N.A. Vetrova1, K.P. Pchelintsev2, V.D. Shashurin3

1–3 Bauman Moscow State Technical University (Moscow, Russia)

Ensuring the required level of reliability of modern radioelectronic devices based on semiconductor heterostructures is currently an extremely difficult task. The analysis of variants of their design and technological realization, taking into account the time factor and external influences requires the implementation of multi-iteration calculations of electrical characteristics. Current computational algorithms demand a significant amount of computational resources to operate, as well as a number of additional empirical coefficients. That in fact makes their application within the limits of the decision of a task of ensuring reliability of radio-electronic devices on the basis of semiconductor heterostructures for the prediction of electrical characteristics of RTD taking into account the time factor and external influences impossible.

Purpose – to design and train of artificial neural network capable of approximating complex, including hidden, functional dependences of RTD electrical characteristics from design-technological parameters with an acceptable accuracy and performance taking into account the time factor and external influences.

An artificial neural network for calculation of electrical characteristics of RTD taking into account the time factor and external influences is developed.

Using the trained neural network model can significantly reduce (up to two orders of magnitude) the calculation time of electrical characteristics of RTD taking into account the time factor and external influences. Use of experimental data for training of neural network allows to increase model accuracy by taking into account influence of technological errors (appearing during manufacturing), and also various degradation changes (proceeding with time and under the influence of external factors during exploitation of resonant-tunnel diodes), and also to exclude from degradation models empirical correcting coefficients.

Vetrova N.A., Pchelintsev K.P., Shashurin V.D. A Perseptron network for prediction of the electrical characteristics of a resonanttunnel diode. Nanotechnology: development and applications – XXI century. 2021 V. 13. № 4. P. 5–9. DOI: https://doi.org/10.18127/ j22250980-202104-01 (in Russian)

1. Wang J., Al-Khalidi A., Zhang C., Ofiare A., Wang L., Wasige E., Figueiredo J.M.L. Resonant tunneling diode as high speed optical/electronic transmitter. 10th UK-Europe-China Workshop on Millimetre Waves and Terahertz Technologies. 2017. 1–4 Doi: 10.1109/UCMMT.2017.8068497
2. Kasagi K., Suzuki S., Asada M. Large-scale array of resonant-tunneling-diode terahertz oscillators for high output power at 1 THz. Journal of Applied Physics. 2019. V. 125. № 15. P. 151601.1. Doi: 10.1063/1.5051007
3. Yang F. Room Temperature Terahertz SubHarmonic Mixer Based on GaN Nanodiodes. Journal of Computational and Theoretical Nanoscience. 2017. V. 14. № 4. P. 1766. Doi: 10.1166/jctn.2017.6501
4. Baba R., Stevens B.J., Mukai T., Hogg R.A. Epitaxial Designs for Maximizing Efficiency in Resonant Tunneling Diode Based Terahertz Emitters. IEEE Journal of quantum electronic. 2018. V. 54. № 2. Doi:10.1109/JQE.2018.2797960
5. Shashurin V.D., Vetrova N.A., Pchelincev K.P., Kuimov E.V., Kozij A.A. Effektivnyj vychislitel'nyj algoritm rascheta elektricheskih harakteristik nanorazmernyh geterostruktur na osnove formalizma Landauera-Buttikera. Nanotekhnologii: razrabotka, primenenie – XXI vek. 2019. T. 11. № 1. S. 34–43 (in Russian).
6. Nadar S., Zaknoune M., Wallart X., Coinon C., Emilien P., Ducournau G., Gamand F., Thirault M. High Performance Heterostructure Low Barrier Diodes for Sub-THz Detection. IEEE Transactions on Terahertz Science and Technology. 2017. V. 7. № 6. P. 780. Doi:10.1109/TTHZ.2017.2755503
7. Anantram M.P., Lundstrom M.S., Nikonov D.E. Modeling of Nanoscale Devices. Proceedings of the IEEE. 2008. V. 96. № 9. P. 1511. Doi:10.1109/JPROC.2008.927355
8. Koziy A.A., Vetrova N.A., Pchelintsev K.P., Shashurin V.D., Meshkov S.A. Quantum-mechanical models for calculating the electrical characteristics of semiconductor 2-d structures for technological optimization of nanoelectronics devices based on them. Journal of Physics. 2019. V. 012194. Doi: 10.1088/1742-6596/1410/1/01219
9. Schmidhuber J. Deep learning in neural networks: An overview. Neural Networks. 2015. V. 61. P. 85. Doi: 10.1016/j.neunet.2014.09.003
10. Vetrova N.A., Pchelintsev K.P., Shashurin V.D. An artificial neural network as a predictor of electrical characteristics of nanoelectronic device channel based on a low-dimensional heterostructure. J. Phys.: Conf. Ser. 2020. V. 1695. Doi: 10.1088/1742-6596/1695/1/012152
11. Vetrova N.A., Pchelincev K.P., SHashurin V.D., Gudkov A.G., Solov'ev YU.V. Nejrosetevye metody dlya TCAD-sred modelirovaniya geterostrukturnyh nanoelektronnyh priborov s poperechnym tokoperenosom. Nanotekhnologii: Razrabotka, primenenie –XXI vek. 2021. № 1. S. 5–11 (in Russian).
12. Laudani A., Lozito G.M., Fulginei F.R., Salvini A. On Training Efficiency and Computational Costs of a Feed Forward Neural Network: A Review. Computational Intelligence and Neuroscience. 2015. V. 2015. Doi: 10.1155/2015/818243
13. Dua V. Finite element solution of coupled-partial differential and ordinary equations in multicomponent polymeric coatings. Computers & Chemical Engineering. 2011. V. 35. № 3. P. 545. Doi: 10.1016/j.compchemeng.2010.06.005
14. Jamili E., Dua V. Parameter estimation of partial differential equations using artificial neural network. Computers & Chemical Engineering. 2019. V. 147. P. 107221. Doi: 10.1016/j.compchemeng.2020.107221
15. Vujicic T., Tripo M., Jelena L., Balota A., Sevarac Z. Comparative Analysis of Methods for Determining Number of Hidden Neurons in Artificial Neural Network. Central European Conference on Information and Intelligent Systems. 2016. P. 219.
16. Gnana Sheela K., Deepa S.N. Review on Methods to Fix Number of Hidden Neurons in Neural Networks. Mathematical Problems in Engineering. 2013. V. 6. Doi: 10.1155/2013/425740