Cr coatings on Zr alloys suffer from Zr–Cr interdiffusion-induced Kirkendall voiding, blistering, and eutectic-related degradation under high-temperature steam oxidation. Herein, a Mo diffusion barrier was introduced using multi-arc ion plating to retard interfacial diffusion and improve oxidation resistance. The oxidation kinetics, cross-sectional microstructure, interfacial phase evolution and diffusion behavior in the Cr–Mo–Zr system were investigated during steam oxidation. The Mo layer does not change the intrinsic oxidation mechanism of Cr coating because of both coatings exhibiting similar activation energies (∼220 kJ/mol). Compared to the Cr coating, the parabolic oxidation stage of Cr/Mo coating is prolonged, the parabolic rate constant (kp) is halved in 1200–1400℃, and the effective protection time is extended from 60 min to 180 min at 1200℃. Microstructural and phase analyses reveal that the Mo layer progressively evolves into a continuous Zr(Cr, Mo)2 interlayer, which acts as a secondary diffusion barrier. The Cr diffusion coefficient in Zr(Cr, Mo)2 is one order of magnitude lower than that in ZrCr2, thereby suppressing Zr-Cr interdiffusion and Kirkendall void formation, and further inhibiting blistering and eutectic-like microstructure, which are the primary degradation mechanism of Cr coating. The eventual degradation of the Cr/Mo coating is governed by progressive reduction by outward-diffusing Zr and degraded oxide layer.

Abstract Deep learning can extract quantitative measurements from microscopy images that are inaccessible to classical analysis, but developing these models requires machine learning expertise that most imaging scientists do not have. Here we present a framework in which a researcher describes their microscopy problem to a large language model (LLM) agent in under ten minutes of conversation—specifying what they image, what they want to measure, and what success looks like—and the agent autonomously handles the rest: designing physics-based training data, implementing a neural network, training, diagnosing failures, and iterating without human intervention. A researcher can start the agent before leaving the lab; overnight, it tests tens to a hundred model variations, each one an experiment that would otherwise demand active attention. We validate the framework across six microscopy modalities and four problem types. On the BBBC039 nuclear segmentation benchmark, the agent autonomously trains a U-Net with 3-class semantic segmentation and morphological post-processing, achieving pixel-level Dice of 0.97 and object-level F1 of 0.84—within 7% of the published baseline—while diagnosing a data pipeline bug that no amount of hyperparameter tuning could resolve. On single-protein holographic microscopy, the agent reads a published paper, designs a simulator, and develops an optimized model in a single session. On PatchCamelyon histopathology classification, the agent autonomously evolves through four optimization phases—from scratch training through transfer learning and regularization to inference-time ensembling—completing 97 iterations on 262,144 images to reach 89.3% test accuracy and 96.3% AUC, nearly matching the published rotation-equivariant baseline. This framework enables microscopy researchers to use deep learning-based image analysis without machine learning domain knowledge.

Detecting and characterizing aggregation of therapeutic monoclonal antibodies is critical for quality assessment, as aggregation can reduce therapeutic effectiveness and increase immunogenicity. Established methods characterize aggregation by size, providing only ensemble-averaged measurements and lacking single-molecule resolution. Here, we demonstrate the application of a label-free single protein oscillator method to simultaneously measure the size and charge of therapeutic monoclonal antibodies, including adalimumab, bevacizumab, and panitumumab, and differentiate different aggregation levels of UV-stressed adalimumab. We tethered single proteins to a sensor surface via a flexible polymer, drove them into oscillation by an alternating electric field, and imaged the process through near-field optical imaging. The results align with those from size exclusion high-performance liquid chromatography (SEC-HPLC) and imaged capillary isoelectric focusing (icIEF) methods, which are commonly employed during biopharmaceutical development. Our approach detects a broader range of molecular sizes beyond the upper limits of SEC-HPLC. Additionally, simultaneously measuring size and charge at single protein resolution enables two-dimensional mapping of the charge/size distribution of aggregates. Our results reveal heterogeneous charge distribution among adalimumab aggregates with similar size, indicating structural changes in monomers and conformers that are not readily accessible by existing methods. The method offers an integrated approach for evaluating the size and charge characteristics of therapeutic antibodies at single-molecule resolution, further supporting therapeutic optimization.

