Publications in Journals and Book Chapters

This chapter gives an overview of the most important methods in artificial intelligence (AI). The methods of symbolic AI are rooted in logic, and finding possible solutions by search is a central aspect. The main challenge is the combinatorial explosion in search, but the focus on the satisfiability problem of propositional logic (SAT) since the 1990s and the accompanying algorithmic improvements have made it possible to solve problems on the scale needed in industrial applications. In machine learning (ML), self-learning algorithms extract information from data and represent the solutions in convenient forms. ML broadly consists of supervised learning, unsupervised learning, and reinforcement learning. Successes in the 2010s and early 2020s such as solving Go, chess, and many computer games as well as large language models such as ChatGPT are due to huge computational resources and algorithmic advances in ML. Finally, we reflect on current developments and draw conclusions.

Current fingerprint identification systems face significant challenges in achieving interoperability between contact-based and contactless fingerprint sensors. In contrast to existing literature, we propose a novel approach that can combine pose correction with further enhancement operations. It uses deep learning models to steer the correction of the viewing angle, therefore enhancing the matching features of contactless fingerprints. The proposed approach was tested on real data of 78 participants (37,162 contactless fingerprints) acquired by national police officers using both contact-based and contactless sensors. The study found that the effectiveness of pose correction and unwarping varied significantly based on the individual characteristics of each fingerprint. However, when the various extension methods were combined on a finger-wise basis, an average decrease of 36.9% in equal error rates (EERs) was observed. Additionally, the combined impact of pose correction and bidirectional unwarping led to an average increase of 3.72% in NFIQ 2 scores across all fingers, coupled with a 6.4% decrease in EERs relative to the baseline. The addition of deep learning techniques presents a promising approach for achieving high-quality fingerprint acquisition using contactless sensors, enhancing recognition accuracy in various domains.

Background: The optimal indication, dose, and timing of corticosteroids in sepsis is controversial. Here, we used reinforcement learning to derive the optimal steroid policy in septic patients based on data on 3051 ICU admissions from the AmsterdamUMCdb intensive care database. Methods: We identified septic patients according to the 2016 consensus definition. An actor-critic RL algorithm using ICU mortality as a reward signal was developed to determine the optimal treatment policy from time-series data on 277 clinical parameters. We performed off-policy evaluation and testing in independent subsets to assess the algorithm’s performance. Results: Agreement between the RL agent’s policy and the actual documented treatment reached 59%. Our RL agent’s treatment policy was more restrictive compared to the actual clinician behavior: our algorithm suggested withholding corticosteroids in 62% of the patient states, versus 52% according to the physicians’ policy. The 95% lower bound of the expected reward was higher for the RL agent than clinicians’ historical decisions. ICU mortality after concordant action in the testing dataset was lower both when corticosteroids had been withheld and when corticosteroids had been prescribed by the virtual agent. The most relevant variables were vital parameters and laboratory values, such as blood pressure, heart rate, leucocyte count, and glycemia. Conclusions: Individualized use of corticosteroids in sepsis may result in a mortality benefit, but optimal treatment policy may be more restrictive than the routine clinical practice. Whilst external validation is needed, our study motivates a ‘precision-medicine’ approach to future prospective controlled trials and practice.

We present a novel setup for treating sepsis using distributional reinforcement learning (RL). Sepsis is a life-threatening medical emergency. Its treatment is considered to be a challenging high-stakes decision-making problem, which has to procedurally account for risk. Treating sepsis by machine learning algorithms is difficult due to a couple of reasons: There is limited and error-afflicted initial data in a highly complex biological system combined with the need to make robust, transparent and safe decisions. We demonstrate a suitable method that combines data imputation by a kNN model using a custom distance with state representation by discretization using clustering, and that enables superhuman decision-making using speedy Q-learning in the framework of distributional RL. Compared to clinicians, the recovery rate is increased by more than 3% on the test data set. Our results illustrate how risk-aware RL agents can play a decisive role in critical situations such as the treatment of sepsis patients, a situation acerbated due to the COVID-19 pandemic (Martineau 2020). In addition, we emphasize the tractability of the methodology and the learning behavior while addressing some criticisms of the previous work (Komorowski et al. 2018) on this topic.

Nowadays, the various omics disciplines such as genomics, proteomics, metabolomics, metagenomics, and transcriptomics generate a plethora of data. At the same time, a multitude of omics markers may be accompanied by a multitude of diseases. Hence, finding relationships between omics markers and disease in their early stages is a challenge that is at the very core of predictive or personalized medicine. In this chapter, an overview of algorithms for solving these problems of supervised learning is given, and challenges in this problem domain are discussed. Questions of learnability should be considered, and the quality and precision of the predictions should be assessed critically and quantitatively. Therefore, quality metrics for the assessment of the predictions are discussed as well.

In this work, a method for automatic hyper-parameter tuning of the stacked asymmetric auto-encoder is proposed. In previous work, the deep learning ability to extract personality perception from speech was shown, but hyper-parameter tuning was attained by trial-and-error, which is time-consuming and requires machine learning knowledge. Therefore, obtaining hyper-parameter values is challenging and places limits on deep learning usage. To address this challenge, researchers have applied optimization methods. Although there were successes, the search space is very large due to the large number of deep learning hyper-parameters, which increases the probability of getting stuck in local optima. Researchers have also focused on improving global optimization methods. In this regard, we suggest a novel global optimization method based on the cultural algorithm, multi-island and the concept of parallelism to search this large space smartly. At first, we evaluated our method on three well-known optimization benchmarks and compared the results with recently published papers. Results indicate that the convergence of the proposed method speeds up due to the ability to escape from local optima, and the precision of the results improves dramatically. Afterward, we applied our method to optimize five hyper-parameters of an asymmetric auto-encoder for automatic personality perception. Since inappropriate hyper-parameters lead the network to over-fitting and under-fitting, we used a novel cost function to prevent over-fitting and under-fitting. As observed, the unweighted average recall (accuracy) was improved by 6.52% (9.54%) compared to our previous work and had remarkable outcomes compared to other published personality perception works.

Silicon nanowire field-effect transistors are promising devices used to detect minute amounts of different biological species. We introduce the theoretical and computational aspects of forward and backward modeling of biosensitive sensors. Firstly, we introduce a forward system of partial differential equations to model the electrical behavior, and secondly, a backward Bayesian Markov-chain Monte-Carlo method is used to identify the unknown parameters such as the concentration of target molecules. Furthermore, we introduce a machine learning algorithm according to multilayer feed-forward neural networks. The trained model makes it possible to predict the sensor behavior based on the given parameters.

In distributional reinforcement learning, the entire distribution of the return instead of just the expected return is modeled. The approach with categorical distributions as the approximation method is well-known in Q-learning, and convergence results have been established in the tabular case. In this work, speedy Q-learning is extended to categorical distributions, a finite-time analysis is performed, and probably approximately correct bounds in terms of the Cramér distance are established. It is shown that also in the distributional case the new update rule yields faster policy evaluation in comparison to the standard Q-learning one and that the sample complexity is essentially the same as the one of the value-based algorithmic counterpart. Without the need for more state-action-reward samples, one gains significantly more information about the return with categorical distributions. Even though the results do not easily extend to the case of policy control, a slight modification to the update rule yields promising numerical results.

A truly meshless numerical procedure to simulate stochastic elliptic interface problems is developed. The meshless method is based on the generalized moving least squares approximation. This method can be implemented in a straightforward manner and has a very good accuracy for solving this kind of problems. Several realistic examples are presented to investigate the efficiency of the new procedure. Furthermore, compared with other meshless methods that have been developed, the present technique needs less CPU time.

We study uncertainty quantification for a Boltzmann-Poisson system that models electron transport in semiconductors and the physical collision mechanisms over the charges, using the stochastic Galerkin method in order to handle the randomness associated with the problem. In this study we choose first as a source of uncertainty the phonon energy, taking it as a random variable, as its value influences the energy jump appearing in the collision integral for electron-phonon scattering. Then we choose the lattice temperature as a random variable, since it defines the value of the collision operator terms in the case of electron-phonon scattering by being a parameter of the phonon distribution. Finally, we present our numerical simulations for the latter case. We calculate then with our stochastic Discontinuous Galerkin methods the uncertainty in kinetic moments such as density, mean energy, current, etc. associated to a possible physical temperature variation (assumed to follow a uniform distribution) in the lattice environment, as this uncertainty in the temperature is propagated into the electron PDF. Our mathematical and computational results let us predict then in a real world problem setting the impact that possible variations in the lab conditions (such as temperature) or limitations in the mathematical model (such as assumption of a constant phonon energy) will have over the uncertainty in the behavior of electronic devices.

