Effect of Transition Curvature and Track Edge ... - IEEE Xplore

2 downloads 0 Views 565KB Size Report
Masato Shiimoto, Hiroyuki Katada, Yosuke Urakami, Mitsuhiro Hashimoto, Mikito Sugiyama,. Takeshi Nakagawa, Takayuki Ichihara, and Kazuhiro Nakamoto.
1428

IEEE TRANSACTIONS ON MAGNETICS, VOL. 46, NO. 6, JUNE 2010

Effect of Transition Curvature and Track Edge Fluctuation on Track Edge Noise for Narrow Track Recording Masato Shiimoto, Hiroyuki Katada, Yosuke Urakami, Mitsuhiro Hashimoto, Mikito Sugiyama, Takeshi Nakagawa, Takayuki Ichihara, and Kazuhiro Nakamoto Central Research Laboratory, Hitachi, Ltd., Odawara 256-8510, Japan We investigated the effect of track edge noise on transition curvature and track edge fluctuation through experiments and calculations. We measured and calculated two types of recording heads that had similar magnetic write widths, but different transition curvatures. In both the measurements and calculations, track edge noise increased more at high linear density with head with larger transition curvature. This was due to wider erase band, which is considered to be noise source at the track edge. We also clarified that reducing the transition curvature and track edge fluctuation was effective to improve bit error rate, because reducing the transition curvature increases signal-to-noise ratio (SNR), especially at high linear density, and reducing the track edge fluctuation also increases SNR, especially at low linear density. Therefore, reducing both transition curvature and track edge fluctuation is necessary in order to reduce the track edge noise and consequently achieve high areal density. Index Terms—Magnetic recording, recording head, track edge fluctuation, track edge noise, transition curvature. TABLE I MEASURED AND CALCULATED HEAD SPECIFICATIONS

I. INTRODUCTION

I

N order to increase the areal density of the hard disk drives, it is necessary to increase both linear density and track density. A key to increasing the linear density is reducing the transition noise. Increasing head field gradient of the down-track direction, reducing head to media spacing, reducing crystal grain size in media, and reducing magnetic cluster (magnetic clustering of crystal grains in media) are well known approaches to reduce the transition noise [1]. To achieve high track density, it is necessary to reduce the track width of the recording head and to suppress degradation of recording properties associated with the reduced track width. However, it was reported that signal-to-noise ratio (SNR) at the track center decreases as the track width becomes narrower while keeping the head field gradient constant [2]. Therefore, clarifying the cause of track edge noise is important in order to achieve high track density. One cause of track edge noise is fluctuation in track edge line [3], [4]. Two effective techniques to reduce this fluctuation are reducing the magnetic cluster [5] and increasing cross-track field gradient [2]. On the other hand, transition curvature [6] is also considered to be important for controlling the track edge magnetization in the recording pattern. In this study, we investigated the effects of the transition curvature and the cross-track field gradient on track edge noise, SNR, and bit error rate (BER) through measurements and calculations. II. EXPERIMENTAL AND CALCULATION PROCEDURE

Read and write characteristics were measured using a spinstand tester. We evaluated the track edge noise and erase band width (EBW) for two kinds of heads that had similar magnetic write widths (MWW) but different transition curvatures (c) using triple-track method [7]. The background of the triple

Manuscript received October 28, 2009; revised December 16, 2009; accepted December 25, 2009. Current version published May 19, 2010. Corresponding author: M. Shiimoto (e-mail: [email protected]). Digital Object Identifier 10.1109/TMAG.2010.2040249

track was AC erased state. Table I lists MWWs at 235 kfci, normalized transition curvatures (c/MWW), pole width (PW), and throat height (TH) for Heads A and B. The transition curvature (c) was calculated using the actual writer structure evaluated in spinstand testing. The c/MWW was controlled by controlling PW and TH, because the c/MWW increased as PW reduced and TH increased at the same MWW. All the written recording patterns were read by an identical read head that had a very narrow magnetic read width (MRW) of 31 nm, which enabled precise evaluation of the write heads. We used capped granular perpendicular media [8]. The head to media spacing was fixed at 10 nm for both writing and reading. The head field distribution was calculated using 3-D finite-element method. The effective head field was calculated as follows:

is the function of the angle between the perpendicular direction for the medium plane and the head field. A coefficient of 1.0 was used, which was determined by fitting to the angular dependence of medium remanence coercivity [9]. One example of a calculated head field contour is shown in Fig. 1. Head field width was defined as the width at the medium dynamic coercivity of 7 kOe. The head field width corresponds to MWW. The head field gradient was defined from a field contour corresponding to the medium dynamic coercivity. The down-track gradient at the track edge was defined as the field gradient in the down-track direction at cross-track position of the field width. The cross-track field gradient was defined as the field gradient in the cross-track direction at the edge of the

