V. All rights reserved.
Keywords: Nanoparticle Exchange-bias Superparamagnetism Activated carbon
1. Introduction
2. Experimental
It is well established that when the size of magnetic particles is drastically reduced down to the nanometer length scale, independently of the shapes or forms of the samples, surface and interface effects become relevant, and superparamagnetic (SPM) effects come into view [1–3]. Moreover, the presence, morphology and composition of a non-magnetic matrix surrounding the nanoparticles (NPs) play an important role in those aspects related to the magnetic interactions between NPs [4]. Furthermore, these systems could display novel physical-chemical properties resulting in its suitability to be used in a large number of technological applications. However, the experimental techniques required for precisely determining the interfacial behaviour at the atomic scale have been recently developed. Hence, an intense research effort, focused in the accurate characterization of these NP-systems, is under way to completely understand the new emerging phenomena [5]. In this article, we report on the correlation between magnetic behaviour and morphology of two composites in which Fe nanoparticles (Fe-NPs) have been randomly dispersed in an amorphous porous carbonaceous matrix.
The carbon matrix employed for the insertion of Fe-NPs is a commercial activated carbon (AC) (M30; Osaka Gas, Japan) [6]. Two types of Fe-carbon composites, denoted as Fe-AC and Fe–S–AC, were synthesized. For the preparation of the Fe-AC sample, the AC was impregnated with a solution of iron nitrate in ethanol, dried at 80 1C and then heat treated under N2 up to 900 1C (heating rate: 3 1C/min) for 3 h. In the case of the Fe-S-AC composite, the synthesis took place in two steps: (i) the AC was impregnated with iron nitrate and heat treated at 900 1C, (ii) the sample obtained from the previous step was impregnated with a solution of sucrose in water, dried at 120 1C and heat treated under N2 at 600 1C for 2 h. The percentages of Fe in both samples were determined by means of thermogravimetric analysis (TGA) in air up to 700 1C, being 16.8% for the Fe-AC sample and 8.8% for the FeS-AC sample. The introduction of sucrose was previously found to be useful to protect Ni NPs against acid environments [7,8]. The textural properties of the AC sample and the composites (Fe-AC and Fe-S-AC) determined by nitrogen physisorption (ASAP 2010) are listed in Table 1. The crystal structure of Fe-NPs was analysed by means of room temperature X-ray powder diffraction (XRD) using Cu Ka radiation (l ¼ 1.5418 A˚) in the 2y range 401–1601. Transmission electron microscopy (TEM) was used to study the morphology of the samples and to estimate the NP size distributions. Magnetization
Corresponding author. Tel.: +34 985105014; fax: +34 985103324.
E-mail address:
[email protected] (M.P. Ferna´ndez). 0304-8853/$ - see front matter & 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.jmmm.2009.04.058
ARTICLE IN PRESS ´ndez et al. / Journal of Magnetism and Magnetic Materials 322 (2010) 1300–1303 M.P. Ferna
2350 750 360
1.5 0.6 0.4
2.5 2.5 2.5
5
Intensity (x 10 3 counts)
50nm
Fe-S-AC
0
15 10 5
Fe-AC
0 40
60
80 100 120 angle, 2θ θ (deg)
140
160
Fig. 1. Observed (points) and calculated (solid line) room temperature X-ray powder diffraction patterns of Fe-S-AC and Fe-AC composites. Positions of Bragg reflections are represented by vertical bars. The observed-calculated difference pattern is depicted at the bottom of each figure. In Fe-S-AC, the first series of vertical marks corresponds to a-Fe. In Fe-AC, the first series of vertical marks is associated with g-Fe and the second with a-Fe. Insets (a) and (c) show the detail in 2y-range around 451 to clearly appreciate the highest intensity Bragg reflections of the spectra. Insets (b) and (d) show TEM images of Fe–S–AC and Fe-AC, respectively.(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
vs. temperature, M(T), curves in zero-field-cooling (ZFC) and fieldcooling (FC) regimes, and under an applied magnetic field, m0H ¼ 1 mT, were measured using a SQUID magnetometer in the temperature range from 5 to 370 K. Magnetization vs. magnetic fields, M(H) cycles, were measured up to m0H ¼ 72 T at several selected temperatures. Low temperature M(H) cycles were performed following two different procedures: cooling from 300 K down to 10 K under m0Hcool ¼ 2 T (M(H)@FC) and m0Hcool ¼ 0 T (M(H)@ZFC).
