An emerging tool in healthcare: wearable surface-enhanced Raman Spectroscopy

This perspective explores the progressive domain of wearable surface-enhanced Raman spectroscopy (SERS), underscoring its potential to revolutionize healthcare. As an advanced variation of traditional Raman spectroscopy, SERS offers heightened sensitivity in detecting molecular vibrations. Applied in wearable technology, it provides a mechanism for continuous, non-invasive, real-time monitoring of chemical and biomolecular processes in the human body through biofluids such as sweat and tears. This underscores its immense potential in enabling early disease detection and facilitating personalized medicine. However, the adoption of wearable SERS is not without challenges, which include device miniaturization, reliable biofluid sampling, user comfort, biocompatibility, and data interpretation. Nevertheless, this perspective emphasizes that the fast-paced advancements in nanotechnology and data sciences render these challenges surmountable. In summary, the perspective presents wearable SERS as a promising innovation in healthcare’s future landscape. It has the potential to enhance individual health outcomes significantly and lower healthcare costs by promoting a preventive health management approach.


Introduction
As our world continues its rapid evolution, technology profoundly impacts every facet of our daily lives.One such innovation, wearable surface-enhanced Raman spectroscopy (SERS), is gaining considerable traction in healthcare .Until recently, wearable technology was largely characterized by devices such as fitness trackers and smart watches, which enable real-time monitoring of physical activities and essential biometrics [40].These devices have laid the groundwork for a more health-conscious society, promoting proactive health management.The integration of SERS technology into wearable devices represents a significant leap in this area .This revolutionary tool facilitates continuous, non-invasive, real-time analysis of chemical and biomolecular processes within the human body through biofluids like sweat and tears [41].Its capacity to detect substances even in minute concentrations paints an optimistic picture for early disease diagnosis and personalized medicine, propelling healthcare from a reactionary to a preventive approach.
SERS, an analytical method, magnifies the vibrational spectra of molecules when they are in close proximity or directly attached to a nanoscale roughened or nanostructured metal surface [42][43][44][45].The amplification mechanism stems from the localized surface plasmon resonance induced by the metal surface when exposed to light, typically near-infrared laser light.This optical excitation results in a dramatically enhanced Raman scattering signal, allowing for the exceptionally sensitive detection of target substances.The technological feat of incorporating SERS into wearable devices opens up the possibility of constant monitoring of specific chemical changes within the human body.This provides an extensive pool of data that could be invaluable to understanding an individual's health status.Given that SERS can detect a broad spectrum of substances, including toxins, metabolites, and disease biomarkers, its wearable format could pioneer a new epoch in healthcare and biomedical research .
The transformative potential of wearable SERS devices in healthcare is indeed compelling.The ever-present need for quick, accurate, and timely disease detection in healthcare could be significantly addressed by wearable SERS devices .Their capability to detect subtle changes in biomarker levels in real time could enable early disease detection, potentially even before symptoms manifest.This could bring about a paradigm shift in the diagnosis and management of chronic diseases like diabetes, heart disease, and cancer, where early detection often leads to more successful treatments and better patient outcomes [1][2][3][4].Furthermore, the real-time and continuous monitoring capability of wearable SERS can pave the way for truly personalized medicine.Here, treatments could be adjusted dynamically based on the real-time data collected about an individual's body chemistry [46][47][48], particularly beneficial for patients with chronic conditions or those on long-term pharmacological treatments.
In contrast to conventional SERS devices which are typically rigid and designed for laboratory settings, wearable SERS devices are characterized by their flexibility and conformability, allowing them to comfortably fit the contours of the human body .These devices are highly sensitive, enabling the detection of trace amounts of biomolecules, and are designed to be highly selective, distinguishing specific molecules amidst complex backgrounds [20][21][22]31].They are compact, lightweight, and often integrate with optical and electronic components for spectroscopic measurement, data processing, and wireless communication, facilitating real-time data analysis and remote monitoring [16,25,27,28,39].Durability is a key feature, with resistance to environmental factors such as moisture and temperature changes, ensuring reliable performance [20,21,24,25,28,36].The non-invasive nature of wearable SERS devices makes them suitable for continuous health monitoring.Customizable substrates and materials allow for a wide range of applications, including medical diagnostics and environmental monitoring [20,36].Additionally, some devices offer multiplexing capabilities to monitor multiple analytes simultaneously [16,21,28,29,36].Energy efficiency is also a significant aspect, with designs consuming minimal power for extended use without frequent recharging or battery replacement [39].Wearable SERS devices thus represent a unique intersection of nanotechnology, material science, and biotechnology, offering versatile platforms for various applications.
Despite the tremendous potential of wearable SERS technology, its broad adoption presents challenges.Miniaturizing SERS devices, particularly optical spectrometers, while maintaining their sensitivity and specificity, is a significant hurdle [49].Moreover, developing non-invasive, reliable methods for biofluid sampling for real-time analysis is a complex task.Other factors such as user comfort, data privacy, the biocompatibility of SERS substrates with human skin, and data interpretation require careful consideration [50,51].Even so, with the pace of technological evolution and progress in nanotechnology and data sciences, these challenges seem increasingly surmountable.The future of wearable SERS technology in healthcare is promising, marking a potential dawn of a new era in medical diagnostics and disease management [52][53][54].Wearable SERS could allow us to predict health issues before they become significant, enabling a shift towards preventive care.Such a transition would not only enhance individual health outcomes but also significantly reduce healthcare costs.
In this perspective, we explore the exciting field of wearable SERS and its potential to transform healthcare.This powerful tool, which provides continuous, non-invasive, and real-time monitoring of chemical and biomolecular processes within the human body, has far-reaching implications for early disease detection and personalized medicine (figure 1).This perspective also delves into the challenges in integrating wearable SERS into healthcare, including device miniaturization, reliable biofluid sampling, user comfort, biocompatibility, and data interpretation.However, we posit that the rapid advancements in nanotechnology, including various types of flexible SERS substrates, indicate these hurdles are not insurmountable.Thus, wearable SERS emerges as a promising component of the future healthcare landscape.Its potential to significantly enhance individual health outcomes and curtail healthcare costs by promoting a preventive approach to health management is indeed compelling.

