Device Simulation Research Articles

A machine learning approach for optimizing and accurate prediction of performance parameters for stacked nanosheet transistor

In this article, the possibilities of accurate prediction of wide range of parameters and optimizing the same through machine learning (ML) approach have been demonstrated for the multi stacked nanosheet transistor (NSFET). The machine is trained by the generated data of the tedious calibrated technology computer aided (TCAD) simulations. An innovative strategy is employed that combines ML with device simulations. Numerous devices are simulated with different geometric parameters like height, width, length and equivalent oxide thickness of the channel. The input, output, and CV characteristics are extrapolated from the simulation which is predicted by ML models. The DC, Analog and RF parameters are derived with a domain expertise approach. The device parameters are anticipated with actual values by the ML approach. Random forest regression, linear regression, polynomial regression and decision tree regression are employed for the prediction of the performance parameters. Random forest regression models provide significant R2 score with minimal error percentage. It indicates that TCAD-augmented ML can be considered an alternative to device simulation due to its reduction in computational cost. This work can also be treated as a benchmark for accurate prediction of the NSFET.

Lead-free perovskite solar cell based on methyl ammonium tin iodide: Possible power conversion efficiency enhancement by device simulation

Although solar cells have the potential to create an endless amount of electrical power, their comparatively low power conversion efficiency draws the curiosity of both academics and industry. The primary goal of this research is to examine the performance of lead-free perovskite solar cells that use methyl ammonium tin iodide (CH3NH3SnI3) as the active material. The SCAPS-1D programme is employed for the simulation and analysis of the solar cells. The solar cell structure was analyzed computationally, and the resulting output was examined by comparing the I-V characteristic curves across different combinations of parameters. Various materials and combinations of parameters were explored in an effort to enhance the power conversion efficiency of solar cells based on CH3NH3SnI3. A series of comprehensive evaluations were undertaken to examine the influence of technical factors, including absorber layer thickness, defect density, and acceptor density, on performance. The acceptor density that yielded the best results was selected in order to improve the performance of the solar cell. The simulation results were acquired across a temperature range spanning from 300K to 450K, revealing enhanced stability within the range of typical ambient temperatures. The study produced encouraging findings, which included a short circuit current density of 30.81mA/cm2, an open circuit voltage of 1.25V, a fill factor of 80.35%, and a power conversion efficiency of 31.11%. This study has the potential to facilitate the development of environmentally sustainable, highly efficient, and cost-effective thin film solar cells based on CH3NH3SnI3, which exhibit noteworthy physical characteristics as alternatives to lead-containing perovskite solar cells.

Gradient-descent hardware-aware training and deployment for mixed-signal neuromorphic processors

Mixed-signal neuromorphic processors provide extremely low-power operation for edge inference workloads, taking advantage of sparse asynchronous computation within spiking neural networks (SNNs). However, deploying robust applications to these devices is complicated by limited controllability over analog hardware parameters, as well as unintended parameter and dynamical variations of analog circuits due to fabrication non-idealities. Here we demonstrate a novel methodology for offline training and deployment of SNNs to the mixed-signal neuromorphic processor DYNAP-SE2. Our methodology applies gradient-based training to a differentiable simulation of the mixed-signal device, coupled with an unsupervised weight quantization method to optimize the network’s parameters. Parameter noise injection during training provides robustness to the effects of quantization and device mismatch, making the method a promising candidate for real-world applications under hardware constraints and non-idealities. This work extends Rockpool, an open-source deep-learning library for SNNs, with support for accurate simulation of mixed-signal SNN dynamics. Our approach simplifies the development and deployment process for the neuromorphic community, making mixed-signal neuromorphic processors more accessible to researchers and developers.

Modeling of fluctuations in dynamical optoelectronic device simulations within a Maxwell-density matrix Langevin approach

We present a full-wave Maxwell-density matrix simulation tool including c-number stochastic noise terms for the modeling of the spatiotemporal dynamics in active photonic devices, such as quantum cascade lasers and quantum dot structures. The coherent light–matter interaction in such devices plays an important role in the generation of frequency combs and other nonlinear and non-classical optical phenomena. Since the emergence of nonlinear and non-classical features is directly linked to the noise properties, detailed simulations of the noise characteristics are required for the development of low-noise quantum optoelectronic sources. Our semiclassical simulation framework is based on the Lindblad equation for the electron dynamics, coupled with Maxwell’s equations for optical propagation in the laser waveguide. Fluctuations arising from the interactions of the optical field and quantum system with their reservoirs are treated within the quantum Langevin theory. Here, the fluctuations are included by adding stochastic c-number terms to the Maxwell-density matrix equations. The implementation in the mbsolve dynamic simulation framework is publicly available.