0018-9464/$26.00 © 2010 IEEE

SHIIMOTO et al.: EFFECT OF TRANSITION CURVATURE AND TRACK EDGE FLUCTUATION ON TRACK EDGE NOISE

1429

Fig. 1. Typical calculated head field contour. Field width, field gradient, and field curvature were defined as shown in this figure.

field width. The field curvature, which corresponds to the transition curvature, was defined as the curvature of 7-kOe contour at the field width as shown in Fig. 1. cross-track position of Read and write characteristics for Heads C and D were calculated by a micro-magnetic simulation based on a Landau-Lifshitz-Gilbert equation [10]. BER was also calculated by simulation using a generated waveform that takes the transition curvature and track edge noise into account [11]. A partial response maximum likelihood system was used for signal processing.

Fig. 2. Corss-track noise profiles for Heads A and B at (a) 235 and (b) 940 kfci.

III. RESULTS AND DISCUSSION A. Effect of Transition Curvature on Track Edge Noise by the Measurement The track edge noise for two kinds of recording heads, Heads A and B, which had different transition curvatures was measured. The measured cross-track noise profiles at (a) 235 kfci and (b) 940 kfci are shown in Fig. 2. Fig. 3 plots the measured dependence of peak noise at the track edge and the EBW on linear density for Heads A and B. The peak noise values at the track edge for Heads A and B were almost the same at low linear density below 300 kfci. However, the peak noise value for Head A, which had larger transition curvature, became larger than that of Head B as linear density increased above 300 kfci. The EBW for Head A at low linear density was almost the same as that for Head B. However, the EBW for Head A became larger than that for Head B as linear density increased above 300 kfci. These behaviors of the EBW were very similar to that of the peak noise value at the track edge. Accordingly, the erase band is considered to be well correlated with the track edge noise. B. Effect of Transition Curvature on Track Edge Noise by the Calculation We investigated the effect of the transition curvature on the track edge noise and the erase band through calculations. First,

Fig. 3. Dependence of peak noise at track edge and erase band width for Heads A and B.

we analyzed why the EBW depended on the transition curvature. Fig. 4 represents the relationship between the field gradient at the track edge and the normalized transition curvature. We found that degradation of the field gradient at the track edge increased as the transition curvature increased. These results indicate that the increase in transition length at the track edge became larger as the transition curvature increased because transition length generally increases as the field gradient decreases. Fig. 5 shows schematic images of recording patterns with different transition curvatures. As shown in the figure, larger transition curvature results in narrower magnetic write width and wider EBW at high linear density. This is because inter-transition interference at the track edge increases as the transition

1430

IEEE TRANSACTIONS ON MAGNETICS, VOL. 46, NO. 6, JUNE 2010

Fig. 4. Relationship between calculated field gradient at track edge and normalized transition curvature.

Fig. 6. Calculated medium noise profiles for Heads C and D with different transition curvatures using (a) wide read head and (b) narrow read head.

erase band, which is considered to be noise source [12]. Therefore, larger track edge noise with larger transition curvature is explained by wider EBW. Fig. 5. Schematic images of recording pattern with different transition curvatures at high linear density. (a) Small transition curvature; (b) large transition curvature.

length at the track edge increases. Therefore, the larger EBW for heads with larger transition curvature is explained by the increase in transition length at the track edge due to degradation of the field gradient at the track edge. Next, we calculated cross-track noise profiles to clarify the relationship between track edge noise and the track edge magnetization. Fig. 6 plots the calculated cross-track noise profiles for Heads C and D, which had similar MWW but different transition curvature. Two read heads with different MRW were used. One had a relatively wide-width of 50 nm and the other was very narrow-width of 15 nm. When the MRW was 50 nm, the noise peak for head with larger transition curvature was higher than that with smaller transition curvature. These calculated results are consistent with the measured results. When the MRW was 15 nm, the noise peak values at the track edge for Heads C and D were almost the same, but the noise at the track edge for head with larger transition curvature was positioned more closely to the track center. These behaviors correspond to results that larger transition curvature increases the width of the

C. Effect of Transition Curvature and Track Edge Fluctuation on SNR Through the Calculation We investigated the difference between the effects of the transition curvature and the track edge fluctuation on signal quality. Fig. 7 plots the dependence of (a) SNR at 235 and 940 kfci and (b) BER on the transition curvature. In all conditions, the MWW at 235 kfci and the MRW were 64 and 40 nm, respectively. Reducing the transition curvature was effective to increase SNR especially at high linear density. This is because the track edge noise at high linear density was reduced, and furthermore, resolution was increased, as the transition curvature reduced. We also found that reducing the transition curvature improved the BER due to the increase in SNR at high linear density. Fig. 8 plots the dependence of (a) SNR at 235 and 940 kfci and (b) BER on the cross-track field gradient. As the cross-track field gradient increased, the track edge fluctuation was reduced. Increasing the cross-track field gradient was effective to increase SNR, especially at low linear density. This result indicates that increase in track edge fluctuation as the linear density increases is smaller than that in transition noise. Accordingly, the effect of reducing the track edge fluctuation on SNR is decreased at higher linear density because the ratio of the track edge fluctuation noise to the total noise becomes low. We also clarified