Methods for wearable SERS
In this section, we elucidate the methods employed in the construction of wearable SERS devices and deliberate upon the advantages and drawbacks associated with each wearable SERS device.Wearable SERS devices encompass plasmonic nanostructures and flexible substrates.The plasmonic nanostructures, mainly composed of silver and gold, are affixed to the flexible substrates through three distinct techniques: the decoration of the metal nanostructures, the synthesis of the metal nanostructures, and the deposition of the  metal onto nanostructures.In relation to the flexible substrates, they can be categorized into three types: film, microfluidic channel, and mesh.To summarize, plasmonic nanostructures are integrated into three classes of flexible substrates (film, microfluidic channel, and mesh) through the processes of decoration, synthesis, and deposition (table 1).

Film substrates
For wearable SERS devices utilizing a film, we introduce three distinct methods for the construction of plasmonic nanostructures, namely (1) decoration, (2) synthesis, and (3) deposition.In the case of the decoration of the plasmonic nanostructures, a suspension encompassing silver nanowires was adeptly dropped onto the dual surface of the piezoelectric-modulated layer by spin-coating [18].Moreover, self-assembled gold nanoparticles (NPs) or a monolayer comprising silver nanocubes were judiciously transposed onto polyethylene terephthalate sheets or the hydrogel film, respectively [15,25].Regarding the chemical synthesis of plasmonic nanostructures, there exists a documented process where silver NPs were meticulously synthesized upon a poly(methyl methacrylate) microneedle array employing Tollen's method (figure 2) [19].Through the deposition of gold, a flexible film composed of silk fibroin-anodic aluminum oxide (AAO), conceived through the etching of rigid AAO, or a hierarchically porous meta-surface constructed from polycarbonate, was enveloped in a layer of gold NPs [20,21].In a separate instance, gold was thermally deposited onto a periodically arranged monolayer of polystyrene, following which the plasmonic nanostructures were skillfully transferred onto a contact lens [22].In a different way to the conventional technique of metal deposition onto nanostructures, heart-shaped gold nanodimers were intricately forged upon polydimethylsiloxane film, achieved by milling and selectively peeling the deposited gold film [23].Finally, we shall deliberate upon the advantages and drawbacks associated with wearable SERS devices utilizing a film.The construction of wearable SERS devices using a film substrate is straightforward and amenable to scalability.Conversely, such devices are beset with issues pertaining to flexibility and stretchability.When this flexible film, housing the plasmonic nanostructures, undergoes flexion and extension, the interstice dimensions amidst the plasmonic nanostructures within the film may be dynamically altered.This phenomenon begets a corresponding modulation in the SERS intensity, engendered through the localized plasmon resonance condition [44][45][46][47].Indeed, some wearable SERS devices, implementing a flexible film, manifest alterations in SERS intensity contingent upon the imposition of bending, stretching, and applied pressure [18,23,24], diverging from other wearable SERS devices utilizing flexible films [20,24,25].Briefly, these wearable SERS devices incorporated into a film through the techniques of decoration, synthesis, and deposition are convenient to manufacture and possess scalability, but they exhibit limitations in terms of flexibility and stretchability.