W coatings with a preferential (110) orientation are critical for thermionic emitters due to their high work function. However, the growth structure and orientation evolution of W coatings prepared by arc ion plating remain unclear. In this work, the effects of ion bombardment regulated by deposition parameters on the microstructure and crystallographic orientation of W coatings were systematically investigated through experimental and theoretical analysis. A pronounced structural transition was observed with decreasing ion bombardment intensity, from (110)-oriented coarse equiaxed grains under single-sided deposition to (211)-oriented fine columnar grains under double-sided deposition. The enhancement of (110) orientation at higher bias voltages and arc currents was driven by surface energy minimization, whereas the formation of (211) orientation under weakened bombardment was governed by strain energy minimization and kinetic growth processes. Additionally, the self-sputtering behavior of W coatings exhibited a strong dependency on bias voltage and deposition mode, with the relative sputtering yield increasing with bias voltage but decreasing by ∼50% in double-sided mode. The variations in work function can be rationalized by the combined influence of crystallographic orientation and surface roughness. These findings elucidate the intrinsic correlations between deposition parameters, microstructural evolution and the resulting work function of W coatings.

The Cr2N composite coating was fabricated on 316H stainless steel using an innovative combination of plasma chromizing and gas nitriding (PLC-GN). A comprehensive comparative analysis was conducted against the coating produced by the conventional pack chromizing and gas nitriding (PAC-GN) method. The study evaluated the differences in microstructure, phase evolution, surface hardness, and tribological properties. The results demonstrate that the PAC-GN process (1090 degrees C/20h chromizing + 1070 degrees C/5h nitriding) produced a similar to 100 mu m thick coating comprising Cr2N, CrN, alpha-Fe, and gamma-Fe. In contrast, the PLC-GN process (1050 degrees C/5h chromizing + 1070 degrees C/5h nitriding) formed a similar to 80 mu m thick coating consisting of Cr2N, alpha-Fe, and gamma-Fe, with the notable absence of the CrN phase. This simplified phase structure suggests enhanced stability for high-temperature liquid sodium environments. Although the surface hardness of the PLC-GN coating (796.7 HV) was slightly lower than that of the PAC-GN coating (923.3 HV), its wear rate was on the same order of magnitude while maintaining a stable and lower coefficient of friction. More importantly, the PLC-GN process achieved this performance while significantly reducing the chromizing temperature by 40 degrees C and shortening the processing time by 80 %, alongside the complete elimination of dust pollution. This study confirms that PLC-GN is a highly efficient, environmentally friendly, and industrially viable alternative for fabricating high-performance Cr2N coatings, showing great potential for applications in advanced nuclear reactors.

Chimeric antigen receptor (CAR) T-cell therapy is an effective treatment for hematologic malignancies. However, it is limited by high costs, risk of severe toxicities such as cytokine release syndrome and neurotoxicity, and heterogeneous patient responses. The current therapy monitoring depends largely on subjective symptom assessment, routine laboratory tests, and basic vital signs, without real-time, quantitative evaluation of CAR T-cell expansion or activation in clinical practice. This lack of timely immune monitoring hampers individualized care and contributes to increased treatment costs. To address this need, we present a proof-of-concept, label-free rapid optical imaging (ROI) biosensor with automated machine learning analysis for direct quantification of CAR T-cells from whole blood. This microfluidic platform integrates red blood cell (RBC) removal, CAR T-cell capture, and imaging-based quantification on a single chip, eliminating the need for centrifugation, staining, and operator-dependent interpretation. For validation, 50 μL whole blood samples spiked with Jurkat cells expressing CD19 CARs underwent RBC depletion by agglutination and microfiltration. The remaining blood components were then incubated on a sensor chip functionalized with recombinant CD19 protein. Captured CAR T-cells were imaged by brightfield microscopy and automatically enumerated using a machine learning algorithm trained on fluorescence-validated cells. The CD-19 cells' capture performance was validated by flow cytometry and fluorescence imaging. The trained machine learning model validated at 88% sensitivity and 96% specificity. Buffer and whole blood calibration curves were established across clinically relevant concentrations (1-1000 cells/µL) with triple replicates. The results showed high correlation (0.975 and 0.990 R2) between the spiked concentration and the detected CAR T-cells, with a 95% certainty limit of detection (LOD) and quantification (LOQ) of 0.6 and 1.1 cells/µL for spiked buffer, and 14 and 67 cells/µL for spiked whole-blood, respectively.

