Nickel nanoparticles deposited into an activated porous carbon: synthesis, mi-‐ crostructure and magnetic properties Pedro Gorria*, 1, María Paz Fernández-‐García1, Marta Sevilla2, Jesús A. Blanco1 and Antonio B. Fuertes2 1 2
Departamento de Física, Universidad de Oviedo, Calvo Sotelo, s/n, 33007 Oviedo, Spain Instituto Nacional del Carbón (CSIC), P.O. Box 73, 33080 Oviedo, Spain
phys. stat. sol. (RRL), (2008) / DOI 10.1002/pssr.200802216 PACS 75.20.-g, 75.50.Tt, 81.05.Rm *
Corresponding author: e-mail
[email protected], Phone: +34 985 102 899, Fax: +34 985 103 324
Activated carbons (AC) are commonly used as efficient adsorbents to remove contaminants. The incorporation of a magnetic material into the AC could greatly enhance its manipulation through magnetic separation. However, the composite material will need to have sufficiently saturated magnetization, and as low as possible coercivity to be easily attracted by commercial permanent magnets. In this letter we report on the correlation between microstructure and magnetic behaviour of Ni nanoparticles (NPs) embedded in an amor-
phous activated porous carbon (Ni-AC). The Ni-AC powders have been synthesized by means of an easy and low-cost procedure. The addition of sucrose during the preparation process provides effective protection in acid media. This Ni-AC composite has a microstructure composed of crystalline NPs with diameters in the range 7 – 25 nm, and exhibits superparamagnetic behaviour at room temperature, with saturation magnetization values around 3 Am2kg-1 under applied magnetic fields of 200 mT.
1 Introduction Activated carbons (AC) are being extensively used for adsorption and catalytic purposes, mainly due to their outstanding efficiency together with a wide availability and low cost [1]. Frequently, they are exploited in liquid phase for a number of applications, such as, catalyst or catalytic supports, to remove contaminants or for recovering specific products. When the selective manipulation of valuable substances associated with AC is pursued, magnetic separation could be the most effective strategy for achieving this task [2, 3]. However, the commonly used magnetic adsorbents have poor porous characteristics, and the conventional synthesis procedures are too complex and expensive for large production compared with those of AC porous materials [4]. Two important conditions must be fulfilled by a composite made by an AC and magnetic nanoparticles (NP), namely, enough saturation magnetization and low coercivity values (or even zero, in the case of superparamagnetic, SPM, behaviour, which is the optimal situation) in order to be easily manipulated using conventional permanent magnets [5]. In this letter, we present a new and easy-to-follow synthesis procedure to prepare magnetically separable porous
carbons. We will also discuss the correlation between microstructure, morphology and magnetic response of these Ni-AC powders. 2 Experimental The starting material is a commercial activated carbon (M30, Osaka Gas) with a high pore volume (1.47 cm3 g-1) and BET surface area (2350 m2 g-1), which after being impregnated with an aqueous solution containing sucrose and nickel nitrate, and a subsequent heat-treatment (at 600 ºC under nitrogen for 3 hours), results in the formation of nickel nanoparticles dispersed along the porous AC matrix. The addition of sucrose allows the formation of a carbon layer surrounding the nickel nanoparticles and protecting them against acid corrosion. The final amount of Ni in the sample is around 16 wt. %, and it decreases to 12 wt.% after washing with HCl (0.1 mol l-1) for 7 days. More details about sample preparation can be found elsewhere [5-7]. Room temperature x-ray powder diffraction (XRD) has been used to check the crystalline structure of the powders, as well as for the estimation of the mean crystalline size through the analysis of the diffraction peak-broadening [8].
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58
The distribution of NP sizes has been obtained from transmission electron microscopy images. The magnetic properties of the Ni-AC sample have been investigated using a vibrating sample magnetometer, through the measure of both the temperature (from 2 K to 300 K) and applied magnetic field (up to ±1.5 T) dependencies of magnetization.
[10]: TB = KeffV/25kB
(1) 4
-3
Assuming a value of Keff = 0.5 × 10 J m for Ni [10] and a NP mean diameter of 16 nm (TEM), we can estimate a value of TB ≈ 31 K.
