Squeeze Casting for the Product of Metallic Parts and Structures

Thomas Wei Jie Kwok , ... Beng Wah Chua , in Encyclopedia of Materials: Metals and Alloys, 2022

Case Study three – Helium Hd Casing

Hard disk drive (HDD) engineering has been steadily improving throughout the years. One of the more recent innovations is a helium filled HDD. Past replacing air with helium, the disks experience less drag when spinning, assuasive for thinner disks (allowing for more than disks per HDD) and faster spin speeds. Helium filled HDDs therefore accept a larger capacity, faster read and write speeds and are also quieter (Seagate, 2016). Still, the claiming facing the industry was to develop a method to keep the helium in the HDD without leaking out over time.

HDD base plates are traditionally manufactured from cast Al alloys via HPDC. However, it was found that cast Al HDD base plates contained interconnected porosity microchannels which would let helium to leak out (Seagate, 2016). Traditional HDD casings are not hermetic and therefore not suitable for this awarding.

Possible methods to create a fully dense, cypher porosity base of operations plate include full machining from extruded wrought alloy feedstock or forging with subsequent machining but both processes are considerably more expensive than HPDC. Squeeze casting, however, is an alternative method with a costing level similar to HPDC. While it is certainly possible for squeeze casting to meet the requirements for a hermetic base plate, the preform pattern is not without challenges. The cross-department of the hard disk baseplate shows a thin base with a thick wall effectually it. Thin-to-thick and thick-to-thin sections are challenging from a die filling perspective but also affects the uniform application of force per unit area (Fig. 7(e) and (f)).

To ensure laminar die filling properties, the dial speed can be slowed to allow even filling of the thick walls. If the punch speed is also fast, then jetting would occur in the thin-to-thick sections, resulting in severe entrapped gas porosity. However, the slow filling speed also allows the thin base to fully solidify before the walls can be filled. The result is that the punch 'bottoms out' and when force per unit area is applied, the punch presses against an incompressible solid base of operations and the thick walls receive picayune to no pressure. This results in galling and extreme dice wear along the base and a large number of casting defects such as shrinkage porosity in the walls which cannot exist tolerated in this particular awarding.

The solution to this puzzler consists of many incremental changes. Firstly, the preform design should be altered to permit better flow and minimize jetting fifty-fifty with a fast punch speed. This included rounding out sharp corners and merging several thick sections together for example. Secondly, the punch can exist divided into several inserts. Die springs or disk springs tin be inserted behind each insert to regulate the amount of forcefulness on the area nether each insert. When the punch "bottoms out" on the sparse base, the springs on the punch insert directly above volition compress, allowing the punch to go on pushing downwards and apply pressure on the thick walls. In several shrinkage prone spots, a 2d hydraulic organisation may also be employed.

Read full affiliate

URL:

https://www.sciencedirect.com/science/article/pii/B9780128197264000387

Mitigating criticality, part I: Material exchange

Alexander King , in Critical Materials, 2021

Nine years from discovery to commercialization

Hard disk drive drives (HDDs) were first introduced as a medium for computer random access memory (RAM) by the IBM Corporation in 1954. They became the dominant data storage format in mainframe computers in the 1960s and quickly took over personal computers later they were first introduced into this sector by IBM, in 1983. The data density of HDDs has increased steadily over time, through the adoption of a variety of new technologies: The commencement hd in a PC had a chapters of 10   MB in a five¼   in. format, and today'due south units offer multiple terabytes in drives every bit modest as two½   in. The ability to detect magnetic field changes on a tiny scale has been a major contributor to this improvement in information density: Pocket-sized magnetic domains can be packed closely together on the surface of the hard disk, but the magnetic flux associated with each binary digit of information also depends on the domain size. Tiny information bits produce correspondingly infinitesimal amounts of magnetic flux, and they can only exist detected by highly sensitive detectors.

Early on read/write systems for HDDs—along with other magnetic recording devices—relied upon the current induced in a coil of wire as it moved through the changing magnetic field produced by the recorded data. This system is limited at small-scale sizes because the induced current in the field gyre depends on the amount of the magnetic flux and the number of turns in the coil. As the information $.25 shrank, the flux declined, and the number of turns in the field coils likewise had to be reduced to brand them smaller, and the readable current in the read head threatened to shrink below the detection limit.

An alternative technology for reading data from a magnetic storage unit is the apply of magnetoresistance; an effect first discovered by Lord Kelvin in 1856. The electrical resistance of some materials tin can change when they are placed in a magnetic field, depending on the alignment of the current with the field management. The resistance of a small conductor is too easier to measure, using an externally applied voltage, than is the current induced in a very small electromagnetic generator. The resistance detection method gains the advantage over electric current detection when the devices are scaled down to very pocket-sized sizes.

IBM began developing HDD systems based on anisotropic magnetoresistance (AMR) in 1969 and brought its first 1 into production 14   years later, in 1983. This is a moderately rapid material-dependent technology adoption, but information technology is also fair to note that it was based on a discovery made 127   years earlier, and the materials involved and the nuts of sensor blueprint were already well known before IBM began its development efforts. The evolution fourth dimension reflects the complexity of bringing new technologies to market in the loftier-applied science sector, and the fact that the MR drives were, in reality, radically unlike devices than the induction electric current HDDs that they replaced.

