3-D magnetic recording (3DMR) technologies, such as 3-D heat-assisted magnetic recording (3-D HAMR) and 3-D bit-patterned magnetic recording (3-D BPMR), enhance areal density by stacking multiple recording layers and utilizing heat-assisted writing to encode multi-bit data through vertically combined magnetization states. Current 3DMR systems primarily employ dual-layer media [also referred to as dual-layer magnetic recording (DLMR)], where the readback signals from a single head consist of superimposed responses from both the top and bottom layers, leading to significant 3-D interference, including inter-symbol interference (ISI), inter-track interference (ITI), and inter-layer interference (ILI). To improve bottom-layer detection reliability in 3DMR, this article proposes a scaling-based signal extraction method that refines the bottom-layer signal using detected top-layer data. This approach enables efficient 1-D detection while significantly reducing the bottom-layer bit error rate (BER). Compared to the previously proposed dual-layer PRML detection, the method maintains the same top-layer BER while achieving an 80% reduction in the bottom-layer BER at the areal density of 3.3 Tb/in(2 ). Moreover, it demonstrates superior bottom-layer detection performance across areal densities ranging from 1 to 4.5 Tb/in(2 )per layer (2-9 Tb/in(2 )for dual-layer systems).

Three-dimensional magnetic recording (3DMR) technologies, such as three-dimensional heat-assisted magnetic recording (3D HAMR) and three-dimensional bit-patterned magnetic recording (3D BPMR), enhance areal density by stacking multiple recording layers and leveraging heat-assisted writing to encode multi-bit data through vertically combined magnetization states. Current 3DMR systems primarily employ dual-layer media (also termed dual-layer magnetic recording), where the readback signal from a single head comprises superimposed responses from a top layer and a bottom layer, introducing severe three-dimensional interference—including inter-symbol interference (ISI), inter-track interference (ITI), and inter-layer interference (ILI). To improve bottom-layer detection reliability in 3DMR, this paper proposes a scaling-based signal extraction method that refines the bottom-layer signal using detected top-layer data. This approach enables efficient one-dimensional detection while significantly reducing the bottom-layer bit error rate (BER). Compared to the previously proposed dual-layer PRML detection, the method maintains the same top-layer BER while achieving a 80% reduction in the bottom-layer BER under the areal density of 3.3 Tbpsi. Moreover, it demonstrates superior bottom-layer detection performance across areal densities ranging from 1 to 4.5 Tbpsi per layer (2–9 Tbpsi for dual-layer systems).

Because so much information is currently being shared online, there has been a sharp rise in the need for data storage devices over the past ten years. The main storage option is the hard disk drive (HDD), which is less expensive than some other types of data storage. Physical constraints, such as the superparamagnetic limit, are difficult to overcome using existing HDD technology. Consequently, bit-patterned magnetic recording (BPMR) has emerged as a potential solution, offering higher areal densities whilst maintaining thermal stability. Nevertheless, BPMR poses new challenges, such as inter-symbol interference and inter-track interference. Consequently, a number of approaches, such as staggered island layouts and array-reader magnetic recording, have been proposed to overcome these issues. However, this article proposes a three-reader one-track detection method to enhance data retrieval in a staggered BPMR system. Leveraging three-track reading for one-track detection, we obtain three readback signals that function as mutual data sequences. This substantially enhances the detection process in one-dimensional partial-response maximum-likelihood channels. Next, using these mutual data sequences, four novel techniques are presented to enhance bit-error rate (BER) performance and detection accuracy: hard-information flipping, maximum soft-information finding, bit-summation detection, and multilayer perceptron (MLP). This study shows that these proposed techniques can provide better BER performance compared with conventional methods and that the MLP is the most effective technique in enhancing system performance.

