Investigations on advanced magnetic recording media 2

Chapter 5 Ion Implantation in Magnetic Media: Effect of Mass of Ion Species CHAPTER 5 5 Ion Implantation in Magnetic Media : Effect of Mass of Ion Species 5.1 Introduction The previou

Trang 1

Chapter 5 Ion Implantation in Magnetic Media: Effect of Mass of Ion Species

CHAPTER 5

5 Ion Implantation in Magnetic Media : Effect of Mass of Ion Species

5.1 Introduction

The previous chapter focused on the effect of energy and species of the implanted

ions on the magnetic and structural properties It was shown that lower energy

implantation led to smaller lateral straggle from the TRIM calculations [195] As the

main objective of ion implantation technique is for fabricating patterned media and

lateral straggle is one of the main issues that needs to be understood and controlled for

extremely high areal densities, the results on lateral straggle from the previous chapter

shed new light on the subject It was further speculated from the previous study that

another parameter which might play a significant role in controlling the lateral range

and straggle is the mass of the implanted species, based on the ion-matter interaction

principles Thus, such a study might be useful for understanding the lateral straggle

and identifying a suitable species In this chapter, therefore, the effects of

implantations of ion species of variable masses have been investigated on the

magnetic and structural properties of the media

5.2 Patterned Media Requirements

As mentioned earlier, the focus of this thesis is on understanding and fabrication of

ion-implantation-based patterned medium towards 10 Tbits/in2 At 10 Tbits/in2, each

magnetic island would occupy an area of about 64 nm2, including the 4 nm spacing as

Trang 2

well If granular medium is used, this area could accommodate 4 grains of 4 nm

diameter In the case of patterned medium, such a bit-cell could accommodate a 4 nm

island with a spacing of 4 nm on all sides, as depicted in Figure 5-1

Figure 5-1 Schematic of granular media and BPM for 10 Tbits/in2

The first step to fabricate such a recording medium is to form resist patterns on the

media The densest patterns fabricated by electron beam lithography that has been

reported so far is at about 3 Tbits/in2 [196] Although an areal density of about 10

Tbits/in2 has been claimed to be achieved using block copolymers, the results can

only be achieved on single crystal substrates with a specific cut Furthermore, the

variable height of the wedges would lead to non-uniform pattern transfer [183]

Therefore, at this point of time, it is very difficult to fabricate patterns at 10 Tbits/in2

Studying the lateral straggle of ions at this areal density is equally important as seen

in Figure 5-2 If the lateral straggle is not controlled the whole area may get implanted

instead of only the unmasked region Figure 5-2 (c) A study on lateral straggle that

has been reported in the literature is only on 78.6 nm patterns at a pitch of 213 nm

[195] It is very difficult to study lateral straggle at a very small scale such as 10

Tbits/in2

Trang 3

Figure 5-2 Schematic of magnetic islands of (a) ideal BPM and BPM fabricated by ion implantation with (b) minimal lateral straggle and (c) very high lateral straggle.

We have used an innovative approach as an alternate route to study lateral straggle

in patterned dots, by emulating similar effects in granular films Figure 5-3 shows the

in-plane TEM image of a conventional perpendicular recording medium in a granular

structure The magnetic grains are surrounded by non-magnetic grain boundaries It

can be observed that the grain boundary is much narrower (1-2 nm) than the

bit-boundary in 10 Tbits/in2 patterned media Ion implantation in magnetic medium will

cause ions to move horizontally from magnetic grains to grain boundaries or vice

versa Therefore, a study on such a thin film with 1-2 nm grain boundaries could shed

light on the lateral straggle effects at densities such as 10 Tbits/in2 and above

A drawback that is associated with this technique is that there is no selective

implantation, but it is thought that some preliminary understanding from such a study

could shed light on lateral straggle and other issues The effect of lateral straggle was

studied from the measurement of exchange interaction in granular films and from the

TRIM simulations Exchange interactions could be studied from hysteresis loops to a

certain extent However, the first-order reversal curves (FORC) method can provide

more details such as the type of interaction field and the magnetic phase, in addition

to regular information obtainable from the hysteresis loops [149]–[151], [197] The

FORC contour is basically plotted with coercivity field in the x-direction and

Trang 4

interaction field along the y-ordinate The peak position along the x-axis in this

contour shows the switching field or coercivity value and width refers to SFD A

positive interaction field refers to magnetostatic interactions and a movement towards

negative values shows an increase in exchange coupling From the MH loops, even

though the slope does give an idea of the interactions present, it cannot separate the

possibility of different interactions present FORC has been explained in detail in

Chapter 3 Therefore, we have studied FORC on ion-implanted thin films Lateral

range and straggle, which are essential for achieving high-density patterns, have also

been studied for various ions species/masses using TRIM and have been correlated

with the results obtained from FORC for the first time

Figure 5-3 Plan-view TEM of conventional PMR showing grain boundary ~1-2 nm

5.3 Experimental Details

Figure 5-4 shows the structure of granular Co69.6Cr7Pt17.4-(SiO2)6 media fabricated on

the glass slides without any soft underlayer No soft underlayer was deposited in this

set of samples, with the purpose of obtaining magnetic signals solely from the

recording layer As a result, more information such as saturation magnetization and

