Fct Feptag
This document was uploaded by user and they confirmed that they have the permission to share it. If you are author or own the copyright of this book, please report to us by using this DMCA report form. Report DMCA
Overview
Download & View Fct Feptag as PDF for free.
More details
- Words: 3,642
- Pages: 8
ARTICLE IN PRESS
Journal of Magnetism and Magnetic Materials 266 (2003) 49–56
Magnetic properties of self-organized L10 FePtAg nanoparticle arrays S. Wanga, S.S. Kanga, D.E. Niklesa,b, J.W. Harrella,*, X.W. Wuc a
Department of Physics and Astronomy, Center for Materials for Information Technology, University of Alabama, Tuscaloosa, AL 3548, USA b Department of Chemistry, Center for Materials for Information Technology, The University of Alabama, Tuscaloosa, AL 35487, USA c Seagate Research, 2403 Sidney Street, Pittsburgh, PA 15203, USA Received 30 September 2002; received in revised form 16 December 2002
Abstract The magnetic properties of chemically synthesized high anisotropy L10 [Fe49Pt51]88Ag12 nanoparticle arrays have been studied as a function of annealing temperature. Particles were prepared by the simultaneous polyol reduction of platinum acetylacetonate and silver acetate and the thermal decomposition of iron carbonyl, yielding monodispersed particles of diameter B3.5 nm. Addition of Ag lowers the ordering temperature of self-assembled arrays by B150 C. After annealing at Ta ¼ 500 C for 30 min in an Ar/H2 atmosphere, the coercivity was 13,800 Oe. TEM and delta-M measurements indicate weak particle aggregation up to Ta ¼ 400 C, with evidence of sintering at higher temperatures. Large ratios of remanent to hysteresis coercivity indicate a large distribution in anisotropy energies. Anomalously large thermal stability constants, KV =kB T; and switching volumes were measured, even in samples with very little evidence of sintering. Zero field viscosity versus remanence curves show evidence of exchange interactions. r 2003 Elsevier B.V. All rights reserved. PACS: 75.75.+a; 75.50.Bb; 75.50.Ss; 75.50.Tt; 75.50.Vv Keywords: FePt nanoparticles; L10 phase; Chemical ordering; Magnetic recording materials; Self-assembly; Magnetic viscosity
1. Introduction There has been considerable interest in ordered arrays of high-anisotropy magnetic nanoparticles for ultra-high density recording media since a recent report by Sun and co-workers on the synthesis and self-assembly of L10 FePt nanoparticles [1]. FePt nanoparticles with a disordered *Corresponding author. Tel.: +1-205-348-9404; fax: +1205-348-2346. E-mail address: [email protected] (J.W. Harrell).
FCC structure and with an organic coating were produced by chemical synthesis. Because the particles had a narrow size distribution, a nonmagnetic coating, and were superparamagnetic, self-assembled arrays could be produced by slow evaporation of a dispersion of the particles on a substrate. In order to obtain a high degree of L10 ordering, the arrays had to be annealed at temperatures of 550 C or greater. These high annealing temperatures, however, degrade the quality of the arrays and would be undesirable for processing of media. At high temperatures, the
0304-8853/03/$ - see front matter r 2003 Elsevier B.V. All rights reserved. doi:10.1016/S0304-8853(03)00454-2
ARTICLE IN PRESS 50
S. Wang et al. / Journal of Magnetism and Magnetic Materials 266 (2003) 49–56
surfactant coatings break down, leading to a reduction in interparticle spacing and eventually to particle coalescence. We have recently shown that the addition of Ag can reduce the ordering temperature by B150 C [2,3]. In this paper, we present a detailed study of the magnetic properties of [Fe49Pt51]88Ag12 nanoparticles as a function of annealing temperature.
2. Experimental procedure [Fe49Pt51]88Ag12 (hereafter referred to as FePtAg) particles were produced by a modification of the technique of Sun et al. for producing FePt nanoparticles [1]. The technique involved the simultaneous thermal decomposition of iron pentacarbonyl and the reduction of platinum acetylacetonate and silver acetate in 1,2-hexadecanediol in phenyl ether. Oleic acid and oleylamine were added to provide a surfactant coating to the particles. The particles were precipitated and rewashed with ethanol and separated by centrifugation. The particle precipitants were purified by redispersing in hexane and reprecipitated by adding ethanol. The particles were then dispersed in a mixture of hexane and octane that included small amounts of oleic acid and oleylamine. In order to prepare films for transmission electron microscopy (TEM) analysis and magnetic measurements, the FePtAg dispersion was diluted with a mixture of hexane and octane, and a small amount of the diluted dispersion was dropped onto a carbon-coated TEM grid or a silicon substrate and allowed to dry. The particle composition was controlled by varying the concentration of the reactants. Details of the particle synthesis are reported elsewhere [2,3]. The composition of the particles was determined by energy dispersive X-ray analysis. Films were annealed in an Ar/2%H2 atmosphere for 30 min at temperatures from 300 C to 500 C. The degree of particle order of the arrays was determined by plane-view TEM images and by small-angle X-ray diffraction (XRD), and the degree of chemical order (transformation to the L10 phase) was determined by ordinary high-angle XRD. Magnetic measurements were made on films on silicon
substrates at room temperature with an alternating gradient magnetometer (AGM) and with a SQUID magnetometer.
3. Results Fig. 1 shows plane-view TEM images of particle arrays before annealing and after annealing at 400 C. The particles have assembled into a 3D hexagonal array and have maintained a high degree of order after annealing. The average particle size before annealing is B3.5 nm.
Fig. 1. Plane view TEM of FePtAg array: (a) before and (b) after annealing at 400 C for 30 min.
