• Open Access

Importance of Nanoparticle Size in Colorimetric and SERS-Based Multimodal Trace Detection of Ni(II) Ions with Functional Gold Nanoparticles

Authors

  • Željka Krpetić,

    1. Centre for Molecular Nanometrology, WestCHEM, Pure and Applied Chemistry, University of Strathclyde, 295 Cathedral Street, Glasgow, G1 1XL, UK
    Search for more papers by this author
  • Luca Guerrini,

    1. Centre for Molecular Nanometrology, WestCHEM, Pure and Applied Chemistry, University of Strathclyde, 295 Cathedral Street, Glasgow, G1 1XL, UK
    Search for more papers by this author
  • Iain A. Larmour,

    1. Centre for Molecular Nanometrology, WestCHEM, Pure and Applied Chemistry, University of Strathclyde, 295 Cathedral Street, Glasgow, G1 1XL, UK
    Search for more papers by this author
  • John Reglinski,

    1. WestCHEM, Pure and Applied Chemistry, University of Strathclyde, 295 Cathedral Street, Glasgow, G1 1XL, UK
    Search for more papers by this author
  • Karen Faulds,

    1. Centre for Molecular Nanometrology, WestCHEM, Pure and Applied Chemistry, University of Strathclyde, 295 Cathedral Street, Glasgow, G1 1XL, UK
    Search for more papers by this author
  • Duncan Graham

    Corresponding author
    1. Centre for Molecular Nanometrology, WestCHEM, Pure and Applied Chemistry, University of Strathclyde, 295 Cathedral Street, Glasgow, G1 1XL, UK
    • Centre for Molecular Nanometrology, WestCHEM, Pure and Applied Chemistry, University of Strathclyde, 295 Cathedral Street, Glasgow, G1 1XL, UK.
    Search for more papers by this author

Abstract

Colorimetric detection of analytes using gold nanoparticles along with surface-enhanced Raman spectroscopy (SERS) are areas of intense research activity since they both offer sensing of very low concentrations of target species. Multimodal detection promotes the simultaneous detection of a sample by a combination of different techniques; consequently, surface chemistry design in the development of multimodal nanosensors is important for rapid and sensitive evaluation of the analytes by diverse analytical methods. Herein it is shown that nanoparticle size plays an important role in the design of functional nanoparticles for colorimetric and SERS-based sensing applications, allowing controlled nanoparticle assembly and tunable sensor response. The design and preparation of robust nanoparticle systems and their assembly is reported for trace detection of Ni(II) ions as a model system in an aqueous solution. The combination of covalently attached nitrilotriacetic acid moieties along with the L-carnosine dipeptide on the nanoparticle surface represents a highly sensitive platform for rapid and selective detection of Ni(II) ions. This systematic study demonstrates that significantly lower detection limits can be achieved by finely tuning the assembly of gold nanoparticles of different core sizes. The results clearly demonstrate the feasibility and usefulness of a multimodal approach.

1. Introduction

Colorimetric assays using gold nanoparticles (GNPs) have become a widely used analytical tool for sensing a plethora of analytes1, 2 due to their ability to monitor small changes occurring at the molecular level.3, 4 Their use is attractive owing to their sensitivity, ease of measurement and the simplicity of the sensor construction. The choice of gold nanoparticles in the fabrication of colorimetric sensors is justified by their highly tunable size and shape-dependent optical properties, remarkable stability against oxidation, relative ease of preparation and the possibility of surface modification with virtually any functionality.5–7 This makes gold nanoparticles a preferable choice amongst other metallic nanoparticles for a range of different applications.8, 9

Color changes of gold colloidal dispersions are visible to the naked eye and are directly associated with alterations of surface plasmon resonances (SPR) that may occur upon changes in particle size, surface chemistry or interparticle distance during the formation of aggregation features.10 Moreover, surface plasmon resonance is the rationale for large enhancement in the Raman scattering from molecules that are in close proximity with the metal surfaces. This phenomenon is well known as surface enhanced Raman scattering (SERS).11 Although significant electromagnetic enhancement can originate from the individual nanoparticles in solution, a dramatic increase in SERS intensification, up to several orders of magnitude higher, is achieved when the nanoparticles are positioned in close proximity to each other.12 For SERS applications, silver and gold are mainly used as primary materials of choice, owing to their optical response in the visible region of the electromagnetic (EM) spectrum. In particular, silver particles provide a larger SERS enhancement in a wider spectral range, from near-UV to near-IR, compared to a more modest enhancement from gold nanoparticles. The reason for this is the large absorption of Au(0) below ca. 600 nm, causing strong damping of the SPR in this spectral region.13