Microfluidic substrates
SERS devices based on microfluidic channels are normally comprised of an adhesive layer, a microfluidic layer, plasmonic nanostructures, and an encapsulation layer [24,[26][27][28].The adhesive layer bears orifices as ingress points to facilitate the conveyance of biofluids from the skin into the plasmonic nanostructures via the microfluidic conduit.The microfluidic channel can be realized through the excavation of trenches [26][27][28] or the delineation of hydrophilic zones upon a pliable substrate [24].The plasmonic nanostructures can be fabricated in the microfluidic channel through three types of methods: (1) decoration, (2) synthesis, and (3) deposition.Upon the paper-based microfluidic channel, gold nanorods were decorated, playing a pivotal role in the constitution of the plasmonic nanostructures (figure 3) [26].To the best of our knowledge, there may be no report on SERS devices utilizing microfluidic channels through the synthesis of the metal.Through the deposition of silver NPs, a hierarchical plasmonic structure, akin to compound eyes, was meticulously fashioned within an all-encompassing arrangement of nanovoids on an array of microparticles, all sheathed in a layer of gold [24].Furthermore, arrays resembling nano mushrooms, constructed from either gold or silver, were crafted via the deposition of metal upon silicon nanopillars, and subsequently repurposed for the fortification of the plasmonic nanostructures [27,28].Within these plasmonic nanoarrays, the NPs were predominantly configured in the fashion of hexagonal close packing.Upon the encapsulation layer, apertures function as egress points, serving to facilitate the efficient transport of biofluids through the plasmonic nanostructures via evaporation.This process aids in the concentration of biofluids before commencing the SERS detection [28].Finally, we shall deliberate upon the advantages and drawbacks associated with wearable SERS devices utilizing microfluidic channels.The utilization of a microfluidic channel emerges as a more advantageous technique for the conveyance of biofluids on the skin [24][25][26][27][28], in contrast to the porous film employed in wearable SERS devices [15,20,21,29].Conversely, the manufacturing of wearable microfluidic SERS devices is a protracted process and lacks scalability.Briefly, these wearable SERS devices integrated into microfluidic channels via the methods of decoration and deposition prove beneficial for sweat detection on the skin but face challenges related to scalability.

Mesh substrates
For wearable SERS devices utilizing a mesh, we introduce three distinct methods for the construction of plasmonic nanostructures, namely (1) decoration, (2) synthesis, and (3) deposition.In the endeavor to contrive flexible SERS devices employing mesh substrates, the suspension of silver nanowires was meticulously dropped onto the silk fibroin layer, with these nanowires assuming the role of a mesh [30,31].Furthermore, the suspension of sea urchin-shaped hollow gold-silver NPs, conceived to embody the silver core-gold shell structure, alloy inclusive, was delicately dropped upon nonwoven fabrics adorned with carbon nanotubes [32].Self-assembled gold NPs were transferred not solely onto films and porous membranes but also onto supple textiles [15].In an alternative scenario involving the metal NPs, sodium alginate underwent dissolution within a suspension of gold NPs, giving rise to the fabrication of SERS fibers, denoted as calcium alginate infused with gold NPs.The synthesis of mesh was accomplished through a wet-spinning technique, exploiting the coagulation of calcium ions [33].In the synthesis of the metal NPs, polyhedron-shaped silver NPs were formed upon electrospun porous styrene-butadiene-styrene (SBS) nanomesh, employing microplasma as the instrumental medium [34].By virtue of the deposition of gold upon electrospun thermoplastic polyurethane or polyvinyl alcohol (PVA) nanomesh, the realization of wearable SERS devices was consummated [35].The gold-PVA nanomesh, acting in the capacity of an adhesive, could be affixed to a moist target sample, facilitated by the application of water spray to dissolve the PVA (figure 4) [36,37].Further innovation in this realm entailed the deposition upon the nanograting of a polymer replica, an operation that facilitated the construction of an array of silver nanowires.These nanowires were subsequently transferred onto a substrate, whereupon they were stacked, forming a structure reminiscent of a mesh sheet.The assemblage of these stacked nanowire arrays culminated in their transfer onto a contact lens and graphene by means of a polymer film [38].Finally, we shall deliberate upon the advantages and drawbacks associated with wearable SERS devices utilizing mesh.Mesh substrates, renowned for their aptitude to imbibe biofluids from the skin, present themselves as uncomplicated to fabricate and exhibit remarkable scalability when juxtaposed with their wearable SERS counterparts harboring microfluidic conduits, which entail multiple strata.Conversely, it is imperative to note that the gaps in the metal-deposited nanowires or nanofibers, serving as the focal points from which SERS light predominantly emanates, within these three-dimensional (3D) arrays are meticulously aligned in a vertical orientation.This alignment contrasts with the horizontal orientation of the polarization direction of epi-excitation light.It is critical to consider that this misalignment between the orientation of the gaps and the direction of polarization can potentially give rise to suboptimal SERS enhancement [42][43][44][45].Briefly, wearable SERS devices integrated into a mesh using the procedures of decoration, synthesis, and deposition offer significant scalability but may not be fully optimized for SERS enhancement.In summary, the wearable SERS devices incorporated into a film, mesh, and microfluidic channel were fabricated through the techniques of decoration, synthesis, and deposition.The wearable SERS film and mesh are convenient to manufacture and possess scalability, while the wearable microfluidic SERS device is a state-of-the-art technology for the detection of biofluids on the skin.

Applications of wearable SERS
In this section, we present the application of wearable SERS devices.Wearable SERS devices provide valuable chemical insights into the biofluids of individuals who wear them.To monitor metabolites and pharmaceutical compounds within the human body, wearable SERS devices are adhered to the skin, enabling the SERS analysis of sweat.We not only discuss two applications for detecting metabolites and drugs in sweat, but also introduce other potential applications of wearable SERS devices.