ARTICLE IN PRESS S. Wang et al. / Journal of Magnetism and Magnetic Materials 266 (2003) 49–56
peak appears at B400 C, indicating that the Ag is phase segregating from the FePt particles. A comparison of XRD spectra with and without Ag suggests that the Ag reduces the chemical ordering temperature by B150 C, a result which is substantiated by coercivity measurements. More details of the effect of Ag on chemical ordering, including the dependence on composition, are reported elsewhere [2,3]. Fig. 3 shows hysteresis loops and DCD curves measured with a SQUID magnetometer with a saturation field of 50 kOe. The coercivity (Hc ), remanent coercivity (Hcr ), coercivity ratio (Hcr =Hc ), and loop squareness (SQ) are given as a function of the annealing temperature (Ta ) in Table 1. For 3D (2D) randomly oriented, non-interacting, uniaxial, single-domain particles with identical magnetic anisotropy constants, and with thermal effects excluded, Hcr =Hc ¼ 1:09 (1.07) [4]. It has recently been shown, however, that the coercivity ratio increases monotonically with increasing anisotropy distribution width [5]. The effect of interactions on the coercivity ratio has been
(200) (002)
Intensity
*(011)
*(001)
(111)
Fig. 2 shows XRD spectra before and after annealing. The spectra show that the particles have a disordered FCC structure before annealing and begin to transform to the L10 phase at an annealing temperature of B350 C. The Ag (1 1 1)
#
(f) (e) (d) (c) (b) (a)
20
30
40 2θ (deg.)
50
60
Fig. 2. XRD spectra of FePtAg nanoparticles: (a) as-made and after annealing for 30 min at (b) 300 C, (c) 350 C, (d) 400 C, (e) 450 C, and (f) 500 C. L10 peaks are noted by (), and the Ag (1 1 1) peak is noted by (#).
350 C 1
M, M d
400 C hyst dcd
hyst dcd
0.5 0 -0.5 -1 -10
-5
0 H (kOe)
10 -30
5
-20
-10
0
10
20
30
H (kOe) 500 C
450 C 1
hyst dcd
hyst dcd
M, M d
0.5 0 -0.5 -1 -40
-20
0 H (kOe)
20
51
40
-40
-20
0 H (kOe)
20
40
Fig. 3. Hysteresis and DC demagnetization curves for annealed FePtAg nanoparticles.
ARTICLE IN PRESS S. Wang et al. / Journal of Magnetism and Magnetic Materials 266 (2003) 49–56
52
from 400 C to 500 C is consistent with increasing exchange interactions due to particle sintering. The squareness values for the 350 C and 400 C annealed samples are similar to that expected for non-interacting particles randomly oriented in 2D (0.63) and higher than that expected for 3D orientation (0.5). Although the TEM images do not show strong evidence of sintering in these particular samples, exchange interactions would be expected to increase the squareness. Thus, the degree of particle orientation (3D versus 2D random) cannot be ascertained. Delta-M measurements made using a SQUID magnetometer are shown in Fig. 4. The delta-M curve is defined by
calculated using a mean field model [5] (Hint ¼ gM) and the effect of exchange and magnetostatic interactions have been calculated using a Monte Carlo technique [6]. Either exchange interactions or a positive mean field constant (g) will decrease the coercivity ratio, while either magnetostatic interactions or a negative mean field constant will increase the coercivity ratio. The effect of magnetostatic interactions is relatively weak for the large anisotropy values that are typical of highly ordered FePt. Based on these calculations, the large coercivity ratios shown in Table 1 suggest large anisotropy distributions in the FePtAg nanoparticles. The decrease in coercivity ratio
dM ¼ 2Mr 1 þ Md ;
where Mr is the isothermal remanent magnetization obtained starting with a virgin sample in a demagnetized state, and Md is the DC demagnetization obtained after applying a saturating field [7]. The DCD measurements were made with an initial saturation field of 50 kOe, and the field was brought to zero between each reverse field without re-saturating the sample. This sequence has been shown to minimize thermal effects in the delta-M
Table 1 Hysteresis and remanence properties of annealed FePtAg arrays Ta ( C)
Hc (Oe)
Hcr (Oe)
Hcr =Hc
SQ
350 400 450 500
690 3430 8690 13,790
1032 5924 11,733 16,465
1.50 1.73 1.35 1.19
0.659 0.609 0.715 0.795
ð1Þ
0
delta-M
0.1 -0.1
Ta = 400 C
Ta = 350 C 0
-0.2 -0.1 -0.3 0
2
4
6
8
10
12
0
5
H (kOe)
delta-M
0.1
0
0
-0.1
-0.1 10
20
30
H (kOe)
15
0.1
Ta = 450 C
0
10
20
H (kOe)
40
50
Ta = 500 C
0
10
20
30
H (kOe)
Fig. 4. Delta-M curves for annealed FePtAg nanoparticles.
40
50
ARTICLE IN PRESS
The appropriate value of f0 in Eq. (2) depends on the anisotropy, which is different for each sample, and on magnetic field at short times. For simplicity, we have used f0 ¼ 109 Hz for all samples, although a higher value might be more appropriate for the highly ordered samples. The 2/3 exponent is appropriate for randomly oriented easy axes. Fig. 6 shows H0 and KV =kB T as a function of the annealing temperature. As expected, both the intrinsic switching field and the thermal stability factor increase with annealing temperature and degree of chemical ordering. It should be noted, however, that these parameters are not representative of the entire energy barrier and switching field distribution of the particles. Those particles with energy barriers below the superparamagnetic limit and those with switching fields above the maximum applied field will not be switched. It is useful, nevertheless, to examine the significance of the fit parameters. If we assume
16
500°C
12
450°C
8
400°C
4 0
53
T (anneal) = 350°C -9
-6
-3 log (t)
0
3
Fig. 5. Dynamic coercivity of annealed FePtAg nanoparticles.
