Aberration-Corrected Transmission Electron Microscopy of the

Here we present the first images of a magnetic recording medium produced using a spherical aberration-corrected TEM showing the true amorphous IP stru...
1 downloads 0 Views 2MB Size
Letter pubs.acs.org/NanoLett

Aberration-Corrected Transmission Electron Microscopy of the Intergranular Phase in Magnetic Recording Media Faraz Hossein-Babaei,*,† Ai Leen Koh,† Kumar Srinivasan,‡ Gerardo A. Bertero,‡ and Robert Sinclair† †

Department of Materials Science and Engineering, Stanford University, California 94305-4034, United States Western Digital Corporation, 1710 Automation Parkway, San Jose, California 95112, United States



ABSTRACT: In perpendicular hard disk memory media, nanometric magnetic Co-rich grains are separated by a ∼1 nm thick nonmagnetic and preferably amorphous intergranular phase (IP). Attempts at observing the IP structure at high resolution using TEM have been obstructed by the superposition of lattice fringes from the crystalline grains extending into the IP region in images. Here we present the first images of a magnetic recording medium produced using a spherical aberration-corrected TEM showing the true amorphous IP structure in contrast to the crystalline grains, allowing the accurate determination of the grain−IP interface and the grain and IP dimensions. It is shown that these aberration-corrected TEM images are functionally superior for analyzing certain features of the ultrahigh capacity data recording media. KEYWORDS: PMR media, nanostructure, intergranular phase, high-resolution TEM, delocalization, spherical aberration correction

E

not have a significant contribution in the SAED pattern other than producing a diffuse background. As observed in the planview transmission electron microscope (TEM) images of the ML (see Figure 2), the IP region has a width of about 1 nm on average and serves to control the jitter noise by reducing exchange coupling between the ML grains,2,3,16 which leads to reduction of the magnetic cluster size.2,17 Therefore, proper characterization of the IP structure is of critical importance. Because of the nanometric dimensions involved, the TEM is the only suitable tool to study the IP structure. However, when imaged in high-resolution mode using conventional TEMs, artifacts caused by contrast delocalization thwart proper discrimination of the IP from the ML crystalline grains.18−20 As a result of this effect, the lattice fringes of the crystalline grains extend beyond their boundaries into the IP, making the grain−IP interface obscure. In Figure 3, for instance, the {1010̅ } lattice fringes are visible extending from the grain into the IP. Recent advancements in spherical aberration (Cs) corrected TEMs have helped compensate for many of the instrumental limitations responsible for the delocalization effect.18−21 There have been studies trying to eliminate the contrast delocalization in images obtained by spherical aberration correction of the TEM.18,19 Here, we present the first aberration-corrected TEM image of the ML nanostructure in high resolution depicting no noticeable delocalization artifact. The image confirms the amorphous structure of the IP and facilitates the proper characterization of its structure. The presented image allows the identification of the IP boundaries. Compared to conventional high-resolution TEM images, the produced image affords a more precise analysis of

ver increasing demand for higher storage capacity memory devices drives the search for media of higher data storage densities.1−4 Presently, high storage density commercial hard disks share the same basic structure comprising a magnetic layer (ML) with grains smaller than 10 nm separated by an intergranular phase (IP).3,5−8 The ML structure of a common perpendicular magnetic recording (PMR) medium is schematically shown in Figure 1 in cross section. The “write-head” orients the magnetization of grains located in rectangular regions in the ML uniformly to one direction perpendicular to the hard disk surface defining bits of data physically delineated by magnetic transition regions.3,4 Owing to the finite ML grain dimensions, it is thought that the transition region boundaries are not straight, but along the grain boundaries.2 Specified by both the IP width and grain size, this imperfection causes jitter noise during data reading.3,4 To obtain read signals of adequate signal-to-noise ratio (SNR), the ML grains are required to be small in relation to the physical bit size and to have a narrow size distribution.2−4,9 The ML studied in this work is made of a CoCrPt-TiO2 alloy with cobalt-rich magnetic grains and an oxygen-rich nonmagnetic IP, which is ideally of a different phase and composition than the ML grains.3,4,10,11 (The IP is referred to as the “grain boundary” in the background literature.)2−4 Co and Pt are the main constituents of the HCP-structured crystalline grains, as depicted in the selected area electron diffraction (SAED) pattern given in Figure 2, while Ti and O are segregated to the IP during grain growth.10,12,13 The strongly favored growth direction of the grains creates a texture whereby c-axes of all grains are nearly perpendicular to the disk surface.3,9,14 This is clear from the SAED pattern shown in Figure 2 which is produced solely by the crystal plane systems belonging to the [0001] zone, i.e., the {hki0} systems. The IP, on the other hand, is believed to be amorphous4,13,15 and would © 2012 American Chemical Society

