D E S C R I P T I O N
Patterned Media Optimized for Data Rea ac
Field of the Invention
The present invention generally relates to magnetic recording media, and more particularly to patterned magnetic recording disks with discrete magnetic regions or islands . More specifically, the invention relates to a patterning process for such magnetic recording disks. Still more specifically, the invention pertains to optimizing the patterned medium for data readback.
Background of the Invention
Conventional magnetic recording media, such as the magnetic recording disks in hard disk drives, typically use a thin film granular ferromagnetic layer, e.g., a sputter-deposited cobalt-platinum (CoPt) alloy, as the recording medium. During the write process the bit cells are defined in the magnetic layer being comprised of many small magnetic grains (FIG. 1A) . The transitions between bit cells represent the information of the recorded data and can be read back by a read head. IBM's US patents 4,789,598 and 5,523,173 describe this type of conventional rigid magnetic recording disk.
The challenge of developing continuous granular films as magnetic media will grow with the trend toward higher areal storage densities. Reducing the size of the magnetic bit cells while maintaining a satisfactory signal-to-noise ratio (S/N- ratio) , for example, will require also to scale down the size of the grains. Unfortunately, significantly reducing the size of weakly magnetically coupled magnetic grains without increasing their anisotropy will make their magnetization
unstable at normal operating temperatures, i.e., their magnetization could change spontaneously below a certain grain volume. To postpone the - arrival of this fundamental „superparamagnetic'Λ limit and to avert other difficulties associated with extending continuous granular media such as extremely high write fields that become necessary with increasing media anisotropy, there has been interest in patterned magnetic media.
With patterned media, the continuous granular magnetic film that covers the disk substrate is replaced by an array of spatially separated discrete magnetic regions or islands (FIG. IB), each of which carrying the information of a single bit. In the conventional form of patterned media the magnetic islands are surrounded by a non-magnetic matrix. The primary approach for producing patterned media has been the use of lithographic processes to selectively deposit or remove magnetic material from a magnetic layer on the substrate so that magnetic regions are isolated from one another and surrounded by areas of nonmagnetic material . Examples of patterned media made with these types of lithographic processes are described in U.S. patents 5,587,223; 5,768,075 and 5,820,769.
From a manufacturing perspective, an undesirable aspect of the process for patterning media that requires the deposition or removal of material is that it requires potentially disruptive processing with the magnetic media in place. In particular the surface topography may be significantly modified. Processes required for the effective removal of resists and for the reliable lift-off of fine metal features over large areas can damage the material left behind and therefore lower production yields. Also, these processes must leave a surface that is clean enough and has a suitable surface topography so that the magnetic read/write head supported on the air-bearing slider
of the disk drive can fly over the disk surface at very low flying heights, typically around 10 nanometers (nm) .
A technique for stabilizing a two-dimensional magnetic array by pinning domain walls is disclosed in US-A-4, 274, 935. Here, the magnetic separation between the elements of the array is defined by ion implantation.
An ion-irradiation patterning technique that also avoids the selective deposition or removal of magnetic material, but uses a special type of perpendicular magnetic recording media, is described by Chappert et al . , in „Planar patterned magnetic media obtained by ion irradiation", Science, Vol. 280, June 19, 1998, pp. 1919-1922. In this technique, Co/Pt multilayer sandwiches which exhibit perpendicular magnetocrystalline anisotropy are irradiated with ions through a lithographic patterned mask (FIG. 2) . The ions mix the Co and Pt atoms at the layer interfaces and reorient the easy axis of magnetization to be in-plane so that the irradiated regions no longer have perpendicular magnetocrystalline anisotropy.
U.S. patent 6,331,364 describes an ion-irradiated patterned disk that uses a continuous magnetic film of a chemically- ordered Co (or Fe) and Pt (or Pd) alloy with a tetragonal crystalline structure. The ions cause disordering in the film and produce regions in the film that have no magnetocrystalline anisotropy, which means they are magnetically „soft" .
A potential disadvantage of the disks patterned by ion radiation according to US-A-4, 274, 935, Chappert et al . and the US-A-6, 331, 364 ion-irradiated patterned disks is that only the regions separating the discrete magnetic regions from one another have to be ion irradiated whereas the discrete magnetic bit cells have to remain non-irradiated. As a result of heavy ion irradiation the Curie Temperature c at which
ferromagnetism disappears • is decreased within the surrounding matrix below the ambient temperature Troom. Besides the fact that such high irradiation doses may be impractical for a volume production of patterned media, another problem arises for the stencil masks used. A mask of opposing polarity is required where the mask openings reflect the surrounding matrix in which free standing non-transparent features have to be imbedded reflecting the bit cells. This is impossible to realize in a single mask. This problem is well known in lithography and is called the „donut problem" , since this problem arises in its simpliest form with "donut-like" geometries, i.e., ring structures.
