Supporting Information

Size: px
Start display at page:

Download "Supporting Information"

Transcription

1 Supporting Information Understanding the Role of Nitrogen in Plasma-Assisted Surface Modification of Magnetic Recording Media with and without Ultrathin Carbon Overcoats Neeraj Dwivedi 1, Reuben J. Yeo 1, Nalam Satyanarayana 1, Shreya Kundu 1, S. Tripathy 2, C. S. Bhatia 1, * 1 Department of Electrical and Computer Engineering, National University of Singapore, Singapore Institute of Materials Research and Engineering (IMRE), A*STAR (Agency for Science, Technology, and Research), 3 Research Link, Singapore *Corresponding Author elebcs@nus.edu.sg (C. S. B.) Tel:

2 S1. Samples Description and Nomenclature Table S1: Description and nomenclature of the samples used in this work. Samples Starting Substrate Surface Modification COCs (nm) S-1 Specially prepared commercial media with no COC and no lubricant Using 100%Ar Plasma No S-2 Specially prepared commercial media with no COC and no lubricant Using 70%Ar+30%N 2 plasma No S-3 Specially prepared commercial media with no COC and no lubricant Using 50%Ar+50%N 2 plasma No S-4 Specially prepared commercial media with no COC and no lubricant Using 100%Ar plasma Yes, 1.5 nm S-5 Specially prepared commercial media with no COC and no lubricant Using 70%Ar+30%N 2 plasma Yes, 1.5 nm S-6 Specially prepared commercial media with no COC and no lubricant Using 50%Ar+50%N 2 plasma Yes, 1.5 nm S-7 Specially prepared commercial media with COC but no lubricant None Yes, 2.7 nm S2. Determination of Etched Thicknesses of Media The determination of etched thickness and etching rate in each plasma condition was performed on untreated bare magnetic media (BMM) samples. Before placing the samples in the processing chamber, a mask was used on each sample to obtain the modified and unmodified surfaces after the plasma treatment. The plasma modifications were performed using three different plasma conditions, namely 100%Ar, 70%Ar+30%N 2 and 50%Ar+50%N 2, while keeping other parameters constant. The total treatment time was kept at 3 min in each case. After the plasma modifications, the masks were removed, leading to the formation of steps. The step heights were then measured using tapping mode AFM to determine the thicknesses. We found that the etched thickness reduced from 2.3 nm for 100%Ar plasma to 1.3 nm for 70%Ar+30%N 2 plasma and 0.7 nm for 50%Ar+50%N 2 2

3 plasma. Table S2 summarizes the etched thickness and etching rate for each plasma condition. Table S2: Determination of etched thickness and etching rate for different plasma conditions. Plasma Condition Etched thickness after 3 min Etching rate (nm/min) 100%Ar 2.3 nm %Ar+30%N nm %Ar+50%N nm 0.23 The determination of etched thickness and etching rate in each plasma condition was performed on untreated bare magnetic media (BMM) samples. Before placing the samples in the processing chamber, a mask was used on each sample to obtain the modified and unmodified surfaces after the plasma treatments. The plasma modifications were performed using three different plasma conditions, namely 100%Ar, 70%Ar+30%N 2 and 50%Ar+50%N 2, while keeping other parameters constant. The total treatment time was kept at 3 min in each case. After the plasma modifications, the masks were removed, leading to the formation of steps. The step heights were then measured using tapping mode AFM to determine the thicknesses. We found that the etched thickness reduced from 2.3 nm for 100%Ar plasma to 1.3 nm for 70%Ar+30%N 2 plasma and 0.7 nm for 50%Ar+50%N 2 plasma. Table S2 summarizes the etched thickness and etching rate for each plasma condition. 3

