Analog monolayer SWCNTs-based memristive 2D structure for energy-efficient deep learning in spiking neural networks.

Advances in materials science and memory devices work in tandem for the evolution of Artificial Intelligence systems. Energy-efficient computation is the ultimate goal of emerging memristor technology, in which the storage and computation can be done in the same memory crossbar. In this work, an analog memristor device is fabricated utilizing the unique characteristics of single-wall carbon nanotubes (SWCNTs) to act as the switching medium of the device. Via the planar structure, the memristor device exhibits analog switching ability with high state stability. The device's conductance and capacitance can be tuned simultaneously, increasing the device's potential and broadening its applications' horizons. The multi-state storage capability and long-term memory are the key factors that make the device a promising candidate for bio-inspired computing applications. As a demonstrator, the fabricated memristor is deployed in spiking neural networks (SNN) to exploit its analog switching feature for energy-efficient classification operation. Results reveal that the computation-in-memory implementation performs Vector Matrix Multiplication with 95% inference accuracy and few femtojoules per spike energy efficiency. The memristor device presented in this work opens new insights towards utilizing the outstanding features of SWCNTs for efficient analog computation in deep learning systems.

A surface treatment method for improving the attachment of PDMS: acoustofluidics as a case study.

A method for a permanent surface modification of polydimethylsiloxane (PDMS) is presented. A case study on the attachment of PDMS and the lithium niobate (LiNbO3) wafer for acoustofluidics applications is presented as well. The method includes a protocol for chemically treating the surface of PDMS to strengthen its bond with the LiNbO3 surface. The PDMS surface is modified using the 3-(trimethoxysilyl) propyl methacrylate (TMSPMA) silane reagent. The effect of silane treatment on the hydrophilicity, morphology, adhesion strength to LiNbO3, and surface energy of PDMS is investigated. The results demonstrated that the silane treatment permanently increases the hydrophilicity of PDMS and significantly alters its morphology. The bonding strength between PDMS and LiNbO3increased with the duration of the silane treatment, reaching a maximum of approximately 500 kPa. To illustrate the effectiveness of this method, an acoustofluidic device was tested, and the device demonstrated very promising enhanced bonding and sealing capabilities with particle manipulation at a flow rate of up to 1 L/h by means of traveling surface acoustic waves (TSAW). The device was reused multiple times with no fluid leakage or detachment issues. The utility of the presented PDMS surface modification method is not limited to acoustofluidics applications; it has the potential to be further investigated for applications in various scientific fields in the future.

Acoustofluidic lysis of cancer cells and Raman spectrum profiling.

The lysis of cancer cells inside a sessile droplet was performed using traveling surface acoustic waves (SAWs) without any chemical reagents. Raman spectrum profiling was then carried out to explore detailed cell-derived data. The Rayleigh waves formed by an interdigital transducer were made to propagate along the surface of an LiNbO3 substrate. Polystyrene microparticles (PSMPs) were used to establish mechanical cell lysis effectively, and gold nanoparticles (AuNPs) were added to enhance the Raman signals from the lysed cells by SAWs. The lysis efficiency was evaluated according to the size and concentration of the PSMPs in experiments where the frequency was varied. Lysis occurred mainly by mechanical collision using PSMPs in a high-frequency domain, and the lysis efficiency was improved by increasing the application time and the energy density of the SAWs. Raman signals from the lysed cells were greatly enhanced by nanogaps formed by the AuNPs, which were evenly distributed irrespective of the SAWs through the frequency-independent behavior of the AuNPs. Finally, detailed Raman spectra of MDA-MB-231, malignant breast cancer cells, were acquired, and various organic matter-derived peaks were observed. The 95% confidence region for cells subjected to lysis was more widely distributed than that of cells not subjected to lysis. The proposed SAW platform is expected to facilitate the detection of small quantities and to be applied in biomedical applications.

A review on microfluidic-assisted nanoparticle synthesis, and their applications using multiscale simulation methods.

