WO2011101837A1

WO2011101837A1 - Substrate for surface enhanced raman spectroscopy

Info

Publication number: WO2011101837A1
Application number: PCT/IE2011/000013
Authority: WO
Inventors: Damian John Aherne; John Moffat Kelly; Deirdre Marie Ledwith; Iouri Kuzmich Gounko; Matthew Gara; Steven Ernest John Bell
Original assignee: The Provost, Fellows, Foundation Scholars, And The Other Members Of Board, Of The College Of The Holy And Undivided Trinity Of Queen Elizabeth, Near Dublin; The Queen's University Of Belfast
Priority date: 2010-02-16
Filing date: 2011-02-16
Publication date: 2011-08-25

Abstract

A substrate for Surface Enhanced Raman Spectroscopy (SERS) comprising an enclosed gold- silver hollow nanostructure. The nanostructure may be triangular in shape or the nanostructure may be disc-like in shape. The nanostructure may be formed from a silver nanoprism template. The nanostructure may comprise a gold-silver alloy.

Description

"Substrate for Surface Enhanced Raman Spectroscopy"

Introduction

Anisotropic noble metal nanoparticles have been the focus of intense research activity over the last decade due to unique optical properties such as the Localized Surface Plasmon Resonance (LSPR).^['^"5]

The LSPR can result in very large enhancements of the local electric field at close proximity to the nanoparticle surface and forms the basis of interesting phenomena such as Metal-Enhanced Fluorescence (MEF)^[6' ^7] and Surface-Enhance Raman Spectroscopy (SERS).^[8] The spectral position of the LSPR is controlled by nanoparticle size, composition, shape and aspect ratio; but if a metal nanoparticle is hollow, then the position of the LSPR can be controlled by adjusting the wall thickness without much change to the nanoparticle 's size, shape or aspect ratio J¹' ^9] In addition, the existence of the hollow cavity raises the exciting possibility of encapsulation and transport of molecules and materials of interest/¹⁰¹

In recent years, galvanic replacement reactions have been employed to produce bimetallic hollow nanostructures of a range of shapes. The galvanic replacement reaction is driven by the deposition of a metal of higher redox potential onto a template nanoparticle of material with a lower redox potential. The difference in redox potentials drives the oxidation of the template material by the metal salt precursor of the metal being reduced resulting in a hollow nanostructure. The most common example is the formation of hollow AuAg nanostructures by epitaxial deposition of gold on the edges and faces of a silver nanoparticle template/⁹' ^{1 l l 4]} This proceeds by reduction of AuCU ions, alloying of the deposited gold with the underlying silver and oxidation of the remaining silver nanoparticle template leaving an almost fully enclosed hollow nanostructure.^[9] In addition, Pt- and Pd-containing hollow nanostructures can also be prepared by this approach. ^{[l 16]} Researchers have refined this approach to produce a whole range of enclosed AuAg nanostructures such as cubic nanoboxes and nanocages ⁹' ¹³¹ cylindrical nanotubes,^{[, ]} spherical nanoshells^{[, 7]} and even multi-walled hollow nanostructures. ^{[, 8]}

The reaction conditions of galvanic replacement can be modified so that deposition of metal occurs only on the edges and corners of the template nanoparticle. After oxidation of the template nanoparticle, an open frame-like structure is left behind. For example, the synthesis of cubic nanoboxes and nanocages can be modified to result in the formation of cubic nanoframes.^[19i Moreover, nearly all previous attempts at utilizing galvanic replacement to produce a AuAg triangular nanobox from a silver nanoprism template has resulted in either the formation of triangular nanoframes,^[20"22] or, with deposition of increasing amounts of gold, back-filling of the triangular nanoframes to form a solid AuAg alloy nanoprism. ^[20] Until now, the formation of an enclosed triangular nanobox from a silver nanoprism template has remained very elusive. The closest example thus far is that of an incomplete nanobox from thick silver nanoprisms.^{[1 ]} Statements of Invention

According to the invention there is provided a substrate for Surface Enhanced Raman

Spectroscopy (SERS) comprising an enclosed gold-silver hollow nanostructure.

The nanostructure may be a flat nanoparticle (i.e. reduced size in one dimension only) and may be shaped, for example the nanostructure may be hexagonal shape or disc-like in shape, or triangular shape.

The nanostructure may be triangular in shape, the nanostructure may be disc-like in shape.

The nanostructure may be formed from a silver nanoprism template. The nanostructure may comprise a gold-silver alloy.

The nanostructure may have an aspect ratio of greater than 1 : 1 . The nanostructure may have an aspect ratio of between about 1 .3: 1 and about 100: 1. The nanostructure may have an aspect ratio of between about 1.3: 1 and about 50: 1. The nanostructure may have an aspect ratio of between about 1.3: 1 and about 20: 1. The nanostructure may have an aspect ratio of between about 1 .3: 1 and about 10: 1. The nanostructure may have an aspect ratio of between about 1.3: 1 and about 5: 1. The nanostructure may have an aspect ratio of between about 1.8: 1 and about 10: 1 .

The nanostructure may have an average perpendicular outer diameter of between about 10 nm and about 500 nm. The nanostructure may have an average perpendicular outer diameter of between about 20 nm and about 250 nm. The nanostructure may have an average perpendicular outer diameter of between about 25 nm and about 150 nm. The nanostructure may have an average perpendicular outer diameter of between about 30 nm and about 100 nm. The nanostructure may have an average perpendicular outer diameter of between about 40 nm and about 80 nm.

The hollow may have an average perpendicular diameter of between about 21 nm and about 249 nm. The hollow may have an average perpendicular diameter of between about 26 nm and about 149 nm. The hollow may have an average perpendicular diameter of between about 31 nm and about 99 nm. The hollow may have an average perpendicular diameter of between about 41 nm and about 79 nm. The hollow may have an average thickness of between about 0.5 nm and about 50 nm. The hollow may have an average thickness of between^" about 1 nm and about 30 nm The hollow may have an average thickness of between about 1.5 nm and about 15 nm. The hollow may have an average thickness of between about 2nm and about l Onm.

The nanostructure may comprise a wall defining the hollow. The wall may have an average thickness of between about 0.5 nm and about 100 nm. The wall may have an average thickness of between about 1 nm and about 50 nm. The wall may have an average thickness of between about 3 nm and about 40 nm. The wall may have an average thickness of between about 5 nm and about 30 nm.

The nanostructure may have an in-plane dipole Localised Surface Plasmon Resonance (LSPR) spectral position between about 390nm and 2000nm. The nanostructure may have an in-plane dipole Localised Surface Plasmon Resonance (LSPR) spectral position between about 400nm and 1500nm. The nanostructure may have an in-plane dipole Localised Surface Plasmon Resonance (LSPR) spectral position between about 420nm and 1 l OOnm. The nanostructure may have an in-plane dipole Localised Surface Plasmon Resonance (LSPR) spectral position between about 450nm and 950nm. The nanostructure may have an in-plane dipole Localised Surface Plasmon Resonance (LSPR) spectral position between about 500nm and 850nm.

The invention further provides a method for probing a sample by SERS comprising the steps of: (a) adding a SERS substrate as described herein to a test sample; and

(b) generating a SERS spectrum by recording light which scatters from the sample when the sample is illuminated by a Raman probe beam.

The SERS substrate may be substantially non-aggregated. The SERS substrate may comprise substantially non-aggregated enclosed gol-silver hollow nanostructures.

The sample may be selected from one or more of a biological sample, a chemical sample, a food sample, or an environmental sample. The biological sample may be a cellular sample or a body fluid sample.

The invention also provides for the use of a substrate as described herein for biomedical screening of disease states.

The invention further provides for a molecular sensing device comprising a substrate as described herein. The molecular sensing device may be used for the detection of low concentrations of compounds in a sample. The sample may be selected from one or more of a biological sample, a chemical sample, a food sample, or an environmental sample. The biological sample may be a cellular sample or a body fluid sample.

« The invention also provides an enclosed triangular gold-silver hollow nanostructure. The nanostructure may be formed from a silver nanoprism template. The nanostructure may comprise a gold-silver alloy.

The nanostructure may have an aspect ratio of greater than 1 : 1 . The nanostructure may have an aspect ratio of between about 1.3: 1 and about 100: 1 . The nanostructure may have an aspect ratio of between about 1 .3: 1 and about 50: 1 . The nanostructure may have an aspect ratio of between about 1.3: 1 and about 20: 1 . The nanostructure may have an aspect ratio of between about 1 .3 : 1 and about 10: 1 . The nanostructure may have an aspect ratio of between about 1.3: 1 and about 5: 1 .

The nanostructure may have an average perpendicular outer diameter of between about 10 nm and about 500 nm. The nanostructure may have an average perpendicular outer diameter of between about 20 nm and about 250 nm. The nanostructure may have an average perpendicular outer diameter of between about 25 nm and about 150 nm. The nanostructure may have an average perpendicular outer diameter of between about 30 nm and about 100 nm. The nanostructure may have an average perpendicular outer diameter of between about 40 nm and about 80 nm. The nanostructure may have an average perpendicular outer diameter of between about 10 nm and about 500 nm.

