Periodic Reporting for period 2 - MAGNEPIC (Magnetic Insulators: An Enabling Platform for Innovative Spintronic Concepts)
Período documentado: 2022-08-01 hasta 2024-01-31
Our society increasingly benefits from digital technologies thanks to the ever-improving performance and storage capacity of computers. Developments of, e.g. personal electronics and artificial intelligence are in full swing and directly impact our daily lives. However, the advances in digital technologies substantially increase global energy consumption, posing sustainability issues. We thus need long-term mitigation plans to alleviate these technologies' precariously growing energy demand. A practical approach is developing energy-efficient data storage and processing platforms for more sustainable computing. Spintronics –use of electron spin as an active variable for data storage and processing- offers innovative solutions in this quest. In spintronics, the main challenge is to exploit spin-dependent phenomena and develop magnetic systems where the magnetization state (e.g. up/down representing 0/1 in binary code) is effectively controlled and detected by electrical means in nanoscale devices.
MAGNEPIC aims to develop disruptive spintronic device concepts using magnetic insulators as an enabling material platform.
For over three decades, magnetic conductors have been predominantly considered in spintronics. This is due to their historical significance, easy-to-measure, well-understood magnetic and transport properties, and convenience of deposition and fabrication. However, they represent only a small group within those possessing magnetic ordering and offer little room for engineering their properties in a device context. Magnetic insulators, however, constitute a more diverse and abundant material family, offering broadly tunable structural and magnetic properties not easily achievable in their conducting counterparts. However, due to the difficulties associated with addressing their magnetization electrically, they have been greatly overlooked until recently.
Provided the state of affairs, in MAGNEPIC, we will
- Develop a set of magnetic insulators, namely rare-earth iron garnets (REIGs), with tunable magnetic properties suitable for spintronic device implementation.
- Explore physical effects that enable us to manipulate and detect REIGs' magnetization state by electrical stimuli (current injection, electric field gating, piezoelectricity, etc.).
- Design, fabricate, and characterize nanometer and micrometer-scale devices based on magnetic insulators for spin-based memory, logic, and interconnect applications.
At the end of the project, we aim to develop a set of novel and groundbreaking device concepts where data can be stored, manipulated, and processed by all-electrical means with unprecedented efficacy and speed. MAGNEPIC will thus establish the ground for next-generation computing platforms based on magnetic insulators.
MAGNEPIC aims to develop disruptive spintronic device concepts using magnetic insulators as an enabling material platform.
For over three decades, magnetic conductors have been predominantly considered in spintronics. This is due to their historical significance, easy-to-measure, well-understood magnetic and transport properties, and convenience of deposition and fabrication. However, they represent only a small group within those possessing magnetic ordering and offer little room for engineering their properties in a device context. Magnetic insulators, however, constitute a more diverse and abundant material family, offering broadly tunable structural and magnetic properties not easily achievable in their conducting counterparts. However, due to the difficulties associated with addressing their magnetization electrically, they have been greatly overlooked until recently.
Provided the state of affairs, in MAGNEPIC, we will
- Develop a set of magnetic insulators, namely rare-earth iron garnets (REIGs), with tunable magnetic properties suitable for spintronic device implementation.
- Explore physical effects that enable us to manipulate and detect REIGs' magnetization state by electrical stimuli (current injection, electric field gating, piezoelectricity, etc.).
- Design, fabricate, and characterize nanometer and micrometer-scale devices based on magnetic insulators for spin-based memory, logic, and interconnect applications.
At the end of the project, we aim to develop a set of novel and groundbreaking device concepts where data can be stored, manipulated, and processed by all-electrical means with unprecedented efficacy and speed. MAGNEPIC will thus establish the ground for next-generation computing platforms based on magnetic insulators.
The departing idea of MAGNEPIC is to use rare-earth iron garnets as a tunable magnetic material platform and create novel spintronic device concepts based on these materials. To this end, we created two new laboratories, one dedicated to thin film fabrication based on the magnetron sputtering technique and the other for high-precision optical and electrical device characterization. We optimized the growth of TbIG, TmIG, and YIG ultrathin films. These materials possess high perpendicular anisotropy, low perpendicular anisotropy, and in-plane anisotropy, respectively, which are beneficial to test different device ideas. Out of these three materials, we specifically focused on TbIG. We characterized its structural, chemical, magnetic, and interfacial spin transport properties in a broad range of temperature and preparation conditions, which we reported in a paper. We can now grow TbIG down to 3 nm, which is hugely beneficial to test some device ideas we propose in MAGNEPIC, where interfacial effects are dominant.