Recent years have witnessed an increased interest in the development of nanoparticles (NPs) owing to their potential use in a wide variety of biomedical applications, including drug delivery, imaging agents, gene therapy, and vaccines, where recently, lipid nanoparticle mRNA-based vaccines were developed to prevent SARS-CoV-2 causing COVID-19. NPs typically fall into two broad categories: organic and inorganic. Organic NPs mainly include lipid-based and polymer-based nanoparticles, such as liposomes, solid lipid nanoparticles, polymersomes, dendrimers, and polymer micelles. Gold and silver NPs, iron oxide NPs, quantum dots, and carbon and silica-based nanomaterials make up the bulk of the inorganic NPs. These NPs are prepared using a variety of top-down and bottom-up approaches. Microfluidics provide an attractive synthesis alternative and is advantageous compared to the conventional bulk methods. The microfluidic mixing-based production methods offer better control in achieving the desired size, morphology, shape, size distribution, and surface properties of the synthesized NPs. The technology also exhibits excellent process repeatability, fast handling, less sample usage, and yields greater encapsulation efficiencies. In this article, we provide a comprehensive review of the microfluidic-based passive and active mixing techniques for NP synthesis, and their latest developments. Additionally, a summary of microfluidic devices used for NP production is presented. Nonetheless, despite significant advancements in the experimental procedures, complete details of a nanoparticle-based system cannot be deduced from the experiments alone, and thus, multiscale computer simulations are utilized to perform systematic investigations. The work also details the most common multiscale simulation methods and their advancements in unveiling critical mechanisms involved in nanoparticle synthesis and the interaction of nanoparticles with other entities, especially in biomedical and therapeutic systems. Finally, an analysis is provided on the challenges in microfluidics related to nanoparticle synthesis and applications, and the future perspectives, such as large-scale NP synthesis, and hybrid formulations and devices.

Numerical Modeling Using Immersed Boundary-Lattice Boltzmann Method and Experiments for Particle Manipulation under Standing Surface Acoustic Waves.

In this work, we employed the Immersed Boundary-Lattice Boltzmann Method (IB-LBM) to simulate the motion of a microparticle in a microchannel under the influence of a standing surface acoustic wave (SSAW). To capture the response of the target microparticle in a straight channel under the effect of the SSAW, in-house code was built in C language. The SSAW creates pressure nodes and anti-nodes inside the microchannel. Here, the target particle was forced to traverse toward the pressure node. A mapping mechanism was developed to accurately apply the physical acoustic force field in the numerical simulation. First, benchmarking studies were conducted to compare the numerical results in the IB-LBM with the available analytical, numerical, and experimental results. Next, several parametric studies were carried out in which the particle types, sizes, compressibility coefficients, and densities were varied. When the SSAW is applied, the microparticles (with a positive acoustic contrast factor) move toward the pressure node locations during their motion in the microchannel. Hence, their steady-state locations are controlled by adjusting the pressure nodes to the desired locations, such as the centerline or near the microchannel sidewalls. Moreover, the geometric parameters, such as radius, density, and compressibility of the particles affect their transient response, and the particles ultimately settle at the pressure nodes. To validate the numerical work, a microfluidic device was fabricated in-house in the cleanroom using lithographic techniques. Experiments were performed, and the target particle was moved either to the centerline or sidewalls of the channel, depending on the location of the pressure node. The steady-state placements obtained in the computational model and experiments exhibit excellent agreement and are reported.

A review of active and passive hybrid systems based on Dielectrophoresis for the manipulation of microparticles.

