Nanoscale Semiconductor Devices Research Articles

Spatially resolved random telegraph fluctuations of a single trap at the Si/SiO2 interface

We use electrostatic force microscopy to spatially resolve random telegraph noise at the Si/SiO2 interface. Our measurements demonstrate that two-state fluctuations are localized at interfacial traps, with bias-dependent rates and amplitudes. These two-level systems lead to correlated carrier number and mobility fluctuations with a range of characteristic timescales; taken together as an ensemble, they give rise to a [Formula: see text] power spectral trend. Such individual defect fluctuations at the Si/SiO2 interface impair the performance and reliability of nanoscale semiconductor devices and will be a significant source of noise in semiconductor-based quantum sensors and computers. The fluctuations measured here are associated with a four-fold competition of rates, including slow two-state switching on the order of seconds and, in one state, fast switching on the order of nanoseconds which is associated with energy loss.

Characteristics of clean SiO2 atomic layer etching based on C6F6 physisorption

SiO2 atomic layer etching (ALE) techniques are widely used to improve the etch issues related to nanoscale semiconductor device etching such as self-aligned contact (SAC) etching requiring high etch selectivity of SiO2/Si3N4. For more precise and cleaner SiO2 ALE, cryo-ALE methods using physisorption of fluorocarbon gas at a cryotemperature are also being investigated. In this study, to avoid the use of cryotemperature for clean SiO2 ALE, a high boiling point (HBP) gas such as C6F6, having 80 °C of boiling point (global warming potential: 7), has been used as the physisorption gas without using cryotemperatures. At a substrate temperature lower than −15 °C, due to the adequate physisorption of C6F6 on the SiO2 and Si3N4 surfaces, high SiO2 etch depth per cycle (EPC) and high etch selectivity of SiO2/Si3N4 could be achieved through the higher chemical reaction of the fluorocarbon layer with SiO2 than that of Si3N4. For the etching of Si3N4 line patterned SiO2, an O2 descumming step was added after every 50 cycles of ALE process to prevent fluorocarbon layer deposition on the sidewall of the Si3N4 mask. Finally, anisotropic SiO2 etch profiles with high etch selectivity of SiO2/Si3N4 could be obtained without contamination of chamber sidewalls.

Microwave characterization of nanomaterials using planar slot resonator

Nanomaterial characterization using microwaves is needed in nanoscale semiconductor devices, microwave imaging, EM shielding, and wireless communication. Many nanomaterials are used as metallic or dielectric layers in these applications. In this paper, we report the characterization of nanomaterials using planar Microwave Slot Resonator (MSR) which was designed and studied using 3D EM simulation tool. The response of MSR is parameterized which offers a platform to calculate relative permittivity (ε r) and conductivity (σ) from measured high frequency response of nanomaterial loaded MSR. With simplified method, this microwave characterization offers accurate and faster results which be used in design, calculation and numerical analysis of nanomaterial based electronic/optoelectronic devices and sensor/shielding applications.

Atomic Layer Etching of SiO2 for Nanoscale Semiconductor Devices: A Review

Atomic Layer Etching of SiO₂ for Nanoscale Semiconductor Devices: A Review

Two-Step Magnetic Ordering in Intercalated Niobium Dichalcogenide MnXNbS2

Transition metal dichalcogenides are studied due to the possibility of creating nanoscale semiconductor devices, as well as fundamental issues of magnetic ordering. We researched the crystal structure and magnetic properties of niobium dichalcogenide Mn0.30NbS2. The results of the X-ray study showed the possible existence of an intermediate 23a0·23a0 structure between the “basic” superstructures. Also, two local maximums were found in the temperature dependence of the dynamic magnetic susceptibility. These features can indirectly confirm the presence of a transition superstructure and reflect the two-step nature of the magnetic ordering.