1
0.2
0
T=10 K -0.2
μ H(T) 0
-0.05
0
0.05
0 0.2
2T
AC Fe-AC Fe-S-AC
M/M
Maximum of the pore size distribution (nm)
2T
Pore volume (cm3 g1)
M/M
BET surface area (m2 g1)
2T
Sample
substance). The whole profile fit of the diffraction patterns (following the procedure explained in [11] for nanostructured systems) has been carried out by considering the same background contribution. The diffraction peaks displayed by the Fe-AC sample were indexed as the Bragg reflections of a bodycentred cubic (bcc) and a face-centred cubic (fcc) crystal structures with lattice parameters a ¼ 2.867(1) and 3.587(1) A˚, respectively (see inset c in Fig. 1). The former corresponds to an a-Fe phase while the latter can be identified as a g-Fe phase, with relative percentages around 45(5)% (a) and 55(5)% (g). However, the XRD pattern of Fe-S-AC composite suggests that only the a-Fe phase is present, with the same value for the lattice parameter, although we cannot completely discard the existence of traces (o5%) of g-Fe phase (see Fig. 1a). The average diameter value of the Fe-NPs in both composites has been estimated after the evaluation of several TEM images, from which a large number of NPs (41500) have been counted. Representative TEM images are shown in the insets (b) and (d) of Fig. 1. The analysis of TEM images reveals that the Fe–AC sample contains a unique broad size distribution of Fe-NPs that follows rather well a log-normal function with an average NP size /tS ¼ 15.4 nm, and a standard deviation s ¼ 6 nm, as it has been previously reported [12]. On the other hand, the analysis of TEM images of Fe–S–AC sample clearly shows the existence of Fe-NPs with two different particle–size ranges: a sharp distribution (2–8 nm) with an average NP size /t(s)S ¼ 3(1) nm and another broad size distribution ranging from 10 to 70 nm and centred around /t(s)S ¼ 38(22) nm. In Fig. 2 M(H)@ZFC of Fe-AC and Fe-S-AC composites recorded at 10 K are displayed. To better compare the magnetic field dependence of the magnetization in both samples, the values of M have been normalized to those measured at m0H ¼ 2 T, M/M2T. From the saturation magnetization, Ms, at T ¼ 10 K, we estimate a value of MsE200 Am2 kg1, for a-Fe phase in both Fe-AC and Fe-SAC samples, taking into account: (i) the different percentages of Fe
Normalized magnetization, M / M
Table 1 Textural properties of activated carbon and the synthesized composites.
1301
Fe-AC Fe-S-AC
0
-0.2
μ H (T) 0
3. Results and discussion Fig. 1 shows the room temperature XRD patterns corresponding to the Fe–S–AC (upper panel) and Fe–AC (bottom panel). The Rietveld refinement of the patterns has been performed using the Fullprof package [9]. Both XRD patterns exhibit broad diffraction peaks together with a large background contribution coming from the disordered and complex AC matrix [10] (note that both composites contain more than 80% of that
-0.05
-1 -2
-1
0 1 Magnetic field, μ H(T)
0
0.05
2
0
Fig. 2. M(H)@ZFC measured at 10 K for Fe-AC (red full circles) and Fe-S-AC (blue open circles) samples. M2T is the value of the magnetization for m0H ¼ 2 T. Insets (a) and (b) display enlarged views of the M(H)@ZFC and M(H)@FC around M/ M2T ¼ 0, respectively (see text).(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
ARTICLE IN PRESS ´ndez et al. / Journal of Magnetism and Magnetic Materials 322 (2010) 1300–1303 M.P. Ferna
40
Fe-S-AC
30
0
c
Coercive field, μ H (mT)
Fe-AC
20 1390(200) K 10 195(10) K
0
0
10
30
20
T
1/2
(K
1/2
40
0.12 FC 0.08 -1
ZFC
Fe-AC
Tmax
2
