Two-inch-Diameter Magneto-optical Disk System with 3 GB Capacity and 24 Mbps Data Transfer Rate Using a Red Laser

Transcription

1 Trans. Magn. Soc. Japan, 2, (2002) Two-inch-Diameter Magneto-optical Disk System with 3 GB Capacity and 24 Mbps Data Transfer Rate Using a Red Laser Masahiro Birukawa*, Yasumori Hino*, Keiji Nishikiori*, Kiyoshi Uchida*, Tsutomu Shiratori**, Tomoyuki Hiroki**, Yasuyuki Miyaoka** and Yasushi Hozumi** *Storage Media Systems Development Center, Matsushita Electric Industrial Co.,Ltd. 1006, Kadoma, Kadoma City, Osaka, , Japan ** DM Projecț Core Technology Development Headquarters, Canon Inc. 30-2, Shimomaruko 3-Chome, Ohta-ku, Tokyo, , Japan Abstract - High-density recording four times that of a current DVD-RAM has been achieved by applying Domain Wall Displacement Detection (DWDD) technology using a conventional optical head which has the same wavelength of 660 nm and NA of 0.6 as the DVD head. It has been confirmed by using a sampled servo and PRML that a recording density of 15 Gbit/inch2 with a track pitch of 540 nm and bit length of 80 nm has sufficient system margin. This means that a capacity of 3 GB on a two-inch-diameter disk is now available for portable use. If applied to a digital video camera, one hour of MPEG2 video contents can be recorded at a transfer rate of 6 Mbps. Key words: magneto-optical recording, domain wall displacement detection (DWDD), high density recording, sampled servo, annealing, system margin, digital video camera 1. Introduction Table 1 Specifications As high-density optical disk technology progresses, a small-size disk, enabling mobile usage, has been awaited. A two-inch-diameter disk called "id-photo" with a capacity of 730 MB [1] was introduced to the market last year. Using the same diameter, a 1 GB-capacity disk using a red laser and a 2 GB disk using a blue laser have since been proposed by applying advanced Center Aperture Detection (CAD) to a magneto-optical (MO) disk [2,3]. Although video compression technology has progressed, a transfer rate of 6 Mbps or more is needed for high quality motion pictures, combined with a capacity of 3 GB or more to achieve one hour of recording time. The most important and advantageous feature of Domain Wall Displacement Detection (DWDD) technology [4] is the potential for extremely high linear recording density, which makes it possible to achieve higher transfer rates on a small-size disk at comparatively lower rotating speeds. This is a key factor when realizing a mobile disk system to reduce the power consumption and track offset of the system. In this paper, high-density recording, achieved by combining DWDD technology and the sampled servo method is discussed in the context of the endeavor to realize 3 GB of capacity on a two-inch-diameter disk. 2. Keys to High Density Recording 2.1 Disk Structure and Specifications The specifications of the proposed disk are shown in Table 1. The disk diameter is 50.8 mm. The read/write area is 11.5 `23.8 mm in radius. The areal recording density of the disk is approximately 15 Gbit/inch 2, i. e. 80 nm/bit and 540 nm/track. In this configuration, laser-pulsed MFM Figure 1 Schematic film structure Trans. Magn. Soc. Jpn., Vol. 2, No. 4,