Received: March 5, 2012 Published: April 23, 2012 2595

dx.doi.org/10.1021/nl301274x | Nano Lett. 2012, 12, 2595−2598

Nano Letters

Letter

Figure 1. Schematic diagram showing the top layers of the layer stack of a CoCrPt-TiO2 alloy PMR medium. Different layers deposited help produce the desired ML nanostructure or protect it from the environment.10

Figure 2. Bright-field (BF) diffraction contrast TEM image of ML in a CoCrPt-TiO2 PMR media plan-view specimen. The SAED at top right of the figure is produced from contribution of plane systems in the [0001] zone of both the ML and the Ru seed layer present in the specimen showing their HCP crystal structures.

Figure 3. High-resolution phase contrast TEM image of ML in the CoCrPt-TiO2 PMR media specimen obtained using the Tecnai TEM taken near Scherzer defocus (−60 nm relative to Gaussian focus at Cs of about 1 mm), while the microscope was operated at 200 kV beam acceleration. The instrument information limit is 0.14 nm. The superposition of the lattice fringes of {101̅0} type planes can be seen as bright spots in most granular regions of the image. The lattice fringes observed in IP are due to the delocalization effect of the TEM.

the ML nanostructure and reveals features not previously identifiable. The CoCrPt-TiO2-based PMR medium sample is produced with a well established fabrication technique: A hard Ni−P coating is electrodeposited on the Al−Mg alloy substrate. The upper thin films of the recording medium are deposited in a series of 20 DC magnetron sputter deposition chambers. These layers include a 200 nm thick soft magnetic under-layer, a ∼5 nm thick amorphous Ta layer, a Ni−W layer, and two 10 nm Ru layers grown at different ambient pressures. The ML is deposited on the Ru layer, and a diamond-like amorphous carbon layer coats the ML to protect it from the environment and improve the tribological properties of the surface. The deposition of an additional lubricant common to commercial products is omitted in these experiments. We have shown that the ML grain structure produced in this way has a one-to-one agreement with the Ru seed layer grains.10 Plan-view TEM specimens are prepared using a conventional method:10,22 After the initial grinding and polishing of each sample from the back side, standard 3 mm diameter TEM specimen disks are punched out. These samples are then further ground using a Gatan-656 dimple grinder from the same side to reach a thickness of less than 15 μm at the center.

The specimens are then ion milled with Ar+ ions at 4−5 keV at an incident angle of 4°−5° using a Gatan-691 precision ion polishing system. Electron transparent regions in the 5−15 nm thickness range found near holes perforated during the ion milling are used for TEM studies. The total thickness of the ML, 10 nm, is directly determined by measurements on crosssectional specimens. The appearance of Moiré fringes in the plan view indicates the presence of both the ML and the Ru seed layer10 and helps locate regions of about 10 nm thickness which are used for the analysis of all samples for the sake of consistency. The same specimen was used for both conventional and aberration-corrected microscopy. Conventional high-resolution imaging of the specimen was carried out using FEI Tecnai G2 F20 at 200 kV, while the Cs aberration-corrected highresolution images were obtained with FEI Titan 80−300 at 300 kV, both equipped with a field emission gun. The highresolution phase contrast images are given in Figures 3 and 4, respectively. Captions of the figures describe the imaging conditions used. Resolutions of both microscopes are sufficient to illustrate the crystalline structures of the grains in highresolution phase contrast images at high magnification in the 2596

dx.doi.org/10.1021/nl301274x | Nano Lett. 2012, 12, 2595−2598

Nano Letters

Letter

Figure 4. High-resolution phase contrast TEM image of the ML in the CoCrPt-TiO2 PMR media specimen obtained using the Titan TEM at about 10 nm defocus relative to the Gaussian focus with a Cs value of about −10 μm after Cs aberration correction operated at a 300 kV beam acceleration voltage. The instrument information limit is 0.07 nm. The amorphous structure of the IP is clear and can be studied more reliably against the crystalline grains.