A Mo diffusion barrier was introduced into Cr-coated Zr-4 alloy by multi-arc ion plating to clarify the role of oxidation-induced interlayer evolution in suppressing interdiffusion-driven coating failure during 1200°C steam oxidation. Compared with the single Cr coating, the Cr/Mo coating reduced the parabolic rate constant from 0.080 to 0.043 (mg·cm-2)2·min-1 and delayed the onset of accelerated oxidation from ~60 to ~180 min. Microstructural and phase analyses reveal that the initially deposited Mo layer progressively evolves into a continuous Zr(Cr, Mo)2 interlayer, which acts as a secondary diffusion barrier after partial Mo consumption. The Cr diffusion coefficient in Zr(Cr, Mo)2 is one order of magnitude lower than that in ZrCr2, thereby suppressing Zr-Cr interdiffusion and Kirkendall void formation, and further inhibiting blistering, which is the primary degradation mechanism of Cr coating. The eventual degradation of the Cr/Mo coating is governed by micropore coalescence within the Cr2O3 scale and its progressive reduction by outward-diffusing Zr. These findings identify the oxidation-induced Zr(Cr, Mo)2 interlayer as the key barrier phase responsible for delaying interdiffusion-driven failure of Cr-coated Zr alloys.

The oxidation behavior and microstructural evolution of the Cr-coated Zr-Sn-Nb alloy in high-temperature steam were investigated through oxidation tests at 1000-1200 degrees C. The mass change per unit area of the oxidized specimens was calculated using the mass gain method, and the microstructure was characterized using X-ray diffractometry and scanning electron microscopy. The results demonstrated that the oxidation weight gain of the Cr-coated Zr-Sn-Nb alloy was approximately 50% lower than that of the uncoated alloy. The oxidation kinetics of the coated alloy followed the parabolic law. Following steam oxidation at 1000-1150 degrees C, the Cr-coated Zr-Sn-Nb alloy exhibited a layered structure of Cr2O3 layer, residual Cr layer, Cr-Zr diffusion layer, and the beta-Zr substrate. At 1200 degrees C, the protective coating failed, and the microstructure was composed of Cr2O3, ZrO2, alpha-Zr(O), and beta-Zr.

Endothelial cell (EC) activation, characterized by upregulation of adhesion molecules that drive monocyte recruitment, contributes to plaque progression while also providing an opportunity for targeted therapeutic delivery. Leveraging the cell membrane cloaking strategy, we recently developed a monocyte-mimetic nanoparticle (MoNP) platform that exploits the natural inflammatory tropism of monocytes for site-specific delivery to atherosclerotic vessels. Recognizing that integrin activation is a key determinant of monocyte adhesion to ECs, this study investigates whether pre-activating integrins on MoNP enhances their binding affinity and accumulation at atherosclerotic lesions. Mouse bone marrow-derived monocytes were pretreated with CCL2 or Mn2⁺ to activate membrane integrins. Isolated monocyte plasma membranes were cloaked onto fluorescently labeled polymeric cores to generate integrin-activated MoNPs (IA@MoNPs). The targeting capability of IA@MoNPs toward endothelial ligands, inflamed ECs, and atherosclerotic lesions was evaluated using in vitro and in vivo models. IA@MoNPs exhibited markedly enhanced binding to VCAM1, the primary endothelial ligand mediating integrin-dependent monocyte adhesion, and significantly increased uptake by ECs under atheroprone conditions compared to standard MoNPs. In vivo, IA@MoNPs demonstrated enhanced accumulation in atherosclerotic arteries without increasing nonspecific binding, and blocking β1-integrins on IA@MoNPs abolished this targeting effect. Importantly, integrin activation on IA@MoNPs did not compromise circulatory stability or induce immune or organ toxicity. Integrin activation represents a simple yet effective strategy to enhance MoNP targeting to inflamed ECs and atherosclerotic lesions. This mechanism-driven approach improves targeting performance while maintaining specificity and safety, advancing the translational potential of the biomimetic nanomedicine platform for atherosclerosis.