A revolution in the science of magnetoresistance occurred in 1988 when giant magnetoresistance (GMR) was discovered independently by teams working in France and Germany [15, 16], a discovery for which Albert Fert and Peter Grünberg shared the 2007 Nobel prize in physics. GMR occurs in conductive sparse-movie superlattices of fe and chromium with individual layers near x atomic layers thick, and it produces much larger changes of resistance than AMR, leading to greater sensitivity and the ability to scale magnetic memory to much smaller sizes. GMR was commercialized in IBM'southward HDDs in 1997, ix   years afterwards information technology was first discovered, and it quickly became the manufacture standard for these pervasive memory devices.

IBM's rapid adoption of GMR technology doubtless benefited from its previous work on adapting AMR to work in HDDs, but the replacement of AMR by GMR still required a significant R&D endeavor.

The speed with which the new material systems involved in both AMR and GMR drives were adopted owes a cracking deal to a big economic driving forcefulness, with very potent competition among the HDD manufacturers of the time. IBM besides benefitted from close links to between its manufacturing divisions and its own world-class bones inquiry chapters (a program that was demonstrably capable of producing its own Nobel Prize winners).

Read full chapter

URL:

https://www.sciencedirect.com/science/article/pii/B9780128187890000050

Applications of Magnetic Materials

Giuseppe Florio , in Encyclopedia of Smart Materials, 2022

Hard Disk Drives

Hard disk drives are among the most mutual technologies used nowadays because of their capacity and performances (Spaldin, 2011). The device is constituted by several components where the magnetic layers used for storage is coupled to a rigid substrate (prevention of shocks, reduced vibrations and thermal stability), an underlayer (isolation of magnetic grains, prevention of inter-particle interactions, control of magnetic flux during the writing process) and an overcoat (protecting interface between the magnetic layer and the read or write heads).

Data storage is performed using the domains of the magnetic fabric with the orientation of magnetization representing the physical realization of a single bit with ii logical states. The storage layer is constituted by a thin film (xx nm thick or less) with magnetically anisotropic grains having a very small bore (10 nm).

The write caput is typically made of a ferromagnetic or ferrimagnetic textile surrounded by a scroll with a time varying electric electric current flowing through it. The current generates a variable magnetic flux that penetrates into the magnetic layer. When the magnetic field is removed the magnetization of the medium remains, thus storing the bit. The recording of data tin can be perpendicular or longitudinal with respect to the aeroplane of the magnetic material. The erstwhile method has been introduced in club to increase the amount on information stored. As a matter of fact, i of the main drawbacks in magnetic storage is the fact that reducing the size of the grains in the recording layer increases the probability of reversing the magnetization of a region due to thermal fluctuation (superparamagnetic result). This event is more pronounced in the case of longitudinal storage. Perpendicular storage tin circumvent this result by increasing the density of magnetic elements and the utilize of hard magnetic materials with larger value of coercivity, thus less sensitive to thermal fluctuation.

In order to perform the reading functioning, one tin can employ two different methods. In principle, the inductive technique used for writing tin be besides used for retrieving the information. On the other manus, the fields in the domains of the magnetic layer are small and, thus, the signal inductively generated in the read caput is small. The use of Anisotropic magnetoresistance (see Section "Magnetoresistance") has led to the evolution of a new generation of reading elements called dual-stripe heads that solve this trouble: the reading head tin can sense the variation of the magnetic field in the storage layer through the modify in the electrical resistance of the magnetoresistive chemical element. Finally, the appearance of devices based on Giant magnetoresistance (equally spin valves) used as extremely sensible field sensors have furtherly improved the reading adequacy of heads and improved speed of data transfer.

Read full chapter

URL:

https://world wide web.sciencedirect.com/science/commodity/pii/B978012815732900067X

Magnetic Materials: Domestic Applications

S.J. Collocott , in Reference Module in Materials Science and Materials Engineering, 2016

5.i Reckoner Disk Drives

Computer hard disk drives are a prime example of where operation enhancements are traceable to improvements in permanent magnets. In a stock-still disk drive permanent magnets are used in the spindle motor, in the actuator that drives the read/write heads (mostly termed the vocalism coil motor (VCM)), and the latch assembly (run across Figure iv). Deejay drives nowadays have a 2.5, 3, or iii.5   in class factor, every bit emphasis has switched from miniaturization to making the storage chapters of the drive every bit big every bit possible.

Effigy 4. Hard disk bulldoze unit of measurement; the VCM actuator can be seen in the top left hand corner.

