Surface morphology of the MgO layer and magnetic properties of FePt-C layer deposited on the MgO were investigated for the FePt-based heat assisted magnetic recording media. Stacking structure of the underlayer for the FePt-C layer was MgO (0-5 nm)/Cr80Mn20 (0-30 nm)/Cr50Ti50 (0-50 nm)/glass sub.. Surface observation result for the MgO film by using an atomic force microscope revealed the existence of nodules with a height of about 2 nm and a network-like convex structure with a height difference of about sub nm (boundary wall, BW) on the MgO crystal grain boundary. Density of the nodules largely depends on the surface roughness of the CrTi layer, RaCrTi and it is suppressed from 10 to 2/0.5 μm2 by reducing RaCrTi from 420 to 260 pm. Height of the BW depends on thickness of the MgO layer, tMgO and it can be suppressed by reducing tMgO to less than 4 nm. From the cross-sectional energy dispersive x-ray mapping, it is clarified that the BW is formed by atomic diffusion of Ti atoms from CrTi layer due to the substrate heating process, and a compound consists of Mg, Ti and O atoms. This BW can be used as a template to magnetically isolate the FePt column in the FePt-based granular film, such as FePt-SiO2, if the size of the BW is reduced to less than 10 nm. M-H loop of the FePt-C granular film deposited on the underlayer showed that the nodule and BW induce oxidation of the FePt grains, and reduction of intergranular exchange coupling.
I. INTRODUCTION
Heat assisted magnetic recording (HAMR) is a promising candidate to overcome the trilemma issue of perpendicular magnetic recording hard disk. For this magnetic recording medium, L10 - ordered FePt alloy is used for the recording layer to overcome the thermal agitation by its high uniaxial magnetocrystalline anisotropy (Ku). To obtain the high Ku FePt grains, heteroepitaxial growth of the FePt on an MgO underlayer (UL) is very important to realize low c-axis angular distribution FePt grains. Previously, we reported that low angular distribution of 002 axis of the MgO grains was realized by depositing the MgO layer on a seed layer of Cr80Mn20/CrTi.1 It was pointed out that one of the current issues of the HAMR media was angular distribution of c-axis of the L10 FePt grains, which came from grain growth at grain boundaries of the MgO UL.2–6 However, it is still unclear how the surface structure at the grain boundary of the MgO UL affects the c-axis angular distribution. In this study, we have investigated the relation between the deposition process of the MgO UL and its surface morphology to make guidelines for realizing the optimum MgO surface morphology.
II. EXPERIMENTAL PROCEDURE
All samples used in this study were fabricated by using in-line magnetron sputtering system (Canon Anelva C-3010). The stacking structures of the samples were composed of MgO (0-5 nm)/Cr80Mn20 (0-30 nm)/Cr50Ti50 (0-50 nm)/nanocrystalline glass substrate. To induce the lateral (002 plane) growth of bcc-Cr80Mn20 seed layer, substrate temperature was elevated from RT to 510, 400 oC before deposition of the CrMn seed layer and the MgO layer, respectively, to expand the grain diameter.1 CrTi amorphous texture induce layer (TIL) was fabricated to realize the lateral growth of the seed layer.1 A magnetic layer of Fe50Pt50-40vol% C (5 nm) was also deposited on the samples to evaluate the magnetic properties. The base pressure of all chambers was kept at approximately 5×10-6 Pa. Surface structure analysis was carried out by an atomic force microscopy (AFM). Crystal structure was evaluated by an X-ray diffractometry (XRD). Composition distribution of the film was observed by cross-sectional energy dispersive x-ray (EDX) mapping. Magnetization curve was measured by a vibrating sample magnetometer using superconducting magnet.
III. RESULTS
A. Surface structure of the underlayer
Figures 1 show the surface morphology of (a) MgO/CrMn/CrTi/sub., (b) CrMn/CrTi/sub., (c) CrTi/sub., and (d) glass sub. For (a), nodules with a height of about 2 nm are observed at the surface. And a network-like convex structure with a height difference of about sub-nm (boundary wall, BW), mesh size of around 40 to 80 nm is also observed at the surface. Position of the BW is thought to correspond to the grain boundary of the MgO grains, considering that valley on the grain boundary is not observed at the surface. For (b), approximately 24 nodules are observed in the area of 1 μm2. On the other hand, for (c) and (d), nodule cannot be confirmed on the surfaces. Therefore, it is understood that the nodules are generated on the CrMn seed layer. Based on this result, the nodule is thought to grow on the small irregular protrusions in the CrTi surface. And the BW cannot be observed in (b), (c), (d). This result suggests that the BW exists on the MgO layer.
AFM images of the surface of (a) MgO/CrMn/CrTi/sub., (b) CrMn/CrTi/sub., (c) CrTi/sub., and (d) sub..
AFM images of the surface of (a) MgO/CrMn/CrTi/sub., (b) CrMn/CrTi/sub., (c) CrTi/sub., and (d) sub..
B. Deposition condition and surface structures
Next, relationship between the nodules and the BW and structure of its underlying layer was investigated. For the nodule, dependence of its occurrence frequency on roughness of the CrTi TIL, RaCrTi was confirmed. RaCrTi was varied by changing input power during CrTi film deposition between 500 and 1500 W, because generally high-power sputtering process induces the amorphous film due to its high-sputtering rate. For the BW, dependence between its height on thickness of the MgO, tMgO and cross-sectional composition map were confirmed.