The electrical characterization setup is our primary tool to test the spintronic devices. In this setup, we have characterized the current-induced spin-orbit torques in an archetypical spintronic system, Pt/YIG. To quantify the torques more precisely, we improved the existing methodology by performing the measurements in an unconventional geometry and combining our findings with numerical simulations. Aside from the precise determination of the torque values, to our surprise, we found that the damping-like torque loses efficiency at high current injection due to the rise in device temperature. We reported the improved methodology and the above findings in a paper. In a parallel effort, we have realized a novel spin valve device where one magnetic layer is TbIG instead of a typical conducting magnet. Here, we used TbIG|Cu|TbCo trilayers, where TbCo serves as the reference layer, and Cu is a nonmagnetic spacer. For a fixed polarization of TbCo, the resistance of Cu|TbCo displayed two distinct levels for the “up” and “down” polarization of TbIG, respectively. This leads to a simple and powerful two-terminal device concept where binary data could be stored and electrically accessed in a magnetic insulator. These exciting findings represent a significant milestone and breakthrough for the MAGNEPIC project and the entire spintronics research community.
We developed two optical characterization setups. The first one is a wide-field magneto-optic Kerr effect microscopy. This setup uses an LED light shone onto the substrate surface, and depending on the magnetization state (up/down) of the film, the reflected light changes its polarization. The analysis of the reflected light provides magnetic imaging of the surface. This setup is compelling for measuring nonuniform magnetic textures (domain walls, skyrmions) with a lateral resolution of <1 micrometer. In this setup, we have been performing a comprehensive study to understand how the domain walls respond to the current injection in a metal/TbIG bilayer where we use different types of metals that can produce spin-orbit torques. The second optical setup is still in progress. Here, we use a laser instead of an LED, and the final goal is to focus the beam down to 1 micrometer spot and get information on the local magnetization state of the films. This setup acquires data at a much faster rate, enabling time-resolved dynamical characterization of, e.g. domain walls. However, the focusing unit has not yet been implemented, and the spot size is currently in the sub-mm range, which is still helpful for inspecting the basic properties of continuous films.
The electrical characterization setup is our primary tool to test the spintronic devices. In this setup, we have characterized the current-induced spin-orbit torques in an archetypical spintronic system, Pt/YIG. To quantify the torques more precisely, we improved the existing methodology by performing the measurements in an unconventional geometry and combining our findings with numerical simulations. Aside from the precise determination of the torque values, to our surprise, we found that the damping-like torque loses efficiency at high current injection due to the rise in device temperature. We reported the improved methodology and the above findings in a paper. In a parallel effort, we have realized a novel spin valve device where one magnetic layer is TbIG instead of a typical conducting magnet. Here, we used TbIG|Cu|TbCo trilayers, where TbCo serves as the reference layer, and Cu is a nonmagnetic spacer. For a fixed polarization of TbCo, the resistance of Cu|TbCo displayed two distinct levels for the “up” and “down” polarization of TbIG, respectively. This leads to a simple and powerful two-terminal device concept where binary data could be stored and electrically accessed in a magnetic insulator. These exciting findings represent a significant milestone and breakthrough for the MAGNEPIC project and the entire spintronics research community.
We developed two optical characterization setups. The first one is a wide-field magneto-optic Kerr effect microscopy. This setup uses an LED light shone onto the substrate surface, and depending on the magnetization state (up/down) of the film, the reflected light changes its polarization. The analysis of the reflected light provides magnetic imaging of the surface. This setup is compelling for measuring nonuniform magnetic textures (domain walls, skyrmions) with a lateral resolution of <1 micrometer. In this setup, we have been performing a comprehensive study to understand how the domain walls respond to the current injection in a metal/TbIG bilayer where we use different types of metals that can produce spin-orbit torques. The second optical setup is still in progress. Here, we use a laser instead of an LED, and the final goal is to focus the beam down to 1 micrometer spot and get information on the local magnetization state of the films. This setup acquires data at a much faster rate, enabling time-resolved dynamical characterization of, e.g. domain walls. However, the focusing unit has not yet been implemented, and the spot size is currently in the sub-mm range, which is still helpful for inspecting the basic properties of continuous films.
Having developed the critical materials and setups, the future of MAGNEPIC is bright. The discovery of a novel magnetoresistive phenomenon, enabling us to detect the magnetization reversal of the magnetic insulator even though no current is flowing through, is significantly beyond the state-of-the-art. We are currently working on, and/or planning to tackle soon, challenges associated with the efficient control of magnetization by all-electrical means. These include ultrafast current-induced magnetization switching and domain wall motion, low-power magnetization control using ionic gating or piezoelectricity, and a novel current-driven incoherent magnon generation concept for long-distance spin-information transport. These findings are expected to give rise to a new set of device concepts for memory, logic, and interconnect circuits, significantly beyond the state-of-the-art with implications for the post-CMOS computing platforms.