Major Findings 1. FePd thin films The structural characterization of the films grown at various substrate temperatures (RT- 700 o C) was performed ex-situ using X-Ray Diffraction (XRD). The optimum substrate temperature required to obtain long range ordered L1o FePd thin films was found by analyzing the relative intensity of various diffraction peaks in the XRD spectra (Figure 1). FePd (216 Å) grown on (001) MgO I (arb. units) 10 14 10 13 Ordered Phase L10 10 12 10 11 10 10 10 9 10 8 10 7 10 6 10 5 10 4 10 3 10 2 10 1 10 0 10-1 700ºC 600ºC 500ºC 450ºC 400ºC 200ºC RT 20 45 50 55 60 2θ(deg) Figure 1: XRD spectra of samples grown with the substrate held at different temperature ranging from 200 o C to 700 o C. The growth rate was calibrated from XRD studies at low angle on one of the samples (RT, 15 min growth) and found to be 14.4 Angstroms/min (Figure 2). 10 8 10 7 10 6 10 5 Experimental Fit 10 4 10 3 10 2 10 1 10 0 10-1 -2 0 2 4 6 8 10 12 14 16 2θ(deg) Figure 2. Low angle XRD for FePd film grown at RT.
The magnetic properties of the samples were characterized using longitudinal and polar magneto-optical Kerr effect. (Figure 3) a) RT f) MOKE (a.u.) b) 200ºC g) c) 450ºC h) d) 500ºC i) ROTATION (a.u.) e) 700ºC f) -500-250 0 250 500-10000 0 10000 H (Oe) Figure 3. Hysteresis loops measured with longitudinal Kerr geometry (left column) and polar Kerr geometry (right column) for FePd samples (~ 21 nm thick). The above data can be summarized in the following table: Growth temperature Growth mode Film structure RT-200ºC 2D growth Mostly disordered 400ºC-450ºC 2D growth Long range order 500ºC 3D growth Order + disorder 600ºC-700ºC 3 D growth Mostly disordered Thus, the correlation between XRD and MOKE established that the optimum substrate temperature for films with significant long range L1o order and perpendicular magnetic anisotropy was 450 o C.
A series of samples of varying thickness were grown at 450 o C and later characterized with ex-situ XRD, AFM/MFM and MOKE. A set of samples was not capped while to other sets were capped with 20 Angstroms of Pd and 30 Angstroms MgO respectively. From the integrated diffraction-peak intensities of several peaks [e.g. (100) and (200)] the degree of long-range order was estimated to be very similar in all the samples. Detailed cross-sectional TEM studies are currently in progress to correlate with the XRD data. (Figure 4). (a) (b) (c) (d) Figure 4. (a) HRTEM image showing the MgO substrate (top) and the FePd film (bottom). We notice that the film has crystalline structure following the epitaxy on the MgO substrate and that there are regions in the film that exhibit a periodic structure parallel to the film/substrate interface that has twice the period of other regions, indicating high chemical order with L1o structure. FFT transforms of various regions confirm this. (b) FFT corresponding to the substrate; (c) FFT corresponding to the highly ordered region. We observe only one direction where there are intermediate spots in the four-fold symmetric pattern, corresponding with the fringes parallel to the interface; (d) FFT corresponding to a crystalline but chemically disordered region in the film.
The magnetic characterization of these samples was done using polar MOKE. Thus, hysteresis cycles as well as the spectral dependence of θ K were determined. (Figures 5,6) R/R (a.u.) 30 A 75 A 225 A -10000 0 10000 H(Gauss) Figure 5. Hysteresis loops in polar Kerr geometry for a series of MgO-capped FePd samples of thickness 30, 75 and 225 Angstroms respectively. 0.2 FePd Kerr Rotation (º) 0.0-0.2-10000 0 10000 H(Gauss) 50A 100A 200A 300A Figure 6. Hysteresis loops in polar Kerr geometry measured for a series of epitaxial FePd films (50-300 Angstroms thick) grown at 450 o C on (001) MgO and capped with 20 Angstroms Pd.