4 S3. High Resolution Transmission Electron Microscopy (HRTEM) The COC thicknesses in samples S-4, S-5, S-6 and S-7, as measured by cross-section HRTEM, are labelled in their respective HRTEM images in Figures S3a 3d. The thicknesses of the COCs in samples S-4 to S-6 were each measured to be 1.5±0.1 nm, which confirmed the accuracy of our deposition rate calibration for pulsed DC sputtering of COC on media. The thickness of commercial COC in sample S-7 was measured to be 2.7±0.1 nm, which corresponds to the thickness value provided by our industrial collaborators. (a) S-4 (b) S-5 (c) S-6 (d) S-7 Figure S3: Cross-sectional HRTEM images showing the measurement of COC thicknesses in samples (a) S-4, (b) S-5, (c) S-6) and (d) S-7. 4

5 S4. Surface Roughness Surface Roughness (nm) R q R a S-1 S-2 S-3 S-4 S-5 S-6 S-7 Samples Figure S4: Variation of surface roughness, in nm, for different samples. The average roughness (R a, red bars) and root-mean-square roughness (R q, blue bars) of all the samples were measured by tapping mode AFM. Samples S-2, S-3, S-4, S-5 and S-6 are seen to have similar roughness. Commercial COC in sample S-7 gave the maximum roughness, which was nearly double the roughness value of samples S-2 to S-6. S5. Preparation of Additional Samples S-8 and S-9 and Their Tribological and Roughness Characteristics As can be seen, the etched thickness of media in samples S-3 and S-6 were found to be ~ 0.7 nm, while the etched thickness of media in samples S-2 and S-5 were measured to be 1.3 nm. To compare the effect of having similar etched thickness on roughness and tribological properties, two additional samples were prepared by performing 50%Ar+50%N 2 plasmaassisted surface modification of media, but with increased plasma treatment time so as to etch the ~ 1.3 nm thickness of the media. Thus, the plasma treatment time for sample S-8 (surface modification using 50%Ar+50%N 2 plasma but without COC) and sample S-9 (surface modification using 50%Ar+50%N 2 plasma nm COC) was kept at 5 min 34s, based on 5

6 the calibrated etch rate for this particular composition, to etch ~ 1.3 nm thickness of the media. After preparation, the surface roughness of samples S-8 and S-9 (Figure S5.1) was measured using tapping mode AFM (same equipment, which was used for other samples). The measured roughness of samples S-8 and S-9 was almost similar to that of samples S-3 and S- 6, respectively. Thus, increased etched thickness (and etching time) showed no influence on the roughness of the samples. Surface Roughness (nm) R q R a S-8 S-9 Samples Figure S5.1: Variation of surface roughness, in nm, for samples S-8 and S-9. R a is represented by red bars and R q is represented by blue bars. Furthermore, tribological measurements on samples S-8 and S-9 were also performed using the same equipment, characterization conditions and parameters (Figure S5.2). Again, the tribological properties of samples S-8 and S-9 in terms of friction and materials wear were almost similar to those of samples S-3 and S-6, respectively. Thus, increased etched thickness (and etching time) showed no influence on the tribological properties of the samples. If we include these samples for comparison, the following conclusions can be drawn: 1) Among samples S-1, S-2, S-3 and S-8, which were plasma-assisted surface modified media without 6

7 COCs, the best sample in terms of low friction and high wear resistance was sample S-2 which was prepared using 70%Ar+30%N 2 plasma treatment. 2) Among samples S-4, S-5, S-6 and S-9, which were plasma-assisted surface modified media with 1.5 nm COCs, the best sample in terms of stable and low friction, and high wear resistance was sample S-5, which was prepared using 70%Ar+30%N 2 plasma treatment followed by 1.5 nm COC deposition. 3) Overall, sample S-5 (even when compared with sample S-7, which was commercial media with a thicker COC of 2.7 nm) showed the best tribological properties, indicating that the composition of 70%Ar+30%N 2 is optimum for plasma treatment. Given the fact that samples S-8 and S-9 showed no improvement in roughness and tribological properties (they even showed slightly poorer properties), we ignored these samples for further study and concentrated on samples S-1 to S-7. Hence, the comprehensive study was conducted only on samples S-1 to S-7, out of which the constant 3 min plasma treatment time was used for samples S-1 to S-6. Coefficient of Friction S-8 Coefficient of Friction S Ball Number of Cycles Number of Cycles (S-8) Wear Track (S-8) Ball (S-9) Wear Track (S-9) Figure S5.2: Frictional curves and optical images of balls and wear tracks for samples S-8 and S-9 after ball-on-disk tribological tests. 7

