Sperm sorting plays a pivotal role in assisted reproductive technologies (ART), ensuring the selection of functionally competent spermatozoa for enhanced fertilization, implantation, and pregnancy outcomes. Traditional approaches such as swim-up, density gradient centrifugation, and simple wash techniques, though widely applied, are often limited by mechanical stress, reduced efficiency, and compromised sperm integrity. Parallel efforts in sex-based sorting using gradients, surface charge, or immunological strategies have also faced challenges, including high cost, DNA damage, and limited clinical translation. Recent advances in microfluidics have emerged as a game-changer in ART, offering biomimetic micro-environments that replicate sperm guidance mechanisms, including chemotaxis, thermotaxis, and rheotaxis, while enabling precise motility-based sorting with minimal manipulation. Microfluidic platforms not only improve the isolation of morphologically superior and genetically intact spermatozoa but also provide scalable, rapid, and less invasive alternatives to conventional methods. Here, we review the evolution of sperm sex and motility sorting technologies, highlight recent progress in microfluidic platforms, and consider their implications for livestock breeding. Innovations in microfluidic device design, including microchannel geometries, integrated flow cytometry, photonic separation, and dielectrophoresis, demonstrate transformative potential for sex sorting. Finally, we discuss future directions, including the integration of automation, biosensing, and artificial intelligence, that could establish next-generation platforms for precision sperm selection in reproductive medicine and biotechnology.

Organ-on-chip systems are designed to recapitulate microphysiological environments for in vitro evaluation of drug safety and efficacy, offering a promising alternative to animal models in pharmaceutical research. However, in vivo organs are separated by distinct barriers yet interconnected through systemic circulation. Inspired by modular interlocking systems, we developed a reconfigurable microfluidic chip array unit for constructing a multi-organ-on-a-chip (MoC) platform. Each unit features a bicompartmental design with integrated microchannels, supporting static or bidirectional dynamic perfusion. Units can be assembled into linear arrays for high-throughput experiments or matrix arrays for systemic analysis. As a proof of concept, we assembled a liver-tumour system and found that varying the liver-to-tumour cell ratio significantly altered drug efficacy and toxicity, demonstrating inter-organ crosstalk. This modular platform offers a scalable tool for studying systemic drug responses and inter-organ interactions, with strong potential to reduce the reliance on animal models in pharmaceutical research.

We engineered a compact methanol steam reforming (MSR) system tailored to power a 1 kW High-Temperature Proton Exchange Membrane (HT-PEM) fuel cell. The unit integrates an evaporator, reformer, and burner within a cylindrical titanium-alloy vacuum flask to minimize parasitic heat loss. Guided by an Artificial Intelligence Complex System Response (AICSR) framework, we developed a segmented catalyst architecture that positions an optimized Pd/ZnO/Al2O3 catalyst downstream of a commercial Cu–Zn catalyst bed. This spatial configuration reduces palladium consumption by >50% while maintaining a hydrogen generation rate of 8000 sccm at 250 °C. During a 40 h stability test, the system exhibited a low deactivation rate of 0.235% h−1, with methanol conversion decaying gradually from 98.1% to 88.7%. The downstream PdZn intermetallic phase actively promoted the water–gas shift (WGS) reaction, restricting CO concentration to an average of 3.9% (minimum 2.5%). Achieving a system thermal efficiency of 88.589% and a 20 min startup time, this design validates AI-assisted spatial catalyst distribution as a highly viable strategy for compact hydrogen generation.

Sex sorting of bovine sperm is a critical technique in the livestock industry to predetermine the sex of offspring, thereby enhancing herd management, breeding efficiency and economic value. Commercial methods such as flow cytometry, though effective, are expensive, complex, and may compromise sperm viability and health due to highpressure and staining processes. The proposed novel bionic microfluidic platform utilizes silica-coated magnetic nanoparticles with selective binding affinity to Y chromosome-bearing sperm cells. By embedding a magnetic strip underneath the microfluidic channel, Ysperm cells are efficiently trapped, while X-sperm cells continue unimpeded toward collection with an efficiency of 70% in X and $80 \% \mathrm{Y}$.