-0.04-0.06 FePd 50A 100A 200A 300A Theta Kerr (deg.) -0.08-0.10-0.12-0.14 Bulk -0.16-0.18 1.5 2.0 2.5 3.0 3.5 4.0 4.5 Energy (ev) Figure 7. Spectral dependence of θ K for FePd highly ordered films (Pd-capped) and bulk Fe 0.5 Pd 0.5 for comparison. The experimental spectra obtained for the samples were compared to published spectra for bulk Fe 0.5 Pd 0.5. We notice that the thinner films exhibit maxima at similar energy than bulk while the thickener films have maxima shifted towards higher energies. We are currently analyzing these data. Regarding the hysteresis loops, we notice that there is different trend depending on the capping of the samples. The MgO-capped samples exhibit larger coercive fields scaling in inverse relation with film thickness, while the coercive fields observed for the Pdcapped samples are smaller but scale in direct proportion with film thickness. We also observed with longitudinal MOKE that the magnetization exhibits a component parallel to the surface of the films, and therefore the magnetization vector at the surface of the films is canted. We postulate that the Pd in the capping layer may be polarized in the latter set, facilitating closure domains with magnetization component parallel to the film surface. This was not the case for MgO capped films, where we could not saturate the films in the direction parallel to the plane of the films for fields as large as 1.2 Tesla. We postulate that in this case the oxide capping layer does not facilitate closure domains at the surface and therefore highly anisotropic nano-regions in the film are harder to reverse and hence the larger coercivity. Modeling of the data to address this trends is in progress.
The surface morphology of the films was studied with ex-situ AFM. (Figure 8, 9) Figure 8. Surface morphology of non-capped FePd films of thickness 30, 75 and 225 Angstroms respectively. 200nm 200nm 200nm Figure 9. Surface morphology of capped FePd films of thickness 50, 100 and 250 Angstroms respectively. We notice the typical island size and inter-island distance scales with the thickness of the films, regardless of the capping material used, thus studies regarding scaling properties of the surface features are in progress. We also notice that capped films tend to be smoother than non-capped ones. An important aspect of our studies is to try also to find correlation between surface features and magnetic properties. Thus, correlated magnetic force microscopy images were obtained on the same regions of the samples where the topographic images discussed above were obtained. The domain size was measured to perform correlated studies between domain-size, surface morphology and microstructure of the samples. (Figure 10).
Figure 10. Magnetic force imaging of capped FePd films of thickness 50, 100 and 250 Angstroms respectively. Domain-size measurements where estimated from the MFM images using two methods: FFT analysis and a graphic method (from Hubert and Schafer, Magnetic Domains, Springer, 2000). The measurements are summarized in the following table: sample image D_FFT (nm) D_ster (nm) 2 150 120504A 4 120 25 nm thick reman4 150 130504A 20 nm thick 120504B 10 nm thick 130504B 5 nm thick reman3 140 11 180 12 140 reman3 170 reman4 130 22 390 23 260 reman4 290 reman2 160 13 550 12 790 reman1 590 13 130 We notice inverse scaling of the domain size with increasing film thickness. These results are currently being analyzed and correlated with the additional information obtained on the microstructure and surface morphology of the samples. Modeling of the data is in progress.