H0 (kOe)
20 15
300 H 0 (kO e)
200
10 KV/kT
5 0 300
KV/kT
measurements [8,9]. The delta-M curve is negative for the samples annealed at 350 C and 400 C, is nearly zero for the sample annealed at 450 C, and has a positive component for the sample annealed at 500 C. This trend can be interpreted in terms of competing magnetostatic interactions, which produce a negative delta-M, and exchange interactions due to particle aggregation, which produce a positive delta-M. These results are consistent with the change in coercivity ratio and squareness with annealing temperature shown in Table 1. The sign and shape of the delta-M curves can depend strongly on the maximum field applied during the DCD measurements if this field is not sufficient to completely saturate the magnetization. For example, when the delta-M measurements are made using an AGM with a maximum field of 18 kOe (not shown), the curves for the samples annealed at 450 C and 500 C are completely negative. Dynamic remanent coercivity measurements made using an AGM with a maximum field of 18 kOe are shown in Fig. 5. The measurements were fit to a Sharrock formula of the form [10] ( 2=3 ) kB T ln ðf0 tÞ Hcr ðtÞ ¼ H0 1 : ð2Þ KV
HCR (kOe)
S. Wang et al. / Journal of Magnetism and Magnetic Materials 266 (2003) 49–56
100
0 350
400
450
500
550
Ta (C) Fig. 6. Intrinsic switching field and thermal stability parameter of annealed FePtAg nanoparticles as a function of annealing temperature.
coherent rotation (which may not be fully valid for the exchange-coupled samples), then K ¼ 3 1 [11] 2Ms Hk BMs H0 : Using Ms ¼ 1140 emu/cm 3 gives K ¼ 1:1 107 emu/cm for Ta ¼ 400 C and K ¼ 1:9 107 emu/cm3 for Ta ¼ 500 C. These values can be compared with the reported bulk value of 6.6–10 107 emu/cm3 [11]. A comparison of AGM and SQUID hysteresis loops suggests that the AGM saturation field is sufficiently high for the 400 C sample but is not high enough for the 500 C sample. Thus, these values of K may not be unreasonable. On the other hand, using these values of K in the measured values of the thermal stability parameter gives very large values of V : These switching values are shown in Fig. 7 along with ‘‘activation volumes’’ that are calculated
ARTICLE IN PRESS S. Wang et al. / Journal of Magnetism and Magnetic Materials 266 (2003) 49–56
Vsw
rs
0.8
350 C
V0
0.6 0.4 0.2 0
400
450
2
500
-1
Ta (C)
0.02 1
400 C
10 log (t)
100
450 C 0
particle volume = 2.24 x 10-20 cc
350
0.04
4
r
Vac
M /M
1
S0 (%/dec)
volume (10
-18
cc)
54
500 C
0 M r0
1
Fig. 7. Magnetization reversal volumes for FePtAg nanoparticles as a function of annealing temperature.
Fig. 8. Zero field viscosity versus initial remanence for annealed FePtAg nanoparticles. Inset shows representative time decay measurement.
using the relationship [12]
interactions [14–17]. In this experiment the magnetization decay is measured in zero field after applying a saturation-reverse-zero field sequence. This measurement differs from the more commonly used method in which the viscosity is measured in the presence of a reverse field. The measurements were made with an AGM with a ‘‘saturating’’ field of 18 kOe. The viscosity coefficient, S ¼ d Mr =d logðtÞ; was calculated from the initial part of the decay (using the data from 1 to 20 s). The measurements are as shown in Fig. 8. For all samples the viscosity changes sign at high remanence values, a feature that is typical of all such measurements. The change in sign of the viscosity with remanence is a consequence of the fact that the remanent state consists of a distribution of rapidly relaxing moments with positive viscosity and slowly relaxing moments with negative viscosity. The overall sign depends on the relative amounts of these ‘‘soft’’ and ‘‘hard’’ components. Interparticle interactions can affect where this sign transition takes place. As expected, the magnitude of the viscosity increases with decreasing annealing temperature because of decreasing thermal stability. The change in the shape of the viscosity curves with annealing temperature, however, is not well understood. For the samples annealed at high temperature, the viscosity curve is nearly flat until it begins to change sign at high remanence values; however, as the annealing temperature decreases the curves become more peaked. This variation of the shape
Vac ¼
kT ; Ms H f
ð3Þ
where Hf ¼
d Hcr : d lnðtÞ
ð4Þ
Eq. (4) is equivalent to Hf ¼ Sr =wirr ; where Sr and wirr are the remanent viscosity and irreversible susceptibility measured at the remanent coercivity [13]. The ‘‘switching’’ and ‘‘activation’’ volumes are similar in magnitude and range from 15 to 44 times the volume of a 3.5 nm particle. Larger magnetic reversal volumes would be expected for the samples annealed at 450 C and 500 C because of particle aggregation, as suggested by the deltaM curves; however, based on TEM and delta-M measurements significant aggregation does not seem to occur for the samples annealed at 350 C and 400 C. It should be noted; however, that the magnetic measurements were made on samples coated on silicon for which no structural data is available to determine the extent of aggregation. Zero field viscosity measurements have also been made on the annealed nanoparticle arrays as a function of the initial remanent magnetization obtained after partial DC demagnetization. These types of measurements have been recently made on a variety of recording media, including tapes, longitudinal, and perpendicular media, and have been shown to be very sensitive to interparticle
ARTICLE IN PRESS S. Wang et al. / Journal of Magnetism and Magnetic Materials 266 (2003) 49–56
of the curves with annealing temperature is opposite to what is expected. Modeling shows that the flat behavior is expected for weak interparticle interactions while the peaked behavior is characteristic of exchange interactions [16,18]. A possible explanation for this anomalous behavior is that the large anisotropy in the highly ordered films masks the effect of exchange interactions on the relaxation. Further modeling is needed, taking into account the wide range of anisotropy distributions and the difference in the relative strength of the exchange and anisotropy energy among the samples.
