Microfluidic devices containing artificial cellular membranes have been developed for biology, biosensors, and bioengineering. Although forming a lipid bilayer between two microchannels has advantages for integrating cellular membrane functions into continuous chemical processing and evaluating cellular membrane responses under fluid-flow-induced shear stress, it typically suffers from complex operation of external pumps and contamination of the lipid bilayer by residual organic solvent. This study presents a simplified formation process of an ultrasmall lipid bilayer between parallel micrometer channels leveraging dominant surface tension effects. Utilizing 4 μm channels with partial hydrophobic modification, the simple operation of a parallel two-phase flow involving an aqueous phase and a lipid-containing organic phase (chloroform) was achieved under a controlled 101 kPa Laplace pressure. After parallel two-phase flow was formed to create a lipid monolayer at the aqueous/organic interface, the organic phase was replaced with the aqueous phase, maintaining the interface position via Laplace pressure. A 2.2 × 3.7 μm2 lipid bilayer with a thickness of 4.2–6.4 nm and reduced contamination was successfully formed, as verified by fluorescence and capacitance measurements. To demonstrate the functionality of this parallel-channel platform, the stability of the lipid bilayer was evaluated under shear stress induced by pressure-driven flows. The results indicated that the bilayer partially ruptured at a shear rate higher than 0.27 μs−1, suggesting a critical shear rate that induces membrane instability. This work demonstrates the feasibility of the developed in vitro platform for studying cellular membranes at the subcellular level and conducting fL-chemical operations using cellular membrane functions.
Precise spatiotemporal control of chemical gradients is crucial for studying microbial chemotaxis. However, traditional microfluidic platforms using poly(dimethylsiloxane) and soft lithography present a high barrier to entry due to their reliance on cleanrooms, specialized molds, and complex external pumping systems. To address these challenges, we introduce a rapid, cleanroom-free, 3D-printed microfluidic assay that simplifies the quantification of microbial chemotaxis and motility. Manufactured in under 4 h using high-resolution stereolithography 3D printing, the device maintains a standard 25 × 75 mm2 footprint for effortless microscope integration. Its monolithic, three-chamber design incorporates a central truncated-cone agar well that mechanically secures a hydrogel barrier, generating well-defined, diffusion-based transient chemical gradients without the need for external equipment. Physical characterization via visual flow tests with high-contrast dye confirmed robust leak prevention and unidirectional transient gradient formation within a quantified operational time window. As a biological proof-of-concept, the platform successfully quantified the concentration-dependent chemotactic response of Chlamydomonas reinhardtii to sodium bicarbonate, revealing peak sensitivity at 10−3M (p < 0.001) while capturing population heterogeneity. Additionally, parallel evaluations with the morphologically distinct Euglena gracilis demonstrated the platform’s architectural adaptability to diverse cell sizes and shapes. These evaluations confirmed the assay’s efficacy in isolating active chemotaxis from passive fluid drift and context-dependent motility. Ultimately, this accessible, rapid-prototyping platform provides a highly customizable tool for investigating complex microbial behavioral dynamics.
Emerging diagnostic protocols impose constraints on blood sample volume, necessitating efficient microscale separation of plasma. We present a novel method for extracting plasma from a single blood droplet of only 25 μL (with and without dilution), using a variable-gap lifted Hele-Shaw cell. The device consists of (i) an elastomeric top surface cyclically deformed by a linear actuator to form a continuously varying spherical cap and (ii) a flat, stationary bottom chip containing a microfilter interfaced with a microfluidic channel outlet. The blood droplet is confined within a periodically varying micro-gap between these two surfaces and is recirculated by radial stretching and contraction of the film on the filter surface. Dead-end and tangential filtration are, thus, coupled to mitigate membrane blocking by cells: a critical bottleneck of blood microfiltration. Controlled shearing of the squeezed blood film induces cell migration away from the filter surface, thereby depleting the filter cake layer. The filtrate plasma is directed from the sub-membrane space to the outlet channel by capillary action. The imposed oscillatory shear field responsible for cell redistribution is tuned to maintain high plasma flux. Free hemoglobin levels in plasma were monitored through spectrophotometry. An optimal operating condition is found, under which the device extracted a plasma yield of 60.51%±0.05%. Protein recovery in separated plasma was assessed against centrifuged plasma using a C-reactive protein assay. The proposed approach demonstrates the utility of dynamically sheared blood flow in a micro-confinement, either toward a standalone plasma separator or in integration with biosensors.
