Application of ordered nanoparticle self-assemblies in surface-enhanced spectroscopy

Abstract

Surface-enhanced spectroscopy (SES), including surface-enhanced Raman spectroscopy (SERS), surface-enhanced infrared absorption (SEIRA), surface-enhanced fluorescence (SEF), surface-enhanced hyper-Raman spectroscopy (SEHRS), surface-enhanced second harmonic generation (SESHG), and surface-enhanced electrochemiluminescence (SEECL), is an emerging subject that has received significant attention from the research community over the past 40 years. The confined and strong enhanced electromagnetic fields generated from the surface of noble metal nanostructures are considered as the main cause of SES. More importantly, it has been proven that SES is strongly related to the orientation and inter-nanoparticle coupling of different nanostructures. This review article mainly focuses on the application of ordered nanoparticle self-assemblies in SES. Different strategies for the fabrication of ordered nanoparticle self-assemblies, the mechanisms of SES, and the application of SES have been reviewed in detail.

---

Jiaming Chen

Jiaming Chen is a PhD candidate at the College of Chemistry at Fuzhou University. His research interests focus on surface-enhanced Raman scattering (SERS) and the fabrication of different functional nano-optical devices for rapid SERS detection.

Longhua Guo

Longhua Guo is currently a full professor at the College of Chemistry at Fuzhou University. He obtained his PhD from Fuzhou University in 2009. He then worked as a postdoctoral research fellow at the School of Chemical and Biomedical Engineering, Nanyang Technological University, from 2009 to 2011. He then joined the Ministry of Education Key Laboratory for Analytical Science of Food Safety and Biology (Fuzhou University) and has worked there since 2011. His research interests focus on the fabrication of nano-optical-based analytical devices for environmental and biological applications.

Tie Wang

Tie Wang is a professor at the Institute of Chemistry, Chinese Academy of Sciences (ICCAS). He received his BSc (2002) from Xi’An Jiaotong University and PhD (2007) from Changchun Institute of Applied Chemistry, China. Then he worked as a postdoctoral fellow at the Rensselaer Polytechnic Institute, Troy, NY, and the University of Florida, Gainesville, FL, both USA. He joined ICCAS and was awarded the “Thousand Youth Talents Plan” in 2013. His research focuses on nanoparticle assemblies and their applications.

---

1. Introduction

Where to go for inspiration? Perhaps the best answer should be nature. Over billions of years, organisms in nature have evolved a variety of biological forms, which display many miraculous biological phenomena and mysteries. Among these, ordered micro- and nanostructures are abundant in biological systems, which exhibit remarkable performance including in optics and sensing.^1–3 Many insects, such as mosquito, have compound eyes composed of a large number of integrated optical units called ommatidia^1,4,5 (Fig. 1a and b). Compound eyes can improve the sensitivity of light vision because of the final periodic compound eye structure. Coincidentally, unique features of ordered microstructures in nature can be observed, including the colourful feather or wings of some butterflies, which result from the ordered arrangement and arrays of vertically aligned net-like skeleton structures.^6–9 Moreover, these ordered micro- and nanostructures can contribute to form a fine sensing system of higher sensitivity, like in the silk moth Bombyx mori, which has hundreds of sensilla on the surface of the antennal branches of one of its antennas.^10,11

Fig. 1 (a) Many creatures in nature have gorgeous colors and outstanding senses, which are based on some special microstructures of their organs; (Reprinted from ref. 5 and 8–11; Copyright 2007, John Wiley and Sons; Copyright 2003, National Academy of Sciences; Copyright 2012, John Wiley and Sons; Copyright 2005, John Wiley and Sons.) (b) SEM images of the organs’ microstructures from the creatures in (a). These special organs are made of ordered and sophisticated microstructures. (c) Various bio-inspired photonic structures have been explored, which can be summarized into photonic crystals in three orthogonal directions. (Reprinted from ref. 1, 6 and 10; Copyright 2012, John Wiley and Sons; Copyright 2009, John Wiley and Sons; Copyright 2006, American Chemical Society.) (d) Based on the periodic structures of these photonic crystals, concepts, like the Bloch theorem, energy band, and Brillouin zones, can be applied to photonic crystals; simultaneously, introducing the defect mode into the photonic band gap and modifying the photon density of states in the photonic crystals can influence the optical characteristics.

In some cases, these abovementioned examples exhibit sophisticated and hierarchic photonic structures, which can provide researchers an ideal reference point. Inspired by nature, various materials and optical devices have been explored to achieve desired optical features. These materials and devices have similar periodic structures, such as multilayer structures, cylinder arrays or stacking spheres, which can be defined as photonic crystals in three orthogonal directions¹² (Fig. 1c). Similar to the propagation of electrons through a periodic arrangement of atoms, the propagation of light occurring in the photonic crystals can be discussed using a similar explanation such as Bloch theorem or by the energy band and Brillouin zones. Additionally, the defect mode is an important concept that appears in the photonic band gap. When photonic crystals are applied to devices and sensors, and when the defect mode wavelength and the emission wavelengths of the active probes (dye or quantum dots) match, emission can be enhanced, due to a change in the density of states in the defect mode. Photonic crystals can be modified by changing the photon density of states; where the photon density of states can affect the optical response of active species near the photonic crystals^12,13 (Fig. 1d). Notably, these photonic crystals have remarkable potential in optical spectroscopy and surface-enhanced spectroscopy. With the development of nanomaterial sciences, photonic crystals have gained more and more attention.

In the past decades, the development of nanotechnology has attracted considerable attention.^14–16 The properties that nanoparticles exhibit differ from those of natural bulk samples of the same materials. This aspect has captivated researchers in various fields and seemingly provides almost unlimited promise. Besides the synthesis of individual nanostructures, the focus on nanotechnology extends to organizing the nanoparticles into an ordered and controllable formation. It is a great challenge for researchers to go from the macroscopic bulk cases to microscopic particles, then back to the macroscopic superstructures under scientific conditions, but undoubtedly, achieving controlled approaches for organizing such “building blocks” of tiny particles to construct specially designed structures would offer great potential for many applications. What has emerged in this aim is the self-assembly of nanoparticles which offers an effective and promising method for controlling the fabrication of superstructures with unique effects induced by the collective interaction of excitation, electromagnetic, optical and other coupling surface plasmons.¹⁵ In particular, these ordered self-assemblies display electronic, magnetic and optical properties, resulting in their great potential in different fields, including plasmon sensors, photocatalysis and applications of surface-enhanced spectroscopy (SES).

Surface-enhanced spectroscopy is an emerging branch of spectroscopy that combines plasmonic nanostructures and spectroscopy. Since 1974, a tremendous enhancement of Raman scattering was observed by Fleischmann et al.¹⁷ from pyridine absorbed on a roughened silver electrode. Surface-enhanced Raman spectroscopy (SERS) has become the subject of intensive study and an effective enhanced method for the application of Raman in widespread areas. In addition, after the SERS effect was discovered, surface-enhanced hyper-Raman spectroscopy (SEHRS),^18,19 surface-enhanced fluorescence (SEF),^20,21 surface-enhanced second harmonic generation (SESHG),^22–24 surface-enhanced infrared absorption (SEIRA),^25–27 surface-enhanced electrochemiluminescence (SEECL)^28–33 and other enhanced spectroscopies have emerged as powerful techniques, collectively known as surface-enhanced spectroscopy (SES), for the study of surface phenomena and the interactions between molecules and nanoparticles. A series of surface-enhanced spectroscopies were observed that have similarities in their mechanisms and effects, especially relying on the collective effects of nanoparticles. Numerous experiments have demonstrated that assemblies of nanoparticles generate coupling plasmons that can be used to further improve the performance of surface-enhanced spectroscopy. Current interests in the assemblies of nanoparticles focus on organizing the nanoparticles into an ordered geometry and developing the effects based on these assemblies that give rise to the sensitivity and reproducibility of these spectroscopies.

Herein, we start by categorizing the main driving forces and strategies of nanoparticle assemblies for constructing plasmonic nanostructures. Then, a theoretical paradigm of plasmon and coupling plasmon is briefly provided. Moreover, the similarities and differences among SERS, SEF and SEIRA based on plasmonic nanostructures are discussed. Finally, we review some recent works on the applications of nanoparticle assemblies for surface-enhanced Raman spectroscopy, speculating about ordered nanoparticle self-assemblies in future scientific research and technological applications.

2. Mechanisms and strategies of ordered nanoparticle assemblies

Self-assembly is depended on several key issues, including the interparticle weak forces, the properties of individual nanoparticles (size distribution/shape), the modified ligands and the assisted external methods. These interactions between nanoparticles, such as van der Waals force, electrostatic force, hydrophobic interaction and so on, are the common driving forces in the self-assembly process.³⁴ The different interactions are considered as one of the fundamental mechanisms that lead to individual nanoparticles self-assembling into diverse nanostructures.³⁵ In some cases, there is not only one interaction existing in an assembly process, multiple interactions in assembly can exist in a complex and interdependent manner. The layering of molecules (surfactant, ligand or biomolecules), usually capped around colloidal nanoparticles, can be used for controlling the aggregation process in a solvent.³⁶ Thus, the interactions and modified molecules play important roles in the particles or particle/environment. In addition, the size and shape of individual nanoparticles can influence the structure and morphology of assemblies.³⁷

Based on these factors and mechanisms, various strategies have been developed for fabricating ordered and plasmonic nanostructures. Herein, we give a brief illustration of the main interactions and the related issues, which may help us to review the mechanism and strategies of self-assembly.

2.1 Driven forces in self-assembly

2.1.1 van der Waals force. van der Waals (vdW) force, as a fundamental driving force, is widely used for guiding the self-assembly process.³⁸ The vdW forces are the residual attractive or repulsive forces originating from electromagnetic fluctuations due to the incessant movements of positive and negative charges within all types of atoms and molecules.³⁴ Typically, vdW forces include three different types: Keesom force, Debye force and London dispersion force, which appear respectively between permanent dipoles, between a permanent dipole and a corresponding induced dipole and between instantaneously induced dipoles.³⁹ The vdW force between atoms and molecules can be expressed by the following equation:where r is the distance between atoms/molecules and C_vdW is a constant describing the interacting species and the surrounding medium.

To understand and calculate the vdW force, different theories have been proposed. Among these, for spherical particles with a smooth surface, a simple and ideal model was used by Haymaker in 1937 to estimate the vdW force between pairs of atoms located within two particles; whereby the van der Waals' interaction energy between spherical bodies of radii R₁ and R₂ and with smooth surfaces could be approximated by:⁴⁰

where A is the Hamaker coefficient depending on the material properties (it can be positive or negative in sign depending on the intervening medium), and z is the centre-to-centre distance; i.e. the sum of R₁, R₂, and r (the distance between the surfaces):

z = R₁ + R₂ + r (3)

In the limit of the close-approach, the spheres are sufficiently large compared to the distance between them; i.e. r ≪ R₁ or R₂.

Eqn (2) for the potential energy function can be simplified to:

Note that, pairwise summation is based on microscopic theory assumptions. However, it inevitably neglects some complex effects, such as many-body interactions and retardation. Therefore, more approaches accounting for these effects were proposed by Lifshitz and Langbein.^41,42

The vdW-force-driven self-assembly, like the drop-evaporation method, induces the nanoparticles to form close-packed structures. Monolayer or multilayer superlattices have been obtained by vdW-force-driven self-assembly.^43,44 By using the vdW force, close-packed monolayer or multilayer superlattices can be organized^37,78 (Fig. 2). The droplet-evaporation method based on vdW force is widely used for the fabrication of close-packed superstructure assemblies. To highly control the geometry of a gold superlattice, the assembly procedure was carefully monitoring by tuning the experimental parameters, namely, the dispersing solvent, deposition temperature, gold nanoparticle (Au NP) concentration and chemistry of the supporting substrate.^44,45 The short-range interactions, such as vdW forces, contribute to the formation of the superstructure. Also, the capping agent chain length and composition, size and concentration ratio can strongly affect the final geometry.^46,47

Fig. 2 Self-assembly of close-packed monolayer and multilayer superlattices by using van der Waals forces. SEM images of monolayer superlattices: (a) octahedral gold nanocrystals; (b) cubic gold nanocrystals. (Reprinted from ref. 37; Copyright 2011, John Wiley and Sons) SEM images of multilayer superlattices: (c) top view of gold nanorods. (d) Side view of gold nanorod superlattices with 14 layers. (Reprinted from ref. 78; Copyright 2009, John Wiley and Sons.) (e) Self-assembled rhombic dodecahedral gold nanocrystals. (Reprinted from ref. 37; Copyright 2011, John Wiley and Sons.)

2.1.2 Electrostatic force. Electrostatic force also plays an important role in the self-assembly process. As we know, the nanoparticles are usually modified with protonation or deprotonation molecules or absorbed as charged ions on the surfaces, which may lead to the aggregation of nanoparticles by electrostatic interaction.⁴⁸ In an aqueous solution, the electrostatic interaction is strongly depended on factors like the pH value, ionic strength or type of the electrolyte.⁴⁹ Understanding the electrostatic attractive and electrostatic repulsive forces is beneficial for designing organized aggregates.⁵⁰ Electrostatic repulsions commonly protect colloidal nanoparticles against aggregation or coagulation between charged nanoparticles, which improves the colloidal stability. Conversely, electrostatic attraction occurring between oppositely charged nanoparticles can be helpful to control the concentration of the nanoparticles and the self-assembly of nanostructures.⁵¹ Electrostatic attraction is one of the most used tools in solution-based self-assemblies. Furthermore, by controlling the amount of surface charges, the ions, polyelectrolytes and molecules, different self-assembled structures can be obtained.⁵² Different types of clusters, such as dimers, trimers or larger aggregates, like core–satellites, can be fabricated by electrostatic forces⁵³ (Fig. 3).

Fig. 3 Self-assembly of a core–satellite structure by using electrostatic forces. (a) Schematic showing the steps involved in the self-assembly of core–satellite clusters. (b) TEM images showing different sizes of gold nanoparticles, which are employed as the core and satellites. (c) TEM images showing the successful assembly of core–satellite structures. (Reprinted from ref. 53; Copyright 2012, American Chemical Society.)

2.1.3 Other driving forces. To maintain the stability of colloids or the induced self-assembly of nanoparticles, different types of molecular interactions between nanoparticles or a physical magnetic field have been demonstrated, such as hydrogen bonding, hydrophobic interactions, covalent bonding, depletion forces, dipole–dipole interactions and magnetic interactions. Based on these interactions, many methods have been used for assembling the nanoparticles. For example, DNA-induced self-assembly is a common method dependent on hydrogen bonding with the DNA base.⁵⁴ Besides, hydrophobic self-assembly has been used for tailoring the surface chemistry of nanoparticles. Ag nano-octahedra were assembled into a larger-scale plasmonic monolayer at the oil/water interface, resulting in a square superlattice with Ag nano-octahedra standing on their vertices.⁵⁵ By ligand exchange with alkylthiols of increasing chain lengths, octahedron silver nanoparticle assemblies showed different forms (Fig. 4). Clearly, the diverse surface modification of nanoparticles has become one of the most useful tools in self-assembly.

Fig. 4 Nanoparticle self-assemblies by using other driving forces. (a) Schematic illustration of the fabrication of gold nano-dimers by using DNA-directed self-assembly. (Reprinted from ref. 54; Copyright 2013, American Chemical Society.) (b) Using increasingly hydrophobic Ag octahedra for self-assembly with various thiols, resulting in the different formations of superlattices. (Reprinted from ref. 55; Copyright 2015, Springer Nature.)