MOSFET Physics-Based Compact Model Mass-Produced: An Artificial Neural Network Approach

The continued scaling-down of nanoscale semiconductor devices has made it very challenging to obtain analytic surface potential solutions from complex equations in physics, which is the fundamental purpose of the MOSFET compact model. In this work, we proposed a general framework to automatically derive analytical solutions for surface potential in MOSFET, by leveraging the universal approximation power of deep neural networks. Our framework incorporated a physical-relation-neural-network (PRNN) to learn side-by-side from a general-purpose numerical simulator in handling complex equations of mathematical physics, and then instilled the "knowledge'' from the simulation data into the neural network, so as to generate an accurate closed-form mapping between device parameters and surface potential. Inherently, the surface potential was able to reflect the numerical solution of a two-dimensional (2D) Poisson equation, surpassing the limits of traditional 1D Poisson equation solutions, thus better illustrating the physical characteristics of scaling devices. We obtained promising results in inferring the analytic surface potential of MOSFET, and in applying the derived potential function to the building of 130 nm MOSFET compact models and circuit simulation. Such an efficient framework with accurate prediction of device performances demonstrates its potential in device optimization and circuit design.

Nanoscale Semiconductor Devices Fabricated using Adhesion Lithography at Low Cost

Nanoscale Semiconductor Devices Fabricated using Adhesion Lithography at Low Cost

Förster Resonance Energy Transfer Assisted Enhancement in Optoelectronic Properties of Metal Halide Perovskite Nanocrystals.

Regulated excited state energy and charge transfer play a pivotal role in nanoscale semiconductor device performance for efficient energy harvesting and optoelectronic applications. Herein, we report the influence of Förster resonance energy transfer (FRET) on the excited-state dynamics and charge transport properties of metal halide perovskite nanocrystals (PNCs), CsPbBr3, and its anion-exchanged counterpart CsPbCl3 with CdSe/ZnS quantum dots (QDs). We report a drop in the FRET efficiency from ∼85% (CsPbBr3) to ∼5% (CsPbCl3) with QDs, inviting significant alteration in their charge transport properties. Using two-probe measurements we report substantial enhancement in the current for the blend structure of PNCs with QDs, originating from the reduced trap sites, compared to that of the pristine PNCs. The FRET-based upshot in the conduction mechanism with features of negative differential resistance and negligible hysteresis for CsPbBr3 PNCs can add new directions to high performance-based photovoltaics and optoelectronics.

Significance of activation functions in developing an online classifier for semiconductor defect detection

In anomaly detection problems for advanced semiconductor devices, non-visual defects occur frequently. Machine learning (ML) algorithms have the advantage of identifying such defects. However, in this real-world problem, data comes sequentially in a streaming fashion, thus, we may not have sufficient data to train an ML model in batch mode. In such a scenario, online ML models are useful to detect defects immediately since they work in a single-pass mode. Besides, when data is collected from more realistic non-stationary monitoring environments, online ML models with evolving architecture are more practical. Thus, evolving and online ML models are developed in this work to detect defects in technology computer-aided design (TCAD)-based digital twin model of advanced nano-scaled semiconductor devices such as a fin field-effect transistor (FinFET) and a gate-all-around field-effect transistor (GAA-FET). Activation functions (AFs) in deep neural networks (DNNs) and membership functions (MFs) in neuro-fuzzy systems (NFSs) play an important role in the performance of those ML models. This work focuses on analyzing the effects of various AFs/MFs in our developed online ML models while detecting defects in real-world nano-scaled semiconductor devices, where significant training samples are not available. From various semiconductor datasets having fewer samples, it has been observed that the proposed evolving neuro-fuzzy system (ENFS) with Leaky-ReLU MF performs better (improvement in the range of 1.9% to 30.8% considering overall classification accuracy) than the other DNN or ENFS-based online ML models. Having an evolving architecture and online learning mechanism, besides anomaly detection, the proposed model’s performance has also been evaluated for handling large data streams problems with concept drift. The performance of the proposed method has been compared with some recently developed baselines under the prequential test-then-train protocol. The classification rates of the proposed method has an improvement in the range of 1.1% to 65.9% than the existing methods. The code of this work has been made publicly available at https://github.com/MdFerdaus/LREC.