Infectious diseases and antimicrobial resistance pose escalating threats to global public health, demanding rapid and accurate pathogen detection to guide timely clinical decision-making. Conventional diagnostic methods, including microbial culture, immunoassays, and laboratory-based molecular testing, require specialized equipment, trained personnel, and lengthy workflows that delay results and limit accessibility, particularly in resource-constrained settings. Microfluidic nucleic acid amplification testing (NAAT) platforms offer a promising route toward miniaturized, automated sample-to-answer molecular diagnostics with reduced reagent consumption and accelerated turnaround times. This review analyzes microfluidic NAAT platforms through a unified comparative framework encompassing on-chip lysis, nucleic acid purification, amplification, and detection.
In this Perspective, we discuss the utility and flexibility of elastomeric pneumatic valving strategies and the challenges in applying these approaches to ultra-thin microfluidic devices. We use the strict material requirements for in situ x-ray analysis as a framework for introducing the challenges in leveraging pneumatic valves and highlight a range of capillary-based fluid handling and valving strategies that have been leveraged in x-ray applications as well as point-of-care diagnostics and biomedical applications where ultra-thin devices are critical. Rather than aiming to reproduce the full programmability of elastomeric valve networks, we highlight the ability of these platforms to facilitate event-based fluid control strategies that balance functional utility with manufacturability and robustness.
Activated platelets are key players in many thrombotic and hemostatic disorders. However, conventional platelet function tests often fail to capture how platelets behave under dynamic flow conditions that closely mimic physiological blood flow. Advances in 3D printing and microfluidic design now enable fabrication of more physiologically relevant microvascular constructs that support controlled investigation of platelet activation under defined hemodynamic environments. When integrated with artificial intelligence (AI) approaches, including deep-learning-based image analysis and physics-informed modeling, these platforms move beyond descriptive measurements toward automated, quantitative, and mechanistically interpretable assessment of platelet behavior. This review critically synthesizes recent progress at the intersection of 3D printing, microfluidic platelet assays, and AI-enabled analytics, with an emphasis on microfluidic design principles, detection strategies, benchmarking requirements, and translational considerations specific to platelet mechanobiology. It highlights how geometric- and shear-resolved microfluidic assays generate high-dimensional datasets that motivate AI-based analysis and address current biological and clinical limitations. Emerging applications include mechanistic studies of shear-mediated thrombosis, high-throughput drug screening under flow, and exploratory approaches to thrombotic risk stratification and patient-specific platelet phenotyping. Key challenges for translation include standardizing benchmarking against reference assays, rigorously reporting fabrication and hemocompatibility parameters, and validating AI models across multiple devices and patient cohorts. While these technologies are best viewed as complementary to established platelet function tests, their integration with AI-driven analytics may have important implications for advancing vascular diagnostics and thrombosis modeling.
The space of Disse is a key liver microenvironment positioned between endothelial cells and hepatocytes, filled with an extracellular matrix, where molecular exchange occurs. Its structure dynamically shifts between physiological and pathological conditions, influencing liver function and disease progression. Due to its critical role, in vitro models of the space of Disse are essential for studying liver physiology, disease mechanisms, and therapeutic interventions. This review provides a comprehensive analysis of current bioengineering strategies for modeling the space of Disse. First, we examine different cellular sources and tissue configurations developed in these models, ranging from 2D cultures and sandwich systems to layered 3D tissues and spheroids. Next, we explore recent advancements in engineering approaches, including the integration of microfluidics, microfabrication techniques, and novel biomaterials to enhance model reliability. Finally, we discuss five key applications of these models: investigating the space of Disse physiology and cellular interactions, studying oxygen gradients, modeling liver pathologies, improving hepatotoxicity prediction in the drug screening process, and developing transplantable liver grafts. By synthesizing these advancements, this review highlights the current state, challenges, and future directions of the space of Disse in vitro modeling.
