- Email: [email protected]
Damage layer study of the overcoat deposited on the top magnetic layer of hard disks
Journal of Magnetism and Magnetic Materials 351 (2014) 76–81
Contents lists available at ScienceDirect
Journal of Magnetism and Magnetic Materials journal homepage: www.elsevier.com/locate/jmmm
Damage layer study of the overcoat deposited on the top magnetic layer of hard disks R.H. Yeh a, T.M. Chao b, A.H. Tan c,n a
Department of Photonics and Communication Engineering, Asia University, Taichung 41354, Taiwan Department of Applied Geomatics, Chien-Hsin University, JungLi 32097, Taiwan c Department of Mechanical Engineering, Chien-Hsin University, JungLi 32097, Taiwan b
art ic l e i nf o
a b s t r a c t
Article history: Received 5 June 2013 Received in revised form 8 September 2013 Available online 25 September 2013
This paper presents the effect of nanoscale damage layer on magnetic results obtained from the optimization of media overcoat deposition parameters on the top magnetic layer on magnetic hard disks. We have investigated the effects of interface interaction between the overcoat deposition parameters on the top magnetic layer on the media by using a plasma enhanced chemical vapor deposition (PECVD) on next generation hard disks. The goal is to achieve a reduced damage layer, lower head media spacing (HMS) and a higher spectrum of signal to noise ratio (SpSNR) optimized by using Taguchi experimental design with a four-factor three-level (L9) orthogonal array. An analysis of variance (ANOVA) was carried out to interpret the measured coercivity (Hc), HMS and SpSNR. It was found that source gas type is the most significant factor with a percentage contribution effect of 59.8% on HMS and 51.7% on SpSNR. The bias voltage is the second most significant factor with its percentage contribution being 24.2% on HMS and 31.0% on SpSNR. Overall, the optimum SpSNR was obtained using a C2H2 source gas, 100 V bias voltage, 50 V anode voltage and 20 sccm gas flow rate, respectively. & 2013 Elsevier B.V. All rights reserved.
Keywords: Damage layer Diamond-like carbon Top magnetic layer Hard disk Taguchi design
1. Introduction Magnetic storage hard disk drives (HDD) have been and are used widely to provide storage of digital information in modern computer systems and consumer electronic products. The disk arial recording density of hard disk drives will continue at a 50% growth rate per year as the video-audio digital applications and data storage capacity demands keep increasing [1]. Consequently, to be able to develop a higher recording density disk with good reliability performance is becoming more and more important in order to meet future storage requirements. Further increases in the magnetic storage density and head read/write performance for future HDD will necessitate an improvement in magnetic layer design [2,3] and a reduction in the head media read/write spacing (HMS) [4,5]. A lower HMS will improve the media as this results in a lower head fly height, thinner diamond-like-carbon (DLC) overcoat thickness and a reduced damage layer thickness (as defined and illustrated in Fig. 1). The damage layer results from the interface interaction between the protective overcoat and top magnetic layer (TML). The deposition of DLC overcoat damages the TML, by producing a nanoscale damage layer which in turn results in an increase of the head/media read/write interface gap.
n
Corresponding author. Tel.: þ 886 345 81196; fax: þ 886 325 03872. E-mail address: [email protected] (A.H. Tan).
0304-8853/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jmmm.2013.09.038
The higher thicknesses of the overcoat film and damage layer are usually associated with a higher HMS which can lead to a reduction in head media read/write sensitivity [6]. In general, HMS decreases and wear resistance decreases with a decrease of the DLC overcoat thickness [7]. The HMS of media correlates well with its magnetic performance [8]. This also indicates that a lower HMS results in a higher spectrum and better signal to noise ratio (SpSNR). Therefore, developing a thinner DLC, and providing better reliability, and a thinner damaged layer with better magnetic performance is a necessity to meet the challenges of future HDD technology. Li and Bhushan studied the micromechanical and tribological characterization of hard amorphous carbon coatings as thin as 5 nm for magnetic recording heads [9]. Disk DLC with 3–4 nm thickness is the most common currently used protective layer between the head/ disk interfaces [10]. Tribology and wear resistance will encounter more challenges when the overcoat thickness is lowered to an ultra-thin level (o3 nm). Hence, regarding the development of disk DLC overcoats, how can we avoid the wear between a head and disk? These are the important topics addressed in this study. Regarding the mechanical properties of DLC films, as studied by many researchers [11,12]. The source gas with a different C/H ratio will result in different sp2/sp3 ratio of the DLC films, which correlated to their friction coefficients and wear rates [13]. Sharma [14] illustrated that the lower hydrogen content in DLC films will enhance the scratch resistance and tribological properties. Niakan [15] illustrated that increasing the deposition bias voltage resulted in a higher density and hardness of DLC films.
R.H. Yeh et al. / Journal of Magnetism and Magnetic Materials 351 (2014) 76–81
77
2. Experimental procedure
Fig. 1. HRTEM cross-section image of various deposition layers on the substrate (illustrates the damaged layer between the overcoat and top magnetic layer).
Magnetic properties may be improved by forming two or more layers of the magnetic layers [16,17], and by providing each layer with different functions. The TML of magnetic layers are also coated with DLC overcoats to protect against chemical reactions and mechanical impacts. The DLC overcoats design concept is not only considered better for the reliability properties but also reduces the damaged layer thickness which results in better magnetic properties. Most of the literatures mention only the mechanical damage [18,19] or the corrosion effects [20–22] of the DLC overcoat films. Until now, the damage layer interaction between overcoat deposition and TML had not been investigated. Plasma enhanced chemical vapor deposition (PECVD) [23] is a good candidate to further improve the durability performance of the media by using thinner overcoats because it has the advantage of offering better control of the chemical composition and bonding structure in the deposited thin films. But it can also cause excessive media surface roughness if the source gas used has too high of a molecular weight. On the other hand, if a source gas with too small of a molecular weight is used, the deposited layer will penetrate into the TML, resulting in a damaged layer which results in an increase of the read/write gap in the head/disk interface. Except for the deposition source gas type of PECVD, there are many deposition parameters which will be investigated. It has also not yet been investigated whether reduced damage layer (enhanced magnetic performance) and remained wear properties can be achieved through manipulation of the deposition conditions, i.e. optimization of source gas, substrate bias voltage, anode voltage and gas flow rate. The need for improved product manufacturing has driven engineers to quickly find an optimal recipe for processing. Therefore, the Taguchi method [24,25] was introduced as a useful engineering methodology to optimize the process conditions. Selecting a proper orthogonal array, the study of 4 factors at 3 levels as mentioned above can be conducted with only 9 experimental runs. This results in less experiments being required in order to study all levels of input parameters and also filters out some effects due to statistical variation [26]. Therefore, the purpose of this research is to study the effects of interface interaction between the ultra-thin overcoat PECVD parameters (source gas, substrate bias voltage, anode voltage and gas flow rate, overcoat film thickness) on TML in next generation hard disks utilizing the Taguchi method. The coercivity (Hc), nucleation field (Hn), HMS, spectrum of signal to noise ratio (SpSNR), nanohardness and microstructure, which is correlated to damage layer, magnetic and wear resistance, shall be investigated. Finally, this optimization of deposition parameters of DLC films on TML will be confirmed, and the media magnetic characteristics and wear resistance of optimized DLC film deposition will be analyzed.