Trang 5

FORC contour plots can be obtained using force magnetometry The media samples

contained two ruthenium intermediate layers (Ru1 and Ru2) for inducing texture and

grain segregation in the recording layer [198], [199] A seed layer, Tantalum (Ta),

was deposited to provide a smooth surface and desired hcp (0002) texture to the Ru1

and Ru2 intermediate layers

The samples were homogeneously implanted with helium (4He+), carbon (12C+),

nitrogen (14N+), argon (40Ar+), cobalt (59Co+) and antimony (121Sb+) ions over the

entire media surface by means of ion beam scanning The sources for the ions were:

He (gas), CO2 (gas), N2 (gas), Ar (gas), CoClx (solid; x=2 or 3) and SbO3 (solid)

respectively The interval between the doses was reduced in order to increase the

number of data points by which the effect of ion dose could be clearly seen The

implantation doses chosen were 1014, 5×1014, 1015, 5× 1015, 1016 ions/cm2 and beyond

up to 5×1016 ions/cm2 From the literature, it is known that the projected depth of

implantation depends on the ion’s mass and energies Ion implantation was done in

the recording layer such that the maximum peak of implantation profile was in the

middle of the recording layer and the ruthenium interlayer was affected the least as

shown in Figure 5-5 This was done keeping in mind the application for patterned

media The TRIM ion profile is a well-established method and has been proven to

correlate well with the secondary ion mass spectrometry (SIMS) ion profile but is

shallower by 2-2.5 nm [200] Also as previously mentioned, TRIM does not consider

the morphology of the layers while calculating the various collisions

Trang 6

Figure 5-4 Schematic of media layer structure

The requirement of energy to implant the ions is calculated first, using the most

comprehensive TRIM (Transport of Ions in Matter) program included in the SRIM

(The Stopping and Range of Ions in Matter) software [111], [190] In order to implant

4He+, 12C+, 14N+, 40Ar+, 59Co+ and 121Sb+ into the recording layer, the respective

energies were fixed to 2.5 keV, 5.8 keV, 6.5 keV, 13 keV, 17 keV and 37 keV

Trang 7

Irradiated and as-deposited samples were tested for magnetic properties by a polar

magneto-optic Kerr (MOKE) magnetometer and alternating gradient force

magnetometer (AGFM) MOKE was primarily used for preliminary understanding of

the hysteresis loops and the AGFM was used for obtaining magnetization and FORC

contours The crystallographic structure was evaluated by a Philips X-Pert X-ray

diffractometer (XRD) using Cu Kα radiation (l.540Å) FORC studies were also

conducted as it has been shown in the literature that a FORC diagram provides more

detailed information than the standard hysteresis loops about the exchange

interactions It will be seen that the heavier the ion species, the higher would be the

change in interaction field The change in interaction field and lateral straggle showed

a correlation

5.4 Magnetic Properties

5.4.1 Hysteresis Loops

In order to investigate the effect of ion implantation on the magnetic properties,

hysteresis loop measurements were carried out (Figure 5-6) MH loops have been

normalized to see clear differences in the change of the slope

Trang 8

Magnetic field (Oe)

Figure 5-6 Kerr loops on the sample implanted with various doses and implanted species – 4He+, 12C+, 14N+, 40Ar+, 59Co+ and 121Sb+

Trang 9

Figure 5-6 shows the hysteresis loops of the samples as measured by MOKE The

reference samples show rectangular hysteresis loops with H c of about 4000 Oe and H n

of about 2000 Oe The implantation leads to a reduction in the H c and H n values as a

function of implantation dose When different species are compared, 4He+ shows the

least change and 121Sb+ the most dramatic change The implantation of species with

higher mass appears to cause a more dramatic decrease of H c and H n The loop slopes also show a change as a function of the ion dose Figure 5-7 shows the SFD calculated

by obtaining full width half maximum value of dM/dH obtained from the hysteresis

loops as a function of dose for different species as explained in Chapter 3 Switching

of magnetic grains at one magnetic field provides a delta peak Hence, the width of

dM/dH shows the variation in the switching field known as SFD It can be noticed

that the SFD kept on narrowing as the fluence increased A narrow SFD is a possible

indication of increase in exchange coupling The loop looks very similar to the

samples in the previous chapter, which were sputtered without Ru2 and where the

grains were exchange coupled It is quite likely that the implantation of 59Co+ that

occurs at the grain boundary causes a stronger exchange coupling that leads to such a

change However, it should be noted that saturation magnetization is also reduced for

carbon, nitrogen, argon and antimony at high fluence values, as seen in Figure 5-9

Hence, both an increase in exchange coupling and reduction in saturation

magnetization (which may cause a decrease in magnetostatic interactions) plays a role

in changing SFD This will be studied again using FORC which can separate the two

types of interactions very easily

Trang 10

Figure 5-7 Switching field distribution (SFD) plotted as a function of various ion fluences

Figure 5-8 shows the coercivity (H c) as a function of dose or fluence (in

logarithmic scale) for all the ion implanted samples Key information extracted from

this figure includes: (i) with increasing dose, coercivity of the sample decreases; (ii)

the decrease in coercivity is dependent on the mass of ion species implanted The

changes in coercivity are similar to what has been reported in the previous chapter