Lab-on-a-Disc (LoaD) platforms have gained increasing attention as practical microfluidic tools capable of automating complex analytical workflows on a compact, low-cost centrifugal disc. This design reduces contamination risks and makes the workflow more practical, especially for point-of-care use. LoaD platforms have been applied in different fields in recent years, mainly molecular diagnostics, foodborne pathogen testing, and environmental analysis. For clinical purposes, several discs were developed that combine lysis, nucleic acid purification, and isothermal amplification, and they were able to detect viruses and bacteria such as influenza, SARS-CoV-2, and some cancer-associated genes. Similar approaches have been used for food samples, where LoaD devices helped identify common pathogens even in complex matrices like milk or meat. Environmental studies also benefit from LoaD systems, for example, in monitoring heavy metals, algal toxins, pesticides, or microbial contamination in water and soil. Although the technology has advanced, practical challenges remain, including stable thermal management, valve reliability, variability introduced by viscous samples, and overall fabrication consistency. Ongoing developments in isothermal and CRISPR-based assays, digital nucleic-acid methods, improved disc materials, and simplified optical readers (including smartphone-based ones) are expected to overcome some of these limitations. For these reasons, LoaD platforms are gradually becoming a realistic option for rapid point-of-care diagnostics in biomedical, environmental, and food-safety applications.
Infectious diseases remain one of the leading causes of morbidity and mortality worldwide, exacerbated by emerging pathogens, antimicrobial resistance, and limitations of conventional diagnostic and therapeutic approaches. Micro- and nanomachines—artificially engineered micro- and nanoscale devices capable of controlled motion, sensing, and actuation—have emerged as transformative tools in infectious disease management. These systems offer unprecedented opportunities for rapid pathogen detection, targeted drug delivery, biofilm disruption, and minimally invasive therapy. This review summarizes recent advances in micro- and nanomachine technologies applied to infectious disease diagnostics and therapeutics, highlighting design principles, propulsion mechanisms, functionalization strategies, and representative biomedical applications. Current challenges related to biocompatibility, scalability, regulatory approval, and clinical translation are critically discussed. Last, perspectives for the future are discussed, emphasizing the integration of micromachines with artificial intelligence, lab-on-a-chip platforms, and precision medicine to combat infectious diseases more effectively.
The pursuit of physiologically relevant preclinical models has driven the emergence of three-dimensional tumor spheroids and organ-on-chip (OOC) platforms, which recapitulate the structural, biochemical, and functional complexity of human tissue microenvironments. Unlike conventional two-dimensional cultures and animal models, these systems provide human-specific insights into disease biology, drug delivery, and therapeutic response, addressing critical translational gaps in biomedical research. They reproduce cell–cell and cell–matrix interactions, vascularization, nutrient and drug gradients, and organ-level functions, offering new opportunities to evaluate drug penetration, efficacy, and safety. Their complexity generates high-dimensional imaging, omics, and functional datasets that are increasingly difficult to analyze using traditional approaches, setting the stage for the integration of artificial intelligence (AI) and machine learning (ML) as transformative tools. Recent advances demonstrate how AI-driven methods can automate spheroid image segmentation, quantify spatial heterogeneity, and extract predictive biomarkers of treatment response with minimal human bias. Deep learning algorithms are also being applied to predict drug transport dynamics, optimize dosing strategies, and analyze complex interactions within OOC systems, enhancing experimental precision and translational relevance. Moreover, AI integration with Multi-OOC platforms is beginning to capture systemic pharmacokinetics and pharmacodynamics, linking in vitro performance with in vivo outcomes and reshaping drug delivery research by coupling biological fidelity with computational intelligence. This review critically analyzes AI-enhanced spheroid and OOC platforms for drug delivery, highlighting progress and key challenges, such as data standardization, interpretability, and reproducibility, while outlining future directions, including the convergence of digital twins, ML, and multi-scale modeling toward intelligent, personalized, and clinically translatable drug delivery systems.