2.2 Shape effects for tuning the morphology of nanoparticle assemblies

Recent nanotechnology advances can provide nanoparticles of various sizes, shapes and dimensions, which is regarded as an opportunity to use them for fabricating the ordering of nanomaterials.^56,57 For different purposes, designing and fabricating a suitable morphology of nanoparticle assemblies is necessary.⁵⁸

The nanospheres or quasi-spherical nanoparticles are the simplest and the most common nanoparticles used as building blocks to construct more complex nanostructures. Simply, nanospheres can self-aggregate into many diverse structures, including dimers, trimers and larger number aggregates of nanospheres, which can then be applied for nanoprobes or nanotags.⁵⁹ With the aggregating number increasing, larger nanoparticle aggregates, such as nanochains or even monolayer and multilayer superstructures, can be achieved^60–66 (Fig. 5). Due to the symmetry of a nanosphere, the geometrical morphology of the assemblies is oriented with a close-packed plane, which results in the formation of fcc (face-centred cubic) and hcp (hexagonal close-packed) structures.³⁵ In addition to the shape effect of nanoparticles, the surface ligands also play a key role in the formation of a superlattice.⁶⁷ These close-packed nanostructures are widely used in fabricating enhanced substrates, because of the unique optical properties, and for coupling plasmon nanostructures. Notably, the close-packed structures have excellent sensitivity and reproducibility in surface-enhanced spectroscopies based on the short-range interparticle gap.⁶⁸

Fig. 5 Nanochains and nanoclusters, which aggregate with the increasing number of nanoparticles. (a) Chemical structures of three hydrophobic ligands 1–3. (b–d) TEM images of the single-, double- and multiple-line chains obtained in DMF/H₂O = 7 : 3 solution and after purification, where (b) 1 = stabilized AuNPs, (c) 2 = AuNPs, (d) 3 = AuNPs coated in PS₁₅₄-b-PAA₄₉ shells used as the monomers. (Reprinted from ref. 66; Copyright 2014, American Chemical Society.) (e–j) SEM images of individual gold nanoclusters with different diameter values and numbers; all the size bars are 100 nm. (Reprinted from ref. 65; Copyright 2010, John Wiley and Sons.)

By contrast, rod-shape nanoparticles have anisotropy so that they commonly exhibit two formations: ss (side-by-side) and ee (end-to-end). Generally, nanorods are organized to produce a variety of structures, ranging from a simple dimer to a chain, in the ss or ee form.⁶⁹ Additionally, rod-shaped nanoparticles can be assembled into superstructures.^70,71 There are more changes in close-packed nanorod assemblies than possible in nanosphere assemblies.^67,72 Close-packed rod-shaped nanoparticle assemblies can be categorized as a crystal, liquid crystal and more complex three-dimensional geometries, such as a needle-like shape.⁷¹ Noteworthily, vertical arrays of nanorods have been observed to assemble into the close-packed formation of hcp.^73,74 Furthermore, other anisotropic nanomaterials, which have a two-dimensional pattern, like a triangle, square or hexagon of the cross-section, can be assembled into ordered nanostructures for maximization of the tiling density⁷⁵ (Fig. 6).

Fig. 6 The two different formations: ss (side-by-side) and ee (end-to-end), of self-assembly of rod-shape nanoparticles. TEM images of hybrid cylindrical micelles formed from PS51k-b-P4VP18k (PDP)2.0 with various NRs (diameter, 6 nm; length, 12 nm) content: (a and c) 11 vol%; (b and d) 21 vol%, resulting in ss and ee, respectively. (Reprinted from ref. 69; Copyright 2013, American Chemical Society.) (e) SEM image of a gold nanorod array. (f) SEM image recorded at a tilt angle of 52°. (Reprinted from ref. 74; Copyright 2012, John Wiley and Sons.)

Additionally, the self-assembly of two nanoparticle species is also possible, rather than forming just size- and shape-uniform nanoparticles. The mixtures of nanoparticle shapes include spheres with spheres, spheres with rods, spheres with plates and so on, which is discussed in Section 2.3.4.

2.3 Strategies for the fabrication of plasmon nanostructures

To obtain the desired structures and functionalities of nanoparticles, various strategies and rules employing diverse shapes and different interactions, functional molecules and other external assisted methods have been extensively studied.⁷⁶

2.3.1 Self-assembly via surface modification. The prepared nanoparticles are usually capped by a layer of surface ligands, such as a surfactant, polymer, DNA or antibody/antigen.^77–80 Modifying the nanoparticle surface has become one of the most important approaches to nanoparticle assembly for achieving various structures.

DNA strands are perhaps the most versatile surface ligands for finely controlling nanoparticle assemblies.⁷⁷ Different structures can be assembled by varying the oligonucleotide sequence,⁸¹ while at the same time the experimental procedure remains unchanged (Fig. 7). Basically, the assembly scheme starts with the particles, which are functionalized with two pre-designed complementary single strands, respectively. A following hybridization of the complementary DNA strands leads to the assembly of the nanoparticles.^82,83 It was demonstrated that weak forces, like hydrogen bonding, electrostatic and hydrophobic interactions, cooperatively drive one DNA strand to assemble with its complement into a double helix. In these ways, extended assemblies, such as 2D or 3D nanoparticle aggregates, can be obtained.^77,84,85 Therefore, DNA-based self-assembly is emerging as a useful tool using a highly promising template for organizing nanomaterials in a pre-designed way.

Fig. 7 Schematic of how plasmonic nanoparticle assemblies can be designed and synthesized by using DNA nanotechnology and a range of plasmonic nanostructures. (Reprinted from ref. 81; Copyright 2011, Springer Nature.)

In addition to the DNA linkers, some biomolecules based on enzyme-linked immunosorbent assay (ELISA) have been used as the linker for the conjunction of nanomaterials.⁸⁶ These biomolecule linkers, such as an antibody with antigen, biotin with streptavidin, a virus and protein, exhibit excellent specific binding for self-organized aggregation.⁷⁹

Besides these biomolecular ligands, polymers, as versatile and diverse functional ligands, also play an important role in surface modification.⁸⁷ Utilizing polymer and polymer properties can enable obtaining not only polymer functional nanoparticle composites but also polymer-nanoparticle assemblies. Many polymers, such as linear, branched and block copolymers, have been applied for the process of self-assembly. These polymer-nanoparticle assemblies can be tuned based on the polymer's molecular weight, chemical nature, architecture, persistence length and surrounding solvent for obtaining pre-designed geometrical morphologies.⁷⁹ Generally, most of polymer-nanoparticle assemblies can follow three strategies: (1) assembly of the polymer-media nanoparticles into aggregates (e.g. tuning the aggregating by utilizing dendrimers of different generations⁸⁸), (2) utilizing the polymer as a matrix to induce nanoparticles into an ordered and anisotropic oriented formation (e.g. nanoparticles grafted with polystyrene and orderly embedded in a polystyrene matrix⁸⁹) or (3) acting as a functional element (e.g. possessing an electronic property⁹⁰).

In most cases, surfactants have been utilized as surface stabilizers or templates in the synthesis of nanoparticles. The research evidence has demonstrated that an excess of surfactant is beneficial for ordered superlattice structures by controlling the dewetting of the solvent.⁹¹ In the droplet-evaporation self-assembly process, these surfactants, such as cetyltrimethylammonium bromide (CTAB), play a key role in the formation of ordered superstructures.⁹² Generally, this approach is simply done by casting a droplet of the nanoparticle solution onto a solid support (e.g. a glass slide or silicon wafer) and keeping it at a controlled temperature and humidity until it is dried.⁹² Besides surfactants, the effect of ions makes a contribution to the nanoparticle aggregation during the process of solvent evaporation, and can lead to a change in the surface charge of the nanoparticles.⁴⁹ A salt (e.g. sodium chloride) can be added into the nanoparticle solution for the aggregation, and this reduces the stabilizing energy barrier in the interaction potential between charge-stabilized colloidal particles.⁹³

2.3.2 Self-assembly via tuning the environmental conditions. Besides modifying the surface ligands, tuning the environmental conditions is another effective approach for organizing nanoparticle assemblies. These environmental conditions include the solvent and applying an external field, such as a magnetic field or electric field.

Oil–water interfacial self-assembly was employed for organizing particles to fabricate larger nanostructures at the oil–water interface, which was based interactions of the surface ligands of nanoparticles and their environment.^94,95 This liquid–liquid assembly, analogous to the case of Pickering emulsions, can generate a resistant film at the interface between two immiscible phases.⁹⁶ For example, Wang and co-workers⁵⁵ assembled three wafer-scale 2D plasmonic superlattices at the oil–water interface by tailoring the surface chemistry of Ag octahedra. Tuning the solvent and the ligands can strongly influence the shape of the final nanostructures at the interface, which is an effect driven by the reduction in interfacial energy^55,97–99 (Fig. 8).

Fig. 8 Different strategies of interfacial self-assemblies. (a) Schematic of the steps involved in the interfacial self-assembly at the oil–water interface by controlling the surface wettability of Ag octahedra. (Reprinted from ref. 55; Copyright 2015, Springer Nature.) (b) Cross-section of the three-phase interface (left). Schematic representation of the stages of film formation of the Ag nanowires at the three-phase interface (right). (Reprinted from ref. 94; Copyright 2010, John Wiley and Sons.)

Applying an external field, like an electric field or magnetic field, can also facilitate the assembly of nanostructures. By the use of an external field, nanoparticles, especially the anisotropic sort, can be directly oriented from individual building blocks to ordered assemblies.^100–102 Zhu et al. and co-workers controlled the assembly of polystyrene-tethered gold nanorods in anodic aluminium oxide channels by the assistance of an electric field.¹⁰³ The electric field played a key role in determining the ordered assemblies, which provided the driving force to reorientate the anisotropic NRs. A magnetic field, analogously, can influence the orientation of magnetic nanoparticles (i.e. with a permanent or field-induced magnetic dipole moment).¹⁰⁴ Wang et al. reported¹⁰⁵ the synthesis of ellipsoidal colloidal particles with anisotropy in both morphology and magnetic properties, which were used as building blocks for assembling 3D ordered structures with unique tuneable photonic properties.

2.3.3 Self-assembly via templates. Self-assembly via the use of a template is another way to facilitate the assembly of nanoparticles in a prepared confined geometry. Typically, the template-assisted assemblies could be classified into two broad categories: (1) the materials are physically deposited in a specific geometry or through chemically directing the growth to occur at a specific location or in a specific direction; (2) the prepared nanoparticles as building blocks are utilized for assembling larger, well-defined nanostructures in a prefabricated, ordered geometry.

With the geometrically constrained features (e.g. pores, channels), different assembly patterns of nanostructures can be fabricated and controlled by the use of pre-designed templates.^{69,73,106,107} For example, nanorod arrays are extensively explored based on the nanoporous templates.¹⁰⁸ Lithographic techniques have been used to create suitable topographical templates and patterns, which is beneficial for organizing prepared nanomaterials into designed structures. These spatially selective assembly of nanoparticles at the desired location can be achieved by an electrostatically mediated assembly of particles and confined deposition nanoparticles.¹⁰⁹ Nepal et al. reported a method involving the assembly of gold nanorods on chemically nanopatterned surfaces with a simultaneous control of both the interparticle spacing and local orientation.¹¹⁰ The combination of different driving forces and lithographic techniques, such as capillary force, could organize nanostructures into ordered assemblies in the confined area.^74,111 Flauraud et al. reported a method involving the insertion of nanorods into the traps as resilience against the receding suspension front, followed by drying of the residual solvent, which could deterministically direct the capillary assembly of Au nanorods to attain a simultaneous control of the position, orientation and interparticle distance at the nanometre level¹¹¹ (Fig. 9).

Fig. 9 The strategy of evaporating self-assembly via a pre-designed template. (a) Schematic of the capillary assembly of nanoparticles onto topographical traps of a low-wetting substrate. The three sequential stages of capillary assembly are illustrated from left to right: nanoparticles are successfully assembled if they (1) are inserted into a trap, (2) are not removed from the trap by capillary doctor blading or by friction with the sliding accumulation zone and (3) withstand displacement and reorientation within the trap during the final solvent drying. (b) Examples of 2D patterns of Au nanorods fabricated by a topographically templated capillary assembly. All scale bars, 250 nm. (Reprinted from ref. 111; Copyright 2016, Springer Nature.)

2.3.4 Self-assembly of simplex and binary nanoparticles.
Simplex nanoparticle assemblies. Simplex nanoparticles, which have a similar composition and shape, can assemble into new organized assemblies. These ordered assemblies of nanoparticles not only exhibit the function of the individual nanoparticles, but also the collective physical effects from the integral superstructure. The forms of nanoparticle assemblies can be classified by the diversity of assemblies based on: non-dense-packed structures and dense-packed structures.

Non-dense-packed assemblies range from dimers of nanoparticles with different individual shapes to more complex and larger aggregates, such as chains and core–satellites. Most studies of these assemblies have been performed with regard to the structural control and the tuning of the interparticle distance, which provide the possibility to generate new properties that can be further used in optical sensing at the single nanostructure level. Dimers of nanoparticles are the most extensively studied among these assemblies, due to their simplex structure. Moreover, to attain the more complex assemblies, either nanochains or nanoclusters, the study of dimers as the basic unit is necessary. In the early stage, large electromagnetic field enhancements at the junctions were observed with pairs of spherical nanoparticles, when the surface plasmons are excited. The unique physical properties confined between the neighbouring nanoparticles have attracted plenty of attention. For the self-assembly of nanoparticles, most attempts have been done not only on the nanoparticle material, size and shape but also on the interparticle spacing, the number of nanoparticles and the structures of the assemblies.

Controlling the number and structures of particles for linear chains of metal nanoparticles, such as one-dimensional (1D) assemblies, offer novel means for exploring complex nanostructures and new properties.^66,112 To effectively control the assemblies, surface modification of the nanoparticles is widely used. Barrow and co-workers made gold nanoparticles assemble into chains of differing lengths (1–6 particles) by using thiolated single strand oligonucleotides as a linker.¹¹³ Based on the experimental data, an exponential model was developed that allowed determination of the asymptotic maximum resonance at a chain length of 10–12 particles. More complex structures of clusters have also been designed, which are not confined to 2D forms. Nanoparticles as building blocks could be precisely fabricated into pre-designed 3D structures via molecular linkers, such as DNA, due to its highly customizable structure. For example, Tian et al.¹¹⁴ used the DNA origami octahedron as the frame to fabricate clusters with various symmetries and particle compositions. Notably, this work could be extended to designed arrangements of nanoparticles in 1D and 2D arrays by prescribing specific vertices of the octahedron as nanoparticle connecting sites.

The studies of nano-dimers have extended to other shapes of nanoparticles, such as rod-, cube- and prism-shaped, which exhibit more outstanding EM optical performances rather than the individual nanoparticles.

Binary nanoparticle assemblies. Different nanoparticles, which may have differences in the material, size and shape, can be assembled into so-called binary nanoparticle assemblies involving more complex structures. Studies of these binary nanoparticle assemblies have received relatively less attention compared to the simplex nanoparticle assemblies. On the other hand, multiple properties can be attained due to the different organization of assemblies.

Non-dense-packed assemblies are not only fabricated by particles with the same properties but also by different types. The blocks for these assemblies may have different properties from more than two types, such as the component, size and shape. Typically, heterocomponent dimers of gold and silver, silver and copper nanoparticles have been observed and explored in recent years.¹¹⁵ For example, Sheikholeslami and co-workers fabricated a silver/gold nanoparticle heterodimer via the DNA assembly method and systematically investigated the effects of symmetry breaking by scattering spectroscopy from single dimers.¹¹⁶ Lombardi et al. demonstrated the fabrication of asymmetric dimers formed from a gold nanoparticle bound to a silica-coated silver nanoparticle (Ag@SiO₂ NP), which were separated with the silica shell of Ag@SiO₂ NP.¹¹⁷ To achieve smaller gaps between these heterodimers, a DNA origami platform was used as a mediate linker to build a hetercomponent dimer that involved the same size (40 nm) gold and silver nanospheres by Weller et al.¹¹⁸ These asymmetric dimers open up many possibilities for nanoscale manipulation of the electromagnetic field and study of the collective plasmon modes by playing with the constituent particle composition, size, shape and interparticle distance (Fig. 10).

Fig. 10 Schematic of an asymmetric nano-dimer. (a) TEM images of asymmetric dimers composed of a 30 nm silver nanoparticle and 40 nm gold nanoparticle. (Reprinted from ref. 116; Copyright 2010, American Chemical Society.) (b) Model morphology of an Au@Ag@SiO₂ hetero-dimer formed by spherical particles. (c) TEM image of the investigated dimer formed by quasispherical particles with R_Au = 29 nm, R_Ag = 20 nm and a SiO₂ shell thickness R_t–R_Ag = 16 nm. (Reprinted from ref. 117; Copyright 2013, American Chemical Society.) (d) Schematic (left) and SEM image (right) of an individual Ag–Au heterodimer (backscatter contrast). (Reprinted from ref. 118; Copyright 2016, American Chemical Society.)

