Search

In the context of Industry 4.0, many production scenarios utilise sensors to monitor manufacturing processes, resulting in massive data that can be leveraged to build machine-learning models to improve production efficiency and quality. In semiconductor manufacturing, where many sensors are employed to monitor the production process, the resulting multi-sensor data poses challenges for constructing a virtual metrology (VM) model to assess wafer quality. Furthermore, building separate prediction models for each sensor can result in a loss of redundancy present in the multi-sensor data. Therefore, this paper proposes a multi-source AdaBoost with a cross-weight sampling method for predicting the critical dimension required in VM. Compared with the existing AdaBoost algorithm in regression, the proposed method adapts well to multi-source data by incorporating a cross-weight sampling distribution. Thus, when updating the sampling weight distribution, the impact of both individual and multi-source data is considered to re-sample high-quality observations. Based on the numerical results from the synthetic datasets and case study in semiconductor manufacturing, the proposed method outperforms benchmark methods. Additionally, owing to the redundancy of multi-sensor data, the proposed method demonstrates robust performance regardless of the noise level in the semiconductor VM data. [ABSTRACT FROM AUTHOR]

The semiconductor industry is a resource-intensive sector that heavily relies on energy, water, chemicals, and raw materials. Within the semiconductor manufacturing process, the diffusion furnace, ion implantation machine, and plasma etching machine exhibit high energy demands or operate at extremely high temperatures, resulting in significant electricity consumption, which is usually carbon-intensive. To address energy conservation concerns, the industry adopts batch production technology, which allows for the simultaneous processing of multiple products. The energy-efficient parallel batch scheduling problem arises from the need to optimise product grouping and sequencing. In contrast to existing heuristics, meta-heuristics, and exact algorithms, this paper introduces the Deep Q-Network (DQN) algorithm as a novel approach to address the proposed problem. The DQN algorithm is built upon the agent's systematic learning of scheduling rules, thereby enabling it to offer guidance for online decision-making regarding the grouping and sequencing of products. The efficacy of the algorithm is substantiated through extensive computational experiments. [ABSTRACT FROM AUTHOR]

Ferroelectric tunnel junctions (FTJs) utilizing hafnium zirconium oxide (HZO) have attracted interest as non-volatile memory for microelectronics due to ease of integration into back-end-of-line (BEOL) complementary metal oxide semiconductor fabrication. This work examines asymmetric electrode NbN/HZO/Nb devices with 7 nm thick HZO as FTJs in a memory structure, with an output resistance that can be controlled by read and write voltages. The individual FTJs are measured to have a tunneling electroresistance of 10 during the read state without significant filament conduction formation and reasonable ferroelectric performance. Endurance and remanent polarizations of up to 105 cycles and 20 μC/cm2, respectively, are measured and are shown to be dependent on the cycling voltage. Electrical measurements demonstrate how magnitude of the write pulse can modulate the high state resistance and the read pulse influences both resistance values as well as separation of resistance states. Then, by using two opposite switching FTJ devices in series, a programmable nonvolatile resistor divider is demonstrated. Measurements of these two FTJ unit memory cells show wide applicability to a BEOL microfabrication process for a re-readable, rewritable, and nonvolatile memory cell. [ABSTRACT FROM AUTHOR]

Radiation damage in semiconductor materials is a crucial concern for electronic applications, especially in the fields of space, military, nuclear, and medical electronics. With the advancements in semiconductor fabrication techniques and the trend of miniaturization, the quality of semiconductor materials and their susceptibility to radiation-induced defects have become more important than ever. In this context, machine learning (ML) algorithms have emerged as a promising tool to study minor radiation-induced defects in semiconductor materials. In this study, we propose a sensitive non-destructive technique for investigating radiation-induced defects using multivariate statistical analyses combined with Raman spectroscopy. Raman spectroscopy is a contactless and non-destructive method widely used to characterize semiconductor materials and their defects. The multivariate statistical methods applied in analyzing the Raman spectra provide high sensitivity in detecting minor radiation-induced defects. The proposed technique was demonstrated by categorizing 100–500 kGy irradiated GaAs wafers into samples with low and high irradiation levels using linear discrimination analysis ML algorithms. Despite the high similarity in the obtained Raman spectra, the ML algorithms correctly predicted the blind testing samples, highlighting the effectiveness of ML in defect study. This study provides a promising approach for detecting minor radiation-induced defects in semiconductor materials and can be extended to other semiconductor materials and devices. [ABSTRACT FROM AUTHOR]

Concerning the high-mixed nature in modern semiconductor manufacturing, the original 'tool-based' exponentially weighted moving average (EWMA) controller is proved to be unstable when the plant-model mismatches are diverse for different products, leading to large variations and lots of wastes. Hence, this paper proposes a general multi-product model for 'tool-based' EWMA control. The core of the model lies in the algorithm on how to estimate the noise disturbance accurately, i.e. the noise is considered as the sum of the following two parts: the estimated noise after the thread has been processed in the last cycle and the cumulative noise during the period when the thread is not on production between the two cycles. In this way, the production for a particular thread is isolated from that of others to avoid instability existing in the traditional tool-based EWMA control. Superiorities of the proposed model are demonstrated through theoretical analysis, simulation experiment and industrial case study. [ABSTRACT FROM AUTHOR]

In this paper, a practical relevant operational production planning problem in complex manufacturing systems is addressed. In this problem, lots are planned individually to provide a more detailed plan than approaches that only consider production quantities. A three-step approach, which is currently fully integrated and used in a Decision Support System, is then introduced. This work follows the one of Mhiri et al. [2018. "Heuristic Algorithm for a WIP Projection Problem at Finite Capacity in Semiconductor Manufacturing." IEEE Transactions on Semiconductor Manufacturing 31 (1): 62–75] who addressed this problem. We push the approach a step further by introducing new optimisation possibilities through new smoothing rules, whose performance is studied according to different indicators. Furthermore, we present the production planning process in which the decision support tool is embedded and how it bridges the gap between the upper and lower planning levels. [ABSTRACT FROM AUTHOR]

The lithography process in semiconductor dynamic random-access memory (DRAM) fabrication plants (fabs) is usually a major bottleneck, and reticle scheduling is complicated by process-specific constraints such as diversity of the product mix and rapidly changing deadlines. This study proposes a three-phase scheduling framework to solve the multi-objective reticle scheduling problem, reducing capacity loss, maintaining line balance and meeting customer demand. The genetic algorithm (GA) and the technique for order preference by similarity to ideal solution (TOPSIS) are used to optimize the scheduling with a trade-off among the multiple objectives. This study tests the proposed framework based on a real semiconductor DRAM (fab) in Taiwan, which has two unrelated parallel machines in its lithography process. The study finds that the proposed framework results in better scheduling reports, reduced computation times and improvements in the overall performance of lithography equipment. [ABSTRACT FROM AUTHOR]