Simple Summary: Blood-based tests for cancer detection are minimally invasive and could be useful for screening asymptomatic patients and high-risk populations. Since a single molecular biomarker is usually insufficient for an accurate diagnosis, we developed a multi-analyte liquid biopsy-based classification model to distinguish cancer patients from healthy subjects. The combination of cell-free DNA mutations, miRNAs, and cell-free DNA methylation markers improved the model's performance. Moreover, we demonstrated that the androgen receptor mutation p.H875Y is not only relevant in prostate cancer but had a strong predictive value for colorectal, bladder, and breast cancer. Our results, although preliminary, showed that a single liquid biopsy test could detect multiple cancer types simultaneously.

Abstract: Liquid biopsy-based tests emerge progressively as an important tool for cancer diagnostics and management. Currently, researchers focus on a single biomarker type and one tumor entity. This study aimed to create a multi-analyte liquid biopsy test for the simultaneous detection of several solid cancers. For this purpose, we analyzed cell-free DNA (cfDNA) mutations and methylation, as well as circulating miRNAs (miRNAs) in plasma samples from 97 patients with cancer (20 bladder, 9 brain, 30 breast, 28 colorectal, 29 lung, 19 ovarian, 12 pancreas, 27 prostate, 23 stomach) and 15 healthy controls via real-time qPCR. Androgen receptor p.H875Y mutation (AR) was detected for the first time in bladder, lung, stomach, ovarian, brain, and pancreas cancer, all together in 51.3% of all cancer samples and in none of the healthy controls. A discriminant function model, comprising cfDNA mutations (COSM10758, COSM18561), cfDNA methylation markers (MLH1, MDR1, GATA5, SFN) and miRNAs (miR-17-5p, miR-20a-5p, miR-21-5p, miR-26a-5p, miR-27a-3p, miR-29c-3p, miR-92a-3p, miR-101-3p, miR-133a-3p, miR-148b-3p, miR-155-5p, miR-195-5p) could further classify healthy and tumor samples with 95.4% accuracy, 97.9% sensitivity, 80% specificity. This multi-analyte liquid biopsy-based test may help improve the simultaneous detection of several cancer types and underlines the importance of combining genetic and epigenetic biomarkers.

The transport of molecules through nanoscale confined space is relevant in biology, biosensing, and industrial filtration. Microscopically modeling transport through nanopores is required for a fundamental understanding and guiding engineering, but the short duration and low replica number of existing simulation approaches limit statistically relevant insight. Here we explore protein transport in nanopores with a high-throughput computational method that realistically simulates hundreds of up to seconds-long protein trajectories by combining Brownian dynamics and continuum simulation and integrating both driving forces of electroosmosis and electrophoresis. Ionic current traces are computed to enable experimental comparison. By examining three biological and synthetic nanopores, our study answers questions about the kinetics and mechanism of protein transport and additionally reveals insight that is inaccessible from experiments yet relevant for pore design. The discovery of extremely frequent unhindered passage can guide the improvement of biosensor pores to enhance desired biomolecular recognition by pore-tethered receptors. Similarly, experimentally invisible nontarget adsorption to pore walls highlights how to improve recently developed DNA nanopores. Our work can be expanded to pressure-driven flow to model industrial nanofiltration processes.

We homogenize the Boltzmann--Poisson system where the background medium is given by a periodic permittivity and a periodic charge concentration. The domain is the half-space with a periodic charge concentration on the boundary. Hence the domain consists of cells in R³ that are periodically repeated in two dimensions and unbounded in the third dimension. We obtain formal results for this homogenization problem. We prove the existence and uniqueness of the solution of the Laplace and Poisson problems in the fast variables with periodic and surface charge boundary conditions generating an electric field at infinity, obtaining formal solutions for the potential in terms of Magnus expansions for the case where the diagonal permittivity matrix depends on the vertical fast variable. Further on, splitting the potential into a stationary part and a self-consistent part, performing the two-scale homogenization expansions for the Poisson and the Boltzmann equations, and applying a solvability condition, we arrive at the drift-diffusion equations for the boundary-layer problem.

Optimal design of electronic devices such as sensors is essential since it results in more accurate output at the shortest possible time. In this work, we develop optimal Bayesian inversion for electrical impedance tomography (EIT) technology in order to improve the quality of medical images generated by EIT and to put this promising imaging technology into practice. We optimize Bayesian experimental design by maximizing the expected information gain in the Bayesian inversion process in order to design optimal experiments and obtain the most informative data about the unknown parameters. We present optimal experimental designs including optimal frequency and optimal electrode configuration, all of which result in the most accurate estimation of the unknown quantities to date and high-resolution EIT medical images, which are crucial for diagnostic purposes. Numerical results show the efficiency of the proposed optimal Bayesian inversion method for the EIT inverse problem.

In the recent decade, meshless methods have been handled for solving some PDEs due to their easiness. One of the most efficient meshless methods is the element free Galerkin (EFG) method. The test and trial functions of the EFG are based upon the special basis. Recently, some modifications have been developed to improve the EFG method. One of these improvements is the variational multiscale EFG (VMEFG) procedure. In the current article, the shape functions of interpolating moving least squares (IMLS) approximation are applied to the variational multiscale EFG technique to numerical study the Navier–Stokes equations coupled with a heat transfer equation such that this model is well-known as two-dimensional nonstationary Boussinesq equations. In order to reduce the computational time of simulation, we employ a reduced order model (ROM) based on the proper orthogonal decomposition (POD) technique. In the current paper, we developed a new reduced order model based on the meshless numerical procedure for solving an important model in fluid mechanics. To illustrate the reduction in CPU time as well as the efficiency of the proposed method, we investigate two-dimensional cases.

Mathematical modeling of epidemiological diseases using differential equations are of great importance in order to recognize the characteristics of the diseases and their outbreak. The procedure of modeling consists of two essential components: the first component is to solve the mathematical model numerically, the so-called forward modeling. The second component is to identify the unknown parameter values in the model, which is known as inverse modeling and leads to identifying the epidemiological model more precisely. The main goal of this paper is to develop the forward and inverse modeling of the coronavirus (COVID-19) pandemic using novel computational methodologies in order to accurately estimate and predict the pandemic. This leads to governmental decisions support in implementing effective protective measures and prevention of new outbreaks. To this end, we use the logistic equation and the SIR (susceptible-infected-removed) system of ordinary differential equations to model the spread of the COVID-19 pandemic. For the inverse modeling, we propose Bayesian inversion techniques, which are robust and reliable approaches, in order to estimate the unknown parameters of the epidemiological models. We deploy an adaptive Markov-chain Monte-Carlo (MCMC) algorithm for the estimation of a posteriori probability distribution and confidence intervals for the unknown model parameters as well as for the reproduction number. We perform our analyses on the publicly available data for Austria to estimate the main epidemiological model parameters and to study the effectiveness of the protective measures by the Austrian government. The estimated parameters and the analysis of fatalities provide useful information for decision makers and makes it possible to perform more realistic forecasts of future outbreaks. According to our Bayesian analysis for the logistic model, the growth rate and the carrying capacity are estimated respectively as 0.28 and 14974. Moreover for the parameters of the SIR model, namely the transmission rate and recovery rate, we estimate 0.36 and 0.06, respectively. Additionally, we obtained an average infectious period of 17 days and a transmission period of 3 days for COVID-19 in Austria. We also estimate the reproduction number over time for Austria. This quantity is estimated around 3 on March 26, when the first recovery was reported. Then it decays to 1 at the beginning of April. Furthermore, we present a fatality analysis for COVID-19 in Austria, which is also of importance for governmental protective decision making. According to our analysis, the case fatality rate (CFR) is estimated as 4% and a prediction of the number of fatalities for the coming 10 days is also presented. Additionally, the ICU bed usage in Austria indicates that around 2% of the active infected individuals are critical cases and require ICU beds. Therefore, if Austrian governmental protective measures would not have taken place and for instance if the number of active infected cases would have been around five times larger, the ICU bed capacity could have been exceeded.

A new theoretical model for the dielectrophoretic (DEP) fabrication of single-walled carbon nanotubes (SWCNTs) is presented. A different frequency interval for the alignment of wide-energy-gap semiconductor SWCNTs is obtained, exhibiting a considerable difference from the prevalent model. Two specific models are study, namely the spherical model and the ellipsoid model, to estimate the frequency interval. Then, the DEP process is performed and the obtained frequencies (from the spherical and ellipsoid models) are used to align the SWCNTs. These empirical results confirm the theoretical predictions, representing a crucial step towards the realization of carbon nanotube field-effect transistors (CNT-FETs) via the DEP process based on the ellipsoid model.

