Flexible mechano-optical sensors from mechanoluminescence to mechanoplasmonics: designs, applications, and prospects

Abstract

Flexible mechano-optical sensors (FMOS) achieve quantitative sensing of mechanical stimuli by monitoring changes in optical response, and due to the incorporation of a polymeric matrix/substrate, they exhibit high flexibility, elasticity, and biocompatibility. This wireless and visualized sensing capability offers potential for both in situ and in vivo applications. In this review, we delve into the mechanisms and developments of two types of FMOS: “active” mechanoluminescence (ML) and “passive” mechanoplasmonics (MP). The focus is on how ML particles and polymers can be combined in various configurations (such as bulk, laminar, and woven blending systems) to yield robust, multifunctional, and hybrid optical/electrical properties, exploring their potentials in engineering, information, and wearable/implantable applications. Additionally, the tunability of ML intensity and emission color under mechanical and various environmental stimuli is summarized, leading to a discussion on the versatile MP nanostructures. With their sophisticated artificial design, MP demonstrates promise for both small-scale sensing and high-level control over spectral wavelength and intensity. Lastly, based on current research on ML and MP, challenges and prospects for combining these two technologies to advance the field of FMOS are proposed.

---

Wei Tao

Wei Tao is a lecturer at the DGUT-CNAM Institute of Dongguan University of Technology in China. Wei earned his joint PhD degree in 2023 from the Department of Materials, Mechanics, Optics, and Nanotechnologies (M2ON) at the Université de Technologie de Troyes (UTT) and the Institute for Applied Physics at the Eberhard Karls University Tübingen (EKUT). His research interests focus on nanostructure-based flexible sensors: mechanoplasmonics, nanogenerators, and optoelectronics for sensing applications.

Ye Liu

Ye Liu received her BS degree in optical information technology from Beijing Jiaotong University, and PhD degree in Optics from Institute of Physics, Chinese Academy of Science. Then she worked as an assistant professor, associate professor in Anhui Institute of Optics and Fine Mechanics, Chinese Academy of Science. Now she is a professor in Dongguan University of Technology. Her research interests include advanced optical sensors, optical fiber sensing technique, and laser spectroscopy.

Thomas Maurer

Thomas Maurer is currently Centrale Lille's Dean and general director. After studying engineering at the Mines de Nancy and a master's degree in solid state physics at the Université Paris-Saclay, Thomas specialised in nanomagnetism and neuron scattering, completing in 2009 at the Laboratoire Léon Brillouin (CNRS-CEA) his doctorate from Université Paris Saclay. After a post-doctorate at the Max-Planck Institute in Stuttgart, he held the CNRS Chair of Excellence in Optical Nanosensors at the Université de Technologie de Troyes (UTT) from 2010 to 2015, before becoming a full professor in 2018 at the UTT. His main research domain concerns mechanoplasmonics.

Monika Fleischer

Monika Fleischer is a professor at the Institute for Applied Physics and the BioNanoPhysics Center of Eberhard Karls University of Tübingen in Germany, where she also earned her PhD degree in physics. She serves as a member of the board of directors at the Center for Light-Matter-Interaction, Sensors and Analytics. Her research interests focus on nanofabrication and optical spectroscopy of hybrid nanoantenna configurations, amongst others for plasmonic sensing purposes.

---

1. Introduction

As advanced manufacturing techniques rapidly evolve, flexible mechanical sensors have emerged as a promising technology that efficiently converts mechanical stimuli, such as pressure, stress, and vibration, into electrical or optical signals.¹ These sensors, distinguished by their exceptional flexibility, effortless bending and folding capabilities, and superior biocompatibility, are well-suited for application in flexible and biological environments. In contrast to traditional, adhesive-based electrical sensors that suffer from slow response times and limited precision, optical signals inherently possess advantages such as wireless transmission and resistance to electromagnetic interference (see detailed comparison in Section 4).² This enables the straightforward implementation of non-contact, high-precision, in situ, and even in vivo mechanical sensing.³ Consequently, flexible mechano-optical sensors (FMOS) have been developed through the integration of cutting-edge fields such as materials selection, structural design, sensing mechanisms, and optical detection methods. These sensors achieve quantitative conversion of mechanical stimuli into a variety of optical signals, including luminescence, reflection color changes, light intensity variations, and multiscale spectral values.

Based on the wireless transmission characteristics of optical signals, two types of FMOS have recently attracted significant attention. These include “actively luminescent” FMOS, which rely on mechanoluminescence (ML) – the ability to emit light under mechanics, and its derivative, mechanochromic luminescence (McL) materials; and “passively non-luminescent” FMOS, which utilize artificial mechanoplasmonic structures to manipulate incident light through mechanics, further divided into near-field enhancement (MP-n) and far-field scattering (MP-f) effects. Fig. 1 depicts schematic representations of two typical FMOS: ML particles (e.g., ZnS:Mn²⁺) embedded within a polydimethylsiloxane (PDMS) matrix, and MP structures (e.g., Au metasurfaces) distributed on the PDMS surface. Although these sensors possess distinct operational mechanisms and sensing approaches, they both demonstrate remarkable flexibility, compatibility, reversibility, and sensitivity in response to mechanical stimulation. By capitalizing on the inherent benefits of optical sensing, these sensors also exhibit exceptional attributes such as being contactless, wireless, and easy-to-read. As a result, they hold great potential for applications spanning mechanical engineering, information technology, and wearable/implantable devices. Specifically, their luminescent properties indicate applications in ML structural diagnosis, while their high flexibility and biocompatibility render them ideal for use in photonic skins, wearable optoelectronics, smart healthcare, as well as plasmonic metasheets. Furthermore, the reduced sensing scale and high-level spectral manipulation of MP sensors enable nanogap antennas, strain mapping, and information encryption as nanopixels.

Fig. 1 Different types of flexible mechano-optical sensors and their versatile applications. The insert figures contain structure diagnosis (Reproduced with permission.⁴ Copyright 2023, Wiley-VCH GmbH); photonic skin (Reproduced with permission.⁵ Copyright 2018, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim); wearable optoelectronics (Reproduced with permission.⁶ Copyright 2021, Wiley-VCH GmbH); smart healthcare (Reproduced with permission.⁷ Copyright 2023, Elsevier B.V. and Science China Press); small-scale sensing (Reproduced with permission.⁸ Copyright 2013, Royal Society of Chemistry); plasmonic metasheets (Reproduced with permission.⁹ Copyright 2020, Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim); nanogap antennas (Reproduced with permission.¹⁰ Copyright 2018, Royal Society of Chemistry); strain mapping (Reproduced with permission.¹¹ Copyright 2017, Elsevier Ltd); and information encryption (Reproduced with permission.¹² Copyright 2023, American Chemical Society).

However, despite their immense potential, the diversity of FMOS also poses challenges. Currently, research in this field remains fragmented, with inadequate integration of various sensors' advantages. Furthermore, the absence of unified standards for assessing mechano-optical performance impedes the progression of FMOS toward achieving lightweight and integrated designs. Therefore, this review explores the mechanisms of photosensitive elements such as luminophores, photonic crystals, and metasurfaces, with a focus on their composition and interaction with flexible polymers and discusses how these photosensitive elements generate or manipulate light when the substrate is subjected to mechanical stimuli. Additionally, we highlight recent progress in integrating FMOS with other electrical sensors to achieve multifunctional and optical/electrical sensing capabilities,^6,13,14 like distinguishing/detecting various mechanical stimuli and environmental parameters. Lastly, by summarizing the mechanisms and development trends of ML and MP, this review aims to provide perspectives for the multifunctional integration of FMOS in the future.

2. From mechanoluminescence to inorganic elastic mechanoluminescence

Perhaps the earliest recorded observation of “mechano-optical” conversion can be traced back to 1605 when Francis Bacon noted the bright light emitted from the surface of a lump of sugar when scraped with a sharp spoon, which is now widely recognized as mechanoluminescence (ML). Despite its discovery over four centuries ago, early progress in ML research was sluggish due to limitations in luminescent material preparation, characterization conditions, and the narrow application scope of such “one-time” luminescence. It was not until 1999 that Xu et al. pioneered the development of high-brightness and recoverable ML materials, such as ZnS crystals doped with Mn²⁺ (ZnS:Mn²⁺),¹⁵ SrAl₂O₄:Eu²⁺,¹⁶ and Ca₂Al₂SiO₇:Ce³⁺,¹⁷ marking a significant breakthrough in this field. Over the past two decades, ML research has exhibited remarkable diversity. Depending on the type of mechanical force applied, ML is primarily classified into three main categories: elastic mechanoluminescence (EML), plastic ML, and fracture ML, while some literature also reported luminescence phenomena under ultrasonic and impact force conditions.^18,19 Distinct from other forms of luminescence, EML is particularly noteworthy for its high linear relationship between luminescence intensity and the magnitude of applied force, along with its exceptional stability, reversibility, and self-recoverability. Additionally, EML can also be classified into inorganic, organic,²⁰ and metal–organic framework EML based on the types of luminophores.²¹ Inorganic EML materials offer high luminescence brightness, efficient mechano-optical conversion, sensitive mechano-response, low luminescence threshold, and good stability, holding broad application prospects in various fields such as visual mechanical sensing, novel light sources, information storage, and smart electronic skins.²² Typically, inorganic ML particles are often compounded with elastomeric polymers to achieve luminescence reversibility and biocompatibility. Given the wide variety of inorganic EML phosphors, most reviews tend to emphasize the mechanisms of phosphors while neglecting the crucial impact of their compounding methods with polymers on the overall mechanical and luminescent properties. Therefore, as illustrated in Fig. 2a, this review systematically discusses the compositions and fabrication methods of inorganic EML particles, along with their blending strategies with polymeric matrices/substrates. The mechanisms of EML are reviewed from the perspectives of luminescent centers (i.e., host:dopant) and their interfacial interactions with polymers. Additionally, we review recent efforts on the development of multifunctional and even optical/electrical dual-mode sensors prepared based on different dispersion states of EML particles, and discuss the ML performance of these sensors, including intensity variations and emission color tunability, under various mechanical stimuli and external factors.

Fig. 2 Composition and mechanism of inorganic EML materials. (a) Schematic diagram. (b) Hydrothermal methods for preparing wurtzite ZnS nanorods. Reproduced with permission.²³ Copyright 2023, American Chemical Society. (c) ML mechanism for ZnS/CaZnOS heterojunction with Mn²⁺ and lanthanide (Ln³⁺) doping. Reproduced with permission.²⁴ Copyright 2020, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

2.1. EML particles in phosphors:dopants

2.1.1. Mechanism. The past two decades have witnessed a wide array of inorganic EML phosphors, including aluminates (e.g., SrAl₂O₄), perovskites (CaTiO₃), silicates (Sr₂MgSi₂O₇), gallates (SrGa₁₂O₁₉), sulfide derivatives (ZnS, CaZnOS), and fluoride derivatives (MgF₂, NaLuF₄), among others. By doping these phosphors with a small amount (e.g., 1 mol%) of transition metals (Cu²⁺, Mn²⁺, etc.) or rare earth elements (Tb³⁺, Er³⁺, Dy³⁺, Yb³⁺, etc.), these ML particles can significantly enhance luminescence intensity and offer versatile color tunability across the visible and even near-infrared regions (see Table 1 for details). In general, the luminophores are classified into two major categories: “rechargeable” and “non-rechargeable,” with aluminates and sulfides serving as representatives, respectively, based on whether UV pre-irradiation is required before ML. Notably, some ML particles classified as “non-rechargeable” require also repeatable UV excitation to achieve stable luminescent performance under dynamic stimulation, such as stretch-recovery processes.^25,26

Table 1 Recent progress in EML studies

Phosphor: dopant	Space group	Preparation method, temperature [°C], atmosphere	Polymer, blending method	Energy activation before ML?	ML peak wavelength [nm]	Highlight	Source
a HTSP: high-temperature solid-phase.
ZnS:Mn^2+	P63mc	Hydrothermal, 700, N2	PVDF-TrFE, bulk blending	No	585	Nanorod ZnS	23
ZnS:Mn^2+		Magnetron sputtering, 350, Ar	—, laminar blending	No	585	Ultrathin ML layer: 340–580 nm	27
ZnS:Mn^2+		Purchased; hydrothermal, 300, Ar	PDMS/epoxy, bulk blending	No	585	Different mechanisms	28
ZnS:Mn^2+		HTSP/^a, 1100, Ar and O2	—	No	585	Enhanced ML intensity due to O2	29
ZnS:Mn^2+ (pairs)		HTSP, 1050, thermal carbon-reduction atmosphere	PDMS, bulk blending	No	585, 650	Dual emission wavelengths	30
ZnS:Cu^2+		Purchased	PDMS, laminar blending	No	—	Multifunctional sensors	13
ZnS:Cu^2+		Purchased	PDMS, bulk blending	No	520	3D printing	31
ZnS:Cu^2+		Purchased	PVDF-TrFE, fiber-reinforced and woven blending	No	525	Multifunctional sensors	6
ZnS:Cu^2+		—/hot-injection	PDMS/EVA, laminar blending	UV irradiation (365 nm)	510–710	Doped with perovskite quantum dots to shift ML peak wavelength	26
ZnS/CaZnOS:Mn^2+	P63mc	HTSP, 1100, Ar	PDMS/PET, laminar blending	No	∼590	Enhanced ML intensity	24
ZnS/CaZnOS:Mn^2+		HTSP, 1100, Ar	PDMS/acrylic/copper electrode/PEDOT:PSS, laminar blending	No	596	Multifunctional sensors	14
ZnS/CaZnOS:Mn^2+		HTSP, 1100, N2	Urethane acrylate/acrylic/photoinitiator resin, bulk blending; cellulose, fiber-reinforced and woven blending	UV irradiation (365 nm)	∼590	3D printing	25
SrZnOS:Bi^3+		HTSP, 1050, Ar	Epoxy resin, bulk blending via blade coating	No	—	Ultrafast ML response	32
CaBaZnOS:Bi^3+	CaZnOS: P63mc	HTSP, 1000, Ar	PDMS/PET, laminar blending	No	∼550, 630	Enhanced ML intensity	33
	BaZnOS: Cmcm
Sr3Al2O5Cl2:Dy^3+	P212121	HTSP, 1200, H2/N2 (10%/90%)	PDMS, bulk blending	No	490, 580, 674	No pre-irradiation	34
Sr3Al2O5Cl2:Eu^2+,Tb^3+,Ce^3+		HTSP, 1200, H2/N2 (10%/90%)	PDMS, bulk blending	UV irradiation (365 nm)	Eu^2+: ∼625	Prominent thermal stability	35
					Tb^3+: ∼490, 550, 590, 625
					Ce^3+: ∼448
SrAl2O4:Eu^2+,Dy^3+	P21/n	Purchased	rGO/PU, bulk blend coating	—	—	Multifunctional sensors	36
Sr2P2O7:Eu^2+/3+,Y	Pnma	HTSP, 1150, H2/N2 (10%/90%)	PDMS, bulk blend coating	UV irradiation (254 nm)	Eu^2+: ∼423	Strain-induced McL	37
					Eu^3+: ∼608
SrGa12O19:Cr^3+	P63/mmc	HTSP, 1400, air	PDMS, bulk blending	UV irradiation (365 nm)	∼750	Near-infrared ML	38
Sr2MgSi2O7:Eu^2+	P421m	HTSP, 1400, reducing atmosphere	Epoxy resin, bulk blending	UV irradiation (365 nm)	—	Long afterglow	39
Sr3Sn2O7:Nd^3+,Yb^3+	A21am	HTSP, 1500, N2	PMMA, bulk blending	UV irradiation (365 nm)	∼776, 814, 889, 945	Multifunctional sensors	40
CaAl2O4:Eu^2+,Sm^3+	P21/n	HTSP, 1380, H2/N2 (5%/95%)	PDMS, bulk blending	Sunlight	∼438	Achieving stress memory	41
MgF2:Yb^2+,Mn^2+	P42/mnm	HTSP, 1160, Ar	PDMS, bulk blending	UV irradiation (—)	Yb^2+: 480	3D printing	4
					Mn^2+: 585
ZnF2/ZnO:Mn^2+ heterojunction	ZnF2: P63mc	HTSP, 600, Ar	PDMS, laminar blending	No	585	Enhanced ML intensity	42
	ZnO: P42/mnm
CsNaLuF4:Tb^3+	Cmmm	Hydrothermal coprecipitation, 295, Ar	Epoxy resin, bulk blending	X-Ray	∼342, 441, 514	3D printing	43
CaF2:Tb^3+	Fm3̄m	Melting-quenching techniques, 1200, air	PDMS, bulk blending	X-Ray	545	Tribo-ML	44
CaF2:Tb^3+	Fm3̄m	HTSP, 1250, air	PDMS, bulk blending	No	∼570	No pre-irradiation	45
NaMgNbO3:Pr^3+ heterojunction	Pbma	HTSP, 1200, —	—	UV irradiation (365 nm)	610	Enhanced ML intensity	46
LiTaO3:Tb^3+	R3c	HTSP, 1050, air	Epoxy resin, bulk blending	UV irradiation (254 nm)	∼510, 585, 644, 692	Multifunctional sensors	47
SrGa12O19:Cr^3+	P63/mmc	HTSP, 1400–1460, —	PET/PDMS, bulk blending	UV irradiation (365 nm); sunlight	750	Low load, persistent ML	48
Ca6BaP4O17:Ce^3+	C2/m	HTSP, 1280, H2/N2 (5%/95%)	PDMS, bulk blending	No	488	Self-recoverable ML	49
Na2ZnGeO4:Mn^2+	P1n1	HTSP, —, —	Epoxy resin, bulk blending	UV irradiation (254 nm)	524	Enhanced ML intensity	50

---

Research hotspots on inorganic EML have identified multiple selections in luminophores, yet they also underscore the challenge of explaining the luminescence mechanism within a unified and concise theoretical framework. Currently reported mechanisms include: charged dislocation,⁵¹ piezo-induced electron bombardment,⁵¹ tribo-induced electron bombardment,³⁸ piezo-induced carrier de-trapping,⁵² piezo-induced electroluminescence,⁵³ and interfacial interaction with their polymeric matrices (as discussed in Section 2.2). Despite this diversity, most researchers attribute the emission of EML particles to their piezoelectricity and energy trap levels. Specifically: (1) in the case of non-rechargeable luminophores, external mechanical stimuli generate a localized piezoelectrification effect due to the crystal impurities, which leads to a reduction in trap-depth and subsequent de-trapping of electrons that migrate to the conduction band (i.e., carrier releasing). These electrons then recombine with holes, releasing energy that excites dopant ions, ultimately resulting in the characteristic luminescence; (2) rechargeable luminophores, on the other hand, possess trap levels that can capture carriers under UV-irradiation, allowing for energy storage beforehand. When subjected to thermal perturbations, these carriers are released relatively slowly, producing a persistent luminescence known as afterglow. However, the external mechanics can lead to dislocations or piezoelectric fields within the EML particle, releasing carriers more rapidly and leading to a faster energy release and intense luminescence.