The availability of high-remanence rare-world magnets has enabled the disk drive designer to reduce dramatically the size of the VCM, equally only a small volume of magnet fabric is required, and to decrease access times. Disk drives use a high-grade sintered Nd2Atomic number 2614B, showing typically a remanence of 1.4   T, an intrinsic coercivity of 1190   kA   m-i, and an energy product of 360 kJ   yard-3. 1 or ii magnets may be employed, with a full weight of 15.five   grand, and render poles are fabricated of depression-carbon steel. Spindle motors may use sintered or polymer-bonded material, with the latter often favored as it tin can exist easily molded to shape. Spindle motor speed is of import every bit the college the speed the shorter the seek time. High-capacity drives found in file-servers take a spindle speed of ten   000   rpm, while drives used in desktops and notebooks have speeds of 7200   rpm and 5400   rpm, respectively. Removable cartridge drives of 100–250   MB capacity use linear motors and polymer-bonded Nd2FefourteenB, to keep their price down. It should be noted that small-scale disk drives used in video cameras and lap-top computers are themselves beingness replaced by solid land memory, in many high-operation products.

Read full chapter

URL:

https://www.sciencedirect.com/science/article/pii/B9780128035818018464

Applications and instance studies

Ian Hutchings , Philip Shipway , in Tribology (2d Edition), 2017

ix.6 Magnetic Data Storage

The magnetic hard deejay drive, a key component for data storage in desktop and laptop computers too as larger-calibration information servers, has increased in capacity by a astounding extent since it was start introduced in the 1950s. The binary data is stored as information encoded into magnetized regions on the surfaces of rapidly spinning disks. In modern drives the data is written to the disk and read from it past a read/write head that is supported on a very thin film of air—a gas bearing—as the deejay spins beneath it. The first disk bulldoze to use aerodynamically-supported heads, the IBM 1302 of 1965, was the size of a minor car and stored 120  MB (1 megabyte   =   106 bytes) at a cost (in 2017 terms) of well-nigh US$2 one thousand thousand; in contrast, a palm-sized drive of the present day stores several TB (1 terabyte   =   1012 bytes) for around $100, and data chapters is probable to proceed to increase. Current data densities exceed one terabit per square inch (approximately one.five gigabits per square millimetre; 1 byte   =   8 bits), equivalent to each fleck of information taking upwards a square area 25   nm by 25   nm. This dramatic increase in storage capacity has been achieved by a progressive reduction in the size and a change in orientation of the magnetized domains in which the data is stored as well as by advances in the physics of the read–write process, just it has also involved a remarkable reduction in the 'flying top' of the head over the deejay surface. A small-scale gap between the head and the magnetic material is needed to achieve high lateral resolution in the reading and writing processes. The earliest drives had a head flying top of most 6   μm, just in modern drives this has been reduced to less than x   nm, a feat that demands not only precise modelling of the gas film and head dynamics but also the production of disk surfaces of extreme flatness and smoothness.

Figure 9.33 shows the typical internal arrangement of a hd, with the caput at the stop of its support arm; each surface of the disks in a multi-disk stack is accessed by a separate head. The disk typically has a bore of 63.5 or 88.9   mm (2.5 or 3.5 inches) and rotates at 7200   rpm. The read/write function is performed by a very small active region embedded into a larger slider, 1–2   mm beyond. The design of the slider tin be complex, equally illustrated in Fig. ix.34 which shows one instance. The surface of the slider is patterned with steps and ledges typically 0.5–two   μm in acme that are designed to create a distribution of aerodynamic force per unit area across the slider surface which raises the slider above the disk in a stable manner. Aerodynamic lubrication acts in the wedge-shaped air film between the stationary slider and the moving deejay surface. The internet pressure difference betwixt the lower and upper surfaces of the slider results in a vertical force that is counterbalanced past the angle of a flexural element in the support arm, and the flying elevation and attitude are controlled past the equilibrium between the aerodynamic and bending forces. Feedback mechanisms are used to command the head position very accurately. Although most drives currently operate in air, there are potential advantages in terms of reduced flying top and lower frictional drag in using a gas with lower viscosity, and helium is therefore used in a few applications.

Fig. 9.33. Internal organisation of a magnetic hd drive, showing a read/write head fastened to the stop of its back up arm. The arm is moved beyond the surface of the deejay so that the caput tin access circular tracks at various distances from the heart of the deejay. Each surface of the 3 disks in this instance has its ain read/write caput

Eric Gaba, Wikimedia Eatables, CC-BY-SA-3.0

Fig. nine.34. Example of a slider and magnetic deejay read/write caput, showing: (a) the lower surface of the slider, patterned with raised areas and recesses to command the airflow across its surface and the resulting pressure distribution and aerodynamic lift; and (b) a cross-section through the slider and head, showing (exaggerated) the wedge-shaped air movie formed between the slider and the deejay

based on patent US 8174795

The disk platters themselves provide an splendid example of surface engineering for functional applications, combining mechanical, data storage and tribological properties. Multilayer structures, equally shown in Fig. 9.35, are deposited on a substrate made from rolled aluminium–magnesium alloy (AA5083; 96% Al, four% Mg) which is highly polished, with a typical roughness Ra of fifteen–25   nm. Electroless plating (run across Section seven.4.i) is used to deposit a relatively thick (10   μm) nickel–phosphorus layer (ninety% Ni, ten% P) which acts as a hard substrate for the functional magnetic layers and is also polished (to 0.5–2   nm Ra). The phosphorus increases the hardness of the deposit. A sequence of very sparse layers is then deposited by PVD processes: a sublayer of chromium (~   50   nm) to bond the magnetic layer to the Ni–P layer; the magnetic layer in which the data is stored (~   30   nm of a circuitous alloy, often Co-based, sometimes itself with a multilayer structure); and an even thinner (~   15   nm) hydrogenated diamond-similar carbon (DLC) overcoat which protects the magnetic layer from mechanical damage. Finally an extremely thin (~   1   nm) layer of lubricant (typically perfluoropolyether, with very low vapour pressure) is deposited by dip-blanket from solution, again to protect the disk surface from damage if the head touches it. Both the lubricant layer and the DLC layer provide some corrosion protection to the magnetic picture show.