The hollow may have an average perpendicular diameter of between about 21 nm and about 249 nm. The hollow may have an average perpendicular diameter of between about 26 nm and about 149 nm. The hollow may have an average perpendicular diameter of between about 31 nm and about 99 nm. The hollow may have an average perpendicular diameter of between about 41 nm and about 79 nm. The hollow may have an average thickness of between about 0.5 nm and about 50 nm. The hollow may have an average thickness of between about 1 nm and about 30 nm The hollow may have an average thickness of between about 1.5 nm and about 15 nm. The hollow may have an average thickness of between about 2nm and about 1 Onm.

The invention also provides for the use of a nanostructure as described herein for the

encapsulation of a material. The material may be selected from one or more of a biologically active compound such as a drug, a contrast agent such as a radioisotope, a magnetic material, or a dye such as a fluorophore.

The invention further provides for the use of a nanostructure as described herein as a substrate for surface enhanced Raman spectroscopy (SERS). In a further aspect, the invention provides a solution phase process for forming an enclosed hollow triangular nanobox from a silver nanoprism template comprising the steps of: (a) providing a silver nanoprism in an aqueous solution; and

(b) adding a reducing agent and a gold salt to the silver nanoprism solution wherein the gold salt is added at a rate of between 0.1 and l Oml/min. The gold salt may be added at a rate of between 0.2 and 5 ml/min, the gold salt may be added at a rate of between 0.3 and 3 ml/min, for example the gold salt may be added at a rate of between 0.5 and 2 ml/min such as at a rate of 1 ml/min

The process may be performed between -5 °C and 100 °C, the process may be performed between 0 °C and 50 °C, the process may be performed between 0 °C and 40 °C, the process may be performed between 5 °C and 35 °C, the process may be performed between 10 °C and 30 °C, for example the process may be performed at room temperature.

The gold salt may be HAuCU. The gold salt may be added at a mole ratio of between about 0.1 : 1 to about 100: 1 gold : silver, the gold salt may be added at a mole ratio of between about 0.5: 1 to about 50: 1 gold : silver, the gold salt may be added at a mole ratio of between about 1 : 1 to about 10: 1 gold : silver, the gold salt may be added at a mole ratio of between about 1 : 1 to about 5: 1 gold : silver. The reducing agent may be ascorbic acid. The gold salt may be added at a mole ratio of between about 0.02: 1 to about 5: 1 gold : reducing agent, the gold salt may be added at a mole ratio of between about 0.05:1 to about 2: 1 gold : reducing agent, the gold salt may be added at a mole ratio of between about 0.1 : 1 to about 1 : 1 gold : reducing agent, the gold salt may be added at a mole ratio of between about 0.2: 1 to about 0.5: 1 gold : reducing agent.

The terms "nanobox" and "enclosed nanostructure" as used herein refer to a nanostructure comprising walls (flat faces of the nanostructure) defining a hollow interior. The walls may have a degree of porosity (% surface area of the nanostructure that consists of pores) of less than about 20%, such as less than about 15%, or less than about 10%, such as less than about 5%, for example less than about 2%. In some embodiments, the walls of the nanostructure may be considered as substantially pore fee. The terms "nanobox" and "enclosed nanostructure" as used herein do not encompass nanoframe structures which have a peripheral frame defining a hollow interior or nanocage structures which have porous walls with a degree of porosity of greater than about 50% defining a hollow interior.

Brief Description of the Drawings

The invention will be more clearly understood from the following description of an embodiment thereof, given by way of example only, with reference to the accompanying drawings, in which:-

Fig. 1 is a UV-Vis extinction spectra of Sample 1 and Samples 3 to 8. Sample 2 is omitted for clarity. The volumes of the samples have been taken into account and thus the spectra have been adjusted accordingly to represent samples of the same nanoparticle concentration. Samples 3 to 8 are prepared by addition of 0.5, 1.0, 2.0, 3.0, 4.0 and 5.0 μιηοΐ of HAuCU (0.5 mM) respectively to the aliquots of silver nanoprisms in the presence of excess ascorbic acid;

Figs. 2A to H are TEM images of flat-lying nanoparticles from Samples 1 to 8. (A) Sample 1 : silver nanoprisms. (B - H) Samples 2 to 8: Results of adding 0.25, 0.5, 1 .0, 2.0, 3.0, 4.0 and 5.0 μιηοΐ of HAuCI₄ respectively to 1 .5 μιηοΐ of silver in the form of nanoprisms, in the presence of excess ascorbic acid. Completely enclosed nanostructures are visible in panels F, G and H (Samples 6, 7 and 8 respectively);

Figs. 3A to C are TEM images showing edge-orientated hollow AuAg nanostructrues from (A) Sample 6, (B) Sample 7 and (C) Sample 8;

Fig. 4 are plots of evolution of percentage composition of gold in metal nanoparticle samples for two theoretical situations, (·) complete galvanic replacement of silver with gold and (x) no galvanic replacement for each sample. Also plotted are the experimental data (□) as determined from EDS for comparison. The small discrepancy between the experimental data and the no galvanic replacement data points may be accounted for by the presence of AgCl;

Fig. 5A is a schematic of the formation of triangular, hollow AuAg nanoboxes from Ag nanoprisms and Fig. 5 B is a schematic of the growth of bridging material between walls of nanoboxes; Figs. 6A to E are TEM images of flat-lying nanoparticles from Samples 10 to 14 (A to E respectively). These have been prepared by addition of 1.0, 2.0, 3.0, 4.0 and 5.0 μπιοΐ HAuCl₄, respectively, to silver nanoprisms without sufficient ascorbic acid to reduce all the added HAuCl₄. The HAuCl₄ is in excess in sufficient quantities to oxidize all the silver present;

Figs. 7A and B are TEM images of edge-orientated AuAg nanoboxes prepared without sufficient ascorbic acid to reduce all the added HAuCl₄ from (A) Sample 13 and (B) Sample 14; Fig. 8 (A) is a TEM image of flat-lying AuAg nanoboxes from Sample 8 with dark spots illustrating material in the nanobox cavity; (B) is a TEM image of flat-lying AuAg nanoboxes from Sample 8 with a torus-like structure due to a hole through the material occupying the nanobox cavity; (C, D) are TEM images of edge-orientated AuAg nanoboxes from Sample 6 (left) and Sample 7 (right) showing material bridging the flat walls of the nanoboxes;

Fig. 9 is a HAADF-STEM image of AuAg alloy nanoboxes from Sample 7. The gray lines are the paths taken for the intensity profile measurements plotted in 10. The high contrast between the interior and the edge of each nanostructure cannot be simply explained by the presence of a silver core, but rather by the presence of a cavity;

Fig. 10 is a plot showing the line profiles of relative intensity of HAADF signal along paths indicated by the gray lines near the top (Top profile) and bottom (Bottom profile) of Fig. 9;

Fig. 1 1A and B are cut-away diagrams showing representation of the dipole arrangement for the two models used in the DDA calculations for simulations of the extinction spectra for Sample 6. Each sphere represents one point dipole, with dark gray for gold dipoles and light gray for silver dipoles. The insets show the models before being cut in half. (A) shows the silver core-gold shell model. (B) shows the hollow (water-filled) alloy shell model. The alloying of silver and gold is achieved by a random assignment of 1 nm silver and gold dipoles in a 31 % to 69 % ratio;

Fig. 12 shows plots of experimental extinction spectra and calculated extinction spectra of silver core-gold shell and alloy shell nanostructures for Sample 6 (top), Sample 7 (middle) and Sample 8 (bottom); Fig. 13 is a plot of maxima of calculated and experimental spectra for Sample 6, 7 and 8;

Fig. 14. is a UV/vis absorption spectra of the unaggregated Au colloids used in Example 6, solutions were the as-prepared samples diluted 2x with water;

Fig. 15 is a Uv /vis absorption spectra showing no indication of aggregation following addition of SERS test compound (adenine) but large broadening of the surface plasmon band as aggregation is induced through the addition of 0.02 mol dm^"3 MgS0₄; Fig. 16 is a Uv /vis absorption spectra showing no indication of aggregation following addition of SERS test compound (thiophenol) but large broadening of the surface plasmon band as aggregation is induced through the addition of 0.02 mol dm^"3 MgSC^;

Fig. 17A to C are TEM Images of AuAg nanoboxes (1.5:5 μιηοΐ Ag:Au, Sample 8) and adenine within Polycarboph.il. Images recorded at 29,000x (A), 7,000x (B), and 2,550x (C)

magnification;

Fig. 18 is a plot showing SERS enhancements obtained using unaggregated AuAg nanoboxes with the amount of Au increasing through samples 1 .5:0.5 μηιοΐ - 1.5:5 μηιοΐ Ag:Au (Samples 3, 4, 5, 6 and 8 respectively) using thiophenol as SERS test compound. Intensity scale is the same for all spectra, apart from an offset for clarity;

Fig. 19 is a plot showing SERS enhancements obtained using unaggregated AuAg nanoboxes with the amount of Au increasing through samples 1 .5:0.5 μιηοΐ - 1.5:5 μιηοΐ (Samples 4, 5, 6 and 8 respectively) Ag:Au using adenine as SERS test compound. Intensity scale is the same for all spectra, apart from an offset for clarity;