Piezoelectricity in inorganic materials can be explained by the displacement of ions within the crystal structure.^54,55 When subjected to pressure, changes in the atomic structure of the crystal lead to alterations in the ionic equilibrium, resulting in the generation of a dipolar moment. To achieve polarization, these dipoles cannot be canceled out by others within the unit cell. Therefore, most ML phosphors are non-centrosymmetric crystals prone to generating piezoelectric potentials. However, recent works have also shown the existence of centrosymmetric (non-piezoelectric) crystals that do exhibit intense ML,^56,57 indicating the current incompleteness of EML mechanisms. Here, we summarize the manifestations of this incompleteness in different aspects, aiming to provide readers with critical insights into the mechanisms: (1) undoped ZnS does not exhibit EML, while transition metals doped ZnS exhibits EML in the micro-crystalline but not in the single-crystal state;^23,58 (2) the contribution rates of stress concentration effects from ML particles, and triboelectric and piezoelectric effects of polymers to the luminescence intensity in the blends remain unclear; (3) EML describes a linear relationship between luminescence intensity and mechanical magnitude within a certain range, but the mechanism transitions between EML and plastic ML or fracture ML in the nonlinear region remain widely unexplored.²⁸

2.1.2. High-temperature synthesis. The luminescent properties of inorganic ML composites are intricately tied to the crystal types of the phosphors. For instance, ZnS exists in two distinct crystal structures: zinc blende (cubic phase) and wurtzite (hexagonal phase). When the temperature exceeds 1020 °C,⁵⁹ ZnS undergoes a phase transition from the cubic to the hexagonal phase, forming the wurtzite crystal with a non-centrosymmetric structure. Similarly, other prevalent ML particles, such as SrAl₂O₄, CaZnOS, perovskites, and fluorides, also necessitate synthesis temperatures above 1000 °C to facilitate the desired crystal transformation (see Table 1 for details in recent work). Consequently, the high-temperature solid-phase method has emerged as the most common synthesis method for inorganic ML particles. This method involves a series of steps, including milling, high-temperature sintering, and re-milling, to ensure thorough mixing and structural transformation of the phosphor host and dopant. It is acknowledged that ML particles synthesized via the sintering method are prone to agglomeration, and conventional physical milling often results in particles with relatively large sizes (ranging from several to tens of micrometers) and uneven dispersion. Recent research has demonstrated alternative high-temperature synthesis techniques, such as hydrothermal,^23,28,43 microwaves,⁶⁰ and spray pyrolysis.⁶¹ Notably, the hydrothermal method for synthesizing ML particles not only reduces the preparation temperature but also effectively decreases the particle size and even allows for the tailored design of ML particle structures. Wang et al. synthesized wurtzite ZnS nanorods by employing an ethylenediamine (EN)/water mixture as the solvent and thiourea and zinc acetate as the precursors.²³ Their study revealed that the type of chelating agent can influence the microstructure of ZnS; specifically, the introduction of poly(acrylic acid) (PAA) alters the morphology of ZnS into tiny nanospheres, while EN plays a crucial role in forming both the wurtzite phase and nanorod morphology. The lone electron pairs of the nitrogen atom in EN can accept protons generated from thiourea hydrolysis, thereby promoting the growth of ZnS along the (001) direction into a rod structure with a yield almost approaching unity (Fig. 2b). Ultimately, such ZnS nanorod structures exhibit ML phenomena when calcined at a temperature of only 700 °C.

During the high-temperature synthesis of ML particles, the introduction of dopants with “activator” properties can enhance the luminescent intensity of the composites by increasing trap depth or improving the crystallinity of ML particles. Lv et al. reported that the introduction of 5 mol% Zr⁴⁺ ions into SrAl₂O₄:Eu²⁺/Dy³⁺ significantly enhanced luminescent intensity.⁶² This enhancement was attributed to the proximity of the Zr⁴⁺ ion radius to that of Al³⁺, allowing to occupy the site of Al³⁺ in the SrAl₂O₄ host matrix. The resulting defects served as new electron or hole traps, capturing photo-induced electrons or holes during the UV excitation process, ultimately leading to improved luminescent performance. Similarly, the introduction of Li₂CO₃ or H₃BO₃ acts as a fluxing agent, prompting CaZnOS:Mn²⁺ to exhibit a c-axis orientation morphology with smooth surfaces, ultimately enhancing the ML intensity more than twofold.^63,64

Furthermore, blending different phosphors to form heterojunction structures during high-temperature synthesis can also substantially enhance luminescent intensity. Peng et al. proposed a heterojunctioned ZnS/CaZnOS piezophotonic system,²⁴ which exhibits significant ML intensity, being two times higher than that of ZnS and 3.5 times higher than of CaZnOS. Density functional theory calculations reveal that the high-performance ML originates from efficient charge transfer and recombination facilitated by the offset between the valence band and conduction band in the heterojunction interface region. As shown in Fig. 2c, the band offset induced by interface bonding not only lowers the electron excitation barrier from the interstitial level to the conduction band but also enhances the efficiency of electron–hole recombination from the activator (such as Mn²⁺) interfacial electronic energy levels. Consequently, the energy released from the recombination also enables the emission of tunable-color photons from Ln³⁺ dopant levels via energy transfer. The superior performance of the heterojunction architecture of ZnS/CaZnOS has been consistently validated in subsequent studies.^14,25,65,66

2.2. Preparation of ML particles in polymeric matrix/substrate

The luminescent properties of inorganic EML materials are intricately tied not only to the crystalline structure and morphology of the ML particles, but also to the composite formation and interactions between these particles and their flexible matrices/substrates, which require low modulus, stretchability, recoverability, high insulation, and importantly, transparency. Commonly employed flexible polymers encompass PDMS, polymethyl methacrylate (PMMA), polyurethanes (PU), polyethylene terephthalate (PET), epoxies, silicones, hydrogels, among others. By selectively combining these substrates or designing unique architectures, significant enhancements in the luminescent performance and multifunctionality of the resulting composites can be achieved. Taking ZnS as an example, its tendency to hydrolyze in the presence of water can be mitigated through integration with PDMS, not only restoring the ML properties but also bolstering waterproofing, chemical stability, and biocompatibility.⁶⁷ Herein, we provide a comprehensive overview of three distinct dispersion strategies for incorporating ML particles into polymeric matrices/substrates: bulk blending, laminar blending, and fiber-reinforced and woven blending. Furthermore, we delve into the tunability of properties, multifunctional sensing capabilities, and potential applications enabled by these various structural designs.

2.2.1. Bulk blending. Bulk blending represents a straightforward yet effective preparation method that involves thoroughly mixing EML particles with polymer precursors, subsequent degassing, and curing to yield homogeneously dispersed elastomeric composites. Given the substantial variations in viscoelasticity, transparency, and polarity among polymers, the same EML particles embedded in different matrices often exhibit distinct mechanical responsiveness and luminescent behaviors. For instance, as shown in Fig. 3a, Song et al. demonstrated that when ML particles (CaF₂:Tb³⁺) are combined with epoxy resin (ER), polyurethane (PU), silicone (SG), and PDMS,⁶⁸ only PDMS exhibits robust ML phenomenon. This disparity stems from the lower triboelectric potentials generated by ER, PU, and SG compared to the significantly higher potential at the inorganic–organic interface in PDMS during contact and separation processes (as discussed later). Furthermore, the authors identified both linear stress–strain and ML intensity-strain relationships for the PDMS blending system. The inset figure illustrates the luminescence phenomenon as a function of increasing strain, thereby demonstrating elastic ML behavior. Although PDMS is generally regarded as an optimal ML polymeric matrix owing to its high elasticity, transparency, and chemical stability, studies have shown that certain ML particles may perform superiorly in alternative polymers. Bian et al. replaced PDMS with PU for the luminophore of ZnS:Mn²⁺,⁶⁹ concluding that PU's internal multipolar groups and ease of binding with transition metals facilitate a more pronounced interfacial effect, leading to a 40% increase in ML intensity compared to PDMS.

Fig. 3 Performances, mechanisms, and applications of the bulk blending system. (a) ML performance of CaF₂:Tb³⁺ in powder form under milling, as bulk blends within epoxy resin under compression, and in polyurethane, silicone, and PDMS under stretching. Elastic ML behavior is demonstrated for ML intensity vs. strain. Reproduced with permission.⁶⁸ Copyright 2024, Royal Society of Chemistry. (b)–(d) Mechanism of stress concentration (Reproduced with permission.⁷⁰ Copyright 2021, American Chemical Society), contact electrification (Reproduced with permission.⁴⁵ Copyright 2024, The Authors, under the Creative Commons Attribution (CC-BY) license), and triboelectric effects (Reproduced with permission.⁷¹ Copyright 2024, The Authors, under CC-BY license) in bulk blends. (e)–(g) Applications of stress visualization in 3D printing blends, such as structural diagnosis (Reproduced with permission.⁴ Copyright 2023, Wiley-VCH GmbH), luminescent textiles (Reproduced with permission.³¹ Copyright 2022, The Authors, under CC-BY license), and precise visualization (Reproduced with permission.⁷² Copyright 2023, American Chemical Society).

Furthermore, recent years have witnessed the incorporation of a third kind (i.e., neither phosphor nor dopant) of nanoparticles (NPs) into ML bulk blends via stress concentration effects to enhance luminescent performance. For instance, Song et al. prepared ZnS:Mn²⁺/Cu²⁺@Al₂O₃/PDMS composites by incorporating SiO₂ rigid particles.⁷³ This addition not only improved mechanical properties but also induced local stress concentrations under strain, enhancing ML intensity by ∼2 times when 100 nm SiO₂ particles were added at 0.3 wt%. In addition, the inclusion of water-soluble particles like salt or sugar,⁷⁴ followed by curing and ultrasonic dispersion to create porous morphologies, has been shown to facilitate ML performance. For instance, Yin et al. reported a blended porous ML material with an initial weight ratio of ZnS : Mn²⁺ to PDMS to NaCl of 1 : 1 : 1.⁷⁵ The porous structure can effectively enhance the compressive elasticity of PDMS, thereby increasing the ML intensity and sensitivity as a pressure sensor.

Note that the stress concentration effect is also manifest in the microcrack system. Ji et al. synthesized Al₂O₃-coated ZnS:M²⁺(Mn/Cu) using solid-phase sintering and hydrothermal methods.⁷⁰ Inspired by the spider's slit organ, they adopted the stereolithography appearance technique to integrate slit microstructures into a blend of ML particles and PDMS (Fig. 3b). Leveraging the stress concentration effect at microcracks, the resulting template exhibited a 10–30 times enhancement in ML intensity under weak strain lower than 10%, effectively addressing the conflict between the high strain requirement for intense ML and the stress dispersion of soft materials.

More fundamentally, from a perspective of band transition during luminescence, the ML performance of particles in polymeric bulk blends is governed by contact electrification and triboelectric effects. Addressing the issue of luminescence degradation under dynamic/repeatable mechanical stimulation, Wang et al. reported a self-recoverable ML phenomenon in CaF₂:Tb³⁺/PDMS blending elastomer.⁴⁵ The mechanism is illustrated in Fig. 3c. Prior to blending, the potential wells of ML particles and PDMS (marked with black lines) are completely separated (Fig. 3c panel I). Upon blending and moderate stretching after curing, these wells merge into an asymmetric configuration, resulting in a reduced energy barrier. Consequently, even without interfacial friction, electrons at higher energy levels transfer from PDMS to the ML particles (marked with red arrows) to maintain energy level equilibrium, thereby generating luminescence (Fig. 3c panel II). Finally, as the blend is subjected to a critical strain, the particles and PDMS gradually separate. Due to the stronger electronegativity of CaF₂:Tb³⁺ compared to PDMS, the electrons on the surface of the phosphor can be attracted back to the PDMS by opposite charges within a short time (Fig. 3c panel III). Therefore, this contact electrification model effectively describes the stable self-recoverable ML in the CaF₂:Tb³⁺/PDMS elastomer, where charge transfer occurs during contact and charge generation occurs during separation. The work of Pan et al. further reveals that triboelectric charges at the interfaces of phosphors and the polymer are the key factor determining the ML intensity (Fig. 3d).⁷¹ For example, due to the relative triboelectric series difference between the triboelectrically positive materials ZnS and Ca_2.5Sr_2.5(PO₄)₃Cl, and the tribo-negative PDMS being 36.23 nC cm⁻² and 0.39 nC cm⁻², respectively, the triboelectricity-induced ML (tribo-ML) intensity of the ZnS/PDMS is approximately 12 times that of Ca_2.5Sr_2.5(PO₄)₃Cl/PDMS.

From an application standpoint, the bulk blending strategy facilitates not only the preparation of uniform luminescent bulk materials or thin films, but also the customization of ML materials through 3D printing and subsequent curing, leveraging the fluidity of blended precursors. Research work employed blends of MgF₂:Yb²⁺/Mn²⁺ and ZnS:Cu²⁺ with PDMS as 3D printing materials,^4,31 creating a series of hollow structural components capable of visually sensing stress overload areas (Fig. 3e and f). For instance, Guo et al.,⁴ inspired by the structures of honeycombs and artificial architectures, fabricated hexagonal and triangular ML blends that exhibited luminescence in stress-concentrated areas. In comparison to finite element analysis, this approach is notably straightforward and practical for the safety protection of mechanical structures. Similarly, Zhao et al. employed an analogous method to create auxetic and honeycomb ML blends,³¹ thereby demonstrating wearable luminescent textiles capable of visualizing sensing for finger joint movements. Notably, epoxy resins, owing to their increased hardness compared to PDMS, have the capability to produce finer 3D-printed structures, thereby enhancing sensing precision. Peng et al. demonstrated this by showcasing 3D-printed structures of human hearts and various hollow cubes using ML printing ink composed of Cs_0.02Na_0.98LuF₄:Tb³⁺ dissolved in dimethylformamide (DMF),⁷² as shown in Fig. 3g. These works underscore the versatility and potential of bulk blending approaches.

2.2.2. Laminar blending. Laminar blending refers to the process of integrating ML particles with polymer substrates or other functional thin-film materials in a laminar fashion, forming structures akin to sandwiches or multilayers. This dispersion technique is particularly prevalent in electroluminescent (EL) materials due to its convenience in inserting conductive layers. In ML applications, a common laminar strategy encapsulates ground luminescent powders between transparent polymer films, such as PET and polyvinyl chloride (PVC). While increasing the relative proportion of ML particles to flexible matrices can enhance luminescent properties in bulk blends, an excessively high ML particle content often compromises overall flexibility, rendering the material prone to fracturing under intense mechanical stimuli. Studies have shown that blending ML particles with PDMS at a high ratio like 5 : 1,³³ with PDMS serving as a binder, and sandwiching them between PET layers, can effectively mitigate powder slippage under mechanical excitation. As illustrated by Wang et al.,⁷⁶ the laminar blending strategy, despite compromising tensile properties, safeguards the luminophores and significantly enhances the bending durability and abrasion resistance of composites, demonstrating its outstanding performance in triboluminescent sensing and handwriting authentication (Fig. 4a). Furthermore, their study showed that even with the laminar blending approach, a comparable elastic ML behavior was observed where the luminescent intensity increased linearly with increasing pressure.

Fig. 4 Performances, enhancements, and applications of the laminar blending system. (a) ML performance in a sandwiched laminar system, indicating bending durability and handwriting authentication capability. Reproduced with permission.⁷⁶ Copyright 2016, Elsevier B.V. (b) and (c) ML enhancement by incorporating piezoelectric PVDF (Reproduced with permission.⁷⁷ Copyright 2019, Elsevier Ltd) and ZnO film (Reproduced with permission.²⁷ Copyright 2024, Elsevier B.V.) through laminar dispersion methods. (d) Optical/electrical dual-mode sensing. Reproduced with permission.⁷⁸ Copyright 2022, Elsevier Ltd. (e) Temperature/force multifunctional sensing. Reproduced with permission.¹⁴ Copyright 2022, American Chemical Society. (f) and (g) Multifunctional TENG-ML systems with single-tribolayer (Reproduced with permission.⁷⁹ Copyright 2020, The American Association for the Advancement of Science) and double-tribolayer (Reproduced with permission.⁸⁰ Copyright 2022, Elsevier Ltd).

In Section 2.1.1, we discussed how the piezoelectricity of asymmetric piezoelectric materials leads to band transitions, carrier release, and subsequent luminescence. When incorporating piezoelectric thin films with ML layers (i.e., piezo-ML), external piezoelectric fields can be induced. These fields have two potential effects: they may further bend the energy band, facilitating the escape of electrons from traps located near the conduction band and thereby easing ML emission (akin to the behavior observed in heterojunction structures in Fig. 2c).⁸¹ Alternatively, they may elicit an additional EL response, enhancing the overall luminescent intensity. β-phase polyvinylidene fluoride (PVDF) is a prototypical piezoelectric polymer, which is commonly synthesized through stretching, electric field poling, or high-pressure treatment.⁸² Wang et al. reported a multilayer luminescent enhancement device fabricated via hot-press encapsulation, comprising a ZnS:Mn²⁺/Cu²⁺ ML film, a PVDF film, and ethylene-vinyl acetate (EVA)/PET encapsulation layers (Fig. 4b).⁷⁷ The composite device demonstrated an ∼85% increase in ML intensity relative to a single-layer ML film. This enhancement was attributed to the external piezoelectric field (E_EPF) of PVDF under pressure, which reduces the overall trap depth. This reduction facilitates the migration of more charges to the bands, thereby increasing the energy levels capable of exciting additional Mn²⁺ within ZnS.