8 S6. Magnetic Properties In this work, CoCrPt-oxide-based bare magnetic media without any plasma treatment and without COC (labelled as BMM) were used as the starting substrate for the preparation of all samples. Since the samples were prepared as a consequence of plasma-assisted surface modification followed by COC deposition, it is hence important to examine any change in the magnetic properties in the samples of interest. The magnetic properties of samples S-1 to S-6 and sample BMM were measured using a custom-made magneto-optic Kerr effect (MOKE) system. Here, sample BMM was used as the reference sample for evaluating the magnetic properties of the modified media samples. BMM S-1 S-4 BMM S-2 S (a) 1.0 (b) Kerr Rotation Kerr Rotation Magnetic Field (T) BMM S-3 S Magnetic Field (T) BMM S-4 S-5 S (c) 1.0 (d) Kerr Rotation Kerr Rotation Magnetic Field (T) Magnetic Field (T) Figure S6: MOKE hysteresis loops of samples S-4 to S-6 and reference sample BMM under an applied switching magnetic field. 8

9 The magnetic properties of the samples with COCs were first compared with their uncoated counterparts which were subjected to the same plasma-assisted surface modification conditions. The magnetic hysteresis loops of samples S-1 and S-4, S-2 and S-5, as well as S-3 and S-6 are presented in Figures S6a, S6b and S6c, respectively. In addition, the magnetic hysteresis loop of reference sample BMM is included in each plot for comparison purpose. Finally, the magnetization hysteresis loops of the plasma-modified samples with COCs (samples of interest) and reference sample BMM are plotted together in Figure S6d. From Figures S6a, S6b and S6c, it is evident that the magnetic properties are consistent with and without COCs for the same plasma treatments. When the samples with COCs and reference sample BMM are plotted together (Fig. S6d), we found that the hysteresis loops for samples S-5 and S-6 are almost identical and overlap with each other. Compared to sample BMM, the magnetic properties of samples S-5 and S-6, such as the coercivity and switching field distribution, show negligible differences. In contrast, sample S-4 shows a visible change in magnetic properties with respect to BMM. Similar results were observed for plasma-modified media without COCs, i.e. samples S-1 to S-3. The change in magnetic properties may be attributed to relatively more etching in 100%Ar plasma as compared to the 70%Ar+30%N 2 and 50%Ar+50%N 2 plasma. Thus, inclusion of nitrogen in argon plasma during the surface modification process did not adversely affect the magnetic properties of the media in our experiments. S7. Surface Energy and Surface Polarity Measurements Contact angle measurements with two liquids water and diiodomethane were performed to estimate the surface energy and surface polarity of different samples. The measurements were carried out using a VCA Optima (AST Product Inc.) goniometer at room temperature. The contact angles of 0.5 l droplets for each liquid were measured 10s after the 9

10 droplets were formed on the sample surfaces after being expelled from the needle, which is sufficient time for the interaction of droplets with the solid surfaces. A total of five measurements were performed on each sample for repeatability and accuracy. Contact Angle (Degree) (a) Water Diiodomethane Dispersion Polar Total S-4 S-5 S-6 S-7 Samples Surface Energy (mj/m 2 ) (b) (c) S-4 S-5 S-6 S-7 Samples Surface Polarity (%) S-4 S-5 S-6 S-7 Samples Figure S7: Variations of (a) water and diiodomethane contact angles, (b) dispersion, polar and total surface energies, and (c) surface polarities for surface modified media samples with COCs and commercial media sample with its original COC. Figure S7a shows the water and diiodomethane contact angles for the different samples. The water contact angles were found to be always higher than the diiodomethane contact angles for all the samples. To calculate the dispersion and polar components of the surface energy of solids, the contact angles of diiodomethane (which has only the dispersion component) and water (which has both polar and dispersion components) were analyzed. Since the surface 10