W coatings deposited by DCMS, MFMS and HiPIMS under identical discharge power were systematically compared in terms of microstructure, texture and mechanical properties. Waveform analysis revealed that HiPIMS mode generated the highest peak power density of 0.24 kW cm−2, far exceeding those of the conventional sputtering modes. The intensified plasma discharge in HiPIMS produced a denser and smoother W coating with a refined columnar microstructure, although accompanied by a reduced deposition rate. In contrast to the dominant (211) texture observed in the DC and MF coatings, the HiPIMS coating exhibited a pronounced (110) preferred orientation with a texture coefficient as high as 3.87, resulting in the highest work function. Furthermore, the HiPIMS mode significantly improved the hardness, H/E and H3/E2 values of the W coating. Such enhancement was mainly attributed to grain refinement, enhanced compressive residual stress, and ion-bombardment-induced lattice defects. The established sputtering mode-microstructure-performance correlation provides valuable insight into the controlled growth of thick W coatings.

Cr-O coatings were successfully prepared by high-power arc ion plating under varying oxygen flow rates, and their microstructure, mechanical properties and oxidation behavior were systematically investigated. With increasing oxygen flow rate, the coatings underwent a distinct microstructural transformation, from a Crdominant columnar structure to a Cr/Cr2O3 nano-equiaxed structure, and ultimately to a Cr2O3 nano-columnar structure. This evolution was accompanied by an enhancement in coating hardness due to grain refinement and the formation of hard oxide phase. The Cr/Cr2O3 nanocomposite coating achieved a desirable combination of hardness and crack resistance. However, it exhibited an inferior oxidation resistance at 900 degrees C due to the formation of porous oxide scale. In contrast, Cr-dominant coating showed oxidation behavior similar to pure Cr coating, while Cr2O3 coating displayed excellent oxidation resistance and interfacial thermal stability.

Label-free plasmonic cell force microscopy is developed to reveal cell exerted force at diffraction-limited spatial resolution. By quantifying cell-substrate interaction dynamics in real-time through plasmonic scattering imaging, the spatial and temporal evolutions of cellular forces are accurately mapped. To demonstrate the capability of the technology, cardiomyocyte force evolution and loading rates are measured with millisecond resolution. Furthermore, cell force responses to nicotinic receptor activation are monitored and observed heterogenic cell force changes among a population of cells, underscoring the versatility and potential impact of this label-free approach.

Ketones, key biomarkers of fat oxidation, are clinically relevant for metabolic health maintenance and disease development, making continuous monitoring crucial. Here, we present a novel colorimetric sensor for non-invasive, continuous acetone detection in breath and skin for point-of-care applications. The sensor comprises a polydimethylsiloxane (PDMS) shell encapsulating a highly sensitive and specific liquid-core acetone-sensing probe. Microsphere sensors were characterized by analyzing their size, PDMS shell thickness, colorimetric response, and sensitivity under realistic conditions (100% relative humidity and CO₂ interference). The microsphere size and sensor sensitivity can be controlled by modifying the fabrication parameters. Critically, the sensor showed high selectivity for acetone detection, with negligible interference from CO₂ concentrations up to 4%. Furthermore, the sensor enabled real-time, continuous, non-invasive monitoring. In addition, the sensor displayed excellent reproducibility (CV < 5%) and stability under realistic storage conditions (over two weeks at 4°C). Finally, the accuracy of the microsphere sensor was validated against a gold standard gas chromatography-mass spectrometry (GC-MS) method using simulated and real breath samples from Type 1 diabetic patients. The correlation between the microsphere sensor and GC-MS rendered a linear fit of slope equal to 0.94 and an R-squared adjustment of 0.9527. Thus, the liquid-core microsphere-based sensor offers a promising platform for continuous, non-invasive, and cost-effective acetone monitoring, potentially revolutionizing point-of-care diagnostics for metabolic disorders and health management.

