Monolithic integration of nanorod arrays on microfluidic chips for fast and sensitive one-step immunoassays

Fluorescent immunoassays are widely used for quantitative detection of protein biomarkers but typically rely on 96-well formats with complex workflows, long incubations, and bulky instruments, limiting point-of-care applicability. Microfluidic lab-on-a-chip technologies have addressed speed, cost, and operational simplicity, yet sensitivity remains a core bottleneck. Metal-enhanced fluorescence (MEF) using plasmonic nanostructures can dramatically improve detection limits, but many nanofabrication approaches (e-beam lithography, nanoimprinting, colloidal lithography, bottom-up synthesis) are costly, low-throughput, or cumbersome. The authors aim to develop an easily fabricated plasmonic substrate—Au nanorod arrays produced by one-step oblique angle deposition (OAD)—monolithically integrated into a microfluidic chip to achieve fast, one-step, flow-through immunoassays with ultrahigh sensitivity, demonstrated for cardiac troponin I (cTnI) and extended to multiplex cancer biomarkers.

Prior microfluidic immunoassay advances include plastic chips enabling one-step assays detecting C-reactive protein <1 ng mL^-1 in 5 min, and volumetric/digital bar-chart chips achieving ultralow limits (~0.1 pM) with simple readouts. Other miniaturized platforms include lab-on-paper, lab-on-disc, lab-on-a-syringe, and lab-on-a-stick. Sensitivity enhancement via MEF has used plasmonic Au/Au films, nanoparticles, nanopillars, nanorods, and nanowells. However, conventional nanofabrication methods (e-beam lithography, nanoimprinting with mold preparation and etching, colloidal lithography with uniformity and etching challenges, solution syntheses with complex operations) limit manufacturability. Reported MEF substrates can reach fM-level detection (e.g., ~5 fM with nanostructured gold films; 0.3 fM with nanoimprinted nanorod arrays), but simpler, scalable fabrication routes are needed. The presented OAD-fabricated nanorod arrays address these gaps by eliminating advanced lithography and enabling monolithic integration in microfluidic devices.

Design and simulation: COMSOL Multiphysics simulations modeled localized surface plasmon resonance (LSPR) and near-field enhancements for Au nanorod arrays under polarized illumination. Arrays with diameters 25–100 nm and pitches 200–400 nm were evaluated. For 75 nm diameter and 200 nm pitch (matching fabricated structures), strong electric field enhancement localized within ~10 nm around nanorods was observed. Increasing rod diameter or pitch induced redshifts in resonance. Experimental absorbance of 75 nm-diameter arrays showed a resonance peak at 495 nm (FWHM 120 nm), aligning with Alexa 488 (ex/em 493/519 nm). Nanorod fabrication (OAD): Dense Au nanorod arrays were deposited on specific regions of glass substrates using a shadow mask and oblique angle deposition, avoiding advanced lithography. Metal atoms incident at a large angle relative to the substrate normal nucleate and grow into nanorods via shadowing. SEM characterized morphologies for different nominal Au thicknesses: 1 µm (tilt ~15°, diameter 59.9±9.2 nm, length 376.0±52.4 nm), 1.5 µm (77.1±8.0 nm, 608.8±10.2 nm), and 2 µm (103.6±19.6 nm, 682.7±38.1 nm). OAD-produced rods exhibit irregularities (protrusions, bifurcations, fusion) that can create additional electromagnetic hot spots. Surface functionalization and static sandwich assay: Nanorod arrays were coated with a self-assembled monolayer of 3,3'-dithiodipropionic acid di(N-hydroxysuccinimide ester) (DSP) (~1.5 nm, n≈1.40) to covalently immobilize capture antibodies (cAb; monoclonal mouse anti-cTnI). After washing and blocking, targets (human cTnI) were incubated (2 h) at serial concentrations. Detection used Alexa 488-labeled goat anti-cTnI (detection antibody, dAb) to form sandwich complexes. Estimated spacer thickness (DSP + cAb + antigen) was ~9.36 nm, within the strong near-field enhancement zone. Microfluidic flow-through device: A monolithically integrated microfluidic chip incorporated: (i) a dAb deposition zone where Alexa 488-labeled dAb was prepatterned by pipetting 20 µL into stencils and incubating 2 h at room temperature; (ii) a serpentine micromixer for rapid dAb–antigen complex formation; (iii) a reaction chamber with Au nanorod arrays bearing immobilized cAb via DSP; and (iv) a capillary pump made of parallel microchannels controlling capillary pressure and flow. Human serum spiked with targets was introduced through the inlet; prepatterned dAb dissolved into the flow to bind targets, complexes were captured on the nanorods, and fluorescence was read via a microscope with CCD. Average flow rate was ~0.6 nL s^-1 with total filling time ~6 min. The capillary pump geometry sets flow and sample volume; washing steps typical of 96-well assays were eliminated by continuous flow-through. Controls and references: Identical assays on poly-L-lysine-modified glass served as references for performance comparison across cTnI, PSA, and CEA. Dose–response curves were fit with five-parameter logistic regression; LoD defined as concentration yielding signal equal to mean blank + 3×SD under Gaussian assumptions. Interassay CV% computed from five independent samples.