In this work, we propose a parameter estimation framework for fracture propagation problems. The fracture problem is described by a phase-field method. Parameter estimation is realized with a Bayesian approach. Here, the focus is on uncertainties arising in the solid material parameters and the critical energy release rate. A reference value (obtained on a sufficiently refined mesh) as the replacement of measurement data will be chosen, and their posterior distribution is obtained. Due to time- and mesh dependencies of the problem, the computational costs can be high. Using Bayesian inversion, we solve the problem on a relatively coarse mesh and fit the parameters. In several numerical examples our proposed framework is substantiated and the obtained load-displacement curves, that are usually the target functions, are matched with the reference values.

We present an adaptive multilevel Monte Carlo algorithm for solving the stochastic drift–diffusion–Poisson system with non-zero recombination rate. The a-posteriori error is estimated to enable goal-oriented adaptive mesh refinement for the spatial dimensions, while the a-priori error is estimated to guarantee linear convergence of the H¹ error. In the adaptive mesh refinement, efficient estimation of the error indicator gives rise to better error control. For the stochastic dimensions, we use the multilevel Monte Carlo method to solve this system of stochastic partial differential equations. Finally, the advantage of the technique developed here compared to uniform mesh refinement is discussed using a realistic numerical example.

Nonlinear partial differential equations (PDEs) play an important role in the modeling of the natural phenomena as they have great significance in real-world applications. This investigation proposes a new algorithm to find the numerical solution of the nonlinear extended Fisher–Kolmogorov equation. Firstly, the time variable is discretized by a second-order finite difference scheme. The rate of convergence and stability of the semi-discrete formulation are studied by the energy method. The existence and uniqueness of the solution of the weak form based on the proposed technique have been proved in detail. Furthermore, the interpolating element free Galerkin approach based on the interpolation moving least-squares approximation is employed to derive a fully discrete scheme. Finally, the error estimate of the full-discrete plan is proposed and its convergence order is O(τ² + δ^m+1) in which τ, δ and m denote the time step, the radius of the weight function and smoothness of the exact solution of the main problem, respectively.

We develop an electrical-impedance tomography (EIT) inverse model problem in an infinite-dimensional setting by introducing a nonlinear elliptic PDE as a new EIT forward model. The new model completes the standard linear model by taking the transport of ionic charge into account, which was ignored in the standard equation. We propose Bayesian inversion methods to extract electrical properties of inhomogeneities in the main body, which is essential in medicine to screen the interior body and detect tumors or determine body composition. We also prove well-definedness of the posterior measure and well-posedness of the Bayesian inversion for the presented nonlinear model. The new model is able to distinguish between liquid and tissues and the state-of-the-art delayed-rejection adaptive-Metropolis (DRAM) algorithm is capable of analyzing the statistical variability in the measured data in various EIT experimental designs. This leads to design a reliable device with higher resolution images which is crucial in medicine for diagnostic purposes. We first test the validation of the presented nonlinear model and the proposed inverse method using synthetic data on a simple square computational domain with an inclusion. Then we establish the new model and robustness of the proposed inversion method in solving the ill-posed and nonlinear EIT inverse problem by presenting numerical results of the corresponding forward and inverse problems on a real-world application in medicine and healthcare. The results include the extraction of electrical properties of human leg tissues using measurement data.

We propose a mathematical model based on a system of partial differential equations (PDEs) for biofilms. This model describes the time evolution of growth and degradation of biofilms which depend on environmental factors. The proposed model also includes quorum sensing (QS) and describes the cooperation among bacteria when they need to resist against external factors such as antibiotics. The applications include biofilms on teeth and medical implants, in drinking water, cooling water towers, food processing, oil recovery, paper manufacturing, and on ship hulls. We state existence and uniqueness of solutions of the proposed model and implement the mathematical model to discuss numerical simulations of biofilm growth and cooperation. We also determine the unknown parameters of the presented biofilm model by solving the corresponding inverse problem. To this end, we propose Bayesian inversion techniques and the delayed-rejection adaptive-Metropolis (DRAM) algorithm for the simultaneous extraction of multiple parameters from the measurements. These quantities cannot be determined directly from the experiments or from the computational model. Furthermore, we evaluate the presented model by comparing the simulations using the estimated parameter values with the measurement data. The results illustrate a very good agreement between the simulations and the measurements.

We develop a meshless numerical procedure to simulate the groundwater equation (GWE). The used technique is based on the interpolating element free Galerkin (IEFG) method. The interpolating moving least squares (IMLS) approximation produces a set of functions such that they are well-known as shape functions. The IEFG technique employs the shape functions of IMLS approximation. The shape functions of IMLS approximation vanish on the boundary and also they satisfy the property of the Kronecker Delta function. Thus, Dirichlet boundary conditions can be exactly imposed. In this paper, we check the unconditional stability and convergence of the proposed numerical scheme based on the energy method. The numerical results confirm the theoretical analysis.

We present a simulation framework for computing the probability that a single molecule reaches the recognition element in a nanopore sensor. The model consists of the Langevin equation for the diffusive motion of small particles driven by external forces and the Poisson-Nernst-Planck-Stokes equations to compute these forces. The model is applied to examine DNA exo-sequencing in α-hemolysin, whose practicability depends on whether isolated DNA monomers reliably migrate into the channel in their correct order. We find that, at moderate voltage, migration fails in the majority of trials if the exonuclease which releases monomers is located farther than 1 nm above the pore entry. However, by tuning the pore to have a higher surface charge, applying a high voltage of 1 V and ensuring the exonuclease stays close to the channel, success rates of over 95% can be achieved.

Nanowire field-effect sensors have recently been developed for label-free detection of biomolecules. In this work, we introduce a computational technique based on Bayesian estimation to determine the physical parameters of the sensor and, more importantly, the properties of the analyte molecules. To that end, we first propose a PDE-based model to simulate the device charge transport and electrochemical behavior. Then, the adaptive Metropolis algorithm with delayed rejection is applied to estimate the posterior distribution of unknown parameters, namely molecule charge density, molecule density, doping concentration, and electron and hole mobilities. We determine the device and molecules properties simultaneously, and we also calculate the molecule density as the only parameter after having determined the device parameters. This approach makes it possible not only to determine unknown parameters, but it also shows how well each parameter can be determined by yielding the probability density function (pdf).

We investigate possible connections between two different implementations of the Poisson-Nernst-Planck (PNP) anomalous models used to analyze the electrical response of electrolytic cells. One of them is built in the framework of the fractional calculus and considers integro-differential boundary conditions also formulated by using fractional derivatives; the other one is an extension of the standard PNP model presented by Barsoukov and Macdonald, which can also be related to equivalent circuits containing constant phase elements (CPEs). Both extensions may be related to an anomalous diffusion with subdiffusive characteristics through the electrical conductivity and are able to describe the experimental data presented here. Furthermore, we apply the Bayesian inversion method to extract the parameter of interest in the analytical formulas of impedance. To resolve the corresponding inverse problem, we use the delayed-rejection adaptive-Metropolis algorithm (DRAM) in the context of Markov-chain Monte Carlo (MCMC) algorithms to find the posterior distributions of the parameter and the corresponding confidence intervals.

This work gives analytical results for a system of transport equations which is the underlying mathematical model for nanopore sensors and for all types of affinity-based nanowire sensors. This model consists of the Poisson equation for the electrostatic potential ensuring self-consistency and including interface conditions stemming from a homogenized boundary layer, the drift-diffusion equations describing the transport of charge carriers in the sensor, the Nernst–Planck equations describing the transport of ions, and the Stokes equations describing the flow of the background medium water. We present existence and local uniqueness theorems for this stationary, nonlinear, and fully coupled system. The existence proof is based on the Schauder fixed-point theorem and local uniqueness around equilibrium is obtained from the implicit-function theorem. The maximum principle is used to obtain a priori estimates for the solution. Due to the multiscale problem inherent in affinity-based field-effect sensors, a homogenized equation for the potential with interface conditions at a surface is used.

Dielectric structures composed of many inclusions that manipulate light in ways the bulk materials cannot are commonly seen in the field of metamaterials. In these structures, each inclusion depends on a set of parameters such as size and orientation, which are difficult to ascertain. We propose and implement an optimization-based approach for designing such metamaterials in two dimensions by using a fast boundary element method and a multiple-scattering solver for a given set of parameters. This approach provides the backbone of an automated process for the design and analysis of metamaterials that does not rely on analytical approximations. We demonstrate the validity of our approach with simulations that converge to optimal parameter values and result in substantially better performance.

