Periodic Reporting for period 1 - SPINHALL (Computing with mutually synchronized topological insulator based spin Hall nano-oscillators)
Reporting period: 2020-09-01 to 2022-08-31
Spin Hall nano-oscillators (SHNOs) are revolutionary nanoscopic, ultra-tunable, and ultra-rapidly modulated microwave oscillators and have direct compatibility with industry-standard CMOS technology. While their first target applications are ultra-wide frequency tunable microwave signal generators/detectors for cell phones, wireless networks, vehicle radar, and ultrafast spectral analysis applications, the rapidly improved understanding of their non-linear properties and demonstration of mutual synchronization of large numbers of SHNOs make them promising candidates for large-scale oscillator networks for unconventional computing.
SPINHALL uses recent breakthroughs in spin Hall devices and materials to improve the performance and applicability of SHNOs. The primary goal is to use the latest breakthrough in compensated ferrimagnets such as GdFeCo and GdCo with their high ferromagnetic resonance frequencies and high spin-orbit torque efficiency to improve the SHNOs operating frequency by an order of magnitude and the power consumption by orders of magnitude. SPINHALL will also use the magnetic Weyl semimetal Co2MnGa to have better electric field control of SHNOs.
SPINHALL uses recent breakthroughs in spin Hall devices and materials to improve the performance and applicability of SHNOs. The primary goal is to use the latest breakthrough in compensated ferrimagnets such as GdFeCo and GdCo with their high ferromagnetic resonance frequencies and high spin-orbit torque efficiency to improve the SHNOs operating frequency by an order of magnitude and the power consumption by orders of magnitude. SPINHALL will also use the magnetic Weyl semimetal Co2MnGa to have better electric field control of SHNOs.
The work was divided up into four work packages.
WP1
The project started by growing thin films of MnGa at the host. After difficulties in growing these films, I decided to collaborate with Prof. Shigemi Mizukami (SM) on these films. The SM group provided me with epitaxial MnGa and bilayer (CoGa/MnGa) thin films. Unfortunately, we could not see any meaningful BLS signal from any high-frequency resonances in these samples. At the same time, our collaborators at the University of Minnesota could not provide us with BiSe topological insulator films due to COVID-19, as their labs were closed for almost 2 years.
I, therefore, devised a new strategy to achieve the project goals. Compensated ferrimagnets such as GdFeCo and GdCo were identified as suitable alternatives as they show high resonance frequencies and high spin-orbit torque efficiency. To use these materials in spintronic devices, it is important to grow them in the ultrathin range (< 10 nm). Gdx(FeCo)1-x [GdFeCo hereafter] thin films in the thickness range of 2-20 nm were grown on industry-compatible high resistance Si(100) [HR-Si(100)] substrates and studied using broadband ferromagnetic resonance (FMR) measurements at room temperature. By tuning the stoichiometry, a nearly compensated behavior was observed in 2 nm GdFeCo for the first time, with an effective magnetization of about 0.02 T and a low effective Gilbert damping constant of about 0.0078 comparable to the lowest values reported so far in 30 nm films. HR-Si(100)/Pt(5nm)/GdFeCo(2-10nm)/SiO2(4nm) stacks were developed to fabricate devices for spin-torque FMR (STFMR) measurements, hall measurements, and SHNO auto-oscillation measurements. The effective spin-orbit torque efficiency of Pt/GdFeCo layers was estimated using STFMR measurements. To understand the spin-orbit torque contribution from GdFeCo, GdFeCo/Cu/NiFe/SiO2-based stacks were developed and studied using STFMR. Analyzing the STFMR data, large values of anti-damping-like and field-like torques were obtained for GdFeCo.
HR-Si(100)/Pt(5nm)/GdFeCo(2-10nm)/SiO2(4nm) stacks were prepared to explore further the compensated ferrimagnet-based SHNOs, and steady-state magnetization auto-oscillations were obtained with SHNO constriction widths in the range of 50-150 nm.