Particle separation is essential in a broad range of systems and has several biological applications. Microfluidics has emerged as a potentially transformational method for particle separation. The approach manipulates and separates particles at the micrometer scale by using well-defined microstructures and precisely managed force fields. Depending on the source of the principal manipulating forces, particle manipulation and separation in microfluidics may be classified as active or passive. Passive microfluidic devices depend on drag and inertial forces and microchannel structure, while active microfluidic systems rely on external force fields. Active microfluidics, in general, can properly control and place particles of interest in real time. Due to the low flow rate, the residual time required to apply an appropriate external manipulating force to the target particles is reduced, thereby limiting overall throughput. Passive microfluidics, on the other hand, has a simple architecture, robustness, and high throughput. Hybrid techniques, which combine active and passive processes, have been created to address the shortcomings of each while maximizing the benefits of each. Numerous hybrid techniques for particle separation have been developed. This study reviews the most recent developments in the field of hybrid devices based on dielectrophoresis. Dielectrophoresis-passive and dielectrophoresis-active hybrid approaches are described and evaluated. Dielectrophoresis-inertial, dielectrophoresis-hydrophoresis, dielectrophoresis- deterministic lateral displacement, and insulator-based dielectrophoresis are examples of dielectrophoresis-passive hybrid devices. Dielectrophoresis with acoustophoresis, magnetophoresis, and optophoresis are examples of dielectrophoresis-active devices. Each hybrid system will be assessed based on its operating principles, advantages, and disadvantages. Following that, a comprehensive explanation of dielectrophoresis physical concepts and operating procedures will be offered. As part of this review, the advantages and disadvantages of DEP-based separation devices will be examined. All these hybrid devices will be thoroughly examined and evaluated. Finally, a summary of present difficulties in the hybrid separation sector will be offered, as well as future suggestions and aspirations.

A deep learning-driven low-power, accurate, and portable platform for rapid detection of COVID-19 using reverse-transcription loop-mediated isothermal amplification.

This paper presents a deep learning-driven portable, accurate, low-cost, and easy-to-use device to perform Reverse-Transcription Loop-Mediated Isothermal Amplification (RT-LAMP) to facilitate rapid detection of COVID-19. The 3D-printed device-powered using only a 5 Volt AC-DC adapter-can perform 16 simultaneous RT-LAMP reactions and can be used multiple times. Moreover, the experimental protocol is devised to obviate the need for separate, expensive equipment for RNA extraction in addition to eliminating sample evaporation. The entire process from sample preparation to the qualitative assessment of the LAMP amplification takes only 45 min (10 min for pre-heating and 35 min for RT-LAMP reactions). The completion of the amplification reaction yields a fuchsia color for the negative samples and either a yellow or orange color for the positive samples, based on a pH indicator dye. The device is coupled with a novel deep learning system that automatically analyzes the amplification results and pays attention to the pH indicator dye to screen the COVID-19 subjects. The proposed device has been rigorously tested on 250 RT-LAMP clinical samples, where it achieved an overall specificity and sensitivity of 0.9666 and 0.9722, respectively with a recall of 0.9892 for Ct < 30. Also, the proposed system can be widely used as an accurate, sensitive, rapid, and portable tool to detect COVID-19 in settings where access to a lab is difficult, or the results are urgently required.

Dielectrophoresis-field flow fractionation for separation of particles: A critical review.

Dielectrophoresis-field flow fractionation (DEP-FFF) has emerged as an efficient in-vitro, non-invasive, and label-free mechanism to manipulate a variety of nano- and micro-scaled particles in a continuous-flow manner. The technique is mainly used to fractionate particles/cells based on differences in their sizes and/or dielectric properties by employing dielectrophoretic force as an external force field applied perpendicular to the flow direction. The dielectrophoretic force is the result of a spatially non-uniform electric field in the microchannel that can be generated either by exploiting microchannel geometry or using special arrangements of microelectrode arrays. Several two-dimensional (e.g., coplanar interdigitated, castellated) and three-dimensional (e.g., top-bottom, side-wall) microelectrode designs have been successfully utilized to perform fractionation of heterogeneous samples. Although originally introduced as a separation technique, DEP-FFF has attracted increasing interest in performing other important operations such as switching, focusing, dipping, and surface functionalization of target particles. Nonetheless, the technique still suffers from limitations such as low throughput and joule heating. By comparatively analyzing recent developments that address these shortcomings, this work is a step forward towards realizing the full potential of DEP-FFF as an ideal candidate for point-of-care (POC) devices with diverse applications in the fields of biomedical, chemical, and environmental engineering.

Multiple Particle Manipulation under Dielectrophoresis Effect: Modeling and Experiments.