Furthermore, the laminar dispersion method finds application in ultrathin luminescent materials. Wang et al. employed magnetron sputtering to deposit a 50 nm thick ZnO piezoelectric layer onto a quartz glass substrate,²⁷ followed by ZnS:Mn²⁺ films of varying growth times of 45, 60, and 75 minutes, corresponding to thicknesses of approximately 340, 450, and 580 nm, respectively. Results indicated that ML intensity increased with film thickness, though no ML was observed for the 45-minute film (Fig. 4c). The spectral fluctuations also suggested an overall low intensity, potentially due to the quenching effects from thin film thickness or small particle size during depositing. However, the advanced thin-film preparation techniques, including magnetron sputtering, chemical vapor deposition, and atomic layer deposition, enable precise control of film thickness and tailored macroscopic mechanical properties, representing a promising direction for ML materials research.

Crucially, laminar blending offers a versatile approach to effectively stack functional layers with distinct electrical sensing capabilities both on and beneath the ML layer. This technique is pivotal for developing composite sensors capable of multifunctional environmental factors sensing, such as detecting various mechanical stimuli, temperature, and humidity, while also enabling simultaneous optical/electrical dual-mode output. Electrical sensors, such as resistive or capacitive ones, are common choices for the layers to be integrated. For example, Zhou et al. reported a transparent and high-stretching polyacrylic acid/cellulose nanofibrils-glycerol hydrogel with enhanced ionic conductivity from LiCl,⁷⁸ enabling strain and pressure sensing via electrical resistance changes. Benefiting from the high transparency of this hydrogel, when combined with ZnS:Cu²⁺/PDMS in a bilayer structure, this hybrid sensor ensures the simultaneous acquisition of changes in resistance signals and luminescent trajectories during pressing and sliding in various directions (Fig. 4d). In addition, Ma et al. proposed a laminar structure incorporating a poly(3,4-ethylenedioxythiophene):poly(styrene sulfonate) (PEDOT:PSS) film and a copper interdigital electrode,¹⁴ sandwiched by heterojunctioned ZnS/CaZnOS ML layers. The temperature-resistant PEDOT:PSS film exhibited a high sensitivity (−0.6%/°C from 21 to 60 °C), while the ML layers visualized applied forces (≥2 N) (Fig. 4e). Thus, the pressure/temperature multifunctional sensor was established by distinguishing the optical/electrical signals, representing a significant advancement in flexible tactile sensors.

On the other hand, electrical sensors, such as “self-powered” piezoelectric or triboelectric nanogenerators (PENG and TENG), can be also integrated with ML layers via laminar blending. Given that PVDF and ZnO are piezoelectric materials capable of generating electrical signals under dynamic pressure, the aforementioned piezo-ML system can be further enhanced by integrating with electrodes and wires, enabling the conversion of pressure into “self-powered” electrical and “contactless” optical signals, thereby achieving PENG-ML dual-mode sensing capability. Unlike piezoelectricity, triboelectric signals primarily stem from electrostatic induction and triboelectrification, requiring dynamic contact/separation between two friction layers with different triboelectric polarities.⁸³ The triboelectric polarity of the ML layer is mainly determined by the flexible substrate: common materials such as PDMS and PVDF exhibit tribo-negative characteristics. Given the tribo-positive nature of human skin, it is feasible to develop a TENG-ML sensor through the laminar integration of the luminescent layer, electrode, and polymer encapsulation, thereby preserving an overall tribo-negative state. This single-tribolayer TENG-ML sensor facilitates the simultaneous characterization of hybrid TENG and ML signals in response to a single mechanical stimulus, revealing its potential applications in optoelectronic skin and tactile sensing technologies. Zhao et al. demonstrated such designs by incorporating a ML layer (ZnS:Cu²⁺ mixed with various polymer binders), an aluminum electrode/shielding layer, and PI/PDMS packaging (Fig. 4f).⁷⁹ Quantitative research revealed that Ecoflex as the binder yielded superior optical signals compared to ethyl cellulose, methylcellulose, and polyvinyl alcohol. This composite sensor exhibited touch-light emission at a low trigger pressure threshold of 20 kPa and demonstrated the ability to simultaneously generate ML signals representing trajectory and respond to triboelectric signals when a finger slides over it. A similar single-tribolayer TENG-ML system is demonstrated by Liang et al.,⁸⁰ where they presented an all-polymer design that encompasses a laminar structure consisting of a ZnS layer, ionic conductor composed of NaCl-polyacrylic acid (PAA) hydrogel, and Ecoflex elastomer layer. Further to this, Zhang et al. reported a double-tribolayer TENG-ML sensor with ZnS:Cu²⁺ layer serving as the tribo-negative layer and Ag nanowires as the positive one.⁸⁴ This sensor was capable of distinguishing different mechanical stimuli such as pressing and stretching based on varying optical and electrical signals (Fig. 4g), while exhibiting stable ML stability over 3000 stretching-releasing cycles.

In summary, the laminar blending of ML films with other functional layers exhibits optical/electrical dual-mode output characteristics for a single stimulus, as well as multifunctional discrimination capabilities towards mechanical stimuli and environmental parameters. These composite sensors qualitatively distinguish and visualize force trajectories, and quantitatively analyze force magnitudes through optical/electrical intensity variations,⁸⁵ holding immense potential in wearable sensors and human–machine interaction, such as gesture control, augmented reality, and smart prosthetics.

2.2.3. Fiber-reinforced and woven blending. Inspired by the fiber and woven structures in textiles, recent research has explored a novel composite filling system for ML particles, distinct from bulk and laminar blending approaches. The integration of nanofibers with high aspect ratios and ML particles offers unique reinforcements, bolstering the overall mechanical stability of the composites such as resistance to deformation and tearing. Additionally, the high specific surface area of nanofibers promotes interfacial interactions with the luminescent layer, thereby enhancing optical performance. Jeong et al. pioneeringly reported a robust ML system consisting of fiber-reinforced ZnS/PDMS (Fig. 5a).⁸⁶ The core of this system was a cross-shaped fiber fabricated via the melt-spinning process, utilizing a combination of thermoplastic polyurethane (PU) and crosslinked silicon rubber as the raw materials. Compared to conventional circular cross-section fibers,⁸⁷ the cross-shaped design increases its contact area with the ML layer and effectively prevents the formation of perpendicular cracks during stretching. Liang et al. advanced the fiber-reinforced strategy by developing a core–shell stretchable sensor incorporating ZnS:Cu²⁺ coating and an elastic optical fiber (Fig. 5b).⁸⁸ The highly transparent and flexible core fiber, composed of silicone elastomer optical encapsulant and PDMS, enabled long-distance transmission of ML signals generated by human mechanical movements. This capability, along with accurate and quantitative analysis of signal amplitude, holds promise for wearable applications. Beyond core–shell structures, Song et al. introduced silver-coated nylon fibers into ZnS/PDMS bulk composites (Fig. 5c).⁸⁹ The embedded conductive fibers not only strengthened the mechanical stability but also enabled the application of an alternating electric field to the ML system, opening avenues for EL applications.

Fig. 5 Reinforcements, applications, and developments of the fiber-reinforced and woven blending system. (a)–(c) ML performance using fiber-reinforced strategies, leveraging the fibers’ large surface contact area (melt-spinning fiber, reproduced with permission.⁸⁶ Copyright 2017, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim), transmission characteristics (optical fiber, reproduced with permission.⁸⁸ Copyright 2021, The Authors, under CC-BY license), and conductive characteristic (conductive fiber, reproduced with permission.⁸⁹ Copyright 2019, Elsevier Ltd). (d) and (e) Wearable applications enabled via ML nanofiber biaxial (Reproduced with permission.⁶ Copyright 2021, Wiley-VCH GmbH) and weft-knitted woven blending (Reproduced with permission.⁹⁰ Copyright 2023, Tsinghua University Press). (f) and (g) Multifunctional TENG-ML sensors (Reproduced with permission.⁹¹ Copyright 2021, Elsevier Ltd) and tribo-ML enhancements (Reproduced with permission.⁹² Copyright 2019, American Chemical Society) realized by woven blending of fiber-like ML rods with an additional friction layer in coaxial yarn.

Furthermore, fiber-reinforced ML composites, characterized by their robust luminescence, high aspect ratios and mechanical strength, have prompted the exploration of a woven blending strategy. This approach not only preserves the advantages of fiber reinforcement but also significantly enhances breathability and elasticity for wearable applications. Yang et al. designed a ML composite featuring a helix core–shell yarn structure (Fig. 5d).⁶ The core layer was prepared using Fermat-twisted spinning to encapsulate nickel–copper fiber electrodes with polyvinylidene fluoride-trifluoroethylamine (PVDF-TrFE) nanofibers, while the sheath layer comprised ZnS:Cu²⁺@Al₂O₃ and poly(styrene-b-(ethylene-co-butylene)-b-styrene) (SEBS) via melt-spinning. The insert band diagram reveals that the piezoelectric PVDF-TrFE nanofiber, in conjunction with the sliding friction between the sheath and core layers, induces a hybrid piezo-tribo-ML effect. This effect results in a further tilting of the ZnS energy band, leading to an increased excitation and subsequent relaxation of electrons to the luminescence emitter. Consequently, the ML intensity is enhanced by 30% compared to the piezo-ML system. Besides, the spatial charge polarization of the phosphor also enlarges the charge transfer behavior at the friction interface, achieving a maximum improvement in self-powered TENG performance of about 100%. By biaxially weaving the TENG-ML fibers and spandex as the warp and cotton as the weft, the resulting textile generates synergistic optical/electrical dual-mode signals under mechanical stretching. Distinct from the TENG-ML system in laminar states, this system exhibits significantly improved wearability and comfort. Furthermore, Chang et al. fabricated ML fibers with a thin PU core and ZnS:Cu²⁺/PDMS shell using adhesion coating technology,⁹⁰ exhibiting 6.5 times stronger luminescence intensity than traditional dip-coated ML fibers (Fig. 5e). Their work also demonstrated an elastic ML behavior, akin to bulk and laminar blending, where the ML intensity exhibited a linear response to the applied strain of 10–100%. Besides, the fiber possessed also a small diameter and smooth surface, enabling the development of a continuous synthesis technology for meter-scale ML fibers. Compared to conventional biaxial ML fabrics, the weft-knitted textile produced using this technology exhibited lower stress during deformation and enhanced comfort, marking progress in integrated ML woven fabric applications.

In the realm of TENG-ML applications, the woven hybrid morphology is also applicable to some high aspect ratio and non-fibrous ML structures, as well as friction materials with robust triboelectric polarity. Compared to the laminar systems, the woven state effectively increases the dynamic contacting area between the two friction layers, enhancing both tribo-ML effects and wearability. He et al. introduced a flexible and stretchable coaxial TENG yarn, using a coil spring as the inner support/conductive core, and layer-by-layer dip-coated ZnS:Cu²⁺/PDMS as the outer friction sheath.⁹¹ By weaving this coaxial TENG yarn with wool or polytetrafluoroethylene (PTFE), a dual-mode TENG-ML fabric capable of simultaneously generating both ML and triboelectric signals was realized to monitor physical movement in the human body. For example, Fig. 5f illustrates that, as the degree of finger joint flexion increases, the hybrid sensor exhibits both an elevation in open-circuit voltage and an enhancement in luminescence intensity, underscoring the advantages of self-powering and dual-mode sensing characteristics. Similarly, Park et al. fabricated a biaxial weave composite consisting of luminescent ZnS:Cu²⁺/PDMS and PTFE,⁹² which provided continuous surface friction between warp and weft while enhancing tensile durability in various directions (Fig. 5g). Comparative experiments demonstrated that under 30% strain, the woven structure composed of ZnS:Cu²⁺/PDMS and PTFE exhibited approximately double the luminescent intensity compared to that composed of biaxially alternating ZnS:Cu²⁺/PDMS. In line with the mechanism posited by Yang et al.,⁶ the authors attributed the enhanced ML of the PTFE system to triboelectricity effects, which further tilt the energy bands of ZnS. Notably, they considered the ML mechanism of ZnS:Cu²⁺ with PDMS to stem from triboelectricity-generated EL, a perspective that diverges from the more widely accepted theory of localized piezoelectrification effects, as discussed in Section 2.1.1.

2.3. ML characterization

2.3.1. ML intensity regulations. As mentioned in Section 2.1.1, the ML process is fundamentally tied to the de-trapping of charge carriers within the bandgap, where the strain-induced internal electric potential facilitates the release of these carriers under mechanical loading. The regulation of ML intensity is governed by the quantity and rate of carrier release, which in turn are modulated by trap depths, densities, and the local piezoelectric potentials. In the realm of ML applications, the ability to achieve real-time, in situ, and readily distinguishable visualization of mechanical stimuli holds the potential to significantly expand the utility of ML materials. With this in mind, we now delve into two key aspects of ML intensity regulation: (1) strategies for enhancing emission and (2) mechanical stimulation for controlling intensity.

(1) ML intensity enhancement strategies

In Fig. 6, we expand upon and summarize the ML intensity enhancement strategies previously discussed. The strategies are categorized into several key aspects:

Fig. 6 ML intensity enhancement strategies via structural modifications of (a) luminophores (Reproduced with permission.⁹³ Copyright 2021, Wiley-VCH GmbH) and (b) polymer matrices, as well as changes in (c) environmental (Reproduced with permission.⁸⁹ Copyright 2019, Elsevier Ltd) and (d) mechanical stimuli (Reproduced with permission.³¹ Copyright 2022, The Authors, under CC-BY license).

Firstly, (a) structural modulation of luminophores is explored. This involves several key strategies: adjusting the concentration of doping elements (such as transition metals, rare earths, and fluxing agents),^63,64 incorporating heterojunction structures (e.g., CaZnOS, ZnF₂/ZnO),^24,42 and modifying the structural morphology (e.g., nanorod crystals, core–shell structures).²³ For instance, Lange et al. deposited a 3 nm thick layer of Al₂O₃ onto the surface of ZnS particles using the rotary atomic layer deposition (ALD) technique.⁹³ This ALD-treated ZnS effectively prevents the formation of photo-corrosive ZnO under UV irradiation. Microscopic and spectroscopic analyses of particle cross-sections provide further evidence of a protective layer that uniformly covers the ZnS surface. The inset figure illustrates the integrated photoluminescence (PL) peak intensities of ZnS particles before and after ALD treatment at temperatures of 100 °C, 150 °C, and 200 °C. Compared to untreated ZnS, the coating applied at 100 °C enhances photostability. Notably, at 150 °C or 200 °C, the treatment successfully inhibits the formation of ZnO, resulting in the highest PL intensity. Given the widespread use of UV-pre-irradiation in ML process, this sheathing approach offers valuable insights into enhancing ML intensity.

Secondly, (b) regulation of the interaction between luminophores and polymers is addressed, focusing on polymer types and blending strategies. Key considerations include the concentration of luminophores (excessive concentration compromises overall flexibility) and the nature of their interaction, which encompasses electrification,⁶⁵ local piezoelectric potentials in piezoelectric polymers (such as PVDF),⁷⁷ and triboelectric potentials between luminophores and polymers.⁶⁶

Additionally, the influence of (c) environmental stimuli is examined. This includes light (e.g., PL effects or pre-irradiation of ML materials)²⁶ or thermal stimuli (e.g., thermoluminescence). As illustrated by Chen et al.,³² during the ML process, some parts of excited charge carriers are temporarily trapped during migration and thermally released to produce persistent emission. This phenomenon leads to a ML dependence on temperature, as detailed in Section 2.3.3. Electrical stimuli, such as electroluminescence (EL), are also encompassed within this category. The piezoelectric or triboelectric potentials generated by polymers are also pulsed alternating currents, which could be classified under alternating currents EL (ACEL). Some researchers have also used EL phenomena to elucidate the mechanisms of ML.^92,94Fig. 6c illustrates the work of Song et al. (the structural design is depicted in Fig. 5c),⁸⁹ wherein they demonstrated that the direct application of an external AC field (300 V, 1 kHz) to ML materials enhances the total luminescence.

Lastly, (d) mechanical stimuli are considered. These encompass stress concentration effects (e.g., microcracks, incorporation of third-party NPs like SiO₂),^68,73 the nature of mechanical action (e.g., stretching, friction, compression),⁹⁵ the magnitude of the applied force,⁴⁷ and the frequency of mechanical stimulation. Zhao et al. conducted cyclic luminescence tests on the ZnS:Cu²⁺/PDMS system during stretching-releasing processes.³¹ The inset figure reveals that the emitted ML intensity increases with the stretching-releasing rate, ranging from 100 to 300 cycles per minute. Furthermore, their work demonstrates that the system maintains stable ML intensity over 5000 stretching-releasing cycles (data not shown), providing a strong practical foundation for the applications.

(2) ML intensity regulations via mechanics

In the ML process, the quantity and rate of carrier release not only dictate the emission intensity but also influence the triggering and saturation thresholds, as well as the sensitivity and linearity when utilized as FMOS. Evidently, the carrier release process is modulated by the type and magnitude of the mechanical stimulus applied. For instance, Fig. 7a illustrates the ML performance of ZnS:Mn²⁺/PDMS blends under stretching (50% strain), rubbing (1 N load), and compressing (300 N load).⁹⁵ The differing ML performances in response to different types of external forces have also been demonstrated in numerous research studies.^4,37 Notably, the ML emission intensity of luminophores at identical filling concentrations varies and does not adhere to a linear relationship with concentration. These behaviors can potentially be attributed to disparities in micro-stress transfer processes under distinct stimuli, which subsequently affect the quantity and rate of carrier release.