Fig. nine.35. Schematic cross-section through the surface region of a typical hard disk drive for magnetic information storage (not to scale), showing the multiple coatings required for magnetic, mechanical and tribological functions

Read full chapter

URL:

https://www.sciencedirect.com/scientific discipline/article/pii/B978008100910900009X

Estimator and network applications

Atsufumi Hirohata , ... Mathias Kläui , in Nanomagnetic Materials, 2021

vii.2.1 Hard disk bulldoze read head

The first HDD, RAMMAC 305, was introduced by IBM in 1956 to store 4.iv MB on 50 disks (2 kbit/in2) [one]. Data writing was performed past introducing a stray magnetic field onto a polycrystalline magnetic recording media from a U-shaped electromagnetic ringlet [ii]. The data was read past detecting anisotropic magnetoresistance (AMR) depending on the magnetization directions of the recorded information. Past property a read head to a higher place the rotating media at a speed of 5400–15,000 rpm, the change in the AMR signals can be detected and converted into a differential indicate to identify two states, "0" and "1," of each data chip. In 1997, a GMR spin-valve HDD read head has been introduced past IBM, followed past a TMR head, introduced by Seagate, Hitachi, and Toshiba independently in 2005 [1,3]. The MgO-based TMR applied science is still used in current HDDs, at an areal density of >600 Gbit/in2. However, the resistance–area (RA) product of the MTJ tunnel barrier has to subtract as the storage areal density keeps on increasing. Indeed, the size of the MTJ reader decreases equally the areal density increases, which imposes a reduction in its RA product since the sensor impedance must remain of the lodge of a few tens of Ohms to match the pre-amp impedance.

For above 2 Tbit/in2 HDD development, information technology becomes increasingly difficult to employ MgO-based MTJs since the bulwark thickness becomes ultralow (below 1 nm), which impacts its reliability and reduces its TMR [4]. It has been proposed past Toshiba to use spin-valves embedding a nano-oxide spacer layer (NOL) formed of an oxide AlOx layer comprising metallic Cu pinholes. These pinholes are intended to locally restrict the current path perpendicular to the GMR stack. This NOL layer is formed by oxidizing an Al1−x Cu x alloy ( ten = two –3%) spacer layer [5]. A Co0.5Iron0.5 (2.5)/NOL/Co0.fiveFe0.5 (two.5) (thickness in nm) junction has demonstrated RA = 0.5–one.5 Ω⋅μmtwo and magnetoresistance (MR) = 7–10% at RT. These values are below the requirement for the 2 Tbit/in2 HDD and the reliability of these devices is very uncertain due to the extremely high electric current density flowing through the pinholes. Therefore, further improvement in these junctions is crucial.

Intense research has been carried out on spin-valves based on one-half-metal Heusler alloys and Ag spacer layers [6] which could found an alternative. Still, the growth of these materials oft requires high-temperature deposition or required annealing conditions may not exist compatible with the head fabrication process [7]. Besides, the total thickness of the reader in HDDs must exist below 20 nm since it determines the reader shield-to-shield spacing and therefore the linear downtrack resolution. This adds another constraint in the blueprint of the reader MR stack.

The chemical compound almanac growth rate (CAGR) of the global need for storage is nigh 40% due to the Internet of Things (IoT), 5th-generation (5G) communications, etc., while the growth of HDD storage density recently slowed downwardly to 25% CAGR (Fig. seven.2.1) [8]. For the farther improvement in the HDD areal density, a trilemma has to be overcome between areal density (grain size), thermal stability, and writability (Fig. 7.2.2) [11]. To improve the writability, possible solutions consist in assisting the magnetization switching during write by transferring additional energy to the local magnetization through heat transfer or microwave. Heat-assisted magnetic recording (HAMR) was proposed [12,13]. A light amplification by stimulated emission of radiation axle is employed to locally transfer estrus to the media. This is achieved through a plasmonic antenna which allows to create a thermal slope as high equally 10 K/nm. The resulting local heating of the media reduces the thermal stability of the information bit to be written, which eases the switching of its magnetization. A successful demonstration has been reported [14]. In 2012, TDK demonstrated a new HAMR head with an areal density of one.5 Tbit/intwo and a bit-error rate of x−2. Seagate also demonstrated a new HAMR drive in 2012. A problem which had to be solved is the reliability of the caput and peculiarly of the plasmonic antenna, which reaches quite elevated temperatures. A head lifetime in a higher place 1000 hours was demonstrated [xv]. In 2020, Showa Denko announced that they developed HAMR media with an areal density of v–half-dozen TB/in2, achieving up to eighty TB [16].