3 Results and Discussion The room temperature XRD pattern (not shown) can be indexed using the Bragg reflections of a face-centred cubic crystal structure with a lattice parameter (a = 3.525 Å) close to that of bulk Nickel (a = 3.52 Å). From the diffraction peak broadening we have estimated a value of XRD = 18(3) nm for the mean crystalline size. Transmission electron microscopy (TEM) images show quasi-spherical Ni NPs uniformly dispersed throughout the AC matrix (see Fig. 1).
Figure 2 Magnetization vs µ0H/T curves showing that M follows a universal f(µ0H/T) law for T > 200 K. The inset shows the normalized magnetization vs applied magnetic field measured at T = 300 K, the solid line is the fit to the Langevin function.
Figure 1 Histogram of the NP diameters together with the fit (solid line) to a log-normal function. The inset shows a typical TEM image of the Ni-AC powders.
The histogram corresponding to the distribution of NP diameters (see Fig. 1) follows rather well a log-normal function with a mean diameter value of TEM = 16(1) nm and a standard deviation of σ = 4(1) nm. However, it is worth noting that TEM images show a rather broad distribution of NP diameter values (7 - 25 nm). If we consider single-domain NPs, due to their small size, then, SPM behaviour above certain temperature can be expected [9]. In Figure 2, we can observe that the magnetization of the Ni-AC powders follows a universal f(µ0H/T) law for T > 200 K. The latter together with the observed reversibility of the M(H) curves (absence of magnetic hysteresis) at T = 300 K (see Fig. 2, inset) indicate that the Ni-AC powders are in SPM regime at room temperature. From the fit of the M(H) curve to a Langevin function (solid line in Fig. 2 inset), values of 13(1) nm and 30(2) A m2 kg-1 for the mean NP-diameter and the saturation magnetization, Ms, respectively, have been estimated. Moreover, there will be a characteristic temperature (the blocking temperature), TB, below which the NPs do not behave as SPM and the effective magnetic anisotropy constant, Keff, the NP volume, V, and the TB, are related by
It is well-known that for T < TB the temperature dependence of the coercivity, µ0Hc, for fine particles of equal-size is proportional to 1 - (T/TB)1/2 [10]. For this reason, we have measured M(H) curves at 21 different temperatures between 2 and 300 K in order to estimate the value for TB from the µ0Hc vs. T dependence (see Fig. 3).
Figure 3 Temperature dependence of the coercivity, µ0Hc, the inset shows the plot of µ0Hc vs. T1/2 (see text).
If we plot µ0Hc vs. T1/2 (see inset in Fig. 3), an almost linear dependence is observed for T < 65 K, and the fit gives a value of TB = 62(2) K. However, this value for TB doubles that previously estimated from Eq. (1). For T > 65 K the coercivity does not vanish due to the existence of a distribution of NP-sizes, and then a unique blocking tem-
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58
perature can not be considered. On the other hand, the zero-field-cooling (ZFC) magnetization vs temperature curve, MZFC(T), under low applied magnetic fields, for a system of equal-size NPs is expected to display a clear maximum at T = TB. However, the MZFC(T) curve for the Ni-AC sample, measured under µ0H = 2.5 mT applied magnetic field, exhibits a broad maximum as a consequence of the distribution of NP sizes (see Fig. 4). Moreover, the maximum occurs at a temperature well above 31 K and close to the value for TB estimated from the temperature dependence of the coercivity. Making use of the Wohlfarth model [11], we have performed a fit (solid line in figure 4) of the MZFC(T) curve (see [12-14] for details), and we obtain values of about 24 A m2 kg-1 and 104 J m3 for Ms and Keff, respectively. The value of Ms is significantly lower than that of pure Ni (57.5 A m2 kg-1, [10]), but similar to previously reported values in other Ni NP systems [15]. However, it must be pointed out that the effective magnetic anisotropy is double to that of pure Ni [10]. Therefore, if Keff = 104 J m3 is used in Eq. (1), we find a value of TB ≈ 62 K in excellent agreement with that estimated form the temperature dependence of the coercivity.