Fig. 7 ML intensity regulations via (a) mechanical stimuli types, such as stretching, rubbing and compressing (Reproduced with permission.⁹⁵ Copyright 2021, American Chemical Society), as well as (b) force magnitudes in uniaxial stretching, based on (c) experimental studies. Reproduced with permission.⁴⁷ Copyright 2024, Wiley-VCH GmbH.

In general, for homogeneous bulk blends composed of ML particles and their matrix, sustained mechanical stimulation (e.g., with a constant loading rate) leads to a linear relationship between luminescence intensity and the applied force within a specific range. This phenomenon is termed EML. To further investigate this relationship, we constructed a comprehensive diagram, based on the classical discussion by Zhang et al. and the experimental work by Yang et al.,^47,96 illustrating the relationship between luminescence intensity and force magnitude for EML composites under uniaxial tension. As shown in Fig. 7b, the ML intensity exhibits a typical trigger threshold when the stress (or stretching time) is too small. This is primarily due to the inability of the minimal force to transfer from the substrate to the luminescent crystals, coupled with factors such as insufficient local piezoelectrification near the luminophores to induce electron de-trapping from filled electron traps. However, as the stress continues to increase, a significant linear relationship between luminescence intensity and stress emerges. Particularly, the force at this stage induces a stable release of charge carriers at a constant rate. Therefore, the strict definition of EML refers to the linear relationship between luminescence intensity and concentration of released charge carriers. As strain further increases, yielding, softening, and saturation phenomena in ML intensity can be observed. These are attributed to: (1) the mechanical response of the polymer shifting from elastic deformation to plastic deformation at the yield point, with “segmental motion” of the polymer leading to inefficient micro-stress transfer to the ML particles; and (2) a reduction in the number of carriers allowed to de-trap within the ML particles. Note that, according to principles of polymer physics,⁹⁷ a polymer substrate that has been deformed beyond its elastic limit (or yield stress) may not recover its initial shape after the removal of external force, but complete recovery can be achieved by heating above its glass transition temperature, allowing the chain segments to regain mobility and relax back towards their original configuration. Ultimately, upon stress release or material fracture, EML materials often exhibit a brief afterglow effect. Interestingly, as shown in Fig. 7c, Yang et al. also experimentally revealed that as the rate of force application increases,⁴⁷ not only does the slope of the luminescence intensity in the linear region increase (indicating that a smaller force can generate greater ML intensity), but the slope in the “yield-saturation” region also correspondingly increases. This finding provides new insights into understanding the mechanical-optical response characteristics of EML materials.

2.3.2. ML color regulations. The remarkable luminescent performance of ML materials is manifest in two aspects: firstly, by enhancing the luminescent intensity under identical mechanical stimuli through lowering the activation threshold or increasing sensitivity; secondly, by exhibiting the capacity to regulate the luminescent color, a phenomenon known as mechanochromic luminescence (McL).⁹⁸ Akin to intensity variations, color manipulations in ML serve as vital carriers of mechanical information, often bolstering the precision and visualization capabilities of FMOS sensing applications. Here, we review recent progress in McL studies and categorize them into two main areas: (I) “static” color regulation achieved by manipulating inherent factors such as ML composition or crystal structure; and (II) “dynamic” color regulation facilitated by modifying the mode, frequency of mechanical stimulation, or environmental conditions.

(1) Static regulations for achieving McL

In general, the emission colors of ML are dependent on substitutional dopants, which serve as emitting centers. Upon mechanical stimulation, these emitting centers, characterized by distinct energy levels, are capable of capturing charge carriers and subsequently recombining with the ground state to generate luminescence. Therefore, static regulations concerning the color design before or during the fabrication process, encompass the type and concentration of luminophores, synthesis atmosphere, and blending of multiple dopants or phosphors. It is evident that for virtually all phosphors, the ML color can be tailored by altering the dopant species. For instance, ZnS/CaZnOS:Mn²⁺ doped with lanthanide (such as Pr³⁺, Sm³⁺, Eu³⁺, Tb³⁺, Dy³⁺, and Ho³⁺) achieves ML colors spanning green, yellow, and red-light bands, attributed to the rich energy levels of rare-earth elements.²⁴ Similarly, modifying the composition of the phosphor also leads to color variations in ML composites. Chen et al. prepared a bilayer ML laminar system by spin-coating a layer of perovskite CsPbX₃ (X = Cl, Br, and I) quantum dots (QDs)/EVA onto the surface of ZnS/PDMS and reported a broad color range with narrow emission bands by tuning the halide composition of CsPbX₃.²⁶ The color conversion between ZnS and perovskite QDs involves radiative combination and reabsorption, enabling full-color tuning from 510 to 710 nm due to their tunable bandgap characteristics.

Furthermore, modulating the dopant concentration can alter the ML color. Golovynskyi et al. reported that in a bulk ML system consisting of ZnS/CaZnOS:Mn²⁺ embedded in PDMS,⁶⁵ the luminescent spectrum under a constant 45 N external force exhibits a single peak that shifts from 580 nm to 630 nm as the Mn²⁺ concentration increases from 0.2% to 8% (Fig. 8a). In addition, doping a single phosphor with two or more dopants in specific proportions offers an effective means of regulating color. Since different dopants typically correspond to luminescence peaks at specific wavelengths, adjusting their component ratios enables the coupling (or mixing) of individual colors with varying intensities, rather than shifting ML emission wavelengths. Guo et al. reported on a dual-doped ML composite of MgF₂:Yb²⁺/Mn²⁺ blended into PDMS,⁴ exhibiting two spectral peaks at 480 (Yb²⁺) and 585 nm (Mn²⁺) in the ML spectrum under an applied stress of 20 N (Fig. 8b). With increasing Mn²⁺ ions concentration, the emission intensity of Yb²⁺ monotonically decreases, while that of Mn²⁺ simultaneously increases due to their competitive relationship. Consequently, the corresponding ML color changed from cyan to orange as the concentration of Mn²⁺ increased from 0 to 3%.

Fig. 8 ML color regulations via (a)–(d) static factors, including luminescent peak shifts (Reproduced with permission.⁶⁵ Copyright 2024, Elsevier B.V.) and intensity competitions (Reproduced with permission.⁴ Copyright 2023, Wiley-VCH GmbH) related to dopant concentration, high-temperature sintering atmospheres (Reproduced with permission.³⁰ Copyright 2024, Wiley-VCH GmbH) and blending strategies (Reproduced with permission.⁶⁶ Copyright 2022, Wiley-VCH GmbH). ML color regulations via (e)–(h) dynamic factors, such as stretching-releasing rate (Reproduced with permission.⁹⁴ Copyright 2013, American Institute of Physics), magnetic field (Reproduced with permission.⁹⁹ Copyright 2017, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim), pressure (Reproduced with permission.¹⁰⁰ Copyright 2021, Elsevier Ltd) and temperature (Reproduced with permission.³⁵ Copyright 2022, Elsevier Ltd).

Interestingly, adjusting the synthesis atmosphere during high-temperature solid-phase preparations can modify the crystal structure and thus alter the ML color. Wu et al. studied the ML behaviors of the resultant Sr₃Al₂O₆:Eu/PDMS blending composites,¹⁰¹ where the ML color can be manipulated by adjusting the synthesis atmosphere of the luminophore based on its unique self-reduction characteristic. Specifically, under an air atmosphere in solid-phase reactions, Eu remains in an oxidized state (Eu³⁺) and exhibits a red ML emission. In contrast, in an inert nitrogen atmosphere (with no oxidation or reduction), some Eu³⁺ ions reduce to Eu²⁺, resulting in yellow ML. Furthermore, this self-reduction can be further promoted by introducing a hydrogen reduction atmosphere, leading to an increase in Eu²⁺ ions and ultimately resulting in green ML. Such color tunability can be further adjusted by altering the ratio of nitrogen and hydrogen in the atmosphere. Similarly, Liu et al. reported that the traditional yellow ML emission of ZnS:Mn²⁺ can be adjusted to red ML by introducing a thermal carbon-reduction atmosphere (TCRA),³⁰ which leads to changes in the crystal structure, i.e., the formation of appropriate sulfur vacancies in the ZnS lattice (Fig. 8c). Furthermore, when compared to a pure nitrogen environment, the TCRA-treated sample exhibited a 3.23-fold increase in ML intensity and demonstrated obvious afterglow phenomena. This approach by adjusting the atmosphere during synthesis offers tunable flexibility compared to methods by changing the species/ratio of phosphor and dopant before synthesis.

Finally, blending different types of phosphors can also achieve color modifications. Since different phosphors often exhibit varying degrees of response to different mechanical stimuli, this approach may enable multifunctional FMOS capable of responding to multiple ML colors under various stimuli, demonstrating promising potential for user-interactive applications. For instance, Wei et al. introduced a hybrid ML system comprising ZnS:Cu²⁺ and ZnS/CaZnOS:Mn²⁺ embedded in PDMS,⁶⁶ which shows the capability to emit an orange color response to shear forces, such as fingertip slipping, with a lower limit of 2 N; and a green color response to applied strains spanning from 30% to 70% (Fig. 8d). In addition, incorporating organic luminophores into inorganic ML materials can be an alternative way. Zhang et al. introduced an organic luminophore, bis(adamantyl)-1,2-dioxetane (Ad), into the ZnS/PU blending system.¹⁰² The luminescence mechanism of Ad involves bond scission, followed by fluorescence resonance energy transfer from the excited ketone to 9,10-diphenylanthracene as the acceptor, ultimately resulting in a blue ML signal. When subjected to tensile testing, this hybrid ML material exhibited green emission dominated by ZnS under elastic strain, and blue emission dominated by Ad upon fracture. This approach successfully broadened the range of strain–stress response, endowing the dual-ML responsive FMOS with enhanced sensitivity for applied strain.

(2) Dynamic regulations for achieving McL

Dynamic regulation strategies concern rapid or instantaneous approaches to manipulate ML color by mechanical force, and environmental factors, such as magnetic field, temperature, pressure, and others. The key to achieving dynamic regulation of McL lies in the precise manipulation of the local piezoelectric potential of luminophores via these mechanical or environmental stimuli. In a pioneer study conducted by Jeong et al.,⁹⁴ the ML properties of ZnS:Cu/PDMS composites were investigated under varying stretching–releasing (S–R) rates. Their findings revealed that as the S–R rate increased from 200 to 500 cycles per minute, the ML intensity surged approximately 2.75-fold, accompanied by a subtle blueshift in the ML peak position from 520 to 517 nm. Extending their investigation to the EL characteristics of this system, the authors observed a pronounced blueshift in the EL peak position from 517 to 455 nm as the applied alternating current frequency rose from 10 Hz to 10 kHz (Fig. 8e). Based on these observations, they proposed that the band bending phenomena occur within both the conduction and valence bands of the ML system, and the enhanced S–R rate facilitates the recombination of electrons in the conduction band with holes in the valence band at lower-lying energy levels, ultimately resulting in the observed spectral blueshift. In conjunction with Fig. 6d, it can be concluded that the rate or frequency of mechanical stimulation exerts a dual influence: it not only augments the ML intensity but also slightly alters the emission color.

In addition, luminescence colors can be modulated by various environmental factors. Wong et al. fabricated a composite material comprising an ML system (ZnS:Cu/Al dispersed in PDMS) embedded into a magnetic precursor (ferromagnetic microparticles blended with pre-cured PDMS),⁹⁹ which were subsequently cured to ensure stability and integrity. By applying an external magnet under the ML composite, the composite exhibited a temporal and remote tuning of light-emitting wavelength and color by modulating the frequency of magnetic-field excitation. As the frequency increased from 50 Hz to 470 Hz, the highest ML peak wavelength shifted from 503 nm to 472 nm (from green to blue emission), demonstrating an in situ color tunability (Fig. 8f). The underlying mechanism is attributed to the tunable piezo-phototronic emission, which arises from the band structure tilting of the ZnS phosphor induced by magnetostrictive strain under high-frequency magnetic field excitation. Zhang et al. used the dynamic diamond anvil cell approach introduced by Evans et al., to investigate the ML performance of ZnS:Mn²⁺ dispersed in silicone oil under high pressure.^100,103 As the pressure increased to 7.3 GPa, the luminophore began to emit orange ML, which gradually intensified and eventually turned dark red, accompanied by a spectral redshift of ∼45 nm. The highest intensity was observed at an intermediate pressure of 3.6 GPa, indicating that ML intensity does not always increase with the intensity of mechanical stimuli (Fig. 8g). Additionally, during decompression, the luminophore exhibited a continuous blueshifting ML emission. Their work demonstrated that the ML intensity is dependent on pressure changes rather than magnitudes, while the emission wavelength of ML is governed by the absolute pressure level.

Furthermore, the introduction of ML particles with different thermal stabilities enables the regulation of emission color by temperature. Generally, luminophores such as the ZnS series exhibit typical thermal quenching, where their ML intensity gradually decreases as the ambient temperature rises. However, research reported by Bai et al. indicated that Sr₃Al₂O₅Cl₂:Ln (Ln = Eu²⁺, Tb³⁺, Ce³⁺) doped into PDMS exhibits high thermal stability under stretching conditions for ML.³⁵ The reason is that, although the triboelectric potential of PDMS gradually decreases with increasing temperature (which is detrimental to luminescence), the elevated temperature also significantly increases the probability of electron bombardment under the triboelectric field, thereby contributing to enhanced luminescent efficiency. Therefore, when combining Sr₃Al₂O₅Cl₂:Ln and ZnS:Cu²⁺ into the PDMS blending system, as shown in Fig. 8h, under an increased temperature, the orange luminescence from Sr₃Al₂O₅Cl₂:Ln remained almost unchanged, while the green luminescence from ZnS:Cu²⁺ gradually weakened. This allowed for the regulation of the emission color by temperature based on the color difference.

2.3.3. Further developments. In Section 2.2, we focus on the multifunctional and optical/electrical dual-mode sensing enabled by bulk, laminar, and woven blending strategies of EML phosphors with various polymers, highlighting a crucial trend in the integration and wearable application development of ML in recent years. Furthermore, this review summarizes several less-explored ML systems that ensure potentially impactful ML directions in Fig. 9.

Fig. 9 Further developments in (a) NIR ML (Reproduced with permission.¹⁰⁴ Copyright 2023, Royal Society of Chemistry), (b) ultrafast ML (Reproduced with permission.³² Copyright 2023, Wiley-VCH GmbH), and (c) organic ML (Reproduced with permission.¹⁰⁵ Copyright 2021, Wiley-VCH GmbH).

(a) Near-infrared (NIR) ML. In implantable scenarios such as biomechanical detection, there is an urgent need for luminophores capable of penetrating biological tissues. Given that tissues are transparent (exhibiting weak light absorption) yet cells within scatter light (through Rayleigh or Mie scattering, contingent on cell size as elaborated in Section 3.2), long-wavelength NIR light experiences reduced scattering and enhanced transmission. Consequently, NIR (700–2500 nm) ML emerges as a promising novel optical probe for real-time, in situ, and in vivo imaging. Additionally, it effectively minimizes interference from indoor lighting, and, due to the human eye's significantly lower sensitivity to NIR radiation compared to visible light, NIR ML holds potential applications in anti-counterfeiting and bright-field stress sensing. However, as described by Xiong et al.,¹⁰⁶ progress in NIR ML research has been sluggish, hindered by challenges in spectral detection and unclear emission mechanisms. From the perspective of NIR ML in implantable devices, assessing the ease of ML excitation is paramount. Many reported luminophores necessitate UV pre-excitation, such as MgGeO₃:Mn²⁺ and SrGa₁₂O₁₉:Cr³⁺.^48,107 In contrast, Xiong et al. reported a non-pre-excitation NIR ML in LiGa₅O₈:Cr³⁺, wherein Cr³⁺ substitutes into the six-coordinated Ga³⁺ site, generating a 716 nm ML emission via its ²E → ⁴A₂ spin-forbidden transition.¹⁰⁸ Furthermore, as shown in Fig. 9a, Lei et al. introduced a Mn²⁺ activation strategy that synergizes host sensitization with dopant sensitization to augment the NIR ML of CaZnOS:Nd³⁺.¹⁰⁴ The PL emission spectrum of CaZnOS:Mn²⁺ (⁴T₁(⁴G)–⁶A₁(⁶S)) overlaps with that of CaZnOS:Nd³⁺ (from ⁴I_9/2–⁴G_9/2 to ⁴I_9/2–⁴F_7/2), facilitating efficient energy transfer from Mn²⁺ to Nd³⁺. As a result, an enhanced NIR ML was observed in CaZnOS:Nd³⁺ with a 0.5% doping of Mn²⁺ during the compressive ML test, exhibiting an intense emission peak at ∼910 nm.

(b) Ultrafast ML response. Currently, there is limited literature focusing on the temporal responses of ML under high-frequency (∼kHz) mechanical stimulation. In contrast to EL or PL, where the temporal response of emission is primarily governed by transient electronic excitation or resonant photon absorption, ML involves the non-transient transmission of force, such as the duration of force application and deformation of the ML flexible substrate. Consequently, investigating the ultrafast temporal response of ML remains a challenge. Chen et al. have made significant contributions in this area.³² They generated microsecond-scale pulsed mechanical excitation (6.4 μs) through elastic collisions between a freely falling iron ball and an elastic ML converter. By combining a fast-response photomultiplier tube (20 ns) and converting photocurrent signals into voltage signals recognizable by a digital oscilloscope (50 MHz), they were able to capture the intrinsic decay behavior of ML. Based on ML decay curves, they employed a double-exponential function to analyze the lifetimes of different ML materials and derived several intriguing conclusions: (i) the ML lifetimes of most materials differ from their PL ones (as exemplified by ZnS:Mn²⁺ in Fig. 9b); (ii) as expected, ML intensity is positively correlated with the energy of the impacting ball, but the ML lifetime remains constant across varying impact energies or ball types; (iii) ML lifetime is influenced by thermoluminescence; for instance, in ZnS:Mn²⁺, as shown in Fig. 9b, ML is likely attributed to band-to-band excitation rather than intra-atomic transition. Part of the excited charge carriers move directly to Mn²⁺ ions, generating short-lived emission (τ_ML-1 for R1), while others are temporarily captured in traps during migration and thermally released to produce persistent emission (τ_ML-2 for R2). (iv) The ML lifetime is jointly determined by the phosphors and dopants, with the shortest measured lifetime being observed in SrZnOS:Bi³⁺ (24 μs). Based on this, the SrZnOS:Bi³⁺ ML sensor shows the potential to detect ultrasonic waves (25 kHz), where clear pulse signals are detected with an interval of ∼39.5 μs, which is significantly shorter than that of conventional electrical resistive sensors (125 μs).

