Final Report Summary - MUSTANG (Magnonics Using Spin Torque, spin caloritronics, And Nanoplasmonic engineerinG)
Spintronics refers to devices and phenomena where not only the charge of the electron, but also its spin, is used for functionality. Magnonics refers to devices and phenomena where magnetic excitations - so-called magnons, or spin waves - are used for functionality. Finally, nano-plasmonics refers to metal structures patterend to tailore the electron plasmon resonances in such a way that light-electron interactions can be greatly enhanced.
In the Mustang project, these three emerging research fields have been combined to realize novel device types and functionalities. In particular, a wide range of different types of devices, where spin-polarized currents and pure spin currents are used to excite very high intensity spin waves, have been fabricated and studied. Some of these devices have also been combined with nanoplasmonic structures to realized light-spin wave interactions.
Overall, the project has been a huge success with a high number of high profile results and publications, including many unpublished results being readied for submission. The final project report lists no less than eight different breakthroughs in topics related to all three involved research fields.
To here only describe one of these breakthroughs, the Mustang project was the first to demonstrate mutual synchronization of a new class of spin current driven devices called spin Hall nano-oscillators (SHNOs). Our first paper on this phenomenon, published in Nature Physics (2017), demonstrated that as many as nine such SHNOs connected in series could be mutually synchronized and that mutual synchronization could persist over SHNO separation as large as 4 micrometers. We have since then shown that we can synchronize 21 similar SHNOs in a chain, and 100 similar SHNOs in a two-dimensional array, all so far unpublished results. What is particularly interesting with this particular breakthrough is that such arrays lend themselves to extremely compact oscillator networks which could be used for non-traditional neuromorphic computing and ultrafast pattern recognition. Given the extremely rapid rise of commercial use of deep learning and artificial intelligence in the last few years, an entirely novel computation technology to achieve orders of magnitude speed-up of pattern recognition appears very timely.
It is therefore with great excitement that we will continue to build on this and all other breakthroughs well after the completion of the Mustang project.
In the Mustang project, these three emerging research fields have been combined to realize novel device types and functionalities. In particular, a wide range of different types of devices, where spin-polarized currents and pure spin currents are used to excite very high intensity spin waves, have been fabricated and studied. Some of these devices have also been combined with nanoplasmonic structures to realized light-spin wave interactions.
Overall, the project has been a huge success with a high number of high profile results and publications, including many unpublished results being readied for submission. The final project report lists no less than eight different breakthroughs in topics related to all three involved research fields.
To here only describe one of these breakthroughs, the Mustang project was the first to demonstrate mutual synchronization of a new class of spin current driven devices called spin Hall nano-oscillators (SHNOs). Our first paper on this phenomenon, published in Nature Physics (2017), demonstrated that as many as nine such SHNOs connected in series could be mutually synchronized and that mutual synchronization could persist over SHNO separation as large as 4 micrometers. We have since then shown that we can synchronize 21 similar SHNOs in a chain, and 100 similar SHNOs in a two-dimensional array, all so far unpublished results. What is particularly interesting with this particular breakthrough is that such arrays lend themselves to extremely compact oscillator networks which could be used for non-traditional neuromorphic computing and ultrafast pattern recognition. Given the extremely rapid rise of commercial use of deep learning and artificial intelligence in the last few years, an entirely novel computation technology to achieve orders of magnitude speed-up of pattern recognition appears very timely.
It is therefore with great excitement that we will continue to build on this and all other breakthroughs well after the completion of the Mustang project.