Compositional depth profile analysis of coatings on hard disks by X-ray photoelectron spectroscopy and imaging

Transcription

1 Surface and Coatings Technology 176 (2003) Compositional depth profile analysis of coatings on hard disks by X-ray photoelectron spectroscopy and imaging Jianxia Gao*, Erjia Liu, David Lee Butler, Aiping Zeng Centre for Mechanics of Microsystems, School of Mechanical and Production Engineering, Nanyang Technological University, 50 Nanyang Avenue, Singapore , Singapore Received 19 September 2002; accepted in revised form 10 January 2003 Abstract A hard disk medium is typically composed of several layers including the magnetic recording layer, a buffer layer, as well as a wear protective layer. In the work presented here, the hard disks analysed have a total of five layers with the uppermost layer being the lubricant. The second layer is diamond like coating and this is followed by the magnetic layer consisting of an alloy of cobalt and other elements. The fourth layer is a buffer composed of an alloy of chromium, vanadium and molybdenum with the final layer being a nickel transition layer doped with phosphorus. These multilayers were subjected to numerous etchings by argon ions. The chemical structures of these layers were analysed with an X-ray photoelectron spectroscope (XPS) after each etching. Combining the XPS spectra with XPS imaging it is possible to determine the depth distribution of elements in the hard disk coating. In addition, it is also shown that XPS imaging can be employed to monitor the thickness of all multilayers Elsevier B.V. All rights reserved. PACS: Fy; Ac; qf; 75.5.Cc Keywords: Hard disk; Multilayers; X-ray photoelectron spectroscopy; Image 1. Introduction X-ray photoelectron spectroscopy (XPS) uses highly focused monochromatised X-rays to probe the material of interest w1,2x. The energy of the photo emitted electrons ejected by the X-rays are specific to the chemical state of the elements and compounds present, i.e. bound-state or multivalent states of individual elements can be differentiated. Specifically, the small-spot XPS technique can provide the following information about samples: (1) elemental identification and quantification; (2) chemical functional group identification and quantification; (3) layer-by-layer depth profiling. In addition, using the parallel XPS imaging technique these chemical state images are acquired in only a few minutes with a resolution of approximately 3 mm w3,4x. With parallel imaging, photoelectrons are collected from the whole field of view simultaneously. A combination of *Corresponding author. Tel.: q ; fax: q address: mjxgao@ntu.edu.sg (J. Gao). lenses before and after the energy analyzer focuses the photoelectrons on a channel plate detector. The hemispherical analyzer permits electrons with only a narrow band of energy to reach the detector. In the field of hard disk technology, a hard disk medium is normally composed of a lubricant layer, an overcoat layer, a magnetic recording layer, transition layers and substrate w5,6x. Currently the liquid monolayer organic films Z-dol (X CF O (CF CF O) p (CF O) CF X) or perfluoropolyether are used as 2 q 2 lubricants of hard disks to reduce friction while amorphous hydrogenated diamond-like carbon (DLC) has been the choice of overcoats over the past 10 years w6,7x. DLC coatings combine the advantages of high hardness with low friction and low wear rates. That is, an extremely thin DLC film being deposited on magnetic layer is mainly used in a wear protective role. It is well known that magnetic layers of hard disk are composed of several kinds of elements such as cobalt (Co), platinum (Pt), chromium (Cr) and so on. In order to obtain hard disks that exhibit a long life, it is /03/$- see front matter 2003 Elsevier B.V. All rights reserved. doi: /s (03)