Circulating tumor cells (CTCs) are essential biomarkers for cancer prognosis, yet their extreme rarity and biological heterogeneity pose significant challenges for label-free detection. This study presents an automated, non-invasive classification framework integrating a self-assembly cell array (SACA) microfluidic chip with hyperspectral imaging (HSI) and deep learning. By utilizing the SACA chip’s 5 µm gap design, patient-derived blood samples were organized into a flattened monolayer, ensuring high-purity spectral acquisition by minimizing cell overlapping. We implemented two deep-learning pipelines: an Attention-Based Adaptive Spectral–Spatial Kernel ResNet (A2S2K-ResNet) for pixel-level feature extraction and a modified ResNet50 for structural image analysis. While spectral classification achieved ~80% accuracy for cultured cell lines, its performance on patient-derived CTCs was hindered by subtle spectral overlap with white blood cells (WBCs). To overcome this, a multi-band ensemble strategy using majority voting across seven optimized spectral bands (470–900 nm) was developed. This hybrid approach significantly enhanced detection robustness, achieving an overall accuracy of >93.5% and precision exceeding 92%. These results demonstrate that combining microfluidic spatial control with multi-band deep learning offers a reliable, label-free pipeline for clinical liquid biopsy and real-time cancer monitoring.

Harnessing the full solar spectrum for sustainable hydrogen production remains a major challenge in photoelectrocatalytic (PEC) water splitting. Herein, we present a cascaded microfluidic PEC reactor integrated with a thermoelectric generator (TEG), achieving a Solar‐to‐Hydrogen (STH) conversion efficiency of 28%. The device combines three synergistic elements: (i) Ti 3 C 2 ‐CdS heterostructure catalysts that enhance charge separation and suppress recombination; (ii) a planar microfluidic reactor that ensures uniform light penetration, laminar flow, and efficient mass transport; and (iii) a Bi 2 Te 3 ‐based TEG module that harvests solar waste heat to provide supplemental bias for overcoming kinetic barriers. The cascaded architecture enables sequential light harvesting across four reactors, leading to cumulative hydrogen yields exceeding 10 890 µmol g −1 h −1 , while simultaneously enabling rapid water treatment. This work establishes a scalable, self‐powered, and multifunctional platform for decentralized clean energy generation and water purification by integrating thermal‐electrics and PEC pathways into a single compact device.

Accurate in vivo glucose monitoring is essential for effective diabetes management and for the care of pre-term infants in critical care. Glucose-monitoring techniques are broadly categorized into three types: invasive, minimally invasive, and non-invasive. Each method presents distinct advantages and challenges. Non-invasive glucose sensors, despite impressive advancements in recent years, still face issues with signal interference and accuracy, limiting their widespread clinical application. In contrast, implanted devices offer more reliable and consistent results in clinical settings, making them the current gold standard. This review provides an overview of the leading glucose-sensing technologies, detailing both their advantages and drawbacks. We discuss invasive techniques, such as implanted electrodes, which allow continuous glucose monitoring with high accuracy, but often come with risks of infection and discomfort. Minimally invasive methods, such as fluorescence sensors, Raman sensors, and microneedle arrays, aim to reduce discomfort while providing more precise measurements than non-invasive devices. Additionally, non-invasive methods, such as optical, infrared, and microwave techniques, are explored for their potential to provide pain-free, continuous glucose monitoring. Finally, the review highlights a brief comparison among the current technologies and future directions in the field, particularly the use of signal enhancement algorithms and integration with wearable devices.

The integration of Artificial Intelligence (AI) techniques with medical kits has revolutionized disease diagnosis, enabling rapid and accurate identification of various conditions. We developed a novel deep learning model, namely DeepATsers based on a combination of CNN and GAN to employ a one-pot SERS biosensor to rapidly detect COVID-19 infection. The model accurately identifies each SARS-CoV-2 protein (S protein, N protein, VLP protein, Streptavidin protein, and blank signal) from its experimental fingerprint-like spectral data introduced in this study. Several augmentation techniques such as EMSA, Gaussian-noise, GAN, and K-fold cross-validation, and their combinations were utilized for the SERS spectral dataset generalization and prevented model overfitting. The original experimental dataset of 126 spectra was augmented to 780 spectra that resembled the original set by using GAN with a low KL divergence value of 0.02. This significantly improves the average accuracy of protein classification from 0.6000 to 0.9750. The deep learning model deployed optimal hyperparameters and outperformed in most measurements comparing supervised machine learning methods such as RF, GBM, SVM, and KNN, both with and without augmented spectral datasets. For model training, a whole range of spectra wavenumbers ( $$320 \hbox { cm}^{-1}$$ to $$1650 \hbox { cm}^{-1}$$ ) as well as wavenumbers ( $$1078 \hbox { cm}^{-1}$$ and $$1582 \hbox { cm}^{-1}$$ ) only for fingerprint peak spectra were employed. The former led to highly accurate 0.9750 predictions in comparison to 0.4318 for the latter one. Finally, independent experimental spectra of SARS-CoV-2 Omicron variant were used in the model verification. Thus, DeepATsers can be considered a robust, generalized, and generative deep learning framework for 1D SERS spectral datasets of SARS-CoV-2.