The microscale operation of close-packed hydrogel microbead suspensions (CPHMSs) has attracted widespread attention due to their advantages in reducing randomness and avoiding the challenges of density matching in applications spanning from single-cell sequencing, tissue engineering to drug carrier fabrication, etc. In this paper, the fundamentals, technologies, and applications of CPHMSs are reviewed. Preparation methods of CPHMSs and the fundamentals of flow and rheological properties of CPHMSs are first discussed. Then, a brief overview of the technological development of microscale manipulation of CPHMSs is presented, including pairing, mixing, co-encapsulation, and 3D printing of CPHMSs. Furthermore, a review on the applications of CPHMS manipulation in several fields such as single-cell sequencing and preparation of topological hydrogels is presented. The technical challenges that need to be further addressed on CPHMS manipulation are also outlined.
Antimicrobial resistance represents one of the most urgent global health threats, necessitating a paradigm shift in antimicrobial susceptibility testing (AST) to ensure timely and targeted therapy. Conventional methods, although clinically standardized, suffer from long turnaround times (24–72 h), which delay treatment decisions and contribute to the overuse of broad-spectrum antibiotics. To solve this issue, microfluidic platforms are redefining AST by miniaturizing assays, reducing analysis times to mere hours, and minimizing sample consumption. Recent advances can be broadly categorized into channel-based, droplet-based, and integrated systems, which leverage diverse detection strategies including optical, electrical, and label-free approaches. Channel-based devices provide precise environmental control for direct microscopic observation and stable antibiotic gradient generation, while droplet-based platforms enable massive parallelization at the single-cell level, capturing heteroresistance and phenotypic diversity. Integrated systems increasingly combine microfluidics with smartphones, automation, and artificial intelligence (AI)-driven analytics, thereby accelerating the transition toward point-of-care testing. Beyond speed and throughput, microfluidic AST is also moving toward personalized diagnostics, incorporating direct-from-sample testing, host–pathogen coculture models, and machine learning-assisted image analysis to tailor therapy to individual patients. Despite challenges in standardization, complex sample handling, and regulatory approval, the convergence of advanced materials, automation, and AI with microfluidic platforms positions microfluidic AST as a cornerstone technology for the next generation of infectious disease management and personalized antimicrobial therapy.
Traditional in vitro maturation (IVM) involves manual handling and static culture conditions and lacks automated assessment. These limitations increase variability, introduce stress that may affect oocyte quality, and fail to recapitulate the dynamic in vivo microenvironment. This study presents a system that streamlines, automates, and potentially enhances IVM procedures through the capture, dynamic on-chip cultivation, and evaluation of single oocytes on-chip, termed FemaleTech Enhanced Maturation Intelligence (FEMI). The platform employs a novel microfluidic chip featuring hexagonal micropillars that efficiently trap individual oocytes across a broad size range while inherently excluding low-quality oocytes, ensuring stable positioning throughout culture. Numerical simulations and experimental validation confirmed the trapping principle and high capture efficiency of the proposed design. FEMI supports continuous microfluidic culture under controlled flow conditions, allowing partial recapitulation of physiological transport and mechanical cues while maintaining oocyte immobilization. By integrating deep learning-based image segmentation, the system automatically quantifies cumulus cell expansion, providing single-cell-level insights into oocyte development. The results suggest that FEMI may improve IVM efficiency and provide a useful platform for assisted reproductive technologies.