A glass substrate with a diameter of 65 mm and a thickness of 10 mm were utilized in this study. All substrates were sequentially deposited with under layers, inter layers of Cr, CrMoB and CrMo, a Ru layer, bottom magnetic layer of CoCrB and top magnetic layer of CoCrPtB with various DLC overcoat films (2 nm in thickness) using an Intevac 250B disc DC sputtering system. After sputtering, all media were coated with a 1.2 nm thickness lubrication layer. Except for the overcoat deposition parameters and structure, all the media studied here had the same under layers, inter layers, Ru layer and magnetic layers. The total thickness of all magnetic layers is 14 nm. In this study, a PECVD system for DLC hydrogenated carbon films coating on the TML were carried out the following control factors: (1) source gas; (2) substrate bias voltage; (3) anode voltage; and (4) gas flow rate. The key parameters using the four factors with three levels (L9) orthogonal array of experimental arrangements are listed in Table 1. The Taguchi method is known to be a very good tool for parameter study, and is applied in the present work. This method is powerful and effective in helping researchers design their products and processes as well as to solve troublesome quality problems. After using Taguchi experimental design, the factorial effect and contribution ratio of each factor of each property are presented. Raman spectroscopy was used to study the bonding states of various phases of carbon and the structural quality of the DLC films. In the analysis of the Raman spectrum, both the Gaussian and Lorentzian distribution functions are able to identify the D and G peak positions from the deconvolution of the spectra, and thus one can obtain the intensity ratio (Id/Ig) as a quality characteristic. The Raman spectroscopy is obtained using a Horiba S320C MKII system with Ar laser at 514 nm wavelengths. The thickness of overcoat films was measured with an N&K optical analyzer, which was correlated to transmission electron microscopy cross-sections. In order to measure the nanohardness of a specimen coated with DLC films, a nanoindenter (UMIS, Australia) was applied with the maximum indentation load varying at the scale of milli-Newton (mN). In this study, the 0.015 mN and 30 mN were used for initial contact and maximum load, respectively. Tribological performance of disks was evaluated by contact-start-stop (CSS) testing in Hot/ Wet environments with 30 1C and 80% relative humidity. After CSS durability test, the wear resistance of overcoat films was observed by optical microscopy (OM). A vibrating sample magnetometer (VSM) was used to measure the Hc and Hn for each of the Taguchi experiments. The HMS and the SpSNR were measured by using a Guzik write/read tester. The HMS includes a pole tip recess, DLC overcoat on head slider, head flying height, thickness of lube and DLC overcoat on the media, and around half of the recording layer thickness [27,28]. The HMS value is adjusted by subtracting a same
Table 1 Taguchi experimental layout using L9 orthogonal array of deposition conditions (current process parameters – Condition 2: CH4 source gas, 200 V bias voltage, 100 V anode voltage and 30 sccm gas flow rate). Cond.
Factor A Source gas
Factor B Bias voltage (V)
Factor C Anode voltage (V)
Factor D Gas flow (sccm)
1 2 3 4 5 6 7 8 9
CH4 CH4 CH4 C2H6 C2H6 C2H6 C2H2 C2H2 C2H2
100 200 300 100 200 300 100 200 300
50 100 150 100 150 50 150 50 100
20 30 40 40 20 30 30 40 20
78
R.H. Yeh et al. / Journal of Magnetism and Magnetic Materials 351 (2014) 76–81
estimate fix spacing value (34.2 nm), and then the damage layer thickness was investigated. Except for the DLC overcoat deposition parameters, the rest of the spacing contribution terms are the same. Therefore it is possible to compare the damage layer thickness differences using different DLC overcoat deposition parameters. We believe that with an increase of HMS resulting from a higher damage layer between overcoat and TML, the SpSNR value of magnetic recording performance will be decreased due to the higher HMS or damage layer thickness. An analysis of variance (ANOVA) was carried out to interpret the measured Hc, HMS and SpSNR values. The Taguchi method was then used to analyze the significant factors and the values. Optimized magnetic performance obtained from Taguchi analysis was verified using different DLC overcoat thicknesses of 1.5– 2.5 nm and then compared with the current design as the basis for control.
Fig. 2. Correlation among Hc, Hn and damage layer thickness for each of the Taguchi experiments.
3. Results and discussion 3.1. Main effects and their contribution of DLC deposition parameters on various properties Hc and Hn of all the Taguchi experiments are shown in Table 2, with an increase in the Hc from 5543 to 5856 Oe, the Hn decreases from 2563 to 2103 Oe. As shown in Fig. 2, there is a good correlation between Hc and Hn. In Acharya et al. study [29], there was a similar trend observed when the thickness of the TML is decreased while the Hc decreases and Hn increases. It is impossible to measure the real damage layer thickness directly. But, a higher damage layer thickness results during overcoat deposition condition variations resulting in a lower thickness of TML, which then corresponds to the higher Hc and the lower Hn values. As Fig. 2 illustrates again, there is a good correlation between Hc and damage layer thickness (damage layer thickness ¼ HMS– 34.2 nm) also. We expect that the higher damage layer thickness of the special overcoat deposition conditions will result in a poorer head media read/write sensitivity and magnetic performance. This indicates that a higher damage layer thickness will induce a higher HMS and a lower SpSNR. From the results, Fig. 3 illustrates the correlation of Hc with HMS and SpSNR. Their R square values are 0.93 and 0.84, indicating that there is a very good correlation between Hc and HMS, Hc and SpSNR, respectively. The experimental results of Table 2 are analyzed in order to see the main factors and their effects on the Hc, HMS and SpSNR, using the Taguchi method, as shown in Tables 3–5, respectively. Table 3 elucidates the influence of source gas type, CH4, C2H6 and C2H2, which results in a change in the main effects of Hc in decreases from 5761 to 5558 Oe. The decrease in the main effects of the Hc Table 2 Hc, Hn, HMS, damage layer thickness (value defined as HMS 34.2 nm) and SpSNR for each of the Taguchi experiments. Cond. Hc (Oe)
1 2 3 4 5 6 7 8 9
Hn (Oe)
HMS (nm)
Damage layer thickness (nm)
SpSNR (dB)
Avg
Std Avg
Std Avg
Std
Avg
Std
Avg
Std
5710 5717 5856 5597 5620 5733 5543 5557 5573
40 45 43 25 30 40 49 45 35
17 20 33 11 26 23 30 22 21
0.20 0.10 0.20 0.17 0.18 0.13 0.14 0.19 0.19
0.90 1.04 1.26 0.48 0.59 1.01 0.20 0.37 0.59
0.20 0.10 0.20 0.17 0.18 0.13 0.14 0.19 0.19
13.52 13.23 12.91 13.74 13.62 13.44 13.80 13.67 13.58
0.10 0.05 0.15 0.12 0.13 0.08 0.09 0.14 0.14
2354 2310 2103 2434 2489 2368 2466 2563 2456
35.10 35.24 35.46 34.68 34.79 35.21 34.40 34.57 34.79
Fig. 3. Correlation among Hc, HMS and SpSNR for each of the Taguchi experiments.
value is explained by a higher TML thickness resulting from the lower damage layer thickness using a source gas C2H2 deposition. When compared with C2H6 and C2H2, the lower molecular weight of CH4 source gas results in more to penetration into the layer which damages the TML during deposition. This indicates that it induces more interaction between the overcoat and the TML. Consequently, we find that CH4 source gas will result in a thicker damage layer. The increases of bias voltage, from 100 to 300 V, results in an increase in the main effects of Hc value as it increases from 5617 to 5721 Oe. This increase is due to the higher substrate bias voltage increasing the impact energy between overcoat and TML. Consequently, the higher bias voltage could enhance the damage layer and therefore reduce the TML thickness. The increase of anode voltage, from 50 to 150 V, results in the Hc value only changing from 5667 to 5673 Oe. There is no significant change on the main effect when the anode voltage is changed. The increase of gas flow rate, from 20 to 40 sccm, results in the Hc value slightly increasing from 5634 to 5670 Oe. This variation is not obvious over this gas flow rate range. Based on the analysis, it can be concluded that the source gas type is the most significant factor with a percentage contribution of 64.5% on Hc value. The bias voltage is the second significant factor with its percentage contribution being 19.9% on Hc value. However, anode voltage and gas flow rate are not significant factors in this study. Table 4 elucidates the influence of source gas type, CH4, C2H6 and C2H2, which results in the change in main effects as the HMS value decreases from 35.27 to 34.58 nm. The decrease in the main effects of HMS value is explained by a lower damage layer thickness on TML which results using a higher molecular weight
R.H. Yeh et al. / Journal of Magnetism and Magnetic Materials 351 (2014) 76–81
Table 3 Main effects and their percentage contribution of each factor on Hc value.