Detection of metabolites and biomolecules
Here, we present the identification and quantification of various metabolites (glucose, lactate, lactic acid, urea, uric acid, dopamine, and ascorbic acid) within sweat, employing wearable SERS devices.Diabetes mellitus, a formidable metabolic ailment, necessitates vigilant monitoring of blood glucose levels.The conventional method of glucose detection via blood sampling is marred by numerous drawbacks, including discomfort, susceptibility to viral infections, labor-intensive sample preparation, and so on.Conversely, wearable SERS devices offer a non-invasive avenue for glucose detection, employing sweat and tears as mediums, closely mirroring the fluctuations in blood glucose levels [19,20,29,31].Indeed, the postprandial elevation in tear glucose concentration has been indirectly detected through a discernible diminishment or shift in the SERS spectral peak of the label molecules on the wearable SERS devices, prompted by the interaction between glucose and the label molecule [20,31].Furthermore, the silver-coated microneedle array, functionalized with 1-decanethiol, integrated into the wearable SERS device, facilitates direct measurement of glucose within the interstitial fluid [19].Dopamine, a neurotransmitter, serves as a valuable indicator of neurological disorders and emotional states [55].During periods of high-intensity physical exertion, dopamine levels in sweat can be effectively gauged by employing wearable SERS devices, functionalized with molecules adept at capturing dopamine [24].Lactate and lactic acid, the principal metabolic byproducts of anaerobic glycolysis, are generated in response to the body's demand for rapid energy production, particularly during instances of high-intensity physical activity or when oxygen availability is limited.Consequently, their concentrations bear a direct relationship with physiological conditions.In the context of intense physical exertion, this pathway may contribute to muscle fatigue via a concomitant reduction in pH [56].The rise in lactate concentration within sweat during strenuous exercise is amenable to direct detection via label-free wearable SERS devices worn by individuals (figure 5) [28,39].Urea and uric acid, metabolic waste products stemming from the catabolism of proteins and purines, respectively, are expelled from the organism through their transport in the bloodstream to the kidneys.Herein, these compounds are filtrated from the blood and subsequently excreted in urine.The accumulation of uric acid within the bloodstream is associated with hyperuricemia, a condition that may precipitate the formation of uric acid crystals, giving rise to painful maladies such as gout.To elucidate, the concentrations of these compounds within urine are indicative of kidney function [57].Although the continuous non-invasive monitoring of urine remains a challenge, recent advancements have enabled the continuous tracking of urea concentration variations in sweat following protein and non-protein intake, respectively.These developments have been made possible through the application of label-free wearable SERS devices [28].From simulated sweat samples containing lactic acid, the SERS device has been instrumental in simultaneously generating calibration curves for glucose, urea, uric acid, and ascorbic acid, hinging upon the chemical mechanism [29].Moreover, the real sweat samples have yielded the SERS peak corresponding to urea, concomitant with the concurrent detection of the peaks representing lactate or ascorbic acid through the silver nano mushroom array or the gold nanomesh within the wearable SERS device, respectively [28,36].Briefly, wearable SERS devices exhibit the capacity to simultaneously detect not only individual metabolites but also a multiplicity of metabolites within biofluids.

Detection of drugs
Next, we present the identification and quantification of pharmaceutical compounds (salicylic acid, acetaminophen, methamphetamine, and nicotine) within sweat, employing wearable SERS devices.Salicylic acid-based pharmaceuticals, typified by aspirin, encompass a subclass of non-steroidal anti-inflammatory agents.These compounds are distinguished by their antipyretic and analgesic attributes.Particularly, salicylic acid boasts keratolytic properties while, on occasion, instigating gastrointestinal discomfort and abdominal pain [58].In a bid to explore the feasibility of non-invasively discerning pharmaceutical agents, salicylic acid was detected within a model solution (distinct from authentic sweat) using the wearable SERS device [34].Acetaminophen, a widely employed antipyretic and analgesic remedy, is not devoid of risk, because an overdose of acetaminophen can inflict damage on the liver [59].Consequently, the non-invasive detection of acetaminophen within the human organism, as facilitated through the analysis of sweat, assumes importance in monitoring drug dosages and metabolic parameters.Indeed, the wearable SERS device was harnessed to measure variations in acetaminophen concentrations within sweat during exercise following acetaminophen administration [27].For its utility in forensic analyses, wearable SERS devices have demonstrated their proficiency in detecting abused drugs upon diverse surfaces [36,37].Furthermore, 2-fluoromethamphetamine, an analog of methamphetamine, commonly recognized as 'meth,' 'ice,' or 'speed,' was introduced into simulated sweat, applied to human cadaver skin, and subsequently detected through the wearable SERS device [30].Nicotine, a constituent of tobacco, elicits withdrawal symptoms in smokers.Transdermal nicotine patches serve as a cessation aid, providing relief from these withdrawal symptoms.In the interest of a non-invasive assessment of nicotine patch efficacy, a nicotine patch was affixed to an area of skin distanced from the wearable SERS devices.Subsequent measurements of nicotine concentrations on the skin, conducted through the wearable SERS device, revealed a protracted decrease in nicotine levels following patch removal, with the extent of decrease correlating with the distance between the patch and the wearable SERS device (figure 6) [25].Within the SERS spectra of diverse pharmaceutical compounds ascertained through the wearable SERS device, their distinct spectral features exhibit minimal overlap, enabling the effective discrimination of these varied drugs [25,36].Briefly, wearable SERS devices demonstrate the capability to concurrently identify not only individual pharmaceutical compounds but also a multitude of drugs within biofluids.