4. Summary and conclusions The addition of Ag to the chemical synthesis of FePt nanoparticles significantly reduces the annealing temperature required for good chemical ordering and high coercivity. Hysteresis, remanence, and delta-M curves suggest that the particles are magnetically isolated at annealing temperatures of 350 C and 400 C, but that exchange interactions within multigrain particles occur at annealing temperatures of 450 C and 500 C. Remanent to hysteresis coercivity ratios indicate that the particles have a large distribution of anisotropy energies. This may be related in part to a compositional distribution, which has been found in FePt nanoparticle arrays [19], and to other non-uniformities in the particles. Dynamic coercivity measurements yield anomalously high thermal stability factors, even for the samples annealed at 350 C and 400 C, for which there is little evidence of particle aggregation. Likewise, magnetic reversal volumes in all samples, including those annealed at low temperatures, are much higher than expected. Zero field viscosity curves for the samples annealed at high temperatures are characteristic of weakly interacting particles and for the samples annealed at low temperatures are characteristic of exchanged-coupled particles. The time-dependence measurements—thermal stability parameters, magnetization reversal volumes, zero-field viscosity—in the nanoparticles are not understood. Structural data is needed to determine the extent of particle aggregation in the
55
samples on silicon, which were used for the magnetic measurements. The results, however, may also be related to a large superparamagnetic fraction of particles. Safonov and Bertram have shown that the thermal stability of a ferromagnetic particle can be enhanced by magnetostatic interactions with a superparamagnetic particle [20]. Chantrell has pointed out the importance of the superparamagnetic part of the distribution in modeling the temperature dependence of the coercivity of FePt nanoparticle arrays, and he has shown the Sharrock formula does not accurately give the temperature dependence of the coercivity when a significant superparamagnetic fraction exists [21]. In the absence of interactions a large superparamagnetic fraction would give rise to a low remanence. Interparticle interactions, however, especially those between superparamagnetic particles and ferromagnetic particles with large thermal stability factors, can enhance the remanence. A broad energy barrier distribution with a significant superparamagnetic fraction appears to exist in these samples because of both anisotropy and volume distributions. The anisotropy distribution may be due to non-uniform chemical ordering, and the volume distribution may be enhanced by surface oxidation of the particles during annealing. For a 3.5 nm FePt particle, approximately 1/3 of the atoms are on the surface, so surface oxidation can significantly reduce the volume of the high-anisotropy core. Additional modeling is needed to understand the importance of the superparamagnetic particles on the stability of the ferromagnetic component.
Acknowledgements This work was supported by the NSF Materials Research Science and Engineering Center award number DMR-9809423.
References [1] S. Sun, C.B. Murray, D. Weller, L. Folks, A. Moser, Science 287 (2000) 1989. [2] S. Kang, J.W. Harrell, D.E. Nikles, Nano Lett. 2 (2002) 1033.
ARTICLE IN PRESS 56
S. Wang et al. / Journal of Magnetism and Magnetic Materials 266 (2003) 49–56
[3] S.S. Kang, D.E. Nikles, J.W. Harrell, J. Appl. Phys. 93 (2003) 7178. [4] E.P. Wohlfarth, J. Appl. Phys. 29 (1958) 595. [5] S. Wang, S. Kang, J.W. Harrell, Intermag Conference, Amsterdam, May 2002, paper AP-07. [6] S. Wang, S.S. Kang, J.W. Harrell, X.W. Wu, R.W. Chantrell, Phys. Rev. B, submitted for publication. [7] P.E. Kelley, K. O’Grady, P.L. Mayo, R.W. Chantrell, IEEE Trans. Magn. 25 (1989) 3881. [8] S. Wang, A.F. Khapikov, S. Brown, J.W. Harrell, IEEE Trans. Magn. 37 (2001) 1518. [9] J.W. Harrell, S. Wang, H. Fujiwara, J. Appl. Phys. 91 (2002) 8605. [10] M.P. Sharrock, J.T. McKinney, IEEE Trans. Magn. 17 (1981) 3020. [11] D. Weller, A. Moser, L. Folks, M.E. Best, W. Lee, M.F. Toney, M. Schwickert, J.-U. Thiel, M.F. Dorner, IEEE Trans. Magn. 36 (2000) 10.
[12] E.P. Wohlfarth, J. Phys. F 14 (1961) L155. [13] A. Lyberatos, R.W. Chantrell, J. Phys.: Condens. Matter 9 (1997) 2623. [14] A.F. Khapikov, S. Wang, B. Xu, J.W. Harrell, Phys. Rev. B 63 (2000) 020401. [15] S. Wang, A.F. Khapikov, J.W. Harrell, IEEE Trans. Magn. 37 (2001) 1541. [16] J.W. Harrell, S. Wang, D.E. Nikles, M. Chen, Appl. Phys. Lett. 79 (2001) 4393. [17] S. Wang, T. Zhao, J.W. Harrell, IEEE Trans. Magn. 38 (2002) 2000. [18] R.W. Chantrell, personal communication. [19] M. Chen, S. Wang, D.E. Nikles, J.W. Harrell, MMM Conference, Seattle, November 2001, paper AC-01. [20] V.L. Safonov, H.N. Bertram, MMM Conference, Tampa, Nov. 2002, paper GT-08, preprint. [21] R.W. Chantrell, D. Weller, T.J. Klemmer, S. Sun, E.E. Fullerton, J. Appl. Phys. 91 (2002) 6866.