2 94 J. Gao et al. / Surface and Coatings Technology 176 (2003) necessary to investigate the elemental distribution and interface properties. However, to the best of authors knowledge, no papers discuss the usage of XPS image in researching the depth profile of coating of hard disks. In our work, argon ions beam was used as etching beam, the XPS image and electronic profile obtained by XPS method were used to determine the depth information of coating of a hard disk. 2. Experimental Nickel film, as a transition layer, was deposited on a metal substrate (alloy of aluminium and magnesium) by a radio frequency (RF) magnetron sputtering system equipped with a cryogenic pump. The base vacuum y7 pressure was less than 2.0=10 Torr. The sputtering was conducted with a sintered nickel target that was 3 inches in diameter and 99.99% in purity. The depositions were carried out in pure Ar 99.99%. The deposition rate was measured using quartz crystal of 0.01 nmys accuracy. The applied RF powers were in the range of W. The substrate was initially at room temperature and no effort was made to control its temperature during deposition. The chromium (Cr), molybdenum (Mo) and vanadium (V) were co-sputtered on the nickel layer by the same magnetron sputtering instrument and form the buffer layer. Similarly, the cobalt (Co), platinum (Pt), tantalum (Ta) and chromium were co-sputtered on the buffer layer by the same instrument and form the magnetic layer. The diamond-like carbon (DLC) film was then synthesized on the magnetic layer using the magnetron sputtering system. Finally, Z-dol lubricant film was dip-coated on the DLC layer, and its thickness was controlled by adjusting the concentrations of the lubricant or by adjusting the removal velocity from the solution. So, the coating of the hard disk contains several layer of films such as Z-dol lubricant, DLC overcoat, magnetic, buffer, and the transition layer. These coating had been etched many times by argon ions beam and analysed by XPS method after each etching. The distributions and chemical state of elements in the coating were also measured by XPS image method within a few minutes after each etching. The etching and analysis process were performed in an ultrahigh vacuum chamber of Kratos Ultra XPS system (Kratos company, UK). The base pressure of the y9 chambers is 2=10 Torr. This chamber mainly con- tained a dual anode (AlyMg Ka 1,2) and a monochromatised Al Ka1,2 X-ray source, and a hemispherical electron energy analyzer. During etching, the argon y7 pressure in the chamber is approximately 1=10 Torr. Then the hard disk was exposed to the X-ray irradiation from a monochromated Al Ka1,2 X-ray source (hns ev) operated at 150 W (15 kv, 10 ma). The photoelectrons induced by X-ray were filtered by hemispherical analyzer, and finally recorded by the eight channeltron multi-detector (for XPS spectrum) or by the microchannel plate and CCD camera detector (for imaging). The set of etching time to each layer depends on the brightness of the imaging of an element and the peak height of the element in the XPS spectrum for this layer. For example, the fluorine only exist in the lubricant layer, the XPS imaging and XPS spectrum of fluorine (F 1s) can be obtained after etching this layer for 2 s. This process can be repeated for several times till that its XPS imaging became dark while the peak height of F 1s in the spectrum became very low. Then a relation between etching time and peak height of F 1s can be obtained and the thickness of the lubricant can be determined. It should be mentioned that carbon exists in both the lubricant layer and DLC layer, however, the content of carbon is higher in the DLC layer than in the lubricant layer. If its XPS image (C1s) becomes brighter while its peak height in the spectrum also becomes higher, it means that the lubricant layer is already removed and the DLC layer was revealed, so the measurement of exact thickness of the lubricant depends on XPS imaging and spectrum of both fluorine and carbon. 3. Results and discussion 3.1. Etching and separation of multilayer The multilayers on a hard disk were etched gradually, and the composition in each layer was obtained by XPS technique. The first layer is Z-dol lubricant and the XPS spectrum in Fig. 1a shows that the lubricant is composed of carbon, oxygen and fluorine before etching. The peak position of carbon (C 1s), fluorine (F 1s) and oxygen (O 1s) are , and ev, respectively. Fig. 1b is the XPS spectra of fluorine (F 1s) before etching, and after etching for 6 s. The second layer is DLC overcoat layer and the imaging of the layer whose binding energy is ev is shown in Fig. 2. Fig. 2a c shows the carbon layer before etching, and after etching for 35 and 40 s by argon beam, respectively. With increasing etching time, the depth, which was etched, also increased. After approximately 40 s, the DLC layer was removed and the XPS image of carbon 1s peak was generated (as shown in Fig. 2). The centre region in Fig. 2b changed colour from bright to gray, it shows the area of the DLC layer was nearly removed after etching for 35 s. The dark colour in the centre region in Fig. 2c means the DLC film was removed completely after etching for 40 s. The brightness in this XPS image depends on the content of carbon; for the higher content of carbon, the brightness of the area which contained carbon in this image is stronger. In detail, the colour scale on the right