Massively parallel nanosensor arrays fabricated with low-cost CMOS technology represent powerful platforms for biosensing in the Internet-of-Things (IoT) and Internet-of-Health (IoH) era. They can efficiently acquire “big data” sets of dependable calibrated measurements, representing a solid basis for statistical analysis and parameter estimation. In this paper we propose Bayesian estimation methods to extract physical parameters and interpret the statistical variability in the measured outputs of a dense nanocapacitor array biosensor. Firstly, the physical and mathematical models are presented. Then, a simple 1D-symmetry structure is used as a validation test case where the estimated parameters are also known a-priori. Finally, we apply the methodology to the simultaneous extraction of multiple physical and geometrical parameters from measurements on a CMOS pixelated nanocapacitor biosensor platform.

In this work, a modification procedure for the functionalization of silicon nanowire (SiNW) is applied in biological field effect transistor (BioFET) system. The proposed method precedes the silanization reaction in a manner that the only SiNW and not its substrate is functionalized by (3-Aminopropyl) triethoxysilane (APTES) initiators. This method has an effective role in increasing the sensitivity of BioFET sensors and can be applied in commercial ones. Furthermore, we introduce an efficient computational technique to estimate unknown senor parameters. To that end, Bayesian inversion is used to determine the number of PSA target molecules bound to the receptors in both selective and nonselective SiNWs. The approach is coupled with the Poisson-Boltzmann-drift-diffusion (PBDD) equations to provide a comprehensive system to model all biosensor interactions.

Purpose -- The current paper aims to develop a reduced order discontinuous Galerkin method for solving the generalized Swift–Hohenberg equation with application in biological science and mechanical engineering. The generalized Swift–Hohenberg equation is a fourth-order PDE; thus, this paper uses the local discontinuous Galerkin (LDG) method for it. Design/methodology/approach -- At first, the spatial direction has been discretized by the LDG technique, as this process results in a nonlinear system of equations based on the time variable. Thus, to achieve more accurate outcomes, this paper uses an exponential time differencing scheme for solving the obtained system of ordinary differential equations. Finally, to decrease the used CPU time, this study combines the proper orthogonal decomposition approach with the LDG method and obtains a reduced order LDG method. The circular and rectangular computational domains have been selected to solve the generalized Swift–Hohenberg equation. Furthermore, the energy stability for the semi-discrete LDG scheme has been discussed. Findings -- The results show that the new numerical procedure has not only suitable and acceptable accuracy but also less computational cost compared to the local DG without the proper orthogonal decomposition (POD) approach. Originality/value -- The local DG technique is an efficient numerical procedure for solving models in the fluid flow. The current paper combines the POD approach and the local LDG technique to solve the generalized Swift–Hohenberg equation with application in the fluid mechanics. In the new technique, the computational cost and the used CPU time of the local DG have been reduced.

We apply a previously developed approach for the automated design of optical structures to two cases. This approach reduces the basis of the electromagnetic system to obtain fast gradient-based optimization. In the first case, an existing photonic crystal demultiplexer is optimized for higher power transmission and lower crosstalk. In the second, new optical diodes for plane- and cylindrical-wave incidence are designed using a photonic crystal as a starting point. Highly efficient and aperiodic devices are obtained in all cases. These results indicate that aperiodic devices produced by this automated design method can outperform their analytically-obtained counterparts and encourage its application to other photonic crystal-based devices.

In this paper, we employ the multilevel Monte Carlo finite element method to solve the stochastic Cahn–Hilliard–Cook equation. The Ciarlet–Raviart mixed finite element method is applied to solve the fourth-order equation. In order to estimate the mild solution, we use finite elements for space discretization and the semi-implicit Euler–Maruyama method in time. For the stochastic scheme, we use the multilevel method to decrease the computational cost (compared to the Monte Carlo method). We implement the method to solve three specific numerical examples (both two- and three dimensional) and study the effect of different noise measures.

In this study, the direct meshless local Petrov–Galerkin (DMLPG) method has been employed to solve the stochastic Cahn–Hilliard–Cook and Swift–Hohenberg equations. First of all, we discretize the temporal direction by a finite difference scheme. In order to obtain a fully discrete scheme the direct meshless local collocation method is used to discretize the spatial variable and the Euler–Maruyama method is used for time discretization. The used method is a truly meshless technique. In order to illustrate the efficiency and accuracy of the explained numerical technique, we study two stochastic models with their applications in biology and engineering, i.e., the stochastic Cahn–Hilliard–Cook equation and a stochastic Swift–Hohenberg model.

The stochastic nonlinear Poisson–Boltzmann equation describes the electrostatic potential in a random environment in the presence of free charges and has applications in many fields. We show the existence and uniqueness of the solution of this nonlinear model equation and investigate its regularity with respect to a random parameter. Three popular nonintrusive methods, a stochastic Galerkin method, a discrete projection method, and a collocation method, are presented for its numerical solution. It is nonintrusive in the sense that solvers and preconditioners for the deterministic equation can be reused as they are. By comparing these methods, it is found that the stochastic Galerkin method and the discrete projection method require comparable computational effort and our results suggest that they outperform the collocation method.

ParticleScattering is a Julia (Bezanson et al. 2017) package for computing the electromagnetic fields scattered by a large number of two-dimensional particles, as well as optimizing particle parameters for various applications. Such problems naturally arise in the design and analysis of metamaterials, including photonic crystals (Jahani and Jacob 2016). Unlike most solvers for these problems, ours does not require a periodic structure and is scalable to a large number of particles. In particular, this software is designed for scattering problems involving TM plane waves impinging on a collection of homogeneous dielectric particles with arbitrary smooth shapes. Our code performs especially well when the number of particles is substantially larger than the number of distinct shapes, where particles are considered indistinct if they are identical up to rotation.

The three-dimensional stochastic drift-diffusion-Poisson system is used to model charge transport through nanoscale devices in a random environment. Applications include nanoscale transistors and sensors such as nanowire field-effect bio- and gas sensors. Variations between the devices and uncertainty in the response of the devices arise from the random distributions of dopant atoms, from the diffusion of target molecules near the sensor surface, and from the stochastic association and dissociation processes at the sensor surface. Furthermore, we couple the system of stochastic partial differential equations to a random-walk-based model for the association and dissociation of target molecules. In order to make the computational effort tractable, an optimal multi-level Monte–Carlo method is applied to three-dimensional solutions of the deterministic system. The whole algorithm is optimal in the sense that the total computational cost is minimized for prescribed total errors. This comprehensive and efficient model makes it possible to study the effect of design parameters such as applied voltages and the geometry of the devices on the expected value of the current.

The numerical solution of stochastic partial differential equations and numerical Bayesian estimation is computationally demanding. If the coefficients in a stochastic partial differential equation exhibit symmetries, they can be exploited to reduce the computational effort. To do so, we show that permutation-invariant functions can be approximated by permutation-invariant polynomials in the space of continuous functions as well as in the space of p-integrable functions defined on [0, 1]^s for 1 <=p < . We proceed to develop a numerical strategy to compute cubature formulas that exploit permutation-invariance properties related to multisymmetry groups in order to reduce computational work. We show that in a certain sense there is no curse of dimensionality if we restrict ourselves to multisymmetric functions, and we provide error bounds for formulas of this type. Finally, we present numerical results, comparing the proposed formulas to other integration techniques that are frequently applied to high-dimensional problems such as quasi-Monte Carlo rules and sparse grids.

In this paper, an optimal multilevel randomized quasi-Monte-Carlo method to solve the stationary stochastic drift–diffusion-Poisson system is developed. We calculate the optimal values of the parameters of the numerical method such as the mesh sizes of the spatial discretization and the numbers of quasi-points in order to minimize the overall computational cost for solving this system of stochastic partial differential equations. This system has a number of applications in various fields, wherever charged particles move in a random environment. It is shown that the computational cost of the optimal multilevel randomized quasi-Monte-Carlo method, which uses randomly shifted low-discrepancy sequences, is one order of magnitude smaller than that of the optimal multilevel Monte-Carlo method and five orders of magnitude smaller than that of the standard Monte-Carlo method. The method developed here is applied to a realistic transport problem, namely the calculation of random-dopant effects in nanoscale field-effect transistors.

The basic analytical properties of the drift-diffusion-Poisson-Boltzmann system in the alternating-current (AC) regime are shown. The analysis of the AC case differs from the direct-current (DC) case and is based on extending the transport model to the frequency domain and writing the variables as periodic functions of the frequency in a small-signal approximation. We first present the DC and AC model equations to describe the three types of material in nanowire field-effect sensors: The drift-diffusion-Poisson system holds in the semiconductor, the Poisson-Boltzmann equation holds in the electrolyte, and the Poisson equation provides self-consistency. Then the AC model equations are derived. Finally, existence and local uniqueness of the solution of the AC model equations are shown. Real-world applications include nanowire field-effect bio- and gas sensors operating in the AC regime, which were only demonstrated experimentally recently. Furthermore, nanopore sensors are governed by the system of model equations and the analysis as well.

