Single-Exosome SERS Detection by Means of a Flexible Metasurface Authors: Konstantin Mochalov, Denis Korzhov, Milena Shestopalova, Andrey Ivanov, Konstantin Afanasev, Alexander Smyk, Alexander Shurygin, Andrey K. Sarychev Published: 2025 ABSTRACT Single exosomes are detected via surface-enhanced Raman scattering (SERS) due to electromagnetic field accumulation on a specially designed flexible metasurface. This metasurface is a modulated silver nanofilm deposited on a thin, flexible plastic substrate. An explicit Equation for calculating the local electric field is given. The field reaches extremely high values under plasmon resonance conditions and fills the depressions of the metasurface. The thin, flexible metasurface can be incorporated into automated Lab-On-Chip analytical systems and used for spectroscopic studies of exosomes. We propose a method to distinguish individual exosomes from the HEK293T cell line on the metasurface and then obtain and assign their SERS spectra. An important advantage of the plasmonic metasurface presented in this work is its spatial complementarity to exosomes and other vesicle-like objects. The plasmonic metasurface is fabricated using holographic lithography and further investigated using a correlation approach combining atomic force microscopy, scanning spreading resistance microscopy, and surface-enhanced spectroscopy. 1. Introduction Single exosomes are detected via a Raman signal from a metasurface (MS) specially developed for research. Exosomes are extracellular vesicles with a diameter of a∼ 100 nm. They can reflect the biological information of their parent cells and participate in intercellular communication [1]. Exosomes have become one of the most promising markers for the early diagnosis and prognosis of cancers and other diseases. With the development of modern optical technology and the emergence of high-resolution optical microscopies, the micro/nanostructures of exosomes can be observed optically for the first time. A flexible metasurface (FMS) is being developed to obtain Raman signals from single exosomes. A plasmon FMS accumulates energy from incident light by exciting local plasmon resonators. Energy storage in plasmon resonances generates a giant local electric field, which is important for photovoltaics, photodetectors [2,3], and metalenses [4,5]. MSs are used for near-field power transfer [6]; generation of second, third, and higher harmonics from an incident laser beam [7,8]; and surface-enhanced Raman scattering (SERS) [9,10]. Recently developed MSs generate hotspots with a local photonic density of states, which, in principle, allows the detection of microscopic quantities of biomolecules and bioobjects. Metal–dielectric MSs, which use gold or silver as metal components [11,12,13,14], exhibit stability and reproducibility. Gold or silver FMSs can be integrated into a microfluidic analytical system, thus creating a so-called Lab-On-Chip for medical diagnostics [15,16,17,18,19,20,21,22]. Clinical microfluidic platform for rapid label-free capture and SERS detection of viruses has been reported [23]. A nanoimprinted lithographic FMS was developed for SERS sensors for SARS-CoV-2 viruses [24]. Recently, [25] an MS composed of hydrophobic silver nanowires (H-CuO@Ag NWs) was used for the quantitative detection of trace nanoplastics (down to a∼ 50 nm) via SERS based on a multiplex-feature coffee ring approach. The experimental results revealed a strong correlation between the formation of the coffee ring and the concentration of the nanoplastic dispersion. The authors developed a machine learning model that maps the nanoplastic concentration to the coffee ring characteristics, and plastic nanoparticles in natural river water were detected. The close similarity in SERS detection of nanoplastics and exosomes should be noted. EM field enhancement in an open resonator (e.g., [26]) can be also invoked. Exosomes are nanosized lipid droplets, which, as suggested in the modern literature [27,28], are formed by living cells for intercellular communication. Exosomes include protein, RNA, and DNA molecules, so they are assumed to participate in the regulation of normal as well as pathological processes in living cells [29,30,31]. Exosomes are intensely studied since they are promising agents for oncology diagnostics [32,33]. Raman and IR spectroscopy are employed for the study of exosomes (see, e.g., [1]). Raman