Figure 7.2.1

Figure vii.ii.i. Areal density of HDD and tape laboratory, demonstration, and products. Afterwards Refs. [8–10].

Figure 7.2.2

Figure vii.2.2. Trilemma in HDD evolution. Later Ref. [11].

Microwave-assisted magnetic recording (MAMR) was also proposed past Zhu et al. as another energy-assisted recording approach [17]. MAMR utilizes microwaves produced by a spin transfer oscillator patterned in the write gap of the write head to reduce the switching field by an order of magnitude [18]. Great progress has been made lately on this engineering science. In 2017, Western Digital appear the commercialization of MAMR-based HDDs [19].

Read total chapter

URL:

https://www.sciencedirect.com/science/article/pii/B9780128223499000134

Plasma-Immersion-Implantation

B. Rauschenbach , ... L. Roux , in Reference Module in Materials Scientific discipline and Materials Engineering, 2016

iii.1.1.ii Source/drain (S/D) region doping

S/D HDD p+−type PII doping for PMOS device is too an application that is used in production for DRAM manufacturing. The reward of PII compared to beam line implanter is the high surface concentration of the doping profile. BiiH6 is ordinarily used for the forerunner, since the co-implanted fluorine with BF3 could lead to contact reliability problems. Compared to the use of beam line implanter, plasma doping reduces the contact resistance by ~50% and increases the conducting electric current IDS by ~15% (Qin et al., 2007). Typical process conditions are biases of 4–six   kV and dose around ii×tenxvi  cm−2.

Some other development have also been performed to optimize the S/D BC (buried contact) doping. Here the goal is to recoup impurity loss during contact area etching and cleaning, and increment impurity concentration at the metal/Si interface without negatively impacting junction characteristics such as VT, SVT, breakdown voltages, and leakages. Qin et al. (2012) demonstrated a reduction of ~lxx% of the contact resistance of devices applying a PII implant of B2H6 through the contact opening simply earlier the PVD Ti/CVD W-based metallization.

Read full chapter

URL:

https://www.sciencedirect.com/science/article/pii/B9780128035818037267

Directed cocky-oriented self-assembly of block copolymers using chemically modified surfaces

R. Seidel , ... P.F. Nealey , in Directed Self-assembly of Cake Co-polymers for Nano-manufacturing, 2015

6.six.2 Fleck-patterned media

The evolution of hd relies on scaling the characteristic size of periodic structures to make loftier-density data storage possible. However, superparamagnetism imposed a limit for the smallest grain size in the traditional granular media. Instead, each fleck in bit-patterned media (BPM) is lithographically patterned and the magnetic islands are comparably larger in size than the grains in this conventional instance. This allows for the reduction of signal-to-noise ratio likewise equally provides thermal stability. The challenge shifts to the production of patterns with high area density, across 1   Tbits/in.2, which falls in the resolution range of self-assembling BCPs.

DSA tin be incorporated in the initial steps to create templates for subsequent blueprint transfer or replication. The sparse chemical patterns for DSA of BCPs were prepared past a rotary E-beam tool. Later on DSA, the patterns with density multiplication were transferred to a substrate, serving as a master template for nanoimprinting to create a high volume of replicates and then disks with magnetic media (Figure six.24). Two geometries of BCPs take been explored to encounter the density demand of BPM over one   Tbits/in.2. The showtime are the hexagonally close-packed cylinders (Hellwig et al., 2010) or spherical (Yang et al., 2013; Xiao et al., 2013) BCP domains, which provides the highest surface area density. The 2d geometry uses an orthogonal line cut to fabricate a rectangular unit of measurement.

Figure 6.24. SEM micrographs at various stages of scrap patterned media fabrication: (a) adult cake copolymer design on Si chief template, (b) Si master template, (c) quartz replica working template, (d) imprinted resist on disk, and (e) finished magnetic islands (inset: X-section TEM). An island placement tolerance of nm is accomplished at all steps.

Reprinted with permission from Albrecht et al. (2013). Copyright © 2013 IEEE

Although the close-packed patterns provide the highest expanse density of features, the integration with recording systems has establish some difficulties. It has been suggested to use rectangular bits with a scrap aspect ratio (the ratio of the cross-track to the downwards-rail pitch) greater than 1. With a line cut strategy (Ruiz et al., 2011), rectangular bit cells can be fabricated via a double imprinting procedure. Briefly, one submaster can be patterned with circumferential lines and the 2d is patterned with radial lines (Effigy 6.25).

Figure 6.25. (a) Schematic representation of a double imprint process to produce rectangular features. Ii "submaster" templates, 1 with circumferential stripes and a second one with radial lines, are imprinted onto a split "master" template. The intersection of the 2 sets results in rectangular features. (b) Circumferential block copolymer lines with a 27-nm total pitch. (c) Sidewall spacer line doubling from block copolymer lines with a 41-nm full pitch (top of prototype) to 20.5-nm pitch lines later on line doubling (lesser). (d) Loftier chip aspect ratio rectangular features imprinted in resist by the main template.

Reprinted with permission from Albrecht et al. (2013). Copyright © IEEE.

