Microfluidics, BioMEMS, and Medical Microsystems XXIII

We present a novel setup for liquid flat-jet applications in vacuum. It employs a compact, high-precision nozzle system capable of producing a flat-jet with a thickness well below the micrometer range, expanding the span of materials and processes that can be analyzed with high precision using spectroscopic techniques in the far-UV spectral range. We explore and demonstrate multiple techniques for generating and applying these flat liquid jets. An advanced microfluidic pump system facilitates jet stability by precisely controlling the liquid flow rate and pressure. Application techniques include integration into spectroscopic experiments, where the flat-jet provides high spatial resolution and minimal sample consumption. This makes it ideal for time-resolved studies and high-throughput screening, with the potential for a recycling system. This is demonstrated by its implementation into a compact table-top soft X-ray absorption spectrometer, highlighting the system's suitability for probing dynamic processes at a molecular level.

Microfluidic systems find vast applications in the field of biochemical sciences. Current systems are limited to diluted (5-100X) biological samples whereas most unmodified biological samples are viscoelastic in nature. Transport of bio-colloids of various classes, in such viscoelastic fluids especially under the influence of dynamic electric and magnetic fields can have tremendous implications in Point-of-Care Diagnostics and Drug Delivery Systems. Lyotropic Liquid Crystals (LLCs) are being studied extensively for biomedical applications. Anisotropic viscosity in addition to non-linear elastic, viscous and inertial effects make LLCs a suitable candidate to study and characterize transport phenomena in biological fluids. For example: blood flow in vascular tumors. In this study, we characterize electrophoretic motion of charged particles of complex topologies in microfluidic flows within different channel geometries. The variation in flow-director coupling affects the electrophoretic motion of such particles as the shear stress gradient varies non-linearly.

Overuse of broad-spectrum antibiotics due to slow traditional methods for bacteria identification and antibiotic susceptibility testing (ID & AST) has heightened public health concerns about antibiotic resistance. Despite advances like PCR and FISH, these methods have limitations, including contamination risk and limited accuracy. We introduce the Multi-Dimensional Laser-Induced Microfluidic Valve System, which integrates a microfluidic valve platform, a multi-channel laser-induced fluorescent single-cell resolution optical detection system, thermal permeation assays, and a machine learning algorithm. This system achieves 98% accurate bacteria ID in under 5 minutes and high-throughput AST in under 1 hour, offering a rapid, sensitive, and specific solution to combat antibiotic resistance.

Gallium-based liquid metals are interesting materials due to their high surface tensions, low viscosities, and high electrical conductivities comparable to classical solid metals. They have been used for several applications in microelectromechanical systems (MEMS) and liquid metal microfluidics (LMMF) including micro-actuators and sensors. However, when it comes to the electrochemical actuation of LMs, in particular Galinstan, two main challenges must be tackled: 1) Generation of reproducible Galinstan plugs with controllable size and interspacing, and 2) Tendency to alloy with the most common metals used for electrodes such as gold (Au), platinum (Pt), titanium (Ti), nickel (Ni), and tungsten–titanium (WTi). At NeptunLab, we designed an on-chip liquid metal plug generator for generating reproducible and scalable sub-millimeter plugs. Using constricted channels based on Laplace pressure, we generated reproducible plugs in terms of size and interspacing compared to pulse width modulation (PWM)-based generation in straight channels. To avoid the second challenge, i.e., alloying, we carried out in situ electrodeposition of electrically conductive polypyrrole (PPy) on the electrodes to provide a conductive alloying-barrier layer between Galinstan and electrodes. While Galinstan alloying was a challenge in contact with electrodes, its amalgamation with heavy metal ions (HMI) like lead was an advantage for developing HMI sensors. Common droplet-based LM HMI sensors showed limitations in scalability and the homogeneity of the surface. To overcome these problems, we introduced tungsten oxide (WO)-Galinstan electrodes fabricated via photolithography and galvanic replacement (GR) in a tungsten salt solution. The in-house fabricated electrodes demonstrated enhanced sensitivity in comparison with electrodes structured from pure Galinstan and detected lead at concentrations down to 0.1 mmol·L−1. Further details of this abstract in gallium-based liquid metal microfluidics for applications in microactuators and sensors will be discussed in the presentation.