WP2
For the electric field (E-field) control of SHNOs, we identified Co2MnGa, a magnetic Weyl semimetal. Epitaxial thin films of Co2MnGa were grown (by collaborators at NIMS, Japan) and studied using FMR, STFMR, and 2nd harmonic Hall effect measurements. A very high value of the anomalous Hall conductivity of about 1350 S/cm was obtained for a 30 nm thick Co2MnGa film, which is comparable to the bulk case. The large anomalous Hall conductivity indirectly confirms the topological Weyl semimetal phase of the Co2MnGa film. Strong anisotropy in the Gilbert damping was also observed for the Co2MnGa 10-30 nm films, and the lowest values of about 0.009 and 0.013 were obtained for 20 and 30 nm films, respectively. The spin-orbit torque efficiency of the films was characterized by analyzing the 2nd harmonic Hall resistance and STFMR measurements and a damping-like torque efficiency in the range of 0.08-0.12 was obtained. To study the E-field effect on spin-orbit torque efficiency, STFMR microstrips, hall bars, and SHNOs were prepared with different gate geometries. Investigation of these devices is still ongoing and will be finished in collaboration with the host group.
WP3
Brillouin Light Scattering microscopy was performed to investigate the high-frequency ferromagnetic/exchange modes in Pt(5nm)/GdFeCo(2-10nm) based SHNOs. Auto-oscillations on the ferromagnetic mode (~8 GHz at 4-6 kOe) were observed in GdFeCo based SHNOs; auto-oscillations on any exchange mode were not observed. Very interestingly, different Gd compositions resulted in different signs of the threshold current for auto-oscillations, which indicates different signs of the effective spin-orbit torque as a function of the composition. The BLS measurements on Pt(5nm)/GdCo(8nm) and Co2MnGa devices are in progress and will be finished in collaboration with the host group.
WP4
The project started with scientific training on the tools in the applied spintronics group and MC2 cleanroom at Chalmers. At first, the fellow got ALL the licenses to operate the tools such as AJA DC/RF magnetron sputtering, NanOsc Phase FMR, Alternating Gradient Magnetometer (AGM), Raith Electron Beam Lithography, Optical Lithography, Dry Ion Beam Etching, Scanning Electron Microscopy, steady state magnetization auto-oscillation measurement setup, Brillouin Light Scattering (BLS), etc.
The project results are already presented at 4 international conferences, including 3 contributory oral talks and 1 invited talk at the Joint European Magnetic Symposium 2022 (JEMS2022). There is another invited talk scheduled for November 2022 at the 67th annual Magnetism and Magnetic Material conference in Minneapolis, USA.
WP1
The project started by growing thin films of MnGa at the host. After difficulties in growing these films, I decided to collaborate with Prof. Shigemi Mizukami (SM) on these films. The SM group provided me with epitaxial MnGa and bilayer (CoGa/MnGa) thin films. Unfortunately, we could not see any meaningful BLS signal from any high-frequency resonances in these samples. At the same time, our collaborators at the University of Minnesota could not provide us with BiSe topological insulator films due to COVID-19, as their labs were closed for almost 2 years.
I, therefore, devised a new strategy to achieve the project goals. Compensated ferrimagnets such as GdFeCo and GdCo were identified as suitable alternatives as they show high resonance frequencies and high spin-orbit torque efficiency. To use these materials in spintronic devices, it is important to grow them in the ultrathin range (< 10 nm). Gdx(FeCo)1-x [GdFeCo hereafter] thin films in the thickness range of 2-20 nm were grown on industry-compatible high resistance Si(100) [HR-Si(100)] substrates and studied using broadband ferromagnetic resonance (FMR) measurements at room temperature. By tuning the stoichiometry, a nearly compensated behavior was observed in 2 nm GdFeCo for the first time, with an effective magnetization of about 0.02 T and a low effective Gilbert damping constant of about 0.0078 comparable to the lowest values reported so far in 30 nm films. HR-Si(100)/Pt(5nm)/GdFeCo(2-10nm)/SiO2(4nm) stacks were developed to fabricate devices for spin-torque FMR (STFMR) measurements, hall measurements, and SHNO auto-oscillation measurements. The effective spin-orbit torque efficiency of Pt/GdFeCo layers was estimated using STFMR measurements. To understand the spin-orbit torque contribution from GdFeCo, GdFeCo/Cu/NiFe/SiO2-based stacks were developed and studied using STFMR. Analyzing the STFMR data, large values of anti-damping-like and field-like torques were obtained for GdFeCo.
HR-Si(100)/Pt(5nm)/GdFeCo(2-10nm)/SiO2(4nm) stacks were prepared to explore further the compensated ferrimagnet-based SHNOs, and steady-state magnetization auto-oscillations were obtained with SHNO constriction widths in the range of 50-150 nm.