3 J. Gao et al. / Surface and Coatings Technology 176 (2003) Fig. 1. The XPS spectra of lubricant of a hard disk. (a) Obtained before etching. C 1s: carbon, ev; F 1s: fluorine, ev; the O 1s: oxygen, ev. (b) The XPS spectra of fluorine before etching, and after etching for 6 s. hand side of the image represents an intensity gradient on the XPS image. When in the imaging mode the Ultra uses a microchannel plate and a phosphorus screen to detect the photoelectrons. The photoelectrons hit the phosphor screen causing visible flashes of light which are recorded by a CCD camera, then the lateral distribution and number of light flashes are stored by the computer. At the end of the experiment the image is displayed with an intensity scale which is equal to the pixel intensity (the number of light flashes recorded per pixel). This is then converted to a colour scale with low pixel intensities dark and high intensities bright. The colour is therefore a qualitative indication of the amount of an element present on the sample surface. It should be noted that, although the density of the argon beam used as etchant is highest in the centre of the argon beam, the focus of the argon beam is not ideal. This results in an asymmetric lateral distribution in those centres of all images. The third layer is the magnetic layer used as recording media, that is an alloy of several metals, the XPS spectrum in Fig. 3 shows that layer is composed of cobalt (Co), platinum (Pt), tantalum (Ta) and chromium (Cr). Fig. 4 is the XPS image of Pt 4f peak whose binding energy is 71.3 ev. The magnetic layer was just disclosed in the centre of Fig. 4a. The Fig. 4b and c show the magnetic layer has been etched for 15 and 25 s, respectively. The increasing area of light region means the revealed area of the magnetic layer was increased in Fig. 4b. Clearly, with increasing etching time, this layer also became thinner and thinner, and finally, after approximately 25 s, this layer was removed when a dark region appeared in the centre area of the image showed in Fig. 4c. The XPS image of the Co 3p peak and Ta 4f peak are similar to that of Pt 4f. However, the chromium is contained both in the magnetic layer and in the buffer layer, its image is different from that of Pt 4f, and it will be discussed later. The fourth layer, as a buffer layer, is alloy of chromium, molybdenum (Mo) and vanadium (V). Fig. 5 is the image of Mo 3d peak whose binding energy is ev. Fig. 5a shows the buffer layer was just disclosed in the centre, Fig. 5b and c indicate the buffer layer had been etched for 30 and 60 s, respectively. The image of V 2s peak is similar to that of Mo 3d. Fig. 5b indicates the light region was enlarged and part of light region has changed into grey colour near the centre of the image. Thus, more area of the buffer layer was revealed and part of this layer was nearly removed near the centre. The dark region in the centre of Fig. 5c means that the fourth layer was removed in the area by argon etching for 60 s while the lighter region was still the present layer. In addition, the image of Cr 2p (binding energy: ev) is shown in Fig. 6. It confirmed the element chromium existed both in the magnetic layer and in the buffer layer. It should be mentioned that the image brightness of chromium in Fig. 6a c was weaker than that of Fig. 6d f. It means Fig. 6a c belong to the magnetic layer and Fig. 6d f belong to the buffer layer, and the content of chromium in the magnetic layer is lower than that of the buffer layer. The light region in Fig. 6a means the magnetic layer was just disclosed in the centre, Fig. 6b and c shows that the lighter region became larger after the magnetic layer had been etched for 15 and 25 s, respectively. Thus the disclosed area of magnetic layer increased with increasing etching time. The light region in Fig. 6d means the buffer layer was just disclosed in the centre.