When designing metal nanoparticle-based sensors, it is generally required that nanoparticles must retain extreme stability in the surrounding medium, while exclusively detecting the analyte under examination and avoiding any nonspecific interactions that may occur between the particles in solution. An important factor for controlling the nanoparticle aggregation process is their surface chemistry, i.e., ligands that stabilize the nanoparticles may determine the aggregation state as well as the rate of aggregation in colloidal systems when chemical or physico-chemical interactions between the analyte and stabilizing ligand molecules are triggered.

Surface modification by organic ligands provides stability against aggregation as well as introducing a variety of surface functionalities available for further reactions.14–17 Examples of colorimetric biosensing using GNPs that have been reported include distance dependent detection of DNA,1 reversible aggregation of nanoparticles induced by antibody–antigen,18 aptamer–analyte19 and protein–protein interactions.20 Gold particles were also applied as tools for detection of extremely small amounts of analytes of wider interest such as pollutants,21 toxins,22 and small molecules.23 Similarly, gold nanoparticles have been applied in SERS based detection of enzymes,24 DNA25 and antigens such as prostate specific antigen.26 Detection of heavy metal ions by gold nanoparticles has attracted particular attention and subsequently sensors for Hg(II), Pb(II) and Cd(II) have been developed.27, 28 Sensing of nonactive Raman species, such as metal ions represent a particular challenge for SERS detection.29 In this regard, two general detection strategies have been exploited: i) monitoring changes in intensity of SERS signals from Raman reporters in the presence of the target ions,30, 31 and ii) monitoring changes in the SERS spectrum of chemoselective receptors at the metal surface as a result of their interaction with metal ions.32–35

Here we report a detailed study based on the construction of gold nanoparticle platforms for both colorimetric and SERS sensing purposes. For this study we have chosen Ni(II) ions as a target analyte that represents a challenging choice both because it is a “spectroscopically silent” species36 and because its fast and accurate trace detection is important for patients suffering from the acute dermatitis and skin allergies that Ni(II) often causes.37 To our knowledge, there is currently no extensive study on Ni(II) sensing using gold nanoparticles although Li et al. have reported colorimetric detection of nickel ions using silver nanoparticles.38

Biocompatibility of nickel is determined by its ability to readily form complexes with biomolecules such as amino acids, peptides, proteins, phosphates and nucleic acids.39 However, excess nickel is often related to skin inflammation, as it induces a delayed-type hypersensitivity cellular response when bound to proteins. A specific receptor for nickel in human cell lines was recently discovered by Schmidt et al. explaining how the activation of Troll-like receptor 4 (TRL4) that leads to contact hypersensitivity (CHS) is specifically triggered by Ni(II) ions.40 Interestingly, calcium carbonate and calcium phosphate nanoparticles were recently found to reduce the skin irritation caused by nickel ions.41

In order to construct a simple and inexpensive GNP-based system for detection of Ni(II) ions in aqueous solution, we have exploited the Ni–nitrilotriacetic acid (NTA) chemistry for the design of nanoparticles, significantly simplifying previously used detection methods.42, 43 Interaction between Ni(II) coordinated to NTA and proteins with six consecutive Histidine residues (His tag) has been widely utilized for protein purification,44 and also employed in the construction of gold nanoparticles involved in protein interactions.14, 45, 46 As a source of Histidine, we have employed naturally occurring dipeptide L-carnosine.

