Visualizable detection of nanoscale objects using anti-symmetric excitation and non-resonance amplification

The study seeks to enable direct, visualizable detection of nanoscale objects with conventional, diffraction-limited optical microscopes. Nanoscale objects perturb host media and include fixed-form defects in semiconductors and free-form particles such as viruses and nanoparticles. Detecting and visualizing these perturbations is vital for applications spanning integrated photonics, quantum chips, disease diagnosis, and semiconductor manufacturing. Conventional optical microscopy struggles due to both the diffraction limit and weak Rayleigh scattering (far-field signal scaling steeply with particle size), leading to poor SNR. While advanced techniques like interferometric scattering microscopy (iSCAT), interferometric cross-polarization microscopy (ICPM), photothermal imaging, and electron microscopy offer sensitivity or resolution, each has constraints such as labeling, speed, field-of-view, or complexity. This work proposes anti-symmetric excitation to create an electromagnetic canyon (a background field null) near nanostructure sensors, combined with non-resonant amplification, to visualize sub-diffraction nanoscale objects, including fixed-form defects, using a simple two-beam illumination add-on and nanowire assemblies.

Prior approaches to nanoscale visualization include electron microscopy (high resolution but destructive, slow, costly, and small FOV) and optical super-resolution methods (e.g., STED, STORM/PALM), which often require fluorescence labeling and may be slow or complex. Photothermal imaging achieves label-free single-particle imaging but depends on resonant transitions being pumped. Interferometric methods such as iSCAT enhance weak scattering by interfering it with a strong reference, improving scaling and sensitivity, and ICPM uses cross-polarized paths to suppress background. Digital microarrays with plasmonic nanorods extend sensitivity and dynamic range for single-molecule detection. However, these methods may not be ideal for visualizable detection of fixed-form nanoscale objects embedded in nanostructured backgrounds, nor do they necessarily provide widefield, simple, label-free operation. Structured illumination microscopy (SIM) yields up to ~2× resolution enhancement via moiré patterns, distinct from the anti-symmetric excitation used here, which relies on positioning relative to a standing-wave intensity null to cancel background fields and works irrespective of gap size between nanostructure pairs.

Concept and theory: The framework creates an electromagnetic canyon (EC), a region of near-zero background field, by driving a pair of identical nanostructures (e.g., parallel nanowires) into an anti-symmetric excitation state so that their background fields cancel along the symmetry plane (SP). Non-resonance amplification arises from coupling between the object and the nanowire modes, boosting the object's signal without requiring resonance, and is broadband. An eigenmode expansion with coupled-mode perturbation describes the total electric field near the SP as a sum of contributions from the nanowires, object-nanowire coupling, and the object itself. Under conventional symmetric excitation, background dominates; under anti-symmetric excitation, the background cancels on the SP, leaving object-related terms that can greatly exceed camera noise. If an object lies exactly on the SP, coupling terms cancel and only the object's own (weak) scattering remains; otherwise, coupling yields amplification. The EC exists regardless of nanowire gap by symmetry. Dipolar approximations and two-dipole interference analyses further validate EC properties (Supplementary Notes). Anti-symmetric excitation via two-beam interference: Two in-phase, y-polarized plane waves impinging at equal oblique angles (±θ) along ±x produce a standing wave with field E = E0 cos(2πx/Λ) ŷ, where Λ = λ/sinθ (intensity period Λ/2) and a π phase jump across each intensity null. Positioning nanowire centers symmetrically about an intensity null at x = (m + 1/2)Λ/2 ± p (m integer, 2p is center spacing) excites them out of phase (Esp1 = −Esp2), generating a local EC near the SP. Scanning the sample laterally aligns nanowire pairs with standing-wave nulls across a large area, enabling many EC sites. Simulations: Near-field electromagnetic response was computed using FDTD for sources represented by electric dipoles or localized plane-wave excitation. Far-field projection decomposed fields into plane waves; those within the objective NA were propagated through the imaging system via the chirped z-transform (validated against an equivalent magnetic-dipole method). Localized illumination minimized finite-domain and substrate reflection artifacts. Simulations compared four modalities—proposed framework, darkfield, brightfield, and iSCAT—for object-free and object-perturbed nanowire assemblies, and evaluated contrast versus object size and position for fixed-form and free-form objects, including double- and quad-nanowire arrays. Experimental setup: A 785 nm single-mode, single-frequency laser with integrated isolator was split 50:50 into two free-space beams, collimated by 25 mm lenses, and passed through linear polarizers. Fiber polarization controllers equalized transmitted powers. Beams were set to 60° incidence using precision rotation mounts to form the standing-wave excitation. Blocking one beam enabled conventional darkfield illumination. The sample was mounted on a rotation stage atop a motorized xyz stage; the sample was scanned in x (typically 100 nm steps) to align nanowire pairs with intensity nulls. Imaging used a Mitutoyo M Plan Apo NIR 20×, 0.4 NA objective, a 75 mm tube lens (7.5× magnification), and an Amscope MU1403B 4096×3286 pixel CMOS camera (IR filter removed). A 780 nm bandpass filter reduced stray light; a white LED provided brightfield illumination for navigation. The system captured widefield images over 726 µm × 582 µm. Calibration and performance characterization: Beam overlap and interference were aligned by minimizing the overlap spot on a viewing card. The microscope path included no computational corrections; the low-cost camera exhibited dynamic range 65.3 dB and SNR 35.5 dB, yielding poor conventional performance: Abbe limit 981 nm (λ/2NA) but only ~4 µm gaps resolved in brightfield; isolated 390×120 nm features were not visible in darkfield. Sample fabrication: On 4-inch Si wafers with thermal oxide and LPCVD nitride, e-beam lithography (Raith eLine, 10 kV, PMMA resist, 400 µC/cm² dose) defined Au nanostructures (5 nm Ti/100 nm Au lift-off) including single, double, and quad nanowires with nominal gaps 0–1000 nm (50 nm increments). A 2 nm Au blanket improved conductivity for SEM. Line widths were biased to account for EBL beam size; actual gaps varied slightly due to proximity effects. Measurement protocol: Widefield images were acquired; regions of interest were cropped and compared with post hoc SEM for corroboration. For EC validation, double-nanowire images were recorded while scanning x, observing transitions between EC and non-EC states (central intensity minimum indicating EC). Comparative darkfield and brightfield images were recorded for the same areas. Diverse fixed-form defects (dots, corner objects, line expansions, bumps) and free-form particles were imaged under EC and non-EC conditions.