This whitepaper outlines a production chain demonstrating cost-effective manufacturing for microfluidics components, achieved through collaboration between Swiss firms FEMTOPrint and 3D AG, alongside FHNW University of Applied Sciences and Arts Northwestern Switzerland. Employing selective glass etching, FEMTOPrint uses femtosecond laser writing and wet etching on a transparent substrate to create intricate shapes in a single step. 3D AG employs electroforming to create nickel shims with nanometer precision from this master template, transferred to a tool for injection molding with precision laser cutting. The process ensures high replication fidelity and optimal pattern appearance in materials like PMMA, COC, and COP, highlighting advancements in microfluidics research and applications.

While nanofluidics have been researched, these channels must be fabricated via complicated processes such as precision nano machining and wafer bonding. We successfully fabricated micro and nanometer-sized channels mechanically via diamond tool in the subsurface in glass. In this study, we applied these channels as nanofluidic channels. First, we characterized the channel size and found it can be controlled via mechanical scribing force. Then, we designed and fabricated a glass mini chip, including the formed crack channels connected with the inlet and outlet for fluid. The fluid’s behavior was investigated as a demonstration.

One-step fabrication of stamps with elastic hollow microneedles on the porous membrane is realized by two-photon polymerization (2PP) 3D printing for Capillary Stamping. Microdroplets can be deposited onto the counterpart substrate by supplying the ink from the back side of the stamp. The ink can be supplied at any moment during the printing process therefore allowing multiple successive substrate contacts with each single contact element, addressing the ink depletion problem involved in Microcontact Printing (µCP) and Polymer Pen Lithography (PPL). The fabricated stamps fit a 3D-printed stamp holder compatible with automatic printing devices designed for PPL. The reported method herein may pave the way for further advancing the contact lithography research field.

Silicon photonic (SiP) biosensors promise portable and analytical systems at the point-of-care. An on-chip in-channel flow sensing solution could improve the robustness and accuracy of SiP biosensor-based measurements. This work presents the design, integration, characterization, and validation of an on-chip silicon photonic microfluidic thermal flow-rate sensor (SiP-MTFS) on a silicon-on-insulator (SOI) process to non-intrusively measure the flow rate inside a microfluidic channel. The calorimetric sensor architecture consists of a titanium-tungsten (TiW) microheater placed under the fluidic channel with two silicon photonic microring resonators placed equidistantly upstream and downstream from the microheater. The proposed SiP-MTFS design is compact, inexpensive, translatable to any photonic foundry process, convenient for integration and readout, and comparable in performance to commercial off-chip calorimetric flow rate sensors. This work demonstrates the first successful attempt at integrating a compact, inexpensive SiP microfluidic thermal flow-rate sensing solution on an SOI process.

This study introduces a low-cost, compact 3D printed microfluidic platform for label-free cell analysis using digital holographic microscopy (DHM). By integrating inertial focusing microfluidics, DHM, and deep learning methods, this platform advances cell analysis in real-time. Utilizing the robustness of conventional fused deposition modeling (FDM) printers and the versatility of silicone extrusion-based 3D printing, various channel designs are fabricated, achieving precise cell alignment and efficient imaging. The system leverages deep learning to enhance cell viability assessment from holographic images, demonstrating potential for high-throughput, precise cellular analysis.

Advances in stem cell technology allow the reprogramming of patient-derived cells, obtained from urine samples or skin biopsies, into induced pluripotent stem cells (iPSCs), which can then be differentiated into any cell type. With this goal in mind, we will present how experimental developments, mostly in the field of microfluidics, can be used to extract several relevant spatiotemporal biophysical properties from the quantitative phase signal provided by digital holographic microscopy, thus moving towards the realization of a truly effective optofluidic platform for non-invasive characterization of cell structure and dynamics. Such label-free characterization is highly conducive to the identification of disease-specific cell phenotypes when comparing iPSC-derived cells from control and diseased patients.

3D culture on microfluidic devices holds great potential for in vivo behavior representation, but faces inherent challenges, particularly in distinguishing cancerous from healthy tissue. We propose here to use Raman spectroscopy on microfluidic chips to achieve on-chip tissue differentiation. This study optimizes Raman spectrum acquisition for better signal-to-noise ratio and acquisition speed, examining PDMS parasitic spectrum, thickness, probe distance, sample size, signal specificity, and intensity. Results show an optimal Raman shift window for PDMS acquisitions, with ideal PDMS thickness under 4mm, sample distance of at least 0.50mm, and sample thickness no more than 1mm. These configurations enhance signal detection accuracy.

This work reports on the development and use of our laser-based, direct-write technique, which uses a visible laser (c.w., operating at 405nm) and a cheap off-the-shelf light-sensitive polymer to create in paper and paper-like porous materials, fluidic devices that enable rapid measurement of targeted analytes for disease diagnosis.