The nanoparticle sensing system is designed in such a way that when NTA and L-carnosine functionalized gold nanoparticles are mixed together there is no unspecific interaction between the particles, i.e., in the absence of metal ions that cause formation of metal complexes, both types of functionalized nanoparticles remain stable and aggregation-free in aqueous solution as there are no unspecific reactions between the surface functional groups. However, when Ni(II) ions are added to the nanoparticle mixture, rapid aggregation is induced as Ni(II) is coordinated between the nanoparticles establishing a small interparticle gap. Aggregation of the nanoparticles results in a change of the extinction profile along with an increase in SERS intensity of the Raman reporter. This is similar in principle to the DNA-driven aggregation of oligonucleotide functionalized nanoparticles containing a Raman reporter that turns on the SERS signal.47

Our system for Ni(II) detection is designed in such a way to expressly fit the requirements of multimodal colorimetric and SERS sensing. In particular, we focus our studies on the effect of the nanoparticle size on both colorimetric and SERS sensing performance. Here we show how this can be utilized to create a selective and highly sensitive nickel ion nanosensor. The purpose of this research is to provide robust information on the strategies for the design of nanosensors for multimodal detection. This is shown initially using Ni(II) ions as the model system, that may be readily translated to a wider range of target analytes.

2. Results and Discussion

Citrate stabilized gold nanoparticles of ca. 14 nm in diameter were prepared by a modified Turkevich/Frens method48, 49 and gold nanoparticles ca. 45 nm in diameter were prepared in a second step via a seeded growth method.50 Subsequently nanoparticles were further stabilized and functionalized by thioctic acid.51 Having a carboxy functional group readily available on the surface, allowed both nitrilotriacetic acid (Nα,Nα-bis(carboxymethyl)-L-lysine hydrate) and L-carnosine (β-Ala-His) to be covalently attached to the nanoparticles via EDC/Sulfo-NHS coupling chemistry (EDC = 1-ethyl-3-[3-dimethylaminopropyl]carbodiimide hydrochloride; Sulfo-NHS = N-hydoxysulfosuccinimide]).52 For SERS sensing purposes, we have included in the nanoparticle ligand shell a NIR active reporter molecule (NIR-797 isothiocyanate). This Raman label provides an additional resonant Raman enhancement while operating at 785 nm laser excitation allowing intense SERS signals to be obtained even at low concentrations of dye molecules (ca. 2% surface coverage). This approach avoids destabilization of the particle ligand shell that may cause further aggregation. It is highly important to avoid any nonspecific particle aggregation, as it would result in an increased SERS signal background leading to poorer discrimination when detecting the target analytes. A scheme of the nanoparticle surface chemistry is reported in Figure 1. No aggregation was detected due to the presence of the dye molecule in the ligand shell (see UV–vis spectra in the Supporting Information (SI), S1). Moreover, combining the Au–NTA and Au–L-carnosine nanoparticles in solution did not result in any nonspecific interaction between the particles. Addition of Ni(II) ions caused rapid aggregation of the nanoparticle system, since Ni(II) quickly forms an octahedral complex coordinating between the NTA moiety and the closely packed histidines present on the nanoparticles (Figure 1a). By using Discovery Studio (v.2.5) software we have estimated an approximate 3 nm interparticle distance upon binding of Ni(II). The aggregation of the particles was monitored by following the changes in the extinction profile colorimetrically (SI, Figure S15) and by monitoring the increase in SERS intensity of the Raman reporter signals, as illustrated in Figure 1c.

Figure 1.

a) Aggregation of the nanoparticles is caused by formation of the octahedral complex when Ni(II) is coordinated between NTA and histidine moieties present at the particle surface. b) Schematic representation of Au–NTA@NIR reporter and Au–L-carnosine nanoparticles. c) SERS signal at 785 nm laser excitation (strongest peaks at 523 and 555 cm−1) of the Raman reporter NIR-797 before (bottom arrow) and after (top arrow) aggregation of the nanoparticles.