Numerical simulation and performance optimization of all-inorganic CsGeI3 perovskite solar cells with stacked ETLs/C60 by SCAPS device simulation

Perovskite solar cells, particularly those featuring CsGeI3 as the absorption layer, are distinguished by their non-toxicity and inherent stability. In this work, to optimize the interface contact of the cells, enhance their built-in electric field, and consequently augment the power conversion efficiency, a C60 layer was incorporated into the electronic transport layer. This incorporation led to the formation of a novel dual-electron transport layer structure. Employing SCAPS-1D simulation software, a range of electron and hole transport layer combinations was explored. The optimal configuration was determined to be FTO/SnO2/C60/IDL1/CsGeI3/IDL2/Cu2O/Au. The absorption layer's thickness, doping concentration, defect density, and bandgap were optimized. Furthermore, an investigation was conducted on the performance implications arising from alterations in the dual electron conduction layer SnO2/C60's thickness, bandgap, and doping concentration. Concurrently, adjustments were made to refine the defect density in the interface defect layer, and an investigation was carried out to determine the optimal operational temperature. These endeavors culminated in the achievement of an impressive 30.34% optimal power conversion efficiency, representing a significant 54.25% enhancement over the pre-optimization efficiency of 19.67%. This work stands as a valuable reference for subsequent investigations into the dual electron transport layer structure in Ge-based perovskite solar cells.

A study on sensitivity to an embedded nanostructure in a micrometer-channel-length Si MOSFET

Nano-artifact metrics (NAM) is an information security technology that uses a nano-scale random structure as a unique identifier, and is expected to provide secure authentication in the Internet of Things era. For electrical discrimination of the two-dimensional random nanostructure in terms of NAM, we investigated the sensitivity to the nanostructure in a Si MOSFET with micrometer channel length in a simulation and experiment. The device simulation showed that the sensitivity was increased by decreasing the channel length and increasing the height of the nano-convex structures. It also showed that a device with a 10 μm channel length could detect a nano-convex. On the other hand, the fabricated Si MOSFET with a 50 nm height nano-convex showed lower nanostructure sensitivity than that expected in the simulation. A detailed analysis indicated that the degradation of the sensitivity was attributed to fabrication process issues, including the unintentional reduction of the convex size and high source and drain resistance.

Structure design and deformation analysis of double hydraulic cylinder deep-sea pressure simulator

Because of the particularity of the deep-sea environment and the difficulty of on-site observation and research, it is particularly important to establish a deep-sea simulation system for research. Based on the high measurement accuracy requirement of deep-sea simulation equipment, a design scheme of a deep-sea pressure simulation device with a double hydraulic cylinder that can measure small water pressure changes is proposed. Firstly, the working parameters and pressure principle of the design are introduced, and then the deformation of the deep-sea pressure simulation device is analyzed. The mathematical model of piston feed and pressure without considering the deformation of the sealed pressure cylinder and the mathematical model of piston feed and pressure considering the deformation of the sealed pressure cylinder are established respectively. Finally, the relationship curve is analyzed by using origin software. The results show that when the deep-sea pressure simulator is used to simulate the underwater environment, it is necessary to consider not only the influence of liquid compression on the pressure but also the influence of deformation on the pressure.

Thin-body effects in double-gate tunnel field-effect transistors

Scaling down the body thickness (T b) of double-gate tunnel field-effect transistors (DG-TFETs) is helpful in suppressing short-channel effects, but it may give rise to thin-body effects (TBEs). Based on 2D device simulations, this study examines the mechanisms and influences of TBEs in DG-TFETs as T b is scaled down. Differently from previous beliefs, the on-current degradation in thin-body DG-TFETs is not mainly caused by the volume effect, but rather by a newly defined TBE named lateralization effect. This is because the lateralization of tunneling direction significantly increases tunnel width, whereas the reduction of tunneling volume is quite limited due to narrow tunneling regions. To study the T b-dependence of current, therefore, the vertical tunneling has to be taken into consideration. When considered as a TBE, the fringing field effect caused by reduction in T b is not significant in degrading the on-current of thin-body DG-TFETs because the narrow tunneling regions are strongly gate-controlled. The only TBE that enhances the on-current is the coupling effect, but its role is only significant for low-bandgap bodies in which the coupling effect can efficiently promote the tunneling towards the body center. Not as previously thought that the quantum confinement effect monotonically increased, it even decreases as T b decreases down to sub-10 nm before turning to increase, thanks to the space sharing between proximate local quantum wells. A comprehensive understanding of the TBEs is useful for providing design insight, especially for determining the optimal T b to maximize the on-current.