Nucleic acid amplification is a critical step in many diagnostic workflows from clinical to point-of-care settings. Microfluidics offers miniaturized platforms for handling small sample volumes; however, heat-assisted bioreactions such as digital loop-mediated isothermal amplification (dLAMP) are challenging to implement in poly(dimethylsiloxane) (PDMS)-based devices due to the material's intrinsic porosity and hydrophobicity. The porosity of this polymer network promotes evaporation of the reaction solution and diffusion of small molecules, while its surface hydrophobicity encourages biofouling, leading to loss of enzymatic activity. In this work, we present a facile and straightforward strategy to overcome these problems that avoids additional fabrication steps, specialized materials, or complex post-fabrication treatments. First, the porosity of PDMS was tuned by altering the conventional monomer-to-cross-linker ratio, resulting in significant retention of solution volume even after 2 h of heating at 60 degrees C and a 70% reduction in small-molecule loss via diffusion into the polymer network. Second, to render the PDMS surface hydrophilic, a dry coating of the commercially available surfactant, Tween-20 was employed. The results indicate that the dry coating enhances amplification efficiency, whereas incorporating the same amount of surfactant directly into the reaction solution reduces efficiency. Through this combined approach, we demonstrate a twofold improvement in the limit of detection for nucleic acid amplification by dLAMP, greater uniformity in reaction volumes across the array, and more robust performance-improvements that are retained even after 8 months of device storage. By addressing key material limitations, this study advances PDMS-based microfluidics as a practical platform for high-precision nucleic acid quantification.
This paper presents fabrication process of cell-encapsulating hollow collagen microgel beads, which enables stable and large-scale production of hollow tissue models. The microgel beads feature a core-shell structure, with a liquid core containing a cell suspension and a collagen shell serving as a scaffold. Cells encapsulated within the core adhere to the inner collagen wall, leading to the formation of a defined hollow tissue. We demonstrated this by fabricating beads with encapsulated human umbilical vein endothelial cells, which successfully adhered to the inner surface. These results suggest that this approach geometrically guides the cells into a hollow structure, facilitating the engineered construction of hollow tissue architectures. The cell-encapsulating hollow collagen microgel beads are promising tools for reproducing pathological models of hollow organs for drug evaluation.
The fabrication of integrated lab-on-a-chip devices conventionally relies on disjointed workflows, requiring mask-dependent photolithography for electrode patterning and separate soft lithography processes for fluidic channelling. This study characterizes an integrated manufacturing workflow utilizing a single commercial LCD-based masked stereolithography (SLA) platform (Phrozen Sonic Mighty 8 K) to perform three distinct physical processes: maskless UV lithography for sensor patterning, high-resolution mold fabrication, and functional substrate printing. We quantify the lithographic fidelity of the LCD interface (30 mu m limit), the mechanical shear strength of the device bonding, and the intrinsic surface wettability of the acrylate-based resin, which enables a simplified fabrication protocol with reduced reliance on plasma surface treatment. To validate the workflow, an electrochemical aptasensor for cardiac troponin I (cTnI) was fabricated and compared against standard mask-patterned controls. Electrochemical impedance spectroscopy and differential pulse voltammetry results demonstrate that despite the pixelation inherent to the LCD matrix, the 3D-assisted electrodes maintain excellent linearity (R-2 > 0.99) in Randles-& Scaron;ev & ccaron;& iacute;k diffusion analysis and achieve sensitive cTnI detection (0.1-1000 pg/ml). Furthermore, material cost analysis highlights a significant improvement in resource efficiency compared to traditional protocols. These findings demonstrate that high-resolution consumer LCD-based masked SLA platforms can serve as robust, single-instrument manufacturing stations that bridge the gap between benchtop prototyping and industrial mass production.