Analysis on the Electrostatic Doping and Several Alternative Devices

With the development of semiconductor technology, the size of precision instruments is becoming more and more stringent. The purpose of electrostatic doping is to provide a possibility on nanoscale semiconductor devices and to replace chemical doping, it also replaces donor/receptor doping with Gate-Induced free electron/hole charges in ultra-thin MOS (Metal-Oxide-Semiconductor) structures, and provide some areas with high electron/hole density in semiconductor devices. This paper introduces Electrostatic Doping methods and Several Alternative Devices, emphasizing the functions of metal and semiconductor operation functions, energy band gaps and applied electric fields, and their interaction in induced ED. In addition, this paper discusses the advantages of ED devices and the major potential obstacles to future CMOs, and the modeling and experimental implementation of this approach will help to implement and evaluate the possibility of replacing traditional doping methods for innovative devices for future CMOs.

Scavenger with Protonated Phosphite Ions for Incredible Nanoscale ZrO2-Abrasive Dispersant Stability Enhancement and Related Tungsten-Film Surface Chemical-Mechanical Planarization.

For scaling-down advanced nanoscale semiconductor devices, tungsten (W)-film surface chemical mechanical planarization (CMP) has rapidly evolved to increase the W-film surface polishing rate via Fenton-reaction acceleration and enhance nanoscale-abrasive (i.e., ZrO2) dispersant stability in the CMP slurry by adding a scavenger to suppress the Fenton reaction. To enhance the ZrO2 abrasive dispersant stability, a scavenger with protonate-phosphite ions was designed to suppress the time-dependent Fenton reaction. The ZrO2 abrasive dispersant stability (i.e., lower H2O2 decomposition rate and longer H2O2 pot lifetime) linearly and significantly increased with scavenger concentration. However, the corrosion magnitude on the W-film surface during CMP increased significantly with scavenger concentration. By adding a scavenger to the CMP slurry, the radical amount reduction via Fenton-reaction suppression in the CMP slurry and the corrosion enhancement on the W-film surface during CMP performed that the W-film surface polishing rate decreased linearly and notably with increasing scavenger concentration via a chemical-dominant CMP mechanism. Otherwise, the SiO2-film surface polishing rate peaked at a specific scavenger concentration via a chemical and mechanical-dominant CMP mechanism. The addition of a corrosion inhibitor with a protonate-amine functional group to the W-film surface CMP slurry completely suppressed the corrosion generation on the W-film surface during CMP without a decrease in the W- and SiO2-film surface polishing rate.

Super fine cerium hydroxide abrasives for SiO2 film chemical mechanical planarization performing scratch free

Face-centered-cubic crystallized super-fine (~ 2 nm in size) wet-ceria-abrasives are synthesized using a novel wet precipitation process that comprises a Ce4+ precursor, C3H4N2 catalyst, and NaOH titrant for a synthesized termination process at temperature of at temperature of 25 °C. This process overcomes the limitations of chemical–mechanical-planarization (CMP)-induced scratches from conventional dry ceria abrasives with irregular surfaces or wet ceria abrasives with crystalline facets in nanoscale semiconductor devices. The chemical composition of super-fine wet ceria abrasives depends on the synthesis termination pH, that is, Ce(OH)4 abrasives at a pH of 4.0–5.0 and a mixture of CeO2 and Ce(OH)4 abrasives at a pH of 5.5–6.5. The Ce(OH)4 abrasives demonstrate better abrasive stability in the SiO2-film CMP slurry than the CeO2 abrasives and produce a minimum abrasive zeta potential (~ 12 mV) and a minimum secondary abrasive size (~ 130 nm) at the synthesis termination pH of 5.0. Additionally, the abrasive stability of the SiO2-film CMP slurry that includes super-fine wet ceria abrasives is notably sensitive to the CMP slurry pH; the best abrasive stability (i.e., a minimum secondary abrasive size of ~ 130 nm) is observed at a specific pH (6.0). As a result, a maximum SiO2-film polishing rate (~ 524 nm/min) is achieved at pH 6.0, and the surface is free of stick-and-slip type scratches.