Besides the heterodimers, satellite-shaped binary nanostructures can be assembled by two different types of nanoparticles with a larger number. Yoon and co-workers presented asymmetric core–satellite nanostructures built by a core gold nanoparticle (51 nm) and several smaller gold nanoparticles (13 ± 3 nm) as “satellites”.¹¹⁹ The gap between the satellites and core nanoparticle could be controlled by varying the length of the alkane dithiol linkers. Besides the use of small molecule linkers, core–satellite assemblies have also been built via DNA and hyperbranched polymer linkers, which allows programmable assembly.^120–122 Similarly, different shapes of nanoparticle were assembled to give a core–satellite nanostructure. For example, Xu et al. utilized a gold nanorod and multiple nanospheres as the core and satellites of the nano-assemblies, respectively, to fabricate a core–satellite nanostructure.¹²³ Based on the selective modification of nanorods with DNA oligomers, three types of assemblies with distinct elements of regiospecificity and high synthetic yield can be attained, which provides the possibility for multi-nanosphere, multi-nanorod superstructures with increased topological complexity and regiospecificity.

3. Principle of surface-enhanced spectroscopy on ordered nanoparticle assemblies

Since the discovery of SERS, surface-enhanced technology has already been applied to fluorescence, second harmonic generation and infrared absorption spectroscopies, which led to the development of a new subject called “surface-enhanced spectroscopy (SES)” that has been widely used in various fields in chemistry, physics, biology and new materials over the last 40 years.^124–126 While some of the applications of SES may be surprising, the mechanisms of enhancement are still not clear at all, which captures the attention of researchers worldwide¹⁴ to investigate the mechanism of enhancement and the origin of these plasmons that play such an important role and which is closely related to the nanostructures. These nanostructures range from a single nanosphere, dimer or larger aggregates to monolayer or multilayer close-packed superstrucures.⁵⁸

The nanostructure-based optical properties, including scattering, absorption and extinction, lead to a variety of spectroscopic techniques.^68,124 In particular, the enormous advances in nanotechnology have led to nanoparticles with a controllable size and shape, and complex aggregates and arrays in pre-designed geometrical shapes. For nanoparticle aggregates or arrays, it is possible for the plasmon resonances in each particle to be coupled, leading to shifts in the plasmon wavelengths or changes in the intensities. These effects are strongly influenced by the interparticles’ distance. In this case, one can combine plasmon-resonance effects in the nanoparticles with photonic resonance associated with the nanostructure to produce new coupling plasmon resonance.¹²⁷ The coupling of nanostructures can also influence the spectroscopic intensities, leading to the possibility of photonic contributions to the enhancement effects in combination with plasmonic enhancements.⁵⁸ Therefore, for better improving the performance of SES, understanding the relationship between SES and plasmons is necessary.

3.1 Localized surface plasmon resonance

From classical electromagnetism, internal polarization has been used for describing Maxwell's equations for media. Plasmon polaritons occur in a metal while the energy wave is shared between the electromagnetic field oscillations and the internal excitations of the medium. Common noble metals, such as Au, Ag, Cu and Pd, have exhibited remarkable optical properties and plasmon resonances.¹²⁸ Additionally, other metals have demonstrated that they can generate plasmon resonance. For example, Al exhibited promising plasmon resonance at wavelengths ranging from deep-UV to NIR.¹²⁹ On planar interfaces, surface plasmon polaritons (SPPs) can be simply viewed as electromagnetic surface modes that are relevant to an electromagnetic (EM) field and plasmonics, which are associated with smooth, thin films of silver and gold with thicknesses in the 10–200 nm range. Propagating surface plasmon polariton (PSPP) is found at boundaries between the dielectric and metal conductor. PSPP can propagate over relatively long distances on the surface, which opens up the possibility to make use of the mechanism to manipulate and guide light for the design of plasmonic chips. The coupling of light makes the SPPs sensitive to external parameters, which leads to sharp surface plasmon resonance (SPR).

To understand, by way of analogy, the modes called localized surface plasmon polaritons (LSPs), involving small metallic particles instead of a planar surface, a simple model, involving a spherical metallic particle, was considered for better studying the electromagnetic (EM) field when molecules (probe or luminophore) are in the vicinity of particle. At the particle surface and its local area, the electromagnetic field strength increases, while the surface plasmon resonances are excited by the light. According to Mie theory, an approximation can help with the understanding of the physics involved. For a spherical metal nanoparticle with a size much smaller than the wavelength of light, the surface plasmon oscillation is dominated by the dipolar mode with a polarizability α given by:^130,131

where V is the particle volume, ε₀ is the vacuum permittivity, ε(ω) = ε_r(ω) + iε_i(ω) is the complex frequency-dependent dielectric function of the metal and ε_m is the dielectric constant of the surrounding medium. A resonance occurs roughly at the electromagnetic frequency ω, where ε_r = −2ε_m, thus determining the surface plasmon resonance (SPR) frequency. Obviously, the frequency of LSP of a spherical particle depends on its size, composition and environment.

When the metallic particle is small to nanoscale, the LSPs of the nanoparticle can be excited by an incident wave, which then couple efficiently, resulting in a resonant optical response at the LSP frequency, with the appropriate polarization and frequency. Different from SPR, localized surface plasmon resonance (LSPR) is the collective oscillation of electrons within the nanostructures excited by incident light of a larger wavelength than the size of the nanostructure. The coherent oscillation results in strong light scattering and a unique surface plasmon absorption spectrum. The intensity and peak location of the absorption and scattering spectra are not only highly dependent on the composition, but also on the nanoparticle shape, size and the surrounding medium. For a spherical particle with radius a, the extinction efficiency for the dipole resonance is:¹³²

where ε_r and ε_i are the real and imaginary parts of the dielectric function, respectively, and λ is the wavelength. These resonances for the electromagnetic field are important as they form the basis for most surface-enhanced spectroscopies. When the resonance occurs in metallic nanoparticles, such as gold or silver, it results in amplification of the electric field E in the vicinity of particle surfaces, whereby |E|² can be 100–10 000 times greater in intensity than the incident field.

Note that, compared with spheres, the effect of shape is more complex. Even though spherical particles provide an excellent model system, some evidence has pointed out that some different shapes have stronger plasmon resonance for the electromagnetic field. These aggregates of nanoparticles, regarded as aggregated small spherical particles, exhibit remarkable enhancement for spectroscopy, like SERS.

3.2 Coupling plasmon resonances

Various new designs of surfaces for enhancement have been studied, such as isolated particles, gratings, randomly roughened surfaces and assemblies of particles on surfaces.

The dimer of spherical particles has served as the related simplest structure for studying the principle of plasmon coupling and the influence on spectroscopic enhancements. From these experiments, the dimer exhibited an enhancement of an intense and localized electromagnetic field. The collective oscillation of free electrons occurs in noble metal nanoparticle assemblies, resulting in an enhancement of the electromagnetic field at the junction regions (nanoscale gap), which are referred to as so-called hot spots. The coupling of EM fields strongly influences the optical properties of molecules and non-metal nanoparticles in the vicinity of the hot spots. The coupling of electromagnetic fields depends on the shape, arrangement of the nanoparticles and the interparticle spacing. A dimer of nanoparticles as the model can be used for better understanding the plasmon resonance:^38,133

where n₀ is the positive charge number, l is the azimuthal quantum number and m_e is the mass of the electron. When the size of the overall particle system is smaller than the plasmon wavelength (D + 2R < λ/4), the dynamics of plasmons could be determined by instantaneous Coulombic interactions between the surface charges:

The interaction V^(m)_ij describes the coupling between two noninteraction plasmon modes i and j:

where X_i is the distance between the centre of particle N_i to a point on the surface of particle N_j, while θ_j and θ_i are the corresponding polar angles.

Also, the incident light induced field is polarized along the axis of the dimer or transverse to the sphere, resulting in the LSPR intensity and shift (Fig. 11). When the coupling plasmon is extended to large-scale particle arrays or lattices, these types of nanostructures can exhibit a considerable strong field by varying the spacing, distribution and geometric of particles. Coupled dipoles can help us understand the coupling plasmon and optical properties.

Fig. 11 Plasmon coupling in spherical dimers with small gaps. (a) Constructive dipolar coupling in spheres with light polarized along the dimer axis. (b) Destructive interference between dipoles with light polarized across the dimer axis. Electric field intensity of constructively (c) and destructively (d) interfering 10 nm radius Ag spheres separated by a 5 nm gap. Extinction spectra of a 10 nm radius sphere Ag dimer with decreasing interparticle gaps with light polarized along (e) and across (f) the dimer axis. The interparticle gaps in (e and f) (from darkest to lightest trace) are 50, 40, 30, 20, 10, 5, 4, 3, 2 and 1 nm. The scale bar is 10 nm. (Reprinted from ref. 58; Copyright 2016, American Chemical Society.)

When considering the coupled-dipole approximation, we consider an array of N particles whose positions and polarizabilities are denoted r_i and α_s. For a given wavelength λ, this field is:^58,132

where the local field E_loc,i is the sum of the incident and retarded fields of the other N − 1 dipoles. The induced dipole P_i in each particle in the presence of an applied plane wave field is P_i = α_sE_loc,i (i = 1,2,…,N), where E₀ and k = 2π/λ are the amplitude and wave vector of the incident wave, respectively. The interaction matrix between dipoles A is then expressed aswhere r_ij is the vector from dipole i to dipole j. The polarization vectors are obtained by solving 3N linear equations of the formAfter obtaining the polarization vectors, we can calculate the extinction cross-section using:

In the case of an infinite array of particles, it is possible to generate an analytical solution to eqn (5) for the case where the wave vector is perpendicular to the array axis (or plane). By assuming that the induced polarization in each array element is the same, this leads to the following expression for the polarization and extinction cross-section of each particle, respectively:

where S is the retarded dipole sum related to the geometric arrangement of particles:where θ_ij is the angle between r_ij and the polarization direction, which gives analytical expressions for the dipole sums for 1D arrays, but it is not difficult to evaluate this numerically for 1D or 2D arrays.

Thus, the retarded dipole sum S greatly influences the optical properties of the array. Using the coupled-dipole approximation, a variety of interesting optical phenomena have been identified that can be controlled with the nanoparticle size, spacing and interparticle distance. As such, 2D arrays can provide enhanced sensing or detection opportunities due to the enhanced sensitivity of the lattice plasmon to changes in the dielectric environment of the particle and local fields, which are further amplified compared to the plasmon enhancement with isolated nanoparticles.

3.3 Electromagnetic field and optical effects

The electromagnetic field can be induced by the LSPR of noble metal nanoparticles and nanoparticle assemblies and some factors, like size, shape and arrangement.^134,135 By using the discrete dipole approximation (DDA),¹³² it was found that the largest E² value for an examined dimer was a factor of 10 larger than the for the monomer examined. For individual particles, Schatz and co-worker found that the largest E² values were very similar for triangular prisms, oblate spheroids or cylindrical rods, with E² always being less than 10⁴. For a dimer of nanoparticles, we find E² values closer to 10⁵ for structures where the particle separation is 2 nm. Also, the enhancements are strongly dependent on the interparticle distance and the particle size.

Noble metal nanoparticles interact remarkably with the EM field and their plasmonic resonance frequencies are located in the visible light region due to the d–d band transitions.¹³⁶ From the absorption spectra, a decrease in the plasmon resonance band intensity and an increase in bandwidth with decreasing particle size can be seen.¹³⁷ As the silver nanoparticles are assembled into a monolayer, the plasmon-resonance peak is shifted to energies lower than those obtained for dilute solutions of isolated particles. Furthermore, extending to the 2D and 3D, a more pronounced spectral shift is observed in 3D than in 2D. The coupling of excitonic and plasmonic properties in nanoparticle assemblies leads to some optical effects, such as enhanced emission, a wavelength shift of the photoluminescence emission and a nonlinear Fano resonance.^138–140

3.4 The theory of surface-enhanced spectroscopy on nanostructures

Recently, the tremendous development of nanomaterials and self-assembly technology offers opportunities for research into plasmonics and increasing attention to SES.^124–126 A series of theories and approaches have been developed for attaining more efficient coupling plasmons that could further accelerate the development of SES.

3.4.1 Surface-enhanced Raman scattering. One technique to highlight is SERS as an ultrahigh sensitive method, even for single-molecule-level detection. For SERS enhancement, long-range electromagnetic (EM) enhancement and short-range chemical enhancement (CE) are the primary theoretical mechanism. Schatz discussed the EM mechanism and its dependence on the large-scale features (10–100 nm).¹⁴¹ When the frequency of incident light is resonant with the dipolar plasmon, due to collective oscillations of the nanostructure's surface electrons, the nanostructure radiates a dipolar field, which is a coherent process with the exciting electric field.¹²⁷ With regard the special position around the nanostructure, it was found that the local field was redistributed and the EM field was enhanced sharply – these areas are called “hot spots”. A molecule in the vicinity of a hot spot can be excited and exhibits a great enhancement in Raman intensity. The nanostructure not only enhances the incident laser field, but also the Raman scattered field. Both the incident light and the scattered signal of molecules are in resonance with the plasmon frequency, meaning the SERS signal is maximized, leading to the |E|⁴ enhancement.¹⁴²

To describe the mechanism in the short-range area, the surface electron movements are taken into account in the CE. There is evidence that the CE is a second enhancement mechanism independent of the EM mechanism. The molecules adsorbed at certain surface sites (such as atomic clusters, terraces and steps) of the nanostructure are believed to couple electronically with the surface, leading to an enhancement effect similar to resonance Raman scattering.¹⁴³ One of the explanations for this is that the absorbate's molecular orbitals are broadened into resonances by interaction with the conduction electrons. When the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) of the absorbed molecules are symmetrically disposed in terms of energy with respect to the Fermi level of the metal surface, then excitation by half the energy can make the transition.^142,144 The other explanation claims that the molecule surface interaction induces novel charge transfer intermediates that have higher Raman scattering cross-sections.^59,145

Recently, studies into SERS substrate have focused not only on a single metal particle, but also on the close-packed, well-engineered systems of metal nanostructures. Raman enhancements of 10⁵–10¹⁰ could be obtained for molecules localized in a small volume within the interparticle interstice at an appropriate wavelength and polarization and for a small enough interparticle gap.¹⁴⁶ For a simple dimer model, two nanoparticles can get close enough together to generate an optical field, for which the strength in the interstitial “hot spot” region can reach ∼10¹¹ or more.¹⁴⁷ Here, light of the appropriate wavelength is used and the exciting electric field vector is polarized along the interparticle axis, which we discussed above in Section 3.2.

The observation was made that strongly interacting metal nanoparticles result in much more effective SERS-active systems than isolated single nanoparticles and polarization sensitivity can be shown by nanoparticle aggregates.¹²⁷ In addition, the enhancement relies on the size, shape, component and the surrounding medium of the nanostructures.¹⁴ Nowadays, for attempts to achieve a greater level of structural control than before, most of the assemblies consist of systems of interacting nanoparticles or hot spots based nanostructures.