scattering (RS) has a very small scattering cross-section, and to obtain a pronounced RS, it is necessary to collect a large number of exosomes and obtain a collective RS signal, including traces of exosome compaction. A solution is the study of the exosomes using the SERS method, which enhances RS by many orders of magnitude [34,35]. SERS technology is highly relevant to exosome detection due to its unique ability to provide highly sensitive and specific molecular fingerprints. SERS can be employed to detect exosomes with minimal sample preparation, making it an ideal method for analyzing exosomes, including detection of various proteins, nucleic acids, and lipids [34,35,36,37,38,39,40]. Numerous researchers have employed SERS for the identification of exosomes, utilizing methodologies that can be classified into labeled and label-free approaches. Labeled techniques offer high accuracy and semi-quantitative analysis; however, they are limited to capturing information from the labels [41,42,43,44,45,46]. For example, nanoprobes and magnetic nanobeads were used in [41] for SERS-based detection of tumor-derived exosomes. Magnetic nanobeads and SERS nanoprobes can capture exosomes by forming a sandwich-type complex. The exosome–nanobead complexes are precipitated with a magnet, and thus SERS signals are detected from compacted exosomes. Label-free SERS detection of exosomes requires simultaneous detection of different exosomes. Label-free SERS, in principle, allows one to distinguish between various exosomes. In the seminal work [47], small drops of solution containing exosomes, obtained from either healthy or tumor colon cells, were deposited on a super-hydrophobic surface consisting of periodic silicone micropillars, the tops of which are deposited with silver nanoparticles (NPs). SERS spectra obtained from healthy or tumor colon exosomes are found to be somewhat different. A direct Raman signal was obtained in [48] from purified exosomes. This signal allows one to distinguish exosomes derived from seven cancerous and non-cancerous cell lines. In [49], it is shown that regular recordable disks (CD-R and DVD-R) coated with a silver nanolayer reveal the SERS fingerprint of hemoglobin and exosomes. In [50], gold–silver NP shells were formed around exosomes to obtain their SERS spectra. In a series of investigations [51,52,53], SERS spectra of individual exosomes were obtained using a hybrid SERS substrate, which comprised a graphene-covered quasi-periodic array of gold pyramids (see also simulations in [54]). SERS spectra of lung cancer exosomes were also obtained from graphene-covered gold pyramids [52]. A hybrid porous biomaterial scaffold with silver NPs is suggested in [55] for a liquid biopsy via exosome diagnostics. Acoustofluidics-induced enrichment of NPs is introduced in [55]. CD63 aptamer-conjugated silica NPs + exosomes were pushed by acoustic waves to the surface of the glass capillary covered by ZnO nanorods, which was analyzed via SERS. In [56], TiO2 micro-porous inverse opal honeycomb coated by gold nanofilm was used as an effective SERS substrate. The honeycomb structure captures the exosomes from the plasma of cancer patients. The authors focused on the intensity of the 1087 cm−1 Raman line, which is a signature of phosphoproteins. They speculated that the intensity of the line is a signature of tumor exosomes. In important works by Shin et al. [57,58] gold NPs with a size of a∼ 100 nm were settled onto glass, where they formed fractal clusters. The fractal substrate was used for exosome SERS characterization and detection of six types of early-stage cancers (lung, breast, colon, liver, pancreas, and stomach). A microfluidic device with SERS for the identification of glioblastoma exosomes is considered in [38]. The same group developed a multiplex fluidic device with embedded arrayed nanocavity microchips that achieves confinement of single exosomes in a minute amount of fluid (<10 μL) and provides the SERS spectrum [59]. In [60], a nano-engineered 3D SERS sensor was reported. The sensor is composed of titania (TiO2) NPs (a∼ 10 nm) containing SERS-active Magneli black TiO2. An ultralow limit of detection (10 exosomes per 10 μL) was achieved. The achievement is attributed to the simultaneous detection of multiple exosome signals. The paper [61] proposes rapid and label-free detection of breast-cancer-derived exosomes from SERS