Therefore, " O 01/99100 discloses a magnetic recording disk that is patterned into discrete magnetic and nonmagnetic regions with small ion irradiated magnetic regions serving as the magnetic recording data bits. The magnetic recording layer comprises two ferromagnetic films separated by a nonferromagnetic spacer film. The spacer film material composition and thickness is selected such that the first and second ferromagnetic films are antiferromagnetically coupled across the spacer film. After this magnetic recording layer has been formed on the disk surface, ions are irradiated onto it through a patterned mask. The ions disrupt the spacer film and thereby destroy the antiferromagnetic coupling between the two ferromagnetic films. As a result, in the regions of the magnetic recording layer that are irradiated the first and second ferromagnetic films are essentially ferromagnetically coupled so that the magnetic moments from- the ferromagnetic films are parallel and produce a magnetic moment that is essentially the sum of the moments from the two films. In the non-irradiated regions of the magnetic recording layer, the first and second ferromagnetic films remain antiferromagnetically coupled so that their magnetic moments are oriented antiparallel . The composition and thicknesses of the first and second ferromagnetic films are selected such that essentially no magnetic field is detectable at a
predetermined distance above the non-irradiated magnetic recording layer corresponding to the height the magnetic recording head would be- located a .
The prementioned „donut problem" and the need for very high doses in the case of an irradiated non-magnetic matrix around the bit cells can both be avoided if only the magnetic bit islands are irradiated, and the surrounding matrix remains „as-grown", i.e., does not undergo any irradiation or similar softening process such as for example by ion-irradiation. But as a consequence, the matrix is still magnetic and reversal of the magnetization may occur while writing the bit cells. Thus, the edges of the bit cells are no more well defined, the noise in readback signals increases and the bit cells cannot be closely packed as is necessary for high storage densities.' The reversal in the surrounding matrix will even be enhanced in the presence of softened islands, because they provide already nucleation sites before nucleation starts within the matrix. In the presence of nucleation sites, however, reversed magnetic domains are already existent and the reversal can proceed already by motion of the domain walls. Therefore the reversal of the matrix can be eased, which can, e.g., be seen in G.J. Kusinsky et al . , Appl . Phys . Lett., 79, 2211 (2001). Accordingly, the reversal of the matrix or extension of magnetic domains into the matrix due to nucleation or domain wall propagation during the write process has to be avoided.
Since the ion exposure time determining the fabrication throughput is a very critical parameter for an industrial application, a material system and a patterning technique has to be developed which minimizes process time for the patterning and therefore increases throughput . In case the ion-irradiation or any other process that can magnetically soften the material in predefined regions is not sufficient, the bit cells suffer from poor magnetic contrast. Poor magnetic contrast means that the bit cells can not be fully saturated in either direction without any magnetic reversal of
the surrounding matrix, so that the magnetization of reversed bit cells and the matrix is not very different. Good magnetic contrast, however, is important for acceptable read back performance .
Summary of the Invention
It is therefore an object of the present invention to provide a method for patterning a magnetic storage medium that avoids the above mentioned disadvantages of the state of the art.
It is another object of the present invention to provide a method for the determination of an external write field of a patterned magnetic storage medium having maximum magnetic contrast .
It is still another object of the present invention to provide a method for the determination of the optimum irradiation dose for patterning a magnetic recording medium.
Still another object of the invention is to provide a patterned magnetic recording medium which allows bits of both orientations to be written without an onset of magnetic reversal within the matrix.
These and other objects and advantages are achieved by the method disclosed in claim 1 and the magnetic recording medium disclosed in claim 11.
Advantageous embodiments are laid down in the dependent claims .
Brief Description of the Drawings
The invention will in the following be described in more detail in conjunction with the drawings, in which
Fig. 1A schematically shows the micromagnetics at a transition between bit cells being magnetized in opposing direction as it appears in the magnetic layer of a conventional medium being comprised of many small magnetic grains;
Fig. IB schematically shows an array of spatially separated discrete magnetic regions or islands as they are present in patterned jnedia;
Fig. 2 schematically depicts the local irradiation of Pt- Co-Pt multilayer sandwiches, exhibiting perpendicular magnetocrystalline anisotropy, with ions through a lithographic patterned stencil mask as realized in state of the art ion projection structuring;
Fig. 3 is a graph showing the remanent magnetization curves (magnetic moment vs. the applied field) of Co/Pt multilayers measured by vibrating sample magnetometry for increasing doses of homogeneous irradiation; and
Fig. 4 schematically shows the track patterning for reducing track edge noise as can be realized by a local magnetic softening of the tracks as an intermediate step for patterned media introduction.