(c) Organic ML. Compared to inorganic luminophores, organic ML offers several advantages: emission color tunability, energy-efficient synthesis, and structural diversity. However, a significant challenge in organic ML arises from π–π stacking interactions in the solid state, which often lead to aggregation-caused quenching effect.¹⁰⁹ In 2001, Luo et al. reported the aggregation-induced emission (AIE) effect,¹¹⁰ where organic ML materials exhibit weak or no luminescence in solution but show a substantial intensity increase in the aggregated state. Subsequently, it has been widely accepted that the ML activity of organic compounds is closely related to their molecular packing,¹¹¹ complicating the establishment of luminescent mechanisms for organic ML materials. Recently, Yang et al. introduced an organic host–guest system,¹⁰⁵ wherein triphenylamine (TPA) was selected as the host, and four guests with 2–4 TPA repeating units served as the guests, as illustrated in Fig. 9c. These organic materials not only exhibited exceptional room-temperature phosphorescence (with lifetimes of 200 ms and quantum yields of 30%) but also emitted blue fluorescence spectra consistent with their PL, upon grinding different doped materials together using a glass rod. Further mechanistic investigations revealed that the ML process in TPA-based organic materials is a consequence of their piezoelectric properties. Specifically, the three phenyl rings in TPA adopt a highly twisted conformation, imparting polarity to the molecule, and its crystal structure belongs to the non-centrosymmetric polar space group C1c1. Consequently, upon disruption of the host crystals, the generated host excitons can transfer energy to the guest molecules, leading to fluorescence emission. Similar to the phosphors:dopants systems in inorganic ML, Yang's work demonstrated that organic ML can establish a host–guest system to explain the ML mechanism through local piezoelectricity. Nevertheless, the continued development of organic ML still faces significant challenges, including the need to enhance luminescent intensity and improve solid-state film-forming properties.

2.4. Table revealing recent progress in EML studies

Recent research progress in EML studies is summarized in Table 1.

3. Mechanoplasmonics

In Section 2.3.1, we elaborate on the current methods for enhancing the ML intensity. However, due to insufficient experimental data, we do not delve deeply into an exciting method for ML emission tuning. In physics, surface plasmons describe the collective oscillations generated by the interaction between electron gas and incident photons at the metal–dielectric interface. When metallic particles exist at sub-wavelength scales, if the oscillation frequency of free electrons matches the incident light frequency, a strong yet decaying localized surface plasmon resonance (LSPR) forms on the surface of the NPs.¹¹² The LSPR effect exhibits notable characteristics such as significant enhancement of the local electric field and sub-wavelength confinement of electromagnetic waves in the near field, as well as pronounced extinction properties in the far field, such as selective absorption and scattering of light at specific frequencies.

The near-field enhancement has stimulated the exploration of its positive impact on luminescence. Indeed, as early as 1980, Glass et al. first reported a fluorescence enhancement phenomenon of rhodamine B on Au/Ag/Cu metal thin layers of appropriate thickness.¹¹³ Later, Mertens et al. reported an increase in the PL intensity of Er³⁺ near Ag nanoparticles (NPs).¹¹⁴ The significant luminescence enhancement effect resulting from the interaction between luminophores and metal nanostructures is widely recognized as plasmon-enhanced luminescence (PEL). Park et al. published a nice review comparing the PEL intensity enhancement factors of different plasmonic structures (e.g., NPs, NP arrays, core–shell structures, etc.),¹¹⁵ as illustrated in Fig. 10a. The reported enhancement factors vary widely, ranging from less than one (quenching) to several hundred. Besides robust intensity regulation, note that the PEL effect often demonstrates more powerful nonlinear frequency conversion. For example, in Fig. 10b, Huang et al. demonstrated that plasmonic dimers,¹¹⁶ compared to single NPs, can directly generate new PL emission peaks. This is because plasmonic dimers, relative to monomers, can produce plasmon hybridization, thereby imparting new emission energy levels to PL.

Fig. 10 Plasmon-enhanced luminescence effect in (a) intensity enhancements (Reproduced with permission.¹¹⁵ Copyright 2015, Royal Society of Chemistry) and (b) frequency conversion (Reproduced with permission.¹¹⁶ Copyright 2015, American Chemical Society).

These studies demonstrate high-level control of the LSPR effect over PL intensity and frequency. However, despite the profound physical connection between PL and ML—where many PL emitters (such as the aforementioned Er³⁺) also serve as ML dopants regulating emission color and intensity—to our knowledge, there are no reported studies on plasmon-enhanced ML phenomena. We believe this absence is multifaceted, primarily due to preparation challenges. Although recent advanced manufacturing techniques have showcased the versatility of LSPR structures from top-down and bottom-up approaches, the LSPR effect strongly depends on the regularity of micro/nanostructures, and effectively integrating them with ML emitters remains difficult. More importantly, when introducing plasmonic structures to ML systems, it is crucial to consider the effects of the modulation of mechanical stimuli on the spacing and the resulting impact on the LSPR effect. As demonstrated in Fig. 10b, plasmonic dimers also exhibit a change in the far-field scattering (DFS) spectrum compared to monomers due to near-field hybridization. Therefore, we can envision a scenario in which plasmonic monomers, situated on flexible substrates, undergo mechanical stimuli that bring NPs into closer proximity. This proximity induces plasmon hybridization, subsequently shifting the scattering spectrum and eliciting novel ML emission peaks from the emitters encircling the NPs.

This represents a promising regulation method; however, before understanding plasmon-enhanced ML, we should delve deeper into the effects of modulation of force on the LSPR effect. Indeed, mechanoplasmonics (MP) investigate the dynamic modulation of the optical properties of plasmonic structures by mechanical stimuli. By incorporating plasmonic particles onto flexible substrates, MP systems exhibit also great sensing potential as FMOS. MPs can utilize mechanical stimuli to regulate light manipulation by adjusting particle morphology/mutual gaps, providing potential for small-scale and highly conformal mechanical detection.¹¹⁷ Here, we mainly discuss two types of MP research: flexible surface-enhanced Raman scattering (SERS) studies based on the near-field enhancement (MP-n) effect of metallic particles, demonstrating their nanoscale sensing prospects; and in situ spectral monitoring studies based on the far-field extinction characteristics (MP-f) of nanoparticles or metasurfaces (as well as dielectric structures, although not plasmonic), showcasing their high-level control over spectral wavelengths and intensity as FMOS.

3.1. MP-n: flexible SERS applications

When excited by light of specific frequencies, metallic NPs can generate strongly enhanced electromagnetic near-fields in their vicinity, exhibiting pronounced surface specificity and geometric inhomogeneity. The specificity to conditions at the surface arises from the dependency of the local electromagnetic field amplitude on the distance from the particle surface, while geometric inhomogeneity reflects the spatial localization of the enhanced near-fields, governed by the nanostructures’ morphology and incident light polarization.¹¹⁸ One of the most notable applications of near-field enhancement is SERS.¹¹⁹ In physics, Raman scattering represents a form of inelastic light scattering where photons interact with molecules or chemical bonds in matter, resulting in energy changes manifest as the emission of photons at different frequencies. These spectral features, known as “fingerprint spectra”, possess narrow linewidths and resistance to photobleaching, directly reflecting the molecular composition and chemical properties of the sample. By positioning analyte molecules in proximity to NPs exhibiting electromagnetic enhancement, SERS maintains fingerprint information of Raman spectroscopy with signal enhancements of 10⁴–10⁶, emerging as a highly sensitive and accurate detection technique at the microscale.

Regarding SERS applications, gold nanoparticles (AuNPs) have garnered significant attention due to their exceptional chemical stability and high SERS activity in the visible light region. Curved edges, tips, or nanogaps (<10 nm) between metallic particles exhibit remarkable enhancement coefficients, contributing predominantly to the SERS signal and giving rise to the “hotspot” effect. Compared to rigid substrates (e.g., wafers, glass, and mica) traditionally employed in SERS, a flexible, stretchable, and foldable polymer substrate not only simplifies the sample pretreatment enabling noninvasive, rapid detection of trace substances by adhering to irregular surfaces, but also facilitates in situ and multi-analyte monitoring applications, with different signals being prominent due to tuning of resonances under strain.¹²⁰ More importantly, from the perspective of FMOS, elastic substrates impart tunable interparticle spacing under mechanical stimuli, theoretically enabling sensing at the sub-10 nm deformation scale of nanoparticle dimers through monitoring SERS enhancement variations, thereby enhancing small-scale sensing capabilities. Hossain et al. investigated the SERS signals for rhodamine 6G molecules absorbed on self-assembled AuNPs supported on a flexible thermoplastic polyurethane (TPU) membrane.⁸ They observed maximum SERS intensity at rest, which decreased upon in situ strain application. Through atomic force microscopy (AFM) testing and numerical simulations, they confirmed that this attenuation of SERS signals was attributed to the increase of the particle spacing from 11 to 16 nm under strain (Fig. 11a).

Fig. 11 Flexible SERS applications based on MP-n effects. (a) Small-scale sensing potential demonstrated in Au dimer system. Reproduced with permission⁸ Copyright 2013, Royal Society of Chemistry. (b) and (c) Enhanced stretching sensing capability through nanowire (Reproduced with permission.¹²¹ Copyright 2019, Elsevier B.V.) and nanochain aggregates (Reproduced with permission.¹²² Copyright 2020, Tsinghua University Press and Springer-Verlag GmbH Germany, part of Springer Nature). (d) Bending sensing performance in nanopyramid metasurfaces. Reproduced with permission.¹²³ Copyright 2022, Tsinghua University Press.

To enhance SERS signal intensity and stability, researchers have explored flexible systems that incorporate aggregated NPs or metasurfaces. Ma et al. fabricated vertically oriented Ag-modified forests of Au nanowires on PDMS via seed-mediated growth and magnetron sputtering.¹²¹ This flexible SERS film exhibited a substantial signal enhancement of approximately 23.77% under an applied strain of 10% when detecting 4-nitrothiophenol molecules. However, as depicted in Fig. 11b, a pronounced decrease in the SERS signal was observed when the strain exceeded 10% (pink solid line), a trend that persisted during both the recovery phase (pink dashed line) and subsequent stretching-recovery cycles (orange lines). SEM tests revealed that the maximum SERS signal enhancement stemmed from PDMS film deformation, inducing surface wrinkles perpendicular to the stretching direction. This deformation reduced the distance between AgNPs and Au nanowires, fostering strong plasmon coupling in the narrowed gaps and generating pronounced “hot-spot” effects that intensified the SERS signal. Conversely, further strain application led to the formation of larger cracks on the PDMS surface, causing wrinkles to dissipate and consequently disrupting the enhancement effect. In addition, some other works involve the near-field effect of NPs embedded within bulk blending systems. Yan et al. synthesized Au nanochains via a hydrothermal method and integrated them into a flexible SERS substrate by blending with polyvinylpyrrolidone (PVP), utilizing crystal violet as the analyte.¹²² Their findings indicated that the Au nanochains could reorient upon external stretching, enabling reversible tuning of the interparticle gaps within each chain (Fig. 11c). Similarly, the SERS signal peaked when the strain reached 13.6%, corresponding to a particle spacing of 1.23 nanometers. However, further increased strain weakened the adhesion of the AuNPs to the film, resulting in a subsequent decrease in SERS enhancement. Note that the pattern of initial increase followed by a decrease in SERS intensity vs. strain has also been reported in disc cluster arrays,¹²⁴ and disc-fiber composites,¹²⁵ where Charconnet et al. attributed this phenomenon to the competitive effects of enhanced near-field enhancement and a decrease in the number of SERS emitters within the finite excitation volume under strain.

Beyond uniaxial stretching, mechanical bending has also been shown to modulate SERS signals. Das et al. leveraged electron beam lithography, nano-printing processes, and surface coating techniques to fabricate Au-coated polymer nanopyramid arrays on a PET/epoxy composite substrate.¹²³ Compared to common PDMS substrates, this substrate exhibited weaker stretchability but higher bending-recovery performance. Consequently, inward bending facilitated the approximation of nanopyramid tips, leading to enhanced near-fields and a 1.1-fold enhancement of the SERS signal compared to the initial flat state; while outward bending diminished near-field enhancement, causing a 1.7-fold signal reduction, a phenomenon validated through numerical simulations (Fig. 11d).

These studies underscore the potential of metallic nanostructure-based near-field enhancement for quantitative sensing of mechanical stimuli, including stretching and bending, demonstrating potential in in situ SERS applications and small-scale mechanics sensing. However, the existing research has, to some extent, indicated relatively weak repeatability for flexible SERS films such as FMOS. This limitation can be attributed to the synergistic effects of strain-induced alterations in NP's near-field enhancement and variations in target molecule dispersion density on the polymer surface. In view of this, future developments of flexible SERS sensors may prioritize not only enhancing the mechanical repeatability and biocompatibility of composite materials, but also improving the adhesion of target molecules and optimizing the physicochemical stability of nanostructures.

3.2. MP-f: high-level spectral manipulation

3.2.1. Nanoparticles dispersed in a uniform environment. On the other hand, metallic NPs exhibit pronounced extinction characteristics, i.e. absorption and scattering, that can be measured in transmission or reflection in the far field under resonance conditions. Compared with near-field effects, these far-field spectra provide a relatively straightforward approach to study plasmonic resonance wavelengths, intensities, and damping effects.¹²⁶ In this context, we review several rigorous mathematical descriptions of the far-field properties of metallic particles that are widely accepted in the academic community. These include the quasi-static approximation and Mie theory for single particles, plasmonic hybridization, the plasmon ruler equation, and Fano resonances for dimers or clusters, as well as surface lattice resonances for metasurfaces.
Quasi-static approximation. The quasi-static approximation characterizes the optical responses of isotropic nanospheres under conditions where the nanosphere's size is significantly smaller than the incident wavelength (typically with a radius on the nanometer scale) and embedded in a uniform medium.^127,128 This approximation assumes a steady electric field across the entire nanosphere, disregarding spatial fluctuations and higher-order multipolar modes beyond the dipolar. Consequently, the absorption and scattering cross-sections of the NP (with radius of a) can be expressed as:where k is the wave vector, and ε_s and ε_d denote the dielectric functions of the sphere and surrounding medium, respectively. This approximation highlights that absorption predominates over scattering in smaller particles.
Mie theory. As the particle size (with radius of a) approaches that of the incident wavelength, significant phase variations within the particle necessitate the use of Mie theory. Developed by Gustav Mie in 1908,¹²⁹ this theory provides rigorous solutions for the optical properties of isotropic, homogeneous dielectric spheres illuminated by monochromatic parallel light. The absorption and scattering efficiency factors are given by:where n defines the multipolar order, and a_n and b_n are the Mie scattering coefficients determined by Ricatti–Bessel functions. These coefficients are computed by applying Maxwell's boundary conditions at the particle's surface after expanding the electromagnetic waves. For brevity, the detailed derivations are not presented here. While Mie's theory and the quasi-static approximation are limited to homogeneous spherical scatterers, they provide valuable insights into the dependence of scattering and absorption properties on particle size, incident wavelength, and dielectric permittivity.
Plasmonic ruler equation. Nano-gap structures exhibit significant near-field enhancement, essential for SERS applications. For a stable coupling system, such as a plasmonic dimer displaying dipolar characteristics, the optical response likewise shows a distinct gap dependence. The shifting resonance is a consequence of plasmon hybridization, where the LSPR splits up into a bonding and antibonding mode.¹³⁰ With decreasing dimer separation, the energy of the bonding mode is lowered, such that a shift to longer wavelengths is observed. As an approximation, the plasmonic ruler equation can be used to describe the exponential dependence of the coupling-mode resonance wavelength on the gap and size. Introduced by Jain et al. for Au disc dimers,¹³¹ the plasmonic ruler equation reads:where the spectral shift Δλ and the mutual gap s are scaled by the single disc's maximum plasmonic wavelength λ₀ and diameter D; and A and τ represent the amplitude and decay constant. Subsequently, these ruler equations have been validated across various systems, including bow-tie/ring-shaped structures and multi-polar coupling systems, demonstrating a universal scaling behavior that explains the gap and size dependence within a dimer system.¹³²
Fano resonance. Distinct from Lorentzian-shaped resonances, Fano resonances mostly exhibit asymmetric lineshapes, demonstrating attractive sensing applications with high quality factors and contrast ratios. These asymmetric resonances were initially investigated by Ugo Fano using a perturbation approach that accounts for the coupling between discrete and continuum states,^133,134 described as:where q represents the shape parameter, and the reduced energy ε corresponds to 2(E − E_F)/Γ, where E_F and Γ determine the resonant energy and the width of the autoionized state. Two eigensolutions are thus revealed: the Fano dip σ_min = 0 at ε = −q and the Fano peak σ_max = 1 + q² at ε = 1/q. In plasmonic systems, Fano resonances have been reported in complex hybrids such as ring/disc cavities,¹³⁵ finite clusters,¹³⁶ and metal–insulator–metal configurations.¹³⁷
Surface lattice resonance. When NPs are arranged in uniform arrays with a consistent lattice period, Wood-Rayleigh anomalies (WRAs) emerge within the lattice plane.^138,139 These anomalies are independent of the permittivity of NPs but are influenced by factors such as the lattice period and medium refractive indices. For instance, the WRA wavelength of two common lattice arrangements, square lattices and regular triangular lattices, can be determined byandwhere P refers to the uniform center-to-center distance between NPs in both lattices, and (m, n) denotes the grating coupling order. In metallic nanostructure arrays, surface lattice resonances with high quality factors emerge owing to the coupling between the LSPR of individual NPs and the WRAs. Note that WRAs observed in transmission through a metallic nanohole array are termed extraordinary optical transmission (EOT) modes, with their behavior influenced by surface plasmon polariton dispersion relations.