SHIIMOTO et al.: EFFECT OF TRANSITION CURVATURE AND TRACK EDGE FLUCTUATION ON TRACK EDGE NOISE

1431

IV. CONCLUSION Using both measurements and calculations, we clarified that increase in track edge noise was greater at high linear density with head that had larger transition curvature, due to wider erase band, which is considered to be noise source at the track edge. We also clarified that reducing the transition curvature and the track edge fluctuation was effective to improve the BER, because reducing the transition curvature increases SNR, especially at high linear density, and reducing the track edge fluctuation increases SNR, especially at low linear density. Therefore, reducing both the transition curvature and the track edge fluctuation is necessary in order to reduce the track edge noise and, consequently achieve high areal density. ACKNOWLEDGMENT The authors would like to thank Y. Nishida, H. Ide, Y. Okada, and I. Ishii of Hitachi Global Storage Technologies and Hitachi Ltd. for their valuable inputs and discussions. REFERENCES Fig. 7. Dependence of (a) SNR at 235 and 940 kfci and (b) bit error rate on c/MWW.

Fig. 8. Dependence of (a) SNR at 235 and 940 kfci and (b) bit error rate on cross-track field gradient.

that increasing the cross-track field gradient was effective to improve the BER due to the increase in SNR at low linear density in addition to SNR at high linear density.

[1] K. Miura, H. Muraoka, and Y. Nakamura, “Effect of head field gradient on transition jitter in perpendicular magnetic recording,” IEEE Trans. Magn., vol. 37, no. 7, pp. 1926–1928, Jul. 2001. [2] Y. Urakami, N. Ito, H. Katada, M. Shiimoto, S. Das, M. Hashimoto, M. Sugiyama, and K. Nakamoto, “Analysis of track-edge noise from the view points of write heads with the different dimensions,” IEEE Trans. Magn., vol. 45, no. 10, pp. 3656–3659, Oct. 2009. [3] T. T. Lam and J. G. Zhu, “Experimental study of track edge noise distribution in narrow track recording,” IEEE Trans. Magn., vol. 30, no. 11, pp. 4245–4247, Nov. 1994. [4] T. Korenari, S. Tsuboi, T. Okumura, H. Matsutera, and K. Tagami, “Analysis of track-edge noise in thin-film recording media,” IEEE Trans. Magn., vol. 33, no. 7, pp. 2509–2512, Jul. 1997. [5] S. Das, N. Ito, M. Sugiyama, and K. Nakamoto, “Effect of track-edge medium noise on high-track density recording,” IEEE Trans. Magn., vol. 45, no. 10, pp. 3558–3561, Oct. 2009. [6] M. Hashimoto, M. Salo, Y. Ikeda, A. Moser, R. Wood, and H. Muraoka, “Analysis of written transition curvature in perpendicular magnetic recording from spin-stand testing,” IEEE Trans. Magn., vol. 43, no. 7, pp. 3315–3319, Jul. 2007. [7] T. Kiya, K. Yamakawa, N. Honda, and K. Ouchi, “Track edge noise measurement with suppressed effect of background magnetization for perpendicular magnetic recording,” J. Magn. Soc. Japan., vol. 33, p. 369, 2009. [8] S. Das and H. Suzuki, “Effect of soft under layer and cap layer thickness on the R/W performance of composite perpendicular media,” J. Appl. Phys., vol. 103, p. 07F540, 2008. [9] K. Tanahashi, H. Nakagawa, R. Araki, H. Kashiwase, and H. Nemoto, “Dual segregant perpendicular recording media with graded properties,” IEEE Trans. Magn., vol. 45, no. 2, pp. 799–804, Feb. 2009. [10] M. Igarashi, M. Hara, A. Nakamura, and Y. Sugita, “Micromagnetic simulation of magnetic cluster, thermal activation volume, and medium noise in perpendicular recording media,” IEEE Trans. Magn., vol. 39, no. 7, pp. 1897–1901, Jul. 2003. [11] Y. Okamoto, K. Ichihara, and H. Osawa, “A study or PRML systems by write and read equalization,” IEEE Trans. Magn., vol. 31, no. 3, pp. 1077–1082, Mar. 1995. [12] K. Miura, T. Ogawa, H. Aoi, H. Muraoka, and Y. Nakamura, “New estimation methods of erase band width and track-edge noise in perpendicular magnetic recording,” J. Magn. Magn. Mater., vol. 320, p. 2908, 2008.