APRTC: advanced precise rapid thermal cycling in blow molding by applying fuzzy controller on thermoelectric devices for cooling and heating applications

This study presents an innovative approach to enhance thermal control in the blow molding production process, specifically focusing on achieving rapid and precise temperature regulation. To address the significant time consumption during mold closure, strategically deploying thermoelectric modules (TEMs) arranged according to the mold’s geometry enables meticulous temperature monitoring through integrated thermocouples. The configuration forms a matrix of interconnected TEMs, thermocouples, and controllers, orchestrating precise temperature adjustments across the mold surface. The research integrates fuzzy algorithms to facilitate seamless communication between thermoelectric devices and thermocouples for precise mold temperature control. Simulation of thermoelectric devices using COMSOL Multiphysics and implementation of fuzzy controllers in Matlab, connected through a live-link, enables real-time. I have adjustments and precise control. This integrated approach allows for a comprehensive analysis and optimization of the thermal control system, ensuring effective and adaptive temperature regulation. TEMs, operating as solid-state heat pumps, offer size versatility, adaptability to diverse mold shapes, and maintenance-free operation. Leveraging individual control capabilities significantly enhances heat transfer resolution, enabling precise manipulation of heat flux. Emphasizing the reduction of cooling time and achieving high-gloss surface quality, the study introduces the Advanced Precise Rapid Thermal Cycling in Blow Molding (APRTC) method. APRTC strategically employs TEMs to expedite cooling and elevate temperature determination accuracy, guaranteeing enhanced temperature resolution on the core mold surface. Furthermore, the study optimizes energy utilization through a closed-loop fluid circulation system, harnessing waste energy from other stages. Comparative analysis highlights APRTC's superiority, revealing a remarkable 91%\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$91 \\%$$\\end{document} reduction in temperature differentials during the cooling cycle and an 89%\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$89 \\%$$\\end{document} decrease in the heating stage. These findings underscore APRTC's efficacy in maintaining minimal temperature variations throughout the entire molding process.

Comparative performance analysis of poly (3-hexylthiophene-2, 5-dial) and [6,6]-phenyl-C61 butyric acid methyl ester-based organic solar cells in bulk-heterojunction and bilayer structure using SCAPS

This work is concerned with the design and photovoltaic performance study of organic semiconducting materials-based solar cells using a SCAPS simulator. Organic solar cells are one form that absorbs light and produces electricity. They exhibit high power conversion efficiency, low manufacturing costs, lightweight, and flexibility which makes them a promising technology for renewable energy. We developed and designed organic solar cell devices with the structure ITO/PEDOT: PSS/P3HT: PCBM/PFN-Br/Al for bulk-heterojunction structure and ITO/PEDOT: PSS/P3HT: PCBM/PFN-Br/Al for bilayer structure where PEDOT: PSS and PFN-Br were used as hole and electron transporting layers (HTL and ETL) respectively. The short circuit current density versus voltage (J-V) characteristics as well as photovoltaic parameters namely open circuit voltage (VOC), short circuit current density (JSC), fill factor (FF), and power conversion efficiency (PCE) have been examined. In addition, the quantum efficiency (QE) of the two structures has been also studied. The effectiveness of the photovoltaic simulated systems has also been attempted to be enhanced by varying the absorber layer's thickness for both architectures. Later, this turned out that variations in the band gap of the absorber layer had an impact on the factors that determined how well solar cells performed. It has been found that the simulated bulk-heterojunction device exhibits 20.43% PCE whereas 3.52% PCE has been calculated from the simulated bilayer device.