Plasma etching is a crucial step in semiconductor manufacturing. High cleanliness and wafer‐to‐wafer reproducibility in the etching chamber are essential in order to successfully achieve nanometer‐sized integrated functions on the wafer. The trend toward the application of more aggressive plasma compositions leads to higher demands on the plasma resistance of the materials used in the etching chamber. Due to its excellent etch resistance, yttrium aluminum garnet Y3Al5O12 (YAG) is starting to replace established materials like SiO2 or Al2O3 in this kind of application. In this study, reactive spark plasma sintering (SPS) was used to manufacture highly dense YAG ceramics from the respective oxides. In addition, yttrium was replaced with heavier lanthanoids (Er, Lu), intending to investigate the role of the A‐site cation in the garnet type structure on the plasma erosion behavior. The produced materials were exposed to fluorine‐based etching plasmas mimicking the conditions in the semiconductor manufacturing apparatus and the erosion behavior was characterized by atomic force microscopy (AFM), secondary ion mass spectrometry (SIMS), transmission electron microscopy (TEM), and profilometry. The induced chemical gradient in the samples is limited to a few nanometers below the surface, which makes its characterization challenging. For advanced analysis, we developed a correlative characterization method combining SIMS and scanning TEM (STEM)–energy‐dispersive spectroscopy (EDS) enabling us to examine the structural and chemical changes in the reaction layer locally resolved. In the case of lanthanoid aluminates, an altered reaction layer and reduced fluorine penetration compared to YAG were found. However, a correlation between the characteristics of the induced chemical gradient and the determined physical erosion rates was not evident. [ABSTRACT FROM AUTHOR]

Quantum dot (QD) color conversion‐based displays have emerged as one of the most promising next‐generation devices due to their superior emission properties in terms of color expression. To date, however, existing QD color conversion layer (QD‐CCL) technologies have suffered from low luminance and power efficiency, mainly due to significant light absorption by the bank structure. Here, the color conversion efficiency of QD‐CCL has been significantly enhanced by fabricating a highly reflective metal layer on the side surface of the bank structure. Using a high‐aspect‐ratio silver reflector fabricated through a secondary sputtering lithographic technique involving argon ion bombardment, the fabricated QD‐CCL is combined with a blue organic light‐emitting diode (OLED) serving as a light source. As a result, light recycling from the reflector significantly enhances color conversion efficiency and luminance by up to 4.60‐fold and 4.29‐fold, respectively. Optical simulation reveals that higher pixel resolution provides greater reflection probabilities during extraction. This photomask‐free approach is not only simple but also highly compatible with existing semiconductor fabrication processes, making it a viable commercial alternative for all color‐converting structures that utilize light‐emitting materials with omnidirectional emission characteristics. [ABSTRACT FROM AUTHOR]

Physical unclonable functions (PUFs) have emerged as a favorable hardware security primitive, they exploit the process variations to provide unique signatures or secret keys amidst other critical cryptographic applications. CMOS-based PUFs are the most popular type, they generate unique bit strings using process variations in semiconductor fabrication. However, most existing CMOS PUFs are found to be vulnerable to modeling attacks based on machine learning (ML) algorithms. Memristors leveraging nanotechnology fabrication processes and highly nonlinear behavior became an interesting alternative to the existing CMOS-based PUF technology, introducing cryptographic and resilient random outputs. Memristor-based PUFs are emerging due to the inherent randomness at both the memristor level due to the cycle-to-cycle (C2C) programming variation of the device and the fabrication process level such as the cross-sectional area and variations. Our study focuses on building a machine learning analysis and attack framework of tools on C u / H f O 2 - x / p + + S i memristor-based PUF (MR-PUF). Our objective is to test the resiliency of the security margins of the presented PUF using machine learning analysis tools, on-top of holistic NIST cryptographic randomness testing initially provided, to provide a high level of certainty in predicting the randomness output of the verified Memrister-based PUF. Our main contribution is a holistic study that focuses on attacking the randomness output resiliency based on building randomness predictors using Logistic Regression (LR), Support Vector Machine (SVM), Gaussian Mixture Models (GMM), K-means, K-means + + , Random Forest, XGBoost and LSTM, within efficient time, and data complexity. Our results yield low accuracy and ROC results of within 0.49 - 0.52 and 0.49 - 0.52 respectively, indicating failure in predicting random data demonstrates efficient randomness prediction resiliency of the MR-PUF. The efficient time and data complexities of these attacks illustrated in this study are yielded to be linear and quadratic resulting in attack execution time in seconds and 5032 training samples combined with 2157 testing samples to verify the randomness of PUF. [ABSTRACT FROM AUTHOR]

The extreme ultraviolet (EUV) photolithography process is a cornerstone of semiconductor manufacturing and operates under demanding precision standards realized via nanometer-level overlay (OVL) error modeling. This procedure allows the machine to anticipate and correct OVL errors before impacting the wafer, thereby facilitating near-optimal image exposure while simultaneously minimizing the overall OVL error. Such models are usually high dimensional and exhibit rigorous statistical phenomena such as collinearities that play a crucial role in the process of tuning their parameters. Ordinary least squares (OLS) is the most widely used method for parameters tuning of overlay models, but in most cases it fails to compensate for such phenomena. In this paper, we propose the usage of ridge regression, a widely known machine learning (ML) algorithm especially suitable for datasets that exhibit high multicollinearity. The proposed method was applied in perturbed data from a 300 mm wafer fab, and the results show reduced residuals when ridge regression is applied instead of OLS. [ABSTRACT FROM AUTHOR]

To increase productivity, more sophisticated cluster tools are developed. To achieve this, one of the ways is to increase the number of spaces in a process module (PM) and the number of fingers on a robot arm as well, leading to a cluster tool with multi-space PMs and a multi-finger-arm robot. This paper discusses the scheduling problem of cluster tools with four-space PMs and a four-finger-arm robot, a typical tool with multi-space PMs and a multi-finger-arm robot adopted in modern fabs. With two arms in such a tool, one is used as a clean one, while the other is used as a dirty one. In this way, wafer quality can be improved. However, scheduling such cluster tools to ensure the residency time constraints is very challenging, and there is no research report on this issue. This article conducts an in-depth analysis of the steady-state scheduling for this type of cluster tools to explore the effect of different scheduling strategies. Based on the properties, four robot task sequences are presented as scheduling strategies. With them, four linear programming models are developed to optimize the cycle time of the system and find feasible schedules. The performance of these strategies is dependent on the activity parameters. Experiments are carried out to test the effect of different parameters on the performance of different strategies. It shows that, given a group of parameters, one can apply all the strategies and choose the best result obtained by one of the strategies. [ABSTRACT FROM AUTHOR]