Double-layer magnetic recording (DLMR) development aims to increase areal density (AD) by stacking multiple recording layers on the same platter. This has the potential to revolutionize magnetic recording. On the other hand, bit-patterned magnetic recording (BPMR) uses nanoscale bit islands to enhance storage density, reduce transition noise, and maintain thermal stability. This study focuses on combining BPMR and DLMR to further increase AD. Unfortunately, this approach encounters difficulties in separating data from different layers within a single readback signal, which results in poor performance. Retrieving data from the lower layer is additionally complicated by its distance from the reader, leading to degraded bit-error-rate (BER) performance. To address this, we have developed mutual soft-information improvement techniques (MIITs) such as the maximum soft-information (MSI) and sum soft-information (SSI) schemes. These techniques enhance the soft-information output obtained from a soft-output Viterbi algorithm (SOVA) detector. Additionally, we propose three iterative systems that exchange soft information between SOVA and low-density parity-check (LDPC) codes. Our demonstrations show improved BER performance for both upper and lower layers, compared to conventional systems with single-layer media at the same AD of 4 Tb/in(2) .

This article discusses a micromagnetic analysis of spin-torque oscillators (STOs) with two field generation layers (FGLs) (dual FGL STO) for use in microwave-assisted magnetic recording (MAMR). Dual FGL STOs have superior characteristics with regard to stable oscillation, especially when integrated into the write head gap, due to antiferromagnetic coupling between the two FGLs. The structure and materials, however, can be complicated. Our new dual FGL STO has a simpler structure, composed of FGL/soft magnetic material (SMM)/FGL. In addition, because the oscillation frequency changes with the current density injected into the STO, it is suitable for dual-layer recording. Recording simulations showed that the STO had a high signal-to-noise ratio (SNR)/track width, compared with a dual FGL STO without the SMM layer.

To increase the areal recording density of magnetic recording, we have investigated the effect of the thicknesses of the individual layers in dual-layer bit-patterned magnetic recording (BPMR) systems. The recording media consisted of two, discrete recording layers, separated by a non-magnetic spacer layer. The bottom recording layer thickness and head-medium spacing (HMS) were fixed, while the top recording layer thickness was varied to adjust the spacing between the reader and the bottom layer. By varying the layer thicknesses, the readback signal strengths from the top and bottom layers can be adjusted and balanced. To determine the bit-errorrate improvement needed to reach a target areal density of 4.0 Tbits/in(2), i.e. 2.0 Tbits/in(2) per layer, the readback signal was processed using an iterative method. The results showed that the proposed design could outperform a single-layer BPMR system.

The effect of Curie temperature distributions and average grain size on the user areal density (UAD) of heat-assisted magnetic recording media was calculated for target areal densities of 2, 3 and 4 Tbit/in 2 . A grain switching probability model was used to write tracks and the bit error rate evaluated, allowing the UAD to be calculated. A 10% Curie temperature distribution reduced the UAD by 0.5 - 0.8 Tbit/in 2 , with larger reductions for media with smaller grains. The effect of the reader on the UAD was also evaluated and found to be less significant for media with Tc distributions.

Multi-level magnetic recording is a new concept for increasing the data storage capacity of hard disk drives. However, its implementation has been limited by a lack of suitable media capable of storing information at multiple levels. Herein, we overcome this problem by developing dual FePt-C nanogranular films separated by a Ru-C breaking layer with a cubic crystal structure. The FePt grains in the bottom and top layers of the developed media exhibited different effective magnetocrystalline anisotropies and Curie temperatures. The former is realized by different degrees of ordering in the L10-FePt grains, whereas the latter was attributed to the diffusion of Ru, thereby enabling separate magnetic recordings at each layer under different magnetic fields and temperatures. Furthermore, the magnetic measurements and heat-assisted magnetic recording simulations showed that these media enabled 3-level recording and could potentially be extended to 4-level recording, as the up-down and down-up states exhibited non-zero magnetization.