3.2.2. Nanoparticles supported on flexible substrates. The above discussion underscores the spectral tunability of the peak wavelength in the far-field resonance of nanostructures, depending on their geometric parameters and distribution states, including spacing, periodicity, and lattice arrangement. In the context of FMOS, integrating these particles onto an elastic substrate allows external mechanical stimuli such as stretching to effectively alter the spacing between the particles. This enables in situ modulation of the far-field spectrum and subsequent quantitative sensing of stimulus magnitude. While most luminophores require repeated UV excitation to achieve repeatable luminescence under tensile-recovery cycles,²⁵ the characteristic of the MP-f effects indicate highly reversible and repeatable spectral manipulation under periodic stimuli. Fig. 12a and b shows randomly distributed AuNPs prepared via spin-coating on the surface of PU and blending within the bulk of PDMS, respectively. Tang et al. reported that the extinction spectra of AuNPs/PU exhibited constant LSPR peak positions at ∼550 nm under strain increased to 100%, with the highest spectral intensity occurring at 40% strain.¹⁴⁰ Similarly, the bulk-blending system reported by Jo et al. indicated negligible spectral shifts in transmittance (peaking at ∼520 nm) and reflectance as the strain increased to 40%.¹⁴¹ These weak manipulations of peak wavelength suggest that randomly distributed NPs, due to the synthesis limitations and initially large interparticle spacing, cannot effectively achieve the desired MP-f effect within the elastic strain range of the substrate.

Fig. 12 High-level spectral manipulation based on MP-f effects. (a) and (b) Random distribution of AuNPs on the polymer surface (Reproduced with permission.¹⁴⁰ Copyright 2021, American Chemical Society) and within polymer matrices (Reproduced with permission.¹⁴¹ Copyright 2022, American Chemical Society). (c)–(f) 1D linear (Reproduced with permission.¹⁴² Copyright 2017, American Chemical Society), 2D square (Reproduced with permission.¹⁴³ Copyright 2019, The Authors, under CC-BY license), 2D triangular metasurfaces (Reproduced with permission.¹⁴⁴ Copyright 2020, American Chemical Society), and 3D helical dispersion of AuNPs (Reproduced with permission.¹⁴⁵ Copyright 2022, Wiley-VCH GmbH) dispersed on polymer substrates.

Therefore, the development of well-ordered NP arrangements or metasurfaces represents a promising direction for achieving effective spectral regulation. Current state-of-art fabrication techniques include “direct methods” such as bottom-up colloidal self-assembly,¹⁴⁶ and top-down methods such as optical tweezers,¹⁴⁷ ion beam milling,¹⁴⁸ and electron beam lithography,¹⁴⁹ as well as “indirect methods” such as pattern transfer and nanoprinting.¹⁵⁰ Steiner et al. employed NP assembly and wet-printing transfer to prepare 1D arrays of AuNPs on the surface of PDMS (Fig. 12c).¹⁴²In situ atomic force microscopy (AFM) tests revealed that these arrays undergo regular chain breakup under stretching, and therefore the longitudinal polarization extinction spectra displayed a blueshift of 59 nm in LSPR peaks at 50% strain (see Table 2 for details). Furthermore, the enhanced regularity of the metallic NPs facilitated the spectral shift of SLRs upon in situ stretching. Gupta et al. utilized soft lithography and colloidal self-assembly to create a 2D square lattice of AuNPs on PDMS.¹⁴³ As shown in Fig. 12d, a 40% increase in strain not only caused a blueshift in the LSPR peaks, but also led to the emergence and separation of two distinct SLR peaks as strain develops, due to interspacing changes in both longitudinal and transverse directions as the strain developed. Similarly, the spectral evolution of a 2D triangular lattice is examined by Jia et al. in AuNPs on PDMS.¹⁴⁴ In this case, a 40% strain induced regular deformation of the triangular array, resulting in two distinct redshifts in the transmittance dips (Fig. 12e). Interestingly, Grzelak et al. introduced a liquid crystalline compound (1,3-phenylenebis[4-(4-oleyloxy-phenyliminonetyl)benzoate]) that can form S or R conformations at different temperatures.¹⁴⁵ When serving as the matrix for AuNP organization, it facilitated the formation of nanocomposites with 3D helical structures of different chirality. The authors then used a thermal nanoimprinting process to transfer the nanocomposites onto PDMS and tested the circular dichroism (CD) intensity during stretching and relaxation (Fig. 12f). Although the CD spectra of the chiral plasmonic structures exhibited differences, they all displayed a slight blueshift and a gradual loss of intensity as the strain increased to 37.5%. Upon relaxation, the chiroptical response was fully recovered, highlighting the dynamic nature of the helical structures.

Table 2 Recent progress in MP-f studies

System	Type of resonance	Strain changes [%]	Peak shifts [nm]	Sensitivity [nm/%]	Repeatability [times]	Highlight	Source
a TP: transverse polarization, where the polarization direction is oriented perpendicular to the strain. b LP: longitudinal polarization, which aligns parallel to the strain.
Disordered polystyrene-capped Au nanobipyramids	LSPR	0–90	∼870–798 (varied with orientation)	−0.8	>100	Orientation-dependent LSPRs	151
Disordered polystyrene-capped Au nanocubes	LSPR	0–50	∼586–517	−1.38	1000	Polymer ligands-dependent LSPRs	152
Disordered tilted Au nanorods	LSPR	0–30	∼633–693	2	—	Orientation-dependent LSPRs	153
Disordered ITO spheres	LSPR	0–50	∼3500–3345	−3.10	7	Wearable applications	154
Disordered poly(vinylpyrrolidone)-capped Au spheres	LSPR	0–4 [N] (force applied)	648–688 (under TP)^a/648–639 (under LP)^b/(varied with strain direction)	10 nm N^−1	—	Polarization-dependent LSPRs	155
				−2.25 nm N^−1 (scaled with force)
Disordered Au-polystyrene sphere-Au structures	LSPR-enhanced Fabry–Pérot mode	0–31.8	941–813	−4.02	—	Metal-insulator-metal structure	156
Disordered Au spheres-TiO2 thin film	LSPR	0–13.4	∼640–589	−3.80	300	Different sensitivity vs. strain speed	157
Au bow-tie dimer	LSPR	0–25	—	1.9	40	Dimer structures	10
Si sphere dimer	Magnetic and electric dipolar resonance	0–100	—	—	10	Intensity jumping	12
1D linear Au sphere array	LSPR	0–50	700–759	1.18	4	Nanochain structures	142
2D triangular Al disc array	SLR	0–50	Quadrupolar SLR: ∼509–701	3.84	—	Different SLRs vs. strain directions	158
			Dipolar SLR: ∼612–717 (varied with strain direction)	2.10
2D triangular Au disc array	LSPR	0–40	LSPR1: ∼561–584	0.58	—	Easy fabrication	144
			LSPR2: ∼841–872	0.78
2D triangular Si disc array	Mie resonance	Mechanical bending	679–690	—	—	Dielectric metasurface	159
2D triangular Au bow-tie array	LSPR	0–30 (pre-stretching state)	∼707–900 (under LP)	6.43	—	Easy fabrication	160
			∼710–937 (under TP)	7.57
2D square protein-coated Au sphere array	LSPR and SLR	0–40	LSPR: ∼546–523	−0.58	100	Splitting SLRs	143
			SLR1: ∼564–551	−0.33
			SLR2: ∼570–615	1.13
2D square Au disc array	SLR	2–5	∼851–882	10.3	—	High sensitivity	161
2D square Au elliptical-disc array	LSPR	0–4	801–865	16	—	Asymmetric structures	162
2D square Au nanoring array	LSPR	0–45	675.8–804	2.85	—	Shape changes	163
2D square Au disc-cluster array	LSPR and SLR	0–35	∼700–644 (under TP)	−1.6	8	Tunable red- and blue-shifts	124
			∼700–809 (under LP)	3.1
2D square Al rectangle array	LSPR	0–32	∼495–645 (strain along short axis)	4.69	>100	Full-spectrum color control	164
		0–31	∼496–439 (strain along long axis)	−1.84
2D square TiO2 nanoblock array	Magnetic and electric dipolar resonance	0–40	∼471–578 (under TP)	2.68	>100	Robust color control	165
			∼473–571 (under LP)	2.45
Au wire grating	SLR	1.6–3.5	744–836	48.42	10	Ultrasensitivity	166
Au wire grating	LSPR	0–30	∼615–621	0.20	100	Easy fabrication	167
Al wire grating	LSPR	0–32.4	∼650–470	−5.56	10	Wearable application	9
U-shaped Al nanowire grating	Fano resonance	0–20	∼494–562	3.40	—	Fano resonance	168
3D wavy square Au disc array	LSPR	0–30	∼643–715 (under TP)	2.4	—	Wavy distribution	169
3D helical Au spheres	LSPR	0–37.5	∼619.2–618.3 (chiral-S)	−0.02	—	Chiral response	145
			∼611.1–608.4 (chiral-R)	−0.07

---

These works demonstrate that MP nanostructures with different initial dispersions can achieve in situ and continuous spectral modulation under strain. However, these particles are primarily concentrated in solid, uniform, and isotropic forms, with relatively limited ability to control their morphology and size distribution. Here, we summarize the development directions for advanced MP studies: (1) spectral monitoring based on nanoscale gap deformation. Advanced by top-down electron beam lithography and wet-etching pattern transfer techniques, Laible et al. studied the reflection spectra of a single Au bow-tie nanostructure on PDMS (Fig. 13a).¹⁰ They detected spectral changes in the nano-gap system demonstrating the prospects of MP-f effects for in situ nanoscale strain sensing prospects. (2) Spectral monitoring based on nanoscale shape alternation. Tao et al. further developed the pattern transfer method by preparing Au nanoring metasurfaces on PDMS.¹⁶³In situ spectral monitoring and SEM tests demonstrated that the deformed asymmetric nanorings under strains achieve distinct spectral tuning characteristics (Fig. 13b). In fact, current advanced top-down techniques are not limited to the 2D patterns by lithography, but also utilize 3D printing techniques like two-photon polymerization to fabricate more complex hollow or sculpted structures.¹⁷⁰ This work indicates the FMOS potential of hollow MP structures, as well as applications in flexible displays and strain mapping. (3) Colloidal aggregates via bottom-up fabrications. While top-down methods provide enhanced control over structure designs, bottom-up self-assembly excels in cost-effectiveness and scalability. The work of Grzelak et al. underscores the advancement in designing NP aggregations states.¹⁴⁵ Similar endeavors are evident in the research by Biswas et al.,¹⁷¹ where they demonstrated the plasmonic photothermal activation to uniformly assemble colloidal superstructures, such as Ag nanowires, and Ag nanospheres assembled on SiO₂@Au bipyramids, operating at reduced temperatures and exhibiting accelerated growth kinetics (Fig. 13c). These studies demonstrate a crucial research direction toward attaining both geometric and distributional precision via self-assembly processes, laying the foundation for MP-based FMOS applications. (4) Dielectric structures for spectral manipulation. Under resonant conditions, dielectric particles can excite circular displacement currents, with the resulting magnetic response playing a role in their scattering efficiency. Despite not exhibiting plasmonic properties, dielectric NPs exhibit moderate light localization with reduced dissipative losses. Yan et al. demonstrated that silicon dimers on PDMS can significantly enhance reflection intensity rather than shifting spectral peaks under strain,¹² highlighting their potential as nanopixels for information encryptions (Fig. 13d).

Fig. 13 Further developments in MP-f studies. (a) and (b) Spectral monitoring based on in situ nano-gap (Reproduced with permission.¹⁰ Copyright 2018, Royal Society of Chemistry) and shape deformation under strain (Reproduced with permission.¹⁶³ Copyright 2023, The Authors, under CC-BY license). (c) Bottom-up methods for achieving complex aggregates. Reproduced with permission.¹⁷¹ Copyright 2023, The Authors, under CC-BY license. (d) Spectral manipulation in flexible dielectric structures. Reproduced with permission.¹² Copyright 2023, American Chemical Society.

3.2.3. Table revealing recent progress in MP-f studies. Recent research progress in MP-f studies is summarized in Table 2.

4. Integration of FMOS and their expansion beyond sensing applications

Here, we summarize the primary challenges faced by various FMOS systems currently encompassing: (1) the wide range and intricate mechanisms of ML materials, with luminescence, mechanical properties, and multi-functionalities significantly influenced by their crystalline state, polymeric matrices and blending methods; (2) the ML performance typically manifests as intensity enhancement versus increased mechanical magnitude, yet it may display non-linear behavior and offers limited dynamic control over luminescence color; (3) MP materials exhibit potential for nanoscale sensing and high-level spectral tunabilities, but invariably require an excitation light source; (4) the optical properties of MP materials are highly dependent on structural regularity, posing stringent requirements for nanomanufacturing; (5) due to the dependence of the plasmon ruler behavior on the respective nanostructure shape, individual sensor calibration poses a challenge; (6) currently, the commercialization of ML and MP systems is primarily in the demonstration phase. Key challenges include achieving robust optical intensity for functional materials while ensuring wearable comfort or implantable compatibility, as well as managing costs and enabling large-scale production. Based on the unique characteristics of ML and MP materials, this review proposes the following prospects and potential applications for integrating ML and MP within the realm of FMOS:

ML + MP for active SERS applications. In SERS applications, lasers are the most common excitation source, enhancing testing accuracy while limiting the operational conditions of flexible SERS measurements. Promoting ML particles to exhibit MP-n effect, or integrating the two materials through different blending methods, may eliminate the need for lasers and enable highly portable and in situ detection.¹⁷² However, this also poses challenges to the testing accuracy and noise reduction of the monitoring techniques.

ML + MP for luminescence intensity/color manipulation. Since the luminescent mechanisms rely on the piezoelectrification of phosphors, it remains unclear whether plasmonic properties can enhance luminescence in a bulk blend of ML and MP structures.¹⁷³ Alternatively, laminating ML and MP layers and examining the regulatory capabilities of MP structures on ML color present another avenue of research. These potential endeavors aim to achieve complementary advantages among different FMOS systems, thereby enhancing their application prospects in mechanical visual sensing.

ML + MP for multi-functional FMOS. As for usage environments, ML can be utilized in dark or low-light environments, while MP performs in bright-field conditions. Therefore, the effective combination of both can deliver FMOS performance tailored to different environments, with evident potential applications in wearable devices and human–machine interaction. Furthermore, in terms of mechanical stimulation, the luminescence phenomenon is significantly induced by dynamic stimuli, e.g., changes in pressure rather than absolute pressure levels.¹⁰⁰ Conversely, MP properties of artificial nanostructures are more related to their distributions under the final mechanical state, which are more responsive to static stimuli. Thus, the effective combination of ML and MP may achieve a multi-functional sensing prospect, providing different feedback to various mechanical stimuli, as well as environmental factors. These potential advantages of combining ML and MP are conducive to the development of integrated, informative, and intelligent FMOS systems.

Furthermore, FMOS exhibit potential for seamless integration with electrical sensors (E-sensors), thereby enabling multifunctional and optical/electrical dual-mode sensing capabilities, as discussed in detail in Section 2.2. Here, we schematically outline the working mechanisms of E-sensors (Fig. 14a), categorizing them into self-powered devices, such as piezoelectric nanogenerators (PENGs) and triboelectric nanogenerators (TENGs), which produce pulsed alternating current, and externally powered resistive and capacitive sensors, typically driven by low-voltage direct current.¹⁷⁴Fig. 14b presents a comparative analysis between E-sensors and FMOS. A notable feature of E-sensors is their dependence on wired electrical signal transmission, often necessitating the incorporation of batteries, microcontrollers for signal processing, and wireless communication modules (e.g., NFC, LoRa) onto a printed circuit board, alongside the sensors themselves.¹⁷⁵ This results in a more complex structure compared to FMOS. Optical signal detection, on the other hand, generally requires spectrometers and is prone to background noise, potentially necessitating additional equipment like optical fibers and encoders, making it less convenient for monitoring than E-sensors. However, FMOS possess a unique advantage in visualization if they can modulate light emission or reflection within the visible light spectrum. Currently, research is actively exploring the visualization of E-sensor arrays. As shown in Fig. 14c, Wang et al. integrated a TENG element array with a ZnS:Mn²⁺ luminescent layer and conducted a systematic comparison of their respective advantages and limitations.¹⁷⁶ The ML layer theoretically allows for pixel-level small-scale sensing, but electrical signals experience interference (≥20 mm) from pressure-induced activation of adjacent TENG elements, limiting the precision of trajectory recognition. In terms of sensitivity, the pressure required for ML activation (≥100 kPa) is substantially higher than that for TENG sensor arrays (≤10 kPa). Nevertheless, recent studies indicate that close-range ML monitoring can enhance photon collection efficiency, elevating ML's sensitivity to the ∼20 kPa level.¹⁷⁷ Furthermore, with regard to response time, Chen et al. have demonstrated that specific ML materials exhibit faster response speeds than commercial E-sensors (39.5 μs vs. 125 μs, as shown in Fig. 9b).³² In the context of E-sensor visualization, the signal processing delay and visualization process may lag behind mechanical stimuli by several seconds, rendering them less responsive than FMOS. In general, FMOS and E-sensors exhibit complementary advantages, making their integration a promising avenue for future research.

