In the past 10 years, the density capability of magnetic hard drives has doubled almost every 2 years. In todays IT savvy world, the demand for more information storage space is always increasing and this trend will only continue with time. The market growth is expected to be 10% or more annually for the coming years. Current technology of recording the Hard Disk Drive (HDD) employs the use of Perpendicular Media Recording (PMR) in which the magnetization pattern is written on a thin film with perpendicular magnetic anisotropy axis. The 1st PMR drive to be introduced to the market was developed by Hitachi in 2006, and soon this technology was adopted by all models to record commercialized HDD.
Today's hard disks usually have recording densities of up to 0.5 Tbit/in2. Areal density in a magnetic data storage can be increased by the scaling method, but improving the density beyond 1 Tbit/in2 by using the same granular method will not be possible. This difficulty arises due to two main reasons. The first reason is that the minimum number of grains per bit is limited to maintain the required signal to noise (SNR) ratio. The second is the occurrence of the superparamagnetic phenomenon. As the grain size becomes too small, the magnetization state becomes unstable thermally, and this may lead to the loss of data (Yang, Chen et al. 2011).
The engineering limit of conventional PMR has not been reached yet; however for further increase of recording density with PMR, these technical difficulties need to be resolved. Patterned media is one such technology which promises to sidestep the density limitations imposed by today's conventional technology. In this technology, data is recorded in an ordered array of highly uniform magnetic islands separated by a nonmagnetic matrix or material with altered magnets. Each island behaves like a single magnetic domain, and this may help in significantly reducing the statistical noise and transition noise problems. The second advantage is that the focus shifts from depending on grain size to depending on the volume and the entire magnetic effect to achieve high data densities.
To implement the patterned media technology, an entire paradigm shift is needed in the HDD industry. The main challenges associated are fabrication of large-area arrays of magnetic grains with a size of less than 50nm at low cost, difficulty in addressing of the array with high spatial precision and different recording and writing mechanisms.
The focus of my final year project is on the fabrication and the studying of the magnetic behavior of patterned media. There are many patterning techniques which are being experimented with to make it commercially feasible. The first method is lithography. Conventional process of photolithography does not address the need for a high resolution patterned media, and instead, the focus is shifted to advanced nanofabrication techniques such as nanoimprint lithography electron beam lithography, X-ray lithography and interference lithography. The second technique is natural lithography which makes use of the self-ordered structures found in nature. This method includes anodized alumina, template growth using self-assembly and guided self-assembly. One more way of patterning is nanoparticle self-assembly.
This project gives me an opportunity to study the current process of fabricating patterned media. I aim to study the challenges faced in the fabrication process of patterned media and a way to possibly address these challenges through experimental analysis. Before establishing the project statement and research plan let's shed some light on the previous work that has been done related to this issue.
REVIEW OF FABRICATION METHODS:
The advantages of patterning recording media were demonstrated as early as 1963 by Shew et al. They showed that the cross-talk and noise problems related to head positioning errors could be reduced by discrete patterned tracks and furthermore, allow increased tracking tolerances (Shew 1963).
I will now evaluate some of the advantages and disadvantages of the fabrication techniques in this section, and in the next section, I will state the direction and scope of my project and a timeline by which I aim to achieve my objective.
LITHOGRAPHY METHOD
In this process, a pattern is created in the resist layer by employing methods such as electron beam, x-ray or interference lithography. The pattern in the resist is then transferred over to the magnetic film.
Interference Lithography:
This is a maskless technique and is performed by combing two beams of coherent radiation at the specified angle at the exposure plain. The interference generates a sinusoidal intensity pattern given by p = λ / 2 Sinθ (θ is the half angle). Patterns of dots are produced by rotating the sample by 900 between exposures (M.L. Schattenburg 1995).
A number of patterned islands have been investigated using a variety of materials. Fabrication of arrays of Co and Ni dots have been carried out by A. Fernandez et al (A. Fernandez 1996), Y. Hao et al (Y. Hao 2002) and others (A. Carl 1999, M. Farhoud 1999, T.A. Savas 1999), and these show periodicities ranging from 70nm to 100nm. Magnetic multilayer dots and corresponding multilayer films, such as Co/Pt and CoNi/Pt, have also been fabricated and analyzed (M. Thielen 1998, A. Carl 1999, M.A.M Haast 1999).