4 96 J. Gao et al. / Surface and Coatings Technology 176 (2003) Fig. 2. The imaging of carbon overcoat layer whose binding energy is ev. (a) The carbon layer before etching by argon beam. (b) and (c) The carbon layer after etching for 35 and 40 s, respectively. Fig. 6e and f shows the buffer layer had been etched for 45 and 60 s, respectively. Compared with Fig. 6d, the light region was enlarged and part of the region became grey again in the centre of Fig. 6e, this implies that the disclosed area of buffer layer increased and part of buffer layer was nearly removed. With increasing etching time, more area of light region changed into gray and even dark colour in the centre of Fig. 6f, the buffer layer had therefore begun to be removed in the centre of the image. The fifth layer is nickel transition layer. The XPS images of Auger peak of nickel (Ni LMMa) are shown in Fig. 7. The lighter region in Fig. 7a is small and is surrounded by large black area, it means the Ni layer is

5 J. Gao et al. / Surface and Coatings Technology 176 (2003) Fig. 3. XPS spectrum of magnetic layer used as recording media, and it is composed of Co, Pt, Ta and Cr. just disclosed in the lighter region by etching beam and the dark area still belongs to the fourth layer. Fig. 7b and c shows the nickel layer was etched 1000 and 2240 s, respectively. They indicate that, during the etching process, the lighter region became brighter as the region expands. So the disclosed area of nickel layer was enlarged gradually. The stack of those layers was fabricated on the substrate of aluminium and magnesium alloy, and the etching depth had not yet reached to the substrate Thickness measurement From the above discussion, each layer thickness can be obtained by measuring the depth distribution of an element, which only exists in that specific layer. The first, the etching rate for each layer can be calculated by the transport of ions in matter (TRIM) simulation w8,9x. The second, the etching time for each layer can be calculated and finally the thickness of layers can be measured. TRIM is the most comprehensive program and will accept complex targets made of compound materials with up to eight different layers. It can be used to calculate both the final 3D distribution of the ions and also all kinetic phenomena associated with the ion s energy loss: target damage, sputtering, ionization, and phonon production. The fluorine element only exists in the lubricant layer; it can be used to measure the thickness of lubricant. From the peaks F 1s at ev in Fig. 1b, it is estimated that the lubricant layer was completely removed after etching it for 6 s, the etching rate calculated by TRIM simulation is 0.83 nmys, so the thickness of the lubricant is approximately 5 nm. It should be noted that standard binding energy of F 1s peak is ev and this is related to the fluorine carbon bond. The energy shift of the experimental value is due to two reasons, the first is the charge accumulation of hard disk during XPS measurement; the second is the binding energy of fluorine carbon atoms in Z-dol lubricant is higher than that of fluorine fluorine atoms (standard binding energy of F 1s). The carbon elements exists both in lubricant and DLC layer, so the thickness of DLC layer can be measured by the difference between the depth of fluorine and that of carbon elements. Fig. 2a is the image of the carbon layer after the lubricant layer is just removed. The binding energy is ev. The centre region in Fig. 2c changed colour from light to dark, it means the DLC layer was removed completely in the centre after this layer had been etched for 40 s. The etching rate of the layer calculated by TRIM is 0.33 nmys, thus the thickness of DLC layer can be determined is 13.2 nm. Since the platinum element only exists in the magnetic layer, it was used to trace the thickness of the layer. The result shown in Fig. 4a is the chemical state image of element Pt 4f component whose binding energy is 71.3 ev, this suggests that DLC layer was just removed and the magnetic layer was disclosed in the centre. The bright area in the picture contains high content of platinum while the gray area in the picture has low content of platinum. The black area means there is no platinum in the area. With increment of etching time, the third layer has been removed gradually. Fig. 4c is also the XPS image of Pt 4f component, it shows the centre region in the image became black, so the platinum element on the region of the layer was removed while the magnetic layer was just removed. From Fig. 4a to c, the etching time is 25 s. The etching rate is 0.99 nmy s calculated by TRIM simulation according to the composition of this layer, so the thickness is 24.8 nm according to the etching time. Molybdenum element is only contained in the buffer layer of the hard disk, so it can be used to trace the depth of the layer and the results are shown in Fig. 5. The binding energy of Mo 3d is ev, and it is used as the detecting energy of the XPS image. The bright area in Fig. 5a indicates that the magnetic layer was removed and the buffer layer was shown in the central area of the image. Similar to the magnetic layer, the argon ion beam etched the buffer layer until a black region appeared in the centre of the layer. The imaging in Fig. 5c shows a dark region in the centre of the image and it means that the centre of the layer was removed. The etching rate of the layer is nmys calculated by TRIM simulation. The etching time is 60 s, so the thickness of the layer is 35.7 nm. The fifth layer, nickel as a transition layer between the substrate and the buffer layer, is a thick film. The peak energy of Ni LMMa component used in Fig. 7 is the ev, the bright area means high content nickel

