Design of oscillatory neural networks by machine learning.

We demonstrate the utility of machine learning algorithms for the design of oscillatory neural networks (ONNs). After constructing a circuit model of the oscillators in a machine-learning-enabled simulator and performing Backpropagation through time (BPTT) for determining the coupling resistances between the ring oscillators, we demonstrate the design of associative memories and multi-layered ONN classifiers. The machine-learning-designed ONNs show superior performance compared to other design methods (such as Hebbian learning), and they also enable significant simplifications in the circuit topology. We also demonstrate the design of multi-layered ONNs that show superior performance compared to single-layer ones. We argue that machine learning can be a valuable tool to unlock the true computing potential of ONNs hardware.

Spin-Wave Optics in YIG Realized by Ion-Beam Irradiation.

Direct focused-ion-beam writing is presented as an enabling technology for realizing functional spin-wave devices of high complexity, and demonstrate its potential by optically-inspired designs. It is shown that ion-beam irradiation changes the characteristics of yttrium iron garnet films on a submicron scale in a highly controlled way, allowing one to engineer the magnonic index of refraction adapted to desired applications. This technique does not physically remove material, and allows rapid fabrication of high-quality architectures of modified magnetization in magnonic media with minimal edge damage (compared to more common removal techniques such as etching or milling). By experimentally showing magnonic versions of a number of optical devices (lenses, gratings, Fourier-domain processors) this technology is envisioned as the gateway to building magnonic computing devices that rival their optical counterparts in their complexity and computational power.

Nanoscale neural network using non-linear spin-wave interference.

We demonstrate the design of a neural network hardware, where all neuromorphic computing functions, including signal routing and nonlinear activation are performed by spin-wave propagation and interference. Weights and interconnections of the network are realized by a magnetic-field pattern that is applied on the spin-wave propagating substrate and scatters the spin waves. The interference of the scattered waves creates a mapping between the wave sources and detectors. Training the neural network is equivalent to finding the field pattern that realizes the desired input-output mapping. A custom-built micromagnetic solver, based on the Pytorch machine learning framework, is used to inverse-design the scatterer. We show that the behavior of spin waves transitions from linear to nonlinear interference at high intensities and that its computational power greatly increases in the nonlinear regime. We envision small-scale, compact and low-power neural networks that perform their entire function in the spin-wave domain.

Efficient electromagnetic transducers for spin-wave devices.

This paper presents a system-level efficiency analysis, a rapid design methodology, and a numerical demonstration of efficient sub-micron, spin-wave transducers in a microwave system. Applications such as Boolean spintronics, analog spin-wave-computing, and magnetic microwave circuits are expected to benefit from this analysis and design approach. These applications have the potential to provide a low-power, magnetic paradigm alternative to modern electronic systems, but they have been stymied by a limited understanding of the microwave, system-level design for spin-wave circuits. This paper proposes an end-to-end microwave/spin-wave system model that permits the use of classical microwave network analysis and matching theory towards analyzing and designing efficient transduction systems. This paper further compares magnetostatic-wave transducer theory to electromagnetic simulations and finds close agreement, indicating that the theory, despite simplifying assumptions, is useful for rapid yet accurate transducer design. It further suggests that the theory, when modified to include the exchange interaction, will also be useful to rapidly and accurately design transducers launching magnons at exchange wavelengths. Comparisons are made between microstrip and co-planar waveguide lines, which are expedient, narrowband, and low-efficiency transducers, and grating and meander lines that are capable of high-efficiency and wideband performance. The paper concludes that efficient microwave-to-spin-wave transducers are possible and presents a meander transducer design on YIG capable of launching [Formula: see text]nm spin waves with an efficiency of - 4.45 dB and a 3 dB-bandwidth of 134 MHz.

Experimental demonstration of a concave grating for spin waves in the Rowland arrangement.