The advantages of this technique is that it gives rise to a high throughput and also results in large area regular arrays (2005). However, the limited wavelength gives rise to a disadvantage of achieving the sub-50 nm resolution as essential for patterned media.
The development of the achromatic interference lithography system may help in achieving a higher resolution. For patterned media which may require circular periodicity, a system has been developed by Solak and David (H.H. Solak 2003) which produces circular periodic patterns.
X-Ray Lithography:
This process, also called as EUV process, makes use of a thin membrane mask usually made of Si Nitride or Si Carbide and is covered with an X ray absorber like Gold (Au) or Tungsten (W). To pattern the radiation from the X-Ray source, the mask needs to be held in close contact to the resist.
Since this process is a parallel process, area as large as the mask can be patterned during a single exposure.
However, the EUV process is a very expensive process to conduct. Due to large exposure time, the throughput will be low, and a need for a high intensity x-ray source such as synchrotron arises. Moreover, achieving a sub 100 nm resolution is difficult as it is limited by mask fabrication and a reliable gap setting.
Permalloy cylindrical dots as small as 88nm have been fabricated by C. Miramond et al (C. Miramond 1997). Co dots of around 200 nm have been produced as well (F. Rousseaux 1995, J.E. Wegrowe 1999).
Electron beam Lithography (EBL):
E-Beam Lithography is by far the most used for nano-devices fabrication mainly because of its high-resolution capability. It also remains the most reliable method for sub-100 nm feature sizes. An e-beam system usually consists of a scanning electron microscope (SEM) and a beam drawing system attached to it.
A focused beam of electron is scanned across a substrate covered by a resist that changes its solubility properties according to the energy deposited by the electron beam. Usually, the e-beam lithography is used to pattern the polymethyl methacrylate (PMMA) layer.
The possibility of having a variety of pattern transfer processes can be utilized to create magnetic elements of various shapes and sizes.
The wide use of this technology is because it is the only one which can achieve a resolution as needed for the patterned media. However, for most resists, it is not possible to go below 25 nm, and the limit has been set to 20 nm (J. A. Liddle 2003). Furthermore, since this is a serial process, the throughput is low and this makes the entire process very expensive. Although this system has been and currently is being studied in laboratories (W. Wu 1998), commercializing it can be extremely difficult.
Projection EBL method has been put forward to address the throughput problem. In this method, the patterns on the masks are demagnified and projected onto a wafer while still being scanned in the same way as the optical scanning stepper (Sbiaa and Piramanayagam 2007).
A major obstacle to this method could be the proximity effect, where electron spread to the neighboring features.
SELF ASSEMBLED NANOPARTICLES
This is another way of fabricating by taking advantage of the self-ordered structures found naturally to accomplish high density arrays.
Template growth:
This method makes use of the self-organized porosity of certain microporous membranes to fabricate magnetic nanodots. The first membranes used for magnetic nanoparticles were anodized Al2O3 (Kawai and R. Ueda 1975). An oxide layer is formed when aluminum oxide is anodized. This layer contains a closed pack of pores arranged hexagonally, and the size and spacing of the pore is determined by the anodizing conditions like the current density, voltage and solution pH. Zhang et al (Z.B. Zhang 1999) have shown pores size of nearly 9 nm (Aranda and Garcia 2002) with a packing density of almost 450 Gb/in2. It has been shown that by templating the aluminum surface through the use of focused ion beam, molding or nanoimprinting, the length scales can be extended significantly. This can possibly address the main problem of long range order. Through electrodeposition, or by making use of the template as a shadow mask, it may be possible to develop low aspect ratio dots, as required. Xiao et al (Z.L.. Xiao 2002) propose that structures, which are created by depositing films on top of the alumina mask to produce a film with regular array of holes, may be used as a form of high density patterned bit medium.
Block Co-polymers assembly:
Self-assembled block co-polymers can be used as a template or a mask for patterning magnetic media. This process is based on a chain of polymer made of a mixture of two immiscible monomers A and B. After spin coating the substrate with AB polymer, nanoscale periodic ordering of A and B can be attained, depending on the concentration of each monomer and the substrate surface energy. The periodicities are in the scale of 10-200 nm (Sbiaa and Piramanayagam 2007). Magnetic islands with a size as small as 25 nm (F. Ilievski 2004) have been created too. In case of patterned media, Cheng et al (J.Y. Cheng 2002) demonstrated some useful findings.