determined from TGA in both samples; (ii) the relative percentages between a-Fe and g-Fe phases obtained from room temperature XRD and (iii) a g-Fe phase [12] not contributing to the net saturation magnetization. This value is in good agreement with that given by Lacroix et al. for Fe-NPs of 2 nm [13], and close to MsE221 Am2 kg1 for pure a-Fe [14]. The observed reduction in the value of the saturation magnetization respect to that of pure a-Fe could be attributed to the possible coexistence of topological disorder and frustration of magnetic spins at the NP-surface giving rise to spin-canting and/or spin-glass-like behaviour, as pointed out by Fiorani et al. in similar systems [15]. No significant differences in moHc [E35(2) mT] and Mr (E0.1 Ms) values are observed in M(H)@ZFC cycles between Fe-AC and Fe-S-AC (see Fig. 2a). It is worth noting that the value of moHc(10 K) for both samples is similar to those previously reported in other Fe-NP systems [13]. On the other hand, a small shift of M(H) to negative magnetic field values is observed when measuring M(H)@FC (see Fig. 2b). This effect is larger in the case of the Fe-AC sample [m0Hex(10 K) E8(1) mT] than in the Fe-S-AC one [m0Hex(10 K)E4(1) mT]. In addition, the Mr value of Fe-AC at 10 K (see Fig. 2b) is around 40% larger than that of Fe-S-AC obtained from M(H)@FC, while similar values of Mr were observed in M(H)@ZFC and M(H)@FC for Fe-S-AC sample. However, the origin of this exchange-bias-like effect [16] and the reason for the different values measured between both samples need further investigation. The evolution of the coercive field, moHc, with temperature has been followed through the measurement of M(H) at several selected temperatures (see Fig. 3). For the Fe-S-AC composite, moHc diminishes as the temperature is increased showing an almost linear T1/2-dependence. From the fit of the moHc vs. T1/2 curve, a rather high blocking temperature, TBE1390(2 0 0) K, can be estimated. The temperature dependence of the coercive field for the Fe–AC sample shows clear differences, namely, a maximum value of around 38(2) mT at T ¼ 25 K, with fairly smaller values at lower temperatures. For T425 K, moHc also diminishes as the temperature is increased, however, it is difficult to found a temperature interval where the moHc vs. T1/2 variation
Magnetization (Am kg )
1302
0.04 μ H = 1 mT 0
0.06
FC
ZFC
0.04
Tmax1 Fe-S-AC
0.02 0
100 200 300 Temperature (K)
400
Fig. 4. M(T) curves in the ZFC-FC regime under an applied field of 1 mT for (a) FeAC (red full circles) and (b) Fe-S-AC samples (blue full circles). The arrows are pointing the temperatures of the MZFC(T) maxima (see text).(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
could be considered linear (see Fig. 3). Nevertheless, we have estimated a blocking temperature of TBE195(10) K from the linear fit shown in Fig. 3. In addition, the M(H) cycles measured at 300 K still exhibit some hyteresis [moHc20(1) and 10(1) mT for Fe-S-AC and Fe-AC, respectively], indicating that a significant fraction of the Fe-NPs are still magnetically blocked at this temperature. Fig. 4 represents the low-field magnetization vs. temperature curves measured in ZFC, MZFC(T), and FC, MFC(T), regimes. The MZFC(T) and MFC(T) curves of both samples, do not overlap in the temperature range measured. This behaviour together with the existence of hysteresis in room temperature M(H) cycles, evidence that the NPs of the largest sizes remain blocked above room temperature. In Fig. 4a, the MZFC(T) curve of Fe-AC slowly increases below T60 K, as a consequence of the existence of magnetically blocked Fe-NPs. In contrast, a broad maximum in MZFC(T) is observed at higher temperatures, Tmax170(40) K, related to the broad distribution of Fe-NPs. On the other hand, the MZFC(T) of Fe-S-AC (see Fig. 4b) exhibits a low temperature maximum [Tmax130(3) K] which is associated with the NPs of small size [3(1) nm]. In addition, MZFC in Fe-S-AC showing a rounding-up trend observed above 30 K seems to suggest the existence of another maximum at higher temperatures (Tmax24370 K), that is related to the NPs of largest size [38(22) nm]. The different tendency observed in both samples for MFC(T) curves at low temperatures (see Fig. 4) could be related to the different type of magnetic interactions between Fe-NPs in Fe-AC and Fe-S-AC composites.