This study develops a machine learning approach to detect SARS-CoV-2 variants (Beta, Gamma, Omicron) using surface-enhanced Raman spectroscopy with a barcode-based multiplex assay. A dataset of 54 SERS spectra, comprising 28 Positive and 26 Negative samples, was analyzed, targeting characteristic peak intensities at $380 ~\text{cm}^{-1}$ (Omicron), $540 ~\text{cm}^{-1}$ (Beta), and $1336 ~\text{cm}^{-1}$ (Gamma). Raw spectra were preprocessed with baseline correction, normalization, and smoothing to minimize noise. Features, including peak intensities, their ratios, and total spectral intensity, were selected to distinguish Positive and Negative samples. Three machine learning models SVM, RF, and LR were trained on 80 % of the data (43 samples) and tested on 20 % (11 samples), using five-fold cross-validation to prevent overfitting. SVM achieved the highest accuracy (90.00 %) and perfect precision ($\mathbf{1. 0 0 0 0}$), followed by $\mathbf{R F}(\mathbf{8 1. 8 2 \%})$, while LR recorded the lowest accuracy $(72.73 \%)$, limited by the small dataset and complex, non-linear SERS spectra. The framework supports extension to other pathogens, such as Influenza A/B, using additional reporters. The limited dataset size necessitates further data collection and validation. Future work will explore deep learning models, like convolutional neural networks, and employ Gauss-Lorentz methods for data augmentation to enhance dataset size and model robustness for automated, point-of-care diagnostics.

Tumor acidosis is a consequence of altered metabolism that primarily takes place due to lactate secretion from anaerobic glycolysis. As a result, many regions within the tumors are chronically hypoxic and acidic. To measure the intratumor pH dynamically, we have fabricated a biocompatible pH nanoparticle sensor using surface-enhanced Raman spectroscopy (SERS-pNPS) and monitored continuous pH levels in three-dimensional multicellular spheroids. The 3D multicellular spheroids were cultured using a micro-well array chip made of polydimethylsiloxane (PDMS). The SERS-pNPS were synthesized by linking 4-Mercaptobenzoic acid (4-MBA) to silver nanoparticles (AgNPs) of size 50 nm. The calibration curve demonstrates a linear correlation between the ratio of Raman peak intensities (1378 cm-1/1620 cm-1) with the pH level. The sensor exhibits a detection limit of pH 4.4 and demonstrates linearity within the physiological pH range (pH 4.4-pH 8.23). The SERS-pNPS was applied for pH measurement in different 3D co-cultured spheroid models such as lung cancer (A549-NIH3T3), breast cancer (MCF-NIH3T3), colon cancer (HCT8-NIH3T3) and mono-cultured spheroids using fibroblast (NIH3T3) cells. The detailed analysis indicated that the 3D co-cultured cancerous tumor models have 16% more acidic microenvironment as compared to 3D mono-cultured spheroid model. Also, a presence of a decreasing pH gradient from peripheral to the core region is observed in both the cases indicating acidosis in the core region. The SERS-pNPS platform facilitates a non-invasive and dynamic pH tracking, and thus offers an improved insight into the acidic microenvironment in various tumor models.

Podocytopathy, characterized by proteinuria, contributes significantly to kidney diseases, with hypertension playing a key role in damaging podocytes and the glomerular filtration barrier (GFB). The lack of functional in vitro models, however, impedes research and treatment development for hypertensive podocytopathy. We established a novel constant pressure-driven podocyte-on-chip model, utilizing our previously developed dynamic staining self-assembly cell array chip (SACA chip) and 3D printing. This platform features a differentiated podocyte monolayer under controlled hydrostatic pressures, mimicking the epithelial side of the GFB. Using this platform, we investigated mechanical force-dependent permeability to three sizes of fluorescent dextran under varying hydrostatic pressures, comparing the results with a puromycin aminonucleoside (PAN)-induced injury model. We observed that external pressures induced size-dependent permeability changes and altered cell morphology. Higher pressures led to greater macromolecule infiltration, especially for larger dextran (70 kDa, 500 kDa). Mature podocytes exhibited immediate, pressure-dependent cytoskeleton rearrangements, with better recovery at lower pressures (20 mmHg) but irreversible injury at higher pressures (40, 60 mmHg). These morphological changes were also corroborated by dynamic mRNA expression of cytoskeleton-associated proteins, Synaptopodin and ACTN4. This platform offers a promising in vitro tool for investigating the pathomechanisms of hypertension-induced podocytopathy, performing on-chip studies of the GFB, and conducting potential drug screening.