Level 1 Level 2 Level 3 Effect Rank Contribution (%)
Factor A Source gas
Factor B Bias voltage
Factor C Anode voltage
Factor D Gas flow
5761 5650 5558 203 1 64.5
5617 5631 5721 104 2 19.9
5667 5629 5673 44 3 3.5
5634 5664 5670 35 4 2.3
Table 4 Main effects and their percentage contribution of each factor on HMS value.
Level 1 Level 2 Level 3 Effect Rank Contribution (%)
Factor A Source gas
Factor B Bias voltage
Factor C Anode voltage
Factor D Gas flow
35.27 34.89 34.58 0.68 1 59.8
34.73 34.86 35.15 0.43 2 24.2
34.96 34.90 34.88 0.08 3 0.8
34.89 34.95 34.90 0.06 4 0.5
Table 5 Main effects and their percentage contribution of each factor on SpSNR.
Level 1 Level 2 Level 3 Effect Rank Contribution (%)
Factor A Source gas
Factor B Bias voltage
Factor C Anode voltage
Factor D Gas flow
13.22 13.60 13.68 0.46 1 51.7
13.69 13.51 13.31 0.38 2 31.0
13.54 13.52 13.44 0.10 4 2.3
13.57 13.49 13.44 0.13 3 3.9
of source gas C2H2 deposition. The increase of bias voltage, from 100 to 300 V, results in an increase in main effects of HMS as its value increases from 34.73 to 35.15 nm. This increase is because the higher impact energy between overcoat and TML results from a higher substrate bias voltage. Consequently, the higher bias voltage could enhance the damage layer formation and therefore decrease the TML thickness. The increases of anode voltage, from 50 to 150 V, results in the HMS value slightly decreasing from 34.96 to 34.88 nm. There is no significant change on the main effect when the anode voltage is changed. The increases of gas flow rate, from 20 to 40 sccm, results in the HMS value only slightly changing from 34.89 to 34.90 nm. This variation is not significant over this gas flow rate range. Based on the analysis, it can be concluded that the source gas type is the most significant factor with its percentage contribution being 59.8% on HMS value. The bias voltage is the second most significant factor with its percentage contribution being 24.2% on HMS value. However, anode voltage and gas flow rate are not significant factors in the change in main effects of HMS. Table 5 shows the influence of source gas type, CH4, C2H6 and C2H2, which results in the change in main effects of SpSNR which increases from 13.22 to 13.68 dB. The increase in the main effects of SpSNR value is explained by the lower damage layer and HMS which results from source gas C2H2 deposition. The increases of bias voltage, from 100 to 300 V, results in a decrease in main effects of SpSNR value as it decreases from 13.69 to 13.31 dB. This decrease is due to the higher impact energy between overcoat
79
and TML resulting from the higher substrate bias voltage. Consequently, the higher bias voltage increases the damage layer happens which decreases the TML thickness. The increases of anode voltage, from 50 to 150 V, results in the SpSNR value slightly decreasing from 13.54 to 13.44 dB. There is no significant change on the main effect when the anode voltage is changed. The increases of gas flow rate, from 20 to 40 sccm, results in the SpSNR value slightly decreasing from 13.57 to 13.44 dB. This variation is not significant over this gas flow rate range. Based on the analysis, it can be concluded that the source gas type is the most significant factor with its percentage contribution being 51.7% on SpSNR value. The bias voltage is the second most significant factor with its percentage contribution being 31.0% on SpSNR value. However, anode voltage and gas flow rate are not significant factors in the change in main effects of SpSNR. Taguchi analysis of the experimental results shows that all, Hc, HMS and SpSNR are most affected by source gas and then secondly by bias voltage. Lower molecular weight (CH4 lower than C2H6 and C2H2) of the source gas is makes it easier to penetrate and damage the TML, indicating that it induces more interaction between the overcoat and the TML, which results in a higher Hc and HMS, and lower SpSNR. To obtain the lowest HMS and the highest SpSNR, the optimized source gas is C2H2 and bias voltage is 100 V, respectively. Anode voltage and gas flow rate are not dominant factors to affect the HMS and SpSNR. 3.2. Optimization of DLC deposition parameters simultaneously for HMS and SpSNR The objective of this study was to explore and determine how to optimize the control factors in order to achieve the best magnetic performance for future HDD product requirement. Overall, to obtain the highest SpSNR (lowest damage layer and HMS), the optimized source gas, bias voltage, anode voltage and gas flow rate are C2H2, 100 V, 50 V and 20 sccm, respectively. In general, HMS decreases and SpSNR increases with a decrease of the DLC overcoat thickness. Further to study the effect of overcoat thickness on HMS and SpSNR, by decreasing the overcoat thickness from 2.5 to 1.5 nm. At the same time, changing the overcoat deposition parameters from the current standard process to Taguchi's optimum design was compared. As shown in Fig. 4, at the same overcoat thickness, the optimum overcoat deposition condition provides about a 0.56 nm lower HMS than that of the current standard process condition. This is attributed to the lower damage layer of the TML which is obtained using the optimum overcoat deposition conditions. Low damage layer thickness results
Fig. 4. HMS of current process versus optimum deposition condition at different thicknesses.
80
R.H. Yeh et al. / Journal of Magnetism and Magnetic Materials 351 (2014) 76–81
in low HMS. Meanwhile, Fig. 5 demonstrates the SpSNR comparison between the standard current process and optimum deposition conditions at different overcoat thickness. The 0.51 dB SpSNR increase can be achieved due to the optimum deposition conditions using the same overcoat thickness. We believe that with the decrease of HMS resulting from the lower damage layer between overcoat and the TML, that the SpSNR value of magnetic recording performance is increased due to the lower HMS or damage layer thickness. Compared with the current deposition condition, this means that a 0.56 nm lower HMS and a 0.51 dB higher SpSNR can be achieved by using the optimum overcoat deposition conditions. Compared with the current standard process overcoat deposition condition, a lower Id/Ig (higher sp3 bonds) and a higher nanohardness of the overcoat properties are found from the Taguchi's optimum condition, as shown in Fig. 6. The decrease of Id/Ig ratio in DLC films induces more sp3 content, thus resulting in higher nanohardness. It is expected that this optimum overcoat deposition condition can be presumed to have better wear resistance because of the lower Id/Ig ratio and the higher nanohardness. Fig. 7(a) and (b) shows the CSS trend charts of current process versus optimum overcoat deposited condition with 2 nm overcoat thickness, respectively. Both of current process and optimum overcoat deposited condition can sustain a 100,000 cycles CSS test, but disk with the optimum overcoat deposited condition exhibits the lower stiction force. This may result from the overcoat strength, which affects the CSS performance. Finally, Fig. 8(a) and (b) illustrates the worn
Fig. 7. CSS trend charts of (a) current process versus (b) optimum overcoat deposited condition. (Overcoat thickness is 2 nm.)