3.4.2 Surface-enhanced fluorescence. Fluorescence spectroscopy is widely used for chemical and biological sensing, microscopy imaging and medical diagnostics, due to the high cross-section of the related emission process. It is a great challenge to improve the sensitivity of fluorescence. Surface-enhanced fluorescence is based on plasmonic nanostructures that provide intense local electromagnetic fields. The excitation and emission of fluorophores are strongly influenced by near-field coupling with the nanostructures. If a fluorophore is placed in close proximity to this plasmonic nanostructure, its spectral properties can be altered by both the excitation and emission modifications.^148,149 Lakowicz and co-workers studied these effects and proposed a semi-empirical model.¹⁴⁸ In the proximity of metals, the fluorophore radiative properties are modified, and the quantum yield increases with a decrease in the spontaneous emission lifetime.¹⁵⁰ These can be expressed by the following equations:where Q_m is the quantum yield modified by the nanostructure, Γ is the radiative rate of the isolated fluorophore, k_nr is the non-radiative rate, Γ_m is the additional radiative rate induced by the metal nanostructure and τ_m is the fluorophore lifetime modified by the nanostructure. A lower lifetime translates into a faster cycling of the fluorophore, which leads to an increase in quantum yield. As calculated from electrodynamics, this change is due to the increased density of photonic states in the proximity of metal nanoparticles.^151,152 Additionally, an increase in the density of photonics states can be found in other structures, such as periodic dielectric microstructures, known as photonic crystals, and also high-Q resonant cavities.^153–155

Different parameters that can largely affect the interactions between plasmonic nanostructures and fluorophore molecules have been examined, such as the distance between the fluorophore molecule and the metal nanostructure.^156–159 When the fluorphore is close to the nanostructure by a small distance, non-radiative decay dominates and the emission is quenched.¹⁶⁰ With the distance increasing, i.e. around an intermediate distance (5 nm), in the competition between non-radiative decay and the enhancement effect, the latter begins to dominate.¹⁵⁶ A maximum fluorescence intensity was observed when the molecule was ∼5 nm away from the surface of the Au nanosphere, which resulted from a synergic action of the excitation rate enhancement and the modification of the overall quantum yield.¹⁶¹ In general, the occurrence of molecular absorption-plasmon resonance coupling requires a strong overlap between the absorption and plasmon bands and a high number density of dye molecules that are positioned within ∼5 nm of the metal surface.¹⁶² For greater distances, i.e. 10 nm and above, the amplification tapers off to eventually reach unity. When distance is up to ∼30 nm, there is no overlap between the absorption and plasmon bands. It was found that the detailed model of the distance dependence of the fluorescence enhancement was beyond Lakowicz's model, and so needed to be carried out by numerical simulations using Maxwell's equations.^163,164

Metal nanostructures modify the incident electromagnetic field and increase the field amplitude, especially close to sharp tips and edges as well as in tight junctions between metal nanostructures.¹⁶⁵ The fluorophore emission intensity depends on the square of the local exciting field as well as on the fluorophore cycling rate as described by its lifetime. The plasmonic enhancement of fluorescence can be described as the coupling of fluorophores (sources of radiation) to metal nanostructures, which then act as antennas and relay the radiation outward.^150,166 Based on the plasmonic nanostructure, the radiative scattering efficiency of a fluorophore can be enhanced, which is an effect related to the magnitude of its radiating dipole, and this in turn reflects the antenna size. Individual spherical metal nanoparticles that are tens of nanometres in size produce some increase in the radiating dipole that can be well described by Mie theory calculations of the radiative scattering cross-section.^58,156

Through continuous investigation of fluorescence enhancement, great efforts have been devoted to constructing more complex metal nanostructures and nanoparticle assemblies.

3.4.3 Surface-enhanced infrared absorption. Both infrared and Raman are vibration spectroscopies, often regarded as two faces of the same coin. Since the discovery of SERS, SERS and Surface-enhanced Infrared Absorption (SEIRA) are classed together in a category called surface-enhanced vibrational spectroscopy (SEVS).¹⁶⁷ SEIRA, as a complementary technique to SERS, provides additional fingerprinting information about the analyte in infrared absorption. The noble metals Ag and Au, which show high enhancement factors in SERS, have also been the most widely employed metals in SEIRA measurements. However, SEIRA models predict an enhancement with transition metals as strongly as with noble metals, such as Cu, Sn, Pb and Fe.¹²⁵ By analogy with the mechanisms of enhancement accepted for SERS, it has been suggested that electromagnetic and chemical contributions are responsible for the observed infrared enhancement. Based on the electrodynamics, the EM mechanism can be explained by coupling incident light into plasmon resonances. Experimentally, SEIRA is explained as being the result of the enhanced optical fields at the surface of particles when illuminated at surface phonon resonance frequencies.¹⁶⁸ The optical properties and electromagnetic field enhancement are strongly dependent on the nanoparticles, which have been extensively discussed for radiation in the visible, near-infrared and mid-infrared regions.^169,170 These theoretical models for the EM enhancement of SERS are often suitable for SEIRA, such as the model for isolated particles or aggregates of nanoparticles.^125,168 These models can help us to better understand the phenomena occurring around the particles, and enable us to design an enhanced substrate. Similarly to the factors found for the enhanced absorption based on the EM mechanism, the distance, size, shape and dielectric function all play key roles.^58,124,169

In addition, the electromagnetic enhancement of SEIRA may be calculated using effective medium theories (EMT), which are applied for the treatment of a collection of particles in a thin film. The effective medium theory, regarded as a semi-empirical approach, was developed to predict the response properties of composites, and provides a formal approach to simulate the effective dielectric function of an inhomogeneous medium, for instance, of metal–organic thin films.¹⁷¹ Finding the dielectric function for the surface-enhanced sample involves finding the effective optical properties of the mixture of these components. The Maxwell–Garnet and the Bruggeman methods are widely used for effective medium calculations. The general form of the Maxwell–Garnet model can be given as:¹⁷²

where α is the polarizability of the inclusions and i is the index over different particles. This allows for the distribution of particle shapes and sizes to be included in the calculation, as opposed to simple spheres. Further, [small epsilon, Greek, macron] is the effective dielectric function, and ε_h is the dielectric function of the host material, in which the particles are embedded, while f represents the volume fraction of the inclusions. This fundamental principle of the EMT can be utilized for calculation of the distribution of shapes and sizes of the nanoparticles.

Following, Bruggeman made a great improvement to the Maxwell–Garnet theory.¹⁷³ The Bruggeman model is a self-consistent theory that includes a greater amount of interaction between inclusions.¹²⁵ Commonly, the Bruggeman method is used to estimate the average “effective” optical and/or electrical parameters.¹⁷⁴ Generally, this formalism can be expressed as:¹²⁵

where α′ is the polarizability of the inclusions with the “Bruggeman” condition applied.

Additionally, the chemical contribution is another consideration for SEIRA, similar to SERS, based on the charge oscillations between molecular orbitals and the increase in the absorption coefficient of absorbates.¹⁷⁰

4. Surface-enhanced spectroscopy on nanoparticle assemblies

The self-assembly of nanoparticles as a simple, controlled and effective bottom-up method has become a very active field of research for application in the fabrication of probes, substrates and nanoparticle aggregates for SES. These theories for SERS can be included in the studies of other relevant spectroscopies in SES, like SEF, SEIRA and SEECL, through a comparison of the theories and experiments. Notably, the variety of responsive variables from SES depend on the plasmonic metallic nanostructures, as mentioned in Section 3. The strategies and applications of SES on nanoparticle assemblies are reviewed in the following.

4.1 Surface-enhanced Raman spectroscopy on nanoparticle assemblies

SERS is defined as a phenomenon associated with the remarkable enhancement of the electromagnetic field near very small metal (or other) objects, which are optically excited by dipolar resonance, such as SPP or LSPP. As an effective and practical tool, SERS can produce an enormous enhancement of the Raman signal and is sensitive enough for trace detection in various fields, even low to molecular detection.

The plasmonic nanostructure plays an important role in SERS applications based on a combination of nano-science and Raman spectroscopic technology. The dramatic development of nano-science and nano-technologies offers many new possibilities, notably in terms of plasmonic nanostructure design and fabrication. Among the thousands of literature reports on SERS published every year, the key issues are focused on the strategies and the applications.

The self-assembly of nanoparticles was used as a means of controlling the nanoscale building blocks for building different nanostructures. Developments in the fabrication and synthesis of noble nanostructures of controlled size, composition and shape have led to a class of well-defined nanostructures whose plasmon resonant frequencies can be manipulated through control of the geometry of the individual nanoparticles.¹⁷⁵ As a simple, effective and low-cost method, self-assembly is established for large-scale fabrication of the nanostructure and for controlling the interparticle distance. The gap between particles is vital for plasmon coupling, giving rise to the coupling strength and resonance frequency. Also, the effect of the shape of the nanoparticle influences the coupling behaviour, which has been proved. Thus, all these parameters are taken into account for developing more efficient plasmonic regions. The plasmonic regions comprise namely the hot spots as well as the nano-tips and nano-edges. The hot spots of the coupled nanostructures lead to promising effects for surface-enhanced spectroscopy due to their strong electromagnetic field.

Nanoparticle assemblies for probes. Assembling nanoparticles into new secondary structures can lead to unique physical properties that differ from those of the individual nanoparticles. These nanostructures involve dimers, trimers and higher-order aggregates, even extending to nanochains and larger aggregates, which have been assembled by biomolecule-linked or chemical-linked approaches and that serve as highly sensitive spectroscopic nanoprobes for the detection of molecules for sensing and imaging in solution or in vivo.

Nanoparticle assemblies have been explored as “tags”, which provide a unique optical “signature” for identifying them, due to their combined molecular fingerprint specificity and single-molecule sensitivity. Owing to the plasmonic nanocluster giving remarkable signal enhancements on SERS, considerable effort has been devoted to the development of probes and sensors based on nanoparticle assemblies. As it is currently understood, single-molecule surface-enhanced Raman scattering (SMSERS) is dependent on the probe containing regions of high electromagnetic field (hot spots), which can be observed in the junctions between metallic nanoparticles. Nanoparticle assemblies used for probes have attracted tremendous interests and the understanding of them has grown more and more deeply since they first appeared. Nanoprobes for SERS are those other than the substrates, which are dimerized or aggregated in the wet surroundings and which are more targeted. Most nanoprobes have been assembled by quasi-spherical or anisotropic metal nanoparticles. Generally, there are three research directions: (1) the composition, shape and structure of the constitutive nanoparticles; (2) the means of assembly; (3) controlling the interparticle distance.

Spherical nanoparticles are the most widely used for the self-assembly of nanostructures. Dimers of spherical nanoparticles are the simplest possible clusters serving as nanoprobes. In the early stage of their development, silver spheroidal nanocrystal dimers as the active structure were used in single-molecule detection.¹⁷⁶ Nevertheless, the salt-induced self-assembly occurred on the supports, not in the solution; and so the method is more used for building the aggregates on the solid support rather than as probes. Besides, for single-molecule detection, the sensitivity of this method is usually limited by the distance between the reporter and substrate or the absence of interparticle plasmonic coupling effects. To facilitate an improvement in the sensitivity, Fabris et al. developed apta-tags as an optical reporter, formed by linking silver nanoparticles with an organic dithiol molecule, followed by surface modification with thiolated single-stranded DNA (ssDNA).⁵⁹ To improve the uniformity of the constituent nanoparticles, Xia's group developed a facile method for preparing dimers of well-defined silver spheres by etching the silver nanocubes in an ethanol solvent.¹⁷⁷ However, some limitations hamper the applications of silver nanostructures, such as their higher cytotoxicity.¹⁷⁸ Even though, as nanoprobes, silver nanoparticles are more efficient than other noble metals (Au or Cu), these others remain the most commonly used materials for SERS cellular sensing. Gold nanospheres were chosen as the block for constructing a dimer in one study due to them giving a unique optical response and biocompatibility.¹⁷⁹ Besides, the coupling plasmon is highly sensitive to the size and morphology of the nanoparticles.¹⁸⁰ To better organize the process of assembling, these nanoparticles are mostly directly assembled by a biomolecule (e.g. DNA) and conjugated by chemical linkers (e.g. rigid, multivalent thiol-linkers, phenyl acetylene, thiol-terminated hydrophobic ligands and diamines).¹⁸¹ After dimers, a series of trimers and tetramers (e.g. pyramidal or chiral) of gold nanoparticles were designed as sensitive probes for SERS.^{179,182–185} Moreover, for avoiding dissociation and aggregation problems, dimers or trimers of nanoparticles can be encapsulated by polymers (e.g. PSPAA), SiO₂ or another metallic shell.^184,186 Note that the shells not only provide secure anchoring sites for biofunctionalization but also improve the long-term stability of the nanoparticles’ organization. Another key for nanoclusters is the interparticle spacing. Nam and co-workers developed an approach for SERS-active nanodumbbells, where the nanoscale gap between two nanoparticles could be precisely controlled.¹⁸⁷ The EM coupling in a dimer or larger assemblies results in a huge EM field enhancement at the nanogap. For example, the interparticle distance within nanoparticles is precisely tailored in the order of a few nanometres with changing the molecule length of the DNA bridge.⁵⁴ Lan et al. reported that the E-field EF increased from 7.8 to 23.4 as the interparticle distance decreased from 15 nm to 5 nm; on the other hand, the size increases with the E-field EF increasing.⁵⁴ Further, significantly increases in the EM field were observed as the interparticle distance decreased from 1 to 0.5 to 0.25 nm.¹⁸⁸ Therefore, the interparticle spacing within nanoparticle assemblies is a crucial factor for plasmonic coupling.

According to the LSPR theories, the optical properties of nanoparticle assemblies are strongly dependent on their geometry.^189,190 Larger number of spheres aggregating represents the next level of complexity beyond dimers. Among these, tetrahedral (pyramidal and chiral) nanoparticle assemblies have been organized by DNA with a well-controlled geometry^185,191,192 (Fig. 12). Using the pyramidal assemblies, three disease biomarkers could simultaneously be detected at the atto-molar level.

Fig. 12 Scheme of Ag pyramids self-assembled by DNA frame for SERS analysis of biomarkers. (a) Ag nanoparticles were modified with 4-ATP, PSA aptamers were inserted in each side of DNA frame of pyramid structure. (b) Three Ag NPs in each pyramid were modified with 4-ATP, NTP, and MATT, respectively, and three aptamers were inserted in three sides of the DNA frame of pyramid structure, respectively. (Reprinted from ref. 192; Copyright 2015, John Wiley and Sons.)

Besides, plasmonic core–satellite structures, which result in coupled plasmonic assemblies, are of high interest in surface-enhanced vibrational spectroscopy. For example, a core–satellite nanocluster comprised of shape-controlled plasmonic nanostructures can be achieved through self-assembly using simple molecular cross-linkers.⁵³ The number of satellites around the cores can be controlled by varying the volume of the satellite solution added to the core solution. Li et al. developed an approach where three kinds of hairpin DNA species (denoted as H1, H2 and H3, respectively) were used to immobilize gold nanoparticles¹⁹³ (Fig. 13). Therefore, the in-built electromagnetic hot spots and Raman reporters of core–satellite structures make them excellent candidates for surface-enhanced Raman scattering probes.

Fig. 13 (a) Schematic for the self-assembly of branched DNA–gold nanoaggregates. (b) SEM images of the prepared AuNPs. (c) SEM images of the fabricated DNA–gold nanoaggregates after mixing the linker DNA (S1) and three kinds of hairpin DNA-barcode–Au nanoprobes for 1 h at 20 °C. (Reprinted from ref. 193; Copyright 2014, Royal Society of Chemistry.)

Gold nanorods are one of the most typical particles with an anisotropic shape, which can undergo different forms of assembly.¹⁷⁵ Because of the anisotropic shape, nanorods can be assemblies into end-to-end or side-by-side structures and can be derived from L-shape and T-shape modes via different modifications of the nanorod surface^194–196 (Fig. 14). Plasmonic dimeric nanostructures can be precisely tuned with the orientation of nanorods, and exhibit a well-defined mutual alignment and EM field changes located in the junction section.¹⁹⁵ Similarly, the assembly of nanorods can be performed using chemical methods and biomolecule linkers, and they are usually encapsulated with a “shell” to maintain the stability.^187,197,198 A larger enhancement in the signal intensity was observed from the end-to-end form of dimers rather than from the isolated Au nanorods, which is due to the enhancement of the electric field at the junctions¹⁹⁹ (Fig. 14). Besides, to further control the process of dimerization, the sides or ends are usually encapsulated by polymers with exposed parts that can be modified or linked.^198,200 However, in contrast to the end-to-end form, side-by-side assembled nanorods were demonstrated to not be suitable for use as highly sensitive SERS probes (Fig. 15).²⁰⁰ Extending to larger numbers, Lee et al. reported that the chains of nanorods could be assembled in an end-to-end manner, which provided a direct correlation between ensemble-averaged surface-enhanced Raman scattering and the extinction properties of the chains, as a function of the average aggregation number of nanorod chains during their assembly.²⁰¹ Therefore, gold nanorods are well suited for the fundamental studies of the optical properties of aggregated nanoparticle systems and for use as nanoprobes in solution.