WP2
For the electric field (E-field) control of SHNOs, we identified Co2MnGa, a magnetic Weyl semimetal. Epitaxial thin films of Co2MnGa were grown (by collaborators at NIMS, Japan) and studied using FMR, STFMR, and 2nd harmonic Hall effect measurements. A very high value of the anomalous Hall conductivity of about 1350 S/cm was obtained for a 30 nm thick Co2MnGa film, which is comparable to the bulk case. The large anomalous Hall conductivity indirectly confirms the topological Weyl semimetal phase of the Co2MnGa film. Strong anisotropy in the Gilbert damping was also observed for the Co2MnGa 10-30 nm films, and the lowest values of about 0.009 and 0.013 were obtained for 20 and 30 nm films, respectively. The spin-orbit torque efficiency of the films was characterized by analyzing the 2nd harmonic Hall resistance and STFMR measurements and a damping-like torque efficiency in the range of 0.08-0.12 was obtained. To study the E-field effect on spin-orbit torque efficiency, STFMR microstrips, hall bars, and SHNOs were prepared with different gate geometries. Investigation of these devices is still ongoing and will be finished in collaboration with the host group.
WP3
Brillouin Light Scattering microscopy was performed to investigate the high-frequency ferromagnetic/exchange modes in Pt(5nm)/GdFeCo(2-10nm) based SHNOs. Auto-oscillations on the ferromagnetic mode (~8 GHz at 4-6 kOe) were observed in GdFeCo based SHNOs; auto-oscillations on any exchange mode were not observed. Very interestingly, different Gd compositions resulted in different signs of the threshold current for auto-oscillations, which indicates different signs of the effective spin-orbit torque as a function of the composition. The BLS measurements on Pt(5nm)/GdCo(8nm) and Co2MnGa devices are in progress and will be finished in collaboration with the host group.
WP4
The project started with scientific training on the tools in the applied spintronics group and MC2 cleanroom at Chalmers. At first, the fellow got ALL the licenses to operate the tools such as AJA DC/RF magnetron sputtering, NanOsc Phase FMR, Alternating Gradient Magnetometer (AGM), Raith Electron Beam Lithography, Optical Lithography, Dry Ion Beam Etching, Scanning Electron Microscopy, steady state magnetization auto-oscillation measurement setup, Brillouin Light Scattering (BLS), etc.
The project results are already presented at 4 international conferences, including 3 contributory oral talks and 1 invited talk at the Joint European Magnetic Symposium 2022 (JEMS2022). There is another invited talk scheduled for November 2022 at the 67th annual Magnetism and Magnetic Material conference in Minneapolis, USA.
SPINHALL achieved very low damping in ultrathin films of GdFeCo compensated ferrimagnet, which is comparable to the lowest value reported so far in 30 nm thick films. Magnetization auto-oscillations were observed for the first time in GdFeCo-based SHNOs, and large spin-orbit torques were obtained in GdFeCo. However, the ferromagnetic resonance frequency is not particularly high (around 8-10 GHz at 5kOe) and is not of the exchange type. There is scope to enhance the ferromagnetic resonance frequency further using other compensated ferrimagnets such as GdCo, as evident from my recent ongoing results from Pt/GdCo stacks. Adding Magnetic Weyl semimetal Co2MnGa is expected to provide better control over the SHNOs. These results show great promise for developing ultrafast and energy-efficient ferrimagnetic spintronic devices such as neuromorphic computing using high-frequency SHNOs.
The fellow has developed new skills such as fabrication and characterization of SHNOs, he is now an experienced cleanroom user who can develop SHNOs down to 10 nm constriction width and study STFMR, 2nd harmonic, and steady-state magnetization auto-oscillations of these devices. He also got a basic understanding of the Brillouin Light Scattering microscopy setup. These skills will help him in the future when he builds his own group.
The fellow has developed new skills such as fabrication and characterization of SHNOs, he is now an experienced cleanroom user who can develop SHNOs down to 10 nm constriction width and study STFMR, 2nd harmonic, and steady-state magnetization auto-oscillations of these devices. He also got a basic understanding of the Brillouin Light Scattering microscopy setup. These skills will help him in the future when he builds his own group.