Other than the bits, servo patterns are also required to be created in BPM for the head to locate data precisely. Servo patterns are by and large irregular in shape, and thus need a specific integration strategy with the BCP assembly. For the chevron pattern, first an EBL sparse pattern was defined and the BCPs tended to follow the bends with different angles. Because of the free-energy penalization, defects formed at the chevron apexes. Accordingly, boosted vertical lines were added at the junction of horizontal tracks and chevrons. This helps to minimize the defective area at the interfaces by one-half, as shown in Effigy half-dozen.26 (Liu et al., 2011c). Subsequently, a new integration method, termed "kickoff burst," was proposed and the position error point for track post-obit could be encoded (Lille et al., 2012). From the EBL step, a fix of circumferential lines was intentionally shifted along the cantankerous-track direction. One approach to integrate servo and data patterns is to run a separate patterning process after the DSA of information region. This usually is not favored because the registration of the two patterns would exist extremely difficult.

Figure 6.26. Chevron design schematics and resulting lamellar patterns. For design I and 2 schematics (shown at the top), the solid lines represent the eastward-axle patterns at a catamenia of 2 50 0  =   54   nm, while the dotted lines stand for lamellae interpolation giving 2   × density multiplication. In pattern Two actress vertical e-beam patterns were added. SEM images of the resulting lamellae ordering are shown to the right. The cherry-red dashed lines indicate the boundaries between lacking and nondefective areas. The widths of the design II lacking areas are smaller equally a event of the additional vertical patterns.

Reprinted with permission from Liu et al. (2011c). Copyright © AVS.

Read full chapter

URL:

https://www.sciencedirect.com/scientific discipline/article/pii/B9780081002506000067

Mitigating criticality, function 3: Improving the stewardship of existing supplies

Alexander King , in Critical Materials, 2021

Rare earth magnets

As we await for sources of disquisitional materials, we demand to place targets that contain pregnant amounts of the materials that we seek and are also viable every bit recycling businesses. The largest book of rare globe magnet materials, bookkeeping for about 40% of the market, goes into industrial electric motors and generators, which as well incorporate big quantities of copper and ferrous alloys, with a relatively small amount on nonrecyclable cloth. These devices, however, typically take long service lives and are retired infrequently. Their collection and delivery to recycling centers involve transport over long distances, and the devices are of variable size and class, making their disassembly difficult to automate.

The second largest single component of the market for rare earth magnets is in hard disk drives, accounting for around 16% of the market in 2014. The needs of this industry can modify quite rapidly as HDDs are supplanted past other technologies and the class factors of the HDDs shrink with their ever-increasing memory density.

The replacement of HDDs by solid-state memory predominantly affects consumer and business concern computers, while HDDs remain the engineering of choice in large information centers because they are robust over large numbers of read/write operations: The life of solid-state memory is limited to a stock-still number of such operations, and data centers store and transfer information at rates that would fire out solid-land RAM very apace. Information centers proliferate with the growth of cloud computing services and Net-based social and business applications, and then they represent good targets for urban mining: They contain large amounts of rare-earth magnet fabric in a small geographical footprint, and a typical data eye can retire hundreds of thousands of nearly identical difficult disk drives each year. These HDDs comprise significant amounts of precious metals forth with aluminum and ferrous alloys, so they have potential every bit a recycling business venture.

In that location are two sets of rare earth permanent magnets in a typical difficult deejay drive. One of these is in the spindle motor, which turns the disk associates, and the other makes up the phonation whorl motor, which positions the read/write head (Fig. seven.4). In a 3½   in. HDD, the magnets in the vocalism-coil motor total betwixt 10 and twenty   g of sintered neodymium magnet material with little or no dysprosium added; for a two½   in. drive the voice-curl magnets total around 2.5   g [10]. Spindle motors contain smaller quantities of resin-bonded neodymium magnet material that is unremarkably not considered economical to recover.

Fig. 7.4

Fig. vii.4. The internal components of a hard disk bulldoze, showing the locations of the rare earth magnets.

Reproduced with permission from A. Walton et al., The use of hydrogen to split and recycle neodymium-atomic number 26-boron-type magnets from electronic waste, J. Clean. Prod. 104 (2015) 236–241. Copyright 2015, Elsevier.

Early efforts to recover magnets from HDDs were based on manual disassembly. A typical HDD is opened past removing about 10 screws from its lid, and and so several more must exist removed to extract the voice-coil and spindle motors. The spindle motor is more complicated to excerpt, and because it contains a smaller magnet than the voice-coil motor, it is usually ignored in manual processes. An experienced disassembler with an electric screwdriver may exist able to extract the voice-coil magnet assembly from an HDD in less than ten   min. For a 3½   in. HDD at the high end of the magnet size range, this assembly would comprise approximately $0.45 worth of REEs at 2018 prices, and so extracting REEs from around 10 HDDs per hr does not pay the hourly wage of a low-skilled worker in a developed country, and disassembly is simply the commencement step in recovering the value of the rare earth. Afterward extracting the voice-coil magnet assembly, several further steps are required to recover the magnet materials, including separating the magnets from their μ-metal bankroll, demagnetizing, stripping the coating, and preparing the material for reuse through some procedure such equally grinding or hydrogen decrepitation [11]. Smaller-format HDDs take about the same corporeality of time to disassemble every bit larger ones, merely they incorporate significantly less magnet cloth, so the economics of manual recycling grow progressively worse every bit HDD applied science advances.