Journal of Magnetism and Magnetic Materials 266 (2003) 49–56
Magnetic properties of self-organized L10 FePtAg nanoparticle arrays S. Wanga, S.S. Kanga, D.E. Niklesa,b, J.W. Harrella,*, X.W. Wuc a
Department of Physics and Astronomy, Center for Materials for Information Technology, University of Alabama, Tuscaloosa, AL 3548, USA b Department of Chemistry, Center for Materials for Information Technology, The University of Alabama, Tuscaloosa, AL 35487, USA c Seagate Research, 2403 Sidney Street, Pittsburgh, PA 15203, USA Received 30 September 2002; received in revised form 16 December 2002
Abstract The magnetic properties of chemically synthesized high anisotropy L10 [Fe49Pt51]88Ag12 nanoparticle arrays have been studied as a function of annealing temperature. Particles were prepared by the simultaneous polyol reduction of platinum acetylacetonate and silver acetate and the thermal decomposition of iron carbonyl, yielding monodispersed particles of diameter B3.5 nm. Addition of Ag lowers the ordering temperature of self-assembled arrays by B150 C. After annealing at Ta ¼ 500 C for 30 min in an Ar/H2 atmosphere, the coercivity was 13,800 Oe. TEM and delta-M measurements indicate weak particle aggregation up to Ta ¼ 400 C, with evidence of sintering at higher temperatures. Large ratios of remanent to hysteresis coercivity indicate a large distribution in anisotropy energies. Anomalously large thermal stability constants, KV =kB T; and switching volumes were measured, even in samples with very little evidence of sintering. Zero field viscosity versus remanence curves show evidence of exchange interactions. r 2003 Elsevier B.V. All rights reserved. PACS: 75.75.+a; 75.50.Bb; 75.50.Ss; 75.50.Tt; 75.50.Vv Keywords: FePt nanoparticles; L10 phase; Chemical ordering; Magnetic recording materials; Self-assembly; Magnetic viscosity
1. Introduction There has been considerable interest in ordered arrays of high-anisotropy magnetic nanoparticles for ultra-high density recording media since a recent report by Sun and co-workers on the synthesis and self-assembly of L10 FePt nanoparticles [1]. FePt nanoparticles with a disordered *Corresponding author. Tel.: +1-205-348-9404; fax: +1205-348-2346. E-mail address: [email protected] (J.W. Harrell).
FCC structure and with an organic coating were produced by chemical synthesis. Because the particles had a narrow size distribution, a nonmagnetic coating, and were superparamagnetic, self-assembled arrays could be produced by slow evaporation of a dispersion of the particles on a substrate. In order to obtain a high degree of L10 ordering, the arrays had to be annealed at temperatures of 550 C or greater. These high annealing temperatures, however, degrade the quality of the arrays and would be undesirable for processing of media. At high temperatures, the
0304-8853/03/$ - see front matter r 2003 Elsevier B.V. All rights reserved. doi:10.1016/S0304-8853(03)00454-2
ARTICLE IN PRESS 50
S. Wang et al. / Journal of Magnetism and Magnetic Materials 266 (2003) 49–56
surfactant coatings break down, leading to a reduction in interparticle spacing and eventually to particle coalescence. We have recently shown that the addition of Ag can reduce the ordering temperature by B150 C [2,3]. In this paper, we present a detailed study of the magnetic properties of [Fe49Pt51]88Ag12 nanoparticles as a function of annealing temperature.
2. Experimental procedure [Fe49Pt51]88Ag12 (hereafter referred to as FePtAg) particles were produced by a modification of the technique of Sun et al. for producing FePt nanoparticles [1]. The technique involved the simultaneous thermal decomposition of iron pentacarbonyl and the reduction of platinum acetylacetonate and silver acetate in 1,2-hexadecanediol in phenyl ether. Oleic acid and oleylamine were added to provide a surfactant coating to the particles. The particles were precipitated and rewashed with ethanol and separated by centrifugation. The particle precipitants were purified by redispersing in hexane and reprecipitated by adding ethanol. The particles were then dispersed in a mixture of hexane and octane that included small amounts of oleic acid and oleylamine. In order to prepare films for transmission electron microscopy (TEM) analysis and magnetic measurements, the FePtAg dispersion was diluted with a mixture of hexane and octane, and a small amount of the diluted dispersion was dropped onto a carbon-coated TEM grid or a silicon substrate and allowed to dry. The particle composition was controlled by varying the concentration of the reactants. Details of the particle synthesis are reported elsewhere [2,3]. The composition of the particles was determined by energy dispersive X-ray analysis. Films were annealed in an Ar/2%H2 atmosphere for 30 min at temperatures from 300 C to 500 C. The degree of particle order of the arrays was determined by plane-view TEM images and by small-angle X-ray diffraction (XRD), and the degree of chemical order (transformation to the L10 phase) was determined by ordinary high-angle XRD. Magnetic measurements were made on films on silicon
substrates at room temperature with an alternating gradient magnetometer (AGM) and with a SQUID magnetometer.
3. Results Fig. 1 shows plane-view TEM images of particle arrays before annealing and after annealing at 400 C. The particles have assembled into a 3D hexagonal array and have maintained a high degree of order after annealing. The average particle size before annealing is B3.5 nm.
Fig. 1. Plane view TEM of FePtAg array: (a) before and (b) after annealing at 400 C for 30 min.