Due to its capacity to achieve nanometre-scale machining and lithography, a focused ion beam (FIB) is an extended tool for semiconductor device fabrication and development, in particular, for diamond-based devices. However, some technological steps are still not fully optimized for its use. Indeed, ion implantation seems to affect the crystalline structure and electrical properties of diamond. For this study, a boron-doped ([B] ∼ 1017 atoms·cm−3) diamond layer grown by chemical vapour deposition was irradiated using Ga+ by FIB, with 1 nA current and 5, 20, and 30 keV of acceleration voltage. The Ga+ implanted diamond layer has been analysed through cathodoluminescence (CL) and scanning transmission electron microscopy (STEM)-related techniques. The beam penetration depth has been simulated by Monte Carlo calculations of both Ga+ (FIB) and e− (CL) beams at different energies. The comparative CL analysis of the layer as-grown and after implantation revealed peaks related to defects, such as A band, H3 centre, and defects present in the green band region. The STEM studies for the 30 keV implanted sample showed that the diamond lattice is affected by the damage, evidencing amorphisation in the layer with a sp2/sp3 ratio of 1.37, estimated by electron energy loss spectroscopy. Therefore, this study highlights the effects of the Ga+ implantation on the optical and structural characteristics of diamond, using different methods. [ABSTRACT FROM AUTHOR]

Polyurethane foam (PUF) pads are widely used in semiconductor manufacturing, particularly for chemical mechanical polishing (CMP). This study prepares PUF composites with microcrystalline cellulose (MCC) and nanocrystalline cellulose (NCC) to improve CMP performance. MCC and NCC were characterized using scanning electron microscopy (SEM) and X-ray diffraction (XRD), showing average diameters of 129.7 ± 30.9 nm for MCC and 22.2 ± 6.7 nm for NCC, both with high crystallinity (ca. 89%). Prior to preparing composites, the study on the influence of the postbaked step on the PUF was monitored through Fourier-transform infrared spectroscopy (FTIR). After that, PUF was incorporated with MCC/NCC to afford two catalogs of polyurethane foam composites (i.e., PUFC-M and PUFC-N). These PUFCs were examined for their thermal and surface properties using a differential scanning calorimeter (DSC), thermogravimetric analysis (TGA), dynamic mechanical analyzer (DMA), and water contact angle (WCA) measurements. Tgs showed only slight changes but a notable increase in the 10% weight loss temperature (Td10%) for PUFCs, rising from 277 °C for PUF to about 298 °C for PUFCs. The value of Tan δ dropped by up to 11%, indicating improved elasticity. Afterward, tensile and abrasion tests were conducted, and we acquired significant enhancements in the abrasion performance (e.g., from 1.04 mm/h for the PUF to 0.76 mm/h for a PUFC-N) of the PUFCs. Eventually, we prepared high-performance PUFCs and demonstrated their capability toward the practical CMP process. [ABSTRACT FROM AUTHOR]

The presence of missing components in incomplete instances precludes a kernel-based model from incorporating partially observed components of incomplete instances and computing kernels, including Gaussian kernels that are extensively used in machine learning modeling and applications. Existing methods with Gaussian kernels to handle incomplete data, however, are based on independence among variables. In this study, we propose a new method, the expected Gaussian kernel with correlated variables, that estimates the Gaussian kernel with incomplete data, by considering the correlation among variables. In the proposed method, the squared distance between two instance vectors is modeled with the sum of the correlated squared unit-dimensional distances between the instances, and the Gaussian kernel with missing values is obtained by estimating the expected Gaussian kernel function under the probability distribution for the squared distance between the vectors. The proposed method is evaluated on synthetic data and real-life data from benchmarks and a case from a multi-pattern photolithographic process for wafer fabrication in semiconductor manufacturing. The experimental results show the improvement by the proposed method in the estimation of Gaussian kernels with incomplete data of correlated variables. [ABSTRACT FROM AUTHOR]

Photoacoustic detection has shown excellent performance in measuring thickness and detecting defects in metal nanofilms. However, existing research on ultrafast lasers mainly focuses on using picosecond or nanosecond lasers for large-scale material processing and measurement. The theoretical study of femtosecond laser sources for photoacoustic nondestructive testing (NDT) in nanoscale thin film materials receives much less emphasis, leading to a lack of a complete physical model that covers the entire process from excitation to measurement. In this study, we developed a comprehensive physical model that combines the two-temperature model with the acoustic wave generation and detection model. On the basis of the physical model, we established a simulation model to visualize the ultrafast lasermaterial interaction process. The damage threshold of the laser source is determined, and the effect of key parameters (laser fluence, pulse duration, and wavelength) for AlCu nanofilms on the femtosecond photoacoustic NDT process is discussed using numerical results from the finite element model. The numerical results under certain parameters show good agreement with the experimental results. [ABSTRACT FROM AUTHOR]

Artificial intelligence (AI) is increasingly used to solve multi‐objective problems and reduce the turnaround times of semiconductor processes. However, only brief AI explanations are available for process/device/circuit engineers to provide holistic feedback on the manufactured results. Herein, linear/nonlinear variation decomposition (LVD/NLVD) of neural networks is demonstrated to quantitatively evaluate the influence of unit processes on the figure of merit (FoM) and co‐analyze the unit process influences with device characteristic behaviors. The NLVD can evaluate the output variation from each input of neural networks in an individual sample, although neural networks are not available in an analytic form. The NLVD is successfully verified by confirming that a) the output and summation of all decomposed output variations perfectly coincide and b) the process influences on the FoM are decomposed to 6.01–54.86% more accurately compared with those of LVD in 1Y nm node dynamic random‐access memory test vehicles with a baseline and split tests introducing high‐k metal gates with a minimum gate length of 1 A nm node for further node scaling. The approaches identify defective processes and defect mechanisms in each sample and wafer, which enhance causal analyses for individual cases in diverse fields based on regression artificial neural networks. [ABSTRACT FROM AUTHOR]

Both nitric oxide (NO) and nitrogen dioxide (NO2) gasses are toxic to humans but are commonly found in industrial settings such as semiconductor manufacturing sites. Due to the spontaneous oxidation of NO to NO2 under ambient conditions, individuals working with NO may in fact be exposed to both gasses in the case of an accidental release. Unfortunately, most safety materials provided to NO users do not address the potential for associated NO2 toxicity, and, until now, models developed to predict health consequences following a release of NO have not appropriately considered the oxidation kinetics nor the toxicity of both NO and NO2 in their assessments. This paper describes an improved multi‐module model that addresses these limitations and explores whether facilities using NO should consider adopting measures that can mitigate the simultaneous health effects of both gasses. The model predicts the morbidity (intoxication/injury), mortality (death), and treatment outcomes that may arise following an industrial NO release by first calculating the doses of both NO and NO2 received by exposed individuals and then applying newly defined toxicity parameters for NO and NO2 to assign dose‐dependent probabilities for the onset of intoxication and/or death and the ability of appropriate treatment(s) to save lives. Modeling results indicate low risk to worker health in the likeliest release scenarios while identifying less likely situations that carry substantially higher risk. Moreover, these results indicate that risks to worker health can be mitigated with simple measures like maintaining reliable alarms, adequate ventilation, and on‐site supplies of methylene blue, as well as encouraging quick responses by personnel. With appropriate parameterization, the improved modeling framework is generalizable to any chemical release, especially multi‐hazard releases resulting from the conversion of one toxic compound into another under likely environmental conditions. By directly addressing the toxicities of multiple compounds, the improved model presents a more realistic picture of the potential health consequences of a chemical release. This generalizable framework for modeling of multi‐hazard chemical releases can inform preparedness and risk mitigation strategies for NO release events. Given the spontaneous oxidation of NO to NO2 under ambient conditions, individuals working with NO in industrial contexts may be exposed to both NO and NO2. This paper describes an improved multi‐module model that considers facility ventilation, oxidation kinetics, and the toxicity of both NO and NO2 and applies the model to explore whether facilities using NO should consider adopting measures that can mitigate the potential for adverse health effects from both gasses during a NO release. [ABSTRACT FROM AUTHOR]