Fig. 8. Worn surface of disk with DLC film of (a) current process versus (b) optimum overcoat deposited condition. (Overcoat thickness is 2 nm.)
surface of an OM image for disk wear DLC film of current and optimized DLC deposition films. Compared with current DLC deposition film, no wear track was observed in optimized DLC deposition film because it exhibited the higher nanohardness. 4. Conclusions
Fig. 5. SpSNR of current process versus optimum deposition condition at different thicknesses.
The purpose of this paper was to establish an efficient method for a robust overcoat deposition design process improvement on magnetic performance for hard disk media by utilizing the Taguchi method. The interface interaction study of nanoscale damage layers between PECVD overcoat deposition parameters on the top magnetic layer (TML) was achieved by using the four factors with a three level (L9) orthogonal array of experimental arrangements. Based on the ANOVA analysis, the source gas type is the most significant factor and its percentage contribution is 59.8% on HMS and 51.7% on SpSNR, respectively. The bias voltage is the second most significant factor as its percentage contribution is 24.2% on HMS and 31.0% on SpSNR. The robust process result obtained from the optimization of overcoat deposition parameters can achieve a 0.56 nm lower HMS, a 0.51 dB higher SpSNR and better wear resistance.
Acknowledgments
Fig. 6. Id/Ig and nanohardness of current process versus optimum overcoat deposited condition. (Overcoat thickness is 2 nm.)
The author would like to thank the National Science Committee of Taiwan for their financial support under Grant NSC100-2221-E231-006.
R.H. Yeh et al. / Journal of Magnetism and Magnetic Materials 351 (2014) 76–81
References [1] S.I. Iwasaki, Perpendicular magnetic recording—its development and realization, Journal of Magnetism and Magnetic Materials 324 (2012) 244–247. [2] M. Oshiki, Review of perpendicular magnetic recording research at Fujitsu, Journal of Magnetism and Magnetic Materials 324 (2012) 351–354. [3] K. Tagam, Looking back at perpendicular magnetic recording R&D in NEC, Journal of Magnetism and Magnetic Materials 324 (2012) 355–362. [4] R. Wood, Future hard disk drive systems, Journal of Magnetism and Magnetic Materials 321 (2009) 555–561. [5] S.N. Piramanayagam, K. Srinivasan, Recording media research for future hard disk drives, Journal of Magnetism and Magnetic Materials 321 (2009) 485–494. [6] G. Choe, M. Zheng, E.N. Abarra, B.G. Demczyk, J.N. Zhou, B.R. Acharya, K.E. Johnson, High-performance CoPtCrO perpendicular media: optimizing exchange coupling and anisotropy orientation dispersion, Journal of Magnetism and Magnetic Materials 287 (2005) 159–166. [7] M. Zhong, C. Zhang, J. Luo, X. Lu, The protective properties of ultra-thin diamond like carbon films for high density magnetic storage devices, Applied Surface Science 256 (2009) 322–328. [8] H.J. Richter, Density limits imposed by the microstructure of magnetic recording media, Journal of Magnetism and Magnetic Materials 321 (2009) 467–476. [9] X.D. Li, B. Bhushan, Micromechanical and tribological characterization of hard amorphous carbon coatings as thin as 5 nm for magnetic recording heads, Wear 220 (1998) 51–58. [10] M. Zhong, C. Zhang, J. Luo, X. Lu, The protective properties of ultra-thin diamond like carbon films for high density magnetic storage devices, Applied Surface Science 256 (2009) 322–328. [11] T.M. Chao, A.H. Tan, DLC deposition parameters optimization for head disk design interface with a thermal protrusion slider from tribological point of view, Materials and Design 48 (2013) 58–67. [12] C.A. Charitidis, Nanomechanical and nanotribological properties of carbonbased thin films: a review, International Journal of Refractory Metals and Hard Materials 28 (2010) 51–70. [13] A. Tomala, A. Pauschitz, M. Roy, Nanotribology of pulsed direct current magnetron sputtered diamond like carbon films, Surface Science 616 (2013) 60–70. [14] N. Sharma, N. Kumar, S. Dash, C.R. Das, R.V. Subba, A.K. Tyagi, B. Raj, Scratch resistance and tribological properties of DLC coatings under dry and lubrication conditions, Tribology International 56 (2012) 129–140. [15] H. Niakan, Q. Yang, J.A. Szpunar, Structure and properties of diamond-like carbon thin films synthesized by biased target ion beam deposition, Surface and Coatings Technology 223 (2013) 11–16.
81
[16] J.G. Choi, D.G. Hwang, J.R. Rhee, S.S. Lee, Soft magnetic properties of hybrid ferromagnetic films with CoFe, NiFe, and NiFeCuMo layers, Thin Solid Films 519 (2011) 8394–8396. [17] D. Mishra, A.K. Singh, P.C. Shyni, D. Sharma, A. Perumal, Enhanced soft magnetic properties in multilayer structured amorphous Fe-Ta-C films, Journal of Applied Physics 109 (2011) 07A304. [18] F.R. Marciano, R.P.C. Costa, O.D.A. Lima, A.O. Lobo, E.J. Corat, Trava AVJ. Tribological behavior under aggressive environment of diamond-like carbon films with incorporated nanocrystalline diamond particles, Surface and Coatings Technology 206 (2011) 434–439. [19] T.S. Santra, T.K. Bhattacharyya, P. Patel, F.G. Tseng, T.K. Barik, Structural and tribological properties of diamond-like nanocomposite thin films, Surface and Coatings Technology 206 (2011) 228–233. [20] R. Asakawa, S. Nagashima, Y. Nakamura, T. Hasebe, T. Suzuki, A. Hotta, Combining polymers with diamond-like carbon (DLC) for highly functionalized materials, Surface and Coatings Technology 206 (2011) 676–685. [21] A.S. Hamdy, Electrochemical behavior of diamond-like-carbon coatings deposited on AlTiC (Al2O3 þ TiC) ceramic composite substrate in HCl solution, Electrochimica Acta 56 (2011) 1554–1562. [22] M. Tan, Y. Yan, H. Zhang, X. Han, J. Zhu, J. Han, H. Ma, Corrosion protection of ultra-thin ta-C films for recording slider applications at varied substrate bias, Surface and Coatings Technology 203 (2009) 963–966. [23] J. Wang, G. Liu, J. Xu, X. Deng, L. Wang, Nanomechanical and Electrochemical properties of diamond-like carbon (DLC) films deposited by plasma enhanced chemical vapor, Plasma Science Technology 12 (2010) 461–465. [24] E. Salgueiredo, M. Amaral, M.A. Neto, A.J.S. Fernandes, F.J. Oliveira, R.F. Silva, HFCVD diamond deposition parameters optimized by a Taguchi matrix, Vacuum 85 (2011) 701–704. [25] Y.T. Yen, T.H. Fang, Y.C. Lin, Optimization of screen-printing parameters of SN9000 ink for pinholes using Taguchi method in chip on film packaging, Robotics Computer-Integrated Manufacturing 27 (2011) 531–537. [26] Y.M. Chiang, H.H. Hsieh, The use of the Taguchi method with grey relational analysis to optimize the thin-film sputtering process with multiple quality characteristic in color filter manufacturing, Computers Industrial Engineering 56 (2009) 648–661. [27] X. Liu, W. Clegg, B. Liu, Ultra low head–disk spacing measurement using dual beam polarisation interferometry, Optics Laser Technology 32 (2000) 287–291. [28] Y. Zhou, B. Liu, L. Li, Engineering issue study of triple harmonic method for in situ flying height analysis, Journal of Magnetism and Magnetic Materials 303 (2006) e120–e123. [29] B.R. Acharya, M. Zheng, G. Choe, M. Yu, P. Gill, E.N. Abarra, Oxidation behavior of sputtered CrN/AlN multilayer coatings during heat treatment, IEEE Transactions on Magnetics 41 (2005) 3145–3147.