Trang 11

5x10 16

Figure 5-8 Coercivity plotted as a function of fluence for all the implanted species – 4He+, 12C+, 14N+, 40Ar+, 59Co+ and 121Sb+

5.4.2 Saturation Magnetization

In patterned media, the irradiated regions are required to produce non-magnetic

material, as no signal has to be generated when the read sensor flies over this region

Since the signal is proportional to remanent magnetization (M r) (the magnetization at

zero field), the remanent magnetization has to be reduced to zero [16], [19] For

perpendicular recording media such as CoCrPt-SiO2, saturation magnetization (M s)

and remanent magnetization (M r) are not much different, provided the anisotropy axis

did not change with implantation Hence, M s can be measured to find out if

non-magnetic material is produced in the implantation region

In our samples, M s was measured using AGFM and is plotted in Figure 5-9 as a

function of fluence for all the implanted species – 4He+, 12C+, 14N+, 40Ar+, 59Co+

and 121Sb+ It was seen that, depending on the element species implanted, there were

changes in M s (within ±15% experimental error in Ms values) M s of as-deposited

Trang 12

media was ~465 emu/cc Except for 4He+ and a 59Co+ ion, the decline in M s was

observed after 1015 ions/cm2 of fluencies depending on the ion species Antimony and

argon showed drastic changes in M s compared to other ions In the implantation

involving these elements, M s was reduced to ~ 0 emu/cc at a fluence of 1016 ions/cm2

5x10 16 0

Saturation magnetization, M s, is an intrinsic property of the material and is defined

as magnetic moment per unit volume The magnetic moment depends on the

electronic band structure; hence one of the reasons for changes in M s could be due to

the modification of the electronic band structure of the magnetic material The M s of

an alloy also depends on the inter-atomic spacing, alloy composition, nearest neighbor

atom and to certain extent, on morphology Hence, any change in either of these may

change the saturation magnetization Ion-implantation is expected to cause all these

changes Therefore, saturation magnetization will be reduced In the case of previous

studies on doping Co with elements such as Cr, the rate of decrease in magnetization

with the number of Cr is much higher than that of several other materials, such as Pd

Trang 13

or Pt A similar effect is seen in our case too, where we observe a different rate of

change in magnetization In the case of 59Co+ implantation, however, since Co is

replacing Co atoms from the lattice, the change in magnetization is not clearly evident

as in other species

Another possibility that is peculiar to implantation induced changes in

magnetization (as compared to other doping studies) is the sputter etching of the

magnetic layer during ion implantation, which eventually reduces the thickness of

recording layer This is quite possible when heavy elements such as 40Ar+ and

121Sb+ are involved In fact, Ga+ ions are well known materials for focused ion beam

etching, wherein they are used to etch away materials As the calculation of Ms from

the measured moment (m) did not consider the change in the thickness (t), the M s

would appear to decrease (Ms = M/(t.A)) if the film is etched away (A is the area of

the film) This is because ‘m’ is reduced with etching but it is assumed to be the same,

resulting in a calculated value of lower M s

Figure 5-10 (a, b) shows the depth profile of the as-deposited CoCrPt-SiO2 media

obtained from secondary ion mass spectroscopy (SIMS) and change in M s as a

function of etch time of the as-deposited CoCrPt-SiO2 medium using the ion etch

system, where argon inside the vacuum chamber is ionized by energetic electron and

ionized argon atoms etch the surface It can be seen that the M s reduces with an

increase in etch time or a decrease in the thickness of the recording layer

Hence, it can be ascertained that the big change in M s after implantation may not

be highly dependent on the thickness of the recording layer sputter etched during

implantation, but sputtering etch could be a partial contributor The other reasons

might be significantly dominant in reducing the saturation magnetization

Trang 14

In such a case, however, the resultant patterned media would have topographical

variation and it has to be tackled using innovative experimental designs

Etch time (sec)

5.4.3 First Order Reversal Curves (FORC) Study

From the M-H loop, it was seen that with an increasing dose, there was an increase in

coercive squareness, which meant an increase in exchange coupling between the

Trang 15

grains with increasing dose (Figure 5-6) The FORC diagrams were measured to

obtain more insight about the interaction in these samples as explained before in

Chapter 3

FORC diagrams were plotted for the samples implanted with various ions and

doses ranging from 1014 to 5×1016 ions/cm2 Figure 5-11 shows the FORC contours

for as-deposited (non-irradiated) 4He+ and 121Sb+ implanted samples The FORC of

the reference sample is also shown for comparison It should be noted that the contour

plots do not take absolute value of magnetization into consideration as the partial

reversal loops have been normalized In the reference sample, the FORC distribution

peaked at around 4 kOe In FORC contours, the peak position (in H c axis) occurs at

the coercivity values This result is in agreement with the measured H c of about the

same value In addition, the vertical shift along H u (interaction field) in these FORC

distributions also provides information about the interactions In the case of samples

with single domain particles with no interaction, the shift is towards the positive

values of H u As the reference medium has well-isolated grains, the dominant

interaction is of magnetostatic origin The shift along the positive values of Hu is a

manifestation of magnetostatic interactions [152], [201], [202] It was observed from

the FORC contours that, depending on the mass and dose of ion species implanted in

the recording layer, H u and H c were both reduced The peak coordinate and spread in

the contour along the H c axis are measures of the mean switching field and switching

field distribution (SFD) It was noticed that with increasing fluence and mass, the

mean switching field (coercivity) dropped, showing similarity to the trend of H c as

shown in Figure 5-8

It can also be seen that the SFD narrowed down with increasing fluence for

samples implanted with various ions In fact, for heavier ions like 121Sb+, the SFD of