- Au nanorod arrays (optimized ~75 nm diameter, 200 nm pitch) exhibit strong near-field enhancement localized within ~10 nm; absorbance peak measured at 495 nm (FWHM 120 nm), matching Alexa 488. - Static (batch) cTnI sandwich assay on Au nanorods achieved LoD 33.9 fg mL^-1 (~1.4 fM) with >6×10^5-fold sensitivity improvement versus glass (LoD 22.9 ng mL^-1, ~1.0 nM); dynamic range ~2.3 pg mL^-1 to 1 µg mL^-1 (>6 orders of magnitude). R^2 of logistic fit = 0.998; interassay CV% = 11.4%. - Flow-through microfluidic assay (serum): cTnI LoD 6.9 pg mL^-1 (~0.29 pM) within <6 min at ~0.6 nL s^-1 flow, a ~4000× improvement versus conventional glass devices (27.0 ng mL^-1, ~1.1 nM) while eliminating washing steps. - Multiplex capability: Parallel microfluidic paths enabled simultaneous detection of PSA and CEA with enhanced signal-to-background ratios. • PSA LoD: 36.6 fg mL^-1 (~1.2 fM) on nanorods vs 29.0 ng mL^-1 on glass; dynamic range 1.6 pg mL^-1–1 µg mL^-1; CV% 13.2%. • CEA LoD: 194.3 fg mL^-1 (~2.7 fM) on nanorods vs 98.9 ng mL^-1 on glass; dynamic range 6.2 pg mL^-1–1 µg mL^-1; CV% 10.8%. - OAD provides simple, lithography-free, and potentially scalable fabrication; nanorod surface irregularities likely generate additional hot spots enhancing MEF.

The study addresses the need for highly sensitive yet simple and fast immunoassays by leveraging MEF from plasmonic Au nanorod arrays fabricated via OAD and integrated into a microfluidic architecture. Simulations and spectral measurements guided geometric optimization (75 nm diameter, 200 nm pitch) to align LSPR with Alexa 488, ensuring strong field–fluorophore coupling within ~10 nm. The dense nanorod arrays increase effective surface area for antibody immobilization and amplify fluorescence via enhanced local fields and radiative rates, collectively driving sub-femtomolar LoDs in static assays and sub-picomolar LoDs in rapid flow-through assays. The microfluidic integration enables one-step processing without washing, reducing assay time to minutes and sample consumption while maintaining high analytical performance. Multiplexed layouts further demonstrate scalability for simultaneous biomarker quantification with improved signal-to-background ratios. Compared to conventional glass substrates and more complex nanofabrication approaches, the OAD-integrated device provides substantial sensitivity gains with simplified, mass-manufacturable processing, advancing the practicality of point-of-care diagnostics.

The work demonstrates a monolithically integrated microfluidic immunoassay platform using OAD-fabricated Au nanorod arrays that delivers ultrahigh sensitivity and rapid analysis. Key contributions include: (i) lithography-free fabrication of dense Au nanorod arrays optimized for MEF; (ii) sub-femtomolar LoD for cTnI in static assays and 6.9 pg mL^-1 LoD in <6 min for flow-through assays; and (iii) multiplexed detection of PSA and CEA with femtogram-per-milliliter sensitivity. The simplicity and scalability of OAD, combined with microfluidic one-step operation, position this approach for translation toward on-site point-of-care diagnostics. Future work could expand to other fluorophores by tailoring nanorod geometry, broaden biomarker panels, integrate compact optical readouts, and further optimize flow and surface chemistry for higher throughput and robustness in complex clinical matrices.

- Reported flow-through LoD for cTnI (6.9 pg mL^-1) is higher than the static assay LoD (33.9 fg mL^-1), reflecting a trade-off between speed/throughput and ultimate sensitivity. - Performance depends on maintaining optimal fluorophore–metal spacing (~10 nm) and spectral alignment between LSPR and dye; deviations may reduce enhancement. - Detection currently relies on fluorescence microscopy with a CCD, which may limit fully instrument-free deployment without additional miniaturized readout integration.