Fig. 20 is a SERS spectra comparing unaggregated AuAg nanoboxes (sample containing 1 .5: 5μηιο1 Ag:Au, Sample 8) with adenine as the SERS analyte and particles which have been aggregated with 0.2 mol dm^"3 MgS0₄. Intensity scale is the same for all spectra, apart from an offset for clarity;

Fig. 21 is a plot showing the relationship between the absorption at 785 nm (solid line) and the SERS intensity (dotted line) for samples with increasing Au (Samples 3, 4, 5, 6 and 8); Fig. 22 is a spectra comparing enhancements obtained with unaggregated nanobox samples (Sample 8), aggregated 40 nm Au and aggregated citrate-reduced silver colloid. The colloids were diluted so that irrespective of their orginal concentration, they had 56 mg/dm³ of metal. Intensity scale is the same for all spectra, apart from an offset for clarity;

Fig. 23 is a SERS spectra showing formation of citrate layer on the surface of the nanobox particle (1.5:5 μιηοΐ Ag:Au, Sample 8). (a) before and (b) after addition of Na3Ct . Spectra were collected without aggregation of the particles. Intensity scale is the same for all spectra, apart from an offset for clarity;

Fig. 24 is a plot showing the shift in zeta potential from -29.5 mV to -45.5 mV as the triangular nanobox particles (Sample 8) are stabilised with Na₃Ct; and Fig. 25 are SERS spectra which compares the spectra recorded under equivalent experimental conditions using unaggregated nanoprisms (i.e. solid prisms), aggregated nanoprisms and unaggregated nanoboxes. This demonstrates the increase in signal given by unaggregated nanoboxes over unaggregated nanoprisms. Intensity scale is the same for all spectra, apart from an offset for clarity.

Detailed Description

The invention relates to the formation of hollow triangular AuAg nanostructures from a silver nanoprism template. The invention provides the first example of the synthesis of enclosed triangular AuAg nanostructures (triangular nanoboxes), via galvanic replacement reactions, from silver nanoprisms. These triangular nanoboxes have been studied by TEM and HAADF-STEM imaging, to elucidate their structure.

The invention shows that the nanostructures are hollow and do not consist of a silver core surrounded by a gold shell. In addition, Discrete Dipole Approximation (DDA) calculations for the extinction spectra have been carried out and provide additional evidence that the nanostructures are hollow. It is envisaged that these new triangular nanoboxes are very attractive candidates for encapsulation and transport of materials of interest such as drugs, radioisotopes or magnetic materials. Furthermore, these triangular nanobox structures are potentially an ideal platform for sensing techniques that are based on the enhancement of the local electric field, such - l i as enhanced sensitivity of the plasmon band to changes in the local refractive index and detection of analytes by Surface Enhanced Raman Spectroscopy (SERS).

The present invention discloses employing galvanic replacement reactions to produce completely enclosed hollow nanostructures from a sacrificial silver nanoprism template. With the experimental approach employed here, these structures do not undergo significant dealloying to form pores even upon further addition of gold salt. The invention also shows that the thickness of the nanobox walls is tunable and can be controlled. The formation of these interesting, novel nanostructures has been confirmed by TEM and HAADF-STEM imaging. Extensive DDA calculations which support a hollow vs. core-shell morphology were performed.

The open nature of the structures during early stages of growth and the highly enclosed nature of the final nanoboxes means that the cavity is ideally suited to encapsulation of materials of interest such as drugs, radioisotopes, magnetic materials, fluorophores etc. We believe that these materials and our approach will open the door to a large number of potential applications which could have a major scientific and commercial value.

In order to form a triangular nanobox, it is desirable to minimize oxidation of the silver so that enough silver is present during the reaction to template growth in towards the centre of the nanoparticle. For this reason, one strategy worth pursuing is to maintain a sufficiently high level of reducing agent in the reaction mixture to provide some protection for the silver nanoprisms against oxidation during galvanic replacement. Indeed, Sanedrin et alP^3] and other researchers¹²⁴-¹ have shown that if excess reducing agent, with respect to added AuCU^", is present, it is possible to preserve the silver nanoprism template during the reaction resulting in a silver nanoprism with a coating of gold at the edges. In addition, we have shown that deposition of a very thin layer of gold at the edges has proven to be a very effective way of stabilizing silver nanoprisms against etching.¹²⁵¹ However, it has not yet been demonstrated that this strategy can be pursued to grow a layer of gold across the large flat { 1 1 1 } faces of silver nanoprisms to either coat them with a complete thin shell of gold to form a silver core-gold shell structure or to form a nanobox structure.

We have demonstrated the synthesis of triangular nanoboxes for the first time and provided electron microscopy data that confirms their hollow structure. In addition, Discrete Dipole Approximation (DDA) calculations for silver core-gold shell and alloy nanobox morphologies were carried out to provide additional supporting evidence that the nanoboxes are hollow.

The invention will be more clearly understood from the following non-limiting examples thereof.

Examples

Experimental methods

Silver nanoprisms:

Sample 1 : To 40 ml of Millipore water in a beaker, 800 μΐ_^ of silver seed solution^t26] and 600 iL of 10 mM ascorbic acid were added. This was followed by 24 ml of 0.5 mM AgNC^ at a rate of 5 ml min^"' by syringe pump to form the silver nanoprisms. One eighth of this solution was set aside as Sample 1 .

Samples 2 to 8: The remaining silver nanoprism solution was split into 7 equal volume aliquots of 8.175 ml. The samples were prepared by adding the appropriate amount of ascorbic acid followed by the appropriate volume of 0.5 m HAuCU at 1 ml min^'1 to each aliquot of silver nanoprisms. See Table 1 below for details.

Table 1 . Reagents for Samples 2 to 8.

Sample Vol. 10 mM Vol. 0.5 mM

AA soln. {\iL) HAuCl₄ soln.

(mL)

2 150 0.5 3 150 1.0 4 225 2.0 5 375 _, 4.0 6 525 6.0 7 675 8.0

825 10.0 Sample 9: To 35 ml of Millipore water in a beaker, 700 of silver seed solution^{1 J} and 525 μΕ of 10 mM ascorbic acid were added. This was followed by 21 ml of 0.5 mM AgNC^ at a rate of 5 ml min^"1 by syringe pump to form the silver nanoprisms. One seventh of this solution was set aside as Sample 9.

Samples 10 to 14: The remaining silver nanoprism solution was split into 6 equal volume aliquots of 8.175 mL. The samples were prepared by adding the appropriate amount of ascorbic acid followed by the appropriate volume of 0.5 mM HAuC at 1 ml min^"1 to five of the aliquots of silver nanoprisms (one aliquot spare). See Table 2 below for details.

Table 2. Reagents for Samples 10 to 14.

Sample Vol. 10 mM Vol. 0.5 mM

AA soln. (μί) HAuCU soln.

(mL)

10 150 2.0

1 1 225 4.0

12 375 6.0

13 525 8.0

14 675 10.0

Electron Microscopy: Samples were prepared for electron microscopy by deposition and drying of a drop of an aqueous solution on to a formvar-coated 300 mesh copper grip. TEM measurements were carried out with a JEOL Jem-2100 LaB₆ operating at 200 kV. The HAADF- STEM imaging was carried out with a Titan 80-300 TEM operating at 300 kV with Z-contrast imaging conditions that avoided the collection of any electrons that arise due to Bragg scattering.

Example 1 - Triangular Nanobox Growth

A series of samples were produced by adding increasing amounts of gold to aliquots of silver nanoprism solution in the presence of excess reducing agent, ascorbic acid. Sample 1 is a sample of silver nanoprisms that was produced by previously published procedure. ^[26] This was split into 7 aliquots of equal volume, each containing 1.5 μιηοΐ of silver. Samples 2 to 8 were prepared by initially adding excess ascorbic acid with respect to the amount of gold salt to be added to each sample. This was then followed by addition of 0.25, 0.5, 1 .0, 2.0, 3.0, 4.0 and 5.0 μιηοΐ of HAuCl₄ (0.5 ITIM) respectively, at a rate of 0.5 μηιοΙ min^~\ By preparing a series of samples in this manner it was possible to monitor the progress of formation of the triangular nanoboxes. Reaction progress can be monitored by measuring the evolution of the Localized Surface Plasmon Resonance (LSPR). In Fig. 1 , one can see that at early stages of the reaction, there is an initial red-shift and broadening of the LSPR, this is followed by a progressive blue- shift and sharpening of the LSPR as increased quantities of gold are added. The evolution of the LSPR can be correlated to the changing structure of the nanoparticles during the reaction by looking at Transmission Electron Microscopy (TEM) data of each sample.