Trang 16

the implanted sample is much narrower than samples implanted with lighter ions like

the 4He+ ions Narrow SFD and shift towards H u=0 in perpendicular magnetic

recording media indicates the emergence of exchange interactions A narrow SFD

may also be due to the reduction in saturation magnetization with increasing fluence

because of reduction in magnetostatic interactions, as seen in Figure 5-11 But it

should be noted that Figure 5-12 shows H u as a function of fluence for various ion

species From Figure 5-12, H u was seen to reduce with an increasing dose, which

meant an increase in exchange interaction with increasing dose In fact, the exchange

interaction becomes more pronounced for samples implanted with heavier ions such

as 40Ar+, 59Co+ and 121Sb+ The higher doses for certain ions are missing from

Figure 5-12 because of high noise which made it difficult to quantify the exact values

of the peak position along the H u axis

Increase in exchange interaction has been previously reported in this chapter (from

the hysteresis loops), and it is presumed to be due to the lateral straggle of the Co

atoms from the magnetic grains going into the grain boundary As a result of the Co

Trang 17

migration from magnetic grains to the grain boundary, an increase in the exchange

coupling between the grains could occur This result will be discussed in the

5x10 16 0

The magnetic properties indicated a possible change in the atomic structures Such

changes in the structure can also be detected from XRD measurements Figure 5-13

shows the θ-2θ scans for samples implanted with various ions and ion species It was

seen from θ-2θ measurements that lower implanted doses of 4He+, 12C+, 14N+,

40Ar+, 59Co+ and 121Sb+ (1014 ions/cm2) showed negligible changes in the peak

position of the Co (0002) peaks However with an increasing dose, the Co (0002)

peaks intensity kept on reducing and disappeared after certain fluence for the specific

ion The shift became more prominent, as the dose increased further

Trang 18

Trang 19

The disappearance of the Co (0002) peak started appearing at lower doses as the

mass of ion species increased Such changes have been reported earlier in this work

[203] The disappearance of the Co (0002) peak could be due to the Ru and Co peaks

merging The shifting of peaks towards lower 2θ values corresponded to an increase

in the inter-planar spacing The increased inter-planar spacing in the recording layer

and Ru interlayer may be attributed to lattice expansion due to either the interstitial

position or a mixing of Ru and Co at the interfaces or both, depending on the ion

species as has been seen in the previous chapter

5.6 Calculated Lateral Range and Straggle

As discussed in the earlier chapter, one of the most critical parameters in the

fabrication of high density patterns is ‘lateral range and straggle’ If the lateral range and straggle are high, the non-magnetic ions may enter into the magnetic regions and

deteriorate the magnetic properties Therefore, it is essential to study the lateral

straggle The TRIM program can calculate the lateral straggle for each ion based on

detailed calculation of every recoiling atom [190], [204] Figure 5-14 shows the

lateral range and straggle, obtained from TRIM, as a function of ion species It can be

noted that the lateral range and straggle for 4He+ was the largest Implantation of

heavy atoms such as Ar+, Co+ and Sb+ ions show the least straggle, which means that

the heavy ions produce least straggle in the lateral direction These results suggest that

heavy mass ions could be more suitable to prevent the contamination of magnetic

regions (protected by resist) from the implanted species at high density

nanopatterning Cobalt, however, cannot be used as no change in saturation

magnetization has been seen in the recording layer

Trang 20

Ion Species

Figure 5-14 The lateral range and straggle as a function of ion species

5.7 Relation between Straggle and Exchange Interaction

As mentioned earlier, the primary objective of carrying out FORC was to understand

the exchange interactions and to find out any correlation between the straggle and

exchange interaction For such a correlation, we decided to consider the changes in

the interaction field, ΔH u ΔH u is essentially the change in Hu of the sample implanted

with 1015 ions/cm2 with respect to the as-deposited (reference) sample A reduction in

ΔH u is an indication of increased exchange coupling and vice versa For such a calculation, a higher dose was not chosen as the shift in Hu was not clear Moreover, at

the highest dose, some of the samples lost their magnetization Figure 5-15 shows the

correlation between the lateral straggle and the change in the interaction field ΔH u, as

measured by FORC The species in x-axis are arranged in the order of increasing

atomic mass, although not to scale

Trang 21

Figure 5-15 Correlation between straggle and ΔH u as measured by FORC curves

It should be noted that in absence of any lateral straggle, there would be no change

in exchange coupling and grain boundaries would maintain the same widths Since

exchange coupling is increasing, it is likely that the lattice Co atoms are displaced to

grain boundaries Increase in exchange coupling has also been shown by the MFM

and plan-view TEM showed an increase in grain size, both referring to the increase in

exchange coupling after implantation in addition to data interpreted from FORC The

results clearly indicate a reduction in the lateral range and an increase in ΔH u as the

basis of atomic mass and the collisions encountered by the species as shown in Figure