spectra obtained via the laser ablation of a silver plate. In [62], SERS was performed with the combination of Ag NPs and a single atomic layer of graphic carbon nitride C3N4, which concentrates exosomes due to the π−π interaction between aromatic exosome molecules and the carbon nitride. In [63], exosomes were investigated using tip-enhanced Raman spectroscopy. Contact between the tip and the bovine milk exosomes can be used for drug delivery and therapeutic applications. In [64], SERS from the nanowire microlens was used for cancer diagnosis, and 24 cases of early pancreatic cancer were successfully classified. In [65], it was proposed to perform exosome detection via SERS based on a TiN–Ag@Ag sol composite substrate. A statistical difference was obtained in the SERS of exosomes extracted from the serum of 31 gastric cancer patients and 31 healthy controls. To our knowledge, there are currently no generally accepted solutions for exosome diagnostics. For example, microRNA from exosomes, which is supposed to serve as a tumor marker, varies in different studies (cf. studies of colorectal cancer [66,67]). The membrane proteins of exosomes may be candidates for the role of tumor markers. The tumor marker proteins HER2, EpCAM, CD63, and PSMA have been proposed for the diagnosis of breast and pancreatic cancer [68,69,70,71,72,73,74]. As we mentioned above, there are two approaches to SERS studies. Metal NPs can be added to an exosome solution in such a way that NPs and exosomes form a mixed aggregate [41,43,44,45,75,76]. Collective plasmon resonances in the metal fractal aggregate lead to an increase in the local electric field E. This field interacts with the exosomes incorporated in the aggregate. The exosomes, in turn, generate Raman signals that excite secondary plasmon resonances in the aggregate of metal NPs [77,78]. The emission of secondary plasmons enables SERS, the amplitude of which, G, is proportional to the fourth power of the local field G∼E4 due to the two-stage scattering. The local field is significantly enhanced in the gaps between NPs. However, the gaps between NPs should be much smaller than the particle size (see, e.g., the explicit equation in [79] and the discussion therein). In any case, the NP+exosome aggregate facilitates collective SERS from captured exosomes. We believe that plasmon MSs are promising tools for obtaining SERS signals from exosomes [79,80,81]. In the near future, the efforts of dozens of experimental groups around the world could allow us to diagnose cancer and other diseases from the SERS spectra of exosomes extracted from liquid biopsies. An important step in this direction is to obtain spectra of individual exosomes. In this paper, we present the Raman spectra of single exosomes deposited on a specially designed silver film, which is modulated with a period ≈ 400 nm smaller than the size of an exosome. An exosome at the smooth silver surface can be visualized using, for example, atomic force microscopy (AFM) as discussed below. An explicit equation is suggested for finding the local electric field in this FMS. The modulation is specifically designed to localize the plasmon in the FMS recesses, where the exosomes are typically localized. The structure of the plasmon electric field appears to be optimal for SERS studies of single exosomes. The fabrication process of the FMS, suitable for mass production, is discussed. The method was specially developed for exosome deposition. The resulting FMS with placed exosomes was used to obtain the Raman spectra of individual exosomes. 2. Materials and Methods 2.1. Theoretical Background A thin, modulated, flexible metal film that operates as a flexible metasurface (FMS) is used to detect the Raman spectrum of an exosome or any other microscopic object. This FMS effectively concentrates the electromagnetic field and operates as an SERS substrate [22,80]. To understand the plasmon resonance and the field enhancement in the FMS, we consider the plasmon resonance in the simplest plasmon system, namely, the metal hollow nanocylinder (NC) with an internal radius r=ri and external radius r=re. The thickness d=re−ri of the NC is assumed to be less than the skin depth δ∼(kℑεm)−1≫d, where k=ω/c is a wavevector, ω is the frequency of the external electric field, and εm(ω) is the metal permittivity. It is also assumed that rek≪1, though it is not so critical for the application of the