2 (a) With laser annealing Figure 2 AFM image of a groove track provided on a polycarbonate substrate. (Magnetic Field Modulation) has been employed to record (1,7) RLL-coded data using a conventional magneto-optical head (Ď = 660 nm and NA = 0.6) through a 0.6 mm-thick substrate. Recorded data is reproduced by a PR(1,-1)+Viterbi method. The user data transfer rate of 24 Mbps is achieved. The DWDD film structure previously proposed by Shiratori et al. [4,5] as shown in Figure 1 has been adopted. Data is stored in the memory layer. When a readout beam is irradiated on the disk during readout operation, the magnetic domain walls in the readout layer are displaced by cutting off the exchange- coupling from the memory layer to the readout layer. The control layer suppresses ghost signals. These films are sandwiched by SiN protective layers. Figure 2 shows an AFM image of the substrate. Data is recorded in a grooved area formed on the polycarbonate substrate. The track pitch and the groove depth are 540 nm and 40 nm, respectively. For smooth domain wall motion, the groove has been made as wide and smooth as possible. The groove width is about 400 nm in the flat portion and the roughness of the groove is nm in Ra. Wobble pits for sampled servo tracking are also provided on the disk, since the groove pitch is too narrow to obtain a conventional push-pull signal for tracking. Figure 3 Simulation result of energy absorption when annealing beam with Ď = 405 nm and NA = 0.85 is irradiated from film side. (b) Without laser annealing Figure 4 Magnetic Force Microscope images of 80nm/bit random marks on a DWDD disk. 2.2 Annealing Process The land part, which corresponds to the inter-track area, has to be annealed with an annealing laser beam to ensure smooth wall motion of the magnetic domain. When the beam is irradiated in a convex area as seen by the beam source, the light energy of the beam is concentrated on the corners of the convex area [6,7]. Figure 3 shows a simulation result of energy absorption based on finite-difference time-domain (FDTD) method using vector diffraction calculation. As seen in this Figure, when the beam with a wavelength of 405 nm and NA of 0.85 is irradiated from the film side, most of the light energy is absorbed at both edges of the land part. This result suggests that the irradiation of the beam on the land part from the film side makes it possible to have an annealing width narrower than the beam spot. At the annealed area, it can be predicted that the direction of magnetization easy axis has been changed from perpendicular to in-plane, since the magnetic anisotropy constant Ku has been decreased by annealing. This prevents a perpendicular magnetic domain from being present in the annealed area. Figure 4 shows Magnetic Force Microscope images of 80 nm/bit random marks on the DWDD disk, with laser annealing (a), and without laser annealing (b). As seen in the Figure, it is confirmed that the stripe-shaped unrecordable area has been formed at the inter-track area by the annealing process. The net width of recordable tracks unavoidably narrows with increased annealing power. Figure 5 (a) shows the dependence of the net width of recordable tracks on the annealing power (Pa). The hatched part represents the unrecordable area. On the right side of the Figure, the cross-section of a groove, as measured by an AFM, is illustrated, corresponding to the track width. Jitter of the 2T signal is also shown in Figure 5 (b). The land part has been annealed with an optical head with a wavelength of 405 nm and NA of 0.85, at a scanning 274 Trans. Magn. Soc. Jpn., Vol. 2, No. 4, 2002

3 heat source by converging the light energy of the beam. As a result, a very narrow annealing area, less than 150 nm in width, can be formed reliably. Figure 5 (b) Annealing power dependence of (a) recordable track width and (b) jitter. velocity of 3 m/sec. The spot width (FWHM) is 280 nm. Annealing starts at a Pa of 3.4 mw. When the Pa reaches 3.9 mw, the convex area of the land is completely annealed (dotted lines in the Figure), where the jitter is at a minimum. This is likely to be the most suitable condition for DWDD operation. In the region more than 4 mw, the annealing area extends to the flat part of the groove. The width of the annealed area is proportional to the annealing power. Jitter increases in this region with decreasing net track width. If an annealing laser beam is irradiated on a non-grooved disk, it may be difficult to form a much narrower annealing area than the spot size. But in this case, the land portion creates the effect of a more concentrated 2.3 Read Write Process The shape of a magnetic domain recorded using MFM is typically a crescent, since it is determined by the temperature contour during the cooling process. In the DWDD system, on the other hand, wall motion starts at the cutoff temperature of exchange-coupling force defined by the temperature contour during the heating-up process. Therefore, due to the mismatch between the two factors just described, a crescent-shaped magnetic domain with large curvature is not suitable for the DWDD system. This mismatch would have a disadvantageous effect on the system, not only by decreasing the driving force of the wall motion but also by varying the start position of the wall motion. In this system, the beams are irradiated onto convex areas as seen by the beam source, both in the annealing process on the land with the blue laser from the film side, and in the read/write process in the groove with the red laser through the substrate. Figure 6 (a) shows a simulation result of energy absorption when an optical beam with a wavelength of 660 nm and NA of 0.6 is irradiated through the substrate, and (b) shows the temperature profile as temperature contours. As seen in the Figure, light energy is mostly absorbed at the edges of the groove, and the temperature contour resembles a rectangle rather than an ellipse in shape. This means that the groove sandwiched by the walls of the land helps to form rectangular magnetic domains. In addition, the recording sensitivity at the track edge is higher than that at the track center, because the annealing process appears to reduce the Ku at the track edge. Because of the above two effects, even if a conventional MFM method is applied for recording, rectangular magnetic domains are formed as shown in Figure 4 (a). The recorded domain shapes are the same as if they were written using conventional HDD magnetic recording. Other advantages of concentration of light energy on (a) (b) Figure 6 (a) simulation result of energy absorption when an optical beam with a wavelength of 660 nm and NA of 0.6 is irradiated through the substrate; (b) temperature profile indicated as temperature contours. Trans. Magn. Soc. Jpn., Vol. 2, No. 4,