Fig. 14 (a) SEM images of two rods aligned end-to-end, side-to-side, in a T configuration, and in an L configuration. All scale bar, 100 nm. (Reprinted from ref. 195; Copyright 2009, American Chemical Society.) (b) TEM images of gold nanorods (b–d) and Raman spectrum of bipy-DT (e–g) at various stages of plasmon coupling; incubation (b and f), dimerization (c and f) and oligomerization (d and g) steps. (Reprinted from ref. 199; Copyright 2011, American Chemical Society.)

Fig. 15 The self-assembly of gold nanorods used for SERS tags in a side-by-side way. (a) Schematic illustration of the side-by-side NR assembly. A SH-PS is attached to the ends of CTAB-coated gold NRs in THF via site-specific ligand exchange. After the addition of the Raman reporter, the side-by-side assembly was triggered by the addition of water (10 vol%). (b) Photograph showing a typical change in colour of the solution of self-assembling NRs as a function of time from reddish-purple to blue (top left). Representative scanning transmission electron microscopy images of NRs in various stages of self-assembly. The scale bar is 15 nm. (c) Variation in extinction properties of NR ensembles over time. (Reprinted from ref. 200; Copyright 2012, American Chemical Society.)

These nanoparticles, like nanorods, nanoplates, nanocubes, nano-polyhedrons, nanoflowers and nanostars, are also used for dimeric assemblies with the junction region of nano-corners, nano-tips or nano-edges, which can provide stronger Raman scattering intensities than their rounded counterparts as they allow for a higher concentration of near field.^14,134,202 Ma et al. developed a method to synthesize a dimer of nanostars as a SERS-probe based on the Hg²⁺-mediated T–T base pair of ssDNA.²⁰³ Song et al. reported a benzoic group functionalized gold nanoflower as a bridge probe for both the recognition of target sialic acids and for the assembly of poly(N-acetylneuraminic acid) modified gold nanoparticles²⁰⁴ (Fig. 16). The special core–satellite structure led to plasmonic coupling of two kinds of gold nanoprobes, which produced a sensitive SERS signal for the imaging of sialic acids in living cells. Multiple nanoprobes can be assembled together and offer the potential of simultaneous SERS detection.

Fig. 16 Core–satellites self-assembly of gold nanoflowers and gold nanospheres. Schematic illustration of: (a) preparation of the AuNF probe, (b) preparation of the DAuNP probe, and (c) SERS imaging of the cell surface SAs based on plasmonic coupling of the core AuNF with the satellite AuNPs. 3-MPBA represents 3-mercaptophenylboronic acid. (Reprinted from ref. 204; Copyright 2016, Royal Society of Chemistry.)

Nanoparticle assemblies for substrates. Various approaches have been made to control the aggregation of plasmonic nanoparticles for sensitive and reproducible optical signals. These assemblies were composed of nanoparticles, such as nanospheres, nanorods, nanotriangles, nanocubes, nanostars and more complicated nanostructures, which depend on their shapes, sizes and interactions, and thus they are becoming attractive and promising building blocks for substrates.⁷⁶ To achieve the goals of sensitivity and reproducibility, numerous nanoparticle assemblies, including different geometrical morphologies, have been developed as signal-enhanced substrates. According to the morphology, the substrate assemblies can be divided into three parts, namely those based on: (1) different kinds of cluster arrays; (2) a monolayer of close-packed nanoparticle arrays; (3) the superlattice of 3D structures.

Cluster arrays. Noble metal nanoparticle cluster arrays are a class of engineered substrates for surface-enhanced spectroscopy, in which the nanoparticles interact on multiple length scales to create a multiscale E-field cascade enhancement. In contrast to using the nanoclusters as the probe or tag, assembling in the solution, nanoparticles assembling as cluster arrays on a surface are focused on the reproducibility, uniformity and universality. These cluster arrays fabricated by self-assembly have many merits, such as short time-consuming process, low cost, high efficiency and long-term stability. While these cluster arrays have been successfully applied in many important sensing applications, random substrates lack both basic engineering design rules and reproducibility and therefore do not fulfil the requirements for all SERS sensing applications.^205–207

Assembling colloidal nanoparticles into the dimer structure can be achieved by fine tuning the colloidal stability. Salt, like sodium chloride-induced nanoparticle aggregates were not controlled by centrifugation or by changing the dispersion medium or the ionic strength.^208–212 Xia et al. reported a simple and one-step method that generated dimers without any additional assembly steps, but which depended on the control of the colloidal stability and oxidative etching by optimizing the amount of chloride added to a polyol synthesis.²¹¹ The formation of dimers was achieved due to the change in the colloidal stability. However, the salt-induced method is poor at controlling the number of constitutive nanoparticles and the ordering arrangement.

Besides, nanocluster arrays are designed by manipulating some parameters, such as the average interparticle separation of the nanoparticles within the clusters and the separation between the clusters, resulting in field enhancements. To better achieve an ordered arrangement, self-assembly and other techniques are extensively combined to build nanocluster arrays. Template-guided self-assembly methods are used for designs that organize particles on a pattern of binding sites.⁶⁵ Using a large-area/low-cost nanopatterning method in conjunction with a meniscus force deposition technique, Alexander and co-workers were able to create large arrays of uniformly spaced nanoclusters comprising two 60 nm gold nanospheres.²¹³ The ordered arrays of dimers could be tuned by the size, orientation and placement with respect to the fiducial patterns.

The spin-coating assisted self-assembly of cluster arrays has attracted plenty of attention, because it is an efficient and low-cost method without any expensive equipment. Based on the methods, the coated platform has a wide variety of options, not confined to just a flat support.²¹⁴ There are two ways of spin-coating: first, the core nanoparticle is spin-coated on the chip as the “anchor”, followed by the self-assembly of other nanoparticles to form the structure of the “core–satellite”;⁴³ second, the pre-assembled nanoclusters can be directly spin-coated onto the support.²¹⁵ Besides, Zheng et al. reported a simple approach without using nano-lithographic methods. The gold core–satellite nanostructures could be obtained using electrostatic and DNA-directed self-assembly¹²¹ (Fig. 17).

Fig. 17 (a) Scheme of the hierarchical self-assembly procedure for the fabrication of core–satellite plasmonic nanostructures. (1) APTES functionalization of a glass substrate. (2) Electrostatic immobilization of 30 nm AuNP–DNA conjugates onto the APTES-modified substrate surface. (3) Neutralization of surface-confined amino groups by the formation of covalent amide bonds with sulfo-SMCC; (4) hybridization of 20 nm AuNP–DNA conjugates onto the 30 nm AuNPs, forming the core–satellite nanoparticle assemblies. (5) Removal of the surface-confined molecules through UV-ozone cleaning, yielding pristine gold surfaces. (b) SEM images of an array of core–satellite nanostructures on a silicon substrate obtained by immobilization of 20 nm satellite AuNPs onto the core-AuNPs. (c) SERS spectra of benzenethiol adsorbed on the core–satellite AuNP assemblies (blue line), a commercial Klarite SERS substrate (green line) and 30 nm core-AuNPs only (black line), excited at 782 nm. (Reprinted from ref. 121; Copyright 2013, John Wiley and Sons.)

Monolayer arrays. The self-assembly strategy is simple and effective and can be adopted to generate common plasmonic arrays important for an efficient optical response. Furthermore, the close-packed structures resulting from self-assembly could provide a maximum surface density of hot spots within an illuminated area.

Generally, monolayer nanoparticle arrays were originally prepared by depositing nanocrystal suspensions on substrates (such as carbon-coated copper grids or silicon wafers), followed by evaporating the solvents.⁹¹ However, it has been found that both the evaporation kinetics and the amount of excess capping ligands affect the formation of the monolayer at the liquid–air interface.⁶³ As observed, multiple factors impact the droplet-evaporation method, which results in different forms, like the “coffee-ring”, 2D monolayer film or 3D superlattice.²¹⁶ Based on the coffee-ring effect, Wang et al. developed an efficient SERS platform, which integrated the self-assembly of silver nanospheres and the preconcentration of analytes into one step.⁴⁴ Furthermore, spherical multitip gold mesoparticles are capable of self-assembling into monolayer or multiple layer arrays on Si substrates following a simple droplet-evaporation deposition.²¹⁷ As the key factors of close-packed nanoparticle monolayers, the shape, size and periodicity (for arrays) of the metal features explicitly determine the profile and intensity of the observed optical response.⁷⁵ Lin and co-workers reported a simple and efficient method for fabricating large-area (>1 cm²) 3D gold or silver films²¹⁸ (Fig. 18). In this method, the close-packed monolayers used a layer-by-layer assembly from suspensions of thiolate-passivated gold or silver colloids. Chen and co-workers presented a large-area (over 1 mm²) hexagonal close-packed structure of alkanethiol-stabilized gold nanoparticles.²¹⁹ The interparticle spacings were finely tuned from 2.3 to 3.4 nm by controlling the alkyl chain lengths of the alkanethiols from 12 to 18 carbon atoms. On the other hand, stabilizing surfactants play an important role in the formation of ordered nanostructures.

Fig. 18 (a) Schematic illustration of the plasma-assisted LbL assembly process. (b) FE-SEM micrographs of AuNP supercrystal films. (b) Large-area and (c) high-resolution cross-sectional SEM micrographs of a 10 mL AuNP film deposited on a Si substrate (imaged on the cleaved face). (d and e) Cross-sectional SEM micrographs of 20 and 30 mL AuNP films deposited on Si substrates. (Reprinted from ref. 218; Copyright 2010, American Chemical Society.)

While other nanoparticle monolayer films over large areas can be prepared through various approaches, a major challenge is to reproducibly prepare stable enhanced substrates with uniform hot spots and possibly to push the gaps to the sub-nanometre regime in order to enhance the sensitivity. The droplet-evaporation technique has also been employed to produce ordered assemblies of Au nanostructures with other shapes, including polyhedra, nanocubes and bipyramids.⁹² A ring with ordered assemblies of Au nanostructures is left on the substrate in these cases. However, inside the ring are disordered nanostructure assemblies, which are a few layers in thickness. Droplet-evaporation allows the fabrication of crystallites limited to the micrometre range, while achieving only local order due to the coffee-ring effects. Liz-Marzán and co-workers reported the synthesis of high-quality nanotriangles, followed by assembling at the air–liquid interface to obtain nanotriangle monolayers²²⁰ (Fig. 19). The possibility of fabricating homogeneous substrates over large areas is most relevant, especially when compared with methods such as droplet-evaporation. Using self-assembly at the oil–water interface, Wang's group reported a series of superstructure assemblies, including monolayer and multilayers, by tuning the ratio of hydrophilic/hydrophobic molecules added and the chain lengths.^55,221

Fig. 19 (a) TEM images of gold nanotriangles (Au NTs). (b) Optical photograph of the Au NT@PVP monolayer on a glass slide. (c) SERS spectra of thiolsbenzenethiol excited at 785 nm with concentrations varying between 10⁻⁵ and 10⁻⁸ M. (Reprinted from ref. 220; Copyright 2014, American Chemical Society.)

The anisotropy of gold nanorods renders the vertical arrays with a strong nanoantenna effect. Peng et al. reported an evaporation-induced vertically aligned monolayer with a nominal 0.8 nm gap distance, which exhibited high sensitivity.⁷² The nanorods could be assembled together in a side-by-side fashion to form layers and the layers were then packed nearly parallel to each other into a 3D structure.⁹²

Silane-ligand-modified surfaces can be used for fabricating monolayer substrates by an electrostatic self-assembly. Su et al. performed a gold nanostar layer onto an ITO glass slip, which was prepared by a low-cost electrostatically-assisted APTES-functionalized surface-assembly method for SERS analysis.²²² It was found that the substrate assembled by the nanostars with the longest branches generated the biggest enhancement when other experimental conditions (such as the laser frequency and power) were controlled to be the same.

3D superlattices. In a way, 3D superlattices can be regarded as a close-packed monolayer extended to the 3D structure. Spherical and non-spherical nanoparticles, which have uniform shape and size, can be organized to build multilayer assemblies. In contrast to monolayer assemblies, superlattices contain a greater junction region in the spatial distribution. When the superstructures are irradiated at the wavelength that couples with the plasmon resonance of the inner particles, the junction regions among the adjacent nanoparticles function as hot spots and the local electromagnetic fields in the superstructure are amplified.^37,223 As a result, the Raman scattering of the detected species located at these junctions will be remarkably enhanced.

Typically, the droplet-evaporation method is widely used to fabricate multilayer assemblies. The size and shape are important for the formation of ordered superlattices.²²⁴ In particular, the shape of the nanoparticles determines the structure and morphology of the superlattices during the assembling of the monodispersed nanoparticles.³⁷

To obtain a huge electric field concentration of crystalline structure with intense and controlled antenna effects, Liz-Marzán and co-workers developed superlattices of standing core–shell gold–silver nanorods stabilized by gemini surfactants. Notably, gemini surfactants comprising two hydrophobic tails and two hydrophilic headgroups linked by a molecular spacer have been proven to be excellent stabilizers and binders in nanoparticle synthesis and self-assembly²²⁵ (Fig. 20).

Fig. 20 (a) SEM images of the coffee-ring deposits. (b) Enlarged view of SEM images showing standing 3D superstructures (ring width ∼30 μm). (Reprinted from ref. 225; Copyright 2013, John Wiley and Sons.)

Even though evaporation-induced nanoparticle assembly on solid surface is widely used, the coffee-ring effect cannot be neglected,^44,216,224 and results in ordered nanostructures usually observed at a relatively small area close to the ring edge. Moreover, utilizing confined geometries can achieve evaporative self-assembly in special geometries.²²⁶ Thus, long-range-ordered self-assembly is a still challenge, due to the strong interaction force and low diffusion rate. To overcome this limitation, some strategies for suppressing the coffee-ring effect have been developed via the use of additives and ellipsoidal particles.^227,228 Solvent evaporation-induced interfacial self-assembly was used for avoiding the coffee-ring effects at the solid substrate, which could help minimize the interfacial energy and improve the convection velocity.^229–232 Typically, Wang's group developed a dynamically regulated strategy to achieve highly uniform, long-range-ordered Ag nanowire (Ag NW) arrays in a short time. Additionally, the oriented Ag NW arrays were applied for the SERS monitoring of a cell's adhesion and growth²³³ (Fig. 21).

Fig. 21 (a) Schematic diagram of the selective self-assembly of AgNWs on the liquid surface of chloroform: (i) the flow of mixed solutions, (iia) the nanowires’ selective arrangement, and (iib) large numbers of nanoparticles assembled on the wall. Numbered images: (1) proposed model for the selective self-assembly of nanowires; (2) proposed model for the assembly of nanoparticles on the wall with the decrease of the liquid surface. (b) Typical SEM image of the selective self-assembly of AgNWs at the water–air interface corresponding to model 1 in (a). (c) Typical SEM image of the assembly of AgNPs on the vessel wall corresponding to model 2 in (a). (d) Schematic illustration of cell adhesion and growth on (ia) oriented AgNW arrays and (iia) random AgNW films. (e) Typical Raman spectra of the cells cultured for 1, 2 and 6 h on the oriented AgNW arrays. (f and g) Typical SEM images of the cells cultured for 12 h: (f) on the oriented AgNW arrays and (g) on the random AgNW films. (Scale bars: (b and c) 250 nm and (f and g) 50 μm.) (Reprinted from ref. 233; Copyright 2016, John Wiley and Sons.)

4.2 Surface-enhanced fluorescence on nanoparticle assemblies

SEF is another hot research field in SES. SEF is also referred to as metal-enhanced fluorescence (MEF) and plasmon-enhanced fluorescence (PEF).²³⁴ By analogy with SERS, many plasmonic nanostructures are designed and fabricated for enhancing fluorescence. However, the direct fabrication of dimers or other aggregates in solution are occasionally used for SEF, due to the influence of the separate distance between the plasmon and the fluorophore.¹⁶³ Therefore, utilizing core–shell nanostructures to separate the plasmon and the fluorophore is one of the main strategies used in SEF. Additionally, regular arrays of metallic nanostructures were observed with strong electric fields (hot spots) at the gap between two neighbouring nanostructures with intensified LSPR. Among these strategies, the combination of a core–shell nanostructure and nano-array has attracted considerable attention.