2. FePt thin films The conditions for the growth of epitaxial FePt thin films with long range order were investigated using ex-situ XRD (Figure 11). 7x10 3 Fe 1.85kV Pt 600V 450ºC 6x10 3 Fe 1.85kV Pt 600V 510ºC Fe 1.85kV Pt 550V 600ºC 5x10 3 Fe 1.85kV Pt 500V 510ºC Fe 1.75kV Pt 500V 625º 4x10 3 Fe 1.85kV Pt 400V 510ºC Fe 1.85kV Pt 550V 450ºC 3x10 3 2x10 3 1x10 3 46 47 48 49 50 51 9x10 8x10 2 Fe 1.85kV Pt 600V 450ºC Fe 1.85kV Pt 600V 510ºC 7x10 2 Fe 1.85kV Pt 550V 600ºC Fe 1.85kV Pt 500V 510ºC 6x10 2 Fe 1.75kV Pt 500V 625º Fe 1.85kV Pt 400V 510ºC 5x10 2 4x10 2 3x10 2 2x10 2 1x10 2-1x10 2 0 20 22 24 26 Figure 11. XRD spectra for FePt films grown on (001) MgO. The microstructure of the films is currently being investigated with ex-situ cross sectional TEM. (Figure 12)
(a) (b) (c) (d) (e) Figure 12. (a) Cross sectional TEM image of an epitaxial FePt sample; (b) atomic resolution TEM image showing nano-regions with high chemical order (L1o); (c) FFT transform of the region in the HRTEM image that corresponds to the substrate; (d) FFT transform of the region in the same image that corresponds to the crystalline but chemically disordered FePt thin film; (e) FFT transform of the region in the image corresponding to the highly ordered region. We notice that there are intermediate spots in the four-fold symmetric pattern in both orthogonal directions. The preliminary TEM data seems to indicate that FePt has much better epitaxy than the FePd films (no tilting or micro domains) and the grains are single domain L1o ordered. The magnetic properties of these films were investigated with Kerr magnetometry in the polar configuration. (Figure 13). We notice that despite the fact that we have obtained films with nano-regions with high chemical order (L1o), the anisotropy still favors inplane magnetization. Further studies on these films are in progress.
Theta Kerr (deg) 0.4 0.3 0.2 0.1 0.0-0.1-0.2 FePt Fe 1.85kV Pt 600V 450ºC Fe 1.85kV Pt 600V 510ºC Fe 1.85kV Pt 500V 510ºC Fe 1.85kV Pt 400V 510ºC Fe 1.85kV Pt 550V 600ºC Fe 1.85kV Pt 550V 450ºC Buffer metálico Fe 1.75kV Pt 500V 625ºC Fe 1.85kV Pt 475V 675ºC -0.3-0.4-20000 -10000 0 10000 20000 H (Oe) Figure 13. Polar Kerr magnetometry of epitaxial FePt films. 3. Ion-implantation to achieve highly anisotropic shallow nano-clusters A preliminary sample was prepared using ion implantation at THIA (Toledo Heavy Ion accelerator). The conditions for the ion-implantation were chosen to achieve a shallow (i.e. approximately topmost 50 Angstroms) deposition of Fe nano-clusters. The preliminary sample was (001) Si that was implanted with iron ions at an incident ion energy of 30 kev and a dose of 2.5x10 16 /cm 2. The sample was kept at room temperature during implantation. Afterwards, the sample was characterized with MOKE. A noisy and weak magnetic signal was observed. (Figure 14). 226.5 226.0 I (arb. units) 225.5 225.0-3 -2-1 0 1 2 3 H (koe) Figure 14. Longitudinal Kerr magnetometry of Si ion-implanted sample.