4 Normalized Annealing Power figure 7 Annealing power dependence of ber the convex area are the suppression of crosswrite and crosstalk during the write and read operations, respectively. By decoding the readout signal using the PR(1,-1)+Viterbi method, the crosstalk signal without DWDD motion and the low-frequency groove noise are also suppressed. The additionally sampled servo is helpful for not only tracking narrow pitch tracks but also making the system simple and stable. Notably, even if the tilt of the disk is large, tracking offset is more effectively suppressed than in the push-pull method. 3. Read/Write Characteristics Read/write from a sampled servo DWDD disk is demonstrated at a bit density of 80 nm/bit and a scanning velocity of 2.5 m/sec using an optical head of ƒé = 660 nm and NA = 0.6. For a laser-pulsed MFM recording, the laser-pulse duty and modulating magnetic field are 50 % and 300 Oe, respectively. The modulation code is (1,7)RLL, the clock signal is generated from the clock pit signal during writing, and from the MO signal itself during readout. In this system, the Read Solomon Product Code has been adapted for ECC with a 32 kb block size, so a bit error rate (ber) of less than 1.0E-3 can be 100% corrected even if a burst error occurs. In this study, the ber criterion is set to 5.0E Power Margins Figure 7 shows the annealing power (Pa) dependence of ber. Although an annealing power margin of ± 7% appears Read power [ mw ] figure 9 Relationship between ber and readout power. to be small compared to the write power margin seen in a conventional optical disk, we regard this value as being sufficiently high, since the annealing process is completed in the factory, where the process conditions, such as temperature and vibration, can be fully controlled. The write power (Pw) dependence of ber is indicated in Figure 8 for both cases of 1-track and 3-track read/write operations. Crosswrite and crosstalk are included in the latter case. At the lower limit of Pw, there is no difference in ber between 1-track and 3-track read/write. However, there is a great difference around the upper limit of Pw. Low ber is maintained for 1-track read/write, but ber for 3-track read/write increases with rising Pw. This means that the upper limit of Pw is restricted by crosswrite. As a result, a writing power margin of 15% has been obtained. If the conventional reading mode were applied to the 540 nm track pitch with ƒé = 660 nm and NA of 0.6, crosstalk signal would be estimated at -15 dbm or more. This would represent a serious problem in a conventional system, but it is not a problem in the DWDD system. Figure 9 shows the relationship between ber and readout power. As shown in the Figure, the ber of the 3-track read/write operation is almost the same as that for 1-track read/write operation. This result shows that the crosstalk signal without DWDD motion is completely suppressed as a result of decoding the readout signal by means of the Write Power. Pa [ mw ] figure 8 The write power dependence of ber both cases of 1-track and 3-track read/write operation. Magnetic Field, Hw [ Oe ] Figure 10 Dependence of ber on recording magnetic field with laser-pulsed MFM. 276 Trans. Magn. Soc. Jpn., Vol. 2, No. 4, 2002