6 98 J. Gao et al. / Surface and Coatings Technology 176 (2003) Fig. 4. The XPS image of Pt 4f peak whose binding energy is 71.3 ev. (a) The magnetic layer was just disclosed in the centre; (b) after etching the magnetic layer for 15 s; (c) after etching the magnetic layer for 25 s. while the grey area around the bright area means it is at the edge of the nickel layer which was disclosed. The etching rate for this layer is 4.07 nmys obtained by TRIM simulation. From Fig. 7a to c, it had been etched for 2240 s and the depth, which was etched, is 9111 nm. Fig. 7c also shows high content nickel, so the layer had not been etched completely. The calibration of these thickness values obtained by the above method can be performed by high resolution scanning electron microscopy (HRSEM) technique. That

7 J. Gao et al. / Surface and Coatings Technology 176 (2003) Fig. 5. The XPS the images of Mo 3d peak whose binding energy is ev. (a) The buffer layer was just disclosed in the centre; (b) and (c) the buffer layer was etched for 30 and 60 s, respectively. is, as an XPS imaging of an element belonging to a layer become dark while its XPS spectrum become flat, then the etching region of the layer can be measured by HRSEM and its thickness can be obtained. Comparing this value with that value obtained from XPS imaging and spectrum, the latter can be calibrated by HRSEM technique. 4. Conclusions XPS spectrum and imaging techniques are effective

8 100 J. Gao et al. / Surface and Coatings Technology 176 (2003) Fig. 6. The image of Cr 2p whose binding energy is ev. (a) Magnetic layer was just disclosed in the centre; (b) and (c) after etching the magnetic layer for 15 and 25 s, respectively. (d) The buffer layer was just disclosed in the centre; (e) and (f) the buffer layer was etched for 45 and 60 s, respectively.

9 J. Gao et al. / Surface and Coatings Technology 176 (2003) Fig. 7. The image of Ni LMMa whose binding energy is ev. (a) The transition layer was just disclosed in the centre; (b) and (c) after etching the transition layer for 1000 and 2240 s, respectively. to be used to characterize the depth distribution of the elements in the coating of hard disks. For the coating in this study, the thickness of Z-dol lubricant layer was approximately 5 nm, that of DLC hard coat, CoPtCrTa recording layer, and CrMoV buffer layer were 13.2, 24.8 and 35.7 nm, respectively, and the Ni transition layer was thicker than 9111 nm. References w1x H. Martinez, C. Auriel, M. Loudet, G. Pfister-Guillouzo, D. Gonbeau, Appl. Surf. Sci. 125 (3 4) (1998) w2x A. Cohen Simonsen, M. Schleberger, S. Tougaard, J.L. Hansen, A. Nylandsted Larsen, Thin Solid Films 338 (1 2) (1999)

10 102 J. Gao et al. / Surface and Coatings Technology 176 (2003) w3x K. Artyushkova, J.E. Fulghum, J. Electron Spectrosc. Relat. Phenomena 121 (1 3) (2001) w4x G.C. Smith, M.P. Seah, J. Electron Spectrosc. Relat. Phenomena 42 (4) (1987) w5x B. Cord, O. Keitel, H. Rohrmann, J. Scherer, J. Magn. Magn. Mater. 193 (1999) w6x B. Tomcik, T. Osipowicz, J.Y. Lee, Thin Solid Films 360 (1 2) (2000) w7x M. Neuhaeuser, H. Hilgers, P. Joeris, R. White, J. Windeln, Diamond Relat. Mater. 9 (2000) w8x W. Eckstein, V.I. Shulga, Nucl. Instr. Methods Phys. Res. Sect. B: Beam Interactions Mater. Atoms (2000) w9x N. Inoue, A. Sagara, Y. Nishi, J. Nucl. Mater (1995)