Pt (200) y0 0 ±0 xc1 46.56702 ±0 w1 0.33924 ±0.00086 A1 43886434.92813 ±184669.30282 xc2 46.80923 ±0.0058 w2 0.99727 ±0.00805 A2 19774263.77977 ±218155.79233 FePt (002) xc3 48.09 ±0 w3 1.60735 ±0.0444 A3 7168165.03003±179956.77938 0 6 8 10 12 14 16 18 We have applied the ion implantation process to epitaxial Pt films grown on (001) MgO. Thus, a Pt thin film sample, epitaxially grown on (001) MgO was implanted under similar conditions than the Si sample but the planned dose is 5.0x10 16 /cm 2 in order to enhance the magneto-optical signal. After ion implantation, the sample was annealed at 450 o C for one hour. Correlated studies between structure and magnetic properties are currently in progress. Figure 15 shows polar Kerr hysteresis loops for the implanted sample before and after annealing. 1 annealed non annealed Kerr rotation (a.u.) 0-1 -10000 0 10000 H(Gauss) Figure 15. Polar Kerr hysteresis loops for ion-implanted Pt film. 8 10 Pt (200) 10 7 10 6 10 5 FePt (001) Symetric Scan after 1h annealing FePt (002) FePt (003) I(cps) y0=2.13209±0.04962 y0=3.299±0.06879 xc1=82.34075±0.01898 16 xc=68.36481±0.00497 w1=0.39415±0.04399 w=0.36394±0.01067 A1=1.07713±0.10651 A=4.17511±0.11913 xc2=82.73212±0.00289 w2=0.12659±0.00627 A2=1.13748±0.06328 12 y0=2.14806±0.07678 xc1=73.53981±0.19545 y0=2.69115±0.08248 w1=2.70748±0.30555 xc=86.92498±0.00833 A1=6.58071±0.90731 w=0.52849±0.01893 y0=2.86757±0.16364 xc2=76.69121±0.08522 A=4.54874±0.1699 xc=62.25838±0.01831 w2=2.38495±0.13216 8 w=0.23801±0.0451 A2=11.51564±0.98006 A=0.62374±0.13275 4 10 4 10 20 30 40 50 60 70 80 90 FePt (001) Symetric FePt (001) 10 2 Pt (111) FePt (001) 0 60 65 70 75 80 85 90 y0 9.58268 ±0.24336 xc1 21.47464 ±0.01552 w1 0.82865 ±0.03104 A1 28.88558 ±1.14813 FePt (001) xc2 23.51328 ±0.01177 w2 1.7374 ±0.02753 A2 118.03957 ±1.76426 Rocking FePt (001) y0 3 ±0 xc 12.3808 ±0 w 3.069 ±0.066 5x10 A 178.57066 ±3.29302 4x10 1 3x10 1 2x10 1 10 1 1x10 1 16 18 20 22 24 26 28 FePt (002) Symetric Scan FePt (002) 10 8 5x10 2 Rocking FePt (002) Pt (200) 4 1.5x10 Rocking Pt (200) 10 7 10 6 FePt (002) 4x10 2 3x10 2 2x10 2 y0 1 ±0 xc 24.3093 ±0 w 0.59963 ±0.00447 A 322.38911 ±2.08065 1.0x10 4 5.0x10 3 y0 21 ±0 xc 23.5478 ±0 w 0.59633 ±0.00546 A 8604.14584 ±68.2252 10 5 1x10 2 10 4 0 0.0 50 Asymmetric Scan asimétrico Scan 18 20 22 24 26 28 30 10 5 Phi scans 22 23 24 25 ScanAxis (FirstAngle) MgO Asymetric scan 3 2.0x10 y0 0 ±0 xc1 68.24981 ±0.00166 1.5x10 3 w1 0.60698 ±0.00451 A1 1346.175 ±19.3728 xc2 68.66617 ±0.03182 w2 1.41928 ±0.03512 1.0x10 3 A2 423.94789 ±20.67852 10 4 10 3 10 2 5.0x10 2 10 1 0.0 66 67 68 69 70 71 10 0-50 0 50 100 150 200 250 300 350 400 Figure 16. XRD charts showing the formation of the highly ordered L1o phase.
13µm 13µm Figure 17. Left: Topographic AFM and right: MFM image of the Fe implanted Pt thin film sample after UHV annealing. Structural characterization (Figure 16) as well as surface morphology measurements (Figure 17) indicate that the Fe implanted Pt thin film sample has formed small L1o clusters near the surface with perpendicular magnetization. Further studies are in progress to determine the effect of rapid thermal annealing (RTA) on ion-implanted samples, to obtain smaller clusters.