The method of artificially guiding the self-assembly process is recommended to increase the long range order. It can either be through physical guidance or chemical patterning.
Despite the advantages, these processes are still in the early stages of experimentation and important problems such as bit size distribution, side wall profile and individual alignment needs to be studied and addressed. Furthermore, the distribution of the particle shape and size results in non-uniform magnetic properties (J.Y. Cheng 2004).
Nanoparticle self-assembly:
Self-assembled magnetic nanoparticles of FePt have been extensively studied for the purpose of recording applications. Sun et al (S. Sun 2000) were the first to demonstrate mono-dispersed FePt nanoparticles with diameters as small as 4 nm and the size distribution of particles to be less than 5%. In the L10 phase, these FePt nanoparticles have sufficient magneto crystalline anisotropy to remain thermally stable (D. Weller 2000), and due to this significantly lower superparamagnetism, it makes them extremely attractive to be considered as recording material. Moreover, FePt nanoparticles are produced and assembled without using lithography, and this is a major factor in saving time. The chemical synthesis method gives the best packing densities and magnetic anisotropy. However, there is a problem of particle agglomeration during annealing (Yao 2006), which have not been overcome completely. The uniformity of the chemical ordering due to annealing is another issue which needs to be addressed. This technology is still in the research stage, however, the potential remains enormous.
NANO-IMPRINT LITHOGRAPHY (NIL)
This technology has created a lot of interest as it has the capability to be implemented in magnetic disk patterning with the potential of reducing the cost and improving the yield and reliability. NIL involves a nanopatterned template (mold; usually made by high resolution tool such as EBL) which brought in contact with the resist to deform it (Sbiaa and Piramanayagam 2007). There are many types of NIL depending on the way the resist is embossed. Chou et al (Chou 1998) were the first to propose this methodology in making patterned media.
The two main methods are thermal NIL and step and flash NIL. In Thermal NIL, the mold is first pressed to the resist using high temperature and pressure. Then, the mold is removed after the resist is cured, and this leaves a topographic relief in the now imprinted resist. This topography in the resist usually does not reach the substrate surface, and hence there is a need to perform etching to remove the remaining resist material from the bottom of the features until a desired thickness is reached, and then the resist is completely removed by lift-off (Ling, Montelius et al. 2005). Since the time needed to heat and cool the resist is long, coupled with the fact that high temperature and pressure may have effects like thermal expansion and deformation, new methods were sought after, and step and flash NIL looks like the promising answer to address these issues.
Step and flash NIL was proposed by Willson et al (Wilson and Colburn 2002) and involves a resist with low viscosity at R.T. which fills the mold feature. It is then polymerized (solidified) by UV light. Once the template is released, the patterning is made in the same manner as used in Thermal NIL. The advantages of this process is that the need to have high temperature and high pressure is eliminated since everything is done at room temperature (Chou 1998, Bailey, Choi et al. 2004) . Advantages shared by both the NIL processes are related to mask making, method to pattern large area and mold/resist anti-adhesion (Sbiaa and Piramanayagam 2007).
All in all, vast majority of this review is inclined in suggesting that patterned media, with its advantages compared to continuous recording, promises to be the technology for the future where magnetic recording is concerned.
PROJECT STATEMENT:
The major techniques discussed above seem to suggest that although the future of magnetic recording lies in patterned media, there are major issues which need to be addressed in order to make it commercially viable. NIL, although relying on other high resolution lithography tools for synthesizing the mold, is a very promising area in terms of resolution once the mold is made. This project will focus on the using some of the techniques discussed above to fabricate patterned media, make use of tools such as Magnetic Force Microscopy to understand and analyze the magnetic state and behavior of the magnetic nanoparticles and model ways in which current magnetic properties' issues can be addressed.
PROJECT PLAN:
SEMESTER I
TASK
Understanding basics and familiarization with magnetic recording
Lab Trainings for NTF Lab and ISML
Identifying the process and bottlenecks
Fabricating and analyzing magnetic nanoparticles
Modelling and Testing
MONTH
AUGUST
SEPTEMBER
OCTOBER
NOVEMBER
DECEMBER
SEMESTER II
TASKS
Modelling and Testing
Finalizing Thesis
Final Presentation
MONTH
JANUARY
FEBRUARY
MARCH
APRIL