surrounding IP region.18−20 In Figure 4, the physical boundaries of the magnetic grains are accurately determinable, and hence, the measurements of the actual dimensions of the grains and the thickness of the IP in various locations can be achieved. The delocalization-eliminated TEM image (Figure 4) reveals nanostructural features not possible to identify using the conventional TEM (Figure 3). It is clear from Figure 4 that the IP is largely amorphous. Nanocrystalline bridging, however, is observed between some of the neighboring grains over the IP; examples are indicated by arrows in Figure 5. Such connections may increase the magnetic coupling between the grains and hinder their independent reversal of magnetization.2 The grains appear faceted. While the faceting tendency of the grains is apparent in the BF TEM images as well (see Figure 2), the aberration-corrected TEM in high resolution (see Figure 4) yields clear evidence of the facets. The grain facets indicated in Figure 5 with line segments, all along {101̅0} and {112̅0} type planes, are evidence of faceting along prominent plane systems of the perpendicularly textured HCP structured crystals in the ML. The comparison of the observed facets of the magnetic grains in different ML structures may lead to establishing methods for controlling magnetic grain morphology in the ML. A spherical aberration-corrected TEM was used to observe the ML nanostructure in a CoCrPt-TiO2-based PMR medium. The obtained high resolution image, compared with that obtained using a conventional TEM, confirms the amorphous structure of the intergranular phase and enables clear identification of the boundaries between neighboring crystalline

form of periodic lattice fringes (see Figures 3 and 4). Figure 3 is taken near the Scherzer defocus (at about −60 nm defocus relative to Gaussian focus for this TEM instrument) to reduce the delocalization effect. The {101̅0} lattice fringes of the neighboring magnetic grains are shown as overlapping into the IP region because of the delocalization effect impeding judgment on the position of the grain−IP interface and establishing the true structure of the IP. This artifact prevails even when examining specimens as thin as 5 nm. BF TEM images (see Figure 2, for example) provide resolution levels required for grain size measurements but are insufficient to study the dimensions and structure of the grain− IP interface in the ML in detail. While the delocalization effect is not an issue for diffraction contrast BF images, this shortcoming is partly because of the fuzzy contrast prevailing at the grain−IP interface and partly due to the lack of the atomic structure information in these images. High-resolution imaging using the aberration-corrected TEM, on the other hand, allows accurate measurement of the grain dimensions as well as the widths of IP regions to the atomic scale. In Figure 4, the delocalization effect is eliminated at the atomic scale by better instrument conditions and Cs aberration correction. The delocalization effect arises due to the spherical aberration of the microscope in high-resolution images and is affected by instrumental parameters, such as the electron beam coherence, effective source size, and defocus. For example, the effect is less pronounced when using a LaB6 electron gun than a field-emission electron gun TEM.20,23,24 Owing to this effect, the lattice fringes of a crystalline grain superpose over the 2597