11 roughness of all the samples is very low, hence, we can disregard any effect of surface roughness on the contact angles of the liquid drop. The Good-Girifalco-Fowkes-Young method was used to calculate the respective surface energies of these samples 1-3. Young s equation for the equilibrium of a liquid drop under the action of three surface tensions is given by: sv cos, (1) sl lv where, and are the surface tensions at the solid-vapor, liquid-vapor and solid-liquid interfaces respectively, and is the contact angle of the liquid on the solid surface. Fowkes proposed that the dispersion interaction between two phases such as solid and liquid,, can be given by the geometric mean of the individual dispersion components of two phases (for example, solid,, and liquid, ) as: (2) Since a non-polar liquid with only dispersion surface free energy ( ) interacts with only the dispersion component of the solid, the dispersion component of the surface free energy of the solid can be estimated from the contact angle of the non-polar liquid: d 2 d lv (1 cos ) sv, (3) 4 Similar to the dispersion interaction, the polar interaction between two phases such as solid and liquid,, can be given by the geometric mean of the individual dispersion components of two phases (solid,, and liquid, ) as: (4) 11

12 According to Berthelot, the work of adhesion between a solid and a liquid can be estimated as the geometric mean of the cohesive works of the solid and liquid as: W W W 2 (1 cos ), (5) sl ss ll sv lv sv lv sl lv The liquid with both dispersion ( ) and polar components ( ) interacts with both the dispersion and polar components of the solid surface and follows the relation: d ( 1 cos ) L (6) lv sl P L sl Using equations (2) and (4), equation (6) can be re-written as: d d p p (1 cos ) 2 2, (7) lv sv lv sv lv Since all the other parameters are known, the polar surface free energy of solid,, can be estimated from equation (7). Employing equations (3) and (7), we calculated the dispersion, polar and total surface energies of solid for the samples of interest, which are presented in Figure S7b. It is evident from Figure S7b that samples S-4, S-5 and S-6 exhibited almost similar total surface energy values (~ 44.5 mj/m 2 ), but the surface polarity values (as summarized in Figure S7c) were significantly lower in samples S-5 (~ 6 %) and S-6 (~ 8 %) than in sample S-4 (~ 16 %). Since the COC fabrication method and thickness in samples S-4, S-5 and S-6 were kept constant, the observed reduced surface polarity in samples S-5 and S-6 can be attributed to the introduction of nitrogen in the argon plasma during plasma-assisted surface modification of media. Among samples S-5 and S-6, a slightly lower surface polarity was observed in sample S-5. The surface polarity could influence the corrosion and oxidation of the media. With a lower surface polarity, higher oxidation and corrosion resistance was observed in 12

13 sample S-5 (as confirmed by electrochemical corrosion and XPS analyses, which is discussed in the main manuscript). In comparison, commercial media with conventional COC (S-7) showed slightly higher surface energy (~ 50 mj/m 2 ) and higher surface polarity (~23.7 %). S8. X-ray Photoelectron Spectroscopy (XPS) Analysis Peak fitting results of XPS C 1s, Co 2p 3/2, and N 1s core level spectra of various samples. S8.1. Analysis of C 1s core level spectra Table S8.1.1: C 1s core level spectra. All samples with COCs showed the presence of sp 2 C, sp 3 C, C-O and C=O bonding. Peak Energy (ev) Bond type ± 0.1 sp 2 C ± 0.1 sp 3 C ± 0.1 C-O ± 0.1 C=O Remark All samples with COCs (S-4, S-5 and S-6) All samples with COCs (S-4, S-5 and S-6) All samples with COCs (S-4, S-5 and S-6) All samples with COCs (S-4, S-5 and S-6) Table S8.1.2: C 1s core level spectra. Estimation of sp 3 C bonding fractions for different samples. The sp 3 C bonding fraction was calculated using an area ratio method. Samples sp 3 C (%) S-4 ~ 20.5 S-5 ~ 22.5 S-6 ~ 22.0 Remark All samples with COCs (S-4, S-5 and S-6) seem to possess almost similar sp 3 C bonding fraction. 13