Silicon monoxide nanowires have become ubiquitous in twenty-first century technology due to their superior photoelectric properties, widely utilized in semiconductor manufacturing and emerging energy fields. While numerous studies have concentrated on tailoring material properties, scant attention has been paid to the impact of process parameters on the productivity of silicon monoxide during synthesis. Through response surface methodology, the effects of holding time, heating temperature, and Si/SiO2 molar ratio on the volatilization ratio of silicon monoxide have been investigated. The optimization of silicon monoxide nanowire preparation via vacuum silicothermic reduction was aimed at achieving maximum efficiency. According to the established mathematical model, the volatilization ratio of silicon monoxide reaches 92.057% when the holding time is 126 min, the heating temperature is 1663 K, and the molar ratio is 1. A comprehensive analysis revealed that temperature is the most significant factor affecting silicon monoxide volatilization among the studied parameters. Furthermore, the reduction slag of the system was characterized and analyzed. The results indicate that vacuum conditions can lower the starting temperature of the silicothermic reduction, thereby promoting the volatilization of SiO(g). [ABSTRACT FROM AUTHOR]

Fluoroelastomer (FKM), a thermosetting elastomer, stands out with its excellent friction characteristics and exceptional attributes of chemical, vacuum, and heat resistance, setting it apart from typical rubber materials. Leveraging these qualities, FKM emerges as the optimal material for friction pads used in extreme environments requiring high temperatures and vacuum, such as space and semiconductor processes. In this study, we introduce micropatterned bolt-inserted FKM pads (BIFP) designed to enhance friction, productivity, and usability. To produce BIFPs with improved friction characteristics, various microstructures were designed and tested for their friction performance. Additionally, to meet the demand for extended durability, we conducted 500,000-cycle repetitive tests. In terms of performance, durability, and usability, the developed BIFPs will be utilized widely, especially in industrial settings requiring harsh operating conditions. [ABSTRACT FROM AUTHOR]

In semiconductor manufacturing, classification of defect patterns in wafer bin maps (WBMs) helps engineers detect process failures and identify their causes. In recent studies on WBMs, convolutional neural networks (CNNs) have demonstrated effective classification performance based on their high expressive power. However, previous studies have implicitly assumed that the labels of WBMs used for training CNNs are correct, even though labels are often incorrect. When trained on mislabeled data, CNNs with standard cross-entropy loss can easily overfit mislabeled samples, leading to poor generalization for testing data. To overcome this issue, we propose a novel training algorithm called sample bootstrapping. Sample bootstrapping identifies which samples have clean or noisy labels by using a two-component beta mixture model, and measures the uncertainty of each identified label. Then, samples with low uncertainty of their estimated labels are selected to build mini-batches via weighted random sampling. Finally, CNNs are trained on the selected mini-batches with dynamic bootstrapping loss. In this manner, we can correct only the samples that are highly likely to have noisy labels and prevent the risk of false correction of correctly labeled samples. Experiments on simulated and real datasets demonstrate the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

Defect pattern classification of wafer bin maps (WBMs) assists in identifying the causes of semiconductor manufacturing process failures, thus contributing to the search for appropriate solutions. However, new previously unobserved defect patterns may be generated in real applications, which require classification methods with out-of-distribution (OOD) pattern detection capabilities. In this paper, we propose a novel method called a spectral-normalized neural radial basis function (SNRBF) network to classify defect patterns in WBMs while simultaneously detecting OOD defect patterns by quantifying the predictive uncertainty of a deep neural network. The SNRBF network adds spectral normalization to the weights in each layer of a deep neural network and replaces the output layer with a radial basis function kernel, facilitating high-quality uncertainty quantification and efficient computation. Additionally, we propose a novel regularization method based on singular values of the representation matrix to improve uncertainty quantification. Furthermore, we propose a new noise-filtering method capable of effectively eliminating random defects in WBMs by considering various defect pattern sizes. The proposed method is evaluated on a real WBM dataset and is demonstrated to outperform other competing methods in OOD detection while presenting competitive classification performance. [ABSTRACT FROM AUTHOR]

Novel sensing applications benefit from multifunctional nanomaterials responsive to various external stimuli such as mechanics, electricity, light, humidity, or pollution. While few such materials occur naturally, the careful design of synergized nanomaterials unifies the cross-coupled properties which are weak or absent in single-phase materials. In this study, 2D MoS2 integrated with ultrathin dielectric oxide layers forms hetero-nanostructures with significant impacts on carrier transport. The ternary TiO2/MoS2/ZnO hetero-nanostructures, along with their individual properties, improve the performance of multifunctional sensing devices. The synthesized hetero-nanostructure exhibits a responsivity of up to 16 mA/W to 700 nm light and responds to 5 ppm ammonia gas at room temperature. These enhancements are attributed to interface charge transfer and photogating effects. The ternary TiO2/MoS2/ZnO hetero-nanostructure is compatible with existing semiconductor fabrication technologies, making it feasible to integrate into flexible, lightweight semiconductor devices and circuits. These results may inspire new photodetectors and sensing devices based on two-dimensional (2D) layered materials for IoT applications. [ABSTRACT FROM AUTHOR]

When SiO 2 films undergo oxidation on 4H–SiC, a distinctive crystalline phase, cristobalite, emerges at elevated temperatures. We deeply explored the growth mechanism of the crystallite, focusing on its dependence on the silicon sublimation process (SSO). Activation energy calculations confirmed that the oxidation after SSO above 1350 °C exhibits a lower activation energy. The reduced activation energy suggests the elimination of the solid-phase diffusion process during oxidation. We devised a controllable growth process for the quartz phase to achieve a seamless transition from 0 to 100 % in the area proportion of cristobalite in SiO 2. This strategic approach facilitates the precise preparation of oxide layers and holds potential applications in semiconductor device manufacturing. [ABSTRACT FROM AUTHOR]