The flow channels of bipolar plates are crucial for high-temperature proton exchange membrane fuel cells, as they manage the distribution of fuel to the anode and air to the cathode, while facilitating the discharge of reaction products from the cathode. Traditional serpentine co-current flow plates face challenges like uneven fuel concentration, temperature distribution, and inconsistent load across large areas, limiting power output and longevity. This study proposes a clover-type counter-flow plate to address these issues by using symmetric, lowresistance channels. Experimental results show a 42.8 % power density increase when the reaction area expands from 1 cm2 to 20 cm2 and a 44 % increase when further expanding to 154 cm2. The counter-flow design also offers better long-term stability, with only a 1 % power density decrease after 48 h of operation compared to 2 % for the parallel flow design. These findings indicate that the clover-shaped counter-flow bipolar plate effectively enhances performance and durability, especially in large-area and long-duration HT-PEMFC applications.

Download This Paper Open PDF in Browser Add Paper to My Library Share: Permalink Using these links will ensure access to this page indefinitely Copy URL Copy DOI

Unraveling the intricacies between oxygen dynamics and cellular processes in the tumor microenvironment (TME) hinges upon precise monitoring of intracellular and intratumoral oxygen levels, which holds paramount significance. The majority of these reported oxygen nanoprobes suffer compromised lifetime and quantum yield when exposed to the robust ROS activities prevalent in TME, limiting their prolonged in vitro usability. Herein, the ruthenium-embedded oxygen nano polymeric sensor (Ru-ONPS) is proposed for precise oxygen gradient monitoring within the cellular environment and TME. Ru-ONPS (approximate to 64 +/- 7 nm) incorporates [Ru(dpp)(3)]Cl-2 dye into F-127 and crosslinks it with urea and paraformaldehyde, ensuring a prolonged lifetime (5.4 mu s), high quantum yield (66.65 +/- 2.43% in N-2 and 49.80 +/- 3.14% in O-2), superior photostability (>30 min), and excellent stability in diverse environmental conditions. Based on the Stern-Volmer plot, the Ru-ONPS shows complete linearity for a wide dynamic range (0-23 mg L-1), with a detection limit of 10 mu g mL(-1). Confocal imaging reveals Ru-ONPS cellular uptake and intratumoral distribution. After 72 h, HCT-8 cells show 5.20 +/- 1.03% oxygen levels, while NIH3T3 cells have 7.07 +/- 1.90%. Co-culture spheroids display declining oxygen levels of 17.90 +/- 0.88%, 10.90 +/- 0.88%, and 5.10 +/- 1.18%, at 48, 120, and 216 h, respectively. Ru-ONPS advances cellular oxygen measurement and facilitates hypoxia-dependent metastatic research and therapeutic target identification.

Nanoporous glucose-based active carbon nanospheres (g-ACNS) with high efficiency and stability in hydrogen (H2) storage is synthesized by a hydrothermal method followed by multiple KOH activation processes. For the optimized...

Early prognosis of cancer recurrence remains difficult partially due to insufficient and ineffective screening biomarkers or regimes. This study evaluated the rare circulating tumor microemboli (CTM) from liquid biopsy individually and together with circulating tumor cells (CTCs) and serum CEA/CA19-9 in a panel, on early prediction of colorectal cancer (CRC) recurrence. Stained CTCs/CTM were detected by a microfluidic chip-based automatic rare-cell imaging platform. ROC, AUC, Kaplan-Meier survival, and Cox proportional hazard models regarding 4 selected biomarkers were analyzed. The relative risk, odds ratio, predictive accuracy, and positive/negative predictive value of biomarkers individually and in combination, to predict CRC recurrence were assessed and preliminarily validated. The EpCAM+Hochest+CD45- CTCs/CTM could be found in all cancer stages, where more recurrences were observed in late-stage cases. Significant correlations between CTCs/CTM with metastatic stages and clinical treatment were illustrated. CA19-9 and CTM could be seen as independent risk factors in patient survivals, while stratified patients by grouped biomarkers on the Kaplan-Meier analyses presented more significant differences in predicting CRC recurrences. By monitoring the panel of selected biomarkers, disease progressions of 4 CRC patients during follow-up visits after first treatments within 3 years were predicted successfully. This study unveiled the value of rare CTM on clinical studies and a panel of selected biomarkers on predicting CRC recurrences in patients at the early time after medical treatment, in which the CTM and serum CA19-9 could be applied in clinical surveillance and CRC management to improve the accuracy.