Referring to Figs. 2B and Figs. 2C it is clear to see that the edges of the nanoprisms are mostly intact while many large holes have formed in the centre of the nanoprisms. This is due to significant oxidation of the silver by AuCl₄ ions despite the presence of excess ascorbic acid. However, it should be pointed out that for Sample 3 (Figs. 2C) enough AuCl₄ has been added to oxidize all the silver but we can see that much of the silver still remains despite the significant oxidation. If all the silver had been oxidized away or alloyed with the deposited gold then we would be left with bare triangular nanoframes and that is clearly not the case. The initial deposition of gold at the edges is exactly what one would expect based on the results of previous experimental work by other researchers mentioned above, and is consistent with the general observation that initial deposition of gold preferentially takes place on higher energy faces with continued growth of gold here and oxidation of silver away from the more stable { 1 1 1 } faces.¹²⁷' ^28]

As the reaction proceeds, the gold-coated edges become much more distinct. Unexpectedly, it is shown from Samples 4 to 6 in Figs. 2D-F that the holes in the centre of the nanoparticles, instead of increasing in size, become smaller and eventually disappear. In addition, it is possible to see an increase in the degree of contrast between the edges and middle of the flat-lying nanoparticles, and this contrast is maintained as more gold is added, see Samples 6, 7 and 8 in Figs. 2F-H. These observations could well be explained by the formation of a hollow nanostructure, but the structure of the nanoparticles is certainly not completely clear from these TEM images. The contrast could be explained by a silver core-gold shell structure since silver has a lower scattering cross section than gold for electrons in the electron beam in the TEM. Yet another explanation is that the nanoparticles are simply thicker at the edges than they are in the middle. In order to determine the true structure of the nanoparticles, a TEM analysis was conducted of nanoparticles that are arranged such that they are resting on their edges. In Figs. 3, there are many examples of edge-orientated nanoparticles from Samples 6, 7 and 8. The first thing to note is that the nanoparticles are certainly not thicker at the edges than in the middle. Therefore the nanoparticles are either hollow or have a core-shell structure. There is a high level of contrast between the interior and the flat sides of the nanoparticles, much higher than would be expected if the interior was filled with silver. Furthermore, in many cases, such as in Figs. 3B, it is possible to see the pattern of nanoparticles lying underneath or on top, by looking through the interior. These observations can only be reasonably explained by the existence of a hollow cavity in the interior of the nanoparticles, i.e. the nanoparticles are triangular nanoboxes. This is quite a remarkable result. Indeed, it is the first time that such triangular nanoboxes have been formed from a silver nanoprism template.

Measurements of the dimensions of the nanoboxes have shown that the size of the interior cavity is not much smaller than the silver nanoprism template (see Table 3, below). For example the average perpendicular diameter of the cavity, as measured from TEM images of flat-lying nanoboxes from Sample 7 is 42 ± 2 nm , while the silver nanoprism perpendicular diameter is 45 ± 2 nm This small decrease is a result of alloying of the outer layer of silver at the edge with the deposited layer of gold. In addition, the thickness of the interior cavity has been measured to be slightly smaller than the thickness of the silver nanoprism templates (see Table 3 below). Taking Sample 7 again as an example, the interior cavity has an average thickness of 4.8 ± 0.8 nm, while the silver nanoprisms have an average thickness of 7.0 ± 0.7 nm. This small decrease can be explained by initial alloying of gold with the outer layer of whatever silver still remained within the newly formed gold edges. Overall, the dimensions of the nanobox are clearly defined by the dimensions of the silver nanoprism template.

An important feature of the nanoboxes is their flat morphology. From the discussion above, it is clear that the nanoboxes inherit this flat morphology from the shape of the silver nanoparticle template. This flat morphology of the nanoboxes can be quantified by defining the Aspect Ratio (AR) as the ratio of perpendicular diameter (Outer diameter) to nanobox thickness (Outer thickness). The aspect ratio is important for the SERS properties of the nanoboxes that will be discussed later. The nanoboxes will typically have an AR that is less than^that or the silver nanoparticle template and will decrease upon further deposition of gold onto the nanoboxes. Table 3. Sizing data for perpendicular diameter and thickness of whole nanoparticle and interior cavity obtained from TEM measurements for Samples 6, 7 and 8.

Sample Cavity Cavity Outer Outer Aspect ratio

diameter thickness diameter (nm) thickness (Outer

(nm) (nm) (nm) thickness/Outer diameter)

1 - - 45± 2 7.0 ± 0.7 6.4

6 43 ± 2 5.1 ± 0.9 59 ± 2 13.2 ± 0.9 4.5

7 42 ± 2 4.8 ± 0.8 60 ± 2 14.3 ± 0.9 4.2

8 41 ± 2 4.9 ± 0.8 61 ± 2 16.1 ± 0.8 3.8

We have demonstrated that the nanoboxes can be produced at room temperature, this is a surprising result. The formation of cubic AuAg nanoboxes from silver nanocubes by galvanic replacement has been reported not to work if the reaction is carried out at room temperature; generally temperatures of about 100° C are required. ^[9] One reason for the failure of reactions at room temperature is that the solubility product of AgCI is much lower at room temperature resulting in the precipitation of AgCI, which interferes with the reaction, yet the present invention does not experience such difficulties here.

Energy-dispersive X-ray Spectroscopy (EDS) analysis of each sample has shown that the Au/Ag ratio is the same as that used to prepare each sample. This means all the silver is still present, either in reduced form or as AgCI. No macroscopic quantities of AgCI precipitate are ever observed in experiments, but by detection of the signal for CI, the EDS analysis has shown that there is actually a small amount of AgCI present in Samples 2 to 6, with the vast majority of silver present in reduced form, while for Samples 7 and 8, about one third of the silver present is in the form of AgCI. It is not surprising that there is some AgCI but it is remarkable that it does not interfere with nanoparticle growth. Fig. 4 shows a plot of how the percentage composition of gold in the metal nanoparticles should vary for two theoretical situations: complete galvanic replacement and no galvanic replacement. The experimentally determined mole fraction of gold in the nanoboxes, with Ag in the form of AgCI excluded from the calculation is also plotted. What is remarkable is that the plot for the experimental values very closely matches for the case of no galvanic replacement, meaning that most of the Ag is preserved in metallic form in the nanoboxes. One would expect results similar to this if silver core-gold shell nanostructures were being formed but we have shown that despite the presence of ascorbic acid, there is a lot of oxidation of silver at early stages in the reaction. Also, the TEM analysis indicates the existence of a cavity in the interior of the indicating that the nanoboxes have a hollow structure. The only reasonable explanation for the presence of metallic silver is if an alloy with gold is formed.

This can be explained as follows: since there is an excess of ascorbic acid, it is reasonable to expect that after initial oxidation of silver from the silver nanoprisms by galvanic replacement, there is co-reduction of Ag⁺ and AuCl₄ to form an alloy. Thus, as the reaction proceeds, an alloy layer grows in towards the centre closing up the large holes that had appeared earlier due to oxidation of the silver, to form a complete AuAg alloy shell with a hollow interior. This oxidation of silver and subsequent co-reduction with gold is important as it is a way to explain the alloying of silver with gold to leave a hollow core since, given the sizes of the nanoparticles and the low temperatures employed here (~ 20° C), the atoms do not have enough energy to alloy simply by interdiffusion of atoms, except perhaps at the interface of gold and silver.'⁹' ^{29, 30]}

The co-reduction of Ag with AuCl₄ ^~ by ascorbic acid might also serve to keep the concentration of Ag⁺ low enough to limit the amount of solid AgCI that is formed. The higher amounts of AgCI for Samples 7 and 8 are probably the result of the presence of increased concentrations of chloride resulting in increased scavenging of Ag⁺ and its incorporation into AgCI. Since the nanoboxes are already formed at this point, the Ag⁺ is not coming from the silver nanoprism template so there must be some minor dealloying of the nanoboxes. Nevertheless, it is particularly interesting that addition of increasing quantities of HAuCl₄ ^~, after formation of complete nanoboxes, does not result in the extensive dealloying and subsequent pore formation that has been observed to take place upon addition of additional AuCU- in other experimental approaches such as the synthesis of AuAg cubic nanoboxes from silver nanocubes.^[9' ¹³¹ Rather, the triangular nanoboxes become increasingly thicker as more and more gold is deposited. This is clear from the data presented in Table 3 above. A schematic diagram illustrating the different stages of a growth model for the triangular nanobox synthesis based on all the data presented above is presented Figs. 5A and 5B. Figs. 5A and 5B shows a cross-sectional view of a typical growing AuAg nanobox, illustrating the manner of growth. In Sample 3, enough gold to oxidize all the silver present has deposited at the edges. Some oxidation of the silver is indicated by the hole but most of the silver remains. For Sample 4, more of the silver has been oxidized away as more gold has been deposited along with silver to form an increasingly larger AuAg alloy shell due to co-reduction of Ag⁺ and AuCl₄ ^~. For Samples 5 and 6, the diagram illustrates oxidation of any remaining silver with closure of the box due to growth of the AuAg alloy shell. In Samples 7 and 8, complete nanoboxes grow increasingly thicker due to deposition of increasing quantities of gold. Fig. 5B illustrates the growth of bridging material between walls of nanoboxes.