Sustainable development of bio and nanomaterials is an emerging trend in the general field of materials but also in biomedical engineering. The development of biomaterials includes many, time consuming and complicated steps, from lab synthesis, in vitro cell tests and animal trials to scaling up and eventually clinical translation. In this work we give two examples regarding how microfluidics can contribute to more sustainable biomaterials at two different steps of their development; a) in vitro evaluation; b) to the synthesis of nanoparticles and scaling up.

Neurostimulation devices, cardiac pacemakers, and electrophysiology mapping catheters diagnose, monitor, and treat various neurological and cardiac disorders. Enhancing these devices requires miniaturized electrodes with exceptional electrochemical performance, enabling higher spatial resolution, precision, and reliability for better patient compatibility and comfort. Achieving this miniaturization necessitates reducing the geometric surface area while increasing the electrochemical surface area of the electrodes. Highly electroactive electrode materials or surfaces are crucial for improving electrochemical performance, ensuring sufficient charge delivery across the electrode/tissue interface for stimulation, and maintaining low impedance for sensing and recording applications. This research introduces "hierarchical surface restructuring" using femtosecond laser technology, an innovative, scalable, and commercially viable electrode surface treatment. We demonstrate how this technology improves stimulation precision, increases electrode density, and significantly reduces electrochemical impedance.

In this work we present “sticker moulds” for ultra-rapid prototyping. For fabrication of PDMS devices, these moulds are comprised of a glass substrate with layers of Uline Industrial Tape S-423 patterned by laser ablation. We examine the geometry and surface roughness of these moulds and the resultant PDMS devices using microscopy and profilometry . Channels dimensions as small as 150µm wide and 50µm deep are achieved, with a W-shaped kerf. These moulds have low material cost (< 10 cents per mould), require under 10 minutes to fabricate, and have low surface roughness, resulting in smooth PDMS devices that readily bond. For devices directly printed on textiles, moulds developed from simple sticky paper can also be patterned using laser ablation, and adhered to the textile or existing ink patterns. The low barrier to entry makes sticker mould ideal for rapid proof-of-concept work where designs are frequently changing, or a classroom setting . Furthermore, we employ sticker moulds for development of electronic and microfluidic structures directly on textiles for wearable biomedical systems.

Biological systems often incorporate complex, micron-scale, highly vascularized architectures within soft matrices, which are challenging to replicate with conventional engineering methods. Projection micro-stereolithography (PµSL) patterns high-resolution features, but it typically employs densely crosslinked, high-moduli materials that do not align with the properties of tissues. Volumetric additive manufacturing (VAM) can cure volumes of photosensitive biomaterials in a single step. However, VAM patterning of high-resolution features with biomaterial resins can be challenging due to the natural scattering of proteins. To overcome these limitations, we developed a material strategy to merge both PµSL and VAM print platforms to obtain high-resolution vascular features within biomaterial scaffolds with superior mechanical integrity. We print a high-resolution template with the PµSL printer using a water degradable photocurable resin. The template is overprinted with VAM utilizing bioderived materials. The overprinted structures are then incubated under mild, basic conditions to dissolve the sacrificial templates and release high-resolution vascular features within soft scaffolds.

Nowadays, many (micro)fluidic applications require measurement of more than one parameter. For instance, to monitor the cell cultures in bioreactors (e.g. to produce proteins or pharmaceuticals), multiple parameters, including flow rate, pressure, pH, glucose concentration, conductivity, off-gas concentration (O2, CO2, air), temperature and humidity, have to be measured. Now, the question is, how can we design and realise multiparameter measurement systems in such a way that they can be flexibly adapted to and fulfil the needs of the many diverse microfluidic applications. To answer this question, we will have a look at three important aspects, namely, the system itself, the transducers in the system, and the data processing of the signals that are generated by the sensors in the system. Examples of all aspects will be given for several different (biomedical) applications, including organ-on-a-chip, multi-infusion systems, ventilator systems and flow chemistry. The pros and cons of each approach will be discussed and an outlook on future research will be given.

This paper presents a broadband biosensor based on a high resistive silicon (HR-Si) dielectric waveguide ring resonator (DWGR2). The biosensor can be used to detect various chemical or hazardous biological fluids. The crucial optimization steps for a broadband sensor design between 320 and 400 GHz are presented, as well as the characterization of different fluids and their evaporation times. As a proof of concept, the measurement has been performed using a Vectorial Network Analyzer (VNA) and the relative permittivity of fluids can be estimated over this frequency. Furthermore, an automated tracking algorithm can be designed to extract all resonance information.