Contents lists available at ScienceDirect
Journal of Magnetism and Magnetic Materials journal homepage: www.elsevier.com/locate/jmmm
Damage layer study of the overcoat deposited on the top magnetic layer of hard disks R.H. Yeh a, T.M. Chao b, A.H. Tan c,n a
Department of Photonics and Communication Engineering, Asia University, Taichung 41354, Taiwan Department of Applied Geomatics, Chien-Hsin University, JungLi 32097, Taiwan c Department of Mechanical Engineering, Chien-Hsin University, JungLi 32097, Taiwan b
art ic l e i nf o
a b s t r a c t
Article history: Received 5 June 2013 Received in revised form 8 September 2013 Available online 25 September 2013
This paper presents the effect of nanoscale damage layer on magnetic results obtained from the optimization of media overcoat deposition parameters on the top magnetic layer on magnetic hard disks. We have investigated the effects of interface interaction between the overcoat deposition parameters on the top magnetic layer on the media by using a plasma enhanced chemical vapor deposition (PECVD) on next generation hard disks. The goal is to achieve a reduced damage layer, lower head media spacing (HMS) and a higher spectrum of signal to noise ratio (SpSNR) optimized by using Taguchi experimental design with a four-factor three-level (L9) orthogonal array. An analysis of variance (ANOVA) was carried out to interpret the measured coercivity (Hc), HMS and SpSNR. It was found that source gas type is the most significant factor with a percentage contribution effect of 59.8% on HMS and 51.7% on SpSNR. The bias voltage is the second most significant factor with its percentage contribution being 24.2% on HMS and 31.0% on SpSNR. Overall, the optimum SpSNR was obtained using a C2H2 source gas, 100 V bias voltage, 50 V anode voltage and 20 sccm gas flow rate, respectively. & 2013 Elsevier B.V. All rights reserved.
Keywords: Damage layer Diamond-like carbon Top magnetic layer Hard disk Taguchi design
1. Introduction Magnetic storage hard disk drives (HDD) have been and are used widely to provide storage of digital information in modern computer systems and consumer electronic products. The disk arial recording density of hard disk drives will continue at a 50% growth rate per year as the video-audio digital applications and data storage capacity demands keep increasing [1]. Consequently, to be able to develop a higher recording density disk with good reliability performance is becoming more and more important in order to meet future storage requirements. Further increases in the magnetic storage density and head read/write performance for future HDD will necessitate an improvement in magnetic layer design [2,3] and a reduction in the head media read/write spacing (HMS) [4,5]. A lower HMS will improve the media as this results in a lower head fly height, thinner diamond-like-carbon (DLC) overcoat thickness and a reduced damage layer thickness (as defined and illustrated in Fig. 1). The damage layer results from the interface interaction between the protective overcoat and top magnetic layer (TML). The deposition of DLC overcoat damages the TML, by producing a nanoscale damage layer which in turn results in an increase of the head/media read/write interface gap.
n
Corresponding author. Tel.: þ 886 345 81196; fax: þ 886 325 03872. E-mail address: [email protected] (A.H. Tan).
0304-8853/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jmmm.2013.09.038
The higher thicknesses of the overcoat film and damage layer are usually associated with a higher HMS which can lead to a reduction in head media read/write sensitivity [6]. In general, HMS decreases and wear resistance decreases with a decrease of the DLC overcoat thickness [7]. The HMS of media correlates well with its magnetic performance [8]. This also indicates that a lower HMS results in a higher spectrum and better signal to noise ratio (SpSNR). Therefore, developing a thinner DLC, and providing better reliability, and a thinner damaged layer with better magnetic performance is a necessity to meet the challenges of future HDD technology. Li and Bhushan studied the micromechanical and tribological characterization of hard amorphous carbon coatings as thin as 5 nm for magnetic recording heads [9]. Disk DLC with 3–4 nm thickness is the most common currently used protective layer between the head/ disk interfaces [10]. Tribology and wear resistance will encounter more challenges when the overcoat thickness is lowered to an ultra-thin level (o3 nm). Hence, regarding the development of disk DLC overcoats, how can we avoid the wear between a head and disk? These are the important topics addressed in this study. Regarding the mechanical properties of DLC films, as studied by many researchers [11,12]. The source gas with a different C/H ratio will result in different sp2/sp3 ratio of the DLC films, which correlated to their friction coefficients and wear rates [13]. Sharma [14] illustrated that the lower hydrogen content in DLC films will enhance the scratch resistance and tribological properties. Niakan [15] illustrated that increasing the deposition bias voltage resulted in a higher density and hardness of DLC films.
R.H. Yeh et al. / Journal of Magnetism and Magnetic Materials 351 (2014) 76–81
77
2. Experimental procedure
Fig. 1. HRTEM cross-section image of various deposition layers on the substrate (illustrates the damaged layer between the overcoat and top magnetic layer).
Magnetic properties may be improved by forming two or more layers of the magnetic layers [16,17], and by providing each layer with different functions. The TML of magnetic layers are also coated with DLC overcoats to protect against chemical reactions and mechanical impacts. The DLC overcoats design concept is not only considered better for the reliability properties but also reduces the damaged layer thickness which results in better magnetic properties. Most of the literatures mention only the mechanical damage [18,19] or the corrosion effects [20–22] of the DLC overcoat films. Until now, the damage layer interaction between overcoat deposition and TML had not been investigated. Plasma enhanced chemical vapor deposition (PECVD) [23] is a good candidate to further improve the durability performance of the media by using thinner overcoats because it has the advantage of offering better control of the chemical composition and bonding structure in the deposited thin films. But it can also cause excessive media surface roughness if the source gas used has too high of a molecular weight. On the other hand, if a source gas with too small of a molecular weight is used, the deposited layer will penetrate into the TML, resulting in a damaged layer which results in an increase of the read/write gap in the head/disk interface. Except for the deposition source gas type of PECVD, there are many deposition parameters which will be investigated. It has also not yet been investigated whether reduced damage layer (enhanced magnetic performance) and remained wear properties can be achieved through manipulation of the deposition conditions, i.e. optimization of source gas, substrate bias voltage, anode voltage and gas flow rate. The need for improved product manufacturing has driven engineers to quickly find an optimal recipe for processing. Therefore, the Taguchi method [24,25] was introduced as a useful engineering methodology to optimize the process conditions. Selecting a proper orthogonal array, the study of 4 factors at 3 levels as mentioned above can be conducted with only 9 experimental runs. This results in less experiments being required in order to study all levels of input parameters and also filters out some effects due to statistical variation [26]. Therefore, the purpose of this research is to study the effects of interface interaction between the ultra-thin overcoat PECVD parameters (source gas, substrate bias voltage, anode voltage and gas flow rate, overcoat film thickness) on TML in next generation hard disks utilizing the Taguchi method. The coercivity (Hc), nucleation field (Hn), HMS, spectrum of signal to noise ratio (SpSNR), nanohardness and microstructure, which is correlated to damage layer, magnetic and wear resistance, shall be investigated. Finally, this optimization of deposition parameters of DLC films on TML will be confirmed, and the media magnetic characteristics and wear resistance of optimized DLC film deposition will be analyzed.