Example 2 - Nanobox Synthesis with Insufficient Amount of Reducing Agent

To better understand the growth mechanism, another series of samples were prepared, this time with insufficient ascorbic acid to reduce all of the added AuCl₄ in each reaction; the AuCl ^~~ is in excess in sufficient quantities to oxidize the 1 .5 μιηοΐ of Ag that is present in each aliquot of silver nanoprisms (Sample 9). Fig. 6 shows TEM images of flat-lying nanoparticles from Samples 10 to 14 which correspond to the same amount of added AuCl₄ ^~ as Samples 4 to 8, but with a deficit of ascorbic acid. Samples 10 and 1 1 are essentially triangular nanoframes with a little material within the nanoframes. The nanoframes in Sample 1 1 are clearly much thicker than those in Sample 10 due to the increased amount of gold. There is much less material within the nanoframes in these samples compared to Samples 4 and 5, due to increased etching away of silver by the AuCl₄ , now that AuCl₄ is in excess. This is evidence for the presence of residual silver in the interior of Samples 4 and 5 and therefore supports the growth mechanism outlined in Fig. 5A.

Sample 12 in Fig. 6C shows that there is substantial growth of material between the edges of the nanoframe, partially forming a cavity. Although there is as much gold here as in Sample 6, the removal of silver in the latter stage of growth has resulted in incomplete nanoboxes. But, if one looks at Sample 13 in Fig. 6D, it can see that many more of the nanoboxes are complete, showing that there is either almost enough gold present in the flat walls of the nanoboxes for them to be complete even after removal of silver by excess AuCl₄ ^_, or that deposition of an outer layer of gold has rendered silver in the alloy layer below inaccessible. Lastly, in Fig. 6E, it can see that Sample 14 consists of mostly complete nanoboxes. Once again, the hollow structure of the nanoboxes can be clearly observed by looking at TEM images of edge-orientated nanoparticles. In Fig. 7 one can see TEM images of edge-orientated nanoboxes from Samples 13 and 14. So, it is clear that if samples are prepared with an insufficient ascorbic acid to reduce all the AuCl₄ , then it is still possible to form enclosed nanoboxes as long as enough gold has been added before all the ascorbic acid is consumed.

Example 3 - Bridging Material in Nanobox Cavity

In many of the images of flat-lying nanoboxes (Samples 6, 7 and 8), dark regions are visible, indicating the presence of some material in the interiors of the nanoboxes, see Fig. 8 A. This is attributed to misdirected growth of the flat walls across the middle of the growing nanobox. In the early stages of nanobox formation, the incomplete walls of the nanobox are not completely supported and it is reasonable to imagine that they can flex in towards the interior cavity of the forming nanobox. Continued growth can result in incomplete walls touching and ultimately forming a bridge between the two walls. In some cases this can result in a hole in the middle of the nanoparticle, thus forming a torus-like nanostructure as can be seen in Fig. 8B. The material bridging the walls of the nanoboxes can also be detected in TEM images of edge-orientated nanoboxes, as shown in Fig. 8C and Fig. 8D. The formation of the bridging material between the walls of the nanoboxes is schematically illustrated in Fig. 5B. The presence of this bridging material between the walls of the nanoboxes provides further evidence that the nanoboxes generally have hollow interiors.

Example 4 - High-Angle Annular Dark Field (HAADF) Scanning Tunnelling Electron Microscopy (STEM) imaging

HAADF-STEM imaging, also known as Z-contrast imaging, operates on the basis that higher atomic number atoms incoherently scatter the electrons in the beam to a higher extent and as such is ideally suited for composition analysis of the internal structures of nanoparticles. ^{[3 I ]} The intensity of the signal in a HAADF image is proportional to Z^{1 7}, where Z is the average atomic number of the material that is scattering the electrons. ^[32] With Z values of 47 and 79 for Ag and Au respectively and assuming a linear dependence of the HAADF signal on nanoparticle thickness at the dimensions that are involved here, it should be able to distinguish between silver core-gold shell and hollow alloy shell morphologies. For the nanostructures in Sample 7 that have a thickness of 14.5 nm with a central core/cavity of 5 nm thickness, the presence of a silver core would mean that the intensity of scattered electrons from the interior would be approximately 80 % of the signal at the edge. On the other hand, a cavity would result in the signal from the interior reducing to approximately 60 % of the value at the edge.

HAADF-STEM analysis has been carried on nanoparticles from Sample 7 to determine their internal structure. In Fig. 9, a HAADF-STEM image of nanoparticles from Sample 7 is shown. The bridging material that was discussed earlier is clearly visible in some of the nanoparticles. Line profiles of the HAADF intensity across two of these nanostructures have been measured as indicated by the gray lines in the image. The relative intensity of the signal along each line profile is plotted in Fig. 10. In the top profile, the relative intensity to the right of "d" and between "b" and "c", corresponding to signal from the edge and the bridging material respectively, is set to 1 for comparison to signal from the hollow or silver core regions of the nanoparticle along the line profile between "a" and "b" and between "c" and "d". In these latter regions the signal is at about 0.5 and 0.6 respectively. Similarly, in the bottom profile, the relative intensity of the signal from the edges of the nanoparticle to the left of "e" and to the right off is set to 1 for comparison to signal coming from the hollow or silver core region between "e" and "f. Again we can see that the signal from the middle is about 0.6. Thus, the high contrast between the darker regions in the centre and the bright edges cannot be explained by assuming the presence of a silver core, but rather implies the existence of a cavity. Example 5 - DDA Calculations

To gain further insight into the structure of the nanoparticles of the present invention, numerical calculations have been carried out for model systems to simulate the extinction spectrum that comes from the LSPR. The Discrete Dipole Approximation (DDA) has proven to be a versatile and highly successful approach. ^[1' ^{2> 33"37]} The DDA calculations were run using the code of Draine and Flatau^[38' ³⁹^ with a 5nm stepsize over 120steps, simulating the extinction spectra from 400-1 OOOnm. In each DDA calculation, the model nanoparticle is comprised of an array of point dipoles. These are spaced 1 nm apart in a simple cubic arrangement in space and are assigned the individual dipoles different dielectric constants in accordance with the model's intended composition. ^[40~43] Since the experimental extinction spectra are for nanoparticles dispersed in water, the following relationship for the dielectric constant of the individual dipoles was used so that the calculations would model materials that would behave optically as if they were in water: ε(ω),

Equation 1 Although triangular nanoparticles were produced, there is some truncation and rounding of the corners of the nanoparticles (see Figs. 2). For this reason, it is possible to choose a disc as the model morphology for Samples 6, 7 and 8. Indeed, for a given measured perpendicular diameter and thickness, the volume of a hexagon or significantly rounded or truncated prism is much closer to the volume of a disc than a triangular prism with the same dimensions. For this reason, the disc model morphology can be seen as a valid and good approximation.

For each sample, both a silver core-gold shell structure and a AuAg alloy hollow (water-filled) shell structure were modelled, giving a total of 6 model nanoparticles. For the core-shell structure it is assumed that the cavities observed in the TEM images are silver cores surrounded by a gold shell. The Au/Ag ratio is then defined by the dimensions of the model and it should be noted that this does not match the Au/Ag ratio in the nanoparticles as determined by EDS. For the hollow nanoshell structure, the Au/Ag ratio is set by the amounts of metallic silver and gold as determined by EDS for each of Samples 6, 7 and 8. Typically, alloys are modelled by a composition-weighted linear combination of the dielectric functions of the metals for each dipole,^[15' ⁴⁴' ^45] Here, a random distribution of silver and gold dipoles were setup, which, for small dipoles is an equivalent approach. The approach does not perfectly represent a random distribution of silver and gold atoms since the volume of space per dipole in the model is larger than the volume of an atom, and so therefore more closely represents a dispersion of silver clusters inside a gold matrix. Although the dipoles could be placed at distances equivalent to atoms in a lattice, the number of dipoles needed for such a calculation would be huge, on the order of 10⁶, and therefore prohibitively time consuming. Nevertheless, the approach can be considered to be a good approximation for an alloy. As an example, the distribution of dipoles used for the models of both the core-shell and alloy shell structures for Sample 6 is shown in Fig. 1 1 .

For each sample the dimensions of model are based on the data obtained from TEM analysis as shown in Table 3 above. The model has an outer diameter of 60nm, a cavity/core diameter of 42 nm and a cavity/core thickness of 5 nm since the variation in these parameters between the samples is not statistically significant. The model is varied for each sample by varying the outer thickness. We have chosen outer thickness values of 13 nm, 14 nm and 16 nm for Samples 6, 7 and 8 respectively. From the data in Table 3, it would seem that the thickness varies linearly between samples and we should choose a thickness of 14.5 nm, midway between 13 nm and 16nm, for Sample 7. This was not done simply because the dipoles in our DDA calculations are 1 nm apart and as such can only choose integer values of thickness for the calculations. A full list of all parameters is given in Table 4 below. Table 4. Parameters for DDA calculations.

Outer Outer Core/cavity Core/cavity Number

% %

Sample diameter thickness diameter thickness of

Silver Gold

(nm) (nm) (nm) (nm) dipoles

Core- shell

6 60 13 42 5 19.0 81.0 36764

7 60 14 42 5 17.6 82.4 39592

8 60 16 42 5 15.4 84.6 45248

Hollow

alloy

6 60 13 42 5 31 .0 69.0 29784

7 60 14 42 5 20.5 79.5 38268

8 60 16 42 5 17.8 82.2 32612

The results of the DDA calculations are shown in Fig. 12; for clarity, only the data points between 500 nm and 900 nm are shown. The calculated extinction spectra of the AuAg alloy shell nanostructures consistently show a very good match to the experimental extinction spectra for Samples 6 and 7, while the simulated extinction spectra of the silver core-gold shell nanostructures are consistently blue-shifted with respect to the experimental extinction spectra. For Sample 8, the experimental extinction spectrum lies half-way between the two calculated spectra so the result here is inconclusive. However, it should be noted that the two calculated spectra are much closer together here than they are for Sample 6 or 7, making it more difficult to get a conclusive match. In Fig. 13, the maxima of the experimental and calculated spectra are plotted and the excellent agreement can be seen between the calculated extinction spectra for the hollow alloy shell nanostructures and the experimental extinction spectra, which provides support that the nanostructures are indeed hollow.

