Stochastic Multistable Systems Laboratory (StoLab)

established at UNN in 2018 by order of the rector No. 69-OD dated 06.02.2018 in accordance with the Decree of the Government of the Russian Federation dated April 9, 2010 No. 220 and in accordance with Agreement No. 074-02-2018-330 (2) between UNN and the Ministry of Education and Science of the Russian Federation.

Laboratory head (leading scientist)

Prof. Spagnolo Bernardo
(University of Palermo, Italy)

Science field

Electrical engineering, electronic engineering,
Information Technology.

Direction of scientific research

Comprehensive study of fluctuation phenomena in multistable systems for creating new generations of electronic devices and neuromorphic artificial intelligence technologies based on memristive materials.

Research goal

Investigation of the constructive role of noise and fluctuation phenomena in multistable memristive systems using the latest statistical analysis methods to create new generations of electronic devices and neuromorphic artificial intelligence technologies based on memristive materials.

Research objectives

Study of the influence of external and internal noise on the behavior of multistable systems.

• Study and analysis of phenomena with a constructive role of noise in multistable systems.
• Experimental study of the behavior of memristive nanostructures based on oxide materials under the influence of external and internal noise.
• Development of an adequate physical macromodel of a memristor taking into account the influence of external and internal noise and comparing it with a micromodel of physicochemical phenomena responsible for resistive switching.
• Study of the microscopic nature of the occurrence and influence of flicker and high-frequency noise in memristive nanostructures.
• Experimental demonstration of fundamentally new possibilities for increasing stability, predicting behavior and controlling parameters of memristive devices in prototypes electronic devices and neuromorphic systems of new generation.

The activities of the sector of modern methods of stochastic analysis are aimed at solving fundamental problems of studying the influence of external and internal noises on the behavior of multistable memristive systems, studying and analyzing phenomena with a constructive role of noise in such systems. The sector is managed by the leading Russian specialists in the field of statistical radiophysics A.A. Dubkov. (Doctor of Physical and Mathematical Sciences, Head of the Department "Mathematical Methods in Radiophysics" of the Radiophysical Faculty of the UNN) and Agudov N.V. (Candidate of Physical and Mathematical Sciences, responsible performer of work on the project).

The sector of technology of memristive materials is engaged in the creation of various memristive nanostructures "metal-dielectric-metal" based on oxide films as laboratory samples of nanomaterials for microstructural studies of the nature of internal noise, in the form of separate memristive devices for studying the effect of external noise and in the form of matrices of memristive elements for integration into analog-digital circuits of neural networks. The head of the sector of technology of memristive materials is Mikhailov A.N. (Ph.D., Head of the Laboratory of Physics and Technology of Thin Films, NIPT, UNN).

The sector of microscopic probe research is engaged in the study of the structural and electrophysical properties of various thin-film memristive materials and memristive structures "metal-dielectric-metal" based on the indicated oxide films.

The co-heads of the microscopic probe research sector are O.N. Gorshkov. (Ph.D., associate professor, head of the department of solid-state electronics and optoelectronics of NIPT, director of the REC "Physics of solid-state nanostructures" of the UNN) and Filatov D.O. (Doctor of Physical and Mathematical Sciences, Associate Professor, Senior Researcher, REC "Physics of Solid State Nanostructures", UNN).

The work of the Noise Physics Sector is devoted to the study of the microscopic nature of the appearance and influence of flicker (1 / f) and high-frequency noise in memristive nanostructures. The head of the Noise Physics Sector is A.V. Yakimov. (Doctor of Physical and Mathematical Sciences, Professor of the Radiophysical Faculty of the UNN).

The sector of neuromorphic technologies will be engaged in the development of algorithms and approbation of models of impulse (spike) neural networks based on self-organization processes in arrays of memristive devices. The head of the sector of neuromorphic technologies is V.A. Demin (Candidate of Physical and Mathematical Sciences, Director-coordinator in the direction of nature-like technologies of the National Research Center "Kurchatov Institute").

The scientific subject of the laboratory is purely interdisciplinary, therefore, the successful solution of the project's tasks is possible only at the junction of different fields of knowledge and requires the involvement of employees from different departments of the university, as well as specialists from outside organizations.

On the basis of UNN, active cooperation is carried out with employees of a number of departments of the Research Institute of Neurosciences and the Institute of Biology and Biomedicine, which represent an informal association known throughout the world as the Nizhny Novgorod Neuroscience Center. This cooperation is aimed at the development and creation of new generations of neuromorphic and neurohybrid systems based on the developed memristive circuits and living networks (cultures and tissues) of the brain for solving urgent problems of robotics, artificial intelligence and medicine.

The uniqueness of the laboratory

The uniqueness of the laboratory is associated with the implementation of a unified systematic approach to the formation, study and application of memristive systems, which are associated with a new paradigm in the development of microelectronics, leading to the creation of unique biosimilar information and computing systems, similar in capabilities to the human brain. For the first time in the world, this approach combines a deep knowledge of the nature of microscopic phenomena responsible for the effect of resistive switching on a nanometer scale, corresponding to the expected dimensions of a promising element base, and the latest methods of statistical physics, which make it possible to reveal the fundamental properties of memristors as multistable stochastic systems.

The uniqueness of the study of the microscopic nature of the occurrence and influence of flicker (1 / f) and high-frequency noise in memristive nanostructures lies in the development of methods for non-destructive quality control of manufacturing in order to increase the reliability of the structures under study. The study of noise characteristics opens up wide opportunities for studying the features of the migration of vacancies and other microscopic details of the process of resistive switching in oxide films, depending on the design and microstructure of memristive devices. The developed methods provide an expansion of the diagnostic capabilities for measuring the electrophysical characteristics of the structures being developed, in addition, they expand the arsenal of methods of measuring technology in the actual field of manufacturing memristors, which are the basis for a new generation of nonvolatile memory devices.

This combination provides a correct interpretation of the memristive effect in metal-oxide memristive devices and the ability to control its parameters, which is necessary for potential applications.

The practical significance of research

The laboratory's research is aimed at creating new generations of electronic memory devices and neuromorphic (neurohybrid) artificial intelligence systems. The project will go from fundamental problems of materials science to integrated devices, compact physical and mathematical models and analog-digital systems based on memristors.

This ensures the creation of an element base and computer-aided design tools for the implementation of a number of products based on memristors.

The simplest application of memristive devices is the universal resistive memory RRAM, which combines the qualities of random access memory and read-only memory, and also has a unique radiation resistance, which is important for use in space and atomic reactors.

Promising areas of application of memristive devices include hardware neural networks (neuroprocessors) for IoT, big data processing, robotics, and artificial intelligence.

A breakthrough in neuromedical technologies may be associated with the creation, based on brain-like memristive networks, of promising architectures of computing devices and adaptive neuroprostheses and neurointerfaces for replacing, expanding, or restoring lost brain functionality.

Despite the fact that the project is at the end of the second year of its implementation, an active discussion of joint developments and prospects for the implementation of the project results is already underway with representatives of large Russian and foreign microelectronic companies:

• Research Institute of Measuring Systems. Yu.E. Sedakova
• Samsung Electronics (Advanced Sensor Development Department)
• Huawei Technologies (Chipset & Technology Development Department)