5 figure 11 Changing of ber as a function of track offset. PR(1,-1)+Viterbi method, giving a readout power margin of } 20%. Figure 10 shows the dependence of ber on recording magnetic field (Hw). As seen in the Figure, a Hw strength of 200 Oe or more is required for this system. 3.2 System Margins The fitting of the shape of the written magnetic domain to the shape of the temperature contour during heat-up is a key factor affecting DWDD operation. Any detracking in the same direction during reading or writing causes a mismatch of the above two parameters. Figure 11 shows the Figure 12 Relationship between ber and tracking offset at radial tilt of } 0.5 deg. and of } 1.5 Đm for read/write operation is obtained as shown in Figure 13. The tilt dependence of ber is shown in Figure 14 (a) for the radial direction and (b) for the tangential direction. For readout, radial and tangential tilt margins are } 1.0 and 0.85, respectively. If tilt occurs during both reading and writing, tilt margins of } 0.7 radially and } 0.66 tangentially are obtained. These are also sufficient for variation in ber as a function of track offset. If data is recorded without any track offset, the detrack margin is more than } 60 nm during readout operation. The detrack margin is decreased to } 36 nm when the same offset is added during both reading and writing. However, due to adoption of the sampled servo system, the track offset is negligibly small even if the radial tilt is relatively large. Figure 12 shows the relationship between ber and tracking offset at a radial tilt of } 0.5. As shown in the Figure, the amount of track offset is only 5 nm or less for a radial tilt of 0.5. It can therefore be concluded that a detrack margin of 36 nm is sufficient for practical use. The defocus margin of } 1.8 Đm for readout operation figure 13 Dependence of ber on focus (a) (b) Figure 14 Tilt dependence of ber, (a) for radial and (b) tangential direction. Trans. Magn. Soc. Jpn., Vol. 2, No. 4,

6 Table 2 System Margins R : Read Mode, W : Read/Write Mode practical use. The obtained system margins for the sampled servo DWDD disk are summarized in Table 2, where the linear density is 80 nm/bit and the ber criterion is 5.0E-4. All these margins are acceptable for practical use. 3.3 High Density Recording There are two kinds of restrictions in DWDD systems on high-density recording. One is signal dropout and the other is jitter. Figure 15 shows the mark length dependence of jitter and dropout. It is not clear whether the writing process for the memory layer, the copying process from the memory layer to the readout layer, or the domain wall motion process in the readout layer is the source of dropouts. However, it is clear that dropouts are caused by perturbation of magnetic properties at the micromagnetic structural level. As seen in the Figure, negligible dropouts are observed at a mark lengths of 94 nm or more. This corresponds to 70 nm/bit in the case of (1,7)RLL-coded data. On the other hand, constant jitter can be obtained in the region of more than 85 nm in mark length, corresponding to 65 nm/bit. However, the data window narrows gradually with rising density. Figure 16 shows the bit length dependence of ber: a ber of 7.0E-5 is obtained at a linear density of 80 nm/bit, which is the target of 3 GB capacity on a 2" disk. Additionally, a ber lower than the criterion of 5.0E-4 can be obtained even at 69 nm/bit. In this range of linear density, the dominant factor impeding high-density recording is jitter rather than dropouts. 4. Summary We have successfully demonstrated high-density recording with a track pitch of 540 nm and 80 nm/bit for a sampled servo DWDD disk using a conventional optical head, which uses the same wavelength of 660 nm and NA of 0.6 as a conventional DVD head. It is confirmed that all the system margins are acceptable for practical use. It can be safely concluded that a 3 GB capacity, 2" diameter disk is now feasible for portable use. If the disk is used for a figure 15 Mark length dependence of dropout and 2T Jitter at 2.5 m/sec. figure 16 Bit length dependence of ber. digital video camera, one hour of MPEG2 video can be recorded at a transfer rate of 6 Mbps. References [1] S. Ohnuki, M. Yoshihiro, K. Shimazaki and N. Isoe, J. Magn. Soc. Japan, 25, No. 3-2, p. 339, (2001). [2] Y. Murakami, T. Numata, N. Takamori, S. Maeda, A. Takahashi, Y. Tanaka, Y. Muto, M. Nishida, M. Kaneko, A. Nakaoki, and K. Fujie, Postdeadline Papers for ODS2001, Santa Fe, WC4, (2001). [3] Y. Tanaka and Y. Muto, Postdeadline Papers for ODS2001, Santa Fe, WC8, (2001). [4] T. Shiratori, E. Fujii, Y. Miyaoka. and Y. Hozumi, J. Magn. Soc. Japan, 22. Supp. No. S2, p. 47, (1998). [5] T. Shiratori, E. Fujii, Y. Miyaoka, and Y. Hozumi,; I Magn. Soc. Japan, 23. Supp. No. Si, p. 145, (1999). [6] C. Peng and M. Mansuripur, Tech. Digest of ODS '00, TuAl, Whistler, p. 47, (2000). [7] Y. Nishi, T. Shimano and H. Kando, Tech. Digest of ODS '00, P15, Whistler, p. 141, (2000). Received April 4, 2002; Accepted August 2, Trans. Magn. Soc. Jpn., Vol. 2, No. 4, 2002