Modeling Methods for Nanoscale Semiconductor Devices

Modeling Methods for Nanoscale Semiconductor Devices

Simulation and Modeling of Novel Electronic Device Architectures with NESS (Nano-Electronic Simulation Software): A Modular Nano TCAD Simulation Framework

The modeling of nano-electronic devices is a cost-effective approach for optimizing the semiconductor device performance and for guiding the fabrication technology. In this paper, we present the capabilities of the new flexible multi-scale nano TCAD simulation software called Nano-Electronic Simulation Software (NESS). NESS is designed to study the charge transport in contemporary and novel ultra-scaled semiconductor devices. In order to simulate the charge transport in such ultra-scaled devices with complex architectures and design, we have developed numerous simulation modules based on various simulation approaches. Currently, NESS contains a drift-diffusion, Kubo–Greenwood, and non-equilibrium Green’s function (NEGF) modules. All modules are numerical solvers which are implemented in the C++ programming language, and all of them are linked and solved self-consistently with the Poisson equation. Here, we have deployed some of those modules to showcase the capabilities of NESS to simulate advanced nano-scale semiconductor devices. The devices simulated in this paper are chosen to represent the current state-of-the-art and future technologies where quantum mechanical effects play an important role. Our examples include ultra-scaled nanowire transistors, tunnel transistors, resonant tunneling diodes, and negative capacitance transistors. Our results show that NESS is a robust, fast, and reliable simulation platform which can accurately predict and describe the underlying physics in novel ultra-scaled electronic devices.

Self-stopping slurry for planarizing extremely high surface film topography in nanoscale semiconductor devices

Self-stopping slurry for planarizing extremely high surface film topography in nanoscale semiconductor devices

A Systematic Compact Model Parameter Calibration with Adaptive Pattern Search Algorithm

A systematic device-model calibration (extraction) methodology has been proposed to reduce parameter calibration time of advanced compact model for modern nano-scale semiconductor devices. The adaptive pattern search algorithm is a variant of the direct search method, which explore in the parameter space with adaptive searching step and direction. It is very straightforward, but powerful, in high dimensional optimization problem since adaptive step and direction are decided by simple computation. The proposed method iterates less but shows superior accuracy over the conventional method. It is possible to be applied to a behavioral or empirical model correspond to emerging devices, such as tunneling field-effect transistor (TFET) and negative capacitance field-effect transistor (NCFET) due to its universality in parameter calibration for the model accuracy.

Unraveling the selective etching mechanism of silicon nitride over silicon dioxide by phosphoric acid: First-principles study

Highly selective etching of a target silicon compound is essential in semiconductor fabrication. Searching for alternatives to conventional etchant that contain hazardous chemicals is challenging, largely due to the unclear understanding of the chemical reaction. In this study, we elucidate etching machinery of phosphoric acid and its outstanding selectivity toward silicon nitride (Si3N4) over silicon dioxide (SiO2) surfaces in atomistic level. Ab-initio thermodynamic and kinetic formalisms integrated with density functional theory computation propose that pyrophosphoric acid (H4P2O7), a condensed form of orthophosphoric acid.(H3PO4) at high concentration and temperature, is even more reactive toward Si3N4 than H3PO4 which has been regarded as a dominant species for decades. We demonstrate that the superior etching selectivity derives from reaction mechanism of thermodynamic control, where H4P2O7 is much more exergonic than H3PO4. Notably, we find that water molecules close to the H4P2O7 assist sequential etching process in two ways: catalysis via structural proton transfer, and hydrolysis of diphosphate group in the Si3N4 surface. Our study guides a quick and accurate screening as well as designing efficient and safe etchants, which facilitates the fabrication of nanoscale semiconductor devices that accompanies selective etching of alternately stacked hundreds of atomic layers of Si3N4 and SiO2.