A glass substrate with a diameter of 65 mm and a thickness of 10 mm were utilized in this study. All substrates were sequentially deposited with under layers, inter layers of Cr, CrMoB and CrMo, a Ru layer, bottom magnetic layer of CoCrB and top magnetic layer of CoCrPtB with various DLC overcoat films (2 nm in thickness) using an Intevac 250B disc DC sputtering system. After sputtering, all media were coated with a 1.2 nm thickness lubrication layer. Except for the overcoat deposition parameters and structure, all the media studied here had the same under layers, inter layers, Ru layer and magnetic layers. The total thickness of all magnetic layers is 14 nm. In this study, a PECVD system for DLC hydrogenated carbon films coating on the TML were carried out the following control factors: (1) source gas; (2) substrate bias voltage; (3) anode voltage; and (4) gas flow rate. The key parameters using the four factors with three levels (L9) orthogonal array of experimental arrangements are listed in Table 1. The Taguchi method is known to be a very good tool for parameter study, and is applied in the present work. This method is powerful and effective in helping researchers design their products and processes as well as to solve troublesome quality problems. After using Taguchi experimental design, the factorial effect and contribution ratio of each factor of each property are presented. Raman spectroscopy was used to study the bonding states of various phases of carbon and the structural quality of the DLC films. In the analysis of the Raman spectrum, both the Gaussian and Lorentzian distribution functions are able to identify the D and G peak positions from the deconvolution of the spectra, and thus one can obtain the intensity ratio (Id/Ig) as a quality characteristic. The Raman spectroscopy is obtained using a Horiba S320C MKII system with Ar laser at 514 nm wavelengths. The thickness of overcoat films was measured with an N&K optical analyzer, which was correlated to transmission electron microscopy cross-sections. In order to measure the nanohardness of a specimen coated with DLC films, a nanoindenter (UMIS, Australia) was applied with the maximum indentation load varying at the scale of milli-Newton (mN). In this study, the 0.015 mN and 30 mN were used for initial contact and maximum load, respectively. Tribological performance of disks was evaluated by contact-start-stop (CSS) testing in Hot/ Wet environments with 30 1C and 80% relative humidity. After CSS durability test, the wear resistance of overcoat films was observed by optical microscopy (OM). A vibrating sample magnetometer (VSM) was used to measure the Hc and Hn for each of the Taguchi experiments. The HMS and the SpSNR were measured by using a Guzik write/read tester. The HMS includes a pole tip recess, DLC overcoat on head slider, head flying height, thickness of lube and DLC overcoat on the media, and around half of the recording layer thickness [27,28]. The HMS value is adjusted by subtracting a same
Table 1 Taguchi experimental layout using L9 orthogonal array of deposition conditions (current process parameters – Condition 2: CH4 source gas, 200 V bias voltage, 100 V anode voltage and 30 sccm gas flow rate). Cond.
Factor A Source gas
Factor B Bias voltage (V)
Factor C Anode voltage (V)
Factor D Gas flow (sccm)
1 2 3 4 5 6 7 8 9
CH4 CH4 CH4 C2H6 C2H6 C2H6 C2H2 C2H2 C2H2
100 200 300 100 200 300 100 200 300
50 100 150 100 150 50 150 50 100
20 30 40 40 20 30 30 40 20
78
R.H. Yeh et al. / Journal of Magnetism and Magnetic Materials 351 (2014) 76–81
estimate fix spacing value (34.2 nm), and then the damage layer thickness was investigated. Except for the DLC overcoat deposition parameters, the rest of the spacing contribution terms are the same. Therefore it is possible to compare the damage layer thickness differences using different DLC overcoat deposition parameters. We believe that with an increase of HMS resulting from a higher damage layer between overcoat and TML, the SpSNR value of magnetic recording performance will be decreased due to the higher HMS or damage layer thickness. An analysis of variance (ANOVA) was carried out to interpret the measured Hc, HMS and SpSNR values. The Taguchi method was then used to analyze the significant factors and the values. Optimized magnetic performance obtained from Taguchi analysis was verified using different DLC overcoat thicknesses of 1.5– 2.5 nm and then compared with the current design as the basis for control.
Fig. 2. Correlation among Hc, Hn and damage layer thickness for each of the Taguchi experiments.
3. Results and discussion 3.1. Main effects and their contribution of DLC deposition parameters on various properties Hc and Hn of all the Taguchi experiments are shown in Table 2, with an increase in the Hc from 5543 to 5856 Oe, the Hn decreases from 2563 to 2103 Oe. As shown in Fig. 2, there is a good correlation between Hc and Hn. In Acharya et al. study [29], there was a similar trend observed when the thickness of the TML is decreased while the Hc decreases and Hn increases. It is impossible to measure the real damage layer thickness directly. But, a higher damage layer thickness results during overcoat deposition condition variations resulting in a lower thickness of TML, which then corresponds to the higher Hc and the lower Hn values. As Fig. 2 illustrates again, there is a good correlation between Hc and damage layer thickness (damage layer thickness ¼ HMS– 34.2 nm) also. We expect that the higher damage layer thickness of the special overcoat deposition conditions will result in a poorer head media read/write sensitivity and magnetic performance. This indicates that a higher damage layer thickness will induce a higher HMS and a lower SpSNR. From the results, Fig. 3 illustrates the correlation of Hc with HMS and SpSNR. Their R square values are 0.93 and 0.84, indicating that there is a very good correlation between Hc and HMS, Hc and SpSNR, respectively. The experimental results of Table 2 are analyzed in order to see the main factors and their effects on the Hc, HMS and SpSNR, using the Taguchi method, as shown in Tables 3–5, respectively. Table 3 elucidates the influence of source gas type, CH4, C2H6 and C2H2, which results in a change in the main effects of Hc in decreases from 5761 to 5558 Oe. The decrease in the main effects of the Hc Table 2 Hc, Hn, HMS, damage layer thickness (value defined as HMS 34.2 nm) and SpSNR for each of the Taguchi experiments. Cond. Hc (Oe)
1 2 3 4 5 6 7 8 9
Hn (Oe)
HMS (nm)
Damage layer thickness (nm)
SpSNR (dB)
Avg
Std Avg
Std Avg
Std
Avg
Std
Avg
Std
5710 5717 5856 5597 5620 5733 5543 5557 5573
40 45 43 25 30 40 49 45 35
17 20 33 11 26 23 30 22 21
0.20 0.10 0.20 0.17 0.18 0.13 0.14 0.19 0.19
0.90 1.04 1.26 0.48 0.59 1.01 0.20 0.37 0.59
0.20 0.10 0.20 0.17 0.18 0.13 0.14 0.19 0.19
13.52 13.23 12.91 13.74 13.62 13.44 13.80 13.67 13.58
0.10 0.05 0.15 0.12 0.13 0.08 0.09 0.14 0.14
2354 2310 2103 2434 2489 2368 2466 2563 2456
35.10 35.24 35.46 34.68 34.79 35.21 34.40 34.57 34.79
Fig. 3. Correlation among Hc, HMS and SpSNR for each of the Taguchi experiments.
value is explained by a higher TML thickness resulting from the lower damage layer thickness using a source gas C2H2 deposition. When compared with C2H6 and C2H2, the lower molecular weight of CH4 source gas results in more to penetration into the layer which damages the TML during deposition. This indicates that it induces more interaction between the overcoat and the TML. Consequently, we find that CH4 source gas will result in a thicker damage layer. The increases of bias voltage, from 100 to 300 V, results in an increase in the main effects of Hc value as it increases from 5617 to 5721 Oe. This increase is due to the higher substrate bias voltage increasing the impact energy between overcoat and TML. Consequently, the higher bias voltage could enhance the damage layer and therefore reduce the TML thickness. The increase of anode voltage, from 50 to 150 V, results in the Hc value only changing from 5667 to 5673 Oe. There is no significant change on the main effect when the anode voltage is changed. The increase of gas flow rate, from 20 to 40 sccm, results in the Hc value slightly increasing from 5634 to 5670 Oe. This variation is not obvious over this gas flow rate range. Based on the analysis, it can be concluded that the source gas type is the most significant factor with a percentage contribution of 64.5% on Hc value. The bias voltage is the second significant factor with its percentage contribution being 19.9% on Hc value. However, anode voltage and gas flow rate are not significant factors in this study. Table 4 elucidates the influence of source gas type, CH4, C2H6 and C2H2, which results in the change in main effects as the HMS value decreases from 35.27 to 34.58 nm. The decrease in the main effects of HMS value is explained by a lower damage layer thickness on TML which results using a higher molecular weight
R.H. Yeh et al. / Journal of Magnetism and Magnetic Materials 351 (2014) 76–81
Table 3 Main effects and their percentage contribution of each factor on Hc value.