Detection of other parameters
We now introduce various other prospective applications for wearable SERS devices, encompassing the detection of breath constituents, bacteria, inflammatory biomarkers, biofluid pH levels, and the physiological parameters of the wearer.For the on-site analysis of breath, the wearable SERS device found its placement within the confines of a facial mask.In the SERS spectra of expiratory secretions, ascertained through a portable Raman spectrometer, conspicuous SERS peaks at 723 cm -1 and 1339-1343 cm -1 are assigned to nucleic acids and collagen, respectively [32].To substantiate the expeditious identification of cutaneous bacteria, capable of posing severe health risks, E. coli on simulated dermal tissue was discerned via the wearable SERS device.The SERS peaks originate from nucleic acids, proteins, and related constituents [32].Matrix metalloproteinase-9 (MMP-9) serves as an inflammatory biomarker with the capacity to selectively cleave peptide substrates.A peptide-functionalized gold SERS substrate, integrated into a contact lens, was employed in the detection of MMP-9.The SERS signals originating from the peptide exhibited diminishment at heightened MMP-9 concentrations due to the enzymatic cleavage of the peptide by MMP-9.The wearable SERS device integrated into the contact lens demonstrated reusability, for up to three cycles [22].Within the human body, the blood pH range is meticulously upheld, spanning from 7 to 7.35.In contrast, sweat pH, a parameter providing insights into dehydration, acidosis, cystic fibrosis, and osteoporosis, exhibits a variable range of 4.5-7 [60].For the determination of sweat pH utilizing wearable SERS devices, plasmonic components have undergone functionalization with label molecules like 4-mercaptobenzoic acid (4-MBA) and 4-mercaptopyridine (4-MPY).The emergence of SERS peaks, attributed to the deprotonation of carboxylic acid or pyridine rings within 4-MBA or 4-MPY, respectively, manifests at elevated pH levels.The calibration curves for pH were drawn from minuscule 1 µl droplets,  (c, d) Calibration curves of (c) the ratios of intensities of peak at 1398 cm −1 to those at 1590 cm −1 and (d) the ratios of intensities of peak at 1707 cm −1 to those at 1590 cm −1 .(e) Decrease of the SERS peak at 1707 cm −1 normalized to the reference peak at 1590 cm −1 at higher pH values.Reprinted with permission from [35].Copyright (2021) American Chemical Society.
employing the wearable SERS device (figure 7) [35].The wearable pH device via SERS exhibited a commendable reusability, extending to at least ten iterations [25].The wearable SERS device has facilitated the derivation of pH values from sweat, saliva, and urine, resulting in respective pH measurements of 5.79, 7.01, and 6.33 [33].The sweat pH, as detected by the wearable SERS device, demonstrated a decline during physical exertion [28].These pH values demonstrated concordance with those obtained using a conventional pH meter [28,33].As abovementioned, wearable SERS devices afford valuable chemical insights into wearers' biofluids.Conversely, wearable SERS devices employing a flexible film hold the potential to furnish physiological insights such as step count, heart rate, and blood pressure.Notably, the SERS intensity, derived from some wearable SERS devices incorporating a flexible film, responded to flexion, extension, and pressure-induced alterations attributable to the fluctuations in electromagnetic fields within the devices [18,23,24].Such wearable SERS devices, engineered for physiological data collection, may concurrently provide sweat pH data, given that the pH is derived not from the SERS intensities but from the peak intensity ratios [25,28,33,35].In summary, the potential of wearable SERS devices can be extended to encompass diverse applications, going beyond the multiplex detection of metabolites and drugs.