Spatially resolved sensing is a burgeoning area of electrochemistry that, in contrast to traditional electrochemical techniques, allows for the analysis of heterogeneous systems such as neurotransmitter release from cells. Of these techniques, optical microscopy methods are valued for real-time high throughput sensing. However, improving the sensitivity of many optical techniques remains a challenge. Here, we modify the gold (Au) electrode of the standard plasmonic electrochemical microscopy (PEM) setup with a mesoporous silica film (MSF) to achieve sensitive imaging of the electroactive species. Sensitivity enhancement occurs via species nanoconfinement from the attraction of ions to the negatively charged silica films, thereby increasing the local concentration change and magnifying the PEM signal. The performance of Au-MSF electrodes in the PEM setup was investigated using 1,1'-ferrocenedimethanol, whose oxidized form carries a positive charge. Results revealed enhancement of the sensing signal, with up to 37-fold improvement in the detection limit and up to 23 times improvement in the sensitivity. Importantly, Au-MSF electrodes allowed for the quantification of detected concentrations, in contrast to Au electrodes, for which R2 values were unacceptably low. Furthermore, Au-MSF electrodes also showed increased sensitivity for dopamine detection compared to Au electrodes and were able to visualize localized dopamine release, showing this setup's great promise for biological applications, such as real-time imaging of the neurotransmitter release.

A high-density nano-oscillator platform using self-assembled DNA-barcoded virion sensors is developed to address the critical need for high-throughput label-free measurement of small-molecule binding to membrane proteins. By integrating virion display technology with charge-sensitive plasmonic detection, our platform enables robust, label-free quantification of small-molecule binding kinetics to membrane proteins. Gold nanoparticle-virion conjugates are self-assembled onto a plasmonic sensor chip via a flexible molecular linker to form high-density nano-oscillators. Driven by an alternating electric field, the oscillation amplitudes of the nano-oscillators are precisely measured via widefield plasmonic imaging. This charge-sensitive mechanism can sensitively detect the binding of small-molecule ligands to the membrane proteins displayed on the virions at single-nanosensor resolution, overcoming the sensitivity limit of conventional mass-sensitive techniques. More importantly, the platform employs novel affinity-discriminated DNA barcodes for multistate decoding with exponential multiplexing capacity, enabling high-throughput screening of a library of membrane proteins. For a proof-of-concept demonstration, binding kinetics of five pairs of G-protein-coupled receptors and their corresponding small molecule ligands are measured on a single sensor chip, with all individual nano-oscillators identified by just two affinity-discriminated, quadra-state DNA decoders. This technology advances membrane protein research and drug screening capabilities, offering a practical solution for biomolecular interaction studies and biosensing applications.

Spaser nanoprobes, with their ultranarrow emission line widths and nanoscale sizes, are emerging as leading contenders for the next generation of biological luminescent probes. However, modulating spasing wavelengths across a broad spectral range remains a formidable challenge that constrains their application in multiplexed sensing and imaging. Here, we introduced a novel wavelength-tunable spaser system, successfully creating a nanoprobe family with 9 distinct spasing wavelengths. These nanoprobes exhibit ultranarrow emission line widths of 3-8 nm with a low pump threshold of 0.5 mJ cm-2. We further investigated the influences of energy matching between the gain medium and the cavity on the emission performance of the spaser nanoprobes and successfully realized narrow-band luminescence in both individual and aggregated spaser particles under commercial confocal instruments. Moreover, we demonstrated the multiplexed imaging capability of 6 distinct spaser nanoprobes in a single HeLa cell within an imaging window of 200 nm, free of spectral crosstalk. Our results address the long-standing issue of wavelength tunability in spaser nanoprobes and also clarify previous disputes on the luminescence origin of single nanoparticles, marking a significant advancement in their application for ultramultiplexed biological imaging and sensing.