Coproduction of other recyclable parts and materials, including precious metals, aluminum, and ferrous alloys, helps the economics, but the amounts of these also decline as the drive sizes shrink.

Alternatives to manual disassembly take been adult to reduce the cost of extracting rare world permanent magnets from HDDs. A Hitachi-developed system tumbles HDDs in a drum to loosen and/or remove the screws, and this is reported to increase the disassembly rate into the range of hundreds of units per hour. An automated system developed at Oak Ridge National Lab in collaboration with the Critical Materials Plant identifies the location of the vocalisation-curl associates and slices off the corner of the drive that contains information technology [12]: a unmarried machine system can extract voice-whorl magnets at rates of thousands of units per 60 minutes, with minimal labor costs, although a larger capital investment is required for this type of machinery.

An alternative approach is to shred entire HDDs without separating their components and and so extract the various materials through chemic methods. While this arroyo has lower costs at the front stop of the process, it may involve more than complex chemical separations farther downstream. It likewise suffers from the challenge that shredded magnets accumulate into "hairballs" along with shards of other magnetizable materials and entrapped nonmagnetic ones, and they stick to the ferrous metal components in the shredding machines.

At that place are several technological choices for recycling HDDs, each with its own advantages and disadvantages. The development of a successful commercial arrangement volition depend on optimizing overall value recovery including all of the recoverable materials and disposal of the materials whose value does non justify recovery. The International Electronics Manufacturing Initiative (iNEMI) has convened a project team made upwards of HDD manufacturers, data center operators, and recyclers, along with authorities and university researchers, to identify a organisation that achieves this goal and then build a airplane pilot version of information technology [xiii, 14]. While this is a work in progress, it has several features that lean in favor of its eventual success:

It is a pragmatic approach that focuses on total value recovery from a single class of device, rather than simply recycling critical materials.

The team includes major contributors across the entire life bike of the device and the materials that it contains, and it has admission to leading-edge labs and researchers.

With increasing information density and decreasing HDD sizes, the overall materials demand for HDDs is not growing every bit rapidly every bit in other sectors, and then recycling older, larger-format HDDs may be able to provide a significant fraction of the industry'southward needs for newer, smaller ones. The fraction of the material for new HDDs that can be provided from recycling old ones could be relatively big, and it may, indeed, be possible to make new HDDs out of sometime ones because of their shrinking size.

The HDD market has some unique features that must be considered in designing a value-recovery system, and these tin can bear upon the use of specific technologies. The options for value recovery from failed HDD units are shown schematically in Fig. 7.5.

Fig. 7.5

Fig. seven.5. The chief value-recovery options for failed HDDs. Recycling is just an selection afterwards all of the options higher up it are exhausted, so the higher-value options reduce the book that is available for recycling.

The greatest value that can be obtained from a retired disk is accomplished by repairing information technology and/or reusing it rather than recycling it. Hard disks are retired when they lose access to a predetermined fraction of their capacity, and this can occur considering of software errors that tin can be corrected, assuasive the unit to be returned to service, or it can occur considering of hardware errors that are more than difficult to remediate. A disk that has significant unrecoverable capacity can be reformatted every bit a lower-capacity unit that still has usable life. In that location is a thriving market in recovering and selling retired HDDs for reuse, which provides greater value recovery than any recycling process: it also significantly reduces the volume of HDDs that are available for recycling at any time, which can reduce the fraction of the supply chain that tin be met through recycling.

When an HDD is retired from service, it contains a large corporeality of data that must be protected confronting unauthorized access. This may be achieved by erasing the information if the disk is to exist reused, and this is role of the process of preparing a disk for secondary markets, but the data on disks removed from recycled units can still exist accessed, and the assurance of information destruction is an important consideration. This impacts decisions about how and where HDD recycling is done: information technology generates a preference for systems that physically destroy the disks from the HDDs and for processes that are conducted at the site where the disks are retired, under the supervision of their owner.

There are clearly several technological options to exist considered with regard to recycling HDDs, and the iNEMI projection volition examine them in item. With regard to the nigh prominent disquisitional material in an HDD, the neodymium magnet, the range of options runs from recovering the individual elements in the magnets, to recovering the magnet cloth, to recovering and reusing the magnets themselves, and eventually to recovering and reusing the magnet assemblies. Each of these calls for successively less processing earlier the fabric is reused, and the final iii, in particular, are bonny if the recovered materials, magnets, or magnet assemblies are used for making new HDDs: this avoids the need to adjust the composition that is likely to be required if the magnets or materials become into other applications.

The most attractive option is mayhap the reuse of magnet assemblies, but this requires stability of the HDD design from generation to generation, and it restricts the process to the supply chain of an individual manufacturer unless industry-wide magnet assembly designs can be adopted. Some level of magnet assembly reuse may be possible for voice-gyre motor components, but it is rather less probable that magnet assemblies from spindle motors tin can be successfully harvested, so the economical viability of recycling the different types of magnets in a HDD is likely to be different, and the voice-coil motor is certainly the more bonny target.