Example 6 - SERS enhancements by unaggregated AuAg nanoboxes

Here, we demonstrate that SERS enhancements equivalent to those attainable with the standard benchmark citrate reduced gold and silver particles can be obtained from unaggregated hollow AuAg nanoboxes. The data also show that there is little correlation between the plasmon absorbance at the Raman excitation wavelength and the enhancement which they provide.

Experimental

AuAg nanoboxes were formed via a glavanic replacement reaction from a silver nanoprism template as described in the experimental methods section above. Silver nanoprisms were produced by a seed based synthesis which catalysed the reduction of Ag⁺ by ascorbic acid.[26] Colloidal samples were labelled 1 .5:0 μηιοΐ Ag:Au - 1 .5:5 μηιοΐ Ag:Au which constituted various stages in the formation of AuAg nanoboxes through the addition of increasing volumes of HAuCU. SERS spectra were recorded on an Avalon Instruments/PerkinElmer RamanStation with an excitation wavelength of 785 nm. The samples were held in 96 well microtiter plates, typically 200 μΐ of colloid was mixed with 50μ1 of analyte solution. For experiments where aggregated particles were examined, 50μ1 of the aggregating agent solution which was 0.1 M MgSC^ was added to this mixture and briefly agitated. Spectra were recorded within a few minutes using 30 s accumulation times and 100 m W laser power.

Uv /Vis spectra were recorded on a HP diode array single beam system (HP 8452) in 1 cm optical path quartz cuvettes. Absorption measurements were made in the range 300- 1 100 nm. The size and surface charge of the nanoparticles were measured using a Zetasizer Nano ZS system (Malvern Instruments) and particles were imaged by TEM using a Philips Tecnai F-20 high resolution transmission electron microscope operated at an accelerating voltage of 200 kV. The samples were stabilised within a Polycarbophil gel before depositing on a 300 mesh Formvar supported copper grid for analysis. Results

Fig. 14 shows the Uv/vis absorption spectra of the unaggregated colloids prepared with different amounts of added Au. An initial red shift and broadening of the LSPR is observed at lower Au addition values, followed by a blue shift and sharping of the surface plasmon band as the quantity of Au added is increased.

SERS experiments were carried out on these unaggregated particles using 2 x 10^"5 mol dm^"3 thiophenoi and adenine as SERS test compounds. This concentration was sufficient to provide monolayer coverage yet did not induce aggregation of the particles as shown by UV/vis absorption measurements. Figs. 15 and 16 show absorption spectra of the 1.5:5 μιηοΐ Ag:Au the unaggregated particles under conditions where they were unaggregated, following addition of SERS test compound and finally following aggregation with MgS0₄. A noticable broadening and shift to longer wavelengths is observed as particle aggregation is induced with MgS0₄ but not with the addition of the test compounds with no MgS0₄. Similar data was obtained for samples 1.5:0.5 μιηοΐ - 1.5:5 μιηοΐ Ag:Au.

To confirm that the SERS analyte was not inducing aggregation, TEM was used to image the particles from the adenine SERS experiments (Fig. 17). In general, for SERS experiments involving nanoparticles it is crucial that the colloid is aggregated in order to create strong electric fields between the particles. However, as the SERS spectra of unaggregated colloidal particles with adenine and thiophenoi shown in Figs. 18 and 19 demonstrate the nanoboxes give strong SERS signals even when no external aggregating agent is added.. These data also show that for nanoparticles with increasing quantities of gold, increased SERS enhancement was obtained. The highest enhancement was obtained for particles from 1 .5 : 5 μτηοΐ Ag:Au which had the highest Au content of all the particles tested.

Even more remarkably, for the nanoboxes, little difference in spectral intensity was observed between samples with the same test analyte before and after the colloid was aggregated with a metal salt (Fig. 20). Typically aggregation with salt gives a very significant increase in SERS intensity, indeed it is normally essential that colloids are aggregated to give detectable signals from low concentration test molecules or analytes. This result is clearly not due to a simple shifting of the particle's plasmon absorption closer to the excitation wavelength since, as shown in Fig. 15, aggregation of sample 1 .5:5μιηο1 Ag:Au shifts the plasmon band to the red, so that the plasmon absorbance at 785 nm is larger for the aggregated particles than for unaggregated colloid. It is widely believed that there is a strong correlation between plasmon absorption at the excitation wavelength and the enhancement obta\ned,[46, 47] however we have previously observed little correlation between the Uv /vis of chaotic Au aggregates and their SERS enhancements. [48]

Similarly, Fig. 14 shows the absorption spectra of the simple unaggregated particles with varying amounts of Au. The particles which provide the greatest enhancement have a plasmon absorption approximately 150 nm away from the excitation laser line. In contrast, sample 1 .5:2 μηιοΐ Ag:Au, which has a surface plasmon which peaks near the 785 nm excitation wavelength gives only ca. an eighth of the enhancement given by 1.5:5 μηιοΐ Ag:Au. The overall trend in plasmon absorption and enhancement are shown in Fig. 21 .

The very large intensity of the unaggregated nanobox signals prompted us to attempt to compare directly the enhancing ability of the AuAg nanoboxes and conventional SERS materials. It is one of the weaknesses of this entire area that researchers who prepare nanoparticles do not attempt to benchmark them against known and widely available existing enhancing substrates. This is a significant problem since it means that it is difficult to assess claims by other workers that they can generate SERS signals from their unaggregated materials [51 ]. Here we have carried out a direct comparison between unaggregated 1.5:5μιηο1 Ag:Au (prepared with 98 mg/dm³ Au ) and:

1 . commercially available monodisperse 40 nm diameter Au colloids (56 mg/dm³ Au), [49] 2. the widely used Lee and Meisel citrate reduced silver particles (prepared with 1 13 mg/dm³ Ag), [50]

It is not rational to make a simple direct comparison using the as-prepared colloids because each has a different mass of metal, number of particles and surface area. Since the signal given by any colloid will to a good approximation double with doubling particle density the absolute values obtained from different colloids are essentially meaningless unless they are normalised in some way to account for particle concentration. The most meangingful comparison is based on the intensity per unit surface area available to provide the enhancement. If different samples of particles made from the same metal are to be compared, scaling them to equal mass is equivalent to scaling for surface area proved the diameters are similar. Fig. 22 shows data for unaggregated AuAg nanoboxes and 40 nm Au nanospheres which have been diluted to the same total concentration by mass (56 mg/dm³ of Au). It is clear that the nanoboxes give an absolute signal level which is at least comparable to the monodisperse Au spheres (the differences in relative band intensities within the two spectra make it impossible to give a single overall value). Comparing the nanoboxes to the aggregated Ag colloid (again diluted to 56 mg/dm³ of Ag) shows that the Ag colloid gives ca. 1 .7 x the signal from the nanoboxes. With these Ag particles the diversity of particle shapes and sizes makes it difficult to estimate the surface area so normalisation by mass is the next best approach. However, even this normalisation does not account for the difference in the relative atomic masses of Au and Ag, (197 and 107) which means that at equal mass we have used 1 .84 x the number of Ag atoms in the colloid compared to the Au nanoboxes.

Immediately after synthesis the particles retain a very weakly bound citrate layer in combination with a number of other possible absorbates eg. ascorbic acid. Washing the particles with weak Na₃Ct solution (10^"3M) not only increases the surface charge of the particles, and hence their stability, but also ensures that the particle surface is covered by simple citrate layer (see Figure 10). This has the advantages firstly that the surface chemistry of such systems is now well understood and secondly that the weakly bound citrate ions may be displaced by stronger binding ions which allows surface modification to be carried out on the particles to increase their affinity towards difficult target analytes. Figs. 23 and 24 show spectra recorded for unaggregated particles before and after addition of citrate and data showing the increase in zeta potential following addition of Na₃Ct. Fig. 25 compares the spectra recorded under equivalent experimental conditions using unaggregated nanoprisms (i.e. solid prisms), aggregated nanoprisms and unaggregated nanoboxes. This is to demonstrate the huge increase in signal which unaggregated nanoboxes give over unaggregated nanoprisms and is evidence that the effect of high surface curvature apices on the particles is not in itself sufficient to allow strong SERS signals to be generated without aggregation. The data in Fig. 22 clearly demonstrate that the nanoboxes can give SERS signals comparable to those obtained with the best currently available colloids but without the need for aggregation. Although this is useful in conventional aqueous solutions the major advantage will be in generating SERS signals in environments where aggregation is much more problematic, such as incorporating the particles into biological materials or within polymeric hosts as sensors. One of the problems with intracellular SERS studies is the need to localise numerous particles so that aggregates can form within cells. These particles have the potential to probe cells and other living tissue without the need for aggregation and may be useful not only for studies of physiological processes but also for diagnosis of disease states.