We experimentally demonstrate the operation of a Rowland-type concave grating for spin waves, with potential application as a microwave spectrometer. In this device geometry, spin waves are coherently excited on a diffraction grating and form an interference pattern that focuses spin waves to a point corresponding to their frequency. The diffraction grating was created by focused-ion-beam irradiation, which was found to locally eliminate the ferrimagnetic properties of YIG, without removing the material. We found that in our experiments spin waves were created by an indirect excitation mechanism, by exploiting nonlinear resonance between the grating and the coplanar waveguide. Although our demonstration does not include separation of multiple frequency components, since this is not possible if the nonlinear excitation mechanism is used, we believe that using linear excitation the same device geometry could be used as a spectrometer. Our work paves the way for complex spin-wave optic devices-chips that replicate the functionality of integrated optical devices on a chip-scale.

Nanoantenna-based ultrafast thermoelectric long-wave infrared detectors.

We investigate the generation of electrical signals by suspended thermoelectrically coupled nanoantennas (TECNAs) above a quasi-spherical reflector cavity in response to rapidly changing long-wave infrared radiation. These sensors use a resonant nanoantenna to couple the IR energy to a nanoscale thermocouple. They are positioned over a cavity, etched into the Si substrate, that provides thermal isolation and is designed as an optical element to focus the IR radiation to the antenna. We study the frequency-dependent response of such TECNAs to amplitude-modulated 10.6 µm IR signals. We experimentally demonstrate response times on the order of 3 µs, and a signal bandwidth of about 300 kHz. The observed electrical response is in excellent correlation with finite element method simulations based on the thermal properties of nanostructures. Both experiments and simulations show a key trade-off between sensitivity and response time for such structures and provide solutions for specific target applications.

Cavity-Backed Antenna-Coupled Nanothermocouples.

This paper reports a two-orders-of-magnitude improvement in the sensitivity of antenna-coupled nanothermocouple (ACNTC) infrared detectors. The electrical signal generated by on-chip ACNTCs results from the temperature difference between a resonant antenna locally heated by infrared radiation and the substrate. A cavity etched under the antenna provides two benefits. It eliminates the undesirable cooling of the hot junction by thermally isolating the antenna from the substrate. More importantly, careful cavity design results in constructive interference of the incident radiation reflected back to the antenna, which significantly increases the detector sensitivity. We present the cavity-depth-dependent response of ACNTCs with cavity depths between 1 µm and 22 µm. When constructive interference is maximized, the thermal response increases by 100-fold compared to devices without the cavity.

Nanoscale spectrum analyzer based on spin-wave interference.

We present the design of a spin-wave-based microwave signal processing device. The microwave signal is first converted into spin-wave excitations, which propagate in a patterned magnetic thin-film. An interference pattern is formed in the film and its intensity distribution at appropriate read-out locations gives the spectral decomposition of the signal. We use analytic calculations and micromagnetic simulations to verify and to analyze the operation of the device. The results suggest that all performance figures of this magnetoelectric device at room temperature (speed, area, power consumption) may be significantly better than what is achievable in a purely electrical system. We envision that a new class of low-power, high-speed, special-purpose signal processors can be realized by spin-waves.

Spin-orbit torque-assisted switching in magnetic insulator thin films with perpendicular magnetic anisotropy.

As an in-plane charge current flows in a heavy metal film with spin-orbit coupling, it produces a torque on and thereby switches the magnetization in a neighbouring ferromagnetic metal film. Such spin-orbit torque (SOT)-induced switching has been studied extensively in recent years and has shown higher efficiency than switching using conventional spin-transfer torque. Here we report the SOT-assisted switching in heavy metal/magnetic insulator systems. The experiments used a Pt/BaFe12O19 bilayer where the BaFe12O19 layer exhibits perpendicular magnetic anisotropy. As a charge current is passed through the Pt film, it produces a SOT that can control the up and down states of the remnant magnetization in the BaFe12O19 film when the film is magnetized by an in-plane magnetic field. It can reduce or increase the switching field of the BaFe12O19 film by as much as about 500 Oe when the film is switched with an out-of-plane field.