5-16

In the case of the lighter ions, the implanted ions undergo multiple collisions with

the lattice atoms and straggle over long distances in the recording layer As a result,

the lateral range is larger for ion species with smaller mass In the case of heavier

ions, the implanted atoms may knock out the atoms in the lattice (for example, Co)

Trang 22

As a consequence, the implanted ions do not scatter much and the lateral range is

shorter

According to the description above, there should be a correlation between the

lateral range and the change in the interaction field It was mentioned that the heavier

species of ions would knock out the Co lattice atoms In this case, the Co atoms would

traverse and a few of them would arrive at the grain boundaries As a result, the

exchange coupling would be increased The M-H loops in the earlier sections showed

an increase in exchange coupling Similar results are also shown by a large value of

ΔH u for the heavier species, as shown in Figure 5-12 This result provides a significant design perspective for patterned media When the species is implanted

through the holes of the resist material, heavier species such as 121Sb+ will remain

immobile, making that region non-magnetic From the hypothesis it seems that the

heavy ion species do not move much into the non-magnetic region as lateral straggle

is small They provide maximum damage in the magnetic properties i.e in magnetic

grains Therefore for BPM fabrication using Co-based alloys, heavy ions should be a

good choice

Trang 23

(a) Light Ions

Ions

grain Grain boundary

5.8 Density Functional Theory Calculations (DFT)

The main purpose of doing DFT calculations in this thesis is to explain theoretically

the variation in saturation magnetization of CoCrPt-SiO2 media upon implantation

with high doses of various ion species Saturation magnetization, M s, is an intrinsic

property of the material and is defined as the magnetic moment per unit volume The

Trang 24

magnetic moment depends on the number of uncompensated electrons in an atom;

hence one of the reasons for the change in M s could be that all the ions and recoil

atoms somehow interact with the electronic structure of the magnetic cobalt lattice

atoms In this scenario, in order to simplify the problem, reduce the calculation costs

as well as making it possible to run the simulation on the computer, we studied only

the effects of implanted ions on the cobalt surface as described by the cluster model

The finite cluster model was chosen as it imitated the granularity of the

CoCrPt-SiO2-based magnetic film [172] In actual, a grain size of 8 nm with a height of 14 nm

would consist of ~4334 Co atoms, considering hcp lattice with a= 2.507 Å and c=

4.069 Å (1 Å =10-8 cm) The cluster model was used in the DMOL3 quantum

mechanical package as it is supremely powerful in calculating the electronic

structures of big atomic clusters For the calculations, GGA-BLYP function was used

for a cutoff size of 4.5Å, SCF density 10-6 and 1000 iterations A cobalt cluster with

18 atoms, Co-18, was used for calculating the density of states (DOS) as the number

of atoms could not be increased due to the time constraint Hence for doping

calculations, cobalt atoms in the cluster were substituted with the ion species

implanted to change the ions/lattice atom Such a density of atoms can be correlated

with the density of ions implanted and it occurs that the fluence of 1016 ions/cm2 had

the ion ratio of about an ion per ~13 lattice atoms only, and for 1015 ions/cm2 it is

about an ion per ~100 atoms Figure 5-17 shows the change in moment with inclusion

of one, two and three ions per 17, 16 and 15 cobalt atoms It was observed that with

an increasing number of cobalt atoms substituted with the implanted species, the

moment was reduced (Figure 5-17) The reduction in moment with substitution of

carbon and argon could be due to the change in the electronic band structure as both

the atoms have a different number of electrons in their valence shells

Trang 25

Magnetic properties and lateral straggles were investigated as a function of the mass

of the ion species It was seen that saturation magnetization (M s) was reduced to ~0

emu/cc at a fluence of 2×1016 ions/cm2 for antimony (121Sb+) and 5×1016 ions/cm2

for argon (40Ar+), which could particularly be beneficial for patterned media

application The mass of the implanted ion species played an extremely crucial role in

reducing the coercivity and lateral straggle With increasing fluence, the exchange

interaction between the grains was found to increase due to the damages done to the

target lattice during implantation, which resulted in the downward shift of H u in the

FORC contours The results indicated that the implantation of heavier ions is a better

approach to minimize the straggle for high densities

Trang 26

Chapter 6 Patterned Media Fabrication

CHAPTER 6

6 Patterned Media Fabrication

6.1 Introduction

The previous chapter highlighted the fundamental understanding needed for the

fabrication of patterned media It was observed that ions with heavier mass have a

smaller lateral straggle It was also highlighted that the Sb+ and Ar+ ions help to

achieve zero Ms at doses of the order of 1016 ions/cm2 In this chapter, experiments

were carried out to implant these ions into the recording media through hard masks

such that only selective regions are implanted and thus no signal is produced from

these regions and the masked area produces the signals

The current granular CoCrPt-oxide thin film media used in perpendicular magnetic

recording contain Cr to segregate the grains [205]–[208] The Cr is also found to be

present inside the grains, hence reducing the Ku value and consequently limiting the areal density [67], [209], [210] Therefore, a high K u material is essential for high

areal density magnetic recording media In the current chapter, L10 FePt has been

used for patterned media fabrication and the understanding made therein

6.2 Experimental Methods

For the long range order of the L10 phase of FePt, where ordering is obtained by

stacking alternating atomic layers of Fe and Pt each, the ordering parameter, S o, is

given by:

Fe

Pt Pt

Trang 27

where, x Fe(Pt) is the atomic fraction of Fe(Pt) in the sample, y Fe(Pt) the fraction of Fe(Pt)

sites and γ Fe(Pt) the fraction of Fe(Pt) sites occupied by the correct atom [52] The

ordering parameter S o is unity for the equiatomic concentration of FePt and hence is

said to be the condition of perfect long range order Experimentally, it has been found

that even though S o is unity for Fe50Pt50, the magnetocrystalline anisotropy is

maximum with Fe content increased slightly above the equiatomic compositions of Fe

and Pt; hence Fe60Pt40 was chosen [53] The alloy composition of the deposited FePt

film was confirmed by the X-ray photoelectron spectroscopy (XPS) depth profile, as

Figure 6-1 Alloy composition of the deposited FePt film

Figure 6-2 shows the schematic diagram of the FePt media stack The topmost

layer of 2 nm Si was deposited as a protective layer to the FePt film at room

temperature A 20 nm Cr90Ru10 underlayer was deposited at 200 W, and 400 ºC on

glass substrate to provide (001) texture to the FePt film CrRu was used instead of Cr

itself because Ru doping into Cr leads to an increase in the mismatch between CrRu

Trang 28

(002) and FePt (001) to 6.33%, compared to 5.8%, hence promoting better chemical

ordering in the FePt films [211]–[213] Cr90Ru10 was followed by an MgO buffer

layer of about 3 nm, deposited at 150 W and 400 ºC, to reduce the Cr diffusion into

the L10 FePt as diffusion of Cr will deteriorate its magnetic properties The Fe60Pt40

layer of 10 nm was deposited on the MgO underlayer at 100 W and 600 ºC to obtain

the L10 phase The thicknesses of all layers were calibrated by atomic force

microscopy The base pressure was lower than 5x10-8 Torr

Figure 6-2 Schematic diagram of FePt media stack

6.2.1 FePt Media Optimization

A 20 nm CrRu layer was deposited at 3 mTorr and a temperature of 400 ºC and power

varying between 150 to 250 W No change in the crystallography growth was

observed on power variation as seen from the XRD θ-2θ plot in Figure 6-3 Hence, a structure was fabricated with 20 nm CrRu deposited at 200 W and 400 ºC, followed

by a 2 nm MgO deposited at 150 W and 200 ºC, to obtain the (001) texture The

sputtering conditions were chosen with some variation based on previous work by our

group due to the use of different sputtering chambers for these studies [213] FePt was

further deposited on the MgO underlayer at a deposition power of 100 W and a

Trang 29

temperature of 600 ºC Both the films were deposited in Ar ambience at 3 mTorr

Figure 6-4 (a) shows the hysteresis loop of the FePt film in the in-plane and

out-of-plane directions It can be seen that the out-of-out-of-plane loop shows the largest coercivity

and remanence, indicating perpendicular magnetic anisotropy The coercivity of the

film in the out-of-plane direction was about ~14 kOe However, it must be noted that

the loop is not closed and the coercivity could be much larger, if a higher field that

could saturate the sample was applied The in-plane loop also shows a large

hysteresis, indicating that the out-of-plane texture is not very good

We carried out XRD scans in order to find out the nature of peaks Figure 6-4 (b)

shows the XRD θ-2θ plot for the sample The FePt (001) and FePt (002) peaks are

observed along with the MgO (200) and CrRu (002) peaks, indicating the growth of

the FePt(001)[100]||MgO(001)[100]||CrRu(001)[110] heteroepitaxial relationship

Trang 30

during the deposition process The relatively weak MgO (200) peak was attributed to

the small thickness of the MgO layer [214] The peaks corresponding to any other

phases of FePt are not seen

Figure 6-4 (a) Out-of-plane and in-plane coercivities (b) XRD θ-2θ plot of the FePt media stack

From a different study on FePt films, the deposition pressure of FePt was varied

from 3 mTorr, 5 mTorr and 10 mTorr at 600 ºC and 100 W It was observed that with

increasing deposition pressure of FePt, the intensity of the FePt (001) and FePt (002)

peaks decreased, which was an indication of poor crystallographic growth In this

optimization step, it was necessary to change the heating steps from 400 to 200 for

Trang 31

CrRu and MgO deposition followed by 600 ºC for FePt deposition This required a

longer waiting time and resulted in fewer samples per day In order to speed up the

process, we tried fabricating MgO layer at a deposition temperature of 400 ºC

Based on hysteresis loop measurements and the XRD θ-2θ plot, almost no change

in MgO crystallography growth was visible Figure 6-5 showed no change in out of

the plane and in plane hysteresis loops of the samples with MgO deposition at 200 ºC

and 400 ºC It should be noted that the y-ordinate is the magnetic moment which does

not take the volume of the sample into consideration A small change in sample size

while cutting the sample may cause this change in the moment of the two different