Quantification
We delve into the quantitative detection of metabolites and drugs using wearable SERS devices.The establishment of a quantitative relationship, represented by a calibration curve, between SERS intensities and sample concentrations has been primarily facilitated through the utilization of dye molecules [15,21,22,24,25,30,31,33,34,36].Notedly, extant literature exhibits calibration curves for various metabolites and drugs, including glucose [19,20,31], uric acid [26], urea [16,28], lactate [28], acetaminophen [27], salicylic acid [34], and MMP-9 [22].These were derived under conditions of a single analyte.Conversely, in the labeled detection, the performance of the wearable SERS device in discerning glucose in the sample with interfering agents such as fructose, sucrose, and galactose has been confirmed [20]; specifically, only glucose induces alterations in the corresponding Raman peak.For glucose, the limit of detection (LOD) has been reported to be 168 nM [20], significantly surpassing the diagnostic threshold range for blood glucose levels set by the American Diabetes Association in 2010 (7.0-11.1 mM) [31].For acetaminophen, the LOD is derived to be 130 nM with a measurement exposure time of 15 s [27].Regarding salicylic acid, the LOD is estimated to be 0.1 nM, presenting a marked improvement compared to conventional detection methods with LODs ranging from 13 µM to 23 pM, utilizing techniques such as high-performance liquid chromatography, differential pulse voltammetry, fluorescence spectrometry, and direct voltammetric detection [34].These findings attest to the notable linearity, sensitivity, and throughput of wearable SERS devices, standing in parity with conventional detection methods.Subsequently, we turn our attention to the consistency across measurements conducted with different setups and SERS devices.Distinct LODs have been ascertained with the same SERS device using different measurement setups.Notably, the LOD for R6G on a gold nanomesh SERS substrate has been determined to be 10 nM and 100 µM through a confocal Raman microscope equipped with a 50× objective lens (NA = 0.75) [36] and a wearable Raman spectrometer [37], respectively.The sensitivities of SERS devices fabricated through different methodologies are then compared.Enhancement factors (EFs), which are independent of the measurement setup, for R6G on a gold nanobowl substrate [22] and a gold nanovoid array functionalized with silver NPs [24] have been determined to be 3.09 × 10 5 and approximately 10 8 , respectively, upon excitation at 633 nm.Discrepancies in EFs may be attributed to the presence of silver NPs.In other instances, the EFs for R6G on gold/PVA nanomesh [36] and polyhedron silver NPs deposited onto SBS nanomesh [34] were calculated to be about 10 8 and 10 11-14 upon excitation at 785 nm and 532 nm, deemed suitable wavelengths for SERS with gold and silver, respectively.The discrepant EFs may originate from inherent differences in SERS efficiencies with gold and silver.Thus, the observed variations across measurement results with different setups and SERS devices are inherently plausible.Henceforth, the consistent quantitative detection of metabolites and drugs can be achieved uniformly across measurements executed with varied setups and SERS devices.

Challenges
Although past research  underscores the immense promise of wearable SERS technology, its widespread implementation is not without challenges.A primary concern is the downsizing of SERS devices, especially optical spectrometers, without sacrificing their sensitivity and accuracy [61][62][63].Additionally, creating reliable, non-invasive methods for real-time biofluid sampling poses its set of complexities.Considerations encompassing user comfort, data privacy, skin compatibility with SERS substrates, and data interpretation demand meticulous attention [47,50,51,[64][65][66][67][68].This section delves into these issues, offering insights into potential solutions.

Miniaturization
Since wearable SERS devices are designed for sensing a range of biofluids, including sweat, tears, urine, and saliva, it is imperative for these devices to be compact, lightweight, and flexible to ensure portability and comfort.Although SERS substrates have successfully achieved these attributes, the optical spectrometers used to measure the spectra of biofluids are still awaiting miniaturization [49,[61][62][63].The horizon looks bright for Raman spectrometer miniaturization, with progress anchored in the advancements of nanotechnology, photonics, and microfabrication (figure 8).The push for portable and wearable diagnostic tools, particularly in the healthcare and environmental sectors, amplifies the demand for compact yet potent spectrometers.Several strategies are under way to address this need.Firstly, the size of Raman spectrometers has considerably reduced due to the miniaturization of laser sources, with diode lasers being a noteworthy influence.Secondly, silicon photonic devices, along with other integrated optical components, hold potential for creating ultra-compact, on-chip Raman spectrometers.Thirdly, the advent of compact detectors such as CCD and CMOS detector arrays has been instrumental in the downsizing process.Furthermore, attempts are in progress to incorporate Raman spectroscopy functionalities into smartphones, utilizing their cameras as detectors.Lastly, the application of sophisticated data processing and machine learning algorithms can sift through noisy or faint signals, allowing for the deployment of smaller and possibly less sensitive devices in specific scenarios.The miniaturized Raman spectrometers that have been reported to date exhibit a wide spectral range spanning approximately 200-3000 cm −1 and a relatively low spectral resolution of approximately 10 cm −1 [61,63].The wide spectral range facilitates the sensitive simultaneous measurement of multiple analytes, while the low spectral resolution may hinder the precise differentiation of similar analytes.The latter challenge can be mitigated through advancements not only in spectrometer design, encompassing dispersive optics, narrowband filters, and Fourier transformation-or reconstruction-based systems [49], but also through enhancements in data processing and the implementation of machine learning algorithms.