Detailed Description of the Preferred Embodiment
It is proposed that the magnetic parameters, in particular those that characterize the onset of magnetic reversal given by domain propagation and/or nucleation of the medium, are determined in the "as-grown" state, i.e., prior to any irradiation or magnetic softening process, e.g., by irradiation with ions. From measured magnetization curves the
nucleation field Hn, i.e., the minimum field at which reversed magnetic domains are formed in an otherwise fully saturated magnetic material can be determined as illustrated in Fig. 3 which will be described in more detail later on. The propagation field H i.e., the minimum field necessary for starting magnetic reversal in a magnetic material which provides nucleation sites, such as already reversed domains, can be measured by applying increasing external fields to the material in the demagnetized state with zero net magnetic moment, Hp being the onset field of magnetic reversal.Hp may be lower than Hn because a reversal solely by domain wall motion which may be energetically preferred in comparison to nucleation processes can take place.
Subsequently the minimum softening (irradiation) dose necessary for the maximum magnetic contrast has to be determined. Maximum magnetic contrast means that the bit cells can be magnetized close to the saturation level in the opposite direction of the surroundiung matrix, the latter remaining in the fully saturated state even in the vicinity of the bit cells. This minimum dose can be derived from the magnetization curves measured in remanence after stepwise increasing the doses of homogeneous irradiation.
For realizing a patterning process, media are magnetically softened (irradiated) only locally in predefined areas which results in a local decrease in coercive force H c as compared to the surrounding. By applying an external magnetic field, exactly these areas or bit cells can be magnetically switched and can magnetically carry a binary information. The magnetic medium now consists of bit-cells with reduced coercivity surrounded by an „as grown" matrix of significantly higher coercivity. In order to store information that can be reliably read back, a complete magnetic reversal within the bit cells shall be achieved by write fields that are too low to start reversal in the matrix. Therefore it is necessary to exactly determine the minimum field where a magnetic reversal in the
matrix occurs or starts . It must be taken into account that magnetic reversal in the matrix may be alleviated by the bit cells acting as nucleation sites.
Therefore the magnetization curve of the as grown film is measured after first saturating the film and subsequently countermagnetizing it until the net magnetic moment vanishes, i.e., applying first the saturation field -Hsat(D=0) (the field which is necessary to fully saturate the film) at zero dose (D=0) and subsequently the coercive field +Hc(D=0) in the opposite direction which by definition leads to zero net magnetization. The minimum field required for initial magnetic reversal in the non-irradiated state (D=0), Hinit(D=0), which equals Hp(D=0) in the presence of nucleation sites, can then be determined from the magnetization curve as indicated in Fig 3. On the other hand the field necessary for a reversal close to the saturation level of the film irradiated with dose D, Hsat(D), can be derived from the remanent magnetization curves measured after stepwise increased irradiation doses. Hsat(D) has to be determined as a function of exposure dose for a film grown under process conditions that can be stabily reproduced for the media to be patterned. The minimum dose required for optimum magnetic contrast is then defined by taking the minimum dose that fulfills the condition
Hsat(D) < Hlnit(D=0).
This guarantees that each magnetic bit cell that is written in opposite direction to the previously saturated matrix will induce two sharp transitions in the read-back signal, whereas parallel oriented bit cells will induce a transition signal only if the saturation magnetization Ms has been lowered in the bit cells by the irradiation, i.e., if Ms(D=0) > Ms (D) . In cases where the recording system is more tolerant with respect to the noise level in the read-back signal, it may also be acceptable if Hsat is slightly above Hinit(D=0).
The underlying idea of the invention is therefore to optimize, i.e., minimize the required irradiation dose for a given external write field or. to determine the narrow range for suitable external fields that can be used by a write head to properly write magnetic bits within the bit cells that were predefined by the irradiation for a given irradiation dose. Moreover, the procedure can be repeatd using varied magnetic material properties, in order to optimize the magnetic parameters with respect to target values of the write field and irradiation dose.
In order to avoid migration of the domain walls from the bit cell edges into the matrix the following relation has to be fulfilled:
Hβat(D) < Hwrite ≤ Hinit(D=0),
where Hsat(D) is the saturation field of the irradiated bit or island, Hinit(D=0) is the field required to start magnetic reversal in the matrix by domain wall propagation and/or nucleation, and HWrite is the suitable external write field sought after, i.e., the write field having a good magnetic contrast. Thus, the suitable external write field must lie between Hsat(D) and Hp(D=0) and also between Hsat(D) and Hn(D=0) . As to the relationship between the values Hirlit, Hp and Hn, Hinit is the smaller one of the two values Hp and Hn, respectively. In most cases, Hinit will be equal to Hp.