dx.doi.org/10.1021/nl301274x | Nano Lett. 2012, 12, 2595−2598

Nano Letters

Letter

(10) Hossein-Babaei, F.; Sinclair, R.; Srinivasan, K.; Bertero, G. A. Nano Lett. 2011, 11, 3751−3754. (11) Nolan, T. P.; Risner, J. D.; Harkness, S. D., IV; Girt, E.; Wu, S. Z.; Ju, G.; Sinclair, R. IEEE Trans. Magn. 2007, 43, 639−644. (12) Risner, J. D.; Sinclair, R.; Bentley, J. J. Appl. Phys., 2006, 99, 033905. (13) Wittig, J.; Al-Sharab, J.; Bentley, J.; Evans, N. Microsc. Microanal. 2003, 9, 482−483. (14) Hirayama, Y.; Honda, Y.; Kikukawa, A.; Futamoto, M. J. Appl. Phys. 2000, 87, 6890−6892. (15) Zheng, M.; Acharya, B. R.; Choe, G.; Zhou, J. N.; Yang, Z. D.; Abarra, E. N.; Johnson, K. E. IEEE Trans. Magn. 2004, 40, 2498−2500. (16) Yin, J.; Zhang, H.; Hu, F.; Shen, B.; Pan, L. Q. J. Appl. Phys., 2009, 106, 103901. (17) Wu, X. W.; Peng, Y.; Wang, B. Appl. Phys. Lett., 2008, 93, 152514. (18) Sato, K.; Konno, T. J.; Hirotsu, Y. J. Appl. Phys., 2009, 105, 034308. (19) Hu, X.; Xie, L.; Zhu, J.; Poudyal, N.; Liu, J. P.; Yuan, J. J. Appl. Phys., 2009, 105, 07A723. (20) Otten, M. T.; Coene, W. M. J. Ultramicroscopy 1991, 48, 77−91. (21) Erni, R.; Rossell, M. D.; Nakashima, P. N. Ultramicroscopy 2010, 110, 151−161. (22) Risner, J. D.; Nolan, T. P.; Bentley, J.; Girt, E.; Harkness, S. D., IV; Sinclair, R. Microsc. Microanal. 2007, 13, 70−79. (23) Lentzen, M.; Jahnen, B.; Jia, C. L.; Thust, A.; Tillmann, K.; Urban, K. Ultramicroscopy 2002, 92, 233−242. (24) de Jong, A. F.; Van Dyck, D. Ultramicroscopy 1993, 49, 66−80.

Figure 5. Same as Figure 4 with arrows pointing at the observed crystalline bridges between neighboring grains and line segments indicating grain facets and atomic steps.

and amorphous phases. Aberration-corrected high-resolution TEM images revealed details in the ML nanostructure which were not identifiable in conventional TEM images. These additional details afford the capacity both to observe key features, such as the faceting of the magnetic nanocrystals along prominent crystallographic planes of the ML and the formation of crystalline nanobridges connecting neighboring grains, and to measure the IP thickness at the atomic scale with unprecedented accuracy; advantages which may prove important in the refinement of the related technology. The described method may be employed to allow the correct observation of other multiphase nanostructures in general.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]; [email protected] Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The authors would like to thank Western Digital Media Inc. for financial support and providing the equipment for preparation of the samples.



REFERENCES

(1) Piramanayagam, S. N.; Srinivasan, K. J. Magn. Magn. Mater. 2009, 321, 485−494. (2) Judy, J. H. J. Magn. Magn. Mater. 2001, 235, 235−240. (3) Piramanayagam, S. N. J. Appl. Phys. 2007, 102, 011301. (4) Moser, A.; Takano, K.; Margulies, D.; Albrecht, M.; Sonobe, Y.; Ikeda, Y.; Sun, S.; Fullerton, E. J. Phys. D: Appl. Phys. 2002, 35, R157− R167. (5) Srinivasan, K.; Piramanayagam, S. N.; Chantrell, R. W.; Kay, Y. S. J. Magn. Magn. Mater. 2008, 320, 3036−3040. (6) Jung, H. S.; Kwon, U.; Kuo, M.; Velu, E. M. T.; Malhotra, S. S.; Jiang, W.; Bertero, G. IEEE Trans. Magn. 2007, 43, 615−620. (7) Wang, J.-P.; Shen, W.; Hong, S. IEEE Trans. Magn. 2007, 43, 682−686. (8) Acharya, B. R.; Zhou, J. N.; Zheng, M.; Choe, G.; Abarra, E. N.; Johnson, K. E. IEEE Trans. Magn. 2004, 40, 2383−2385. (9) Kwon, U.; Sinclair, R.; Velu, E. M. T.; Malhotra, S.; Bertero, G. IEEE Trans. Magn. 2005, 41, 3193−3195. 2598

dx.doi.org/10.1021/nl301274x | Nano Lett. 2012, 12, 2595−2598