Concentration of the nanoparticles for Ni(II) detection was optimized in order to achieve the highest sensing performance (See SI for details). To monitor the aggregation process, aliquots of 5 mM aqueous NiCl2 were added to 40 pM gold colloidal dispersions. Upon addition of increasing concentration of Ni(II) the color of the colloidal dispersion gradually changed from raspberry red to purple. Extinction spectra from every combination of nanoparticle core sizes were recorded after 3 min of incubation with Ni(II) and are reported in Figure 2. The effect of nanoparticle size on the sensor response was investigated by following the absorbance at 700 nm and plotting the ratio as Abs700/AbsSPR, as this ratio is sensitive to the concentration of Ni(II). To rule out nonspecific salt-induced aggregation of the particles we have verified the stability of the particles in the presence of NaCl(aq) and no aggregation was observed comparing the same concentrations (SI, S4). Aggregation features at different Ni(II) concentrations were visualized by SEM (Figure 3). Only at a reasonably high concentration of Ni(II) (25 ppm) was a colorimetric change observed when both Au–NTA and Au–L-carnosine of the smallest particle size were combined (14 nm GNPs). An improved colorimetric response (sensitivity) was detected when 14 nm sized particles were combined with larger particles (45 nm in diameter) (Figure 2b), and even further by combining both types of larger nanoparticles (45 nm) (Figure 2c).

Figure 2.

Aggregation of the particles upon addition of Ni(II) in 0–70 ppm concentration range monitored by UV–vis spectroscopy showing spectra of Au–NTA and Au–L-carnosone nanoparticles of a) both 14 nm, b) 14 and 45 nm, c) both 45 nm.

Figure 3.

SEM micrographs showing the nanoparticle aggregation patterns at increasing Ni(II) concentrations. a) At low concentrations of Ni(II) formation of symmetric ‘flower-like’ features is observed while assembling the larger nanoparticles, whereas d) asymmetric ‘satellite-like’ aggregation features are formed assembling nanoparticles of two different core sizes. As the Ni(II) concentration increases significantly, larger and more complex aggregates are formed both for b,c) 45–45 nm systems and e,f) 45–14 nm systems. Scale bars are 200 nm.

Similarly, the SERS intensity increase of the two most intense bands, located at 523 and 555 cm−1, tentatively assigned to skeletal deformations of the benzo[e]indole macrocycles,53 were monitored and reported in Figure 4. When 14 nm sized particles were combined in the presence of Ni(II), no significant SERS increase was observed even for relatively high Ni(II) concentrations (50 ppm) contrary to the colorimetric system. For all other particle combinations examined, the SERS intensity rapidly increased to a maximum intensity followed by a plateau and then a slow decrease reaching a constant value (Boltzmann sigmoid). A similar trend was also observed by Braun et al.54 in controlled nanoparticle assembly via bifunctional molecular linkers and can be explained by initial formation of highly active SERS enhancers (dimers, trimers, small clusters) to the detriment of more subtle enhancing nanoparticles (monomers). As the aggregation of the system proceeds, at higher concentrations of Ni(II), larger clusters are formed. This results in the formation of a higher number of hot spots per cluster, which however does not improve the overall SERS enhancement55 leading to the decrease of cluster density in the colloidal volume interrogated by the laser. Moreover, by forming even larger aggregates (over 40–50 ppm see SEM images in the SI, S5) the overall number of particles in solution reduces due to precipitation and as a result we observed a decrease in SERS signal. Interestingly, depending on the position of the Raman reporter in an asymmetric system, i.e., whether the reporter molecule is attached to a 14 or 45 nm sized Au–NTA nanoparticle, absolute SERS intensity changes reached significantly higher values when the dye molecule was placed on the larger particle. As the average number of Raman reporters per surface area was kept constant in our experimental conditions, it was required to combine 45 nm sized particles with ones of 14 nm diameter in a 1:10 molar ratio.56 This kind of nanoparticle assembly resulted in formation of a satellite-type configuration of the aggregates57 (Figure 3df). For this configuration, it is expected that a larger number of Raman reporters present in the interparticle junctions would profit from the local SERS enhancement when located on the larger particles (SI, S6).

Figure 4.

SERS intensity (left) and colorimetric response (right) as a function of the Ni(II) concentration. SERS intensity is reported as a sum of band peak heights at 523 and 555 cm−1 while colorimetric response is reported as Abs700/AbsSPR ratio. Graphical representation of the particle assembly; dye labeled particles are represented with stars (bottom).

Colorimetric detection of Ni(II) was measured using the same concentration range as for the SERS measurements (Figure 4). As for SERS, 45 nm–45 nm particle systems offered better performance with respect to other investigated nanoparticle systems. However, in our experimental conditions, we have found that SERS analysis offers higher sensitivity, i.e., lower limit of detection (0.5 vs 5 ppm; see SI, S7) while colorimetric sensing offers a wider range of analytical response (from 5 to >80 ppm Ni(II)). The analytical range monitored by SERS can be extended by a change in experimental conditions, specifically, nanoparticle concentration.