We propose a design strategy for affinity-based biosensors using nanowires for sensing and measuring biomarker concentration in biological samples. Such sensors have been shown to have superior properties compared to conventional biosensors in terms of LOD (limit of detection), response time, cost, and size. However, there are several parameters affecting the performance of such devices that must be determined. In order to solve the design problem, we have developed a comprehensive model based on stochastic transport equations that makes it possible to optimize the sensing behavior.

We present a 3D finite element solver for the nonlinear Poisson--Nernst--Planck (PNP) equations for electrodiffusion, coupled to the Stokes system of fluid dynamics. The model serves as a building block for the simulation of macromolecule dynamics inside nanopore sensors. The source code is released online at github.com/mitschabaude/nanopores. We add to existing numerical approaches by deploying goal-oriented adaptive mesh refinement. To reduce the computation overhead of mesh adaptivity, our error estimator uses the much cheaper Poisson--Boltzmann equation as a simplified model, which is justified on heuristic grounds but shown to work well in practice. To address the nonlinearity in the full PNP–Stokes system, three different linearization schemes are proposed and investigated, with two segregated iterative approaches both outperforming a naive application of Newton’s method. Numerical experiments are reported on a real-world nanopore sensor geometry. We also investigate two different models for the interaction of target molecules with the nanopore sensor through the PNP--Stokes equations. In one model, the molecule is of finite size and is explicitly built into the geometry; while in the other, the molecule is located at a single point and only modeled implicitly -- after solution of the system -- which is computationally favorable. We compare the resulting force profiles of the electric and velocity fields acting on the molecule, and conclude that the point-size model fails to capture important physical effects such as the dependence of charge selectivity of the sensor on the molecule radius.

Existence and local-uniqueness theorems for weak solutions of a system consisting of the drift-diffusion-Poisson equations and the Poisson-Boltzmann equation, all with stochastic coefficients, are presented. For the numerical approximation of the expected value of the solution of the system, we develop a multi-level Monte-Carlo (MLMC) finite-element method (FEM) and we analyze its rate of convergence and its computational complexity. This allows to find the optimal choice of discretization parameters. Finally, numerical results show the efficiency of the method. Applications are, among others, noise and fluctuations in nanoscale transistors, in field-effect bio- and gas sensors, and in nanopores.

A basis-adaptation method based on polynomial chaos expansion is used for the stochastic nonlinear Poisson–Boltzmann equation. The uncertainty in this numerical approach is motivated by the quantification of noise and fluctuations in nanoscale field-effect sensors. The method used here takes advantage of the properties of the nonlinear Poisson–Boltzmann equation and shows an exact and efficient approximation of the real solution. Numerical examples are motivated by the quantification of noise and fluctuations in nanowire field-effect sensors as a concrete example. Basis adaptation is validated by comparison with the full solution, and it is compared to optimized multi-level Monte-Carlo method, and the model equations are validated by comparison with experiments. Finally, various design parameters of the field-effect sensors are investigated in order to maximize the signal-to-noise ratio.

This work addresses the development of an optimal computational scheme for the approximation of the first-order corrector arising in the stochastic homogenization of linear elliptic PDEs in divergence form. Equations of this type describe, for example, diffusion phenomena in materials with a heterogeneous microstructure, but require enormous computational efforts in order to obtain reliable results. We derive an optimization problem for the needed computational work with a given error tolerance, then extract the governing parameters from numerical experiments, and finally solve the obtained optimization problem. The numerical approach investigated here is a stochastic sampling scheme for the probability space connected with a finite-element method for the discretization of the physical space.

Conclusions: bSCr values back-estimated using currently available eGFR formulae are inaccurate and cannot correctly classify AKI stages. Our model eSCr improves the prediction of AKI but to a still inadequate extent.

In this work, we develop a 2D algorithm for stochastic reaction-diffusion systems describing the binding and unbinding of target molecules at the surfaces of affinity-based sensors. In particular, we simulate the detection of DNA oligomers using silicon-nanowire field-effect biosensors. Since these devices are uniform along the nanowire, two dimensions are sufficient to capture the kinetic effects features. The model combines a stochastic ordinary differential equation for the binding and unbinding of target molecules as well as a diffusion equation for their transport in the liquid. A Brownian-motion based algorithm simulates the diffusion process, which is linked to a stochastic-simulation algorithm for association at and dissociation from the surface. The simulation data show that the shape of the cross section of the sensor yields areas with significantly different target-molecule coverage. Different initial conditions are investigated as well in order to aid rational sensor design. A comparison of the association/hybridization behavior for different receptor densities allows optimization of the functionalization setup depending on the target-molecule density.

A transport equation for confined structures is used to calculate the ionic currents through various transmembrane proteins. The transport equation is a diffusion-type equation where the concentration of the particles depends on the one-dimensional position in the confined structure and on the local energy. The computational significance of this continuum model is that the (6+1)-dimensional Boltzmann equation is reduced to a (2+1)-dimensional diffusion-type equation that can be solved with small computational effort so that ionic currents through confined structures can be calculated quickly. The applications here are three channels, namely OprP, Gramicidin A, and KcsA. In each case, the confinement potential is estimated from the known molecular structure of the channel. Then the confinement potentials are used to calculate ionic currents and to study the effect of parameters such as the potential of mean force, the ionic bath concentration, and the applied voltage. The simulated currents are compared with measurements, and very good agreement is found in each case. Finally, virtual potassium channels with selectivity filters of varying length are simulated in order to discuss the optimality of the filter.

In this work, we calculate the effect of the binding and unbinding of molecules at the surface of a nanowire biosensor on the signal-to-noise ratio of the sensor. We model the fluctuations induced by association and dissociation of target molecules by a stochastic differential equation and extend this approach to a coupled diffusion-reaction system. Where possible, analytic solutions for the signal-to-noise ratio are given. Stochastic simulations are performed wherever closed forms of the solutions cannot be derived. Starting from parameters obtained from experimental data, we simulate DNA hybridization at the sensor surface for different target molecule concentrations in order to optimize the sensor design.

We review transport equations and their usage for the modeling and simulation of nanopores. First, the significance of nanopores and the experimental progress in this area are summarized. Then the starting point of all classical and semiclassical considerations is the Boltzmann transport equation as the most general transport equation. The derivation of the drift-diffusion equations from the Boltzmann equation is reviewed as well as the derivation of the Navier-Stokes equations. Nanopores can also be viewed as a special case of a confined structure and hence as giving rise to a multiscale problem, and therefore we review the derivation of a transport equation from the Boltzmann equation for such confined structures. Finally, the state of the art in the simulation of nanopores is summarized.

In this work, the multiscale problem of modeling fluctuations in boundary layers in stochastic elliptic partial differential equations is solved by homogenization. A homogenized equation for the covariance of the solution of stochastic elliptic PDEs is derived. In addition to the homogenized equation, a rate for the covariance and variance as the cell size tends to zero is given. For the homogenized problem, an existence and uniqueness result and further properties are shown. The multiscale problem stems from the modeling of the electrostatics in nanoscale field-effect sensors, where the fluctuations arise from random charge concentrations in the cells of a boundary layer. Finally, numerical results and a numerical verification are presented.

We present a convergent finite-difference scheme of second order in both space and time for the 2D electromagnetic Dirac equation. We apply this method in the self-consistent Dirac-Poisson system to the simulation of graphene. The model is justified for low energies, where the particles have wave vectors sufficiently close to the Dirac points. In particular, we demonstrate that our method can be used to calculate solutions of the Dirac-Poisson system where potentials act as beam-splitters or Veselago lenses.

In this work, we present calculated numerical values for the kinetic parameters governing adsorption/desorption processes of carbon monoxide at tin dioxide single-nanowire gas sensors. The response of such sensors to pulses of 50ppm carbon monoxide in nitrogen is investigated at different temperatures to extract the desired information. A rate-equation approach is used to model the reaction kinetics, which results in the problem of determining coefficients in a coupled system of nonlinear ordinary differential equations. The numerical values are computed by inverse-modeling techniques and are then used to simulate the sensor response. With our model, the dynamic response of the sensor due to the gas–surface interaction can be studied in order to find the optimal setup for detection, which is an important step towards selectivity of these devices. We additionally investigate the noise in the current through the nanowire and its changes due to the presence of carbon monoxide in the sensor environment. Here, we propose the use of a wavelet transform to decompose the signal and analyze the noise in the experimental data. This method indicates that some fluctuations are specific for the gas species investigated here.