ARTICLE IN PRESS S. Wang et al. / Journal of Magnetism and Magnetic Materials 266 (2003) 49–56
peak appears at B400 C, indicating that the Ag is phase segregating from the FePt particles. A comparison of XRD spectra with and without Ag suggests that the Ag reduces the chemical ordering temperature by B150 C, a result which is substantiated by coercivity measurements. More details of the effect of Ag on chemical ordering, including the dependence on composition, are reported elsewhere [2,3]. Fig. 3 shows hysteresis loops and DCD curves measured with a SQUID magnetometer with a saturation field of 50 kOe. The coercivity (Hc ), remanent coercivity (Hcr ), coercivity ratio (Hcr =Hc ), and loop squareness (SQ) are given as a function of the annealing temperature (Ta ) in Table 1. For 3D (2D) randomly oriented, non-interacting, uniaxial, single-domain particles with identical magnetic anisotropy constants, and with thermal effects excluded, Hcr =Hc ¼ 1:09 (1.07) [4]. It has recently been shown, however, that the coercivity ratio increases monotonically with increasing anisotropy distribution width [5]. The effect of interactions on the coercivity ratio has been
(200) (002)
Intensity
*(011)
*(001)
(111)
Fig. 2 shows XRD spectra before and after annealing. The spectra show that the particles have a disordered FCC structure before annealing and begin to transform to the L10 phase at an annealing temperature of B350 C. The Ag (1 1 1)
#
(f) (e) (d) (c) (b) (a)
20
30
40 2θ (deg.)
50
60
Fig. 2. XRD spectra of FePtAg nanoparticles: (a) as-made and after annealing for 30 min at (b) 300 C, (c) 350 C, (d) 400 C, (e) 450 C, and (f) 500 C. L10 peaks are noted by (), and the Ag (1 1 1) peak is noted by (#).
350 C 1
M, M d
400 C hyst dcd
hyst dcd
0.5 0 -0.5 -1 -10
-5
0 H (kOe)
10 -30
5
-20
-10
0
10
20
30
H (kOe) 500 C
450 C 1
hyst dcd
hyst dcd
M, M d
0.5 0 -0.5 -1 -40
-20
0 H (kOe)
20
51
40
-40
-20
0 H (kOe)
20
40
Fig. 3. Hysteresis and DC demagnetization curves for annealed FePtAg nanoparticles.
ARTICLE IN PRESS S. Wang et al. / Journal of Magnetism and Magnetic Materials 266 (2003) 49–56
52
from 400 C to 500 C is consistent with increasing exchange interactions due to particle sintering. The squareness values for the 350 C and 400 C annealed samples are similar to that expected for non-interacting particles randomly oriented in 2D (0.63) and higher than that expected for 3D orientation (0.5). Although the TEM images do not show strong evidence of sintering in these particular samples, exchange interactions would be expected to increase the squareness. Thus, the degree of particle orientation (3D versus 2D random) cannot be ascertained. Delta-M measurements made using a SQUID magnetometer are shown in Fig. 4. The delta-M curve is defined by
calculated using a mean field model [5] (Hint ¼ gM) and the effect of exchange and magnetostatic interactions have been calculated using a Monte Carlo technique [6]. Either exchange interactions or a positive mean field constant (g) will decrease the coercivity ratio, while either magnetostatic interactions or a negative mean field constant will increase the coercivity ratio. The effect of magnetostatic interactions is relatively weak for the large anisotropy values that are typical of highly ordered FePt. Based on these calculations, the large coercivity ratios shown in Table 1 suggest large anisotropy distributions in the FePtAg nanoparticles. The decrease in coercivity ratio
dM ¼ 2Mr 1 þ Md ;
where Mr is the isothermal remanent magnetization obtained starting with a virgin sample in a demagnetized state, and Md is the DC demagnetization obtained after applying a saturating field [7]. The DCD measurements were made with an initial saturation field of 50 kOe, and the field was brought to zero between each reverse field without re-saturating the sample. This sequence has been shown to minimize thermal effects in the delta-M
Table 1 Hysteresis and remanence properties of annealed FePtAg arrays Ta ( C)
Hc (Oe)
Hcr (Oe)
Hcr =Hc
SQ
350 400 450 500
690 3430 8690 13,790
1032 5924 11,733 16,465
1.50 1.73 1.35 1.19
0.659 0.609 0.715 0.795
ð1Þ
0
delta-M
0.1 -0.1
Ta = 400 C
Ta = 350 C 0
-0.2 -0.1 -0.3 0
2
4
6
8
10
12
0
5
H (kOe)
delta-M
0.1
0
0
-0.1
-0.1 10
20
30
H (kOe)
15
0.1
Ta = 450 C
0
10
20
H (kOe)
40
50
Ta = 500 C
0
10
20
30
H (kOe)
Fig. 4. Delta-M curves for annealed FePtAg nanoparticles.
40
50
ARTICLE IN PRESS
The appropriate value of f0 in Eq. (2) depends on the anisotropy, which is different for each sample, and on magnetic field at short times. For simplicity, we have used f0 ¼ 109 Hz for all samples, although a higher value might be more appropriate for the highly ordered samples. The 2/3 exponent is appropriate for randomly oriented easy axes. Fig. 6 shows H0 and KV =kB T as a function of the annealing temperature. As expected, both the intrinsic switching field and the thermal stability factor increase with annealing temperature and degree of chemical ordering. It should be noted, however, that these parameters are not representative of the entire energy barrier and switching field distribution of the particles. Those particles with energy barriers below the superparamagnetic limit and those with switching fields above the maximum applied field will not be switched. It is useful, nevertheless, to examine the significance of the fit parameters. If we assume
16
500°C
12
450°C
8
400°C
4 0
53
T (anneal) = 350°C -9
-6
-3 log (t)
0
3
Fig. 5. Dynamic coercivity of annealed FePtAg nanoparticles.