Fig. 14 Further developments in FMOS. (a) Overview of E-sensors; (b) comparison between FMOS and E-sensors; (c) integration of FMOS with E-sensors (Reproduced with permission.¹⁷⁶ Copyright 2017, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim); (d) FMOS applications in blue energy harvesting (Reproduced with permission.¹⁷⁸ Copyright 2021, The Authors, under CC-BY license); and (e) FMOS applications in human–machine interaction (Reproduced with permission.¹⁷⁹ Copyright 2023, The Authors, under CC-BY license).

Beyond sensing applications, FMOS demonstrate considerable potential across various domains, including blue energy harvesting, flexible electronics, and human–machine interaction. As illustrated in Fig. 14d, Wang et al. achieved dynamic modulation of ML color by incorporating different ML materials into ZrO₂ microspheres using a core–shell structure.¹⁷⁸ The inset figure schematically shows ML particles placed in a transparent acrylic container, which can float on the ocean surface and emit light with ocean currents, offering new application possibilities for marine rescue and navigation. Furthermore, Zhou et al. introduced a method to capture ML signals for the machine learning process,¹⁸⁰ enabling intelligent control of a trolley, as shown in Fig. 14e. Similarly, Xu et al. devised an approach to input ML signals into a convolutional neural network (CNN) for deep learning, facilitating the recognition of tactile interaction objects.¹⁷⁹ Collectively, these studies underscore the versatile potential of FMOS in human–machine interaction applications that extend beyond traditional sensing capabilities.

5. Conclusion

This review presents a comprehensive survey of recent advancements in the rapidly evolving field of FMOS, emphasizing their wireless and visualized sensing capabilities. In the realm of ML, we discuss the emission mechanisms of luminophores and intensity enhancements by introducing activated dopants and heterojunction structures during the high-temperature synthesis routes. Notably, we highlight the performance and underlying principles of various blending systems that integrate ML particles with polymeric matrices/substrates, including bulk, laminar, and woven configurations. These blending strategies facilitate prospects for applications such as 3D-printed mechanical visualization, multifunctional dual-mode sensing, and wearable optoelectronics. The versatility of ML composites is further emphasized by their tunable emission intensity and color in response to diverse mechanical and environmental stimuli, underscoring their potential for quantitative sensing of a wide range of mechanical inputs. Turning to MP, we discuss the potential of artificially designed nanostructures to harness both near-field enhancements and far-field scattering effects. Specifically, the flexible SERS application emerges as a promising candidate for small-scale and in situ sensing applications, while MP-f structures demonstrate dynamic spectral control over resonance wavelengths and intensity, supported by rigorous mathematical modeling and experimental studies.

Despite the substantial progress achieved in both ML and MP, there remains a pressing need to address usage limitations and further advance the FMOS strategy. Given the complementary strengths of ML (active luminescence) and MP (high-level spectral manipulation), we envision hybrid FMOS systems integrating both technologies, with potential applications spanning luminescent SERS systems, luminescence intensity/color manipulation via MP effect, and multifunctional sensing. In addition, we provide a detailed comparison between FMOS and electrical sensors, elucidating their complementary strengths, and expand upon the discussion of FMOS potential applications in blue energy harvesting and human–machine interaction. To realize this vision, future research should prioritize material fabrication and optimization efforts aimed at enhancing the overall stability, durability, and sensitivity. Additionally, a deeper understanding of the underlying mechanisms governing sensor performance, coupled with the development of standardized protocols for signal evaluation and comparison, will be instrumental in advancing the field. To conclude, this review aims to serve as a solid foundation for research endeavors in this rapidly evolving and exciting sensing technology of FMOS.

Author contributions

Conceptualization: HQH, LY, YL; methodology: WT, YFX; writing – original draft: WT; writing – review & editing: WT, YL, TM, MF; funding acquisition: LY, YL; supervision: HQH, TM, MF.

Data availability

This is a review article and does not include original data.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors acknowledge the support from the Dongguan Key Laboratory of Digital and Intelligent Equipment for Emergency Industry, Dongguan University of Technology Analytical and Testing Center,  and Guangdong Provincial Key Laboratory of Intelligent Disaster Prevention and Emergency Technologies for Urban Lifeline Engineering (no. 2022B1212010016), the National Natural Science Foundation of China (NSFC) (11874111, 12204101) and Guangdong Basic and Applied Basic Research Foundation (2022A1515012131, 2022A1515140143).

