Conventional hard disk sample

In order to test the resolution limitations of our setup, we have investigated three different types of hard disk drives. The first type of HDD under investigation was an ordinary HDD drive from Western Digital (WDCaviar400) with relative large bit size varying between ∼ 0.7 − 2 µm. The results for room temperature are shown in Fig. 5.1. The working parameters of the various images taken of the HDD are listed in table 5.1.

The imaged bits are ∼ 2 − 3 µm in size.

For the interpretation of the magnetic contrast formation, we assume that we are dealing with a longitudinal magnetized recording medium. If we assume that the tip’s magnetization is perpendicular to the sample’s

Table 5.1: Table of the working parameters used for the HDD images. All of the images have been taken with the Multi75–G cantilever from Budgetsensors (table 4.1)

Fig. Mode T f Aexc⋆ Q Image vscan hlif t♥ Frame Loop

[K] [kHz] [mVp]^⋆⋆ factor [µm²] [µm/s] [nm] size filter^†

5.1 CE lock–in 298 76.0 15 122 14 × 14 15 80 512 × 512 250 Hz

5.2 CE lock–in 77 77.6 2.7 1.52 · 10³ 28 × 12 14 80 512 × 212 200 Hz 5.3 CE PLL^†† 4.2 77.7 1.8 5.57 · 10³ 31 × 31 150 − 512 × 512 200 Hz 5.4 CE PLL^†† 4.2 77.7 1.8 5.57 · 10³ 15 × 15 7.5 − 512 × 512 200 Hz

⋆Applied excitation amplitude values may deviate, due to improper calibration of the IO output. This output should be terminated at 50 Ω. Also see chapter 4.4.2.

⋆⋆The amplitude values used in this chapter are expressed in units of mVp, the subscript here indicates the

‘peak’ value, whereas the subscriptppindicates the ‘peak–to–peak’ value andrmssubscript denotes the

‘root–mean–square value’ of the amplitude.

†Applied filters are 1^storder filters, i.e. 6 dB/Oct.

††Modulation has been disabled for magnetic contrast imaging. For the images taken in lift mode operation, the modulation is disabled during the second scan.

♥Although in our setup lifting the cantilever actually means lowering of the sample with hlif t, it is much more illustrative to talk about the lift height of the cantilever instead of the lowering height of the sample.

1

2 mm

11 nm 39 nm

(a)

2

–167 deg

–171 deg

2 mm

(b)

Figure 5.1: Room temperature topographic image of a conventional HDD (a) and the corresponding magnetic contrast image (b). Image was taken in CE lock–in mode with a lift of 80 nm. The bottom of each picture shows the corresponding line view. The apparent curvature of the bits in the magnetic contrast picture is caused by the hysteresis of the x and y scanner piezos.

surface, we expect to see a large contribution the magnetic contrast formation when we are scanning over the magnetic stray field of the domain walls at the bit boundaries. For longitudinal recording media, the bits itself do not contribute as much to the magnetic contrast formation as the domain walls, since their magnetization

20 nm –23 deg

50 nm

2 2 2 1

5 m m

–105 deg

5 m m

Figure 5.2: Topographic image of the conventional HDD (top) and the magnetic contrast formation of the same region (bottom) at 77 K. Both images are taken in CE lock–in mode. The applied lift height was 80 nm. The right–hand side shows the corresponding line views.

is in–plane, whereas the domain walls have a magnetization that is perpendicular to the surface. In the case of a perpendicular magnetized tip, this direction of magnetization produces the largest deviations of the cantilever’s deflection. From the line view of Fig. 5.1b, we can see that the observed magnetic contrast formation is in agreement with our expectations. Starting from left to right, while looking at the line view of Fig. 5.1b, we first see the contribution of the domain wall (downward magnetization) to the contrast formation, the bit magnetized in x direction (in–plane magnetization), the domain wall (upward magnetization), the bit magnetized in −x direction (in–plane), the domain wall (upward) and so on. The magnetic contrast formation of the bit itself is almost unresolved due to the much larger contributions of the domain walls at the boundaries.

Of course, as we have seen in chapter 3, all of these contributions depend on the tip’s magnetization direction itself, which as previously stated is unknown. In the picture, the bits appear to be curved, this is an artifact stemming from the piezo hysteresis, as discussed in chapter 4.2.2. This hysteresis is even more enhanced due to the large scan speed at which the image has been taken (vscan = 15 µm/s). Fig. 5.2 shows an image of the HDD taken at 77 K. The bit sizes displayed in this picture are comparable to the ones imaged at RT. As one can see in the picture, the magnetization contrast values have increased significantly from ∆φ ∼ 5^◦ at RT to over more than 50^◦ at 77 K as a result of the depreciation in temperature. As expected we also observe a similar magnetic contrast formation as a result of the different magnetization directions within the longitudinal magnetized recording medium.

When going to even lower temperatures, the Q factor increases further and the setup can be operated in PLL mode (chapter 4.4.3). This can be seen in Fig. 5.3 and Fig. 5.4. In contrast to the previous pictures (Fig. 5.1 and Fig. 5.2), both pictures show a different area of the HDD. The average bit size is ∼ 700 nm, which is a factor of 3 smaller than the bit size of the room temperature images. The hysteresis of the piezo scanners decreases at low temperatures, as explained in chapter 4.2.2. In Fig. 5.3 we see that the magnetization direction of the bit itself is unresolved (profile 1). Profiles 2 and 3 of both Fig. 5.3 and 5.4 show a line view scan over a single domain wall with an upward (downward) magnetization. Although we expect a uniform magnetization, we can see that there are fluctuations in the magnetic contrast formation. In both of the pictures, the contribution to the contrast formation is ∼ 1/5^th of the overall signal. This does not necessarily mean that the magnetization over a single magnetic domain wall is non–uniform. These contributions can also be attributed to fluctuations in the detection signal, laser power fluctuations or electronic crosstalk of the high–voltage positioning signals for the piezos.

2.87 Hz

1

2

–2.98 Hz

5 mm

3

Figure 5.3: Magnetic contrast image of the conventional HDD taken at 4.2 K in CE PLL Mode operation with the feedback modulation switched off. The right–hand side shows the corresponding line views.

6.6 Hz

–6.6 Hz 1

3

3 mm

2

Figure 5.4: Close up of the magnetic contrast image of Fig. 5.3 taken at 4.2 K in CE PLL Mode operation with the feedback modulation switched off. The right–hand side shows the corresponding line views. This image has been taken at much slower scan speed vscan= 7.5 µm/s than the previous image (vscan= 150 µm/s) which results in a better spatial resolution of image.