H0 (kOe)
20 15
300 H 0 (kO e)
200
10 KV/kT
5 0 300
KV/kT
measurements [8,9]. The delta-M curve is negative for the samples annealed at 350 C and 400 C, is nearly zero for the sample annealed at 450 C, and has a positive component for the sample annealed at 500 C. This trend can be interpreted in terms of competing magnetostatic interactions, which produce a negative delta-M, and exchange interactions due to particle aggregation, which produce a positive delta-M. These results are consistent with the change in coercivity ratio and squareness with annealing temperature shown in Table 1. The sign and shape of the delta-M curves can depend strongly on the maximum field applied during the DCD measurements if this field is not sufficient to completely saturate the magnetization. For example, when the delta-M measurements are made using an AGM with a maximum field of 18 kOe (not shown), the curves for the samples annealed at 450 C and 500 C are completely negative. Dynamic remanent coercivity measurements made using an AGM with a maximum field of 18 kOe are shown in Fig. 5. The measurements were fit to a Sharrock formula of the form [10] ( 2=3 ) kB T ln ðf0 tÞ Hcr ðtÞ ¼ H0 1 : ð2Þ KV
HCR (kOe)
S. Wang et al. / Journal of Magnetism and Magnetic Materials 266 (2003) 49–56
100
0 350
400
450
500
550
Ta (C) Fig. 6. Intrinsic switching field and thermal stability parameter of annealed FePtAg nanoparticles as a function of annealing temperature.
coherent rotation (which may not be fully valid for the exchange-coupled samples), then K ¼ 3 1 [11] 2Ms Hk BMs H0 : Using Ms ¼ 1140 emu/cm 3 gives K ¼ 1:1 107 emu/cm for Ta ¼ 400 C and K ¼ 1:9 107 emu/cm3 for Ta ¼ 500 C. These values can be compared with the reported bulk value of 6.6–10 107 emu/cm3 [11]. A comparison of AGM and SQUID hysteresis loops suggests that the AGM saturation field is sufficiently high for the 400 C sample but is not high enough for the 500 C sample. Thus, these values of K may not be unreasonable. On the other hand, using these values of K in the measured values of the thermal stability parameter gives very large values of V : These switching values are shown in Fig. 7 along with ‘‘activation volumes’’ that are calculated
ARTICLE IN PRESS S. Wang et al. / Journal of Magnetism and Magnetic Materials 266 (2003) 49–56
Vsw
rs
0.8
350 C
V0
0.6 0.4 0.2 0
400
450
2
500
-1
Ta (C)
0.02 1
400 C
10 log (t)
100
450 C 0
particle volume = 2.24 x 10-20 cc
350
0.04
4
r
Vac
M /M
1
S0 (%/dec)
volume (10
-18
cc)
54
500 C
0 M r0
1
Fig. 7. Magnetization reversal volumes for FePtAg nanoparticles as a function of annealing temperature.
Fig. 8. Zero field viscosity versus initial remanence for annealed FePtAg nanoparticles. Inset shows representative time decay measurement.
using the relationship [12]
interactions [14–17]. In this experiment the magnetization decay is measured in zero field after applying a saturation-reverse-zero field sequence. This measurement differs from the more commonly used method in which the viscosity is measured in the presence of a reverse field. The measurements were made with an AGM with a ‘‘saturating’’ field of 18 kOe. The viscosity coefficient, S ¼ d Mr =d logðtÞ; was calculated from the initial part of the decay (using the data from 1 to 20 s). The measurements are as shown in Fig. 8. For all samples the viscosity changes sign at high remanence values, a feature that is typical of all such measurements. The change in sign of the viscosity with remanence is a consequence of the fact that the remanent state consists of a distribution of rapidly relaxing moments with positive viscosity and slowly relaxing moments with negative viscosity. The overall sign depends on the relative amounts of these ‘‘soft’’ and ‘‘hard’’ components. Interparticle interactions can affect where this sign transition takes place. As expected, the magnitude of the viscosity increases with decreasing annealing temperature because of decreasing thermal stability. The change in the shape of the viscosity curves with annealing temperature, however, is not well understood. For the samples annealed at high temperature, the viscosity curve is nearly flat until it begins to change sign at high remanence values; however, as the annealing temperature decreases the curves become more peaked. This variation of the shape
Vac ¼
kT ; Ms H f
ð3Þ
where Hf ¼
d Hcr : d lnðtÞ
ð4Þ
Eq. (4) is equivalent to Hf ¼ Sr =wirr ; where Sr and wirr are the remanent viscosity and irreversible susceptibility measured at the remanent coercivity [13]. The ‘‘switching’’ and ‘‘activation’’ volumes are similar in magnitude and range from 15 to 44 times the volume of a 3.5 nm particle. Larger magnetic reversal volumes would be expected for the samples annealed at 450 C and 500 C because of particle aggregation, as suggested by the deltaM curves; however, based on TEM and delta-M measurements significant aggregation does not seem to occur for the samples annealed at 350 C and 400 C. It should be noted; however, that the magnetic measurements were made on samples coated on silicon for which no structural data is available to determine the extent of aggregation. Zero field viscosity measurements have also been made on the annealed nanoparticle arrays as a function of the initial remanent magnetization obtained after partial DC demagnetization. These types of measurements have been recently made on a variety of recording media, including tapes, longitudinal, and perpendicular media, and have been shown to be very sensitive to interparticle
ARTICLE IN PRESS S. Wang et al. / Journal of Magnetism and Magnetic Materials 266 (2003) 49–56
of the curves with annealing temperature is opposite to what is expected. Modeling shows that the flat behavior is expected for weak interparticle interactions while the peaked behavior is characteristic of exchange interactions [16,18]. A possible explanation for this anomalous behavior is that the large anisotropy in the highly ordered films masks the effect of exchange interactions on the relaxation. Further modeling is needed, taking into account the wide range of anisotropy distributions and the difference in the relative strength of the exchange and anisotropy energy among the samples.