Figures 2 show (a) relation between nodule density and RaCrTi and (b) BW height as a function of tMgO. For (a), nodule density clearly decreased from 10 to 2/0.5 μm2 as RaCrTi decreases from 420 to 260 pm. This result suggests that amorphization of the TIL reduces the density of the small protrusion on the TIL. For (b), the BW height took 1.6 nm for tMgO of 7 nm. And it decreases with decreasing tMgO, and took 0 nm for tMgO of less than 4 nm. This result indicates that the BW height depends on tMgO and it will not be formed for tMgO less than 4 nm.
(a) Relation between nodule density and surface roughness of CrTi TIL RaCrTi and (b) boundary wall height as a function of MgO thickness tMgO.
(a) Relation between nodule density and surface roughness of CrTi TIL RaCrTi and (b) boundary wall height as a function of MgO thickness tMgO.
Figures 3 show cross-sectional (a) high-angle annular dark field scanning transmission electron microscopy image and EDX-mapping images for (b) Mg, (c) Cr, and (d) Ti atom for MgO (5 nm)/CrMn (30 nm)/CrTi (50 nm)/sub.. For (a), a line is observed in the CrMn layer portion. And convex structure is also observed on the line. According to the result on section III A, this convex structure is thought to correspond to the BW. For (b), the convex structure is also observed. This result suggests that the structure is composed of Mg. For (c), brightness is weak on the line in the CrMn layer portion. This shows that the position of the line corresponds to grain boundary of the CrMn layer. For (d), bright portion is surprisingly observed at the CrMn grain boundary, surface of the CrMn layer, and the convex structure. Position of the Ti atom is overlapped with position of initial portion of the MgO layer by approximately 2.5 nm. Concentrations of Ti are approximately 5 at.% at above positions. These results suggest that Ti atoms diffuses from CrTi layer to the surface of CrMn layer, and forms a convex structure which consists of a compound with Mg, Ti and O atoms. Driving force of the Ti diffusion is most likely the substrate heating at 400 oC before the MgO deposition. Thus, the size and height of the BW can be controlled by changing the size of the grain diameter of the CrMn seed layer, amount of the Ti atoms, and substrate temperature before MgO deposition. If the BW with size of less than 10 nm was realized, it could be used for a template to magnetically isolate the FePt column in the FePt-based granular film, such as FePt-SiO2.7
Cross-sectional (a) HAADF-STEM image and EDX-mapping images for (b) Mg, (c) Cr, and (d) Ti atom for the sample of MgO (5 nm)/CrMn (30 nm)/CrTi (50 nm)/sub..
Cross-sectional (a) HAADF-STEM image and EDX-mapping images for (b) Mg, (c) Cr, and (d) Ti atom for the sample of MgO (5 nm)/CrMn (30 nm)/CrTi (50 nm)/sub..
C. Magnetic properties of FePt based magnetic layer deposited on the BW
FePt-based granular film was deposited on the underlayer. To maintain the surface morphology of the nodule and BW, substrate temperature was set to 410 oC, which is the maximum temperature of the nodule in the process. Figure 4 shows M-H loops for the FePt-40vol%C media with various underlayer deposited at 410 oC. Black, blue, and red lines correspond to the underlayer without nodule and BW, with nodule, and with BW, respectively. All the loops take a shape in which perpendicular component and in-plane component were mixed in almost the same amount. This shape indicates the mixture of low and high Ku grains due to low degree of ordering of the FePt grains by low substrate temperature deposition. Saturation magnetization and coercivity of the red and blue loops were approximately 20% lower and 2 times higher than that of the black loop, respectively. This result suggests that the nodule and BW induce oxidation of the FePt due to increase of contact area with MgO and FePt, and reduction of intergranular exchange coupling due to the surface topology of the nodule and the BW.
M-H loops for the FePt-40vol%C media with various underlayer deposited at substrate temperature of 410 oC.
M-H loops for the FePt-40vol%C media with various underlayer deposited at substrate temperature of 410 oC.
IV. SUMMARY
Surface morphology of the MgO layer and magnetic properties of FePt-C layer deposited on the MgO were investigated for the FePt-based heat assisted magnetic recording media. Stacking structure of the underlayer for the FePt-C layer was MgO (0-5 nm)/Cr80Mn20 (0-30 nm)/Cr50Ti50 (0-50 nm)/glass sub.. Surface observation result for the MgO film by using an atomic force microscope revealed the existence of nodules with a height of about 2 nm and a network-like convex structure with a height difference of about sub nm (boundary wall, BW) on the MgO crystal grain boundary. Density of the nodules largely depends on the surface roughness of the CrTi layer, RaCrTi and it is suppressed from 10 to 2/0.5 μm2 by reducing RaCrTi from 420 to 260 pm. Height of the BW depends on thickness of the MgO layer, tMgO and it can be suppressed by reducing tMgO to less than 4 nm. From the cross-sectional energy dispersive x-ray mapping, it is clarified that the BW is formed by atomic diffusion of Ti atoms from CrTi layer due to the substrate heating process, and a compound consists of Mg, Ti and O atoms. This BW can be used as a template to magnetically isolate the FePt column in the FePt-based granular film, such as FePt-SiO2, if the size of the BW is reduced to less than 10 nm. M-H loop of the FePt-C granular film deposited on the underlayer showed that the nodule and BW induce oxidation of the FePt grains, and reduction of intergranular exchange coupling.