Finally, we have to mention that when magnetic measurements are repeated after 36 months, the same results have been obtained, indicating that the Ni-AC powders a both structural and chemically stable, which is a condition of primary importance in order to use these composites for application purposes. 4 Summary and conclusions A powder composite formed by Ni nanoparticles embedded in an activated porous carbon matrix has been synthesized using an easy and low-cost procedure. The sample displays superparamagnetic behaviour at room temperature. A good correlation between microstrural and magnetic measurements has been pointed out: the Ni-NPs saturation magnetization is roughly one half of that of pure Ni, while the effective magnetic anisotropy is double. The relatively high values for the absolute saturation magnetization (3 A m2 kg-1) together with the absence of magnetic hysteresis at room temperature make these powders well suited for its use as magnetic separable adsorbents, due to its simple manipulation by means of low applied magnetic fields (200 mT), that can be easily attained using conventional permanent magnets. The addition of sucrose provides protection against acid attack. Moreover, the Ni-AC composite exhibits high chemical and structural stability with the absence of aging effects and deterioration of its magnetic properties. Acknowledgements The financial support from FEDER and the Spanish MICINN (former MEC) (MAT2008-00407, MAT2005-06806-C04-01, MAT2008-06542-C04-C03 & NAN2004-09203-C04-03) is gratefully acknowledged.
References
Figure 4 ZFC magnetization vs temperature curve (squares) measured under µ0H = 2.5 mT applied magnetic field. The solid line represents the fit using the Wohlfarth model [11].
It is worth noting that due to the broad spread of NP diameter values (7 – 25 nm), a distribution of blocking temperatures must be expected as well. Hence, we can assume, using Eq. (1), that the smaller NPs enter to SPM regime at T > 5 K (TB ≈ 5 K for 7 nm NPs), and explains the rapid increase of the magnetization at low temperatures (see Fig. 4). In the same way, the non-zero coercivity above 65 K can be understood taking into account that the largest NPs are blocked up to T ≈ 200 K (TB ≈ 237 K for 25 nm NPs). In addition, we have estimated from M(T,H) measurements the magneto-caloric effect across the SPM regime in Ni-AC. These estimations lead to magnetic entropy changes, in the immediacy of TB, two orders of magnitude lower (< 100 mJ kg-1 K-1) than those found in promising candidates for magnetic refrigeration, such as nanostructured Pr2Fe17 [16].
[1] R. C. Bansal, J. Donnet and F. Stoeckli, Active Carbon (Marcel Dekker, New York, 1985), chapter 6. [2] L. Oliveira et al., Carbon, 40, 2177 (2002). [3] D. -W. Wang et al., Carbon, 46, 1593 (2008). [4] A. Lu et al., Angew. Chem. Int. Ed., 43, 4303 (2004). [5] P. Gorria et al., Carbon, 44, 1954 (2006). [6] A. B. Fuertes and P. Tartaj, Chem. Matter, 18, 1675 (2006). [7] A. B. Fuertes and P. Tartaj, Small 3, 275 (2006). [8] D. Martínez-Blanco et al., J. Phys.: Condens. Matter, 20, 335213 (2008). [9] J. L. Dormann and D. Fioranni, Magnetic Properties of Fine Particles (North-Holland, Amsterdam, 1992). [10] B. D. Cullity, Introduction to Magnetic Materials (AddisonWesley, Reading, MA, 1972). [11] E. P. Wohlfarth, Phys. Lett. A 70, 489 (1979). [12] F. Bødker et al., Phys. Rev. B, 61, 6826 (2000). [13] R. García Calderón et al., J. Magn. Magn. Mater., 254-255, 79 (2003). [14] L. Fernández Barquín et al., Phys. Rev. B, 76, 172404 (2007). [15] C. de Julián Fernández, Phys. Rev. B, 72, 054438 (2005). [16] P. Gorria et al., J. Phys. D: Appl. Phys., 41, 192003 (2008).
