The development of 2.5D and 3D crystal integration technologies imposes increased requirements on the housings of microelectronic products. A promising material for the manufacture of housings is anodic aluminum oxide, which allows the formation of interlayer vertical electrical connections of different layers without additional operations of creating holes in the interlayer insulation. This makes it possible to obtain subsequent layers on the surface with good planarization. The article considers a method for manufacturing substrates from porous anodic aluminum oxide with insulated conductive pads using through local anodizing of aluminum. A metho dology for studying the insulating properties of the obtained dielectric substrates with various methods for improving the quality of insulation of conductive channels is presented. The results of a study of leakage currents of isolated conductive channels depending on the applied voltage are presented.

One of the promising areas of label-free analysis is optical biosensors based on 2D photonic crystals created from an array of nanopillars. The disadvantage of such crystals, usually formed on the basis of silicon and its dioxide, is the use of photolithography technology. The use of photonic crystals based on self-organized systems, such as arrays of niobium and tantalum oxide nanopillars obtained by anodizing two-layer Al/Nb and Al/Ta systems, allows us to solve this problem. The optical properties of photonic crystals based on arrays of niobium and tantalum oxide nanopillars were simulated. The sensitivity of the biosensor was determined by the shift of the main reflection peak relative to its position when filling the voids between the nanopillars of photonic crystals with air and biotin-streptavidin. The reflection spectra were estimated for specific and non-specific binding of biotin-streptavidin to the surface of photonic crystals. The intensity of the main reflection peak of the bio sensor based on tantalum oxide with a metallic Ta sublayer was 0.41 rel. units at a wavelength of 353 nm, the shifts of the peaks with a refractive index distribution of 1.46 for specific and non-specific binding were 12 and 24 nm, respectively. The intensity of the main reflection peak of the biosensor based on niobium oxide with a metallic Nb sublayer was 0.51 rel. units at a wavelength of 371 nm, the shifts of the reflection peaks with a refractive index distribution of 1.46 for specific and non-specific binding were 12 and 31 nm, respectively.

The article presents the results of a study of laser-assisted electrochemical deposition of functional thinfilm submicron conductive structures using cyanide electrolytes. The effect of laser radiation with wavelengths of  λ = 600–200 nm on cyanide electrolytes was investigated. This radiation leads to the active dissociation of compounds of the K_iMe(CN)_j, Me_i(CN)_j, and K_iMe_i(CN)_j types. This promotes a local increase in the concentration gradient of active electrochemical species and causes an increase in the local rate of electrochemical deposition as a whole. The influence of such parameters as laser radiation intensity, the absorption coefficient of laser radiation by the electrolyte (effective absorption cross-section), and the dissociation quantum yield, which are determined by the composition of the electrolyte and the selectivity of the laser irradiation, on these processes was analyzed.

The article presents the results of obtaining large-block polycrystals of the semiconductor solid solution system Cu₂FeSn(S_x,Se_1–x)₄, formed in the entire concentration range. It was found that both the compounds Cu₂FeSnS₄, Cu₂FeSnSe₄ and the solid solutions have a tetragonal stannite structure I4̅2m with a ratio of the unit cell parameters с/а ∼ 2. The unit cell parameters varied linearly in accordance with Vegard’s law – from a = (5.704 ± 0.005) Å and c = (11.261 ± 0.01) Å for Cu₂FeSnSe₄ to a = (5.441 ± 0.005) Å and c = (10.72 ± 0.01) Å for Cu₂FeSnS₄. The dependences of the X-ray density and the Debye temperature were determined. Based on the results of differential thermal analysis, the melting temperatures of the samples were determined and a system phase diagram was constructed.

A technique for analyzing the average intensity of electromagnetic background (EMB) created near the earth’s surface by subscriber terminals of satellite communication systems which are based on constellations of low-orbit satellites has been developed. Using previously obtained results of analyzing the EMB created near the earth’s surface by radiations from the space segment of these systems, estimates have been made of the total levels of anthropogenic and natural EMB in the microwave range in areas with intensive information servicing of the population by these systems. The results obtained indicate that radiation from subscriber terminals of these systems makes the main contribution to the microwave EMB intensity near the earth’s surface, exceeding by several orders of magnitude other EMB components created by both natural sources of microwave radiation and radiation from many low-orbit satellites, generally significantly changing the physical characteristics of the habitat.

The paper presents a calculation method and analytical relationships for determining the efficiency of using error-correcting channel coding in systems with a binary block code of Bose – Chaudhuri – Hocquenghem (BCH), M-position modulation and decoding using the Viterbi algorithm with a soft decision (SOVA). A generalized decoding efficiency indicator was used, which took into account the code parameters, modulation format and decoder operating mode. In comparison with the known method, the proposed method does not require knowledge of the weight components of the code spectrum and the use of computer modeling procedures in calculations. The obtained analytical relationships are presented in a compact form and general for research. They are used to calculate the error immunity, energy gain from coding, the correcting ability of the decoder and the information efficiency of communication channels with variations in code parameters and error probabilities at the decoder output.

