Лаборатория

Лаборатория стохастических
мультистабильных систем

Laboratory of Stochastic
Multistable Systems

Scientific results of 2018

1. In 2018, the following scientific results were obtained:


1. The statistical regularities and properties of the behavior of multistable memristive systems are investigated.

1.1.    The well-known models of multistable memristive systems, created on the basis of various approaches, are studied: microstructural approach, stochastic macromodel approach, thermodynamic analysis and some simplified models obtained by synthesizing various approaches.1.2.    A new geometric method is proposed that allows to study phase transitions in quantum memristive systems.

1.3.    A new macro model of memristor is introduced, which takes into account the influence of external and internal fluctuations.

1.4.    Using new technological approaches, thin oxide films were formed on various substrates, as well as control nanostructures "metal-oxide-metal" and "metal-oxide-semiconductor", manifesting a memristive effect.

1.5.    It has been shown experimentally that the parameters of ionic and electron transport differ for different oxide materials, which affects both the sources and characteristics of internal noise and the systemic energy profile that determines the response to external noise.


2. Phenomena with a constructive role of noise in multistable systems are studied and analyzed.

2.1.    Within the framework of thermodynamic models of quantum memristive systems, the phenomena of noise enhancement of the stability of metastable quantum states, resonance-activated exit from a quantum metastable state have been studied.

2.2.    Within the framework of a stochastic macro-model, the phenomena of noise delay in relaxation to a stationary distribution, acceleration and noise delay in switching a memristive system to a state with low resistance have been  studied; increasing and decreasing by noise the variance of the times of switching the memristive system to the state with low resistance.


3. An experimental macroscopic and microscopic study of the effect of external and internal noise on the behavior of multistable memristive structures based on oxide materials has been carried out.

3.1.     Internal noise in multistable memristive structures based on ZrO2 (Y) oxide films has been experimentally investigated. The statistical characteristics of the noise of films in different conducting states have been determined.

3.2.    The resistance reactions of memristive nanostructures to noisy impulse (spike) activity recorded in the culture of living cells of the hippocampus have been studied.

3.3.    The features of resistive switching in ultrathin (about 5 nm thick) films of stabilized zirconium dioxide (SDZ) with embedded single-layer arrays of Au nanoparticles have been studied by atomic-force microscopy.

3.4.    The dynamics of the spatial distribution of the potential induced by an electric charge injected from the AFM probe into thin (<10 nm thick) SDZ / Si films with embedded Au nanoparticles has been studied by scanning Kelvin probe microscopy (SKPM). The SKPM images and profiles of the surface potential induced by electrons trapped in Au nanoparticles have been measured and analyzed as a function of the time elapsed from the moment of charge injection


4. The microscopic nature of the appearance and influence of flicker- and high-frequency noise in memristive structures, including in nanoscale areas, has been studied by atomic-force microscopy with a conductive AFM probe.

4.1.    Noises (caused by fluctuations in the concentration of magnetic defects) in rare-earth metal oxides, which are promising for the creation of memristive structures, have been investigated.

4.2.    The flicker noise in autogenerators made on an element base including memristive structures, caused by fluctuations of their equivalent resistive elements, has been investigated. The possibility of determining microscopic processes in memristive structures through macroscopic measurable quantities is shown.

4.3.    The microscopic nature of the appearance of flicker-noise in memristive structures has been investigated by atomic-force microscopy with a conducting AFM probe. The activation energies for the diffusion of oxygen ions inside a conducting filament in a structure based on ZrO2 (Y) has been determined.


5. Predictive models of the evolution of multistable memristive systems under the influence of noise and fluctuations have been created.

6. The following elements of perspective architectures of impulse (spike) neural networks have been developed and investigated:

5.1.    The forms of probabilistic distributions of fluctuating parameters have been obtained experimentally. It is shown that the observed trimodal profiles qualitatively correspond to the tristable profile of the free energy of filament formation, which appears in the framework of the thermodynamic model of the memristor.

5.2.    The time dependences of the resistive state have been established depending on the parameters of the potential profile and the amplitude of external noise.

6.1.    A hardware and software impulse neural network consisting of 4 input neurons connected to one output neuron of the threshold type through memristive connections based on a nanocomposite material: CoFeB metal granules in a LiNbO3-y dielectric matrix with oxygen vacancies.

6.2.    The parameters of the threshold neuron have been selected and the training of the developed impulse neural network was carried out on a noise signal with a Poisson-type distribution according to local learning rules of the STDP type. The independence of the learning result from the initial state of the network has been demonstrated.


7. 6 articles have been published in scientific journals, indexed in the Web of Science database (including 2 articles in journals from Q1). In addition, 7 articles have been sent to scientific publications indexed in the Web of Science database.

8. The project executors took part in 4 international conferences, scientific seminars, symposia on the subject of the laboratory's work (at the expense of subsidies).

9. 3 events have been held on the topic "Stochastic multistable systems" (at the expense of subsidies).

10.  An application has been filled and registered for a patent for an invention "Method of manufacturing a memristor with electric field nanoconcentrators" (Notification of acceptance and registration of the application, reg. No. 2018146457 dated 26.12.2018).

11. The current repair of the laboratory premises has been carried out.

12. 6 members of the research team were trained.

13. The topology of the layouts of the memristive devices has been developed, and the parameters of the memristive devices have been studied.

13.1. Topological variants of memristive devices layouts were developed in Cadence end-to-end CAD design using the Virtuoso topological editor as a part of a test chip that provides installation in a standard cermet package and integration of memristors into analog-digital electrical circuits, which will be developed to demonstrate the constructive role of noise and new neural network architectures on based on memristive devices. In the course of research, the ability to control resistive switching according to the STDP rule has been demonstrated.


14. The project executors took part in 5 international conferences, scientific seminars, symposia on the subject of the laboratory's work (at the expense of additional funds).

15. A seminar "Neuromorphic and neurohybrid systems" (within the framework of the "Volga Neuroscience Meeting - 2018" symposium) has been held on the topic of the laboratory at the expense of additional funds.

2. Publication in scientific journals indexed in the "web of science" database

№ 

Title of the article

Surname

Name

Title
of the journal

Year, month
(volume, issue)

Impact factor of publication

1

Enhancing metastability by dissipation and driving in an asymmetric bistable quantum system

Spagnolo
Carollo

Bernardo
Angelo

Entropy

2018, April
(20, 4)

2.305

2

Stabilizing effect of driving and dissipation on quantum metastable states

Spagnolo
Carollo

Bernardo
Angelo

Phys. Rev. A
(Q1)

2018, April
(97, 4)

2.909

3

Uhlmann curvature in dissipative phase transitions

Spagnolo
Carollo

Bernardo
Angelo

Scientific Reports – Nature
(Q1)

2018, June
(8, 1)

4.122

4

Symmetric Logarithmic Derivative of Fermionic Gaussian States

Spagnolo
Carollo

Bernardo
Angelo

Entropy

2018, July
(20, 7)

2.305

5

Stabilization by Dissipation and Stochastic Resonant Activation in Quantum Metastable Systems

Spagnolo
Carollo

Bernardo
Angelo

Eur. Phys. J-Special Topics

2018, September
(227, 3-4)

1.947

6

Conductive Atomic Force Microscopy Study of the Resistive Switching in Yttria-Stabilized Zirconia Films with Au Nanoparticles

Filatov
Antonov
Pavlov
Antonov
Gorshkov

Dmitry
Ivan
Dmitry
Dmitry
Oleg

Scanning

2018, 07
(5489596)

0,994