Multi-scale alignment to buried atom-scale devices using Kelvin probe force microscopy

Abstract Fabrication of quantum devices by atomic-scale patterning with scanning tunneling microscopy (STM) has led to the development of single/few atom transistors, few-donor/quantum dot devices for spin manipulation, and arrayed few-donor devices for analog quantum simulation. We have developed atomic precision lithography, dopant incorporation, device encapsulation, ex situ device re-location, and contact processes to enable high-yield device fabrication. In this work, we describe a multiscale alignment strategy using Kelvin probe force microscopy to enable the alignment of buried device components to electronic support structures such as source/drain leads, in-plane and top gates, and waveguides while preserving flexibility in the placement of fabricated STM patterns. The required spatial accuracy to bridge the sub-micrometer scale central region of the device to millimeter scale large wire-bond pads is achieved through a multi-step alignment process at various stages of fabrication, including atom-scale device fabrication using STM, re-location and registration, and electron beam lithography for contact leads and pads. This alignment strategy allows imaging small device regions as well as large-scale fiducial marks, thereby bridging the gap from nanometer STM patterns to the millimeter-scale electrical contact fabrication with a 95% yield on more than 150 devices fabricated to date.

Optimal modeling of perovskite solar cell with graphene oxide as hole transport layer using L32 (28) Taguchi design

Material parameter variations are one of the main contributors affecting the performance of solar cell devices, thus, Taguchi design is employed to optimize the material parameters in attaining maximum power conversion efficiency (PCE). This paper discusses the optimal modeling of the Perovskite solar cell (PSC) with graphene oxide (GO) hole transport layer (HTL) using L32 (28) Taguchi design. The device simulation is conducted using a solar cell capacitance simulator (SCAP), whereas the L32 (28) Taguchi design is used for device optimization. The final results reveal that the L32 (28) Taguchi design has significantly optimized the device parameters in which FTO thickness, FTO donor concentration, TiO2 thickness, TiO2 donor concentration, CH3NH3PbI3-xClx thickness, CH3NH3PbI3-xClx donor concentration, GO thickness and GO acceptor concentration are predictively set to 0.1 µm, 1 x 1020 cm-3, 0.03 µm, 1 x 1020 cm-3, 0.9 µm, 1 x 1020 cm-3, 0.03 µm and 1 x 1020 cm-3 correspondingly. Analysis of variance (ANOVA) reveals that the CH3NH3PbI3-XClX thickness is the most dominant input parameter affecting the PCE of the device. The optimized input parameters yield the maximum attainable PCE of 35.91% with a signal-to-noise ratio (SNR) of 31.11 dB.

Computational Flow Diverter Implantation—A Comparative Study on Pre-Interventional Simulation and Post-Interventional Device Positioning for a Novel Blood Flow Modulator

Due to their effect on aneurysm hemodynamics, flow diverters (FD) have become a routine endovascular therapy for intracranial aneurysms. Since over- and undersizing affect the device’s hemodynamic abilities, selecting the correct device diameter and accurately simulating FD placement can improve patient-specific outcomes. The purpose of this study was to validate the accuracy of virtual flow diverter deployments in the novel Derivo® 2 device. We retrospectively analyzed blood flows in ten FD placements for which 3D DSA datasets were available pre- and post-intervention. All patients were treated with a second-generation FD Derivo® 2 (Acandis GmbH, Pforzheim, Germany) and post-interventional datasets were compared to virtual FD deployment at the implanted position for implanted stent length, stent diameters, and curvature analysis using ANKYRAS (Galgo Medical, Barcelona, Spain). Image-based blood flow simulations of pre- and post-interventional configurations were conducted. The mean length of implanted FD was 32.61 (±11.18 mm). Overall, ANKYRAS prediction was good with an average deviation of 8.4% (±5.8%) with a mean absolute difference in stent length of 3.13 mm. There was a difference of 0.24 mm in stent diameter amplitude toward ANKYRAS simulation. In vessels exhibiting a high degree of curvature, however, relevant differences between simulated and real-patient data were observed. The intrasaccular blood flow activity represented by the wall shear stress was qualitatively reduced in all cases. Inflow velocity decreased and the pulsatility over the cardiac cycle was weakened. Virtual stenting is an accurate tool for FD positioning, which may help facilitate flow FDs’ individualization and assess their hemodynamic impact. Challenges posed by complex vessel anatomy and high curvatures must be addressed.

Experimental study on mine water purification mechanism for broken coal and rock masses in the underground reservoir of ecologically vulnerable mining area.