References

1. Y. Luo, M. Abidian, J.-H. Ahn, D. Akinwande, A. Andrews, M. Antonietti, Z. Bao, M. Berggren, C. Berkey and C. Bettinger, ACS Nano, 2023, 17, 5211–5295.
2. T. Maurer, J. Marae-Djouda, U. Cataldi, A. Gontier, G. Montay, Y. Madi, B. Panicaud, D. Macias, P.-M. Adam and G. Lévêque, Front. Mater. Sci., 2015, 9, 170–177.
3. V. P. Rachim and W.-Y. Chung, Sens. Actuators, B, 2019, 286, 173–180.
4. L. Guo, P. Xia, T. Wang, A. N. Yakovlev, T. Hu, F. Zhao, Q. Wang and X. Yu, Adv. Funct. Mater., 2023, 33, 2306875.
5. X. Qian, Z. Cai, M. Su, F. Li, W. Fang, Y. Li, X. Zhou, Q. Li, X. Feng, W. Li, X. Hu, X. Wang, C. Pan and Y. Song, Adv. Mater., 2018, 30, 1800291.
6. W. Yang, W. Gong, W. Gu, Z. Liu, C. Hou, Y. Li, Q. Zhang and H. Wang, Adv. Mater., 2021, 33, 2104681.
7. C. Wang, H. Hu, D. Zhu and C. Pan, Sci. Bull., 2023, 68, 559–561.
8. M. K. Hossain, G. R. Willmott, P. G. Etchegoin, R. J. Blaikie and J. L. Tallon, Nanoscale, 2013, 5, 8945–8950.
9. H. Kumagai, T. Fujie, K. Sawada and K. Takahashi, Adv. Opt. Mater., 2020, 8, 1902074.
10. F. Laible, D. A. Gollmer, S. Dickreuter, D. P. Kern and M. Fleischer, Nanoscale, 2018, 10, 14915–14922.
11. J. M. Djouda, G. Montay, B. Panicaud, J. Beal, Y. Madi and T. Maurer, Mech. Mater., 2017, 114, 268–278.
12. J. Yan, K. Zhao, T. Wu, X. Liu, Y. Li and B. Li, ACS Appl. Mater. Interfaces, 2023, 15, 38682–38692.
13. S. Hajra, S. Panda, S. Song, B. K. Panigrahi, P. Pakawanit, S. M. Jeong and H. J. Kim, Nano Energy, 2023, 114, 108668.
14. X. Ma, C. Wang, R. Wei, J. He, J. Li, X. Liu, F. Huang, S. Ge, J. Tao, Z. Yuan, C. Ping, D. Peng and C. Pan, ACS Nano, 2022, 16, 2789–2797.
15. C. N. Xu, T. Watanabe, M. Akiyama and X. G. Zheng, Appl. Phys. Lett., 1999, 74, 1236–1238.
16. C. N. Xu, T. Watanabe, M. Akiyama and X. G. Zheng, Appl. Phys. Lett., 1999, 74, 2414–2416.
17. M. Akiyama, C.-N. Xu, H. Matsui, K. Nonaka and T. Watanabe, Appl. Phys. Lett., 1999, 75, 2548–2550.
18. F. Yang, H. Cui, X. Wu, S.-J. Kim and G. Hong, Nanoscale, 2023, 15, 1629–1636.
19. Y. Zheng, K. Wang, J. He, L. Yang, L. Qian and Y. Lu, Sens. Actuators, A, 2024, 369, 115217.
20. Z. Xie, X. Zhang, Y. Xiao, H. Wang, M. Shen, S. Zhang, H. Sun, R. Huang, T. Yu and W. Huang, Adv. Mater., 2023, 35, 2212273.
21. S. Jindal, G. Anjum, V. K. Maka and J. N. Moorthy, Nanoscale, 2021, 13, 9668–9677.
22. B. Chen, X. Zhang and F. Wang, Acc. Mater. Res., 2021, 2, 364–373.
23. Z. Wang, Y. Tai, J. Nam and Y. Yin, Chem. Mater., 2023, 35, 6845–6852.
24. D. Peng, Y. Jiang, B. Huang, Y. Du, J. Zhao, X. Zhang, R. Ma, S. Golovynskyi, B. Chen and F. Wang, Adv. Mater., 2020, 32, 1907747.
25. T. Zheng, M. Runowski, I. R. Martín, K. Soler-Carracedo, L. Peng, M. Skwierczyńska, M. Sójka, J. Barzowska, S. Mahlik, H. Hemmerich, F. Rivera-López, P. Kulpiński, V. Lavín, D. Alonso and D. Peng, Adv. Mater., 2023, 35, 2304140.
26. L. Chen, C. Wu, X. Zhang, M.-C. Wong, L. Pang and J. Zhai, J. Alloys Compd., 2022, 893, 162333.
27. Y. Wang, J. Fan, H. Su, Y. Wu, D. Tu, D. Peng and A. Zhong, Appl. Surf. Sci., 2024, 656, 159583.
28. M. V. Mukhina, J. Tresback, J. C. Ondry, A. Akey, A. P. Alivisatos and N. Kleckner, ACS Nano, 2021, 15, 4115–4133.
29. X. Wang, R. Ling, Y. Zhang, M. Que, Y. Peng and C. Pan, Nano Res., 2018, 11, 1967–1976.
30. H. Liu, Y. Zheng, S. Liu, J. Zhao, Z. Song, D. Peng and Q. Liu, Adv. Funct. Mater., 2024, 2314422.
31. J. Zhao, S. Song, X. Mu, S. M. Jeong and J. Bae, Nano Energy, 2022, 103, 107825.
32. C. Chen, Z. Lin, H. Huang, X. Pan, T. L. Zhou, H. Luo, L. Jin, D. Peng, J. Xu, Y. Zhuang and R.-J. Xie, Adv. Funct. Mater., 2023, 33, 2304917.
33. J. Jia, X. Gao and G. Zou, Adv. Funct. Mater., 2022, 32, 2207881.
34. Y. Bai, X. Guo, B. Tian, Y. Liang, D. Peng and Z. Wang, Adv. Sci., 2022, 9, 2203249.
35. Y. Bai, F. Wang, L. Zhang, D. Wang, Y. Liang, S. Yang and Z. Wang, Nano Energy, 2022, 96, 107075.
36. J. Liu, G. C. Lama, F. Recupido, C. Santillo, G. Gentile, G. G. Buonocore, L. Verdolotti, X. Zhang and M. Lavorgna, Compos. Sci. Technol., 2023, 236, 109993.
37. B. Tian, Z. Wang, A. T. Smith, Y. Bai, J. Li, N. Zhang, Z. Xue and L. Sun, Nano Energy, 2021, 83, 105860.
38. S. Yu, S. Fang, L. Zhao, Y. Bai, R. Wang and Z. Wang, Chem. Eng. J., 2023, 474, 145542.
39. J. Luo, B. Ren, X. Zhang, M. Zhu, T. Liang, Z. Huang, Y. Zheng, X. Li, J. Li, Z. Zheng, B. Chen, Y. Fu, D. Tu, Y. Wang, Y. Jia and D. Peng, Adv. Sci., 2024, 11, 2305066.
40. W. Li, S. Wang, M. Jin, L. Wang, J. Nan, C. Wang, P. Xiong, Q. M. Hu, L. Liu and J. Ren, Small, 2024, 2310180.
41. Y. Cai, S. Liu, L. Zhao, C. Wang, H. Lv, B. Liu, J. Qiu, X. Xu and X. Yu, J. Adv. Ceram., 2022, 11, 1319–1329.
42. L. Li, C. Cai, X. Lv, X. Shi, D. Peng, J. Qiu and Y. Yang, Adv. Funct. Mater., 2023, 33, 2301372.
43. S. Peng, P. Xia, T. Wang, L. Lu, P. Zhang, M. Zhou, F. Zhao, S. Hu, J. T. Kim, J. Qiu, Q. Wang, X. Yu and X. Xu, ACS Nano, 2023, 17, 9543–9551.
44. Y. Song, J. Xiao, L. Zhao, Z. Liu, Y. Ling, Y. Yan, Y. Xu, T. Hu, T. G. Cherkasova and Q. Xu, J. Mater. Chem. A, 2024, 12, 2796–2806.
45. W. Wang, S. Wang, Y. Gu, J. Zhou and J. Zhang, Nat. Commun., 2024, 15, 2014.
46. C. Wang, H.-Z. Shi, W.-H. Li, Q.-M. Hu, L. Liu, J. Ren and J. Zhang, ACS Appl. Mater. Interfaces, 2023, 15, 30891–30901.
47. H. Yang, Y. Wei, H. Ju, X. Huang, J. Li, W. Wang, D. Peng, D. Tu and G. Li, Adv. Mater., 2024, 2401296.
48. S. Liu, Y. Zheng, D. Peng, J. Zhao, Z. Song and Q. Liu, Adv. Funct. Mater., 2023, 33, 2209275.
49. S. Qin, W. Wei, B. Tian, Z. Ma, S. Fang, Y. Wang, J. Zhang and Z. Wang, Adv. Funct. Mater., 2024, 2401535.
50. J. C. Zhang, N. Gao, L. Li, S. Wang, X. Shi, M. Sun, X. Yan, H. W. He, X. Ning, B. Huang and J. Qiu, Adv. Funct. Mater., 2021, 31, 2100221.
51. B. Chandra, V. Chandra, P. Jha, D. Pateria and R. Baghel, Luminescence, 2016, 31, 67–75.
52. B. Chandra, V. Chandra and P. Jha, Phys. B, 2015, 461, 38–48.
53. G. Lee, S. Song, W. H. Jeong, C. Lee, J. S. Kim, J. H. Lee, J. Choi, H. Choi, Y. Kim, S. J. Lim and S. M. Jeong, Small, 2024, 20, 2307089.
54. S. Rahmany, A. Shpatz Dayan, M. Wierzbowska, A. J. Ong, Y. Li, S. Magdassi, A. I. Y. Tok and L. Etgar, ACS Energy Lett., 2024, 9, 1527–1536.
55. S. S. H. Abir, S. Sharma, P. Sharma, S. Karla, G. Balasubramanian, J. Samuel and N. Koratkar, Nat. Commun., 2024, 15, 5768.
56. X. Guo, J. Bian, Y. Bai, Z. Ma, S. Yang and Z. Wang, Chem. Phys. Lett., 2022, 787, 139235.
57. Y. Xiao, P. Xiong, S. Zhang, K. Chen, S. Tian, Y. Sun, P. Shao, K. Qin, M. G. Brik, S. Ye, D. Chen and Z. Yang, Chem. Eng. J., 2023, 453, 139671.
58. B. Chandra, V. Chandra and P. Jha, Phys. B, 2015, 463, 62–67.
59. D. Moore and Z. L. Wang, J. Mater. Chem., 2006, 16, 3898–3905.
60. F. C. Ugbo, S. Porcu, R. Corpino, A. Pinna, C. M. Carbonaro, D. Chiriu, P. F. Smet and P. C. Ricci, Materials, 2023, 16, 3511.
61. J. Y. Byeon and K. Y. Jung, Ceram. Int., 2022, 48, 34837–34847.
62. H. Lv, Y. Wang, Z. Pan, X. Li and Y. Zhang, J. Rare Earths, 2023, 41, 639–648.
63. D. Tu, C.-N. Xu and Y. Fujio, J. Adv. Dielectr., 2014, 4, 1450017.
64. W. Li, S. Wang, M. Jin, L. Wang, J. Nan, C. Wang, P. Xiong, Q. M. Hu, L. Liu, J. Ren and J. Zhang, Small, 2024, 2310180.
65. S. Golovynskyi, Z. Kanwal, I. S. Babichuk, Z. Xie, B. Li, H. Wu and D. Peng, J. Alloys Compd., 2024, 990, 174437.
66. R. Wei, J. He, S. Ge, H. Liu, X. Ma, J. Tao, X. Cui, X. Mo, Z. Li, C. Wang and C. Pan, Adv. Mater. Technol., 2023, 8, 2200757.
67. M. P. Wolf, G. B. Salieb-Beugelaar and P. Hunziker, Prog. Polym. Sci., 2018, 83, 97–134.
68. Y. Song, J. Xiao, L. Zhao, Z. Liu, Y. Ling, Y. Yan, Y. Xu, T. Hu, T. G. Cherkasova, Q. Xu, C. Wang and X. Xu, J. Mater. Chem. A, 2024, 12, 2796–2806.
69. J. Bian, Y. Han, F. Wang, B. Liu, H. Li and Z. Wang, Mater. Res. Bull., 2021, 140, 111295.
70. H. Ji, Y. Tang, B. Shen, X. Qian, Z. Cai, F. Li, M. Su, L. Wu, Y. Ma and Y. Song, ACS Appl. Mater. Interfaces, 2021, 13, 60689–60696.
71. X. Pan, Y. Zhuang, W. He, C. Lin, L. Mei, C. Chen, H. Xue, Z. Sun, C. Wang, D. Peng, Y. Zheng, C. Pan, L. Wang and R.-J. Xie, Nat. Commun., 2024, 15, 2673.
72. S. Peng, P. Xia, T. Wang, L. Lu, P. Zhang, M. Zhou, F. Zhao, S. Hu, J. T. Kim and J. Qiu, ACS Nano, 2023, 17, 9543–9551.
73. X. Qian, Z. Cai, M. Su, F. Li, W. Fang, Y. Li, X. Zhou, Q. Li, X. Feng and W. Li, Adv. Mater., 2018, 30, 1800291.
74. C. Wang, R. Ma, D. Peng, X. Liu, J. Li, B. Jin, A. Shan, Y. Fu, L. Dong, W. Gao, Z. Wang and C. Pan, InfoMat, 2021, 3, 1272–1284.
75. J. Yin, X. Huo, X. Cao, R. Li, Y. Zhou, T. Jiang, L. Wang, Z. Wu and Z. L. Wang, ACS Mater. Lett., 2022, 5, 11–18.
76. W. Wang, D. Peng, H. Zhang, X. Yang and C. Pan, Opt. Commun., 2017, 395, 24–28.
77. F. Wang, F. Wang, X. Wang, S. Wang, J. Jiang, Q. Liu, X. Hao, L. Han, J. Wang, C. Pan, H. Liu and Y. Sang, Nano Energy, 2019, 63, 103861.
78. Z. Zhou, K. Liu, Z. Ban and W. Yuan, Composites, Part A, 2022, 154, 106806.
79. X. Zhao, Z. Zhang, Q. Liao, X. Xun, F. Gao, L. Xu, Z. Kang and Y. Zhang, Sci. Adv., 2020, 6, eaba4294.
80. G. Liang, Z. Ruan, Z. Liu, H. Li, Z. Wang, Z. Tang, F. Mo, Q. Yang, L. Ma, D. Wang and C. Zhi, Adv. Electron. Mater., 2019, 5, 1900553.
81. C. Chen, J. Zhu, Y. Zhang and A. Wang, Coatings, 2022, 12, 449.
82. J. Yang, X. Yao and Z. Meng, Polymer, 2022, 245, 124691.
83. Z. L. Wang, ACS Nano, 2013, 7, 9533–9557.
84. X. Zhang, Z. Li, W. Du, Y. Zhao, W. Wang, L. Pang, L. Chen, A. Yu and J. Zhai, Nano Energy, 2022, 96, 107115.
85. H. Song, S. Panda, S. Hajra, S. Hwang, J. Jo, N. Kim, B. K. Panigrahi, J. Yu, S. M. Jeong and H. J. Kim, Energy Technol., 2024, 2400424.
86. S. M. Jeong, S. Song, H. J. Seo, W. M. Choi, S. H. Hwang, S. G. Lee and S. K. Lim, Adv. Sustainable Syst., 2017, 1, 1700126.
87. J. Zhang, L. Bao, H. Lou, J. Deng, A. Chen, Y. Hu, Z. Zhang, X. Sun and H. Peng, J. Mater. Chem. C, 2017, 5, 8027–8032.
88. H. Liang, Y. He, M. Chen, L. Jiang, Z. Zhang, X. Heng, L. Yang, Y. Hao, X. Wei, J. Gan and Z. Yang, Adv. Intell. Syst., 2021, 3, 2100035.
89. S. Song, B. Song, C.-H. Cho, S. K. Lim and S. M. Jeong, Mater. Today, 2020, 32, 46–58.
90. S. Chang, Y. Deng, N. Li, L. Wang, C.-X. Shan and L. Dong, Nano Res., 2023, 16, 9379–9386.
91. M. He, W. Du, Y. Feng, S. Li, W. Wang, X. Zhang, A. Yu, L. Wan and J. Zhai, Nano Energy, 2021, 86, 106058.
92. H.-J. Park, S. Kim, J. H. Lee, H. T. Kim, W. Seung, Y. Son, T. Y. Kim, U. Khan, N.-M. Park and S.-W. Kim, ACS Appl. Mater. Interfaces, 2019, 11, 5200–5207.
93. T. Lange, S. Reichenberger, M. Rohe, M. Bartsch, L. Kampermann, J. Klein, J. Strunk, G. Bacher, R. Schlögl and S. Barcikowski, Adv. Funct. Mater., 2021, 31, 2009323.
94. S. Moon Jeong, S. Song, S.-K. Lee and B. Choi, Appl. Phys. Lett., 2013, 102, 5.
95. Y. Han, Y. Bai, J. Bian, X. Guo, B. Liu and Z. Wang, ACS Appl. Polym. Mater., 2021, 3, 5180–5187.
96. J.-C. Zhang, X. Wang, G. Marriott and C.-N. Xu, Prog. Mater. Sci., 2019, 103, 678–742.
97. E. J. Bailey and K. I. Winey, Prog. Polym. Sci., 2020, 105, 101242.
98. F. Nie, B. Zhou and D. Yan, Chem. Eng. J., 2023, 453, 139806.
99. M. C. Wong, L. Chen, G. Bai, L. B. Huang and J. Hao, Adv. Mater., 2017, 29, 1701945.
100. L. Zhang, K. Shi, Y. Wang, L. Su, G. Yang, B. Huang, J. Kong, X. Dong and Z. L. Wang, Nano Energy, 2021, 85, 106005.
101. C. Wu, S. Zeng, Z. Wang, F. Wang, H. Zhou, J. Zhang, Z. Ci and L. Sun, Adv. Funct. Mater., 2018, 28, 1803168.
102. J.-H. Zhang, S. Liu, Y. Yuan and Y.-L. Chen, Chin. J. Polym. Sci., 2023, 41, 1162–1168.
103. W. J. Evans, C.-S. Yoo, G. W. Lee, H. Cynn, M. J. Lipp and K. Visbeck, Rev. Sci. Instrum., 2007, 78, 7.
104. J. Lei, W. Li, Y. Tang, Y. Cai, S. Wang, K. Dou and J.-C. Zhang, J. Mater. Chem. C, 2023, 11, 3588–3595.
105. J. Yang, X. Wu, J. Shi, B. Tong, Y. Lei, Z. Cai and Y. Dong, Adv. Funct. Mater., 2021, 31, 2108072.
106. P. Xiong, M. Peng and Z. Yang, iScience, 2021, 24, 1.
107. Y. Xiao, P. Xiong, S. Zhang, K. Chen, S. Tian, Y. Sun, P. Shao, K. Qin, M. G. Brik and S. Ye, Chem. Eng. J., 2023, 453, 139671.
108. P. Xiong, B. Huang, D. Peng, B. Viana, M. Peng and Z. Ma, Adv. Funct. Mater., 2021, 31, 2010685.
109. Y. Tu, Z. Zhao, J. W. Lam and B. Z. Tang, Matter, 2021, 4, 338–349.
110. J. Luo, Z. Xie, J. W. Lam, L. Cheng, H. Chen, C. Qiu, H. S. Kwok, X. Zhan, Y. Liu, D. Zhu and B. Tang, Chem. Commun., 2001, 1740–1741.
111. J.-A. Li, Z. Song, Y. Chen, C. Xu, S. Li, Q. Peng, G. Shi, C. Liu, S. Luo, F. Sun, Z. Zhao, Z. Chi, Y. Zhang and B. Xu, Chem. Eng. J., 2021, 418, 129167.
112. E. Hutter and J. H. Fendler, Adv. Mater., 2004, 16, 1685–1706.
113. A. Glass, P. F. Liao, J. Bergman and D. Olson, Opt. Lett., 1980, 5, 368–370.
114. H. Mertens and A. Polman, Appl. Phys. Lett., 2006, 89, 21.
115. W. Park, D. Lu and S. Ahn, Chem. Soc. Rev., 2015, 44, 2940–2962.
116. D. Huang, C. P. Byers, L.-Y. Wang, A. Hoggard, B. Hoener, S. Dominguez-Medina, S. Chen, W.-S. Chang, C. F. Landes and S. Link, ACS Nano, 2015, 9, 7072–7079.
117. F. Laible, A. Horneber and M. Fleischer, Adv. Opt. Mater., 2021, 9, 2100326.
118. I. Kaminska, T. Maurer, R. Nicolas, M. Renault, T. Lerond, R. Salas-Montiel, Z. Herro, M. Kazan, J. Niedziolka-Jönsson, J. Plain, P.-M. Adam, R. Boukherroub and S. Szunerits, J. Phys. Chem. C, 2015, 119, 9470–9476.
119. A. Campion and P. Kambhampati, Chem. Soc. Rev., 1998, 27, 241–250.
120. Y. Wang, C. Zhao, J. Wang, X. Luo, L. Xie, S. Zhan, J. Kim, X. Wang, X. Liu and Y. Ying, Sci. Adv., 2021, 7, eabe4553.
121. Y. Ma, Y. Du, Y. Chen, C. Gu, T. Jiang, G. Wei and J. Zhou, Chem. Eng. J., 2020, 381, 122710.
122. X. Yan, Q. Chen, Q. Song, Z. Huo, N. Zhang and M. Ma, Nano Res., 2021, 14, 275–284.
123. A. Das, U. Pant, C. Cao, R. S. Moirangthem and H. B. Kamble, Nano Res., 2023, 16, 1132–1140.
124. M. Charconnet, C. Kuttner, J. Plou, J. L. García-Pomar, A. Mihi, L. M. Liz-Marzán and A. Seifert, Small Methods, 2021, 5, 2100453.
125. Y.-J. Yeh, C.-J. Cho, J.-S. Benas, K.-L. Tung, C.-C. Kuo and W.-H. Chiang, ACS Appl. Nano Mater., 2023, 6, 10115–10125.
126. S. A. Lee and S. Link, Acc. Chem. Res., 2021, 54, 1950–1960.
127. S. A. Maier, Plasmonics: fundamentals and applications, Springer, 2007.
128. C. F. Bohren and D. R. Huffman, Absorption and scattering of light by small particles, John Wiley & Sons, 2008.
129. G. Mie, Ann. Phys., 1908, 330, 377–445.
130. P. Nordlander, C. Oubre, E. Prodan, K. Li and M. Stockman, Nano Lett., 2004, 4, 899–903.
131. P. K. Jain, W. Huang and M. A. El-Sayed, Nano Lett., 2007, 7, 2080–2088.
132. C. Tabor, R. Murali, M. Mahmoud and M. A. El-Sayed, J. Phys. Chem. A, 2009, 113, 1946–1953.
133. U. Fano, Il Nuovo Cimento (1924–1942), 1935, vol. 12, pp. 154–161.
134. U. Fano, Phys. Rev., 1961, 124, 1866.
135. W. Tao, H. Hu, O. Guillot, T. Maurer and M. Fleischer, Adv. Photonics Res., 2023, 4, 2300012.
136. S. Jia, Z. Li and J. Chen, Opt. Express, 2021, 29, 21358–21368.
137. S. Tavana and S. Bahadori-Haghighi, Plasmonics, 2022, 17, 2441–2449.
138. A. Maradudin, I. Simonsen, J. Polanco and R. Fitzgerald, J. Opt., 2016, 18, 024004.
139. A. A. Darweesh, S. J. Bauman, D. T. Debu and J. B. Herzog, Nanomaterials, 2018, 8, 809.
140. Z. Tang, J. Wu, X. Yu, R. Hong, X. Zu, X. Lin, H. Luo, W. Lin and G. Yi, ACS Appl. Mater. Interfaces, 2021, 13, 9281–9288.
141. S. Jo, J. Y. Woo, J. H. Oh, C.-W. Song, H. M. Yang and C.-S. Han, ACS Appl. Mater. Interfaces, 2022, 14, 23734–23742.
142. A. M. Steiner, M. Mayer, M. Seuss, S. Nikolov, K. D. Harris, A. Alexeev, C. Kuttner, T. A. König and A. Fery, ACS Nano, 2017, 11, 8871–8880.
143. V. Gupta, P. T. Probst, F. R. Goßler, A. M. Steiner, J. Schubert, Y. Brasse, T. A. König and A. Fery, ACS Appl. Mater. Interfaces, 2019, 11, 28189–28196.
144. P. Jia, D. Kong and H. Ebendorff-Heidepriem, ACS Appl. Nano Mater., 2020, 3, 8242–8246.
145. D. Grzelak, M. Tupikowska, D. Vila-Liarte, D. Beutel, M. Bagiński, S. Parzyszek, M. Góra, C. Rockstuhl, L. M. Liz-Marzán and W. Lewandowski, Adv. Funct. Mater., 2022, 32, 2111280.
146. Z. Li, Q. Fan and Y. Yin, Chem. Rev., 2021, 122, 4976–5067.
147. P. Kabtiyal, A. Robbins, E. Jergens, C. E. Castro, J. O. Winter, M. G. Poirier and E. Johnston-Halperin, Nano Lett., 2024, 24, 3097–3103.
148. M. Semple, A. C. Hryciw, P. Li, E. Flaim and A. K. Iyer, ACS Appl. Mater. Interfaces, 2021, 13, 43209–43220.
149. N. Osaka, M. Ozawa, T. Matsuyama, K. Wada and K. Okamoto, Opt. Mater. Express, 2023, 13, 1504–1512.
150. V. Gupta, S. Sarkar, O. Aftenieva, T. Tsuda, L. Kumar, D. Schletz, J. Schultz, A. Kiriy, A. Fery and N. Vogel, Adv. Funct. Mater., 2021, 31, 2105054.
151. R. Fu, D. E. Gómez, Q. Shi, L. W. Yap, Q. Lyu, K. Wang, Z. Yong and W. Cheng, Nano Lett., 2020, 21, 389–396.
152. R. Fu, T. Warnakula, Q. Shi, L. W. Yap, D. Dong, Y. Liu, M. Premaratne and W. Cheng, Nanoscale Horiz., 2020, 5, 1515–1523.
153. J. Marae-Djouda, A. Gontier, R. Caputo, G. Lévêque, B. Bercu, Y. Madi, G. Montay, P.-M. Adam, M. Molinari, S. Stagon and T. Maurer, ACS Appl. Nano Mater., 2018, 1, 2347–2355.
154. H. Matsui, A. Momose, H. Yoda and A. Fujita, ACS Appl. Mater. Interfaces, 2023, 15, 50447–50456.
155. A. Nauman, H. S. Khaliq, J.-C. Choi, J.-W. Lee and H.-R. Kim, ACS Appl. Mater. Interfaces, 2024, 16, 6337–6347.
156. P. Güell-Grau, F. Pi, R. Villa, O. Eskilson, D. Aili, J. Nogués, B. Sepúlveda and M. Alvarez, Adv. Mater., 2022, 34, 2106731.
157. M. S. Rodrigues, J. Borges and F. Vaz, Sensors, 2022, 22, 1375.
158. A. Yang, A. J. Hryn, M. R. Bourgeois, W.-K. Lee, J. Hu, G. C. Schatz and T. W. Odom, Proc. Natl. Acad. Sci. U. S. A., 2016, 113, 14201–14206.
159. G. Zhang, C. Lan, H. Bian, R. Gao and J. Zhou, Opt. Express, 2017, 25, 22038–22045.
160. Y. F. Liu, R. Zhang and P. V. Braun, Part. Part. Syst. Charact., 2020, 37, 1900463.
161. D. Wang, M. R. Bourgeois, W.-K. Lee, R. Li, D. Trivedi, M. P. Knudson, W. Wang, G. C. Schatz and T. W. Odom, Nano Lett., 2018, 18, 4549–4555.
162. K.-H. Chang, J.-S. Cheng, T.-W. Lu and P.-T. Lee, Opt. Express, 2018, 26, 33215–33225.
163. W. Tao, F. Laible, A. Hmima, T. Maurer and M. Fleischer, Nano Convergence, 2023, 10, 15.
164. M. L. Tseng, J. Yang, M. Semmlinger, C. Zhang, P. Nordlander and N. J. Halas, Nano Lett., 2017, 17, 6034–6039.
165. C. Zhang, J. Jing, Y. Wu, Y. Fan, W. Yang, S. Wang, Q. Song and S. Xiao, ACS Nano, 2019, 14, 1418–1426.
166. W. Chen, W. Liu, Y. Jiang, M. Zhang, N. Song, N. J. Greybush, J. Guo, A. K. Estep, K. T. Turner, R. Agarwal and C. R. Kagan, ACS Nano, 2018, 12, 10683–10692.
167. R. F. Tiefenauer, T. Dalgaty, T. Keplinger, T. Tian, C. J. Shih, J. Vörös and M. Aramesh, Small, 2018, 14, 1801187.
168. F. Lütolf, D. Casari and B. Gallinet, Adv. Opt. Mater., 2016, 4, 715–721.
169. D. Feng, H. Zhang, S. Xu, L. Tian and N. Song, Nanotechnology, 2017, 28, 115703.
170. K. Liu, H. Ding, S. Li, Y. Niu, Y. Zeng, J. Zhang, X. Du and Z. Gu, Nat. Commun., 2022, 13, 4563.
171. A. Biswas, N. Lemcoff, O. Shelonchik, D. Yesodi, E. Yehezkel, E. Y. Finestone, A. Upcher and Y. Weizmann, Nat. Commun., 2023, 14, 6355.
172. M. Runowski, S. Goderski, J. Paczesny, M. Ksiezopolska-Gocalska, A. Ekner-Grzyb, T. Grzyb, J. D. Rybka, M. Giersig and S. Lis, J. Phys. Chem. C, 2016, 120, 23788–23798.
173. M. C. Sekhar, E. Veena, N. S. Kumar, K. C. B. Naidu, A. Mallikarjuna and D. B. Basha, Cryst. Res. Technol., 2023, 58, 2200130.
174. J. Xue, H. Xiang, Y. Zhang, J. Yang, X. Cao and Z. Wang, Energy Environ. Mater., 2024, 7, e12492.
175. L. Kong, W. Li, T. Zhang, H. Ma, Y. Cao, K. Wang, Y. Zhou, A. Shamim, L. Zheng, X. Wang and W. Huang, Adv. Mater., 2024, 2400333.
176. X. Wang, M. Que, M. Chen, X. Han, X. Li, C. Pan and Z. L. Wang, Adv. Mater., 2017, 29, 1605817.
177. Y. Zhuang, X. Li, F. Lin, C. Chen, Z. Wu, H. Luo, L. Jin and R. J. Xie, Adv. Mater., 2022, 34, 2202864.
178. C. Wang, R. Ma, D. Peng, X. Liu, J. Li, B. Jin, A. Shan, Y. Fu, L. Dong and W. Gao, InfoMat, 2021, 3, 1272–1284.
179. J. Xu, L. Peng, S. Yuan, S. Li, H. Zhu, L. Fu, T. Zhang and T. Li, Adv. Funct. Mater., 2024, 2410885.
180. B. Zhou, J. Liu, X. Huang, X. Qiu, X. Yang, H. Shao, C. Tang and X. Zhang, Nano-Micro Lett., 2023, 15, 72.

---