Reliable biofluid sampling
Reliable sampling of biofluids is foundational for accurate SERS measurements.Key features of wearable SERS devices encompass microfluidic channels that direct the biofluid towards the SERS-active region, ensuring consistent and representative sample analysis [64,69,70].Sample volume control is also vital, which can be achieved using techniques like capillary action or other passive and active mechanisms to modulate the biofluid volume that interfaces with the SERS substrate.Equally crucial is the prevention of  sample contamination.This can be addressed using protective coatings or films that can be easily removed or replaced between tests.Additionally, for devices designed for multiple uses, the incorporation of disposable components ensures that new samples are not tainted by residues from previous ones.Maintaining sample integrity is another pivotal aspect, given that biofluids might undergo degradation or chemical alterations over time.Specifically, when analyzing sweat using wearable SERS devices, multiple potential targets emerge, such as sweat from sweat glands, sebum from sebaceous glands, and skin components leached into the adhesive solvent from the stratum corneum.The detection sensitivity of the device hinges on factors such as the distribution and density of sweat glands, which can differ based on skin location, condition, age, individual variability, and environmental temperature.However, areas such as the forehead, soles, palms, dorsum of the hands, and thumbs have a high sweat gland density [71].Notably, the palm, with its thicker stratum corneum, exhibits lower trans-epidermal water loss than the face and arm.In addition, its sweat gland density surpasses that of the face and arm, making the palm a potentially optimal site for sweat sensing, minimizing interference from stratum-corneum components.For effective analysis, the wearable SERS device must sample a significant sweat volume during its attachment to the skin.The bonding surface's choice is crucial.Additionally, as depicted in figure 9, elements other than sweat, such as sebum and skin components, may infiltrate the adhesive solvent of the wearable SERS device [72].The type of solvent selected for attaching these adhesive devices can influence the variety and quantity of skin components, such as amino acids from the stratum corneum [73], presenting challenges to be addressed in future studies.

User comfort
Comfort and health implications for users are paramount when considering the design and application of wearable SERS devices [64,65].These devices can inadvertently compromise the skin barrier function by inhibiting trans-epidermal water loss.This occlusion can lead to a significant increase in the water content of the stratum corneum, the skin's outermost layer of dead, denucleated cells.Ordinarily, the water content is lowest at the skin's surface, which is exposed to air, but it gradually increases deeper towards the living cell layer.However, when the skin's surface is occluded, the moisture level of the upper stratum corneum, which is typically maintained between 20-30 wt%, can surge [74].This moisture imbalance is especially pronounced in areas like the face or rougher skin patches where trans-epidermal water loss naturally occurs at a higher rate.The palms, with their dense population of sweat glands, could be particularly susceptible.The resulting skewed water content profile, akin to the skin's waterlogged state post-immersion, weakens the skin's protective barrier.This renders it more permeable to external contaminants, potentially triggering skin issues for the wearer.For optimal user comfort, designs should also ensure that the solvent used for the device's adhesion does not linger on the skin surface, further exacerbating the moisture imbalance.A balance between device efficiency and user comfort, while maintaining the skin's health, is vital in the next generation of wearable SERS devices.

Biocompatibility
Biocompatibility is of paramount importance for wearable SERS devices [50,51].It is crucial to ensure that trace metals or impurities from these devices are non-toxic to human skin.Beyond immediate toxicity concerns, there is the shadow of metal allergies [75].Prolonged exposure might instigate skin irritations, especially when friction from the device is combined with potential moisture-induced weakening of the skin's natural defenses.Allergic reactions can also stem from unexpected sources, such as the PVA in the gold/PVA nanomesh substrate.Moreover, the solvents and adhesives used to affix these wearables introduce another layer of potential health concerns.With the evolution of wearables, new materials and compounds are being studied, and it is paramount to study their long-term effects on human health.As we integrate wearables more deeply into our daily lives, their continuous interaction with the skin requires a thorough understanding.Assessing not just individual materials, but their combined and synergistic effects, will be the key to ensuring both the safety and success of wearable SERS devices in the wider market.

Data interpretation
When wearable SERS devices are attached to the skin using solvents such as water, it is vital to discern whether the detected components originate from sweat glands, sebaceous glands, or the stratum [47,[66][67][68].The latter can elute free amino acids and intercellular lipids into the solvent.In vivo spontaneous Raman scattering spectra of human skin in the high-wavenumber (figure 10) and fingerprint (figure 11) regions reveal numerous Raman bands, indicating biomolecules from the stratum corneum.Challenges remain in understanding the correlation between sweat components and those from the skin or blood, influenced by factors such as the sweat gland type, the individual's overall health, and their mental state.To establish this device as a standard in healthcare sensing, these complexities need further exploration.In future developments, by understanding the relationship between biomolecules in sweat and skin against the backdrop of overall body and skin health, more precise data interpretation can be achieved.

Conclusions
In the ever-evolving landscape of technology, wearable SERS stands poised to redefine healthcare (figure 12).Transcending the boundaries of conventional wearables like fitness trackers, SERS-equipped devices promise continuous, real-time monitoring of intricate biomolecular processes within the human body.While earlier devices fostered health awareness, SERS' incorporation holds the potential for proactive disease detection, enabling a paradigm shift towards preventive healthcare.The capacity of these wearables to pinpoint even trace amounts of substances paves the way for early disease diagnosis, potentially well before symptoms surface.Challenges, including device miniaturization and biofluid sampling, inevitably arise, yet the strides in nanotechnology and data sciences hint at viable solutions on the horizon.As this perspective underscores, wearable SERS not only epitomizes the future of medical diagnostics but could herald a transformational shift towards predictive and preventive care.By championing early interventions and facilitating truly personalized medicine, it can catalyze enhanced health outcomes while simultaneously streamlining healthcare expenses.The integration of SERS into daily wearables could very well signify a monumental leap in health management, emphasizing its profound implications for the future of global healthcare.