Three-dimensional (3D) magnetic recording with multiple recording layers has garnered considerable attention as a next-generation magnetic recording method that uses dual conventional recording layers and a magnetoresistive (MR) head for reading. In this work, to examine signal processing methods for the 3D magnetic read/write (R/W) channel using heat-assisted magnetic recording (HAMR), we have constructed a HAMR R/W channel using a granular media model. Our 3D HAMR R/W channel model, specialized for signal processing development, can account for the Curie temperatures of the recording layers and the temperature distribution of the heat source.

Microwave assisted magnetic recording was simulated using write heads with dual FGL STOs. Exchange coupled composite recording media were optimised for use with dual FGL STOs. The optimised recording medium showed a higher figure of merit (SNR / written track width) for dual FGL STOs than single FGL STOs due to narrower written tracks. It is supposed that flux closure at the edges of the two FGLs was the reason for this.

We investigate several architectures for a convolutional neural network (CNN) media noise estimator (MNE) that sends media noise estimates to a three-track Bahl-Cocke-Jelenik-Raviv (BCJR) TDMR detector. While novel CNN architectures have previously been proposed for image classification problems, here the CNN estimates media noise based on inputs from the BCJR detector and from a linear partial response equalizer. The CNN architecture used in our previous work on CNN-based TDMR detection is used as a comparison baseline. Experiments show that an optimized CNN MNE with three residual path connections achieves up to a 3.2% reduction in mean squared estimation error and a 71.9% reduction in computational complexity compared to the baseline.

L10-FePt grains with structural inhomogeneity in a granular film was investigated, and their blocking phenomenon was simulated. It was confirmed that: 1) structure analysis reveals that both lattice distorted disordered portion at the bottom region near the interface with MgO underlayer and ordered portion at the upper region exist inside one FePt grain in a granular film and 2) simulation of blocking phenomenon for FePt grains with structural inhomogeneity which consists of a hard (order)/soft (disorder) bilayer shows that switching field ( ${H}_{\mathbf {sw}}$ ) increases steeply near the Curie temperature ( ${T}_{C}$ ) of the disorder layer with decreasing temperature. To generalize blocking phenomenon of the hard/soft bilayer structure, it was found that matching the Curie temperatures of the hard and soft layers leads to enhancement of ${H}_{\mathbf {sw}}$ near the ${T}_{C}$ , which means enhancement of robustness against thermal field in random direction. These findings are useful for the design of cap-layer for heat-assisted magnetic recording (HAMR) media to realize the high spatial resolution.

Arrays of spin-torque oscillators (STOs) can be used in artificial neural networks. The STOs in such arrays can be synchronized via magnetostatic interactions. In this work, we show that variations in the STO diameter can be used to control the flow of information. Unidirectional information flows can be enabled by a consistent change in diameter along a row of STOs. A switch to direct the information flow can also be constructed using STOs with different diameters.

The structural inhomogeneity of an L1 <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">0</inf> –FePt grain in a granular film was investigated, and its blocking phenomenon was simulated. It was confirmed that 1) structural analysis reveals that both lattice distorted disordered portion at near the interface with MgO underlayer and ordered potion at the upper region exist inside one FePt grain in a granular film, 2) simulation of blocking phenomenon for FePt grains with structural inhomogeneity, hard (order)/soft (disorder) bilayer grains, shows that switching field (H <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">sw</inf> ) increases steeply near T <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">C</inf> of the disorder layer with decreasing temperature. To generalize blocking phenomenon of the hard/soft bilayer structure, it was found that matching the Curie temperatures of the hard and soft layers leads to high H <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">sw</inf> near the T <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">C</inf> , which means high robustness against thermal field in random direction. These are useful findings for the design of cap-layer media for HAMR to realize the high spatial resolution.

Hysteresis loops of single grains and recording media were calculated under the influence of high-frequency (HF) magnetic fields, similar to those produced by a spin-torque oscillator (STO) in microwave-assisted magnetic recording (MAMR). The effect of the HF field ellipticity and the strength of magnetostatic interactions between grains were investigated. Elliptical HF fields with a major:minor axis ratio of 6:1 were almost as effective as circular HF fields in reducing the media coercivity. For media with a small grain pitch, i.e., strong magnetostatic interactions between grains, the HF field was able to reduce the coercivity by much more than when the grains were far apart, and magnetostatic interactions were weak.