quasistatic approximation. The NC is excited by an external electric field. The electric field, calculated in the quasistatic approximation, has a textbook form [82] when the NC is excited by the uniform external field E0. The field inside the hollow is uniform, and it is equal to Ei=E0A/Det, where the coefficient A=4α2εeεm, and the determinant is equal to(1)Det=α2εe+εmεi+εm+εm−εeεi−εm, where α=re/ri; εi, εm=εm′+iεm′′; and εe are the permittivities of the internal dielectric, metal NC, and the ambient medium, respectively. The optical metal permittivity εm(ω) depends on the frequency, and its real part εm′(ω)<0 is typically negative in optics. The imaginary part εm′′ of the metal permittivity is small for good optical metals like silver or gold, and the optical quality Qm=ε′m(ω)/εm′′(ω) is large [83]. For any given frequency ω, there is a critical ratio α=re/ri of the external re and the internal ri radii of the cylinder(2)αRes≃εe−εm′εi−εm′εe+εm′εi+εm′, when Det is minimal in absolute value,(3)DetRes≃2iεe+εiεm′εm′2−εeεiQmεe+εm′εi+εm′≃2iεe+εiεm′Qm, where it is assumed that the optical quality Qm(ω)≫1; the last estimate in the above equation holds when εm′≫1, which is typical for red and infrared spectral ranges. Then, the internal field Ei is much enhanced: Ei/E0≃4εeαRes2εm′/DetRes≃2εeQm/(εe+εi)∼Qm≫1. When a molecule is placed inside the NC, it generates a Raman signal at the Stokes frequency ωs≈ω, which in turn excites the secondary plasmon resonance in the cylinder. The sequential plasmon resonances exhibit an effect called surface-enhanced Raman scattering (SERS). The total Raman enhancement is estimated as G∼Qm4≫1 since the Stokes frequency ωs is typically close to the laser frequency ωs≈ω. Let us consider the internal plasmon resonance excited from within the hollow NC. To calculate the local electric field, it is convenient to use the complex potential U, while its real part has physical meaning. We measure all distances in terms of the internal radius ri, and we introduce dimensionless complex coordinates u and v with the origin in the center of the NC and complex variable w=u+iv. Then, in terms of the dimensionless coordinate w, the internal and external surfaces of the NC are given by wi=exp(iφ) and we=αexp(iφ), respectively, where α=re/ri and φ are the polar angles in the u-v coordinate system. Suppose the potential source, which is placed inside the NC, is equal to Uext=iU0log(w−a), where the parameter 0a. The first term in Equation (4) gives the circular electric field, the amplitude of which, ∼1/r, does not depend on metal permittivity. The second, third, etc., terms in Equation (4) excite dipole, quadrupole, etc., plasmon resonances in the NC. The first term in the sum in the r.h.s. of Equation (4) is equal to Ud=−iU0a/w, and it excites the electric field with the potentials Ui=Aw, Um=Bw+C/w, and Ue=D/w inside the hollow NC (r<1), in the metal NC (1α), respectively. The radial Er and the polar Eφ electric fields are obtained from the complex potential as follows: Er−iEφ=−exp(iφ)dU/dw. The coefficients A, B, C, and D are found by matching the polar electric fields and normal electric displacement at the inner and outer interfaces of the NC. The amplitude of the internal “dipole” electric field is constant and is equal to(5)1U0dUidw=A=aα2εe+εmεi−εm−εe−εmεi+εmDet, where the determinant Det is given by Equation (1). Det reaches a minimum when the ratio α=αRes, given by Equation (2). The minimum of the determinant Det=DetRes (Equation (3)) corresponds to the maximum amplitude of the internal field, which is estimated as(6)Amax=2aεiQmεe+εiεm′2−εe2εeεi−εm′2∼Qm where it is assumed that the metal optical quality Qm≫1. The conformal mapping (see [80,84,85,86,87])(7)z1=x1+iy1=−ilog(w−a), transforms the metal NC into a modulated metal metasurface extending along the x1 axis in the space {x1,y1}. The dimensionless coordinates {x1,y1} are equal to {x1,y1}=(2π/L){x,y}, where L is the period of FMS modulation. The inner interface of the cylinder transforms into the right film surface yu(x), which is illuminated by the light falling from the right (from y=+∞), as shown in Figure 1. The outer surface of the cylinder transforms into the left film surface yb(x). That is, the internal space of the cylinder (a0, and n=0,1,2… The depressions in the right interface of the FMS have coordinates xu,min=±nL+L/2 and yu,min=−(L/2π)log(1+a)<0; recall the parameter 0