14 S8.2. Analysis of Co 2p 3/2 core level spectra Table S8.2.1: Co 2p 3/2 core level spectra. All samples showed the presence of Co-Co (metallic cobalt), Co-oxide and Co-oxide/hydroxide bonding. However, additional Co-N x bonding was observed in those samples which were treated using mixed Ar+N 2 plasma with and without COCs. Peak Energy (ev) Bond type Remark ± 0.05 Co-Co (metallic cobalt) All samples (S-1, S-2, S-4, S-5 and S-6) ± 0.05 Co-oxide (as in Co 2 O 3 ) All samples (S-1, S-2, S-4, S-5 and S-6) ± 0.1 Co-oxide/hydroxide (as in Co-O, Co 3 O 4, All samples CoOOH) (S-1, S-2, S-4, S-5 and S-6) ± 0.1 Co-oxide/hydroxide (as in Co-O, Co 3 O 4, All samples Co(OH) 2 ) (S-1, S-2, S-4, S-5 and S-6) 781.8± 0.1 Co-N x Samples S-2, S-5 and S-6 Table S8.2.2: Co 2p 3/2 core level spectra. Estimation of Co-metallic and Co-(oxide + oxide/hydroxide) contents from different samples. An area ratio method was used to determine these fractions. Samples Co (Metallic) (%) Co (Oxide + Oxide/Hydroxide) (%) S-1 ~ 14.1 ~ 85.9 S-2 ~ 24.0 ~ 61.0 S-4 ~ 60.3 ~ 39.7 S-5 ~ 72.0 ~ 24.0 S-6 ~ 68.0 ~ 28.0 Remark This sample is mostly in its oxidized state and has minimum (maximum) metallic (oxide) content of Co. This sample is better than sample S- 1 in terms of relatively higher oxidation protection. This sample showed higher oxidation protection to underlying media than samples S-1 and S-2 due to the presence of 1.5 nm COC. Among all samples, maximum oxidation protection to underlying media was observed in sample S-5. This is the best sample of this work. This sample showed higher oxidation protection than sample S- 4 but lesser than sample S-5. S-7 ~68 ~32.0 Figures are not given in manuscript. 14

15 S8.3. Calculation of the absolute and relative percentage changes in the metallic Co content for the surface modified media samples In order to reveal the role of nitrogen in plasma for surface modification of media, we compare the metallic content of Co in samples S-1, and S-2 for surface modified media without COC, and in samples S-4, S-5 and S-6 for surface modified media with 1.5 nm COCs. We calculated both the absolute and relative differences in the metallic content of Co to distinguish the oxidation protection capability between each sample. 1) For the calculation of % relative difference in the metallic content of Co between samples S-1 and S-2 (surface modified media without COCs), the metallic content of Co of sample S-1 was used as the reference. 2) For the calculation of % relative difference in the metallic content of Co between samples S-4 and S-5 (surface modified media with 1.5 nm COCs), the metallic content of Co of sample S-4 was used as the reference. 3) For the calculation of % relative difference in the metallic content of Co between samples S-4 and S-6 (surface modified media with 1.5 nm COCs), the metallic content of Co of sample S-4 was used as the reference. 4) For the calculation of % relative difference in the metallic content of Co between samples S-5 and S-6 (surface modified media with 1.5 nm COCs), the metallic content of Co of sample S-6 was used as the reference. The formula used for the calculation of the % relative difference of the metallic Co content is given as: 15