4. Summary and conclusions The addition of Ag to the chemical synthesis of FePt nanoparticles significantly reduces the annealing temperature required for good chemical ordering and high coercivity. Hysteresis, remanence, and delta-M curves suggest that the particles are magnetically isolated at annealing temperatures of 350 C and 400 C, but that exchange interactions within multigrain particles occur at annealing temperatures of 450 C and 500 C. Remanent to hysteresis coercivity ratios indicate that the particles have a large distribution of anisotropy energies. This may be related in part to a compositional distribution, which has been found in FePt nanoparticle arrays [19], and to other non-uniformities in the particles. Dynamic coercivity measurements yield anomalously high thermal stability factors, even for the samples annealed at 350 C and 400 C, for which there is little evidence of particle aggregation. Likewise, magnetic reversal volumes in all samples, including those annealed at low temperatures, are much higher than expected. Zero field viscosity curves for the samples annealed at high temperatures are characteristic of weakly interacting particles and for the samples annealed at low temperatures are characteristic of exchanged-coupled particles. The time-dependence measurements—thermal stability parameters, magnetization reversal volumes, zero-field viscosity—in the nanoparticles are not understood. Structural data is needed to determine the extent of particle aggregation in the
55
samples on silicon, which were used for the magnetic measurements. The results, however, may also be related to a large superparamagnetic fraction of particles. Safonov and Bertram have shown that the thermal stability of a ferromagnetic particle can be enhanced by magnetostatic interactions with a superparamagnetic particle [20]. Chantrell has pointed out the importance of the superparamagnetic part of the distribution in modeling the temperature dependence of the coercivity of FePt nanoparticle arrays, and he has shown the Sharrock formula does not accurately give the temperature dependence of the coercivity when a significant superparamagnetic fraction exists [21]. In the absence of interactions a large superparamagnetic fraction would give rise to a low remanence. Interparticle interactions, however, especially those between superparamagnetic particles and ferromagnetic particles with large thermal stability factors, can enhance the remanence. A broad energy barrier distribution with a significant superparamagnetic fraction appears to exist in these samples because of both anisotropy and volume distributions. The anisotropy distribution may be due to non-uniform chemical ordering, and the volume distribution may be enhanced by surface oxidation of the particles during annealing. For a 3.5 nm FePt particle, approximately 1/3 of the atoms are on the surface, so surface oxidation can significantly reduce the volume of the high-anisotropy core. Additional modeling is needed to understand the importance of the superparamagnetic particles on the stability of the ferromagnetic component.
Acknowledgements This work was supported by the NSF Materials Research Science and Engineering Center award number DMR-9809423.
References [1] S. Sun, C.B. Murray, D. Weller, L. Folks, A. Moser, Science 287 (2000) 1989. [2] S. Kang, J.W. Harrell, D.E. Nikles, Nano Lett. 2 (2002) 1033.
ARTICLE IN PRESS 56
S. Wang et al. / Journal of Magnetism and Magnetic Materials 266 (2003) 49–56
[3] S.S. Kang, D.E. Nikles, J.W. Harrell, J. Appl. Phys. 93 (2003) 7178. [4] E.P. Wohlfarth, J. Appl. Phys. 29 (1958) 595. [5] S. Wang, S. Kang, J.W. Harrell, Intermag Conference, Amsterdam, May 2002, paper AP-07. [6] S. Wang, S.S. Kang, J.W. Harrell, X.W. Wu, R.W. Chantrell, Phys. Rev. B, submitted for publication. [7] P.E. Kelley, K. O’Grady, P.L. Mayo, R.W. Chantrell, IEEE Trans. Magn. 25 (1989) 3881. [8] S. Wang, A.F. Khapikov, S. Brown, J.W. Harrell, IEEE Trans. Magn. 37 (2001) 1518. [9] J.W. Harrell, S. Wang, H. Fujiwara, J. Appl. Phys. 91 (2002) 8605. [10] M.P. Sharrock, J.T. McKinney, IEEE Trans. Magn. 17 (1981) 3020. [11] D. Weller, A. Moser, L. Folks, M.E. Best, W. Lee, M.F. Toney, M. Schwickert, J.-U. Thiel, M.F. Dorner, IEEE Trans. Magn. 36 (2000) 10.
[12] E.P. Wohlfarth, J. Phys. F 14 (1961) L155. [13] A. Lyberatos, R.W. Chantrell, J. Phys.: Condens. Matter 9 (1997) 2623. [14] A.F. Khapikov, S. Wang, B. Xu, J.W. Harrell, Phys. Rev. B 63 (2000) 020401. [15] S. Wang, A.F. Khapikov, J.W. Harrell, IEEE Trans. Magn. 37 (2001) 1541. [16] J.W. Harrell, S. Wang, D.E. Nikles, M. Chen, Appl. Phys. Lett. 79 (2001) 4393. [17] S. Wang, T. Zhao, J.W. Harrell, IEEE Trans. Magn. 38 (2002) 2000. [18] R.W. Chantrell, personal communication. [19] M. Chen, S. Wang, D.E. Nikles, J.W. Harrell, MMM Conference, Seattle, November 2001, paper AC-01. [20] V.L. Safonov, H.N. Bertram, MMM Conference, Tampa, Nov. 2002, paper GT-08, preprint. [21] R.W. Chantrell, D. Weller, T.J. Klemmer, S. Sun, E.E. Fullerton, J. Appl. Phys. 91 (2002) 6866.
Related Documents
Fct Feptag
October 2019 11
Feptag
October 2019 11
Fct 1189527088
October 2019 9
Etude Fct
November 2019 22
Fct Apresentar
November 2019 13