This procedure can be illustrated using Fig. 3 showing the remanent magnetization curves of the unpatterned magnetic media film before irradiation (curve I) and after stepwise increasing the irradiation dose (curves II, III, IV and V) . In addition, a typical run of a magnetization curve measured for the non-irradiated media after applying first -Hsat(D=0) and subsequently +Hc(D=0) is displayed (curve without dots) and the propagation field Hp(D=0) is indicated which, in absolute values, is somewhat lower than Hn(D=0) indicated in curve I.
Therefore Hinit(D=0) equals Hp(D=0). The saturation fields for the applied different doses (curves II-V) , Hsat (D= II, III, IV, V) are marked, by dots in the respective curves. It can be seen that only Hsat(D=IV) and Hsat(D=V) lie below Hinit(D=0) n( therefore fulfill the previously mentioned condition. Because irradiation time shall be minimized, the dose which will be used in the patterning process is therefore given by dose D=IV. The write field which generates complete reversal in the irradiated par,s without reversing the "as grown" matrix can also be determined from this figure and lies around 0,3 x 105 A/m, i.e., a little bit above Hsat(D=IV) and just below Hp(D=0) for this specific case.
The outlined procedure can also be carried out with milli-, micro- or nano-patterned samples or a sample that has been exposed to locally graded doses . Such local measurements can as well be used to determine Hsat # Hwrite and Hinit_
The outlined procedure for determining the optimum dose and/or suitable write field range can also be repeated for different thin film growth parameters or materials to optimize the magnetic raw material parameters of the media.
In a special embodiment, the invention provides an ion beam patterned medium generated by minimum irradiation dose for a specific magnetic medium at which a suitable write field can be defined. It has, however, to be mentioned that the invention is not restricted to ion projection or other ion beam patterning, but that any suitable methods that can magnetically soften the magnetic film in the predefined bit cells, such as e-beam or any other projection or irradiation technique or near field interaction (e.g., stamping or microcontact printing) can be used for this patterning and can be optimized in the same way.
The method for determination of a suitable external write field of a patterned magnetic storage medium consists of the
following steps. First, the remanent magnetization of the unpatterned, continuous magnetic storage medium is measured in the "as-grown" state and after increasing irradiation doses, leading to respective magnetization curves from which the parameters Hsat(D) and Hinit(D=0) are determined. Subsequently, the minimum irradiation dose and/or the write field range are determined according to the following formula
Hsat(D) ≤ Hwrite < Hinit(D=0).
Finally, the medium is irradiated with this minimum irradiation dose determined in order to generate a patterned magnetic storage medium.
When using the method according to the invention different signal strenghts are achieved for bit cells oriented parallel and antiparallel to the surrounding matrix. For an antiparallel orientation a signal Vffi proportional to ABS [Msat(D) + Msat(D=0)] where Msat(D)is the saturation magnetization of the irradiated bit cells and Msat(D=0) the saturation magnetization of the unirradiated matrix can be read-back twice. For a parallel orientation V^ is proportional to ABS [Msat (D) -Msat (D=0) ] which can also be read back twice if it is different from zero. In most cases there will be two large signals from the transitions in the antiparallel orientation but only very small signals in the parallel orientation. Nevertheless, since this is a patterned media also the absence of a signal at a predefined location can be interpreted as a binary information.
Since in contrast to a continuous medium also the absence of a signal at a predefined island position can well be interpreted in a patterned medium, one may consider to maximise the signal or better the signal-to-noise ratio for the antiparallel orientation of the bits.
Furthermore the inventive method could be used for intermediate steps for patterned media introduction, as will be described in the following.
It is possible to provide only the servo structures which can be written with a broad magnet before file assembly. This would have the advantage to reduce costs which arise during state of the art servo write processes at the end of the HDD assembly. Data tracks can then ,still be independently written in the unpatterned regions during the normal HDD operation with the advantage that potentially disturbed servo patterns can be recovered by an external field at any time. It is also possible to provide two switching levels on the medium by irradiation at two dose levels, where servo and data structures can be structured and independently written. It is, however, possible to create more than two switching levels, if necessary (e.g., Hsat(bit) < Hwrite(bit) ≤ Hsat(servo) < Hwrite(servo) < Hlnit(D=0)) .
Finally, the inventive method could be used for patterned media introduction in a form where not the bits are patterned but track (and maybe servo) patterning is performed. This is shown in Fig. 4. In this case, a magnetic recording medium is provided that comprises a magnetic track arrangement consisting of reduced HC(D) tracks being intersected by high Hc(D=0) barriers, whereby said tracks and barriers can be manufactured according to the inventive method. The main advantage of this medium is that there will occur' no „donut problem" for the circular tracks because they are intersected by servo informations, and also because the ion field or sample or mask can be rotated during the patterning. With this, a reduction of track edge noise can be achieved.