4. Summary
)
Fig. 3. T1/2-dependence of the coercive field, moHc, obtained from M(H)@ZFC. The solid line shows the linear fit of m0Hc to a T1/2-law (see text) providing TB ¼ 195(10) and 1390(2 0 0) K for Fe-AC (red full circles) and Fe-S-AC (blue open squares) samples, respectively.(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
In this work a structural and magnetic study of two composites that contain Fe nanoparticles embedded in a porous amorphous carbon matrix is presented. One of the samples (Fe–S–AC) was obtained with the aid of sucrose, whereas the other one (Fe–AC) was not. The post-synthesis with sucrose seems to be responsible
ARTICLE IN PRESS ´ndez et al. / Journal of Magnetism and Magnetic Materials 322 (2010) 1300–1303 M.P. Ferna
for morphology changes and for the control of crystalline phases in these composites. As a consequence of these features, different magnetic responses were found in the M(T) curves at low applied magnetic fields. Further work will aim to explore the magnetic coupling of the core/shell structure in these composites. Fe-NPs inserted in activated carbons could be manipulated by means of low external magnetic fields, thanks to the low coercivity values. This fact together with the relative high saturation magnetization values, make these composites attractive candidates for applications as magnetically separable adsorbents to be produced on a large scale and at low cost.
Acknowledgements Financial support from FEDER and the Spanish MICINN (MAT2008-06542-C04-03 and MAT2008-00407) is acknowledged. One of us, M.P.F., thanks MICINN for the award of a FPI Grant cofinanced by the European Social Fund. SCT’s at University of Oviedo, as well as University of Porto and IFIMUP are also acknowledged for the assistance and experimental equipment provided.
1303
References [1] E. Wolf, Nanophysics and Nanotechnology, An Introduction to Modern Concepts in Nanoscience, Wiley, New York, 2004. [2] X. Battle, A. Labarta, J. Phys. D: Appl. Phys. 35 (2002) R15. [3] S. Bedanta, W. Kleemann, J. Phys. D: Appl. Phys. 42 (2009) 013001. [4] J.L. Dorman, D. Fiorani, Magnetic Properties of Fine Particles, North-Holland, Amsterdam, 1992. [5] D.L. Huber, Small 5 (2005) 482. [6] A.B. Fuertes, P. Tartaj, Chem. Mater. 18 (2006) 1675. [7] P. Gorria, M. Sevilla, J.A. Blanco, A.B. Fuertes, Carbon 44 (2006) 1954. [8] P. Gorria, M.P. Ferna´ndez, M. Sevilla, J.A. Blanco, A.B. Fuertes, Phys. Status Solidi (RRL) 3 (2009) 4. [9] J. Rodrı´guez Carvajal, Physica B 192 (1993) 55. [10] P.J.F. Harris, Z. Liu, K. Suenaga, J. Phys.: Condens. Matter 20 (2008) 362201. [11] D. Martı´nez-Blanco, P. Gorria, J.A. Blanco, M.J. Pe´rez, J. Campo, J. Phys.: Condens. Matter 20 (2008) 335213. [12] M.P. Ferna´ndez, D.S. Schmool, A.S. Silva, M. Sevilla, A.B. Fuertes, P. Gorria, J.A. Blanco, J. Non-Crystalline Solids 354 (2008) 5219. [13] L.-M. Lacroix, S. Lachaize, A. Falqui, T. Blon, J. Carrey, M. Respaud, F. Dumestre, C. Amiens, O. Margeat, B. Chaudret, P. Lecante, E. Snoeck, J. Appl. Phys. 103 (2008) 07D521. [14] D. Craik, Magnetism, Principles and applications, Wiley, New York, 1995. [15] D. Fiorani, L. Del Bianco, A.M. Testa, K.N. Trohidou, J. Phys.: Condens. Matter 19 (2007) 225007. ˜ ach, J.S. Mun ˜ oz, M.D. Baro´, [16] J. Nogue´s, J. Sort, V. Langlais, V. Skumryev, S. Surin Phys. Rep. 422 (2005) 65.