Water-rock interaction mechanism and water purification capacity of broken coal and rock masses are very important for the efficient operation of the underground reservoir. In this paper, a water purification simulation device for an underground mine reservoir was designed. The experimental study on the dynamic interaction between broken coal and rock masses and mine water was carried out. The water purification mechanism is analyzed from the changes in rock mineral composition and mine water quality before and after the test. The results show that after the broken coal and rock mass purification, the water turbidity and the concentration of chlorides and suspended solids decreased obviously. The water purification capacities of mudstone and sandstone are stronger than that of coal samples. After 60days of reaction between the working face sewage and the broken samples (mudstone, sandstone, and coal), the turbidity, chromaticity, and residual chlorine decreased by > 90%, 90%, and 60%, respectively; and COD decreased by 35.29%, 30.59%, and 28.99%, respectively. While the TDS and the total hardness increased by about 40%, 30%, and 10% for the mudstone, sandstone, and coal, respectively. It shows that coal also has the worst degradation performance. The water purification effect of broken coal and rock masses has a significant time effect. The early stage of water-rock interaction is dominated by mineral dissolution, and the middle stage is dominated by precipitation and adsorption. The pH value of the solution has a certain influence on the ion change. In the later stage, the water-rock interaction is weak in a dynamic equilibrium state, and the change in the mine water quality index is not obvious. Considering the influence of rock lithology on water quality and the law of water-rock interaction time, the construction site selection and water storage time optimization of underground reservoirs in Jinjie Coal Mine were carried out, respectively.

Mole Fraction and Device Reliability Analysis of Vertical-Tunneling-Attributed Dual-Material Double-Gate Heterojunction-TFET with Si0.7Ge0.3 Source Region at Device and Circuit Level

This paper puts forward a new vertical-tunneling-attributed dual-material double-gate heterojunction-TFET (VTDMDG-HTFET). The device structure of VTDMDG-HTFET includes Si[Formula: see text]Ge[Formula: see text] material for the designing of source region. The mechanism of vertical tunneling in VTDMDG-HTFET provides superior electric field at tunneling interface and leads to improved transfer characteristics. VTDMDG-HTFET also includes metal gate work-function engineering approach to facilitate high ON-current and small OFF-current. Application of high-[Formula: see text] dielectric material HfO2 in the form of gate oxide enhances the capacitive-coupling at channel-source interface. Moreover, the work includes the comparison of proposed VTDMDG-HTFET with the conventional vertical-tunneling based dual-material double-gate-TFET (VTDMDG-TFET). It has been observed from the simulation results that SiGe source-based VTDMDG-HTFET offers better DC characteristics in terms of superior ON-current, smaller OFF-current, steeper subthreshold-slope, lower threshold voltage, higher transconductance along with smaller drain-induced barrier-lowering (DIBL) effects as compared to its counterpart silicon-source-based VTDMDG-TFET. In this work, reliability of VTDMDG-HTFET has also been examined and compared with VTDMDG-TFET in terms of temperature sensitivity and effect of different interface trap charges. The device simulation and investigations have been carried out using technology computer aided design (TCAD) tool. Moreover, device circuit mixed-mode simulation analysis is carried out for conventional and proposed device based resistive load inverters. A comparison is performed between both the inverters in terms of prorogation delay and switching threshold voltages. Further, the mixed mode simulation analysis is also performed for the proposed device based inverter under the influence of different interface trap charges (ITCs). Study shows that reliability performance of the proposed device is better as compared to its counterpart conventional device. Hence, vertical tunneling and SiGe source-based VTDMDG-HTFET can be a better choice for various applications such as analog/RF, analog and digital electronic circuits, energy harvesting, gas-sensing and memory devices.

Skin marker-based versus bone morphology-based coordinate systems of the hindfoot and forefoot

Segment coordinate systems (CSs) of marker-based multi-segment foot models are used to measure foot kinematics, however their relationship to the underlying bony anatomy is barely studied. The aim of this study was to compare marker-based CSs (MCSs) with bone morphology-based CSs (BCSs) for the hindfoot and forefoot. Markers were placed on the right foot of fifteen healthy adults according to the Oxford, Rizzoli and Amsterdam Foot Model (OFM, RFM and AFM, respectively). A CT scan was made while the foot was loaded in a simulated weight-bearing device. BCSs were based on axes of inertia. The orientation difference between BCSs and MCSs was quantified in helical and 3D Euler angles. To determine whether the marker models were able to capture inter-subject variability in bone poses, linear regressions were performed. Compared to the hindfoot BCS, all MCSs were more toward plantar flexion and internal rotation, and RFM was also oriented toward more inversion. Compared to the forefoot BCS, OFM and RFM were oriented more toward dorsal and plantar flexion, respectively, and internal rotation, while AFM was not statistically different in the sagittal and transverse plane. In the frontal plane, OFM was more toward eversion and RFM and AFM more toward inversion compared to BCS. Inter-subject bone pose variability was captured with RFM and AFM in most planes of the hindfoot and forefoot, while this variability was not captured by OFM. When interpreting multi-segment foot model data it is important to realize that MCSs and BCSs do not always align.