Informational Lithography Approach Based on Source and Mask Optimization

Optical lithography is a critical technique to fabricate nano-scale semiconductor devices by replicating the layouts of integrated circuits from the lithography mask onto the silicon wafer. As the critical dimension of integrated circuits continuously shrinks, source and mask optimization (SMO) methods are extensively used to improve the resolution and image fidelity of lithography patterning. However, the theoretical lower bound of the lithography pattern error in the SMO framework is not yet understood. This paper introduces an informational lithography approach to unveil the information transmission mechanism in lithography systems under freeform illumination configurations. The lithography system is regarded as an information channel, where the mask pattern and the print image on the wafer are modeled as the statistical input and output signals, respectively. Subsequently, we derive the optimal information transfer (OIT) of the lithography system, which represents the best information transfer strategy rendering the least image distortion. Based on the OIT, we derive a lower bound for the lithography pattern error, and the corresponding optimal source pattern and optimal mask probability distribution. Finally, we propose a new SMO algorithm based on the information theoretical framework to effectively improve the lithography image fidelity compared to the existing gradient-based SMO algorithm.

Characteristics of silicon nitride deposited by very high frequency (162 MHz)-plasma enhanced atomic layer deposition using bis(diethylamino)silane

Silicon nitrides, deposited by capacitively coupled plasma (CCP)-type plasma enhanced atomic layer deposition (PEALD), are generally applied to today’s nanoscale semiconductor devices, and are currently being investigated in terms of their potential applications in the context of flexible displays, etc. During the PEALD process, 13.56 MHz rf power is generally employed for the generation of reactive gas plasma. In this study, the effects of a higher plasma generation frequency of 162 MHz on both plasma and silicon nitride film characteristics are investigated for the purpose of silicon nitride PEALD, using bis(diethylamino)silane (BDEAS) as the silicon precursor, and N2 plasma as the reactant gas. The PEALD silicon nitride film deposited using the 162 MHz CCP exhibited improved film characteristics, such as reduced surface roughness, a lower carbon percentage, a higher N/Si ratio, a lower wet etch rate in a diluted HF solution, lower leakage current, and higher electric breakdown field, and more uniform step coverage of the silicon nitride film deposited in a high aspect ratio trench, as compared to silicon nitride PEALD using 13.56 MHz CCP. These improved PEALD silicon nitride film characteristics are believed to be related to the higher ion density, higher reactive gas dissociation, and lower ion bombardment energy to the substrate observed in N2 plasma with a 162 MHz CCP.

HAADF-STEM block-scanning strategy for local measurement of strain at the nanoscale

Lattice strain measurement of nanoscale semiconductor devices is crucial for the semiconductor industry as strain substantially improves the electrical performance of transistors. High resolution scanning transmission electron microscopy (HR-STEM) imaging is an excellent tool that provides spatial resolution at the atomic scale and strain information by applying Geometric Phase Analysis or image fitting procedures. However, HR-STEM images regularly suffer from scanning distortions and sample drift during image acquisition. In this paper, we propose a new scanning strategy that drastically reduces artefacts due to drift and scanning distortion, along with extending the field of view. It consists of the acquisition of a series of independent small subimages containing an atomic resolution image of the local lattice. All subimages are then analysed individually for strain by fitting a nonlinear model to the lattice images. The method allows flexible tuning of spatial resolution and the field of view within the limits of the dynamic range of the scan engine while maintaining atomic resolution sampling within the subimages. The obtained experimental strain maps are quantitatively benchmarked against the Bessel diffraction technique. We demonstrate that the proposed scanning strategy approaches the performance of the diffraction technique while having the advantage that it does not require specialized diffraction cameras.