In particular aggregation of the larger particles (45 nm) was monitored at different concentrations of Ni(II) by SEM. At 5 ppm Ni(II), populations of nanoparticle dimers were formed, albeit nanoparticles are prevalently present as monomers. A progressive increase in Ni(II) concentration yielded trimers, tetramers to larger clusters (1 μm) as shown for higher Ni(II) concentrations (50–60 ppm). A gallery of representative images is shown in the SI.

From these findings we suggest that using nanoparticles of 45 nm diameter for nanoparticle-sensor construction offers optimal performance for sensing of the analytes both colorimetrically and by SERS. Here, we have shown this in detail using Ni(II) ions, although a similar affinity and sensor response was detected with the exception of Cu(II) (see SI). Our experimental findings can be compared to theoretical calculations reported for dimeric structures, which are often used as fairly good models for analysis of complex experimental data.58 When two nanoparticles are brought into close proximity, new plasmon resonances appear. Monitoring of the most red-shifted resonance (gap-plasmon resonance (GPR)) is crucial for sensing applications.58 The gap-plasmon peak position depends on many factors such as the metal optical properties, interparticle distance, and particle size. For a dimer with fixed interparticle spacing (ca. 3 nm interparticle gap for our system), the larger the nanoparticle size the more red-shifted the GPR. As a consequence, there will be a specific size threshold value above which the coupling of the two nanoparticles result in GPRs located above 600 nm, as shown by theoretical calculations.58, 59

Two important implications in the case of the larger nanoparticles arise: i) the spectral overlap between the intense plasmons from individual nanoparticles and the weaker gap-plasmon from dimers is drastically reduced. This is of great importance for assembly of colorimetric sensors as this enables an increase in absorbance values at 700 nm resulting in a significant color change (from red to blue) and thus relating the color change directly with the analyte concentration; ii) gap-plasmon resonance is tuned above the critical optical threshold (ca. 600 nm) where gold approaches the SERS-enhancing performance of silver. To further support this claim we have prepared core/shell Au@Ag nanoparticles (45 nm) using a seeded growth method and subsequently functionalized them in the same fashion as already described for gold. These core/shell nanoparticles present ‘silver-like’ optical properties, i.e., the plasmon resonance of 14 nm gold cores is completely damped (Figure S11, SI) with particle size distribution and concentration analogous to gold particles of 45 nm diameter. This allowed us to investigate colorimetric changes along with SERS sensing performance related solely to different optical properties of the two metals. The colorimetric response using Au@Ag functionalized particles exhibited far less sensitivity to Ni(II) as compared to gold nanoparticles of 45 nm diameter (SI, S8). In Figure 5, SERS intensities of Au and Au@Ag nanoparticles (45 nm) are compared in the 0–14 ppm Ni(II) concentration range. Similar sensitivity is observed for both systems with comparable values for the limit of detection (LOD) of 0.5 ppm (S/N ratio = 3). As foreseen, higher absolute SERS intensity values were found in the case of silver nanoparticles, albeit the enhancing abilities of gold particles compared to those of silver greatly increase with advancement of the aggregation state. This is graphically represented in Figure 5b where the ratio between SERS intensities resulting from gold and silver nanoparticles in the same range of Ni(II) concentration is reported. In Figure 5c maximum SERS enhancement factors of monomeric silver and gold nanoparticles (50 nm) along with their corresponding dimers (3 nm gap) are schematically represented in accordance with theoretical calculations13, 60 supporting our findings which clearly show a much larger increase in SERS enhancement in the case of gold nanoparticles (ΔEFAu) compared to that of silver (ΔEFAg) with particle aggregation. These findings clearly demonstrate that gold nanoparticles of 45 nm diameter represent a valuable substitute for silver using 785 nm laser excitation, whenever the Raman reporter signal increase results from analyte-driven aggregation of nanoparticles. In this way gold particles provide SERS sensitivity comparable to silver, while retaining the advantageous intrinsic properties of gold.