In order to improve quality assurance in wafer manufacturing, there are strict process requirements. Besides the well-known residency time constraints (RTCs), a dual-arm cluster tool also requires each robot arm to execute a specific set of tasks. We call such a tool an arm task-constrained dual-arm cluster tool (ATC-DACT). To do this, one of the arms is identified as the dirty one and the other as the clean one. The dirty one can deal with raw wafers, while the clean one can deal with processed wafers. This requirement raises a new problem for scheduling a cluster tool. This paper discusses the scheduling problem of ATC-DACTs with RTCs. Due to the arm task constraints, the proven, effective swap strategy is no longer applicable to ATC-DACTs, making the scheduling problem difficult. To address this problem, we explicitly describe the robot waiting as an event and build a Petri net (PN) model. Then, we propose a hybrid task sequence (HTS) as an operation strategy by combining the swap and backward strategies. Based on the HTS, the necessary and sufficient conditions for schedulability are established; also, a linear programming model is developed. We then develop an algorithm using these results to optimally schedule the system. Industrial case studies demonstrate the application of this method. [ABSTRACT FROM AUTHOR]

Hydrazine‐assisted water electrolysis presents a promising approach toward energy‐efficient hydrogen production. However, the progress of this technology is hindered by the limited availability of affordable, efficient, and durable catalysts. In this study, a feasible strategy is proposed for interface modulation that enables efficient hydrogen evolution and hydrazine oxidation through the construction of n‐type semiconductor heterostructures. The metal–semiconductor contacts are rationally designed using ruthenium nanoclusters and a range of metal oxide (M–O) semiconductor heterostructures, including p‐type semiconductor substrates (NiO, Co3O4, NiCo2O4, CuCo2O4, ZnCo2O4) and n‐type semiconductor substrate (Fe2O3). Intriguingly, Ru nanoclusters supported on p‐type M–O substrates induce a transition from p‐type M–O to n‐type M‐O/Ru. The design of n‐type semiconductor heterostructures can significantly reduce space‐charge regions and increase charge carrier concentration, thereby improving the electrical conductivity of electrocatalysts. Moreover, Ru atoms can serve as highly efficient active sites for hydrogen evolution reaction and hydrazine oxidation reaction. The NiO/Ru heterostructure can drive current densities of 10 and 100 mA cm−2 with only 0.021 and 0.22 V cell voltages for hydrazine‐assisted water electrolysis. This work provides new insights for the development of highly efficient semiconductor catalysts, enabling energy‐saving hydrogen production. [ABSTRACT FROM AUTHOR]

Wet etching of Silicon nitride (Si3N4) by phosphoric acid (H3PO4) is one of the key steps in the semiconductor fabrication process, especially for three‐dimensional integrated circuits (ICs). Unfortunately, few systematic studies of phosphoric acid etching of Silicon nitride are available. Summarizing how various factors affect etching and determining the reaction mechanism behind the etching process will help further optimize the etching process. In this work, how the boiling point of hydrous phosphoric acid solutions depends on phosphoric acid concentration is firstly investigated. Next, the etching of over 100 samples of Silicon nitride by H3PO4 at different concentrations and temperatures is systematically studied. The results show that the etching rate increases nearly linearly in both temperature and H3PO4 concentration, while temperature play more important role as compared to concentration. For each concentration, the reaction rate is maximal at the boiling point of the hydrous H3PO4 solution. Moreover, the etching rate of 85 wt % phosphoric acid is analyzed using the Arrhenius equation, yielding an apparent activation energy Ea=57.28 KJ/mol. Finally, a mechanism of Silicon nitride etching in phosphoric acid is proposed which explains the prior studies well. This work contributes to precisely control semiconductor processing for three‐dimensional ICs. [ABSTRACT FROM AUTHOR]

The global industrial development and increase in the number of transportation vehicles, such as automobiles and ships, have led to a steady increase in the issues related to greenhouse gas emissions. NO2 is a greenhouse gas emitted in large quantities from automobiles and factories, and its emission is unavoidable in the modern world. Therefore, a sensor capable of precise detection of NO2 is required. The most commonly reported types of NO2 sensors are those based on metal oxides. However, their operation at room temperature is impossible owing to their high-temperature operating characteristics, and therefore, a heater must be designed inside or installed outside the sensor for heating. Meanwhile, NO2 sensors based on PbS quantum dots (QDs) are advantageous as they can operate at room temperature and can be easily manufactured through a solution process rather than a complicated semiconductor process. Herein, a NO2 sensor was fabricated by doping PbS QDs with poly(3-hexylthiophene) (P3HT). The as-developed sensor exhibited high responsivity to 100–0.4-ppm NO2 gas with a resolution of 200 ppb owing to the stability of the thin film and high hole mobility of P3HT. [ABSTRACT FROM AUTHOR]

Energy dissipation through physical downscaling towards more complex types of memory and logic devices, loss of ultrapure water and consumption of large amounts of (toxic) chemicals for wafer cleaning processes, as well as high thermal budget of solid-state synthesis and thin film growth of standard semiconductors including the use of rare earth elements – all this poses great challenges for semiconductor materials science and technology. Therefore, research and development of alternative methods for micro- and nanofabrication and chemical functionalization of a new type of resource- and energy-efficient semiconductors as the core component of every computer chip is crucial. One of the promising opportunities is the transformation of today's complementary metal-oxide-semiconductor (CMOS) electronics into ecofriendly and neuroinspired electronics driven by molecular design and multi-level switching mechanisms at room temperature. The sustainable chemical technology of electron transport and switching materials in semiconductor manufacturing and the development of devices with new unconventional nanophysics, improved performance, and augmented functionalities (beyond-CMOS and More-than-Moore) is becoming increasingly important in the context of a gradual transition to a future-oriented concept of Internet of Everything (IoE). In this article, we focus on the technological significance of semiconductor preparation from single-source (molecular) precursors and the prospect of functionalizing semiconductors using DNA origami nanotechnology and stimuli-responsive metal–oxygen cluster ions such as polyoxometalates (POMs). We also describe the advanced characterization of these qualified molecular systems by soft X-rays. We emphasize the technical relevance of using solution-based methods for the bottom-up preparation of novel and hybrid semiconductors as well as their challenging scalability and the compatibility of methods of molecular technology with lithography-based mass production. Our article aims to contribute to the achievement of the United Nations' Sustainable Development Goal 9 (Industry, Innovation and Infrastructure). [ABSTRACT FROM AUTHOR]