The electrical properties of hemoglobin (Hb) solutions are essential characteristics that offer valuable insights into red blood cell health metrics. These properties are influenced by factors such as intracellular ions and Hb. This study uses electrical impedance spectroscopy to analyze Hb solutions prepared from both lyophilized and freshly extracted human Hb across frequency range (40 Hz-110 MHz), aiming to clarify Hb-ion interactions and their effects on solution conductivity across physiological conditions. To determine the contributions of ions, Hb, and the device to the measured impedance value, electrical circuit models combined with finite element analysis in COMSOL Multiphysics, were used and validated with impedance measurements on de-ionized water and ionic solutions of defined composition. In low ionic strength solutions (sub-physiological levels), Hb addition increased conductivity of solutions, whereas in high ionic strength solutions, Hb addition altered conductivity of solutions, with these effects varying by ionic strength. Notably, the extrapolated solution conductivity at the physiological Hb level, 320 mg/ml, were found in a relatively narrow range (0.43-0.59 S/m), despite the marked variations in the background ionic strength (0.074-0.13 mol/l) and pH value (6.8 and 8.85), indicating significant Hb-ion interactions that modulate the overall electrical behavior of Hb solutions. The combined experimental and computational approach provides a robust framework for assessing the influence of Hb and ions on solution conductivity and can be extended to study other Hb variants beyond normal Hb.
While high-throughput combinatorial sample generation plays a critical role in drug discovery and cell-based assays, standard methods have encountered challenges of time or cost-effectiveness when using manual preparation or serial-dispensing liquid handlers. To address these challenges, we introduce a novel parallel combinatorial microfluidic (PCM) system that applies the design of hydraulic-resistance networks to simultaneously generate a prespecified combinatorial map. Unlike conventional serial dispensing, PCM devices integrated superhydrophobic outlets for parallel contactless dispensing, thereby minimizing contamination, reducing processing time, and ensuring an identical reaction time across all samples. Regardless of the number of target samples, the PCM system demonstrated rapid operation by completing all processes within 5 minutes. Devices fabricated for multiple combinatorial maps validated the PCM design principle and high-throughput scalability by reliably producing the predefined maps with droplet sizes of 3.5 μL. A proof-of-concept study of antibiotic cell assays confirmed the biocompatibility of the system and its ability to produce results comparable to those of standard methods, as demonstrated by good overlapping ranges of the half maximal inhibitory concentration (IC 50). This work establishes a novel, cost-effective, high-throughput technology for combinatorial screenings in drug development with future improvements targeting sub-3.5 μL to enhance accuracy.
Droplet microfluidics enables the generation of highly monodisperse picoliter droplets and underpins a wide range of applications, from single-cell analysis to materials synthesis. Despite the maturity of the field, the design and operation of droplet microfluidic systems still largely rely on trial-and-error approaches, driven by the absence of universal predictive models and exacerbated by device-to-device and lab-to-lab variability. In recent years, machine learning (ML) and artificial intelligence (AI) have emerged as powerful tools to address these challenges by extracting structure-property-process relationships from experimental and numerical data. This Perspective provides a critical overview of recent advances in AI-enabled droplet microfluidics, with a particular focus on three interconnected themes of crucial interest in microfluidic applications: prediction of droplet properties, read out of droplet characteristics from experimental images, and closed-loop optimization of device operation and design. We discuss how data-driven models have enabled accurate prediction of droplet size, frequency, and regime across different geometries, how computer vision approaches have transformed high-throughput droplet analysis, and how Bayesian optimization and autonomous frameworks are moving the field toward automated microfluidic platforms. Finally, we highlight current limitations, including data sparsity, generalizability, and the treatment of complex fluids, and outline key opportunities for future research aimed at establishing robust, interpretable, and broadly applicable ML and AI tools for droplet microfluidics.
Cellular mechanical properties and liquid viscosity are critical biophysical indicators for assessing physiological status and diagnosing diseases. For the characterization of cellular mechanics, microfluidics-based high-throughput cell deformation assays have garnered significant attention owing to their high throughput and minimal cell damage. However, such techniques typically require cells to be suspended in a medium with known viscosity, which precludes the measurement of the native sample's own viscosity-a particular limitation for precious clinical specimens. Here, we present a microfluidic chip based on virtual fluidic channels that employs a "split-sample, dual-assay" strategy, enabling the simultaneous acquisition of two key biophysical parameters (liquid viscosity and cellular mechanical properties) from a single sample source with minimal volume consumption.