Figure 5.

a) Comparison of SERS intensities obtained by Au and Au@Ag nanoparticles (45 nm) and b) corresponding SERS intensity ratio between gold and silver NP system in 0–14 ppm Ni(II) concentration range for 785 nm laser excitation wavelength. c) Schematic representation of the increase in maximum SERS enhancement factor (EF) upon coupling of monodisperse Au and Ag nanoparticles of 50 nm diameter resulting in formation of dimers with 3 nm interparticle gaps (in accordance with theoretical calculations reported by Etchegoin and co-workers).13, 60

Furthermore, maintaining the experimental conditions the selectivity of the sensing system in the presence of different metal ions (Co(II), Fe(III), Zn(II) and Cu(II)) was tested and it was shown that Ni(II) ions are preferentially bound to our nanoparticles, while other ions investigated cause very weak or no aggregation, with the anticipated exception of Cu(II) (SI, S9).

Careful tuning of the nanoparticle size offers another important advantage, i.e., asymmetric assembly of the particles can be easily exploited in order to achieve detection limits in the ppb concentration range by employing size selective centrifugation. Asymmetric aggregation features (obtained by interacting 45 nm sized particles and 14 nm sized Raman dye-labeled ones) were separated from the particle mixture by size selective centrifugation (centrifuging the particles at low speed; 4000 rpm) leaving smaller particles unbound in the supernatant. Pellets were then re-dispersed in milli-Q water, concentrating the sample by approximately 8 times, and the SERS intensity was subsequently measured. In this way the SERS signals are related exclusively to dye-labeled particles ergo directly proportional to the amount of Ni(II) detected. The LOD (S/N ratio = 3) was found to be 30 ppb (see Figure S14 in SI) measured by this method.

3. Conclusion

We have developed a multimodal selective nanosensing platform equally suitable for colorimetric and SERS trace detection of Ni(II) based on controlled aggregation occurring as a consequence of Ni(II) coordination to specific sites present on the nanoparticles surface. Symmetric and asymmetric assembly of particles was employed using two different particle core sizes. It was demonstrated that core size of the rationally designed individual particles for colorimetric and SERS-based applications is important for reproducibility of the nanoparticle assembly and tunable sensor response. We believe that our findings are of particular interest to the scientific community working with nanoparticle systems for both SERS and colorimetric detection applications61 as they provide useful tools to maximize sensing performance of gold nanosensors.

4. Experimental Section

Chemicals: All the reagents (of highest available grade) were purchased from Sigma Aldrich and used as received.

Nanoparticle Preparation: AuNPs (14 nm): Gold nanoparticles of 13.8 ± 1.7 nm diameter with a narrow size distribution were prepared following the Turkevich-Frens preparation method.48, 49 Briefly, 38.1 μmol (15 mg) of HAuCl4 trihydrate were dissolved in 150 mL of milli-Q water and heated to boiling. Subsequently, 4.5 mL of the 1% w/w aqueous trisodium citrate solution, previously warmed to ca. 70 °C was added, and the mixture was refluxed for 30 minutes. The solution was then allowed to cool to room temperature under vigorous stirring. The ruby red sol was characterized by UV–vis spectroscopy observing the typical plasmon band at 520 nm and the size distribution determined by SEM.

Nanoparticle Preparation: AuNPs (45 nm): Gold nanoparticles of 44.6 ± 4.7 nm diameter were prepared following the two-step nanoparticle seeded growth preparation method.50 Briefly, 37 μmol of HAuCl4 trihydrate were added to 125 mL of milli-Q water and heated to boiling. Subsequently, 5 mL of gold colloidal dispersion (14 nm sized nanoparticles acting as seeds) was added, followed by addition of 1% w/w aqueous trisodium citrate (21.7 μmol, 638 μL). The mixture was refluxed for 30 min under vigorous stirring. To assure the stability of colloidal dispersion, 4.9 mL of 1% w/w aqueous trisodium citrate was then added and further refluxed for 1 h. The raspberry red sol was characterized by UV–vis spectroscopy observing the plasmon band at 534 nm and the size distribution determined by SEM (see SI). All colloidal dispersions were filtered through a 0.45 μm Millipore filter before characterization.