Trang 32

conditions Moreover, saturation magnetization cannot be compared due to

restrictions on the AGFM system which had a maximum field of only 20 kOe

This was followed by reducing the pressure of CrRu and MgO both to 1.5 mTorr

and finally the pressure of the FePt layer was also reduced to 1.5 mTorr in the third

stack Figure 6-6 shows the hysteresis loops in the out-of-plane and in-plane

(a) Out-of-plane

(b) In-plane

CrRu/ MgO/FePt CrRu/ MgO/FePt (3mTorr) CrRu/ MgO (3mTorr) /FePt (3mTorr)

Figure 6-6 (a) Out-of-plane and (b) In-plane hysteresis loops of the FePt media with CrRu deposited at 1.5 mTorr and MgO and FePt layers deposited at either 3 mTorr or 1.5 mTorr of

Ar pressure

It can be seen that a slight increase in the out-of-plane coercivity from 13.93 to

14.77 kOe was observed for the sample with MgO and CrRu deposited at 1.5 mTorr

Out-of-plane coercivity values were found to be 14.54 kOe when all the layers were

Trang 33

deposited at 1.5 mTorr Also, in-plane coercivity was reduced to 7.82 kOe from 9.51

kOe when both MgO and CrRu were deposited at 1.5 mTorr Furthermore, deposition

of FePt at 1.5 mTorr led to an increase in in-plane coercivity back to 9.2 kOe Hence,

the samples structure with 10 nm Fe60Pt40 films were deposited on the MgO

(3nm)/Cr90Ru10 (20nm) pre-coated corning glass substrate FePt, MgO and CrRu

layers were deposited at 600, 400 and 400 ºC, respectively, in the Ar environment

The Ar pressure during the deposition of FePt, MgO and CrRu were 3, 1.5 and 1.5

mTorr, respectively

From the magnetic hysteresis loop shown in Figure 6-6(b), a strong in-plane loop

can be observed However, the XRD did not show the presence of any other phases

than the preferred L10 (001) Such an observation can be explained based on the

critical thickness as reported by Perumal et al [55] In their study on FePt, they have

reported that the FePt grows up to a thickness of 6 nm in L10 phase However, beyond

this critical thickness, a second layer of FePt is formed which does not exactly grow

in the L10 phase and is softer compared to FePt from the first layer Since the

orientation of the grain in the second layer may be at an angle different from that of

the bottom grain, it leads to an in-plane loop

6.2.2 Ion Species and Dose Optimization

For the purpose of patterned media, it is necessary to optimize the dose of the

implantation in such a way that the saturation magnetization is reduced to zero For

this purpose, an optimization study of doses was carried out on unpatterned thin films

The FePt media samples were homogeneously implanted with Sb+ over the entire

media surface by means of ion beam scanning The source for the ions was SbO3

Trang 34

(solid) The implantation doses chosen were 1015, 5×1015, 1016 and 5×1016 ions/cm2

From the literature, it is known that the projected depth of implantation depends on

the ion’s mass and energies Ion implantation was done in the recording layer such

that the maximum peak of implantation profile was in the middle of the recording

layer and the CrRu interlayer is affected the least This was done keeping in mind the

application for patterned media To obtain the required profile, TRIM calculation was

done to obtain the energies of implantation, which is 9 keV for 121Sb+ ions

Figure 6-7 (a) and (b) show the change in out-of-plane coercivity of the samples

implanted with Sb+ with various ion fluences It can be noticed that the magnetization

loops showed a change in easy axis direction from out of the plane to in plane

direction even for the lowest dose of 1015 ions/cm2 as evidenced by a square loop

along in-plane and slope in out-of plane direction Such a change of easy axis

direction has previously been reported in FeCoBSi film after Co+ implantation [214]

The change in anisotropy was explained by the irradiation induced rearrangement of

the local atomic structure

When the implantation dose was further increased, no significant change was

observed At a fluence of 5×1016 ions/cm2, the in-plane loop vanished and the sample

appears to have switched the anisotropy easy axis back to the direction of

out-of-plane From these results, it appears that the ion implantation on L10 FePt media with

Sb+ ions has caused a phase transformation from chemically ordered ferromagnetic

L10 phase to disordered A1 phase explaining the increase in softness Atomic

rearrangements as a result of implantation have been reported to change the

anisotropy easy axes The changes in atomic disorder will be explained later in this

chapter based on the XRD results

Trang 35

Figure 6-7 (a) Out-of-plane s (b) In-plane hysteresis loops for FePt samples implanted with 121Sb+ ions at various dose

The saturation magnetization (Ms) of the as-deposited sample was found to be

1100 emu/cc It can be seen from Figure 6-8 that saturation magnetization (Ms)

reduced with increasing fluence However, Ms reduced from ~820 emu/cc to ~20

emu/cc, as the implantation dose increased from 1015 ion/cm2 to 5×1016 ion/cm2 The

hysteresis loop for the as-deposited FePt film has not been shown in Figure 6-7 as the

change in the ion implanted samples would not be seen clearly then This reduction in