The dissipative particle dynamics (DPD) technique was employed to design multiple microfluidic devices for investigating the motion of bioparticles at low Reynolds numbers. A DPD in-house FORTRAN code was developed to simulate the trajectories of two microparticles in the presence of hydrodynamic and transverse deflecting force fields via considering interparticle interaction forces. The particle-particle interactions were described by using a simplified version of the Morse potential. The transverse deflecting force considered in this microfluidic application was the dielectrophoresis (DEP) force. Multiple microfluidic devices with different configurations of microelectrodes were numerically designed to investigate the dielectrophoretic behavior of bioparticles for their trajectories and the focusing of bioparticles into a single stream in the middle of the microchannel. The DPD simulation results were verified and validated against previously reported numerical and experimental works in the literature. The computationally designed microdevices were fabricated by employing standard lithographic techniques, and experiments were conducted via taking red blood cells as the representative bioparticles. The experimental results for the trajectories and focusing showed good agreement with the numerical results.

Dissipative particle dynamics for modeling micro-objects in microfluidics: application to dielectrophoresis.

The dissipative particle dynamics (DPD) technique is employed to model the trajectories of micro-objects in a practical microfluidic device. The simulation approach is first developed using an in-house Fortran code to model Stokes flow at Reynolds number of 0.01. The extremely low Reynolds number is achieved by adjusting the DPD parameters, such as force coefficients, thermal energies of the particles, and time steps. After matching the numerical flow profile with the analytical results, the technique is developed further to simulate the deflection of micro-objects under the effect of a deflecting external force in a rectangular microchannel. A mapping algorithm is introduced to establish the scaling relationship for the deflecting force between the physical device and the DPD domain. Dielectrophoresis is studied as a case study for the deflecting force, and the trajectory of a single red blood cell under the influence of the dielectrophoretic force is simulated. The device is fabricated using standard microfabrication techniques, and the experiments involving a dilute sample of red blood cells are performed at two different cases of the actuation voltage. Good agreement between the numerical and experimental results is achieved.

Investigation of DPD transport properties in modeling bioparticle motion under the effect of external forces: Low Reynolds number and high Schmidt scenarios.

We have used a dissipative particle dynamics (DPD) model to study the movement of microparticles in a microfluidic device at extremely low Reynolds number (Re). The particles, immersed in a medium, are transported in the microchannel by a flow force and deflected transversely by an external force along the way. An in-house Fortran code is developed to simulate a two-dimensional fluid flow using DPD at Re ≥ 0.0005, which is two orders of magnitude less than the minimum Re value previously reported in the DPD literature. The DPD flow profile is verified by comparing it with the exact solution of Hagen-Poiseuille flow. A bioparticle based on a rigid spring-bead model is introduced in the DPD fluid, and the employed model is verified via comparing the velocity profile past a stationary infinite cylinder against the profile obtained via the finite element method. Moreover, the drag force and drag coefficient on the stationary cylinder are also computed and compared with the reported literature results. Dielectrophoresis (DEP) is investigated as a case study for the proposed DPD model to compute the trajectories of red blood cells in a microfluidic device. A mapping mechanism to scale the external deflecting force from the physical to DPD domain is performed. We designed and built our own experimental setup with the aim to compare the experimental trajectories of cells in a microfluidic device to validate our DPD model. These experimental results are used to investigate the dependence of the trajectory results on the Reynolds number and the Schmidt number. The numerical results agree well with the experiment results, and it is found that the Schmidt number is not a significant parameter for the current application; Reynolds numbers combined with the DEP-to-drag force ratio are the only important parameters influencing the behavior of particles inside the microchannel.

Continuous-Flow Cell Dipping and Medium Exchange in a Microdevice using Dielectrophoresis.