Modern projection optical systems, such as DUV lithography and high-numerical-aperture objectives, require extremely high surface machining accuracy, making the consideration of mechanical deformations of optical elements particularly critical. This paper presents the results of a computer simulation of the deformation of a 195 mm diameter fused silica biconvex lens under vacuum. The study was performed in CAE (Computer-aided engineering) software suites employing different calculation algorithms. A lens model designed in a CAD (Computer-aided design) system was used. The lens material is fused silica with a Young’s modulus of 72 GPa and a Poisson’s ratio of 0.17. It was established that an operational pressure differential of 15 kPa causes non-uniform deformation of the lens surface, with maximum values ranging from 22.59 to 23.24 nm, depending on the calculation algorithm. The discrepancy between the results was 2.8 %. A linear dependence of deformation on the pressure differential was established: as the pressure differential changes from 0 to 18 kPa, the deformation increases from 0.75 to 27.74 nm. The greatest surface distortion is observed in the central zone of the lens, which is critical for interferometric measurements requiring nanometer-level accuracy. The results underscore the necessity of adjusting vacuum mounting parameters to minimize deformations and improve the quality of optical surface machining.

This article presents the results of solving the problem of recognizing ballistic missile types using neural networks based on trajectory information. Trajectory information from radar stations tracking these ballistic objects was used for recognition. Recognition using neural networks based on the “altitude – energy altitude” parameters is considered. Simulations showed that, with approximately equal solution times, the probability of correct recognition for feedforward neural networks (FFNN) is significantly higher than for the plane-cell-based algorithm used for comparison.

This article examines methods for working with imbalanced data when building machine learning models for classification problems. Balancing methods are studied to determine their impact on the performance of classical and ensemble models. Five datasets of varying sizes and degrees of imbalance are selected and preprocessed. The impact of the imbalanced-learn library’s methods of increasing the smaller class and decreasing the larger class is studied, both when used separately and in combination. The optimal class ratio after balancing is determined (from 1:1 to 2:1, where the first number corresponds to the number of objects in the initially smaller class), and the impact of hyperparameter selection using Optuna is assessed. It is established that hyperparameter optimization does not compensate for the lack of data balancing, and the best model performance is achieved by using an integrated approach combining two different types of balancing methods, using an ensemble, and hyperparameter selection. The greatest impact on model quality was achieved by using a single balancing method in conjunction with ensemble modeling, so this combination is recommended for limited time and computational resources. Adding a larger class reduction method and hyperparameter tuning is advisable when resources are sufficient and model quality requirements are high.

A new scheme for constructing truly random sequence generators using a randomness source control unit is proposed. A generalized controlled bistable element is considered as a source of randomness, for which a lo gical model based on controlled feedback inverters is proposed. It is established that the transition of a bistable element to a metastable state is possible regardless of its internal structure. The metastable state is expressed in oscillation of the output signal with a unique frequency, which will further allow generating unpredictable random sequences. The created software model of a controlled bistable element in the SystemVerilog language proved its viability during testing.

The article describes the architecture and functionality of a cross-platform band structure analyzer for processing the calculation results of the popular specialized OpenMX material properties modeling package. Its operation is demonstrated using a test task as an example. Commercial packages for analyzing and processing data obtained in programs for quantum mechanical modeling of materials are presented. It is shown that the OpenMX package lacks an analyzer program for processing the results of calculations of the structure of materials. Practical problems of analyzing the calculations of the band structure of materials and the functionality of such a program are considered, and requirements for its implementation are formulated. The proposed analyzer is not inferior in performance to the existing analog in terms of overall time costs, while it has a wider functionality, including basic analysis and post-processing of data with the ability to optionally customize the output file saved in a convenient text format. This development is promising in the context of universalization to ensure compatibility with other widely used commercial quantum mechanical modeling packages.

Numerous web services designed for generating interior images using generative design methods currently exist on the market. However, the high cost and limitations of some commercial solutions limit their accessibility to a wide range of users. This article proposes a cost-effective solution without significant sacrifices in functionality and quality. The article presents the results of developing a mobile application for generating images of rooms (interiors) based on a photograph and a user-supplied text description. A generative adversarial neural network was used to generate the images. The process of converting a text query into an image is described, and application testing results are provided.

Today, classification problems are typically solved using neural networks. When choosing a network architecture, particular attention is paid to compression layers, while the multilayer perceptron architecture is chosen intuitively, although the quality of the solution largely depends on the type of separating surface. This article proposes an algorithm for testing the effectiveness of a multilayer perceptron architecture based on an analysis of input data properties. The algorithm is based on a special numerical metric for assessing the effectiveness of a neural network architecture, the skew coefficient, developed by the authors. The coefficient is calculated based on the confusion matrix, which does not require a large amount of calculations. Experiments have shown that choosing the right classifier architecture can improve solution quality by 50 %.