In the past years, various types of metal nanostructures, consisting of special spacer layers, with different shapes and metallic species, have been used for fabricating effective SEF-active nano-arrays.²³⁵ The metallic species used for the fabrication of SEF-responsive plasmonic nanostructures are not only limited to Au and Ag, but also include Cu and Al.^{163,236–238} Among these, Ag and Au have attracted most attention. The effects of metal nanoparticles for enhancing the luminescence of CdSe quantum dots have been compared, by Batteas's group.²³⁹ They developed a single layer of quantum dots on top of arrays of the desired metal nanostructure positioned in a grid pattern, in which the separation between the two was controlled with a polymer spacer formed using layer-by-layer assembly. By changing the thickness of the polymer spacer, the photoluminescence enhancement of CdSe was found to peak at a factor of 2 at distances of ∼11 and 8 nm, when coupled to Au and Ag particles respectively (Fig. 22).

Fig. 22 (a) Confocal scanning microscopy image generated by collecting the 488 nm laser line on nine layers of polymers deposited upon Au-NP patterned GaAs surfaces (the z-scale bar is from 50–850 a.u.). (b) Spectra on the GaAs background (red circle) and Au-NP patterned (white circle) regions. The red line and the black line represent the averaged spectra of the reflected laser light of red circle and white circle areas in panel A, respectively. (c) The corresponding confocal image at the same scanning region generated by collecting the 488 nm laser line on five layers of polymers deposited upon Ag-NP patterned GaAs surfaces. (d) Reflected laser light from the GaAs background (green circle) and the Ag-NP patterned (white circle) regions. The green line and the black line represent the averaged spectra of green circle and white circle areas in panel C, respectively. (e) Schematic cross-sectional view of the sample showing that polymers and 4 nm QDs were deposited onto Au-NP arrays on GaAs in sequence. (Reprinted from ref. 239; Copyright 2009, American Chemical Society.)

For the assembly of nano-arrays, common nanostructures, such as nanospheres, nanorods, nanotriangles and nanocages, have been used.^238–240 Nanospheres are the most widely used for fabricating ordered arrays and close-packed sheets, due to their high controllability and symmetrical structure.²⁴¹ Beyond that, the self-assembly of anisotropic nanostructures was used to fabricate a plasmonic surface of an ordered array. Au NR is a promising anisotropic nanostructure for direct assembly in an ordered array format on a substrate. Mei and co-workers developed a SEF-based DNA detecting platform where fluorescence-labelled hairpin ssDNA was modified on an ordered array of vertical Au NRs for capturing the complementary ssDNA.²⁴⁰ By varying the number of base pairs in the DNA duplex, the fluorescence intensity could be tuned. Furthermore, the exposing tips in the vertical arrays could amplify the SPR signals at neighbouring particles collectively. As such, the nanostructures of nano-tips and nano-corners were demonstrated to be able to generate a significant local enhancement of the surrounding EM field. Xia's group developed Au nanocage arrays as an enhancing substrate through microcontact printing and investigated the potential of the Au nanocages as a novel class of MEF substrates in the NIR region.²⁴² MEF was demonstrated with NIR light-emitting films and electrospun polymer nanofibres, with additional spatial control provided by microprinting of the nanocages.

It is well known that MEF critically depends on the distance between the metal surface and the fluorophores. Thus, the design of the spacer layer is commonly used in MEF substrates. For arrays, there two main ways for fabricating the spacer layer: one is by building a pre-coating shell of the nanostructures before assembling, and the other one is coating, spinning and growing a layer finely, or modifying a molecular layer on the assembled nanostructures.^240,242 The common materials for the spacer layer involve polymer materials, semiconductor nanomaterials and DNA molecules.

4.3 Other surface-enhanced spectroscopies on nanoparticle assemblies

Surface-enhanced infrared absorption (SEIRA) is another important surface-enhanced vibrational spectroscopy, besides SERS.¹²⁸ However, SEIRA has received much less attention than SERS. Generally, SEIRA and SERS complement each other. Similar to SERS, two different mechanisms, namely an electromagnetic effect and chemical effect, are supposed to contribute to the total SEIRA enhancement (refer to Section 3.4.3). SEIRA can also provide several orders of magnitude enhancement in the IR region. Numerous metallic substrates, such as roughened or metal island films, fabricated arrays of small metallic antennas, where molecules are located on or near the structure, are created to provide a sufficiently strong signal for SEIRA detection.^243,244 The self-assembly of nanostructures has been developed for fabricating plasmonic SEIRA substrates, due to the low cost and high throughput.²⁴⁵ For example, Halas and co-workers focused on the application of nanoshells for SEIRA.^246,247 By the droplet-evaporation method, gold nanoshells were assembled into periodic arrays with nanoscale interparticle gaps, which could provide significant enhancements to both SERS and SEIRA, effectively combining both spectroscopies on a single substrate. Simultaneously, molecular detection and characterization could be achieved with high precision and sensitivity, through SERS and SEIRA. Notably, the sub-10 nm interparticle gaps between the nanostructures can induce strong plasmon coupling, which results in the specific electromagnetic properties of this array geometry (Fig. 23).

Fig. 23 (a) Fabrication of Au nanoshell arrays. (b and c) SEM images of nanoshell arrays formed by drying 40 μL of aqueous solutions of CTAB-capped Au nanoshells on silicon wafers. The core radius of the nanoshells is 150 ± 12 nm, and the shell thickness is 22 ± 1 nm. (d) TEM image of the nanoshell arrays formed on a TEM grid. (e) Empirical SERS enhancement factors of para-mercaptoaniline (pMA) adsorbed on nanoshell arrays and the isolated nanoshells. (f) Normal IR spectrum of pure pMA and the SEIRA spectrum of pMA SAMs on the nanoshell arrays. (Reprinted from ref. 246; Copyright 2007, John Wiley and Sons.)

Surface-enhanced electrochemiluminescence (SEECL) is one of the fastest developing methods in the surface-enhanced spectroscopy family. Similar to SEF, SEECL was found to be strongly dependent on the distance between ECL emitters and the plasmonic nanostructures, due to the competition between field enhancement and Förster resonance energy transfer (FRET).³¹ Generally, ECL emitters and plasmonic nanostructures are assembled on the surface of solid electrode in SEECL. The ECL intensity can be altered by controlling the thickness of the silica shell or length of the DNA^30,31,248 (Fig. 24). Wang et al. positioned Au NP@SiO₂ on a modified gold electrode though electrostatic self-assembly.²⁹ They demonstrated that the LSPR-field enhancement of ECL is the dominant factor when the shell thickness increases from 2 nm to 6 nm. Even though some studies have been reported, SEECL is still a young subject in SES.

Fig. 24 (a) Schematic of SEECL of the AuNP@SiO₂-modified electrode. (b) ECL emission of GE (blank), GE/PDDA (red), GE/PDDA/SiO₂ (orange), GE/PDDA/AuNPs (violet) and GE/PDDA/AuNP@SiO₂ (green, the thickness of the silica shell is 4 nm). (c) Relationship between ECL intensity of AuNP@SiO₂-modified electrodes and the silica shell thickness. Error bars represent the standard deviations of 5 replicates; (Reprinted from ref. 29, 2015, Springer Nature) (d) ECL Aptamer sensing platform based on energy transfer between CdS NCs and Au NPs. (Reprinted from ref. 31; Copyright 2011, American Chemical Society.)

5. Summary and perspectives

In summary, in recent years, a lot of efforts have been paid to the fabrication of different kinds of ordered nanoparticles, and the potential applications of these nanostructures for SES have been investigated in detail (Fig. 25). As stated in the text, a number of mechanisms and properties behind these phenomena have been proposed. These efforts have greatly improved the development of SES. Nevertheless, the trends in the near future in nanoparticle assemblies should still focus on the fabrication of new supercrystals with superior properties compared to individual nanoparticles. The following two ways are the most likely to realize such a purpose: first, a diverse strategy to develop multi-functionalities of supercrystals, which could generate interparticle optical, electric or magnetic elements communications; second, the fabrication of nanoparticle supercrystals that consist of binary cooperative complementary materials.

Fig. 25 Summary of the mechanisms of ordered nanoparticle assemblies and the strategies and applications of ordered assemblies for SES.

On the other hand, some issues must be solved to push these promising techniques into real applications: most of all, the mechanism of SES is still unclear and needs to be determined. Although a couple of mechanisms have been proposed to explain the phenomenon of SERS, hitherto none of these mechanisms can explain all the SERS phenomena. In addition, most of the other SES (e.g. SEIRA and SEF) cited the mechanism of SERS; however, little attention has been paid to explain the reason why the signal enhancement factor between different kinds of SES are much different even when using the same substrate; second, although a large number of approaches have been reported for the fabrication of different kinds of SES substrates, most of these substrates have been used in individual laboratories, while few of them have been commercialized. Therefore, the fabrication of economic, repeatable and universal SES substrates is still a tough task.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

T. W. acknowledges the support from the 1000 Young Talents program, the National Natural Science Foundation of China (Grant No. 21635002, 21621062) and ICCAS. L. G. and Y. L. acknowledge the support from the National Natural Science Foundation of China (Grant No. 21675028, 21575027 and 21775026).