16 Table S8.3: Percentage change in the metallic content of Co for different samples to understand the role of composition for oxidation protection. The table provided summarizes the absolute and relative percentage changes in the metallic content of Co between different surface modified magnetic media samples with and without COCs. Sample % Co (Metallic) % Absolute Difference % Relative Difference Remark S S Addition of nitrogen in plasma for surface modification of media enhances the oxidation protection of media. S S Inclusions of nitrogen in mixed Ar+N 2 plasma for surface modification of media assist to reduce the oxidation of media. S (when sample S-4 is used as a reference) 4.0 (when sample S-6 is used as a reference) ~ 19.4 (when sample S-4 is used as a reference) ~ 5.9 (when sample S-6 is used as a reference) The role of nitrogen in plasma for surface modification of media is clear. This is the best proposed modification of this work and provided maximum protection to underlying media against oxidation. Based on the above results, we can make the following comparisons: 1) Comparison between samples S-1 and S-2 (surface modified media with COCs): The absolute and relative differences in the metallic content of Co were found to be ~ 9.9 % and ~ 70.2 %, respectively. Higher metallic content of Co, and hence higher oxidation protection, was observed in sample S-2 compared with sample S-1. 2) Comparison between samples S-4 and S-5 (surface modified media with COCs): The absolute and relative differences in the metallic content of Co were found to be ~ 11.7 % and ~ 19.4 %, respectively. Higher metallic content of Co, and hence higher oxidation protection, was observed in sample S-5 compared with sample S-4. 3) Comparison between samples S-4 and S-6 (surface modified media with COCs): The absolute and relative differences in the metallic content of Co were found to be ~ 7.7 % and ~ 12.8 %, respectively. Higher metallic content of Co, and hence higher oxidation protection, was observed in sample S-6 compared with sample S-4. 16

17 4) Comparison between samples S-5 and S-6 (surface modified media with COCs): The absolute and relative differences in the metallic content of Co were found to be ~ 4.0 % and ~ 5.8 %, respectively. Higher metallic content of Co, and hence higher oxidation protection, was observed in sample S-5 compared with sample S-6. Overall, the sample sequence in terms of the extent of oxidation protection is as follows: S-5 > S-6 > S-4 > S-2 > S-1. S8.4. Analysis of N 1s core level spectra Table S8.4.1: N 1s core level spectra. Samples S-2, S-5 and S-6 showed a variety of bonding of nitrogen with metal, oxygen and carbon. Peak Energy (ev) Bond type Remark 397.2±0.2 Co-N x & Cr 2 -N Samples S-2, S-5 and S ±0.2 Metal-N-pyridyl Samples S-2, S-5 and S Graphitic C=N Samples S-2, S-5 and S ±0.1 N-O x Samples S-2, S-5 and S N-pyridyl/pyridinic-N Samples S-5 and S-6 Table S8.4.2: N 1s core level spectra. Estimation of bonding fraction of pyridine-like structure (peak ev) for samples S-5 and S-6. An area ratio method was used to determine these fractions. Samples Pyridine-like bonding (%) S S

18 References 1. Good, R. J. & Girifalco, L. A. A theory for estimation of surface and interfacial energies.iii Estimation of surface energies of solids from contact angle data. J. Phys. Chem. 64, 561 (1960). 2. Fowkes, F. M. Determination of interfacial tensions, contact angles and dispersion forces in surfaces by assuming additivity of intermolecular interactions in surfaces. J. Phys. Chem. 66, 382 (1962). 3. Van Oss, C. J., Chaudhury, M. K & Good, R. J. Interfacial Lifshitz-van der Waals and polar interactions in macroscopic systems. Chem. Rev. 88, 927 (1988). 18