Read full chapter

URL:

https://www.sciencedirect.com/scientific discipline/article/pii/B9780128187890000074

Ultrahigh-density ferroelectric information storage using scanning nonlinear dielectric microscopy

Yasuo Cho , in Scanning Nonlinear Dielectric Microscopy, 2020

4.2.5 An hard disk scanning nonlinear dielectric microscopy information storage unit for loftier-density ferroelectric recording

As described in a higher place, an HDD test device allowed writing and reading at rates of 20 and two   Mbps, respectively [51]. Still, this system did not allow recording densities on the 1   Tbit/in.2 scale. Withal, subsequent piece of work using a similar organisation provided this level of storage density. In these trials, a CLT single crystal was used as the recording medium [52], because this medium permits the generation of the nanodomain dots required for high-density data storage [35,52]. A very thin ferroelectric layer is believed to be essential when attempting to write small inverted domains while increasing the recording density. This is required because the electric field is concentrated immediately below the tip when using a thinner flick [53]. In the trial discussed hither, ferroelectric recording media were produced via the mechanical polishing of wafers made of single crystals, obtaining thicknesses of well-nigh ane.8   μm. The desired exact thicknesses were obtained via dry out etching using electron cyclotron resonance. Both the thin polished film and the substrate were composed of the same material to make sure that these components underwent the same level of thermal expansion. The resulting recording medium had a surface surface area of 15×15   mmtwo and a thickness of 46±xv   nm and polarization was applied to the unabridged medium in the downward direction.

An SNDM prototype showing a 7×seven assortment of inverted domain dots on the CLT medium is provided in Fig. iv.15A. The voltage pulses applied to form these dots had durations ranging from 10 to 100   ns and pulse amplitudes between 4 and 10   V. Following pulse application, the medium was subjected to a 1   V dc starting time voltage so every bit to stabilize the dots. The dot size exhibited a tendency to decrease with decreases in both the pulse duration and amplitude. A magnified image of the smallest dot (indicated by the blackness square in Fig. 4.15A) is shown in Fig. four.15B, while Fig. 4.15C presents the line profile acquired from this same dot. The dot was generated using a xl   ns, v   V voltage pulse in conjunction with a i   Five offset voltage, and had a bore of 12   nm. This size was sufficiently small-scale such that an array having a recording density in excess of one   Tbit/in.2 was possible.

Figure 4.15. (A) SNDM image of vii×7 inverted domain dot array on single-crystal CLT medium with a thickness of 46   nm, (B) close-up of smallest dot in this array, and (C) line profile of smallest dot.

A demonstration of high-density recording was carried out using this CLT medium together with the HDD-type test instrumentation, and Fig. 4.16 provides the SNDM paradigm of the resulting bits. These bits were written by applying 100   ns, 8.8   V pulses over a fourth dimension span of 80   μs, together with a one   5 showtime voltage. This image demonstrates the precise generation of well separated domain dots, providing a repeating pattern of "one" and "0" bits at thirteen.7   nm intervals, equal to a three.4   Tbit/in.2 memory density.

Figure four.16. SNDM prototype of $.25 written on single-crystal CLT medium with a scrap spacing of 13.seven   nm (corresponding to a memory density of three.44   Tbit/in.2).

A R/Due west test was also performed by tracing the array after writing to produce a readout signal. An SNDM image of the written bits resulting from this trial, made past applying 100   ns, eleven.viii   V pulses together with a 1.five   V offset and a pulse time span of 167   μs, is shown in Fig. 4.17A. This trial verified the correct writing of the bit array, with 25.8   nm spacing, equal to a 0.97   Tbit/in.ii memory density. A representative real-time read point acquired during this test is presented in Fig. 4.17B. Hither, the "1" and "0" signals repeat at 167 μs intervals, like to the time span applied during the scrap writing procedure. This result indicates that the bits were accurately read. Following a fast Fourier transform (see Fig. 4.17C), a signal-to-racket ratio of 3.9   dB was determined.

Effigy 4.17. Read signal respective to bits written on a unmarried-crystal CLT medium at a transfer charge per unit of 12   kbps and a spacing of 25.eight   nm (respective to a retention density of 0.97   Tbit/in.2): (A) SNDM image of dots, (B) existent-time waveform, and (C) Fourier spectrum.

The writing of actual information was also assessed using the same experimental system. This sit-in consisted of recording the letters S, N, D and M every bit an ASCII bit array on the CLT medium. This was performed using 300   ns, eleven.iii   V pulses with 25   nm chip spacing, giving the results in Fig. 4.18. The SNDM image of the resulting scrap information confirms the separate recording of each bit, and shows that the original chip array was readable using suitable bespeak processing.

Figure iv.eighteen. Bodily information bit array written on CLT medium: (A) written fleck data assortment, (B) SNDM prototype, and (C) contour along line in (B).

Read full affiliate

URL:

https://www.sciencedirect.com/science/article/pii/B9780128172469000042