Various data types generated in the semiconductor manufacturing process can be used to increase product yield and reduce manufacturing costs. On the other hand, the data generated during the process are collected from various sensors, resulting in diverse units and an imbalanced dataset with a bias towards the majority class. This study evaluated analysis and preprocessing methods for predicting good and defective products using machine learning to increase yield and reduce costs in semiconductor manufacturing processes. The SECOM dataset is used to achieve this, and preprocessing steps are performed, such as missing value handling, dimensionality reduction, resampling to address class imbalances, and scaling. Finally, six machine learning models were evaluated and compared using the geometric mean (GM) and other metrics to assess the combinations of preprocessing methods on imbalanced data. Unlike previous studies, this research proposes methods to reduce the number of features used in machine learning to shorten the training and prediction times. Furthermore, this study prevents data leakage during preprocessing by separating the training and test datasets before analysis and preprocessing. The results showed that applying oversampling methods, excluding KM SMOTE, achieves a more balanced class classification. The combination of SVM, ADASYN, and MaxAbs scaling showed the best performance with an accuracy and GM of 85.14% and 72.95%, respectively, outperforming all other combinations. [ABSTRACT FROM AUTHOR]

In this study, we conduct a comparative analysis of single-beam laser annealing (SBLA) and dual-beam laser annealing (DBLA) techniques for semiconductor manufacturing. In the DBLA approach, two laser beams were precisely aligned to simultaneously heat a phosphorus-doped silicon (Si) wafer. The main objective was to investigate the impact of the two annealing techniques on the electrical properties, crystalline structure, and diffusion profile of the treated phosphorus-doped Si at equivalent laser powers. Both SBLA and DBLA improved the electrical properties of the phosphorus-doped Si, evidenced by increased carrier concentration and reduced carrier mobility. Additionally, the crystalline structure of the phosphorus-doped Si showed favorable modifications, with no defects and improved crystallinity. While both SBLA and DBLA produced similar phosphorus profiles with no significant redistribution of dopants compared to the as-implanted sample, DBLA achieved a higher activation ratio than SBLA. Although the results suggest improved dopant activation with minimal diffusion, further studies are needed to clearly confirm the effect of DBLA on dopant activation and diffusion. [ABSTRACT FROM AUTHOR]

The article discusses the rationale for industrial policy in the semiconductor industry, focusing on the European context. It highlights the challenges faced by the European semiconductor industry, the importance of semiconductors for achieving climate and digital transformation goals, and the need for targeted industrial policy to address market failures. The text also explores the potential benefits of industrial policy, the role of political economy in shaping policies, and the importance of adaptive strategies that allow for the phasing out of unsuccessful projects. The article emphasizes the need for careful design and consideration of political constraints in industrial policy decisions. [Extracted from the article]

Conventional statistical process control tools monitor either continuous or count data but rarely both simultaneously. While process data are becoming increasingly complex, there will be more data points containing both continuous and count information. In the case of mixed continuous and count data with unknown distributions, the traditional parameter control chart cannot be used to monitor them. It is proposed in this paper a novel nonparametric EWMA control chart to monitor mixed continuous and count data. The mixed continuous and count data are first transformed into categorical data, and then a log-linear model is utilized to analyze correlations between variables, followed by the construction of an EWMA statistic that is used to monitor mixed continuous and count data. Next, the proposed control chart is compared with several improved control charts for monitoring mixed continuous and count data. Based on the numerical simulation results, the control chart presented in this paper provides a superior method of detecting alarm signals in the process compared to some improved control charts. Finally, the proposed control chart is demonstrated to be effective and applicable using the semiconductor manufacturing process dataset from the UC Irvine Machine Learning Repository. [ABSTRACT FROM AUTHOR]

Autonomous preventive maintenance (APM) is important for modern semiconductor manufacturing equipment to enhance yield. APM involves detecting various signals from the equipment's components and/or the workpiece to monitor equipment health in real time and perform adequate maintenance. This paper proposes a vibration sensor module with noncontact sensing capability that is small, simple, and highly sensitive. The module's hardware includes two sensor components, which are a combination of a small infrared light- emitting device (IR LED) and a photodetector (PD). The module performs sensing based on the principle of beating. Compared with conventional direct sensing, the proposed module achieves a fourfold improvement in sensitivity and a 3.5-improvement in signal-to- noise ratio. Moreover, the small size of the IR LED and PD die chips can enable miniaturization. If a sensor system combining the proposed sensor module with a microcontroller unit and a transceiver is integrated into semiconductor manufacturing equipment, independent and remote APM can be easily and widely realized. [ABSTRACT FROM AUTHOR]

Overlay (OVL) is one significant performance indicator for the lithography process control in semiconductor manufacturing. The accuracy of the OVL metrology is extremely critical for guarantee the lithography quality. Currently, diffraction-based overlay (DBO) is one of the mainstream OVL metrology techniques. Unfortunately, the accuracy of the DBO metrology is largely affected by the defect features of the OVL target. Therefore, there is a strong need to investigate the impacts of these target defects on the DBO metrology performance. However, efficiently investigating the statistical and interactive impacts of various DBO target defects remains challenging. This study aims to address this issue through proposing an intelligent sensitivity analysis approach. A cumulative distribution based global sensitivity analysis (GSA) method is utilized to assess the nonlinear influences of multiple defects in the OVL target on the DBO inaccuracy. The scenarios with both known and unknown distributions of the OVL target defects are considered. For the former, a neural network driven forward model is constructed for fast calculating the optical diffraction responses to accelerate the GSA process. For the latter, another neural network based inverse model are built for efficiently estimating the distribution of the target defects. Finally, a series of simulation experiments are conduct for typical DBO targets with multiple common defect features. The results demonstrate the effectiveness and robustness of the proposed approach as well as give valuable insights into the DBO defect analysis. Our study provides a strong tool to assist the practitioners in achieving intelligent and efficient DBO analysis and thus in enhancing OVL metrology performance. [ABSTRACT FROM AUTHOR]

Surface defect detection on wafers is crucial for quality control in semiconductor manufacturing. However, the complexity of defect spatial features, including mixed defect types, large scale differences, and overlapping, results in low detection accuracy. In this paper, we propose a CC-De-YOLO model, which is based on the YOLOv7 backbone network. Firstly, the coordinate attention is inserted into the main feature extraction network. Coordinate attention decomposes channel attention into two one-dimensional feature coding processes, which are aggregated along both horizontal and vertical spatial directions to enhance the network's sensitivity to orientation and position. Then, the nearest neighbor interpolation in the upsampling part is replaced by the CAR-EVC module, which predicts the upsampling kernel from the previous feature map and integrates semantic information into the feature map. Two residual structures are used to capture long-range semantic dependencies and improve feature representation capability. Finally, an efficient decoupled detection head is used to separate classification and regression tasks for better defect classification. To evaluate our model's performance, we established a wafer surface defect dataset containing six typical defect categories. The experimental results show that the CCDe-YOLO model achieves 91.0% mAP@0.5 and 46.2% mAP@0.5:0.95, with precision of 89.5% and recall of 83.2%. Compared with the original YOLOv7 model and other object detection models, CC-De-YOLO performs better. Therefore, our proposed method meets the accuracy requirements for wafer surface defect detection and has broad application prospects. The dataset containing surface defect data on wafers is currently publicly available on GitHub (https://github.com/ztao3243/Wafer-Datas.git). [ABSTRACT FROM AUTHOR]