A 3d FETI method for the drift-diffusion-Poisson system including discontinuities at a 2d interface is developed. The motivation for this work is to provide a parallel numerical algorithm for a system of PDEs that are the basic model equations for the simulation of semiconductor devices such as transistors and sensors. Moreover, discontinuities or jumps in the potential and its normal derivative at a 2d surface are included for the simulation of nanowire sensors based on a homogenized model. Using the FETI method, these jump conditions can be included with the usual numerical properties and the original Farhat-Roux FETI method is extended to the drift-diffusion-Poisson equations including discontinuities. We show two numerical examples. The first example verifies the correct implementation including the discontinuities on a 2d grid divided into eight subdomains. The second example is 3d and shows the application of the algorithm to the simulation of nanowire sensors with high aspect ratios. The Poisson-Boltzmann equation and the drift-diffusion-Poisson system with jump conditions are solved on a 3d grid with real-world boundary conditions.

We apply our self-consistent PDE model for the electrical response of field-effect sensors to the 3D simulation of nanowire PSA (prostate-specific antigen) sensors. The charge concentration in the biofunctionalized boundary layer at the semiconductor-electrolyte interface is calculated using the PROPKA algorithm, and the screening of the biomolecules by the free ions in the liquid is modeled by a sensitivity factor. This comprehensive approach yields excellent agreement with experimental current-voltage characteristics without any fitting parameters. Having verified the numerical model in this manner, we study the sensitivity of nanowire PSA sensors by changing device parameters, making it possible to optimize the devices and revealing the attributes of the optimal field-effect sensor.

Nanowire gas sensors show high sensitivity towards various gases and offer great potential to improve present gas sensing. In this work, we investigate experimental results achieved with an undoped single SnO₂ nanowire sensor device for CO pulses in N₂ atmosphere at different operating temperatures. We calculated the reaction parameters according to the mass action law including frequency factors, activation energies, and numbers of intrinsic as well as extrinsic surface sites. With the values obtained, we then calculated the surface charge of the nanowire sensor by solving the corresponding differential equations. The simulated results agree very well with the experimental values at an operating temperature of 200°C and hence provide good understanding of the chemical reaction. This can be used to simulate the current through the transducer and consequently the sensitivity of the device, and the parameters provided here are useful for computational procedures to provide selectivity.

A book chapter. Contents:
1. Introduction
2. Homogenization
3. The biofunctionalized boundary layer
4. The current through the nanowire transducer
5. Summary

For the development of nanowire sensors for chemical and medical detection purposes, the optimal functionalization of the surface is a mandatory component. Quantitative ATR-FTIR spectroscopy was used in-situ to investigate the step-by-step layer formation of typical functionalization protocols and to determine the respective molecule surface concentrations. BSA, anti-TNF-α and anti-PSA antibodies were bound via 3-(trimethoxy)butylsilyl aldehyde linkers to silicon-oxide surfaces in order to investigate surface functionalization of nanowires. Maximum determined surface concentrations were 7.17×10^-13 mol cm^-2 for BSA, 1.7×10^-13 mol cm^-2 for anti-TNF-α antibody, 6.1×10^-13 mol cm^-2 for anti-PSA antibody, 3.88×10^-13 mol cm^-2 for TNF-α and 7.0×10^-13 mol cm^-2 for PSA. Furthermore we performed antibody-antigen binding experiments and determined the specific binding ratios. The maximum possible ratios of 2 were obtained at bulk concentrations of the antigen in the μg ml^-1 range for TNF-α and PSA.

We present existence and local uniqueness theorems for a system of partial differential equations modeling field-effect nano-sensors. The system consists of the Poisson(-Boltzmann) equation and the drift-diffusion equations coupled with a homogenized boundary layer. The existence proof is based on the Leray-Schauder fixed-point theorem and a maximum principle is used to obtain a-priori estimates for the electric potential, the electron density, and the hole density. Local uniqueness around the equilibrium state is obtained from the implicit-function theorem. Due to the multiscale problem inherent in field-effect biosensors, a homogenized equation for the potential with interface conditions at a surface is used. These interface conditions depend on the surface-charge density and the dipole-moment density in the boundary layer and still admit existence and local uniqueness of the solution when certain conditions are satisfied. Due to the geometry and the boundary conditions of the physical system, the three-dimensional case must be considered in simulations. Therefore a finite-volume discretization of the 3d self-consistent model was implemented to allow comparison of simulation and measurement. Special considerations regarding the implementation of the interface conditions are discussed so that there is no computational penalty when compared to the problem without interface conditions. Numerical simulation results are presented and very good quantitative agreement with current-voltage characteristics from experimental data of biosensors is found.

In order to facilitate the rational design and the characterization of nanowire field-effect sensors, we have developed a model based on self-consistent charge-transport equations combined with interface conditions for the description of the biofunctionalized surface layer at the semiconductor/electrolyte interface. Crucial processes at the interface, such as the screening of the partial charges of the DNA strands and the influence of the angle of the DNA strands with respect to the nanowire, are computed by a Metropolis Monte Carlo algorithm for charged molecules at interfaces. In order to investigate the sensing mechanism of the device, we have computed the current–voltage characteristics, the electrostatic potential and the concentrations of electrons and holes. Very good agreement with measurements has been found and optimal device parameters have been identified. Our approach provides the capability to study the device sensitivity, which is of fundamental importance for reliable sensing.

Field-effect biosensors based on nanowires enjoy considerable popularity due to their high sensitivity and direct electrical readout. However, crucial issues such as the influence of the biomolecules on the charge-carrier transport or the binding of molecules to the surface have not been described satisfactorily yet in a quantitative manner. In order to analyze these effects, we present simulation results based on a 3D macroscopic transport model coupled with Monte-Carlo simulations for the bio-functionalized surface layer. Excellent agreement with measurement data has been found, while detailed study of the influence of the most prominent biomolecules, namely double-stranded DNA and single-stranded DNA, on the current through the semiconductor transducer has been carried out.

A system of diffusion-type equations for transport in 3d confined structures is derived from the Boltzmann transport equation for charged particles. Transport takes places in confined structures and the scaling in the derivation of the diffusion equation is chosen so that transport and scattering occur in the longitudinal direction and the particles are confined in the two transversal directions. The result are two diffusion-type equations for the concentration and fluxes as functions of position in the longitudinal direction and energy. Entropy estimates are given. The transport coefficients depend on the geometry of the problem that is given by arbitrary harmonic confinement potentials. An important feature of this approach is that the coefficients in the resulting diffusion-type equations are calculated explicitly so that the six position and momentum dimensions of the original 3d Boltzmann equation are reduced to a 2d problem. Finally, numerical results are given and discussed. Applications of this work include the simulation of charge transport in nanowires, nanopores, ion channels, and similar structures.

In this work, a Monte-Carlo algorithm in the constant-voltage ensemble for the calculation of 3d charge concentrations at charged surfaces functionalized with biomolecules is presented. The motivation for this work is the theoretical understanding of biofunctionalized surfaces in nanowire field-effect biosensors (BioFETs). This work provides the simulation capability for the boundary layer that is crucial in the detection mechanism of these sensors; slight changes in the charge concentration in the boundary layer upon binding of analyte molecules modulate the conductance of nanowire transducers. The simulation of biofunctionalized surfaces poses special requirements on the Monte-Carlo simulations and these are addressed by the algorithm. The constant-voltage ensemble enables us to include the right boundary conditions; the DNA strands can be rotated with respect to the surface; and several molecules can be placed in a single simulation box to achieve good statistics in the case of low ionic concentrations relevant in experiments. Simulation results are presented for the leading example of surfaces functionalized with PNA and with single- and double-stranded DNA in a sodium-chloride electrolyte. These quantitative results make it possible to quantify the screening of the biomolecule charge due to the counter-ions around the biomolecules and the electrical double layer. The resulting concentration profiles show a three-layer structure and non-trivial interactions between the electric double layer and the counter-ions. The numerical results are also important as a reference for the development of simpler screening models.

Fluctuations in the biofunctionalized boundary layers of nanowire field-effect biosensors are investigated by using the stochastic linearized Poisson-Boltzmann equation. The noise and fluctuations considered here are due to the Brownian motion of the biomolecules in the boundary layer, i.e., the various orientations of the molecules with respect to the surface are associated with their probabilities. The probabilities of the orientations are calculated using their free energy. The fluctuations in the charge distribution give rise to fluctuations in the electrostatic potential and hence in the current through the semiconductor transducer of the sensor, both of which are calculated. A homogenization result for the variance and covariance of the electrostatic potential is presented. In the numerical simulations, a cross section of a silicon nanowire on a flat surface including electrode and back-gate contacts is considered. The biofunctionalized boundary layer contains single-stranded or double-stranded DNA oligomers, and varying values of the surface charge, of the oligomer length, and of the electrolyte ionic strength are investigated.