References

1. Y. Li, J. Zhang, S. Zhu, H. Dong, F. Jia, Z. Wang, Z. Sun, L. Zhang, Y. Li and H. Li, Adv. Mater., 2009, 21, 4731–4734.
2. Z. Xie, K. Cao, Y. Zhao, L. Bai, H. Gu, H. Xu and Z. Z. Gu, Adv. Mater., 2014, 26, 2413–2418.
3. K. Liu and L. Jiang, Nano Today, 2011, 6, 155–175.
4. C. S. Zuker, Science, 1994, 265, 742–744.
5. X. Gao, X. Yan, X. Yao, L. Xu, K. Zhang, J. Zhang, B. Yang and L. Jiang, Adv. Mater., 2007, 19, 2213–2217.
6. J. Huang, X. Wang and Z. L. Wang, Nano Lett., 2006, 6, 2325–2331.
7. R. A. Potyrailo, H. Ghiradella, A. Vertiatchikh, K. Dovidenko, J. R. Cournoyer and E. Olson, Nat. Photonics, 2007, 1, 123–128.
8. W. Peng, S. Zhu, W. Wang, W. Zhang, J. Gu, X. Hu, D. Zhang and Z. Chen, Adv. Funct. Mater., 2012, 22, 2072–2080.
9. J. Zi, X. Yu, Y. Li, X. Hu, C. Xu, X. Wang, X. Liu and R. Fu, Proc. Natl. Acad. Sci. U. S. A., 2003, 100, 12576–12578.
10. D. Spitzer, T. Cottineau, N. Piazzon, S. Josset, F. Schnell, S. N. Pronkin, E. R. Savinova and V. Keller, Angew. Chem., Int. Ed., 2012, 51, 5334–5338.
11. J. Krieger, E. Große-Wilde, T. Gohl and H. Breer, Eur. J. Neurosci., 2005, 21, 2167–2176.
12. R. V. Nair and R. Vijaya, Prog. Quantum Electron., 2010, 34, 89–134.
13. F. Liu, B. Dong and X. Liu, Optical Devices in Communication and Computation, InTech, 2012.
14. K. L. Kelly, E. Coronado, L. L. Zhao and G. C. Schatz, J. Phys. Chem. B, 2003, 107, 668–677.
15. M. E. Stewart, C. R. Anderton, L. B. Thompson, J. Maria, S. K. Gray, J. A. Rogers and R. G. Nuzzo, Chem. Rev., 2008, 108, 494–521.
16. J. N. Anker, W. P. Hall, O. Lyandres, N. C. Shah, J. Zhao and R. P. Van Duyne, Nat. Mater., 2008, 7, 442.
17. M. Fleischmann and P. Hendra, Chem. Phys. Lett., 1974, 26, 123.
18. C. B. Milojevich, D. W. Silverstein, L. Jensen and J. P. Camden, J. Am. Chem. Soc., 2011, 133, 14590–14592.
19. J. Kneipp, H. Kneipp and K. Kneipp, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 17149–17153.
20. A. R. Guerrero and R. F. Aroca, Angew. Chem., Int. Ed., 2011, 50, 665–668.
21. M. D. Furtaw, D. L. Steffens, T. M. Urlacher and J. P. Anderson, Anal. Chem., 2013, 85, 7102–7108.
22. G. F. Walsh and L. Dal Negro, Nano Lett., 2013, 13, 3111–3117.
23. Y. Shen, Nature, 1989, 337, 519.
24. F. M. Geiger, Annu. Rev. Phys. Chem., 2009, 60, 61–83.
25. J. Kozuch, C. Steinem, P. Hildebrandt and D. Millo, Angew. Chem., Int. Ed., 2012, 51, 8114–8117.
26. L. V. Brown, K. Zhao, N. King, H. Sobhani, P. Nordlander and N. J. Halas, J. Am. Chem. Soc., 2013, 135, 3688–3695.
27. R. Bukasov and J. S. Shumaker-Parry, Anal. Chem., 2009, 81, 4531–4535.
28. D. Wang, L. Guo, R. Huang, B. Qiu, Z. Lin and G. Chen, Electrochim. Acta, 2014, 150, 123–128.
29. D. Wang, L. Guo, R. Huang, B. Qiu, Z. Lin and G. Chen, Sci. Rep., 2015, 5, 7954.
30. D. Wang, Y. Li, Z. Lin, B. Qiu and L. Guo, Anal. Chem., 2015, 87, 5966–5972.
31. J. Wang, Y. Shan, W.-W. Zhao, J.-J. Xu and H.-Y. Chen, Anal. Chem., 2011, 83, 4004–4011.
32. Y. Shan, J.-J. Xu and H.-Y. Chen, Chem. Commun., 2009, 905–907.
33. R. Yang, Y. Liu, H. Ye, B. Qiu, Z. Lin and L. Guo, Electroanalysis, 2016, 28, 1783–1786.
34. K. J. Bishop, C. E. Wilmer, S. Soh and B. A. Grzybowski, Small, 2009, 5, 1600–1630.
35. M. A. Boles, M. Engel and D. V. Talapin, Chem. Rev., 2016, 116, 11220–11289.
36. Y. Yin and A. P. Alivisatos, Nature, 2005, 437, 664–670.
37. Z. Zhu, H. Meng, W. Liu, X. Liu, J. Gong, X. Qiu, L. Jiang, D. Wang and Z. Tang, Angew. Chem., Int. Ed., 2011, 123, 1631–1634.
38. D. Luo, C. Yan and T. Wang, Small, 2015, 11, 5984–6008.
39. D. Ben-Avraham, J. Stat. Phys., 2006, 123, 709–710.
40. H. Hamaker, Physica, 1937, 4, 1058–1072.
41. E. Lifshitz, Perspectives in Theoretical Physics: The Collected Papers of E. M. Lifshitz, 1956, pp. 73–83.
42. D. Langbein, Phys. Rev. B: Solid State, 1970, 2, 3371.
43. C. Zhu, G. Meng, Q. Huang and Z. Huang, J. Hazard. Mater., 2012, 211, 389–395.
44. W. Wang, Y. Yin, Z. Tan and J. Liu, Nanoscale, 2014, 6, 9588–9593.
45. M. Corricelli, N. Depalo, E. Fanizza, D. Altamura, C. Giannini, D. Siliqi, R. Di Mundo, F. Palumbo, V. G. Kravets and A. N. Grigorenko, J. Phys. Chem. C, 2014, 118, 7579–7590.
46. E. V. Shevchenko, D. V. Talapin, C. B. Murray and S. O'Brien, J. Am. Chem. Soc., 2006, 128, 3620–3637.
47. H.-J. Yang, S.-Y. He, H.-L. Chen and H.-Y. Tuan, Chem. Mater., 2014, 26, 1785–1793.
48. P. M. Claesson, T. Ederth, V. Bergeron and M. Rutland, Adv. Colloid Interface Sci., 1996, 67, 119–183.
49. R. A. French, A. R. Jacobson, B. Kim, S. L. Isley, R. L. Penn and P. C. Baveye, Environ. Sci. Technol., 2009, 43, 1354–1359.
50. J. T. Han, Y. Zheng, J. H. Cho, X. Xu and K. Cho, J. Phys. Chem. B, 2005, 109, 20773–20778.
51. A. M. Kalsin, M. Fialkowski, M. Paszewski, S. K. Smoukov, K. J. Bishop and B. A. Grzybowski, Science, 2006, 312, 420–424.
52. R. Levicky, T. M. Herne, M. J. Tarlov and S. K. Satija, J. Am. Chem. Soc., 1998, 120, 9787–9792.
53. N. Gandra, A. Abbas, L. Tian and S. Singamaneni, Nano Lett., 2012, 12, 2645–2651.
54. X. Lan, Z. Chen, X. Lu, G. Dai, W. Ni and Q. Wang, ACS Appl. Mater. Interfaces, 2013, 5, 10423–10427.
55. Y. H. Lee, W. Shi, H. K. Lee, R. Jiang, I. Y. Phang, Y. Cui, L. Isa, Y. Yang, J. Wang and S. Li, Nat. Commun., 2015, 6, 6990.
56. N. R. Jana, Angew. Chem., Int. Ed., 2004, 43, 1536–1540.
57. J. Gong, R. S. Newman, M. Engel, M. Zhao, F. Bian, S. C. Glotzer and Z. Tang, Nat. Commun., 2017, 8, 14038.
58. M. B. Ross, C. A. Mirkin and G. C. Schatz, J. Phys. Chem. C, 2016, 120, 816–830.
59. L. Fabris, M. Dante, T. Q. Nguyen, J. B. H. Tok and G. C. Bazan, Adv. Funct. Mater., 2008, 18, 2518–2525.
60. H. Zhang and D. Wang, Angew. Chem., Int. Ed., 2008, 120, 4048–4051.
61. Z. Yin, W. Zhang, Q. Fu, H. Yue, W. Wei, P. Tang, W. Li, W. Li, L. Lin and G. Ma, Small, 2014, 10, 3619–3624.
62. M. Yang, G. Chen, Y. Zhao, G. Silber, Y. Wang, S. Xing, Y. Han and H. Chen, Phys. Chem. Chem. Phys., 2010, 12, 11850–11860.
63. T. P. Bigioni, X.-M. Lin, T. T. Nguyen, E. I. Corwin, T. A. Witten and H. M. Jaeger, Nat. Mater., 2006, 5, 265–270.
64. H. Duan, D. Wang, D. G. Kurth and H. Möhwald, Angew. Chem., Int. Ed., 2004, 43, 5639–5642.
65. L. Yang, B. Yan, W. R. Premasiri, L. D. Ziegler, L. D. Negro and B. M. Reinhard, Adv. Funct. Mater., 2010, 20, 2619–2628.
66. H. Wang, X. Song, C. Liu, J. He, W. H. Chong and H. Chen, ACS Nano, 2014, 8, 8063–8073.
67. K. Thorkelsson, P. Bai and T. Xu, Nano Today, 2015, 10, 48–66.
68. S.-Y. Ding, J. Yi, J.-F. Li, B. Ren, D.-Y. Wu, R. Panneerselvam and Z.-Q. Tian, Nat. Rev. Mater., 2016, 1, 16021.
69. W. Li, P. Zhang, M. Dai, J. He, T. Babu, Y.-L. Xu, R. Deng, R. Liang, M.-H. Lu and Z. Nie, Macromolecules, 2013, 46, 2241–2248.
70. J. Guo, Y. Zhang, L. Shi, Y. Zhu, M. F. Mideksa, K. Hou, W. Zhao, D. Wang, M. Zhao and X. Zhang, J. Am. Chem. Soc., 2017, 139, 17964–17972.
71. T. Wang, J. Zhuang, J. Lynch, O. Chen, Z. Wang, X. Wang, D. LaMontagne, H. Wu, Z. Wang and Y. C. Cao, Science, 2012, 338, 358–363.
72. B. Peng, G. Li, D. Li, S. Dodson, Q. Zhang, J. Zhang, Y. H. Lee, H. V. Demir, X. Yi Ling and Q. Xiong, ACS Nano, 2013, 7, 5993–6000.
73. S.-Y. Zhang, M. D. Regulacio and M.-Y. Han, Chem. Soc. Rev., 2014, 43, 2301–2323.
74. T. Thai, Y. Zheng, S. H. Ng, S. Mudie, M. Altissimo and U. Bach, Angew. Chem., Int. Ed., 2012, 51, 8732–8735.
75. A. Tao, P. Sinsermsuksakul and P. Yang, Nat. Nanotechnol., 2007, 2, 435–440.
76. G. M. Whitesides and B. Grzybowski, Science, 2002, 295, 2418–2421.
77. S. J. Barrow, A. M. Funston, X. Wei and P. Mulvaney, Nano Today, 2013, 8, 138–167.
78. A. Guerrero-Martínez, J. Pérez-Juste, E. Carbó-Argibay, G. Tardajos and L. M. Liz-Marzán, Angew. Chem., Int. Ed., 2009, 48, 9484–9488.
79. Y. Ofir, B. Samanta and V. M. Rotello, Chem. Soc. Rev., 2008, 37, 1814–1825.
80. D. Liu, W. Chen, K. Sun, K. Deng, W. Zhang, Z. Wang and X. Jiang, Angew. Chem., Int. Ed., 2011, 50, 4103–4107.
81. S. J. Tan, M. J. Campolongo, D. Luo and W. Cheng, Nat. Nanotechnol., 2011, 6, 268–276.
82. A. P. Alivisatos, K. P. Johnsson, X. Peng, T. E. Wilson, C. J. Loweth, M. P. Bruchez Jr and P. G. Schultz, Nature, 1996, 382, 609–611.
83. L. Guo, Y. Xu, A. R. Ferhan, G. Chen and D.-H. Kim, J. Am. Chem. Soc., 2013, 135, 12338–12345.
84. A. Kuzyk, R. Schreiber, Z. Fan, G. Pardatscher, E.-M. Roller, A. Högele, F. C. Simmel, A. O. Govorov and T. Liedl, Nature, 2012, 483, 311–314.
85. Y. Li, Z. Liu, G. Yu, W. Jiang and C. Mao, J. Am. Chem. Soc., 2015, 137, 4320–4323.
86. L. Lu and Y. Xia, Anal. Chem., 2015, 87, 8584–8591.
87. M. Grzelczak, J. Vermant, E. M. Furst and L. M. Liz-Marzán, ACS Nano, 2010, 4, 3591–3605.
88. D. Jishkariani, B. T. Diroll, M. Cargnello, D. R. Klein, L. A. Hough, C. B. Murray and B. Donnio, J. Am. Chem. Soc., 2015, 137, 10728–10734.
89. P. Akcora, H. Liu, S. K. Kumar, J. Moll, Y. Li, B. C. Benicewicz, L. S. Schadler, D. Acehan, A. Z. Panagiotopoulos and V. Pryamitsyn, Nat. Mater., 2009, 8, 354.
90. N. Sharma, A. Top, K. L. Kiick and D. J. Pochan, Angew. Chem., Int. Ed., 2009, 48, 7078–7082.
91. X. Lin, H. Jaeger, C. Sorensen and K. Klabunde, J. Phys. Chem. B, 2001, 105, 3353–3357.
92. T. Ming, X. Kou, H. Chen, T. Wang, H. L. Tam, K. W. Cheah, J. Y. Chen and J. Wang, Angew. Chem., Int. Ed., 2008, 120, 9831–9836.
93. R. Klein and P. Meakin, Nature, 1989, 339, 360–362.
94. H. Y. Shi, B. Hu, X. C. Yu, R. L. Zhao, X. F. Ren, S. L. Liu, J. W. Liu, M. Feng, A. W. Xu and S. H. Yu, Adv. Funct. Mater., 2010, 20, 958–964.
95. R. B. Grubbs, Nat. Mater., 2007, 6, 553.
96. L. Hu, M. Chen, X. Fang and L. Wu, Chem. Soc. Rev., 2012, 41, 1350–1362.
97. A. Böker, J. He, T. Emrick and T. P. Russell, Soft Matter, 2007, 3, 1231–1248.
98. F. Reincke, S. G. Hickey, W. K. Kegel and D. Vanmaekelbergh, Angew. Chem., Int. Ed., 2004, 43, 458–462.
99. S. Si, W. Liang, Y. Sun, J. Huang, W. Ma, Z. Liang, Q. Bao and L. Jiang, Adv. Funct. Mater., 2016, 26, 8137–8145.
100. C. Akin, J. Yi, L. C. Feldman, C. Durand, S. M. Hus, A.-P. Li, M. A. Filler and J. W. Shan, ACS Nano, 2015, 9, 5405–5412.
101. S. Gupta, Q. Zhang, T. Emrick and T. P. Russell, Nano Lett., 2006, 6, 2066–2069.
102. A. A. Shah, M. Ganesan, J. Jocz and M. J. Solomon, ACS Nano, 2014, 8, 8095–8103.
103. K. Wang, S.-M. Jin, J. Xu, R. Liang, K. Shezad, Z. Xue, X. Xie, E. Lee and J. Zhu, ACS Nano, 2016, 10, 4954–4960.
104. T. Ding, K. Song, K. Clays and C. H. Tung, Adv. Mater., 2009, 21, 1936–1940.
105. M. Wang, L. He, W. Xu, X. Wang and Y. Yin, Angew. Chem., Int. Ed., 2015, 54, 7077–7081.
106. M. R. Jones, K. D. Osberg, R. J. Macfarlane, M. R. Langille and C. A. Mirkin, Chem. Rev., 2011, 111, 3736–3827.
107. Z. Liu, H. Huang and T. He, Small, 2013, 9, 505–510.
108. D. P. Lyvers, J.-M. Moon, A. V. Kildishev, V. M. Shalaev and A. Wei, ACS Nano, 2008, 2, 2569–2576.
109. L. Jiang, X. Chen, N. Lu and L. Chi, Acc. Chem. Res., 2014, 47, 3009–3017.
110. D. Nepal, M. S. Onses, K. Park, M. Jespersen, C. J. Thode, P. F. Nealey and R. A. Vaia, ACS Nano, 2012, 6, 5693–5701.
111. V. Flauraud, M. Mastrangeli, G. D. Bernasconi, J. Butet, D. T. Alexander, E. Shahrabi, O. J. Martin and J. Brugger, Nat. Nanotechnol., 2016, 12, 73–80.
112. Y. Wang, J. Xu, Y. Wang and H. Chen, Chem. Soc. Rev., 2013, 42, 2930–2962.
113. S. J. Barrow, A. M. Funston, D. E. Gómez, T. J. Davis and P. Mulvaney, Nano Lett., 2011, 11, 4180–4187.
114. Y. Tian, T. Wang, W. Liu, H. L. Xin, H. Li, Y. Ke, W. M. Shih and O. Gang, Nat. Nanotechnol., 2015, 10, 637–644.
115. G. Bachelier, I. Russier-Antoine, E. Benichou, C. Jonin, N. Del Fatti, F. Vallée and P.-F. Brevet, Phys. Rev. Lett., 2008, 101, 197401.
116. S. Sheikholeslami, Y.-w. Jun, P. K. Jain and A. P. Alivisatos, Nano Lett., 2010, 10, 2655–2660.
117. A. Lombardi, M. P. Grzelczak, A. Crut, P. Maioli, I. Pastoriza-Santos, L. M. Liz-Marzán, N. Del Fatti and F. Vallée, ACS Nano, 2013, 7, 2522–2531.
118. L. Weller, V. V. Thacker, L. O. Herrmann, E. A. Hemmig, A. Lombardi, U. F. Keyser and J. J. Baumberg, ACS Photonics, 2016, 3, 1589–1595.
119. J. H. Yoon, J. Lim and S. Yoon, ACS Nano, 2012, 6, 7199–7208.
120. H. Xing, Z. Wang, Z. Xu, N. Y. Wong, Y. Xiang, G. L. Liu and Y. Lu, ACS Nano, 2011, 6, 802–809.
121. Y. Zheng, T. Thai, P. Reineck, L. Qiu, Y. Guo and U. Bach, Adv. Funct. Mater., 2013, 23, 1519–1526.
122. P. Dey, S. Zhu, K. J. Thurecht, P. M. Fredericks and I. Blakey, J. Mater. Chem. B, 2014, 2, 2827–2837.
123. L. Xu, H. Kuang, C. Xu, W. Ma, L. Wang and N. A. Kotov, J. Am. Chem. Soc., 2012, 134, 1699–1709.
124. M. Jahn, S. Patze, I. J. Hidi, R. Knipper, A. I. Radu, A. Mühlig, S. Yüksel, V. Peksa, K. Weber and T. Mayerhöfer, Analyst, 2016, 141, 756–793.