The semiconductor industry is being radically impacted by the placing of greater emphasis on the development of hetero-devices and systems that will act as essential drivers for a wide spectrum of technological applications. The introduction of new materials and their integration with currently used materials are projected to replace integrated circuitry (IC) design and device scaling as the key enablers to the realization of improved device performance and larger density gains. Yet material selection has been constrained by existing fabrication process technology. To date, fabrication processes have dictated material selection by limiting chemical sources or precursors to those that match existing process tools associated with chemically based vapor phase processes and their variants, which in turn limits material compositions in ICs. The processing and integration of new materials compositions and structures will require the introduction of new deposition and etching processes, and manufacturing worthy tool designs and associated protocols that provide new methods for atomic-level control. To this end, a novel manufacturing paradigm is presented comprising a method and system for real-time, closed-loop monitoring and control of synthesis, supply, and consumption of precursors in process intensification techniques including chemical vapor deposition (CVD), atomic layer deposition (ALD), atomic layer etching (ALE), and other IC manufacturing processes. This intelligent automated manufacturing approach is consistent with a central component of the semiconductor industry's recent adoption of Industry 4.0., including vertical integration of IC manufacturing through robotization, artificial intelligence, and cloud computing. Furthermore, the approach eliminates several redundant steps in the synthesis, handling, and disposal of source precursors and their byproducts for CVD, ALD, ALE and other chemically based manufacturing processes, and thus ultimately lowers the manufacturing cost for both conventional and new IC materials. Further, by eliminating the issues associated with precursor thermal, chemical, and pyrophoric instabilities, this new paradigm enables the deposition of a myriad of new thin-film materials and compositions for IC applications that are practically unattainable with existing precursors. Preliminary and planned demonstration examples for the generation and deposition of highly toxic and unstable source precursors are provided. [ABSTRACT FROM AUTHOR]

Within the domain of semiconductor fabrication, which entails progressively complex patterning steps, the significance of plasma stripping processes, particularly to achieve the effective stripping of photoresist (PR) without damaging the underlying substrates via uniform gas distribution across 300 mm wafers, cannot be overstated. The efficacy of plasma stripping is influenced by the design of the components of the process chamber, which is critical for advancing semiconductor manufacturing technologies. In this study, we elucidated the influence of the design of a process chamber, particularly the showerhead nozzle angles, on the plasma chemical reactions of radicals emanating from conventional PR equipment using computational fluid dynamics simulations. We assessed the impact of the showerhead design, which incorporated divergent or angled gas supply nozzles, on the distribution of the supply gas across the wafer within the process chamber. Five distinct angles of showerhead nozzles were investigated, and the uniformity of the oxygen mass fractions was optimized for nozzle angles of 45° and 60°. Additionally, the factors contributing to the low uniformity in a showerhead design were identified and design components were modified, resulting in a 16% improvement in uniformity. This study delineated the relationship between the nozzle position, mass fraction, and flow streamlines, thereby establishing the critical role of equipment design for enhancing the plasma stripping process. Furthermore, this principle transcends the realm of stripping equipment design; it is anticipated that incorporating flow dynamics simulations for designing process components will enhance the overall process performance. [ABSTRACT FROM AUTHOR]

This paper provides a systematic review of research works on deterministic scheduling of diffusion furnaces (D-SDF) in the semiconductor manufacturing industry. After our screening process, we have identified 72 research articles published during the period 1992 to 2021 in various journals, conference proceedings, etc. This study proposes various classification schemes to systematically organise all the identified studies and to get micro-level details of D-SDF research. Further, various simple meta-analyses in the form of summary counts and percentages are carried out w.r.t (a) proposed classification schemes, and (b) various parameters such as data source, the maximum number of jobs in the instances, benchmark procedure considered, number of articles published, number of contributed authors over the years, top researchers in terms of number of publications, highly cited articles, publication outlets and publishers. From the detailed review analyses, this study suggests/shows future research opportunities/unexplored research problems considering problem configuration/objective(s)/solution methodologies in D-SDF research. Further, mapping of all references of 72 papers on D-SDF w.r.t. proposed classifications schemes would ease new researchers in D-SDF, in multiple perspectives. Finally, meta-analyses presented based on proposed classification schemes and various parameters, considered related to D-SDF problem, provide many important inferences to related researchers. [ABSTRACT FROM AUTHOR]

The formation of a double S-scheme heterojunction in the MoS 2 @MoO 3 /(Cu+/g-C 3 N 4) ternary composites is responsible for preserving strong reduction ability of photogenerated electrons on the Cu+/g-C 3 N 4 and strong oxidation ability of photogenerated holes on the MoO 3 , which efficiently participated in the PMS activation and the mineralization the RhB. [Display omitted] • The MoS 2 @MoO 3 /(Cu+/g-C 3 N 4) ternary composites was prepared by a simple calcination method. • The double S-scheme heterojunction was responsible for the enhanced performance in the degradation of RhB. • The reactive species for oxidizing RhB were identified by free-radical trapping experiments and ESR spectra. • The possible degradation pathway was proposed according to DFT calculations and LC-MS tests. The construction of semiconductor heterojunction is an effective way for charge separation in photocatalytic degradation of pollutants. In this study, a novel MoS 2 @MoO 3 /(Cu+/g-C 3 N 4) ternary composites (MMCCN) was prepared via a simple calcination method. The as-prepared composites exhibited exceptional performance in activating peroxymonosulfate (PMS) for the degradation of rhodamine B (RhB). The activity testing results indicated that 99.41 % of RhB (10 mg·L−1, 10 mL) was effectively removed by the synergistic effect of composites photocatalyst (0.1 g·L−1) and PMS (0.1 g·L−1) under visible light irradiation for 40 min. Its reaction rate constant exceeded that of Cu+/g-C 3 N 4 , MoO 3 and MoS 2 by a factor of 3.56, 17.30 and 11.73 times, respectively. The crystal structure, band gap and density of states (DOS) of the semiconductors were calculated according to the density functional theory (DFT). Free radical trapping tests and electron spin resonance spectroscopy validated that 1O 2 , O 2 − and h+ are primary reactive species participating in the decomposition of RhB. The ternary composites demonstrated good stability and maintained excellent degradation efficiency even across four reaction cycles. Furthermore, the activation mechanism and the intermediates produced during the decomposition course of RhB by MMCCN/PMS/vis system were analyzed and elucidated. A double S-scheme heterojunctions was responsible for efficient separation of photo-induced electron-hole pairs. This work presents a novel method in the construction of double S-scheme heterojunctions for PMS activation which is expected to find wide applications in wastewater treatment and environmental remediation. [ABSTRACT FROM AUTHOR]

