Current Projects

Project: MOE Tier 2 ARC 1/19 RSR

$ 831,276.00

Abstract: The recent trend of 3D ultra-large-scale-integration where various systems are stacked on single IC and are interconnected with copper wires suffer from unwanted resistive and inductive losses. To overcome this issue, the implementation of wireless interconnects on 3D IC is much required solution. Nanoscale and broad-range frequency tunable Spin Hall Nano-Oscillators (SHNO) are most suitable candidates for these wireless interconnects for chip to chip wireless communication which also offers the strong compatibility with standard CMOS processing and technology. However, low output power of SHNOs still remain an open challenge. The maximum output power of SHNO is in mW range and needs to be improved to sub-mW range for chip to chip wireless communications. The key target of the project is to improve output power in the array of SHNOs by phase lock scheme using spin Hall effect and Inverse spin Hall effect in high spin-Hall materials like Pt, W and Ta. In this project, we will focus on (i) Design, fabrication and optimization of PMA based Magnetic Tunnel Junction (MTJ) with large MR ratio at low temperature < 300 °C, (ii) Charge to spin conversion in spin-Hall material (Pt/Ta/W) thin film and its integration in single SHNO, and (iii) Realization of high output power and high quality factor in PMA based phase-locked SHNOs array device.

Duration: 5 Years

Role: Principal Investigator

$5.7Million

Neuromorphic computing (NC) mimics the functions of the brain using a network of synthetic neurons interconnected through synaptic devices. NC is gaining attention worldwide due to its potential for AI and big data analysis at low power. The research objectives of this programme will be aligned towards developing spin-orbit coupling based NC hardware elements, such as synthetic neurons and synaptic elements and integrating them into NC architecture for technology translation. In this proposed programme, our research and development will focus on the physics, materials science and device aspects of scalable and enduring synthetic neurons and synaptic elements based on spin-orbit coupling (SOC) and demonstration and technology translation of low-power neuromorphic computing.

Project: MOE Tier 2 ARC 1/17 RSR MOE2017-T2-2-129

$697848.00

Abstract:Pure spin current offers advantages of reduced power dissipation, absence of stray Oersted fields, and decoupling of spin and charge noise; making them highly suitable for magnetization direction manipulation (write operation) in spin transfer torque – magnetic random access memory (STT-MRAM) technology. Current charge to spin conversion methods have limitations. For example (i) method using exchanged coupled magnetic layer in magnetic tunnel junction suffers from very high current density (106 A/cm2) requirement, (ii) Spin Hall Effect (SHE) method in heavy metal shows limited (<33%) conversion efficiency and (iii) the orthogonal locking of electrons’ momentum and spin in topological insulators suffers from reproducibility issues. To explore innovative scheme based on Rashba Effect along with development of new material hetero interfaces, which can generate very high spin current with nearly 100% charge to spin efficiency by achieving large spin orbit interaction, needed for reliable low power magnetization switching. Large spin orbit interaction and correspondingly higher charge to spin interconversion efficiency can be achieved by appropriately designing the Rashba hetero interface along with additional control using external electric field. We will focus on (i) Rashba interface design using metal decorated graphene on epitaxial YIG/GGG system to achieve tunable spin-orbit coupling using external electric field, (ii) charge to spin interconversion efficiency measurement – using spin torque ferromagnetic resonance (ST-FMR) technique and spin-pumping through inverse Rashba Edelstein effect (IREE), and (iii) realization of magnetization Switching in magnetic layer of the fabricated device.