In general, a SERS enhancing material which gives signals as good as those of conventional colloids but without aggregation has the potential to replace these conventional colloids without loss of signal and with considerable simplification in the experimental protocols required. Although aggregation is relatively straightforward to achieve in solution, it is an inherently random process and does cause significant problems for quantitative rather than qualitative analysis since it may give rise to variations in the absolute signal heights which are not related to the concentration of the test analyte. Moreover, aggregation within solid or semi-solid matrices (including biological tissue and within cellular environments) is also difficult or impossible, so that the ability to carry out SERS enhancement without aggregation is a huge advantage. This means that the present invention has fields of application which include detection of low concentrations of compounds for applications in Homeland security, forensic science, environmental pollution monitoring, food safety, body fluid analysis and general analytical chemistry. In addition, the nanoboxes described herein have the potential to be used in protocols for biomedical screening of disease states in intact tissue and biopsy specimens. The nanoboxes may also be incorporated into porous permeable or swellable host materials (such as organic and inorganic polymers) to act as molecular sensing devices for any of the applications listed above.

In the specification the terms "comprise, comprises, comprised and comprising" or any variation thereof and the terms include, includes, included and including" or any variation thereof are considered to be totally interchangeable and they should all be afforded the widest possible interpretation and vice versa.

The invention is not limited to the embodiments hereinbefore described but may be varied in both construction and detail. References

[1 ] B. J. Wiley, S. H. 1m, Z.-Y. Li, J. McLellan, A. Siekkinen, Y. Xia, J. Phys. Chem. B

2006, 110, 15666.

[2] K. L. Kelly, E. Coronado, L. L. Zhao, G. C. Schatz, J. Phys. Chem. B 2003, 107, 668.

[3] L. M. Liz-Marzan, Langmuir 2006, 22, 32.

[4] E. Hao, G. C. Schatz, J. Chem. Phys. 2004, 120, 357.

[5] C. Noguez, J. Phys. Chem. 2007, 111, 3806.

[6] V. K. Komarala, Y. P. Rakovich, A. L. Bradley, S. J. Byrne, Y. K. Gun'ko, N. Gaponik,

A. Eychmuller, Appl. Phys. Lett. 2006, 89, 2531 18.

[7] Y. Chen, K. Munechika, D. S. Ginger, Nano Lett. 2007, 7, 690.

[8] S. E. J. Bell, N. M. S. Sirimuthu, J. Am. Chem. Soc. 2006, 128, 15580.

[9] Y. Sun, Y. Xia, J. Am. Chem. Soc. 2004, 126, 3892.

[10 N. G. Portney, M. Ozkan, Anal. Bioanal. Chem. 2006, 384, 620.

[1 Y. Sun, B. T. Mayers, Y. Xia, Nano Lett. 2002, 2, 481 .

[12 Y. Sun, B. Mayers, Y. Xia, Adv. Mater. 2003, 75, 641.

[13 J. Chen, J. M. McLellan, A. Siekkinen, Y. Xiong, Z.-Y. Li, Y. Xia, J. Am. Chem. Soc.

2006, 128, 14776.

[14 S. E. Skrabalak, J. Chen, Y. Sun, X. Lu, L. Au, C. M. Cobley, Y. Xia, Acc. Chem. Res.

2008, 41, 1 587.

[15 J. Chen, B. Wiley, J. McLellan, Y. Xiong, Z.-Y. Li, Y. Xia, Nano Lett. 2005, 5, 2058.

[16 C. M. Cobley, D. J. Campbell, Y. Xia, Adv. Mater. 2008, 20, 748.

[17 B. G. Prevo, S. A. Esakoff, A. Mikhailovsky, J. A. Zasadzinski, Small 2008, 4, 1 1 83.

[18 Y. Sun, B. Wiley, Z.-Y. Li, Y. Xia, J. Am. Chem. Soc. 2004, 126, 9399.

[19 X. Lu, L. Au, J. McLellan, Z.-Y. Li, M. Marquez, Y. Xia, Nano Lett. 2007, 7, 1764.

[20 G. S. Metraux, Y. C. Cao, R. Jin, C. A. Mirkin, Nano Lett. 2003, 3, 519.

[21 Y. Sun, Y. Xia, Adv. Mater. 2003, 15, 695.

[22 L.-P. Jiang, S. Xu, J.-M. Zhu, J.-R. Zhang, J.-J. Zhu, H.-Y. Chen, Inorg. Chem 2004, 43, 5877.

[23 R. G. Sanedrin, D. G. Georganopoulou, S. Park, C. A. Mirkin, Adv. Mater. 2005, 17, 1027.

[24 B. Rodriguez-Gonzalez, A. Burrows, M. Watanabe, C. J. Kiely, L. M. Liz-Marzan, J.

Mater. Chem. 2005, 75, 1755.

[25 D. Aherne, D. E. Charles, M. E. Brennan-Fournet, J. M. Kelly, Y. K. Gun'ko, Langmuir

2009, 25, 10165. [26] D. Aherne, D. M. Ledwith, M. Gara, J. M. Kelly, Adv.. Fund. Mater. 2008, 18, 2005.

[27] Y. Yin, C. Erdonmez, S. Aloni, A. P.Alivisatos, J. Am. Chem. Soc. 2006,728, 12671 .

[28] Q. Zhang, J. Xie, J. Y. Lee, J. Zhang, C. Boothroyd, Small 2008, 4, 1067.

[29] T. Shibata, B. A. Bunker, Z. Zhang, D. Meisel, C. F. Vardeman 11, J. D. Gezelter, J. Am.

Chem. Soc. 2002, 124, 1 1989.

[30] C. Wang, S. Peng, R. Chan, S. Sun, Small 2009, 5, 567.

[31 ] Z. Y. Li, J. Yuan, Y. Chen, R. E. Palmer, J. P. Wilcoxon, Appl. Phys. Lett. 2005, 87, 243103.

[32] S. J. Pennycook, Ultramicroscopy 1989, 30, 58.

[33] E. M. Purcell, C. R. Pennypacker, Astrophys. J. 1973, 186, 705.

[34] K. L. Shuford, M. A. Ratner, G. C. Schatz, J. Chem. Phys. 2005, 123, 1 14713.

[35] K. L. Kelly, A. A. Lazarides, G. C. Schatz, Comput. Sci. Eng. 2001, 3, 67.

[36] I. Pastoriza-Santos, L. M. Liz-Marzan, J. Mater. Chem. 2008, 75, 1724.

[37] I. O. Sosa, C. Noguez, R. G. Barrera, J. Phys. Chem. B 2003, 707, 6269.

[38] B. T. Draine, P. J. Flatau, J. Opt. Soc. Am. 1994, 77.

[39] B. T. Draine, P. J. Flatau, "User Guide to the Discrete Dipole Approximation Code DDSCAT 7.0"http://arxiv.org/abs/0809.0337v5 (2008).

[40] M. L. Theye, Phys. Rev. B 1970, 2, 3060.

[41 ] P. Winsemius, F. F. van Kampen, H. P. Lengkeek, C. G. van Went, J. Phys. F 1976, 6, 1583.

[42] E. D. Palik, Ed. Handbook of Optical Constants of Solids, Academic Press, Inc., London

1985.

[43] G. Levique, C. G. Olson, D. W. Lynch, Phys. Rev. B 1983, 24, 4653.

[44] K.-S. Lee, M. A. El-Sayed, J. Phys. Chem. B 2006, 770, 19220.

[45] H.-M. Bok, K. L. Shuford, S. Kim, S.-K. Kim, S. Park, Langmuir 2009, 25, 5266.

[46] A. M. Schwartzberg, T. Y. Olson, C. E. Talley, J. Z. Zhang, Journal of Physical Chemistry B 2006, 770, 19935.

[47] A. M. Michaels, J. Jiang, L. Brus, Journal of Physical Chemistry B 2000, 104, 1 1965.

[48] S. E. J. Bell, M. R. McCourt, Physical Chemistry Chemical Physics 2009, 77, 7455.

[49] www.britishbiocell.co.uk.

[50] P. C. Lee, D. Meisel, Journal of Physical Chemistry 1982, 86, 3391.

[51] M. Gellner, B. Kustner, Schlucker, Vibrational Spectroscopy 2009, 50(1 ), 43-47.

Claims

1 . A substrate for Surface Enhanced Raman Spectroscopy (SERS) comprising an enclosed gold-silver hollow nanostructure.

2. A substrate as claimed in claim 1 wherein the nanostructure is triangular in shape.

3. A substrate as claimed in claim 1 wherein the nanostructure is disc-like in shape.

4. A substrate as claimed in any one of the preceding claims wherein the nanostructure is formed from a silver nanoprism template.

5. A substrate as claimed in any one of the preceding claims wherein the nanostructure comprises a gold-silver alloy.

6. A substrate as claimed in any one of the preceding claims wherein the nanostructure has an aspect ratio of greater than 1 : 1

7. A substrate as claimed in any one of the preceding claims wherein the nanostructure has an aspect ratio of between about 1.3: 1 and about 100: 1 .