Ms to ~20 emu/cc may have been due to atomic dilution or change in electronic

structure of the FePt media as a result of implantation

Trang 36

As mentioned earlier, the reduction in Ms could also be partially due to the sputter

etch of the FePt layer by Sb+ ions Our previous studies on Sb+ implantation have

indicated that a dose of 2×1016 ions/cm2 is sufficient to reduce magnetization to zero

in Co-based alloys (Chapter 5) Because the saturation magnetization in FePt is larger

(1100 emu/cc), a slightly higher implantation dose has been found to be necessary to

reduce the magnetization to zero Therefore, for the fabrication of patterned media, a

dose of 5×1016 ion/cm2 was used to ensure that the magnetization could be reduced to

zero

Figure 6-8 Dependence of M s as a function of various ion fluences

Furthermore, 4He+ ions were also implanted into the FePt media Similar results

were observed for 4He+ implantation into the FePt media However, the reduction in

Hc, in both out of plane and in-plane directions, was at a slower rate comparatively as

seen in Figure 6-9 While the irradiation of Sb+ at 1015 ions/cm2 caused a very narrow

loop, a similar dose of He+ caused a reduction in Hc to about 10000 Oe The

difference in the coercivity behavior with ions can be explained based on the effect of

mass causing structural changes in the host lattice

Trang 37

6.2.3 Hard Mask Fabrication

The FePt media structure was deposited as shown in Figure 6-2 20, 30 and 45 nm

tantalum (Ta) was further deposited on these samples as shown in Figure 6-10 Ta

was deposited to serve as a hard metal mask for the samples Metal mask was used on

our samples, as metal etching properties would not change much after implantation

Ta-coated samples were spun coated with the Obducat proprietary Simultaneous

Thermal and UV (STU) resist The resist was first diluted with methiopropamine

(MPA) in a ratio of 2:3 The resist was spun coated at 5000 rpm for 45 seconds and

Trang 38

baked at 100 ºC for three minutes, resulting in an organic layer of ~50 nm thickness

The spun coated resist samples were then used in nano imprint lithography (NIL) A

silicon master mold which had cylindrical pillars of 50 nm diameter at 80 nm pitch

was used to pattern the STU resist

Figure 6-10 Ta metal deposited on FePt media samples with thicknesses of 20, 30 and 45 nm

Combined thermal and UV NIL was used to increase the lifetime of the master

mold through a two-step process Using thermal imprinting under the optimized

imprinting conditions as described in

Table 3-1, the master mold was replicated on an intermediate polymer stamp (IPS)

The daughter mold, IPS, thus formed was used to imprint the UV-curable STU resist

at a much lower constant temperature (Table 3-2) Such a technique solves the

problem related to thermal expansion and provides highly uniform and thin residual

layers

Trang 39

Figure 6-11 Cross-section SEM of FePt media sample with 45 nm Ta when etched for 1 min

30 sec

A 10 nm residual layer was observed after the imprinting, which was later removed

by means of oxygen plasma for nine seconds at 120 mTorr and 50 W RF bias in 30

sccm of oxygen Based on previous optimization, it was seen to remove any residual

layer as shown by the cross-section SEM on the samples To transfer the patterns to

Ta in order to obtain the hard mask, the samples were further etched by means of

reactive ion etching-inductively coupled plasma (RIE-ICP) Some calibration samples

were first deposited with 45 nm of Ta strips on FePt and were exposed to CF4 gas at

flow rates ranging from 30 to 90 sccm 90 sccm of CF4 gas reduced the height of the

Ta strips for the RF power of 50 W and ICP power of 200 W The lowest pressure

that could ignite the plasma was 150 mTorr The nanoimprinted samples were etched

in the CF4 environment for 5 sec, 40 sec and 1 min 30 sec Figure 6-11 presents the

cross-section SEM, which shows a pillar height of about 20 nm for etching time of 1

min 30 sec on samples with 45 nm Ta When the etch time was increased beyond this,

the height of the pillar started to reduce

Trang 40

Figure 6-12 Ta metal on FePt media samples with thicknesses of 20, 30 and 45 nm etched in

CF 4 for 1 min 30 sec to obtain Ta pillars of ~20 nm

Based on the etching experiments, three types of samples with Ta thickness of 20

nm, 30 nm and 45 nm were used The FePt samples with Ta metal masks can be

shown by the schematic in Figure 6-12 The three types of samples (Figure 6.10) were

subsequently implanted with various ions species and their optimized ion dose as

obtained on the plane FePt media samples The metal mask was removed by the same

RIE process after implantation

6.3 Characterization of Patterned Media

6.3.1 Magnetic Properties

All the three types of patterned samples, as shown in Figure 6-12, were subsequently

implanted with antimony (Sb+) at 5×1016 ions/cm2 so as to ensure that the

magnetization reduced close to zero in the implanted regions Figure 6-13 (a) shows

the hysteresis loops of the patterned samples for different values of residual mask

layer thickness in out of plane direction It is surprising to note that none of the

Định dạng
Số trang	89
Dung lượng	6,36 MB