125. R. Aroca, Surface-enhanced vibrational spectroscopy, John Wiley & Sons, 2006.
126. A. J. Haes, C. L. Haynes, A. D. McFarland, G. C. Schatz, R. P. Van Duyne and S. Zou, MRS Bull., 2005, 30, 368–375.
127. K. Kneipp, M. Moskovits and H. Kneipp, Surface-enhanced Raman scattering: physics and applications, Springer Science & Business Media, 2006.
128. R. Aroca and S. Rodriguez-Llorente, J. Mol. Struct., 1997, 408, 17–22.
129. S. K. Jha, Z. Ahmed, M. Agio, Y. Ekinci and J. r. F. Löffler, J. Am. Chem. Soc., 2012, 134, 1966–1969.
130. U. Kreibig and M. Vollmer, Optical Properties of Metal Clusters, Springer, 1995, pp. 13–201.
131. P. K. Jain, X. Huang, I. H. El-Sayed and M. A. El-Sayed, Plasmonics, 2007, 2, 107–118.
132. K. L. Kelly, E. Coronado, L. L. Zhao and G. C. Schatz, J. Phys. Chem. B, 2003, 107, 668–677.
133. P. Nordlander, C. Oubre, E. Prodan, K. Li and M. Stockman, Nano Lett., 2004, 4, 899–903.
134. E. Hao and G. C. Schatz, J. Chem. Phys., 2004, 120, 357–366.
135. J. P. Kottmann, O. J. Martin, D. R. Smith and S. Schultz, Phys. Rev. B: Condens. Matter Mater. Phys., 2001, 64, 235402.
136. J. Gong, G. Li and Z. Tang, Nano Today, 2012, 7, 564–585.
137. M. M. Alvarez, J. T. Khoury, T. G. Schaaff, M. N. Shafigullin, I. Vezmar and R. L. Whetten, J. Phys. Chem. B, 1997, 101, 3706–3712.
138. A. O. Govorov, G. W. Bryant, W. Zhang, T. Skeini, J. Lee, N. A. Kotov, J. M. Slocik and R. R. Naik, Nano Lett., 2006, 6, 984–994.
139. J. Lee, P. Hernandez, J. Lee, A. O. Govorov and N. A. Kotov, Nat. Mater., 2007, 6, 291–295.
140. W. Zhang, A. O. Govorov and G. W. Bryant, Phys. Rev. Lett., 2006, 97, 146804.
141. G. C. Schatz, Acc. Chem. Res., 1984, 17, 370–376.
142. M. J. Banholzer, J. E. Millstone, L. Qin and C. A. Mirkin, Chem. Soc. Rev., 2008, 37, 885–897.
143. W. E. Doering and S. Nie, J. Phys. Chem. B, 2002, 106, 311–317.
144. A. Campion and P. Kambhampati, Chem. Soc. Rev., 1998, 27, 241–250.
145. K. Kneipp, Y. Wang, H. Kneipp, L. T. Perelman, I. Itzkan, R. R. Dasari and M. S. Feld, Phys. Rev. Lett., 1997, 78, 1667.
146. N. Liver, A. Nitzan and J. Gersten, Chem. Phys. Lett., 1984, 111, 449–454.
147. H. Xu and M. Käll, ChemPhysChem, 2003, 4, 1001–1005.
148. J. R. Lakowicz, Y. Shen, S. D'Auria, J. Malicka, J. Fang, Z. Gryczynski and I. Gryczynski, Anal. Biochem., 2002, 301, 261–277.
149. I. Gryczynski, J. Malicka, Y. Shen, Z. Gryczynski and J. R. Lakowicz, J. Phys. Chem. B, 2002, 106, 2191–2195.
150. W. Deng and E. M. Goldys, Langmuir, 2012, 28, 10152–10163.
151. C. D. Geddes and J. R. Lakowicz, J. Fluoresc., 2002, 12, 121–129.
152. V. Giannini, J. A. Sánchez-Gil, O. L. Muskens and J. G. Rivas, J. Opt. Soc. Am. B, 2009, 26, 1569–1577.
153. P. Goy, J. Raimond, M. Gross and S. Haroche, Phys. Rev. Lett., 1983, 50, 1903.
154. N. Ganesh, W. Zhang, P. C. Mathias, E. Chow, J. Soares, V. Malyarchuk, A. D. Smith and B. T. Cunningham, Nat. Nanotechnol., 2007, 2, 515–520.
155. Y.-J. Rao, Opt. Lasers Eng., 1999, 31, 297–324.
156. E. Fort and S. Grésillon, J. Phys. D: Appl. Phys., 2007, 41, 013001.
157. N. Akbay, J. R. Lakowicz and K. Ray, J. Phys. Chem. C, 2012, 116, 10766–10773.
158. T. Corrigan, S.-H. Guo, H. Szmacinski and R. Phaneuf, Appl. Phys. Lett., 2006, 88, 101112.
159. B. Yang, N. Lu, D. Qi, R. Ma, Q. Wu, J. Hao, X. Liu, Y. Mu, V. Reboud and N. Kehagias, Small, 2010, 6, 1038–1043.
160. J. Malicka, I. Gryczynski, J. Kusba and J. R. Lakowicz, Biopolymers, 2003, 70, 595–603.
161. P. Anger, P. Bharadwaj and L. Novotny, Phys. Rev. Lett., 2006, 96, 113002.
162. W. Ni, Z. Yang, H. Chen, L. Li and J. Wang, J. Am. Chem. Soc., 2008, 130, 6692–6693.
163. J. Zhang, Y. Fu, M. H. Chowdhury and J. R. Lakowicz, Nano Lett., 2007, 7, 2101–2107.
164. J. Zhang, Y. Fu, M. H. Chowdhury and J. R. Lakowicz, J. Phys. Chem. C, 2007, 111, 11784–11792.
165. T. Ming, H. Chen, R. Jiang, Q. Li and J. Wang, J. Phys. Chem. Lett., 2012, 3, 191–202.
166. S. Eustis and M. A. El-Sayed, Chem. Soc. Rev., 2006, 35, 209–217.
167. A. Hartstein, J. Kirtley and J. Tsang, Phys. Rev. Lett., 1980, 45, 201.
168. R. F. Aroca, D. J. Ross and C. Domingo, Appl. Spectrosc., 2004, 58, 324A–338A.
169. E. A. Coronado and G. C. Schatz, J. Chem. Phys., 2003, 119, 3926–3934.
170. M. Osawa, Near-Field Optics and Surface Plasmon Polaritons, Springer, 2001, pp. 163–187.
171. D. Ross and R. Aroca, J. Chem. Phys., 2002, 117, 8095–8103.
172. G. C. Papavassiliou, Prog. Solid State Chem., 1979, 12, 185–271.
173. T. C. Choy, Effective medium theory: principles and applications, Oxford University Press, 2015.
174. M. Khardani, M. Bouaïcha and B. Bessaïs, Phys. Status Solidi C, 2007, 4, 1986–1990.
175. X. Huang, S. Neretina and M. A. El-Sayed, Adv. Mater., 2009, 21, 4880–4910.
176. J. P. Camden, J. A. Dieringer, Y. Wang, D. J. Masiello, L. D. Marks, G. C. Schatz and R. P. Van Duyne, J. Am. Chem. Soc., 2008, 130, 12616–12617.
177. W. Li, P. H. Camargo, L. Au, Q. Zhang, M. Rycenga and Y. Xia, Angew. Chem., Int. Ed., 2010, 49, 164–168.
178. S. Abalde-Cela, P. Aldeanueva-Potel, C. Mateo-Mateo, L. Rodríguez-Lorenzo, R. A. Alvarez-Puebla and L. M. Liz-Marzán, J. R. Soc., Interface, 2010, 7, S435–S450.
179. P. Kühler, E.-M. Roller, R. Schreiber, T. Liedl, T. Lohmüller and J. Feldmann, Nano Lett., 2014, 14, 2914–2919.
180. L. V. Brown, H. Sobhani, J. B. Lassiter, P. Nordlander and N. J. Halas, ACS Nano, 2010, 4, 819–832.
181. R. Sardar, T. B. Heap and J. S. Shumaker-Parry, J. Am. Chem. Soc., 2007, 129, 5356–5357.
182. J. A. Fan, C. Wu, K. Bao, J. Bao, R. Bardhan, N. J. Halas, V. N. Manoharan, P. Nordlander, G. Shvets and F. Capasso, Science, 2010, 328, 1135–1138.
183. K. Imura, H. Okamoto, M. K. Hossain and M. Kitajima, Nano Lett., 2006, 6, 2173–2176.
184. G. Chen, Y. Wang, L. H. Tan, M. Yang, L. S. Tan, Y. Chen and H. Chen, J. Am. Chem. Soc., 2009, 131, 4218–4219.
185. A. J. Mastroianni, S. A. Claridge and A. P. Alivisatos, J. Am. Chem. Soc., 2009, 131, 8455–8459.
186. H. Wang, K. Schaefer and M. Moeller, J. Phys. Chem. C, 2008, 112, 3175–3178.
187. D.-K. Lim, K.-S. Jeon, H. M. Kim, J.-M. Nam and Y. D. Suh, Nat. Mater., 2010, 9, 60–67.
188. J. M. McMahon, A.-I. Henry, K. L. Wustholz, M. J. Natan, R. G. Freeman, R. P. Van Duyne and G. C. Schatz, Anal. Bioanal. Chem., 2009, 394, 1819–1825.
189. J. Zhu, J. Nanopart. Res., 2009, 11, 785–792.
190. J. Ye, F. Wen, H. Sobhani, J. B. Lassiter, P. V. Dorpe, P. Nordlander and N. J. Halas, Nano Lett., 2012, 12, 1660–1667.
191. W. Yan, L. Xu, C. Xu, W. Ma, H. Kuang, L. Wang and N. A. Kotov, J. Am. Chem. Soc., 2012, 134, 15114–15121.
192. L. Xu, W. Yan, W. Ma, H. Kuang, X. Wu, L. Liu, Y. Zhao, L. Wang and C. Xu, Adv. Mater., 2015, 27, 1706–1711.
193. Y. Li, X. Qi, C. Lei, Q. Yue and S. Zhang, Chem. Commun., 2014, 50, 9907–9909.
194. C. Sönnichsen and A. P. Alivisatos, Nano Lett., 2005, 5, 301–304.
195. A. M. Funston, C. Novo, T. J. Davis and P. Mulvaney, Nano Lett., 2009, 9, 1651–1658.
196. L. S. Slaughter, Y. Wu, B. A. Willingham, P. Nordlander and S. Link, ACS Nano, 2010, 4, 4657–4666.
197. A. McLintock, N. Hunt and A. W. Wark, Chem. Commun., 2011, 47, 3757–3759.
198. Z. Nie, D. Fava, M. Rubinstein and E. Kumacheva, J. Am. Chem. Soc., 2008, 130, 3683–3689.
199. J. Kumar and K. G. Thomas, J. Phys. Chem. Lett., 2011, 2, 610–615.
200. A. Lee, A. Ahmed, D. P. dos Santos, N. Coombs, J. I. Park, R. Gordon, A. G. Brolo and E. Kumacheva, J. Phys. Chem. C, 2012, 116, 5538–5545.
201. A. Lee, G. F. Andrade, A. Ahmed, M. L. Souza, N. Coombs, E. Tumarkin, K. Liu, R. Gordon, A. G. Brolo and E. Kumacheva, J. Am. Chem. Soc., 2011, 133, 7563–7570.
202. M. Rycenga, M. H. Kim, P. H. Camargo, C. Cobley, Z.-Y. Li and Y. Xia, J. Phys. Chem. A, 2009, 113, 3932–3939.
203. W. Ma, M. Sun, L. Xu, L. Wang, H. Kuang and C. Xu, Chem. Commun., 2013, 49, 4989–4991.
204. W. Song, L. Ding, Y. Chen and H. Ju, Chem. Commun., 2016, 52, 10640–10643.
205. Y. C. Cao, R. Jin and C. A. Mirkin, Science, 2002, 297, 1536–1540.
206. G. L. Liu and L. P. Lee, Appl. Phys. Lett., 2005, 87, 074101.
207. J. Wang, L. Yang, S. Boriskina, B. Yan and B. r. M. Reinhard, Anal. Chem., 2011, 83, 2243–2249.
208. S. Basu, S. Pande, S. Jana, S. Bolisetty and T. Pal, Langmuir, 2008, 24, 5562–5568.
209. L. Sun, Y. Song, L. Wang, C. Guo, Y. Sun, Z. Liu and Z. Li, J. Phys. Chem. C, 2008, 112, 1415–1422.
210. K. L. Wustholz, A.-I. Henry, J. M. McMahon, R. G. Freeman, N. Valley, M. E. Piotti, M. J. Natan, G. C. Schatz and R. P. V. Duyne, J. Am. Chem. Soc., 2010, 132, 10903–10910.
211. W. Li, P. H. Camargo, X. Lu and Y. Xia, Nano Lett., 2008, 9, 485–490.
212. G. Chen, Y. Wang, M. Yang, J. Xu, S. J. Goh, M. Pan and H. Chen, J. Am. Chem. Soc., 2010, 132, 3644–3645.
213. K. D. Alexander, M. J. Hampton, S. Zhang, A. Dhawan, H. Xu and R. Lopez, J. Raman Spectrosc., 2009, 40, 2171–2175.
214. F. L. Yap, P. Thoniyot, S. Krishnan and S. Krishnamoorthy, ACS Nano, 2012, 6, 2056–2070.
215. W. J. Cho, Y. Kim and J. K. Kim, ACS Nano, 2011, 6, 249–255.
216. R. G. Larson, Angew. Chem., Int. Ed., 2012, 51, 2546–2548.
217. J. Fang, S. Du, S. Lebedkin, Z. Li, R. Kruk, M. Kappes and H. Hahn, Nano Lett., 2010, 10, 5006–5013.
218. M.-H. Lin, H.-Y. Chen and S. Gwo, J. Am. Chem. Soc., 2010, 132, 11259–11263.
219. C.-F. Chen, S.-D. Tzeng, H.-Y. Chen, K.-J. Lin and S. Gwo, J. Am. Chem. Soc., 2008, 130, 824–826.
220. L. Scarabelli, M. Coronado-Puchau, J. J. Giner-Casares, J. Langer and L. M. Liz-Marzán, ACS Nano, 2014, 8, 5833–5842.
221. Y. Yang, Y. H. Lee, I. Y. Phang, R. Jiang, H. Y. F. Sim, J. Wang and X. Y. Ling, Nano Lett., 2016, 16, 3872–3878.
222. Q. Su, X. Ma, J. Dong, C. Jiang and W. Qian, ACS Appl. Mater. Interfaces, 2011, 3, 1873–1879.
223. P. Schuck, D. Fromm, A. Sundaramurthy, G. Kino and W. Moerner, Phys. Rev. Lett., 2005, 94, 017402.
224. H. Ma and J. Hao, Chem. Soc. Rev., 2011, 40, 5457–5471.
225. S. Gómez-Graña, J. Pérez-Juste, R. A. Alvarez-Puebla, A. Guerrero-Martínez and L. M. Liz-Marzán, Adv. Opt. Mater., 2013, 1, 477–481.
226. W. Han and Z. Lin, Angew. Chem., Int. Ed., 2012, 51, 1534–1546.
227. P. J. Yunker, T. Still, M. A. Lohr and A. Yodh, Nature, 2011, 476, 308–311.
228. L. Cui, J. Zhang, X. Zhang, L. Huang, Z. Wang, Y. Li, H. Gao, S. Zhu, T. Wang and B. Yang, ACS Appl. Mater. Interfaces, 2012, 4, 2775–2780.
229. J. Pang, S. Xiong, F. Jaeckel, Z. Sun, D. Dunphy and C. J. Brinker, J. Am. Chem. Soc., 2008, 130, 3284–3285.
230. S. Yang, C.-F. Wang and S. Chen, J. Am. Chem. Soc., 2011, 133, 8412–8415.
231. A. Dong, J. Chen, S. J. Oh, W.-k. Koh, F. Xiu, X. Ye, D.-K. Ko, K. L. Wang, C. R. Kagan and C. B. Murray, Nano Lett., 2011, 11, 841–846.
232. S. Xiong, D. R. Dunphy, D. C. Wilkinson, Z. Jiang, J. Strzalka, J. Wang, Y. Su, J. J. de Pablo and C. J. Brinker, Nano Lett., 2013, 13, 1041–1046.
233. Z. Zhang, Y. Fu, W. Yu, X. Qin, Z. Xue, Y. Liu, D. Luo, C. Yan, X. Sun and T. Wang, Adv. Mater., 2016, 28, 9589–9595.
234. M. Li, S. K. Cushing and N. Wu, Analyst, 2015, 140, 386–406.
235. J. Dong, Z. Zhang, H. Zheng and M. Sun, Nanophotonics, 2015, 4, 472–490.
236. K. Sugawa, T. Tamura, H. Tahara, D. Yamaguchi, T. Akiyama, J. Otsuki, Y. Kusaka, N. Fukuda and H. Ushijima, ACS Nano, 2013, 7, 9997–10010.
237. M. H. Chowdhury, K. Ray, S. K. Gray, J. Pond and J. R. Lakowicz, Anal. Chem., 2009, 81, 1397–1403.
238. P. Pompa, L. Martiradonna, A. Della Torre, F. Della Sala, L. Manna, M. De Vittorio, F. Calabi, R. Cingolani and R. Rinaldi, Nat. Nanotechnol., 2006, 1, 126.
239. Y.-H. Chan, J. Chen, S. E. Wark, S. L. Skiles, D. H. Son and J. D. Batteas, ACS Nano, 2009, 3, 1735–1744.
240. Z. Mei and L. Tang, Anal. Chem., 2016, 89, 633–639.
241. S. Masuda, Y. Yanase, E. Usukura, S. Ryuzaki, P. Wang, K. Okamoto, T. Kuboki, S. Kidoaki and K. Tamada, Sci. Rep., 2017, 7, 3720.
242. A. Camposeo, L. Persano, R. Manco, Y. Wang, P. Del Carro, C. Zhang, Z.-Y. Li, D. Pisignano and Y. Xia, ACS Nano, 2015, 9, 10047–10054.
243. F. Neubrech, C. Huck, K. Weber, A. Pucci and H. Giessen, Chem. Rev., 2017, 117, 5110–5145.
244. A. Pucci, F. Neubrech, D. Weber, S. Hong, T. Toury and M. L. de La Chapelle, Phys. Status Solidi B, 2010, 247, 2071–2074.
245. N. J. Halas, S. Lal, W.-S. Chang, S. Link and P. Nordlander, Chem. Rev., 2011, 111, 3913–3961.
246. H. Wang, J. Kundu and N. J. Halas, Angew. Chem., Int. Ed., 2007, 46, 9040–9044.
247. F. Le, D. W. Brandl, Y. A. Urzhumov, H. Wang, J. Kundu, N. J. Halas, J. Aizpurua and P. Nordlander, ACS Nano, 2008, 2, 707–718.
248. W. Zhang, H. Xiong, M. Chen, X. Zhang and S. Wang, Biosens. Bioelectron., 2017, 96, 55–61.

---

Footnote

† These authors contributed equally to this work.

---