Figure 1 .
Figure 1.Transformative potential of wearable SERS devices in healthcare.It enables early disease detection and personalized medicine via continuous, non-invasive, and real-time monitoring of chemical and biomolecular processes within the human body.Reproduced with permission from irasutoya.

Figure 2 .
Figure 2. Fabrication process of the wearable SERS device employing a flexible PMMA film through chemical synthesis of silver nanoparticles.Reprinted with permission from [19].Copyright (2020) American Chemical Society.

Figure 3 .
Figure 3. Structure of the wearable microfluidic SERS device through decoration of gold nanorods (AuNRs).(a) Schematic illustration, (b) top view, and (c) stacked view of the wearable SERS device.(d) Picture of the wearable SERS device.(e) TEM image of the AuNRs on a microfluidic channel.This figure has been adapted from [26].Copyright © 2022 Mogera et al, some rights reserved; exclusive licensee American Association for the Advancement of Science.No claim to original U.S. Government Works.Distributed under a Creative Commons Attribution License 4.0 (CC BY).Reproduced from [26].CC BY 4.0.

Figure 4 .
Figure 4. Fabrication process and characteristics of the wearable SERS device composed of gold nanomesh.(a) Fabrication process.SEM images of (b) PVA nanomesh, (c) gold-deposited PVA nanomesh, and (d) gold nanomesh after removing PVA by spraying water.This figure has been adapted from [36].Copyright © 2022 Liu et al, Advanced Optical Materials published by Wiley-VCH GmbH.Distributed under a Creative Commons Attribution License 4.0 (CC BY).Reproduced from [36].CC BY 4.0.

Figure 5 .
Figure 5. SERS of lactate and urea in sweat using the wearable SERS device.(a) Picture of a portable Raman spectrometer and the wearable SERS sensor on the wearer's skin.(b) SERS spectrum obtained with the Raman spectrometer.(c) Monitoring of the wearer's sweat spectrum during his exercise.This figure has been adapted from [28].Copyright © 2022, He et al.Distributed under a Creative Commons Attribution License 4.0 (CC BY).Reproduced from [28].CC BY 4.0.

Figure 6 .
Figure 6.SERS of nicotine in sweat using the wearable SERS device.(a) Schematic illustration of the experiment.(b), (c) Nicotine concentration detected using wearable SERS devices A and B at different times after removing the nicotine patch.(d) Nicotine concentration detected using the wearable SERS device attached away from the nicotine patch at different distances.This figure has been adapted from [25].Copyright © 2021 Wang et al, some rights reserved; exclusive licensee American Association for the Advancement of Science.No claim to original U.S. Government Works.Distributed under a Creative Commons Attribution NonCommercial License 4.0 (CC BY-NC).From [25].Reprinted with permission from AAAS.

Figure 7 .
Figure 7. SERS of sweat pH using the wearable SERS device.(a) SERS spectra of 4-MBA on the wearable SERS device normalized to the peak at 1590 cm −1 within the sweat pH range.(b) Increase of the SERS peak at 1398 cm −1 normalized to the peak at 1590 cm −1 at higher pH values.(c,d) Calibration curves of (c) the ratios of intensities of peak at 1398 cm −1 to those at 1590 cm −1 and (d) the ratios of intensities of peak at 1707 cm −1 to those at 1590 cm −1 .(e) Decrease of the SERS peak at 1707 cm −1 normalized to the reference peak at 1590 cm −1 at higher pH values.Reprinted with permission from[35].Copyright (2021) American Chemical Society.

Figure 8 .
Figure 8. Need for a miniaturized spectrometer that can be attached to the human skin.(a) Picture of the wearable SERS device (gold/PVA nanomesh [36, 37]) on the wrist of the wearer.(b) Picture of wearable Raman spectrometers (WSERS-785, BaySpec).(c) Picture of a wearable Raman spectrometer attached to the wrist for SERS.(d) Picture of the user smartphone interface of the wearable Raman spectrometer.

Figure 9 .
Figure 9. Vertical cross-section image of human skin.(a) Sebaceous gland.(b) Rough skin.Reproduced from an original image whose copyright is owned by Shiseido Co., Ltd with permission.Reproduced with permission from Shiseido Co., Ltd.

Figure 10 .
Figure10.Spontaneous Raman scattering spectra of the human skin in vivo in the high wavenumber region.The spectra were obtained at different depths in the skin: 0-10 µm from the skin surface.Reproduced from[67] with permission from the Royal Society of Chemistry.

Figure 11 .
Figure11.Spontaneous Raman scattering spectra of human skin in vivo in the fingerprint region.The spectra were obtained at different depths in the skin: 0-10 µm from the skin surface.Reproduced from[67] with permission from the Royal Society of Chemistry.

Figure 12 .
Figure 12.Number of wearable SERS publications per year on the Web of Science.Search query: ALL = (wearable SERS).

Table 1 .
Comparison of wearable SERS devices in fabrication method and substrate type.