Field-effect nanobiosensors (or BioFETs, biologically sensitive field-effect transistors) have recently been demonstrated experimentally and have thus gained interest as a technology for direct, label-free, real-time, and highly sensitive detection of biomolecules. The experiments have not been accompanied by a quantitative understanding of the underlying detection mechanism. The modeling of field-effect biosensors poses a multiscale problem due to the different length scales in the sensors: the charge distribution and the electric potential of the biofunctionalized surface layer changes on the Angstrom length scale, whereas the exposed sensor area is measured in micrometers squared. Here a multiscale model for the electrostatics of planar and nanowire field-effect sensors is developed by homogenization of the Poisson equation in the biofunctionalized boundary layer. The resulting interface conditions depend on the surface charge density and dipole moment density of the boundary layer. The multiscale model can be coupled to any charge transport model and hence makes the self-consistent quantitative investigation of the physics of field-effect sensors possible. Numerical verifications of the multiscale model are given. Furthermore a silicon nanowire biosensor is simulated to elucidate the influence of the surface charge density and the dipole moment density on the conductance of the semiconductor transducer. The numerical evidence shows that the conductance varies exponentially as a function of both charge and dipole moment. Therefore the dipole moment of the surface layer must be included in biosensor models. The conductance variations observed in experiments can be explained by the field effect, and they can be caused by a change in dipole moment alone.

BioFETs (biologically sensitive field-effect transistors) are field-effect biosensors with semiconducting transducers. Their device structure is similar to a MOSFET, except that the gate structure is replaced by an aqueous solution containing the analyte. The detection mechanism is the conductance modulation of the transducer due to binding of the analyte to surface receptors. The main advantage of BioFETs, compared to currently available technology, is label-free operation. We present a quantitative analysis of BioFETs which is centered around multi-scale models. The technique for solving the multi-scale problem used here is the derivation of interface conditions for the Poisson equation that include the effects of the quasi-periodic biofunctionalized boundary layer. The multi-scale model enables self-consistent simulation and can be used with any charge transport model. Hence it provides the foundation for understanding the physics of the sensors by continuum models.

BioFETs (biologically active field-effect transistors) are biosensors with a semiconductor transducer. Due to recent experiments demonstrating detection by a field effect, they have gained attention as potentially fast, reliable, and low-cost biosensors for a wide range of applications. Their advantages compared to other technologies are direct, label-free, ultra-sensitive, and (near) real-time operation. We have developed 2D and 3D multi-scale models for planar sensor structures and for nanowire sensors. The multi-scale models are indispensable due to the large difference in the characteristic length scales of the biosensors: the charge distribution in the biofunctionalized surface layer varies on the Angstrom length scale, the diameters of the nanowires are several nanometers, and the sensor lengths measure several micrometers. The multi-scale models for the electrostatic potential can be coupled to any charge transport model of the transducer. Conductance simulations of nanowire sensors with different diameters provide numerical evidence for the importance of the dipole moment of the biofunctionalized surface layer in addition to its surface charge. We have also developed a web interface to our simulators, so that other researchers can access them at the nanohub and perform their own investigations.

The solutions of the nonlinear Schrödinger equation are of great importance for ab initio calculations. It can be shown that such solutions conserve a countable number of quantities, the simplest being the local norm square conservation law. Numerical solutions of high quality, especially for long time intervals, must necessarily obey these conservation laws. In this work we first give the conservation laws that can be calculated by means of Lie theory and then critically compare the quality of different finite difference methods that have been proposed in geometric integration with respect to conservation laws. We find that finite difference schemes derived by writing the Schrödinger equation as an (artificial) Hamiltonian system do not necessarily conserve important physical quantities better than other methods.

The classical Coulomb potential and force can be calculated efficiently using fast multi-pole methods. Effective quantum potentials, however, describe the physics of electron transport in semiconductors more precisely. Such an effective quantum potential was derived previously for the interaction of an electron with a barrier for use in particle-based Monte Carlo semiconductor device simulators. The method is based on a perturbation theory around thermodynamic equilibrium and leads to an effective potential scheme in which the size of the electron depends upon its energy and which is parameter-free. Here we extend the method to electron-electron interactions and show how the effective quantum potential can be evaluated efficiently in the context of many-body problems. Finally several examples illustrate how the momentum of the electrons changes the classical potential.

In recent years DNA-sensors, and generally biosensors, with semiconducting transducers were fabricated and characterized. Although the concept of so-called BioFETs was proposed already two decades ago, its realization has become feasible only recently due to advances in process technology. In this paper a comprehensive and rigorous approach to the simulation of silicon-nanowire DNAFETs at the feature-scale is presented. It allows to investigate the feasibility of single-molecule detectors and is used to elucidate the performance that can be expected from sensors with nanowire diameters in the deca-nanometer range. Finally the computational challenges for the simulation of silicon-nanowire DNA-sensors are discussed.

Effective quantum potentials describe the physics of quantum-mechanical electron transport in semiconductors more than the classical Coulomb potential. An effective quantum potential was derived previously for the interaction of an electron with a barrier for use in particle-based Monte Carlo semiconductor device simulators. The method is based on a perturbation theory around thermodynamic equilibrium and leads to an effective potential scheme in which the size of the electron depends upon its energy and which is parameter-free. Here we extend the method to electron-electron interactions and show how the effective quantum potential can be evaluated efficiently in the context of many-body problems. The effective quantum potential was used in a three-dimensional Monte-Carlo device simulator for calculating the electron-electron and electron-barrier interactions. Simulation results for an SOI transistor are presented and illustrate how the effective quantum potential changes the characteristics compared to the classical potential.

This paper presents an anisotropic adaptation strategy for three-dimensional unstructured tetrahedral meshes, which allows us to produce thin mostly anisotropic layers at the outside margin, i.e., the skin of an arbitrary meshed simulation domain. An essential task for any modern algorithm in the finite-element solution of partial differential equations, especially in the field of semiconductor process and device simulation, the major application is to provide appropriate resolution of the partial discretization mesh. The start-up conditions for semiconductor process and device simulations claim an initial mesh preparation that is performed by so-called Laplace refinement. The basic idea is to solve Laplace’s equation on an initial coarse mesh with Dirichlet boundary conditions. Afterward, the gradient field is used to form an anisotropic metric that allows to refine the initial mesh based on tetrahedral bisection.

The quality of the numeric approximation of the partial differential equations governing carrier transport in semiconductor devices depends particularly on the grid. The method of choice is to use structurally aligned grids since the regions and directions therein that determine device behavior are usually straightforward to find as they depend on the distribution of doping. Here, the authors present an algorithm for generating structurally aligned grids including anisotropy with resolutions varying over several orders of magnitude. The algorithm is based on a level set approach and permits to define the refined resolutions in a flexible manner as a function of doping. Furthermore, criteria on grid quality can be enforced. In order to show the practicability of this method, the authors study the examples of a trench gate metal-oxide-semiconductor field-effect transistor (TMOSFET) and a radio frequency silicon-on-insulator lateral double diffused metal-oxide-semiconductor (RF SOI LDMOS) power device using the device simulator MINIMOS NT, where simulations are performed on a grid generated by the new algorithm. In order to resolve the interesting regions of the TMOSFET and the RF SOI LDMOS power device accurately, several regions of refinement were defined where the grid was grown with varying resolutions.

In state-of-the-art devices, it is well known that quantum and Coulomb effects play significant role on the device operation. In this paper, we demonstrate that a novel effective potential approach in conjunction with a Monte Carlo device simulation scheme can accurately capture the quantum-mechanical size quantization effects. We also demonstrate, via proper treatment of the short-range Coulomb interactions, that there will be significant variation in device design parameters for devices fabricated on the same chip due to the presence of unintentional dopant atoms at random locations within the channel.

We propose a novel parameter-free quantum potential scheme for use in conjunction with particle-based simulations. The method is based on a perturbation theory around thermodynamic equilibrium and leads to an effective potential scheme in which the size of the electron depends upon its energy. The approach has been tested on the example of a MOS-capacitor by retrieving the correct sheet electron density. It has also been used in simulations of a 25 nm n-channel nanoscale MOSFET with high substrate doping density. We find that the use of the quantum potential approach gives rise to a threshold voltage shift of about 220 mV and drain current degradation of about 30%.