Majority logic gate for 3D magnetic computing.

For decades now, microelectronic circuits have been exclusively built from transistors. An alternative way is to use nano-scaled magnets for the realization of digital circuits. This technology, known as nanomagnetic logic (NML), may offer significant improvements in terms of power consumption and integration densities. Further advantages of NML are: non-volatility, radiation hardness, and operation at room temperature. Recent research focuses on the three-dimensional (3D) integration of nanomagnets. Here we show, for the first time, a 3D programmable magnetic logic gate. Its computing operation is based on physically field-interacting nanometer-scaled magnets arranged in a 3D manner. The magnets possess a bistable magnetization state representing the Boolean logic states '0' and '1.' Magneto-optical and magnetic force microscopy measurements prove the correct operation of the gate over many computing cycles. Furthermore, micromagnetic simulations confirm the correct functionality of the gate even for a size in the nanometer-domain. The presented device demonstrates the potential of NML for three-dimensional digital computing, enabling the highest integration densities.

High-yield transfer printing of metal-insulator-metal nanodiodes.

Nanoscale metal-insulator-metal (MIM) diodes represent important devices in the fields of electronic circuits, detectors, communication, and energy, as their cutoff frequencies may extend into the "gap" between the electronic microwave range and the optical long-wave infrared regime. In this paper, we present a nanotransfer printing method, which allows the efficient and simultaneous fabrication of large-scale arrays of MIM nanodiode stacks, thus offering the possibility of low-cost mass production. In previous work, we have demonstrated the successful transfer and electrical characterization of macroscopic structures. Here, we demonstrate for the first time the fabrication of several millions of nanoscale diodes with a single transfer-printing step using a temperature-enhanced process. The electrical characterization of individual MIM nanodiodes was performed using a conductive atomic force microscope (AFM) setup. Our analysis shows that the tunneling current is the dominant conduction mechanism, and the electrical measurement data agree well with experimental data on previously fabricated microscale diodes and numerical simulations.

Bio-inspired nano-sensor-enhanced CNN visual computer.

Nanotechnology opens new ways to utilize recent discoveries in biological image processing by translating the underlying functional concepts into the design of CNN (cellular neural/nonlinear network)-based systems incorporating nanoelectronic devices. There is a natural intersection joining studies of retinal processing, spatio-temporal nonlinear dynamics embodied in CNN, and the possibility of miniaturizing the technology through nanotechnology. This intersection serves as the springboard for our multidisciplinary project. Biological feature and motion detectors map directly into the spatio-temporal dynamics of CNN for target recognition, image stabilization, and tracking. The neural interactions underlying color processing will drive the development of nanoscale multispectral sensor arrays for image fusion. Implementing such nanoscale sensors on a CNN platform will allow the implementation of device feedback control, a hallmark of biological sensory systems. These biologically inspired CNN subroutines are incorporated into the new world of analog-and-logic algorithms and software, containing also many other active-wave computing mechanisms, including nature-inspired (physics and chemistry) as well as PDE-based sophisticated spatio-temporal algorithms. Our goal is to design and develop several miniature prototype devices for target detection, navigation, tracking, and robotics. This paper presents an example illustrating the synergies emerging from the convergence of nanotechnology, biotechnology, and information and cognitive science.

The role of field coupling in nano-scale cellular nonlinear networks.

We review some of our previous work on field-coupling in nano-scale cellular arrays. Electronic devices based on metallic and magnetic nanoscale dots and molecular structures have been suggested, however, no technologically viable architecture for nanoelectronic circuit integration has emerged so far. A natural architecture on the nanoscale appears to be near-neighbor cellular networking, and we explore promising alternative ways of integrating nanodevices by direct physical field coupling, i.e. either by Coulomb or by magnetic interactions. We review new architectures for such field-coupled nanocircuits.