Level 1 Level 2 Level 3 Effect Rank Contribution (%)
Factor A Source gas
Factor B Bias voltage
Factor C Anode voltage
Factor D Gas flow
5761 5650 5558 203 1 64.5
5617 5631 5721 104 2 19.9
5667 5629 5673 44 3 3.5
5634 5664 5670 35 4 2.3
Table 4 Main effects and their percentage contribution of each factor on HMS value.
Level 1 Level 2 Level 3 Effect Rank Contribution (%)
Factor A Source gas
Factor B Bias voltage
Factor C Anode voltage
Factor D Gas flow
35.27 34.89 34.58 0.68 1 59.8
34.73 34.86 35.15 0.43 2 24.2
34.96 34.90 34.88 0.08 3 0.8
34.89 34.95 34.90 0.06 4 0.5
Table 5 Main effects and their percentage contribution of each factor on SpSNR.
Level 1 Level 2 Level 3 Effect Rank Contribution (%)
Factor A Source gas
Factor B Bias voltage
Factor C Anode voltage
Factor D Gas flow
13.22 13.60 13.68 0.46 1 51.7
13.69 13.51 13.31 0.38 2 31.0
13.54 13.52 13.44 0.10 4 2.3
13.57 13.49 13.44 0.13 3 3.9
of source gas C2H2 deposition. The increase of bias voltage, from 100 to 300 V, results in an increase in main effects of HMS as its value increases from 34.73 to 35.15 nm. This increase is because the higher impact energy between overcoat and TML results from a higher substrate bias voltage. Consequently, the higher bias voltage could enhance the damage layer formation and therefore decrease the TML thickness. The increases of anode voltage, from 50 to 150 V, results in the HMS value slightly decreasing from 34.96 to 34.88 nm. There is no significant change on the main effect when the anode voltage is changed. The increases of gas flow rate, from 20 to 40 sccm, results in the HMS value only slightly changing from 34.89 to 34.90 nm. This variation is not significant over this gas flow rate range. Based on the analysis, it can be concluded that the source gas type is the most significant factor with its percentage contribution being 59.8% on HMS value. The bias voltage is the second most significant factor with its percentage contribution being 24.2% on HMS value. However, anode voltage and gas flow rate are not significant factors in the change in main effects of HMS. Table 5 shows the influence of source gas type, CH4, C2H6 and C2H2, which results in the change in main effects of SpSNR which increases from 13.22 to 13.68 dB. The increase in the main effects of SpSNR value is explained by the lower damage layer and HMS which results from source gas C2H2 deposition. The increases of bias voltage, from 100 to 300 V, results in a decrease in main effects of SpSNR value as it decreases from 13.69 to 13.31 dB. This decrease is due to the higher impact energy between overcoat
79
and TML resulting from the higher substrate bias voltage. Consequently, the higher bias voltage increases the damage layer happens which decreases the TML thickness. The increases of anode voltage, from 50 to 150 V, results in the SpSNR value slightly decreasing from 13.54 to 13.44 dB. There is no significant change on the main effect when the anode voltage is changed. The increases of gas flow rate, from 20 to 40 sccm, results in the SpSNR value slightly decreasing from 13.57 to 13.44 dB. This variation is not significant over this gas flow rate range. Based on the analysis, it can be concluded that the source gas type is the most significant factor with its percentage contribution being 51.7% on SpSNR value. The bias voltage is the second most significant factor with its percentage contribution being 31.0% on SpSNR value. However, anode voltage and gas flow rate are not significant factors in the change in main effects of SpSNR. Taguchi analysis of the experimental results shows that all, Hc, HMS and SpSNR are most affected by source gas and then secondly by bias voltage. Lower molecular weight (CH4 lower than C2H6 and C2H2) of the source gas is makes it easier to penetrate and damage the TML, indicating that it induces more interaction between the overcoat and the TML, which results in a higher Hc and HMS, and lower SpSNR. To obtain the lowest HMS and the highest SpSNR, the optimized source gas is C2H2 and bias voltage is 100 V, respectively. Anode voltage and gas flow rate are not dominant factors to affect the HMS and SpSNR. 3.2. Optimization of DLC deposition parameters simultaneously for HMS and SpSNR The objective of this study was to explore and determine how to optimize the control factors in order to achieve the best magnetic performance for future HDD product requirement. Overall, to obtain the highest SpSNR (lowest damage layer and HMS), the optimized source gas, bias voltage, anode voltage and gas flow rate are C2H2, 100 V, 50 V and 20 sccm, respectively. In general, HMS decreases and SpSNR increases with a decrease of the DLC overcoat thickness. Further to study the effect of overcoat thickness on HMS and SpSNR, by decreasing the overcoat thickness from 2.5 to 1.5 nm. At the same time, changing the overcoat deposition parameters from the current standard process to Taguchi's optimum design was compared. As shown in Fig. 4, at the same overcoat thickness, the optimum overcoat deposition condition provides about a 0.56 nm lower HMS than that of the current standard process condition. This is attributed to the lower damage layer of the TML which is obtained using the optimum overcoat deposition conditions. Low damage layer thickness results
Fig. 4. HMS of current process versus optimum deposition condition at different thicknesses.
80
R.H. Yeh et al. / Journal of Magnetism and Magnetic Materials 351 (2014) 76–81
in low HMS. Meanwhile, Fig. 5 demonstrates the SpSNR comparison between the standard current process and optimum deposition conditions at different overcoat thickness. The 0.51 dB SpSNR increase can be achieved due to the optimum deposition conditions using the same overcoat thickness. We believe that with the decrease of HMS resulting from the lower damage layer between overcoat and the TML, that the SpSNR value of magnetic recording performance is increased due to the lower HMS or damage layer thickness. Compared with the current deposition condition, this means that a 0.56 nm lower HMS and a 0.51 dB higher SpSNR can be achieved by using the optimum overcoat deposition conditions. Compared with the current standard process overcoat deposition condition, a lower Id/Ig (higher sp3 bonds) and a higher nanohardness of the overcoat properties are found from the Taguchi's optimum condition, as shown in Fig. 6. The decrease of Id/Ig ratio in DLC films induces more sp3 content, thus resulting in higher nanohardness. It is expected that this optimum overcoat deposition condition can be presumed to have better wear resistance because of the lower Id/Ig ratio and the higher nanohardness. Fig. 7(a) and (b) shows the CSS trend charts of current process versus optimum overcoat deposited condition with 2 nm overcoat thickness, respectively. Both of current process and optimum overcoat deposited condition can sustain a 100,000 cycles CSS test, but disk with the optimum overcoat deposited condition exhibits the lower stiction force. This may result from the overcoat strength, which affects the CSS performance. Finally, Fig. 8(a) and (b) illustrates the worn
Fig. 7. CSS trend charts of (a) current process versus (b) optimum overcoat deposited condition. (Overcoat thickness is 2 nm.)