Electronegative inductively coupled plasmas (ICPs) are used for conductor etching in the microelectronics industry for semiconductor fabrication. Pulsing of the antenna power and bias voltages provides additional control for optimizing plasma–surface interactions. However, pulsed ICPs are susceptible to capacitive-to-inductive mode transitions at the onset of the power pulse due to there being low electron densities at the end of the prior afterglow. The capacitive (E) to inductive (H) mode transition is sensitive to the spatial configuration of the plasma at the end of the prior afterglow, circuit (matchbox) settings, operating conditions, and reactor configurations, including antenna geometry. In this paper, we discuss results from a computational investigation of E–H transitions in pulsed ICPs sustained in Ar/Cl2 and Ar/O2 gas mixtures while varying operating conditions, including gas mixture, pulse repetition frequency, duty cycle of the power pulse, and antenna geometry. Pulsed ICPs sustained in Ar/Cl2 mixtures are prone to significant E–H transitions due to thermal dissociative attachment reactions with Cl2 during the afterglow which reduce pre-pulse electron densities. These abrupt E–H transitions launch electrostatic waves from the formation of a sheath at the boundaries of the plasma and under the antenna in particular. The smoother E–H transitions observed for Ar/O2 mixture results from the higher electron density at the start of the power pulse due to the lack of thermal electron attaching reactions to O2. Ion energy and angular distributions (IEADs) incident onto the wafer and the dielectric window under the antenna are discussed. The shape of the antenna influences the severity of the E–H transition and the IEADs, with antennas having larger surface areas facing the plasma producing larger capacitive coupling. Validation of the model is performed by comparison of computed electron densities with experimental measurements. [ABSTRACT FROM AUTHOR]

Aiming at the joint optimisation problem of uncertain distributed manufacturing and preventive maintenance (PM) for semiconductor wafers considering multi-energy complementary (UDMPM-MEC), a bi-level integrated scheduling model is constructed with interval makespan, total interval carbon emissions and total interval cost as the optimisation objectives. The upper model is used to obtain the joint optimisation scheduling scheme of distributed manufacturing and PM, and the lower model determines the energy allocation scheme on this basis. Given the problem characteristics, we propose an improved multi-objective whale swarm algorithm (IMOWSA). By designing jobs-to-factories allocation strategy, synchronous scheduling maintenance strategy, multi-energy complementary energy allocation decoding strategy, individual distance calculation strategy, multi-neighbourhood search strategy and deduplication strategy, the local and global search are effectively balanced, and the convergence precision is improved. A large number of simulation experiments and comparative analyses show that the IMOWSA algorithm is feasible and effective to solve this problem. [ABSTRACT FROM AUTHOR]

The unprecedented sanctions policy against Russia forces us to think about the directions of development of the domestic economy. One of the priority directions is the development of the domestic electronic industry — the most knowledge-intensive branch of mechanical engineering, the structural core of the post-industrial wave in accordance with the wave concept of E. Toffler. It is electronic products, including components and products made from them, that are extremely differentiated and, among other things, serve as components for other industrial and consumer goods. This means that the development of other spheres and sectors of the economy becomes difficult without the development of the electronic industry. To determine the basic strategy for the development of the electronic industry in Russia. Tasks of the article — to identify the strategic role of innovations in electronics and microelectronics in Russia’s achievement of technological sovereignty.The article is based on the general theory of strategy and methodology of strategizing by academician, foreign member of the Russian Academy of Sciences V. L. Kvint, developed at the Center for Strategic Studies of the Institute for Mathematical Research of Complex Systems and the Department of Economic and Financial Strategy of Lomonosov Moscow State University.The article provides an overview of global innovation trends and startups that are transforming the electronic industry today. In addition to the review, the author provides statistical information on the level of development of these technologies in Russia.The strategy for the development of the electronic industry in Russia should be based on a reorientation towards innovative products and components, as well as innovative production methods.

In recent years, research on hyperdoped semiconductors has accelerated, displaying dopant concentrations far exceeding solubility limits to surpass the limitations of conventionally doped materials. Nitrogen defects in silicon have been extensively investigated for their unique characteristics compared to other pnictogen dopants. However, previous practical investigations have encountered challenges in achieving high nitrogen defect concentrations due to the low solubility and diffusivity of nitrogen in silicon, and the necessary non-equilibrium techniques, such as ion implantation, resulting in crystal damage and amorphisation. In this study, we present a single-step technique called high-pressure gas immersion excimer laser doping (HP-GIELD) to manufacture nitrogen-hyperdoped silicon. Our approach offers ultrafast processing, scalability, high control, and reproducibility. Employing HP-GIELD, we achieved nitrogen concentrations exceeding 6 at% (3.01 × 1021 at/cm3) in intrinsic silicon. Notably, nitrogen concentration remained above the liquid solubility limit to ~1 µm in depth. HP-GIELD's high-pressure environment effectively suppressed physical surface damage and the generation of silicon dangling bonds, while the well-known effects of pulsed laser annealing (PLA) preserved crystallinity. Additionally, we conducted a theoretical analysis of light-matter interactions and thermal effects governing nitrogen diffusion during HP-GIELD, which provided insights into the doping mechanism. Leveraging excimer lasers, our method is well-suited for integration into high-volume semiconductor manufacturing, particularly front-end-of-line processes. [ABSTRACT FROM AUTHOR]

Hardware-based acceleration approaches for Machine Learning (ML) workloads have been embracing the significant potential of post-CMOS switching devices to attain reduced footprint and/or energy-efficient execution relative to transistor-based GPU and/or TPU-based accelerator architectures. Meanwhile, the promulgation of fabless IC chip manufacturing paradigms has heightened the hardware security concerns inherent in such approaches. Namely, unauthorized access to various supply chain stages may expose significant vulnerabilities resulting in malfunctions including subtle adversarial outcomes via the malicious generation of differentially-corrupted outputs. Whereas the Spin-Orbit Torque Magnetic Tunnel Junction (SOT-MTJ) is a leading spintronic device for use in ML accelerators, as well as holding security tokens, their manufacturing-only security exposures are identified and evaluated herein. Results indicate a novel vulnerability profile whereby an adversary without access to the circuit netlist could differentially-influence the machine learning application's behavior. Specifically, ML recognition outputs can be significantly swayed via a global modification of oxide thickness (Tox) resulting in bit-flips of the weights in the crossbar array, thus corrupting the recognition of selected digits in MNISTdataset differentially creating an opportunity for an adversary. With just 0.05% of bits in crossbar having a flipped resistance state, digits "4" and "5" show the highest overall error rates, and digit "9" exhibit the lowest impact, with recognition accuracy of digits "2," "3," and "8" unaffected by changing the oxide thickness of SOT-MTJs uniformly from 0.75 nm to 1.2 nm without modifying the netlist nor even having access to the circuit design itself. Exposures and mitigation approaches to such novel and potentially damaging manufacturing-side intrusions are identified, postulated, and quantitatively assessed. [ABSTRACT FROM AUTHOR]