Research Laboratory of Magnonics and Radiophotonics named after B.A. Kalinikov

Scientific results:

• Using the method of Hamiltonian formalism, a theory of nonlinear excitation of spin waves by microstrip antennas was constructed. Using the constructed theory and calculations carried out on it, the influence of two types of nonlinearity on the amplitude of the excited spin wave was compared. For the first time, the basic physical mechanism responsible for the nonlinearity of the characteristics describing the excitation of spin waves by microstrip antennas has been established. This mechanism is a four-wave parametric interaction of spin waves in that part of the magnetic film that is located directly under the antenna.
• A theory of nonlinear transmission characteristics of an active ring resonator, taking into account nonlinear dispersion and attenuation of surface spin waves, has been developed.
• A study of the influence on the formation of nonlinear transmission characteristics of such active ring resonator parameters as the open ring gain, the external magnetic field strength and the input dimensionless amplitude of spin waves was carried out.
• It has been shown that increasing the gain to the self-generation threshold leads to an expansion of the frequency range of bistability, while, in the case of a sufficiently small amplitude of spin waves, the influence of nonlinear damping can be neglected. It is shown that an increase in the magnetic field ensures an increase in the nonlinear coefficient, which also leads to an expansion of the frequency range of bistability. It was found that an increase in the input amplitude of spin waves leads to an expansion of the frequency range of bistability; however, the influence of nonlinear damping in this mode is much more pronounced. The results obtained were used to study the nonlinear dynamics of spin waves in magnonic ring reservoir systems in order to improve their performance.
• A numerical model has been developed to calculate the performance characteristics of a reservoir computing system built on a spin-wave active ring oscillator, in which data input is carried out by changing the gain of the feedback circuit. The model is based on the Ginzburg-Landau equation, which describes the nonlinear phase shift and nonlinear damping of operating spin waves.
• Based on the developed numerical model, a set of computer programs was created to determine the efficiency of reservoir systems. The main evaluation criteria were the capacities obtained as a result of the short-term memory test and the parity test. These tests allowed us to evaluate two main efficiency parameters, namely memory and reservoir nonlinearity.
• A physical implementation of a time-delay reservoir computer based on a ring resonator with a spin-wave delay line has been developed. It was shown that this system, due to the dynamics of spin waves, can realize the properties necessary for the reservoir, such as nonlinearity and fading memory. The inclusion of a reference line in the reservoir computer circuit made it possible to improve nonlinear computing power without a sharp reduction in the amount of fading memory. Performance tests of the reservoir computer were carried out and it turned out that the performance of the developed reservoir computer was higher than that known from the literature.
• A model of a magnon-photon physical reservoir, built on a combination of spin-wave and optoelectronic delay lines, has been developed. Its high performance has been demonstrated.
• A magnetic field sensor has been developed based on a magnon reservoir computer. The principle of its operation is based on the dependence of the dynamics of spin waves in the spin-wave delay line on external conditions.

Implementation of research results:

The beneficiaries of the research results are 1) scientific and educational organizations specializing in conducting fundamental and applied research in the field of artificial intelligence and signal processing, as well as 2) organizations involved in the development and implementation of information systems and technologies in such areas as machine learning, computer vision, artificial intelligence, information retrieval, etc.

Scientific organizations will be able to use the scientific results obtained by the research laboratory of magnonics and microwave photonics for the further development of advanced fundamental research in the field of artificial intelligence systems, neural networks, information processing systems and magnonics.

For companies specializing in the development of various artificial intelligence systems, the scientific developments created (and continuing to be created by the laboratory team) will become the mathematical and physical foundation that allows the implementation of optimal information processing algorithms based on neural networks into practical activities.

Organizational and infrastructural changes:

Laboratory of Microwave Photonics and Magnonics (LMPM) is established.

Education and personnel occupational retraining:

• Educational programs have been developed (including courses of lectures, laboratory work, practical classes) in the following areas of scientific research: “Artificial neural networks on the principles of physical electronics”, “Microwave photonics and fiber optics”, “New Generation of Electronic Component Base”, “Physical Electronics”, “Microwave Photonics”. These training programs are introduced into the educational process.
• The laboratory’s scientific team organizes and holds the annual international conference “Microwave Electronics and Microelectronics”.
• An additional professional development program “Magnonics and reservoir calculations” has been developed.
• Educational programs for engineering specialist training in accordance with Federal State Educational Standard 4 have been developed.

Cooperation:

ООО «Инзарус», ООО «Пауплей Системс» (Россия), Российский квантовый центр, Сколтех, ИОФ РАН, ПАО МТС.