Cellular systems achieve precise biomolecular recognition through dynamic regulation of molecular conformation and spatial arrangement, a complexity that is difficult to replicate in vitro, limiting advancements in biosensing technologies. The nanoscale programmability of tetrahedral DNA frameworks (TDFs) offers a compelling solution, enabling precise control over the spatial arrangement and conformation of nucleic acid targets, making TDFs highly effective for biosensor interface engineering. In this study, we developed dimeric TDF capture probes with tunable interprobe distances (25-45 nm), allowing for the precise stretching and ultrafast detection of single-stranded DNA (ssDNA) targets. By integrating auxiliary probes to modulate local target conformation, hybridization efficiency was significantly enhanced, yielding a 2.9-fold improvement in signal intensity. This approach was successfully applied to single-nucleotide polymorphism (SNP) detection, demonstrating a 2-fold improvement in discrimination sensitivity. Furthermore, integration with a microarray fluorescence chip enabled rapid and accurate quantification of IDH1 mutant allele frequency (MAF), highlighting its potential for glioma classification, disease monitoring, and therapeutic evaluation. These findings underscore the transformative potential of TDF-based interface engineering as a platform for high-performance biosensing and diagnostic applications.

Multidrug-resistant (MDR) bacterial infections present a great challenge to healthcare, particularly during life-threatening conditions, such as septic shock. In these critical cases, testing delays are life-threatening, as patient mortality increases by 7.6% for each hour without effective antibiotics. To overcome these issues, we have developed a multichannel Large-Volume Scattering imaging (LVSim) system for rapid and precise phenotypic antimicrobial susceptibility testing (AST) and minimum inhibitory concentration (MIC) determination. This system simultaneously monitors up to eight sample/drug conditions over time. A Bayesian Gaussian process model is employed to analyze the temporal dynamics of bacterial growth for rapid MIC determination. We validated our method with Escherichia coli, a model bacterial strain, and two Pseudomonas aeruginosa strains─one model reference strain and one slow-growing MDR clinical isolate─and achieved rapid MIC determination within 2 h. The multiplexed LVSim system offers a promising rapid AST solution to combat MDR infections.

Ketones, which are key biomarkers of fat oxidation, are relevant for metabolic health maintenance and disease development, making continuous monitoring essential. In this study, we introduce a novel colorimetric sensor designed for potential continuous acetone detection in biological fluids. The sensor features a polydimethylsiloxane (PDMS) shell that encapsulates a sensitive and specific liquid-core acetone-sensing probe. The microsphere sensors were characterized by evaluating their size, PDMS shell thickness, colorimetric response, and sensitivity under realistic conditions, including 100% relative humidity (RH) and CO2 interference. The microsphere size and sensor sensitivity can be controlled by modifying the fabrication parameters. Critically, the sensor showed high selectivity for acetone detection, with negligible interference from CO2 concentrations up to 4%. In addition, the sensor displayed good reproducibility (CV < 5%) and stability under realistic storage conditions (over two weeks at 4 °C). Finally, the accuracy of the microsphere sensor was validated against a gold standard gas chromatography-mass spectrometry (GC-MS) method using simulated and real breath samples from healthy individuals and type 1 diabetes patients. The correlation between the microsphere sensor and GC-MS produced a linear fit with a slope of 0.948 and an adjusted R-squared value of 0.954. Therefore, the liquid-core microsphere-based sensor is a promising platform for acetone body fluid analysis.