Magnetic nanowires can be used to store information. There are various ways to write magnetic domains in nanowires, usually involving the use of an Oersted field. In this work the use of a spin torque oscillator to write domains in magnetic nanowires is investigated using micromagnetic modelling. It is shown that domains can be written using a lower magnetic field than if an Oersted field were used. Writing of multiple domains in a nanowire is also demonstrated.

Transition curvature in heat-assisted magnetic recording can limit the linear recording density. In this work we propose a write head design which can reduce transition curvature. The split-pole write head incorporates a nano coil which generates a field in the opposite direction to the main pole excited by the head coil winding, allowing transition curvature to be controlled.

Magnetic recording in current hard disk drives is basically two dimensional, with information stored in a thin recording structure covering the surface of the disk. This review examines the prospects for three dimensional magnetic recording using multiple recording structures, and the technologies that may help its realisation. In particular, energy-assisted recording techniques are discussed as these appear to have the greatest potential.

Dual field generation layer (FGL) spin torque oscillators (STO) were investigated for use in microwave-assisted magnetic recording. The STOs were integrated into the gaps of write heads for stable oscillation. According to preliminary calculations, a FGL thickness of 5 nm was suitable, while the volume of 7.5 nm and 10 nm thick FGLs was too large and they did not oscillate stably. The most dominant factor influencing the FGL rotation was found to be the strength of the antiferromagnetic coupling between the FGLs. Other parameters were also varied, e.g., the exchange coupling between the negative spin injection layer (nSIL) and the trailing shield (TS), however, no significant effect was found. Unlike single FGL STOs, the oscillation frequency changed drastically, from 38 GHz to 73 GHz, on varying the injected current density to the STO. Recording simulations showed that the signal-to-noise ratio was maximized for an STO oscillation frequencies between 34 GHz and 50 GHz, depending on the recording media used in the calculations. Therefore, the dual FGL STO may be suitable for a wide range of recording media. Alternatively, the dual FGL STO may be used for dual-layer recording, where different resonance frequencies are needed for each of the layers, obviating the need for a second STO.

This article considers a turbo-detection system that includes a convolutional neural network (CNN)-based equalizer, a Bahl-Cocke-Jelinek-Raviv (BCJR) trellis detector, a CNN-based media noise predictor (MNP), and a low-density parity-check (LDPC) channel decoder for two-dimensional magnetic recording (TDMR) in the presence of track misregistration (TMR). The input readings are passed to a 2-D partial response (PR) equalizer, which is either linear or CNN-based. The equalized waveforms are inputs to a 2-D BCJR detector, which generates log-likelihood-ratio (LLR) outputs. The CNN MNP is provided with BCJR LLRs to estimate signal-dependent media noise samples and feed them back to the BCJR. A second pass through the BCJR produces LLRs, which are decoded by an LDPC decoder; achieved areal density (AD) is computed from the LDPC code rate. Spatially varying read- and write-TMR models are developed. We investigate the performance of the proposed system on simulated TDMR readback waveforms generated by grain-switching probabilistic (GSP) simulations. We have two types of GSP datasets. Dataset #1 includes two 10 nm bit length (BL) datasets with 18 and 24 nm track pitch (TP). Dataset #2 has 11 nm BL and 15 nm TP. The comparison baseline is a 1-D BCJR detector with pattern-dependent noise prediction (PDNP) and soft intertrack interference (ITI) subtraction, referred to as 1-D PDNP with LLR exchange. The write-TMR and read-TMR are modeled as cross-track-independent downtrack-correlated random processes. In the presence of joint write- and read-TMR, the proposed turbo-detection system achieves 8.34% and 0.70% AD gain over 1-D PDNP with LLR exchange for TP 18 and 24 nm dataset #1, respectively, and is more robust to TMR compared to the baseline.