Fig. 8. Worn surface of disk with DLC film of (a) current process versus (b) optimum overcoat deposited condition. (Overcoat thickness is 2 nm.)
surface of an OM image for disk wear DLC film of current and optimized DLC deposition films. Compared with current DLC deposition film, no wear track was observed in optimized DLC deposition film because it exhibited the higher nanohardness. 4. Conclusions
Fig. 5. SpSNR of current process versus optimum deposition condition at different thicknesses.
The purpose of this paper was to establish an efficient method for a robust overcoat deposition design process improvement on magnetic performance for hard disk media by utilizing the Taguchi method. The interface interaction study of nanoscale damage layers between PECVD overcoat deposition parameters on the top magnetic layer (TML) was achieved by using the four factors with a three level (L9) orthogonal array of experimental arrangements. Based on the ANOVA analysis, the source gas type is the most significant factor and its percentage contribution is 59.8% on HMS and 51.7% on SpSNR, respectively. The bias voltage is the second most significant factor as its percentage contribution is 24.2% on HMS and 31.0% on SpSNR. The robust process result obtained from the optimization of overcoat deposition parameters can achieve a 0.56 nm lower HMS, a 0.51 dB higher SpSNR and better wear resistance.
Acknowledgments
Fig. 6. Id/Ig and nanohardness of current process versus optimum overcoat deposited condition. (Overcoat thickness is 2 nm.)
The author would like to thank the National Science Committee of Taiwan for their financial support under Grant NSC100-2221-E231-006.
R.H. Yeh et al. / Journal of Magnetism and Magnetic Materials 351 (2014) 76–81
References [1] S.I. Iwasaki, Perpendicular magnetic recording—its development and realization, Journal of Magnetism and Magnetic Materials 324 (2012) 244–247. [2] M. Oshiki, Review of perpendicular magnetic recording research at Fujitsu, Journal of Magnetism and Magnetic Materials 324 (2012) 351–354. [3] K. Tagam, Looking back at perpendicular magnetic recording R&D in NEC, Journal of Magnetism and Magnetic Materials 324 (2012) 355–362. [4] R. Wood, Future hard disk drive systems, Journal of Magnetism and Magnetic Materials 321 (2009) 555–561. [5] S.N. Piramanayagam, K. Srinivasan, Recording media research for future hard disk drives, Journal of Magnetism and Magnetic Materials 321 (2009) 485–494. [6] G. Choe, M. Zheng, E.N. Abarra, B.G. Demczyk, J.N. Zhou, B.R. Acharya, K.E. Johnson, High-performance CoPtCrO perpendicular media: optimizing exchange coupling and anisotropy orientation dispersion, Journal of Magnetism and Magnetic Materials 287 (2005) 159–166. [7] M. Zhong, C. Zhang, J. Luo, X. Lu, The protective properties of ultra-thin diamond like carbon films for high density magnetic storage devices, Applied Surface Science 256 (2009) 322–328. [8] H.J. Richter, Density limits imposed by the microstructure of magnetic recording media, Journal of Magnetism and Magnetic Materials 321 (2009) 467–476. [9] X.D. Li, B. Bhushan, Micromechanical and tribological characterization of hard amorphous carbon coatings as thin as 5 nm for magnetic recording heads, Wear 220 (1998) 51–58. [10] M. Zhong, C. Zhang, J. Luo, X. Lu, The protective properties of ultra-thin diamond like carbon films for high density magnetic storage devices, Applied Surface Science 256 (2009) 322–328. [11] T.M. Chao, A.H. Tan, DLC deposition parameters optimization for head disk design interface with a thermal protrusion slider from tribological point of view, Materials and Design 48 (2013) 58–67. [12] C.A. Charitidis, Nanomechanical and nanotribological properties of carbonbased thin films: a review, International Journal of Refractory Metals and Hard Materials 28 (2010) 51–70. [13] A. Tomala, A. Pauschitz, M. Roy, Nanotribology of pulsed direct current magnetron sputtered diamond like carbon films, Surface Science 616 (2013) 60–70. [14] N. Sharma, N. Kumar, S. Dash, C.R. Das, R.V. Subba, A.K. Tyagi, B. Raj, Scratch resistance and tribological properties of DLC coatings under dry and lubrication conditions, Tribology International 56 (2012) 129–140. [15] H. Niakan, Q. Yang, J.A. Szpunar, Structure and properties of diamond-like carbon thin films synthesized by biased target ion beam deposition, Surface and Coatings Technology 223 (2013) 11–16.
81
[16] J.G. Choi, D.G. Hwang, J.R. Rhee, S.S. Lee, Soft magnetic properties of hybrid ferromagnetic films with CoFe, NiFe, and NiFeCuMo layers, Thin Solid Films 519 (2011) 8394–8396. [17] D. Mishra, A.K. Singh, P.C. Shyni, D. Sharma, A. Perumal, Enhanced soft magnetic properties in multilayer structured amorphous Fe-Ta-C films, Journal of Applied Physics 109 (2011) 07A304. [18] F.R. Marciano, R.P.C. Costa, O.D.A. Lima, A.O. Lobo, E.J. Corat, Trava AVJ. Tribological behavior under aggressive environment of diamond-like carbon films with incorporated nanocrystalline diamond particles, Surface and Coatings Technology 206 (2011) 434–439. [19] T.S. Santra, T.K. Bhattacharyya, P. Patel, F.G. Tseng, T.K. Barik, Structural and tribological properties of diamond-like nanocomposite thin films, Surface and Coatings Technology 206 (2011) 228–233. [20] R. Asakawa, S. Nagashima, Y. Nakamura, T. Hasebe, T. Suzuki, A. Hotta, Combining polymers with diamond-like carbon (DLC) for highly functionalized materials, Surface and Coatings Technology 206 (2011) 676–685. [21] A.S. Hamdy, Electrochemical behavior of diamond-like-carbon coatings deposited on AlTiC (Al2O3 þ TiC) ceramic composite substrate in HCl solution, Electrochimica Acta 56 (2011) 1554–1562. [22] M. Tan, Y. Yan, H. Zhang, X. Han, J. Zhu, J. Han, H. Ma, Corrosion protection of ultra-thin ta-C films for recording slider applications at varied substrate bias, Surface and Coatings Technology 203 (2009) 963–966. [23] J. Wang, G. Liu, J. Xu, X. Deng, L. Wang, Nanomechanical and Electrochemical properties of diamond-like carbon (DLC) films deposited by plasma enhanced chemical vapor, Plasma Science Technology 12 (2010) 461–465. [24] E. Salgueiredo, M. Amaral, M.A. Neto, A.J.S. Fernandes, F.J. Oliveira, R.F. Silva, HFCVD diamond deposition parameters optimized by a Taguchi matrix, Vacuum 85 (2011) 701–704. [25] Y.T. Yen, T.H. Fang, Y.C. Lin, Optimization of screen-printing parameters of SN9000 ink for pinholes using Taguchi method in chip on film packaging, Robotics Computer-Integrated Manufacturing 27 (2011) 531–537. [26] Y.M. Chiang, H.H. Hsieh, The use of the Taguchi method with grey relational analysis to optimize the thin-film sputtering process with multiple quality characteristic in color filter manufacturing, Computers Industrial Engineering 56 (2009) 648–661. [27] X. Liu, W. Clegg, B. Liu, Ultra low head–disk spacing measurement using dual beam polarisation interferometry, Optics Laser Technology 32 (2000) 287–291. [28] Y. Zhou, B. Liu, L. Li, Engineering issue study of triple harmonic method for in situ flying height analysis, Journal of Magnetism and Magnetic Materials 303 (2006) e120–e123. [29] B.R. Acharya, M. Zheng, G. Choe, M. Yu, P. Gill, E.N. Abarra, Oxidation behavior of sputtered CrN/AlN multilayer coatings during heat treatment, IEEE Transactions on Magnetics 41 (2005) 3145–3147.