• Demonstrated direct, widefield visualization of nanoscale objects as small as a 25 nm radius (≈λ/31 at 785 nm) in the near field of nanowire sensors using a conventional microscope (0.4 NA objective) and a low-performance CMOS camera across a 726 µm × 582 µm field of view.
• Generated electromagnetic canyons (background-null regions) via anti-symmetric excitation using two-beam interference; EC formation observed repeatedly as samples were scanned, independent of nanowire gap size.
• Visualized individual nanowires and gaps well below Abbe’s limit (981 nm for the system): resolved double-nanowire gaps as small as 63 nm, and other gaps of 137, 150–188, and 750 nm, despite conventional brightfield/darkfield failing to resolve them on the same setup.
• Achieved highest image contrast for detecting fixed-form nanoscale objects vs. darkfield, brightfield, and initial implementations of iSCAT across object sizes (simulations). Darkfield had better contrast than brightfield/iSCAT but worse than the proposed method; iSCAT/ICPM can have stronger absolute signals for isolated free-form objects but are less suited to fixed-form defects embedded in nanostructures.
• Demonstrated visualization of diverse objects: isolated dot (~25 nm radius) between nanowires; asymmetric corner defects of different sizes (e.g., δ1 = 90 nm vs. δ2 = 50 nm) with intensity correlating to size; central line expansions (Δ ≈ 16–31 nm) with distinguishable brightness; complex quad-nanowire structures with tilted 63 nm gaps and central bumps causing observable asymmetry in EC images.
• Showed that free-form particles are visualizable at various positions near nanowires except when exactly on the symmetry plane, where coupling terms cancel (consistent with anti-symmetric EC theory).
• Validated broadband, non-resonant amplification: object-nanowire coupling enhances signals above camera noise without requiring resonance tuning or environmental stabilization.

The work addresses the core challenge of weak nanoscale scattering masked by background and diffraction by canceling the dominant background near nanostructure sensors through anti-symmetric excitation, thereby forming an electromagnetic canyon. Within the EC, object-induced coupling to nanowire modes non-resonantly amplifies the signal above noise, enabling direct visual discernment of nanoscale objects with a standard, low-NA widefield microscope. This approach maintains high contrast even when objects are embedded within nanoscale patterned backgrounds, where interferometric methods can struggle. The technique offers several practical advantages: (1) direct sub-diffraction visualization on conventional microscopes; (2) broadband operation without stringent environmental control due to non-resonant amplification; (3) straightforward nanowire fabrication; (4) compact sensor footprints scalable in arrays; and (5) large field-of-view imaging in a single shot. Although visualization requires objects to be near the nanowire sensors and fails if objects lie exactly on the symmetry plane, arrays of nanowire pairs can tile the field to provide coverage, and real object shapes reduce the probability of exact cancellation. The demonstrated capability to detect fixed-form defects at deep subwavelength scales suggests strong utility in semiconductor inspection, and the widefield, label-free, non-destructive operation makes the method attractive for biosensing (e.g., viruses, molecular aggregates) and microfluidic monitoring. Future optimization of sensor geometry/materials and imaging hardware can further increase sensitivity and robustness.

This paper introduces and experimentally validates a simple, add-on optical framework that enables visualizable detection of nanoscale objects by creating electromagnetic canyons via anti-symmetric excitation of nanowire pairs and leveraging non-resonant amplification. The method allows direct visualization of λ/31-scale objects and sub-100 nm gaps using a conventional, low-NA widefield microscope across a large field of view, outperforming conventional brightfield and darkfield and offering superior contrast for fixed-form defects relative to initial iSCAT implementations. The approach is broadband, fabrication-friendly, and scalable via arrays. Future work should focus on improving camera dynamic range/SNR, optimizing nanowire geometry and materials (including index matching to biological targets), functionalizing sensor surfaces for specific analyte capture, and deploying arrays to cover entire fields for near-field access. These advances could enable routine label-free visualization of viruses, nanoparticles, and semiconductor defects, providing a complementary route to resonant biosensors for diagnostics and monitoring.

• Requires objects to be in the near field of the nanowire sensors; coverage of the full field requires arrays of nanowire pairs and proper alignment to standing-wave nulls.
• Objects located exactly on the symmetry plane can become invisible due to cancellation of coupling terms under anti-symmetric excitation.
• Sensitivity to fabrication imperfections and surface roughness; while manageable, roughness can distort patterns and may necessitate optimized fabrication for single biomolecule visualization.
• Current demonstration used a low-performance camera with limited dynamic range and SNR; smaller targets (e.g., ~20 nm viruses) may require higher dynamic range detectors and further sensor optimization.
• Alignment complexity: generating uniform ECs over large areas depends on precise two-beam alignment and phase stability; sample scanning is needed to position structures at intensity nulls.
• The approach is contingent on proximity to engineered nanostructures, which may limit certain applications without integrated sensor arrays.