Abstract Expansion Microscopy (ExM) enhances biological imaging to the nanoscale level by enlarging biological specimens, increasing the separation between biomolecules while maintaining their spatial arrangement. By integrating ExM with traditional microscopy, super-resolution volumetric images can be captured efficiently without compromising tissue integrity. Accordingly, fine molecular structures previously indiscernible by means of standard fluorescence microscopy can be visualized. However, one of the major challenges with ExM is that the samples are embedded in the polymerized gel, which limits the use of high numerical aperture objective lenses due to their short working distances. To address this issue, we have developed an approach using carboxymethyl tamarind kernel gum (CMTKG) cross-linked with sodium acrylate to create a superabsorbent hydrogel suitable for ExM. This technique facilitates removal of excess polymers from the expanded sample, resulting in sample exposure to the objective lens. Here, we demonstrate that it is possible to achieve approximately 7X expansion of Drosophila melanogaster brains, free from surrounding hydrogel, in a process we have named Naked-ExM. Followed by iterative expansion, the centimeter-sized fly brain is obtained and the ~15 um sized synapses are measured at 15X expansion.

Evaluation of glucose in cancerous tumours acts as a hall mark for cancer progression and thus serves as a major interest in these days. Here, we have developed a biocompatible, surface enhanced Raman spectroscopy -based glucose Nano Particle Sensor (SERS-gNPS) to measure glucose dynamically in 3D Co -cultured Colon Cancer Tumour Spheroids (3D-CCTS). 3D-CCTS were produced using polydimethylsiloxane (PDMS) based mu-well array chip. SERS-gNPS was fabricated by conjugating 4-Mercaptophenylboronic acid (4-MPBA) to 50 nm silver nanoparticles (AgNP) and used under confocal Raman spectroscopy to measure glucose. The calibration curve shows direct relationship between Raman peak intensity at 1078 cm -1 with increasing glucose levels. The sensor has a limit of detection of 0.1 mM and found to be linear in the physiological range (1 mM - 10 mM) of glucose with an average standard deviation of +/- 5%. SERS-gNPS was applied for continuous glucose measurement in an in -vitro 3D-CCTS and 3D Mono -cultured fibroblast Spheroids (3D -MS). The results revealed that the 3D-CCTS have more prominent glucose gradient (-1 mM) as compared to 3D -MS (-0.4 mM) from core to peripheral regions at 3 days to 5 days of culture. SERS-gNPS provides a platform for accurate and real-time measurement of glucose, and offers an enhanced understanding of glucose distribution in the tumour spheroids.

Physisorption is the general dominating adsorption mode on pure carbonaceous materials. In this study, we examined if chemisorbed hydrogen can be accomplished without incorporating metals on carbon surfaces. We constructed two 100-atom amorphous carbon (a-C) models with different microenvironments (42 % 2-fold, 52 % 3-fold, and 6 % 4-fold coordinated carbon atoms and 18 % 2-fold and 82 % 3-fold, respectively) using ab initio molecular dynamics (AIMD) simulations to mimic activated carbon samples in reality. The structural, energetic, electronic structure and kinetic properties of hydrogen adsorption, migration, and desorption on the a-C surfaces were analyzed using density functional theory (DFT) and AIMD calculations, which showed that: (1) hydrogen molecules could spontaneously dissociate on the convex side of 2-fold coordinated carbon atoms. Physisorption of hydrogen molecules mainly took place on 3-fold coordinated carbon atoms. (2) Regarding the multiple hydrogen adsorption, the entire a-C model could uptake around 156 hydrogen atoms with the adsorption energy per hydrogen atom about-0.70 to-0.60 eV at high hydrogen pressure. In addition, we found that the average individual adsorption energy of a single hydrogen atom on different carbon atoms could be used as a quick prediction for the saturated hydrogen adsorption energy. (3) The migration barrier of a hydrogen atom between two adjacent carbon atoms was about 2 eV at both low and high hydrogen coverage, indicating that hydrogen mobility on this designed a-C model was low. The desorption of a hydrogen molecule can take place at 300 K using AIMD calculation, but not for atomic hydrogen desorption. In this work, we demonstrated that hydrogen chemisorption could take place on this designed a-C without the decoration of metal particles to enhance hydrogen adsorption amounts. The materials design to balance hydrogen chemisorption, migration, and desorption shall be considered in the future.