Urinary tract infections (UTIs), typically caused by E.coli, pose detection challenges in point-of-care settings due to slow and insensitive traditional methods. This study improves UTI detection by integrating microfluidic sample processing for bacterial enrichment with photothermal sensing in lateral flow immunoassays (LFAs). Using polyethylene oxide (PEO) in various microchannel designs, we achieved rapid E.coli enrichment. Photothermal sensing showed enhanced sensitivity, correlating signals to E.coli concentrations, promising a rapid, sensitive, and affordable UTI detection method with broad healthcare and environmental applications.

We demonstrate a low-cost, easily manufacturable multichannel on-chip pump integrated into a microfluidic chip for organ-on-a-chip applications. Using femtosecond laser techniques, we fabricated 3D microfluidic channels from biocompatible materials, including medical-grade tapes, acrylic, polystyrene, and deformable rubber, enabling complex fluid routing and on-chip valves and pumps. A deformable film serves as an elastic membrane to control channel opening and fluid movement. In the active-valve and pump design, the film is actuated by air pressure ranging from -5 to 10 psi, allowing fluid delivery or withdrawal across multiple channels with flow rates from 0.1 µL/min to 100 µL/min. A passive-valve design prevents backflow. Our device was used for iPSC-derived pancreas islet culture and precise drug injection, demonstrating stable operation over 12 days without significant drug absorption. This on-chip pump and valve system offers essential functionalities for organ-on-a-chip applications at low cost.

Water is fundamental for the development and sustenance of life and for anthropical activities. Many diseases can occur in the presence of bacterial strains contamination. The aim of this work is to design, fabricate, and test a low-cost, miniaturized detection system employing the chemiluminescent reaction between luciferin and adenosine triphosphate (ATP), the energy unit in biological systems, to measure the concentration of ATP in water. The ATP–luciferin chemiluminescent solution is introduced in a 3D printed lab-on-chip microfluidics (LoC) system and is faced with a silicon photomultiplier (SiPM) for highly sensitive real-time detection, aiming to detect ATP concentrations down to 0.5 nM.

Solid metal microneedles are a promising technology for transdermal drug delivery and biosensing with minimal pain. However, their complex microfabrication processes are difficult to scale for mass production, hindering widespread clinical adoption. We previously developed a scalable method of fabricating solid metal microneedles using a modified wire bonding process, well-suited for mass production due to its roots in high-throughput semiconductor manufacturing. Here we demonstrate the capability of wire bonded microneedles to be used in drug delivery and biosensing applications through the addition of different coatings on the microneedle surface, underscoring their potential as a versatile platform technology for diverse applications.

The profound impact of stress on health underscores the need for precise and objective measurement techniques, as traditional self-report questionnaires often fall short in reliability. Molecular biosensors, especially those targeting cortisol—a critical stress hormone—offer a compelling alternative. This study presents an innovative wearable sensing system that utilizes a molecularly imprinted polymer-radiofrequency (MIP-RF) mechanism for non-invasive, real-time cortisol detection in sweat. The system is designed to be wireless, flexible, battery-free, reusable, and environmentally stable, ideal for long-term use with an inductance-capacitance transducer that translates cortisol levels into resonant frequency shifts with high sensitivity within a physiological range of 0-1μM. The device is further enhanced with NFC for seamless, battery-free operation and a 3D-printed microfluidic channel for direct sweat collection, facilitating continuous cortisol monitoring during daily activities. Validation through circadian rhythm tracking, which compares morning and evening cortisol levels, confirms its effectiveness for precise molecular stress biomarker detection.

Traditional methods for assessing stress, such as self-reported questionnaires and laboratory assays, have limited practical applications for continuous stress monitoring due to the non-quantitative measurements and lack of standardization, and high operation costs, and are time consuming. To overcome these limitations, we have developed a machine learning based, wireless, battery-free, flexible wearable system for continuous stress monitoring. The galvanic skin response (GSR) sensor is fabricated through cost-effective screen-printing of Ag onto a flexible SEBS substrate, ensuring excellent mechanical stability and low contact impedance due to its epidermal contact with the skin, resulting in a high signal-to-noise ratio. We employed NFC technology for both wireless power and data transmission, to realize a battery-free wireless readout system for signal collection and enhance wearability and comfort. The GSR data is wirelessly transmitted to a mobile phone for real-time processing and analysis. ML models were employed to classify stress events, accurately distinguishing between stress and rest states based on GSR signals collected during a Stroop test.