Novel device concepts such as dual gate SOI, Ultra thin body SOI, FinFETs, etc., have emerged as a solution to the ultimate scaling limits of conventional bulk MOSFETs. These novel devices suppress some of the Short Channel Effects (SCE) efficiently, but at the same time more physics based modeling is required to investigate device operation. In this paper, we use semi-classical 3D Monte Carlo device simulator to investigate important issues in the operation of FinFETs. Fast Multipole Method (FMM) has been integrated with the EMC scheme to replace the time consuming Poisson equation solver. Effect of unintentional doping for different device dimensions has been investigated. Impurities at the source side of the channel have most significant impact on the device performance.

An approach for determining higher order coefficients of the electrical and thermal conductivities for different materials is presented. The method is based on inverse modeling using three-dimensional transient electrothermal finite element simulations for electrothermal investigations of complex layered structures, for instance polycrystalline silicon (polysilicon) fuses or other multi-layered devices. The simulations are performed with a three-dimensional interconnect simulator, which is automatically configured and controlled by an optimization framework. Our method is intended to be applied to optimize devices with different material compositions and geometries as well as for achieving an optimum of speed and reliability.

One of the challenges that technology computer-aided design must meet currently is the analysis of the performance of groups of components, interconnects, and, generally speaking, large parts of the IC. This enables predictions that the simulation of single components cannot achieve. In this paper, we focus on the simulation of backend processes, interconnect capacitances, and time delays. The simulation flows start from the blank wafer surface and result in device information for the circuit designer usable from within SPICE. In order to join topography and backend simulations, deposition, etching, and chemical mechanical planarization processes in the various metal lines are used to build up the backend stack, starting from the flat wafer surface. Depending on metal combination, line-to-line space, and line width, thousands of simulations are required whose results are stored in a database. Finally, we present simulation results for the backend of a 100-nm process, where the influence of void formation between metal lines profoundly impacts the performance of the whole interconnect stack, consisting of aluminum metal lines, and titanium nitride local interconnects. Scanning electron microscope images of test structures are compared to topography simulations, and very good agreement is found. Moreover, charge-based capacitance measurements were carried out to validate the capacitance extraction, and it was found that the error is smaller than four percent. These simulations assist the consistent fabrication of voids, which is economically advantageous compared to low-κ materials, which suffer from integration problems.

We present an algorithm for smoothing results of three-dimensional Monte Carlo ion implantation simulations and translating them from the grid used for the Monte Carlo simulation to an arbitrary unstructured three-dimensional grid. This algorithm is important for joining various simulations of semiconductor manufacturing process steps, where data have to be smoothed or transferred from one grid to another. Furthermore different grids must be used since using ortho-grids is mandatory because of performance reasons for certain Monte Carlo simulation methods. The algorithm is based on approximations by generalized Bernstein polynomials. This approach was put on a mathematically sound basis by proving several properties of these polynomials. It does not suffer from the ill effects of least squares fits of polynomials of fixed degree as known from the popular response surface method. The smoothing algorithm which works very fast is described and in order to show its applicability, the results of smoothing a three-dimensional real world implantation example are given and compared with those of a least squares fit of a multivariate polynomial of degree 2, which yielded unusable results.

We evaluate optimization techniques to reduce the necessary user interaction for inverse modeling applications as they are used in the technology computer-aided design field. Four optimization strategies are compared. Two well-known global optimization methods, simulated annealing and genetic optimization, a local gradient-based optimization strategy, and a combination of a local and a global method. We rate the applicability of each method in terms of the minimal achievable target value for a given number of simulation runs and in terms of the fastest convergence. A brief overview over the three used optimization algorithms is given. The optimization framework that is used to distribute the workload over a cluster of workstations is described. The actual comparison is achieved by means of an inverse modeling application that is performed for various settings of the optimization algorithms. All presented optimization algorithms are capable of evaluating several targets in parallel. The best optimization strategy that is found is used in the calibration of a model for silicon self-interstitial cluster formation and dissolution.

Numerically solving the nonlinear Schrödinger equation and being able to treat arbitrary space dependent potentials permits many application in the realm of quantum mechanics. The long-term stability of a numerical method and its conservation properties is an important feature since it assures that the underlying physics of the solution are respected and it ensures that the numerical result is correct also for small time spans. In this paper we describe symplectic integrators for the nonlinear Schrödinger equation with arbitrary potentials and perform numerical experiments comparing different approaches and highlighting their respective advantages and disadvantages.

The formation and dissolution of silicon self-interstitial clusters is linked to the phenomenon of transient-enhanced diffusion (TED) which in turn has gained importance in the manufacturing of semiconductor devices. Based on theoretical considerations and measurements of the number of self-interstitial clusters during a thermal step, a model for the formation and dissolution of self-interstitial clusters is presented including the adjusted model parameters for two different technologies (i.e. material parameter sets). In order to automate the inverse modeling part, a general optimization framework was used. In addition to solving this problem, the same setup can solve a wide range of inverse modeling problems occurring in the domain of process simulation. Finally, the results are discussed and compared with a previous model.

An algorithm for smoothing results of three-dimensional (3-D) Monte Carlo ion implantation simulations and translating them from the grid used for the Monte Carlo simulation to an arbitrary unstructured 3-D grid is presented. This algorithm is important for joining various process simulation steps, where data have to be smoothed or transferred from one grid to another. Furthermore, it is important for integrating the ion implantation simulator into a process flow. One reason for using different grids is that for certain Monte Carlo simulation methods, using orthogrids is mandatory because of performance reasons.

The algorithm presented sweeps a small rectangular grid over the points of the new tetrahedral grid and uses approximation by generalized Bernstein polynomials. This approach was put on a mathematically sound basis by proving several properties of these polynomials. It does not suffer from the adverse effects of least squares fits of polynomials of fixed degree as known from the response surface method.

The most important properties of Bernstein polynomials generalized to cuboid domains are presented, including uniform convergence, an asymptotic formula, and the variation diminishing property. The smoothing algorithm which works very fast is described and, in order to show its applicability, the resulting values of a 3-D real world implantation example are given and compared with those of a least squares fit of a multivariate polynomial of degree two, which yielded unusable results.

Filling high aspect ratio trenches is an essential manufacturing step for state of the art memory cells. Understanding and simulating the transport and surface processes enables to achieve voidless filling of deep trenches, to predict the resulting profiles, and thus to optimize the process parameters and the resulting memory cells.

Experiments of arsenic doped polysilicon deposition show that under certain process conditions step coverages greater than unity can be achieved. We developed a new model for the simulation of arsenic doped polysilicon deposition, which takes into account surface coverage dependent sticking coefficients and surface coverage dependent arsenic incorporation and desorption rates. The additional introduction of Langmuir--Hinshelwood type time dependent surface coverage enabled the reproduction of the bottom up filling of the trenches in simulations. Additionally, the rigorous treatment of the time dependent surface coverage allows to trace the in situ doping of the deposited film.

The model presented was implemented and simulations were carried out for different process parameters. Very good agreement with experimental data was achieved with theoretically deduced parameters. Simulation results are shown and discussed for polysilicon deposition into 0.1μm wide and 7μm deep, high aspect ratio trenches.

The shape of the hot electron distribution function in semiconductor devices is insufficiently described using only the first four moments. We propose using six moments of the distribution function to obtain a more accurate description of hot carrier phenomena. An analytic expression for the symmetric part of the distribution function as a function of the even moments is given which shows good agreement with Monte Carlo data for both the bulk case and inside n⁺-n-n⁺ test structures. The influence of the band structure on the parameters of the distribution function is studied and proven to be of importance for an accurate description.

The SIESTA framework is an extensible tool for optimization and inverse modeling of semiconductor devices including dynamic load balancing for taking advantage of several, loosely connected workstations. Two gradient-based and two evolutionary computation optimizers are currently available through a uniform interface and can be combined at will. At a real world inverse modeling example, we demonstrate that evolutionary computation optimizers provide several advantages over gradient-based optimizers, due to the specific properties of the objective functions in TCAD applications. Furthermore, we shortly discuss some issues arising in inverse modeling and conclude with a comparison of gradient-based and evolutionary computation optimizers from a TCAD point of view.

Conventional macroscopic impact ionization models which use the average carrier energy as a main parameter can not accurately describe the phenomenon in modern miniaturized devices. Here, we present a model which is based on an analytic expression for the distribution function. In particular, the distribution function model accounts explicitly for a hot and a cold carrier population in the drain region of metal-oxide-semiconductor transistors. The parameters are determined by three-even moments obtained from a solution of a six-moments transport model. Together with a nonparabolic description of the density of states, accurate closed form macroscopic impact ionization models can be derived based on familiar microscopic descriptions.