8. A substrate as claimed in any one of the preceding claims wherein the nanostructure has an aspect ratio of between about 1 .3: 1 and about 50: 1 .

9. A substrate as claimed in any one of the preceding claims wherein the nanostructure has an aspect ratio of between about 1.3: 1 and about 20: 1.

10. A substrate as claimed in any one of the preceding claims wherein the nanostructure has an aspect ratio of between about 1.3: 1 and about 10: 1 .

1 1 . A substrate as claimed in any one of the preceding claims wherein the nanostructure has an aspect ratio of between about 1.3: 1 and about 5: 1.

12. A substrate as claimed in any one of the preceding claims wherein the nanostructure has an average perpendicular outer diameter of between about 10 nm and about 500 nm.

13. A substrate as claimed in any one of the preceding claims wherein the nanostructure has an average perpendicular outer diameter of between about 20 nm and about 250 nm.

14. A substrate as claimed in any one of the preceding claims wherein the nanostructure has an average perpendicular outer diameter of between about 25 nm and about 150 nm.

15. A substrate as claimed in any one of the preceding claims wherein the nanostructure has an average perpendicular outer diameter of between about 30 nm and about 100 nm.

16. A substrate as claimed in any one of the preceding claims wherein the nanostructure has an average perpendicular outer diameter of between about 40 nm and about 80 nm.

17. A substrate as claimed in any one of the preceding claims wherein the nanostructure has an average perpendicular outer diameter of between about 10 nm and about 500 nm.

18. A substrate as claimed in any one of the preceding claims wherein the hollow has an average perpendicular diameter of between about 21 nm and about 249 nm.

19. A substrate as claimed in any one of the preceding claims wherein the hollow has an . average perpendicular diameter of between about 26 nm and about 149 nm.

20. A substrate as claimed in any one of the preceding claims wherein the hollow has an average perpendicular diameter of between about 31 nm and about 99 nm.

21. A substrate as claimed in any one of the preceding claims wherein the hollow has an average perpendicular diameter of between about 41 nm and about 79 nm.

22. A substrate as claimed in any one of the preceding claims wherein the hollow has an average thickness of between about 0.5 nm and about 50 nm.

23. A substrate as claimed in any one of the preceding claims wherein the hollow has an average thickness of between about 1 nm and about 30 nm.

24. A substrate as claimed in any one of the preceding claims wherein the hollow has an average thickness of between about 1.5 nm and about 15 nm.

25. A substrate as claimed in any one of the preceding claims wherein the hollow has an average thickness of between about 2nm and about l Onm.

26. A substrate as claimed in any one of the preceding claims wherein the nanostructure has an in-plane dipole Localised Surface Plasmon Resonance (LSPR) spectral position between about 390nm and 2000nm.

27. A substrate as claimed in any one of the preceding claims wherein the nanostructure has an in-plane dipole Localised Surface Plasmon Resonance (LSPR) spectral position between about 400nm and 1500nm.

28. A substrate as claimed in any one of the preceding claims wherein the nanostructure has an in-plane dipole Localised Surface Plasmon Resonance (LSPR) spectral position between about 420nm and 1 l OOnm.

29. A substrate as claimed in any one of the preceding claims wherein the nanostructure has an in-plane dipole Localised Surface Plasmon Resonance (LSPR) spectral position between about 450nm and 950nm.

30. A substrate as claimed in any one of the preceding claims wherein the nanostructure has an in-plane dipole Localised Surface Plasmon Resonance (LSPR) spectral position between about 500nm and 850nm.

31. A method for probing a sample by SERS comprising the steps of:

(a) adding a SERS substrate as claimed in any one claims 1 to 30 to a test sample; and

32. A method as claimed in claim 31 wherein the SERS substrate is substantially non- aggregated.

33. A method as claimed in claim 31 or 32 wherein the sample is selected from one or more of a biological sample, a chemical sample, a food sample, or an environmental sample.

34. A method as claimed in claim 33 wherein the biological sample is a cellular sample or a body fluid sample.

35. Use of a substrate as claimed in any one of claims 1 to 30 for biomedical screening of disease states.

36. A molecular sensing device comprising a substrate as claimed in any one of claims 1 to 30.

37. Use of a molecular sensing device as claimed on claim 36 for the detection of low

concentrations of compounds in a sample.

38. Use as claimed in claim 37 wherein the sample is selected from one or more of a

biological sample, a chemical sample, a food sample, or an environmental sample.

39. Use as claimed in claim 38 wherein the biological sample is a cellular sample or a body fluid sample.

40. An enclosed triangular gold-silver hollow nanostructure.

41. A nanostructure as claimed in claim 40 formed from a silver nanoprism template.

42. A nanostructure as claimed in claim 40 or 41 comprising a gold-silver alloy.

43. A nanostructure as claimed in any one of claims 40 to 42 wherein the nanostructure has an aspect ratio of greater than 1 : 1

44. A nanostructure as claimed in any one of claims 40 to 43 wherein the nanostructure has an aspect ratio of between about 1.3: 1 and about 100: 1.

45. A nanostructure as claimed in any one of claims 40 to 44 wherein the nanostructure has an aspect ratio of between about 1 .3: 1 and about 50: 1 .

46. A nanostructure as claimed in any one of claims 40 to 45 wherein the nanostructure has an aspect ratio of between about 1.3: 1 and about 20: 1.

47. A nanostructure as claimed in any one of claims 40 to 46 wherein the nanostructure has an aspect ratio of between about 1.3: 1 and about 10: 1 .

48. A nanostructure as claimed in any one of claims 40 to 47 wherein the nanostructure has an aspect ratio of between about 1.3: 1 and about 5: 1.

49. A nanostructure as claimed in any one of claims 40 to 48 wherein the nanostructure has an average perpendicular outer diameter of between about 10 nm and about 500 nm.

50. A nanostructure as claimed in any one of claims 40 to 49 wherein the nanostructure has an average perpendicular outer diameter of between about 20 nm and about 250 nm.

51. A nanostructure as claimed in any one of claims 40 to 50 wherein the nanostructure has an average perpendicular outer diameter of between about 25 nm and about 150 nm.

52. A nanostructure as claimed in any one of claims 40 to 51 wherein the nanostructure has an average perpendicular outer diameter of between about 30 nm and about 100 nm.

53. A nanostructure as claimed in any one of claims 40 to 52 wherein the nanostructure has an average perpendicular outer diameter of between about 40 nm and about 80 nm.

54. A nanostructure as claimed in any one of claims 40 to 53 wherein the nanostructure has an average perpendicular outer diameter of between about 10 nm and about 500 nm.

55. A nanostructure as claimed in any one of claims 40 to 54 wherein the hollow has an average perpendicular diameter of between about 21 nm and about 249 nm.

56. A nanostructure as claimed in any one of claims 40 to 55 wherein the hollow has an average perpendicular diameter of between about 26 nm and about 149 nm.

57. A nanostructure as claimed in any one of claims 40 to 56 wherein the hollow has an average perpendicular diameter of between about 31 nm and about 99 nm.

58. A nanostructure as claimed in any one of claims 40 to 57 wherein the hollow has an average perpendicular diameter of between about 41 nm and about 79 nm.

59. A nanostructure as claimed in any one of claims 40 to 58 wherein the hollow has an average thickness of between about 0.5 nm and about 50 nm.

60. A nanostructure as claimed in any one of claims 40 to 59 wherein the hollow has an average thickness of between about 1 nm and about 30 nm.

61. A nanostructure as claimed in any one of claims 40 to 60 wherein the hollow has an average thickness of between about 1 .5 nm and about 15 nm.

62. A nanostructure as claimed in any one of claims 40 to 61 wherein the hollow has an average thickness of between about 2nm and about l Onm.

63. A nanostructure as claimed in any one of claims 40 to 62 wherein the nanostructure has an in-plane dipole Localised Surface Plasmon Resonance (LSPR) spectral position between about 390nm and 2000nm.

64. A nanostructure as claimed in any one of claims 40 to 63 wherein the nanostructure has an in-plane dipole Localised Surface Plasmon Resonance (LSPR) spectral position between about 400nm and 1500nm.

65. A nanostructure as claimed in any one of claims 40 to 64 wherein the nanostructure has an in-plane dipole Localised Surface Plasmon Resonance (LSPR) spectral position between about 420nm and 1 l OOnm.

66. A nanostructure as claimed in any one of claims 40 to 65 wherein the nanostructure has an in-plane dipole Localised Surface Plasmon Resonance (LSPR) spectral position between about 450nm and 950nm.

67. A nanostructure as claimed in any one of claims 40 to 66 wherein the nanostructure has an in-plane dipole Localised Surface Plasmon Resonance (LSPR) spectral position between about 500nm and 850nm.

68. Use of a nanostructure as claimed in any one of claims 40 to 67 for the encapsulation of a material.

69. Use as claimed in claim 68 wherein the material is selected from one or more of a

biologically active compound such as a drug, a contrast agent such as a radioisotope, a magnetic material, or a dye such as a fluorophore.

70. Use of a nanostructure as claimed in any one of claims 40 to 67 as a substrate for surface enhanced Raman spectroscopy (SERS).