Medium exchange is the process of changing the suspension medium of cells/particles, and has applications in washing, surface modifications, nutrient replenishment, or simply changing the environment of the target entities. Dipping involves diverting the path of target cells in the carrying fluid to immerse them in another fluid for a short duration, and pushing them again into the original medium. In this paper, a simple microfluidic platform is introduced that employs dielectrophoresis to achieve medium exchange and dipping of micro-objects in a continuous manner. The essential feature of the platform is a microchannel that includes two arrays of microelectrodes that partly enter the bottom surface from both sides. In the first step, numerous finite element-based parametric studies are carried out to obtain the optimized geometrical and operational parameters ensuring successful dipping and medium exchange processes. The results of those studies are utilized to fabricate the platform using standard photolithography techniques. The electrodes are patterned on a glass substrate, while the channel, made out of polydimethylsiloxane, is bonded on top of the glass. Trajectories of blood cells from numerical studies and experimentations are reported, and both results exhibited close agreement.

Dielectrophoresis-based 3D-focusing of microscale entities in microfluidic devices.

This article presents a validated mathematical model of a dielectrophoresis (DEP)-based microfluidic device capable of 3D-focusing microscale entities at any lateral location inside the microchannel. The microfluidic device employs planar, independently controllable, interdigitated transducer (IDT) electrodes on either side of the microchannel. The developed model is used for understanding the influence of different geometric and operating parameters on 3D focusing, and it comprises of motion equation, Navier-Stokes equation, continuity equation, and electric potential equation (Laplace equation). The model accounts for forces associated with inertia, gravity, buoyancy, virtual mass, drag, and DEP. The model is solved using finite difference method. The findings of the study indicate that the 3D focusing possible with the proposed microfluidic device is independent of microscale entity's size and initial position, microchannel height, and volumetric flow rate. In contrast, 3D focusing achievable with the microfluidic device is dependent on the applied electric potential, protrusion width of electrodes, and width of electrode/gap. Additionally, the lateral position of 3D focused can be controlled by varying the applied electric potential. The advantage of the proposed microfluidic device is that it is simple to construct while capable of achieving 3D focusing at any lateral location inside the microchannel.

Lateral fluid flow fractionation using dielectrophoresis (LFFF-DEP) for size-independent, label-free isolation of circulating tumor cells.

This short communication introduces a continuous-flow, dielectrophoresis-based lateral fluid flow fractionation microdevice for detection/isolation of circulating tumor cells in the presence of other haematological cells. The device utilizes two sets of planar interdigitated transducer electrodes micropatterned on top of a glass wafer using standard microfabrication techniques. A microchannel with a single inlet and two outlets, realized in polydimethylsiloxane, is bonded on the glass substrate. The two sets of electrodes slightly protrude into the microchannel. Both of the electrode sets are energized with signals at different frequencies and different operating voltages ensuring that the cancer cells experience positive dielectrophoretic force from one set of the electrodes and negative dielectrophoretic force from the other array. Normal cells experience unequal negative dielectrophoretic forces from opposing sets of electrodes. The resultant dielectrophoretic forces on cancer and normal cells push them to flow towards their designed outlets. Successful isolation of green fluorescent protein-labelled MDA-MB-231 breast cancer cells from regular blood cells, both suspended in a sucrose/dextrose medium, is reported in this work.

Design and modeling of electrode networks for code-division multiplexed resistive pulse sensing in microfluidic devices.

A typical microfluidic device sorts, captures or fractionates sample constituents by exposing them to discriminating microenvironments. Direct electronic acquisition of such manipulation by a network of integrated sensors can provide a fast, integrated readout, replacing otherwise required microscopy. We have recently introduced a sensor technology, Microfluidic CODES, which allows us to multiplex resistive pulse sensors on a microfluidic device. Microfluidic CODES employs a network of micromachined coplanar electrodes such that particles passing over these electrodes produce distinguishable code sequences. In this paper, we explain the design process to specifically generate an orthogonal digital code set for an efficient and accurate demultiplexing of the sensor signals. We also introduce an equivalent circuit model for a network of code-multiplexed resistive pulse sensors by utilizing the Foster-Schwan model and conformal mapping, to model dynamic cell-electrode interaction in a non-uniform electric field. Our results closely match with both experimental measurements using cell lines and finite element analysis. The coding and modeling framework presented here will enable the design of code-division multiplexed resistive pulse sensors optimized to produce desired waveform patterns to ensure reliable and efficient decoding.