Deep photonic network platform enabling arbitrary and broadband optical functionality

Expanding applications in optical communications, computing, and sensing continue to drive the need for high-performance integrated photonic components. Designing these on-chip systems with arbitrary functionality requires beyond what is possible with physical intuition, for which machine learning-based methods have recently become popular. However, computational demands for physically accurate device simulations present critical challenges, significantly limiting scalability and design flexibility of these methods. Here, we present a highly-scalable, physics-informed design platform for on-chip optical systems with arbitrary functionality, based on deep photonic networks of custom-designed Mach-Zehnder interferometers. Leveraging this platform, we demonstrate ultra-broadband power splitters and a spectral duplexer, each designed within two minutes. The devices exhibit state-of-the-art experimental performance with insertion losses below 0.66 dB, and 1-dB bandwidths exceeding 120 nm. This platform provides a tractable path towards systematic, large-scale photonic system design, enabling custom power, phase, and dispersion profiles for high-throughput communications, quantum information processing, and medical/biological sensing applications.

Thermal Compact Modeling and Resistive Switching Analysis in Titanium Oxide-Based Memristors.

Resistive switching devices based on the Au/Ti/TiO2/Au stack were developed. In addition to standard electrical characterization by means of I-V curves, scanning thermal microscopy was employed to localize the hot spots on the top device surface (linked to conductive nanofilaments, CNFs) and perform in-operando tracking of temperature in such spots. In this way, electrical and thermal responses can be simultaneously recorded and related to each other. In a complementary way, a model for device simulation (based on COMSOL Multiphysics) was implemented in order to link the measured temperature to simulated device temperature maps. The data obtained were employed to calculate the thermal resistance to be used in compact models, such as the Stanford model, for circuit simulation. The thermal resistance extraction technique presented in this work is based on electrical and thermal measurements instead of being indirectly supported by a single fitting of the electrical response (using just I-V curves), as usual. Besides, the set and reset voltages were calculated from the complete I-V curve resistive switching series through different automatic numerical methods to assess the device variability. The series resistance was also obtained from experimental measurements, whose value is also incorporated into a compact model enhanced version.

Identifying Performance Limiting Parameters in Perovskite Solar Cells Using Machine Learning

Herein, it is shown that machine learning (ML) methods can be used to predict the parameter that limits the solar‐cell performance most significantly, solely based on the current density–voltage (J–V) curve under illumination. The data (11’150 J–V curves) to train the model is based on device simulation, where 20 different physical parameters related to charge transport and recombination are varied individually. This approach allows to cover a wide range of effects that could occur when varying fabrication conditions or during degradation of a device. Using ML, the simulated J–V curves are classified for the changed parameter with accuracies above 80%, where Random Forests perform best. It turns out that the key parameters, short‐circuit current density, open‐circuit voltage, maximum power conversion efficiency, and fill factor are sufficient for accurate predictions. To show the practical relevance, the ML algorithms are then applied to reported devices, and the results are discussed from a physics perspective. It is demonstrated that if some specified conditions are met, satisfying results can be reached. The proposed workflow can be used to better understand a device's behavior, e.g., during degradation, or as a guideline to improve its performance without costly and time‐consuming lab‐based trial‐and‐error methods.

Device simulations of perovskite transistors containing mobile ions and their relevance to experimental data

We present a device simulation of lead-halide perovskite-based thin film transistors (TFTs) containing mobile charged species to provide physical reasoning for the various experimental reports. We study the output characteristics for a range of scan duration (1/speed), average mobile ion densities, and N- and P-channel TFTs. We then directly compare our results to published data by Zeidell et al. [Adv. Electron. Mater. 4(12), 1800316 (2018)] and show that if the transistor’s measurement procedure is such that the ions’ effects are apparent, and then, our model can resolve the sign of the mobile ions in their MAPbI3−xClx TFTs (cations) and provide a good estimate of their density (∼1017 cm−3 at 200 k). Interestingly, we find that effects previously associated with channel screening are due to the ion-blocking of the charge extraction and that the incomplete saturation often reported is due to ion-induced channel shortening. Utilizing the same perovskite materials as in solar cells would allow researchers to improve their understanding of the mechanisms governing solar photovoltaics and improve their performance.