Tactile displays are very popular among visually disabled and blind people for haptic reading and accessing graphs. Navigating through a haptic display would be easier through touch instead of buttons. Therefore, a touch screen is a necessity particularly when dealing with graphics. We developed a flexible capacitive touch screen based on silver nanowires (AgNWs) on a polyimide (PI) substrate with a total thickness of 210 µm. To classify eight hand gestures, we have implemented a hand gesture recognition application using deep learning and achieved an accuracy of 97% acquired from 300 measurements per gesture.

A 3D printed microfluidic device for microscopic worm (C. Elegans) sorting was demonstrated using actuated syringe-based fluid flow up to 11.7 µL/min. Particles (10 µm polystyrene microspheres) were also sorted with sorting efficiency of 62% with 78 particles total or 2.9 beads/min. This demonstrated an inexpensive, quick and facile fabrication approach to microfluidic worm sorting.

Artificial collagen membranes (ACM) are a novel cell-based model system for preclinical drug testing and basic research. Here we show how a custom pneumatic actuation device and a commercial spectrometer-based optical coherence tomography (OCT) system can be used to create and examine in vivo-like ACM conditions in a microphysiological system. Based on static and dynamic pressure changes with synchronized OCT acquisition, the capabilities of the monitoring system for cell culture and testing are demonstrated.

Continuous flow cell sorting based on image analysis is a powerful concept that exploits spatially-resolved features in cells for sorting them. However, most of the systems published so far are very complex, offer only limited image quality and/or are limited to a specific imaging modality. We present a low-complex microfluidic microscope add-on that allows for image-activated cell sorting. The cells are handled using negative dielectrophoresis (DEP) at a sample volume throughput of ca. 1 µl/min, making use of low flow velocities around 1 mm/s, enabling AI-enhanced image acquisition and processing, and reducing shear stress on the cells. Core of the system is a y-shaped microfluidic channel. DEP electrodes align the cells before those enter the microscopic field of view for image analyis using artificial intelligence. For a detected target cell a sorting electrode array downstream of the imaging area is activated, guiding the cell into a dedicated outlet. The system sorts thousands of live T cells with sample volume throughputs in the range of µl/min, achieving purities between 80% and 95% in yield-mode and enabling recovery rates of >85% for targeted cells.

Salmonella is a major cause of foodborne illness worldwide, costing $3.7 billion annually. Despite national efforts to improve detection, infection rates have remained stagnant for 30 years. This study develops an impedance-based microfluidic biosensor for rapid Salmonella detection in raw turkey rinsate. The biosensor includes focusing, trapping, and detection regions to concentrate, capture, and identify Salmonella through antigen-antibody binding, resulting in measurable impedance changes. It detects Salmonella serotypes with high selectivity, achieving a limit of detection of 1 cell/mL in one hour. The sensor distinguishes low concentration of live cells from high concentration of dead ones and non-specific pathogens.

Micro-epidermal actuator is an essential component of flexible conductive hearing aids to generate vibrations on skin and bypass outer and middle ear. The actuators are made of piezoelectric actuators embedded into a soft polydimethylsiloxane (PDMS) substrate. The mechanical properties of the piezoelectric actuators do not match with underlying flexible substrate and soft tissues that affect vibrational transmission to the skin and bone. PDMS mechanical properties including Young’s modulus can be engineered by changing the temperature and base to curing agent ratios to improve the vibrational transmission and increase the vibrations in bone and cochlea. The experimental data showed the velocity of vibration increased by roughly 7 dB at 9 kHz when Young’s modulus of the PDMS was reduced from 1 MPa to 200 kPa. Furthermore, the results showed that attenuation in the flexible substrate can be reduced by engineering PDMS Young’s modulus.

A novel microneedle (MN) sensor coated with a glucose-sensitive composite material was developed to measure glucose levels in interstitial fluid (ISF). The composite combines a polymer containing glucose-responsive phenylboronic acid (PBA) moieties with conductive carbon nanotubes (CNTs). This design leverages the polymer's reversible swelling response to glucose, altering CNT distribution and, consequently, electrical percolation on the MN surface. We optimized PBA content and CNT loading for sensitivity by testing the MN in an ISF-mimicking gel. In animal models, the MN sensor's electrical current responses closely correlated with blood glucose levels, demonstrating clinical-grade accuracy comparable to conventional glucometers.