Nanophotonic sensing and label-free imaging of extracellular vesicles

Abstract

This review examines imaging-based nanophotonic biosensing and interferometric label-free imaging, with a particular focus on vesicle detection. It specifically compares dielectric and plasmonic metasurfaces for label-free protein and extracellular vesicle detection, highlighting their respective advantages and limitations. Key topics include: (i) refractometric sensing principles using resonant dielectric and plasmonic surfaces; (ii) state-of-the-art developments in both plasmonic and dielectric nanostructured resonant surfaces; (iii) a detailed comparison of resonance characteristics, including amplitude, quality factor, and evanescent field enhancement; and (iv) the relationship between sensitivity, near-field enhancement, and analyte overlap in different sensing platforms. The review provides insights into the fundamental differences between plasmonic and dielectric platforms, discussing their fabrication, integration potential, and suitability for various analyte sizes. It aims to offer a unified, application-oriented perspective on the potential of these resonant surfaces for biosensing and imaging, aiming at addressing topics of interest for both photonics experts and potential users of these technologies.

Fig. 1. Outline of the review with separate strands for sensing (bulk assays) and imaging (single particles).

a Nanophotonic sensing of bulk analytes. (i) General schematic of biosensing based on evanescent waves. (ii) Examples of resonant metasurfaces, including plasmonic nanohole arrays (left) and dielectric nanohole arrays (right, shown in a ‘chirped’ configuration). (iii) The readout can take many forms, such as spectral or intensity-based approaches. We limit this review to imaging-based sensing. Images are adapted from^,,,. b Single particle imaging using fluorescence microscopy. Resonant surfaces can be used to enhance fluorescent signals from single particles. Subfigures rearranged and taken from. c Label-free imaging. (i) Interferometric scattering microscopy (iScat) can detect single particles without labels. (ii) Resonant metasurfaces are used for surface-enhanced iScat. Images are from^,. Ab, antibody; PEG, polyethylene glycol; iPSF, interferometric point spread function; CCD, charge-coupled device

Fig. 2. Comparison between plasmonic and dielectric nanohole arrays.

a Schematic of typical spectra. The plasmonic resonance spectrum (dashed line) is based on experimental results in ref. , and the dielectric resonance spectrum (dashed line) on experimental results in ref. . This illustration highlights the practical differences between plasmonic and dielectric metasurfaces for sensing, presenting their respective advantages in terms of resonance Q-factor, amplitude, and sensitivity. The arrows indicate the peak shift towards higher wavelengths with increasing refractive index. This is a conceptual representation illustrating the enhanced refractometric sensitivity of plasmonic systems and is not based on empirical data. b The electric field (E) simulations show plasmonic ‘hotspot’ formation with strong near-field enhancement (top) and more extended confinement of resonances in a dielectric array (bottom). The image was adapted from

Fig. 3. Numerical simulation of resonances in periodic dielectric arrays.

a The Electric field (E), normalized to an incident field (E₀), was calculated for a silicon nitride photonic crystal slab. The top panel shows photonic crystal resonance (TM), and the bottom panel shows a symmetry-protected bound state in the continuum (BIC). Breaking the symmetry via a nonzero incidence angle (here 0.1°) results in a quasi-BIC (Q-BIC) mode with high Q and a ~ 10-fold higher enhancement in the electric field magnitude (compare the values inside red and blue ellipses). b Schematic representation of a biosensing configuration, illustrating the binding of target proteins (antigen active biolayer) to a functionalized surface (antibody fixed biolayer). c Calculated resonance spectra of the silicon nitride photonic crystal slab. Both TM and Q-BIC modes exhibit a redshift upon biolayer addition. The TM mode exhibits a more significant shift (i.e., higher sensitivity) compared to the Q-BIC mode, which is attributed to a greater field overlap with the analyte volume

Fig. 4. Summary of published state-of-the-art in quality factor (Q), sensitivity (S), and resonance peak amplitude (A).

The numbers within the diamond represent the reference numbers from which these values were obtained. Filled (colored) diamonds indicate that A values are directly or indirectly reported in the literature. Unfiled (clear) diamonds indicate that A values were not reported. Plasmonic resonances generally exhibit low Q and A but high S when compared to dielectric array resonances (high Q and A, low S). Some dielectric array resonances demonstrate exceptionally high Q, albeit often at the expense of reduced A. RIU, refractive index unit

Fig. 5. Vesicle detection with metasurface sensors.

a Label-free detection and molecular profiling of EVs with nanoplasmonic sensor. (i) Cancer cells produce exosomes through the fusion of multivesicular bodies with the plasma membrane. These exosomes carry parental cell proteins. (ii) Simulation showing enhanced electromagnetic fields near a periodic gold nanohole surface. (iii) Scanning electron micrograph of gold nanohole array. Images are adapted from. MVB, multi-vesicular body. b Dielectric metasurface for the detection of vesicles using a ‘chirped’ gradient nanohole array. (i) Scanning electron micrograph of a hydrogenated amorphous silicon nanohole array. (ii) Simulation showing mode confinement of a transverse magnetic (TM) mode with a period of 420 nm, thickness of 120 nm, and radius of 60 nm. (iii) Schematic (top) and image (bottom) of nanohole array with a spatial gradient in its periodicity. The yellow band (top) indicates the resonance, and the red line (bottom) shows a corresponding Fano resonance shape. (iv) Principle of imaging-based vesicle detection with a “chirped” nanohole array with a fixed operating wavelength. Images are adapted from^,

Fig. 6. Single EV imaging.

a Holographic fluorescence imaging of extracellular vesicles. Computational refocusing (left) and 3D tracking of fluorescent beads (right). Images are from. b Size photometry of single EVs. Schematic of the microscopy setup with a microfluidic flow cell, EVs on a coverslip, an epifluorescence microscope, sCMOS camera, and a 50:50 beamsplitter for iSCAT imaging. (i) Workflow to suppress background inhomogeneities. A pre-incubation iSCAT image is acquired to capture the background. Subsequently, EVs are immobilized, unbound EVs are washed, and a post-incubation iSCAT image is acquired. (ii) iSCAT images register candidate spots, followed by contrast or fluorescence intensity measurements. Images are from. LED, light emitting diode

Fig. 7. Surface-enhanced fluorescent imaging.

a Plasmonic nanohole arrays were employed to enhance fluorescent signals in EV imaging. (i) Biotinylated EVs were captured on an avidin-coated gold nanohole array and subsequently labeled with fluorescent antibodies. The assay resulted in enhanced fluorescence single EV imaging. (ii) Scanning electron micrograph of gold nanohole array. (iii) Simulated near-field enhancement of gold nanohole array. Images are adapted from. b All-dielectric metasurface fluorescence enhancement for antibody/antigen detection. (i) Scanning electron micrograph of silicon-on-insulator (SOI) nanorod metasurface. (ii) Measured reflectance spectra. (iii) Simulated electromagnetic field intensities at reflectance dip. (iv) Simulated reflectance spectra. Images are adapted from