Nanoparticle Preparation: Au@AgNPs (45 nm): Au@Ag core–shell nanoparticles of 45.4 ± 6.1 nm diameter were prepared following the two-step nanoparticle seeded growth method.50 Briefly, 40 μmol of AgNO3 were added to 125 mL of milli-Q water and heated to boiling, in dark conditions. Subsequently, 5 mL of gold colloidal dispersion (14 nm gold seeds) was added, followed by addition of 1% w/w aqueous trisodium citrate (21.7 μmol, 638 μL). The mixture was refluxed for 30 min under vigorous stirring. To assure the stability of colloidal dispersion, 4.9 mL of 1% w/w aqueous trisodium citrate was then added and further refluxed for 1 h. The nanoparticles produced were then filtered through a 0.45 μm Millipore filter and further centrifuged (1000 rpm, 10 min) to remove any larger aggregates before further conjugation with thioctic acid.51

Surface Modification of GNPs With Functional Ligands: Au–TA Nanoparticles: In a typical reaction, to the citrate stabilized gold nanoparticle dispersion (10 mL) methanolic solution of thioctic acid (4 μL, 40 mM) was added and left stirring gently for 2 h protected from light. The concentration of the thioctic acid was adjusted to 20 500 ligand molecules per nanoparticle of 14 nm diameter and 200 000 for 45 nm diameter GNPs. This excess was deliberately chosen to assure complete coverage of the nanoparticles with ligands. Subsequently particles were purified from the excess ligands by repeated centrifugation (at 13 000 and 4000 rpm, respectively, depending on the particle size) and finally redispersed in milli-Q water. Concentration of the nanoparticles was calculated from the UV–vis spectra and referred according to Haiss et al.56

Surface Modification of GNPs With Functional Ligands: Au–NTA Nanoparticles: The attachment of the NTA functionality was achieved by an EDC/Sulfo-NHS coupling reaction linking the Nα,Nα-bis(carboxymethyl)-L-lysine hydrate to carboxy-functionalized Au–TA nanoparticles.52 In a typical reaction, 6 mM (26.65 μL) aqueous EDC (N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride) and 6 mM (26.65 μL) aqueous Sulfo-NHS (N-hydroxysulfosuccinimide) were added to the Au–TA nanoparticle dispersion and left for 30 min protected from light in order to activate the COOH functionality. Subsequently aqueous NTA (6 mM) was added and left reacting overnight. Particles were purified from the excess ligands by two cycles of centrifugation (at 13 000 and 3500 rpm, respectively, depending on the particle size) and subsequently redispersed in milli-Q water (to a final volume of 1.5 mL).

Surface Modification of GNPs With Functional Ligands: Au–L-Carnosine Nanoparticles: The attachment of L-carnosine (β-Ala-His) was achieved in a similar fashion as described above for Au–NTA nanoparticles. A 6 mM solution of L-carnosine was freshly prepared before use and the solution was kept protected from light during use.

Surface Modification of GNPs With Functional Ligands: Raman Reporter Labeling of the Nanoparticles for SERS Sensor Assembly Measurements: To enable the sensing with gold particles by both techniques at the same time, colorimetric and SERS, additionally NTA-functionalized gold particles were labeled with the Raman reporter molecule (as illustrated in the SI, S2). NIR dye (NIR-797 isothiocyanate = 1,1′-bis(4-sulfobutyl)-11-(4-isothiocyanatophenylthio)-3,3,3′,3′-tetramethyl-10, 12-trimethyleneindotricarbocyanine monosodium salt) was dissolved in ethanol (33.8 μM) and aliquots of 6 and 1.4 μL added respectively to 1 mL Au–NTA nanoparticle dispersions of 14 and 45 nm in diameter (c14nm = 12.4 nM; c45nm = 2.5 nM) and left overnight. Excess dye and unbound ligands were removed by centrifugation and redispersion in milli-Q water. Number of dye molecules added to gold particles of both core sizes per nm2 was kept low (ca. 2% surface coverage) and maintained constant in all cases.

Colorimetric and SERS Assay of Ni(II) Ions: For Ni(II) detection experiments typically the nanoparticle-based sensor was assembled adding 10 μL of each 2.5 nM Au–NTA or Au–L-carnosine particle dispersion to 300 μL of milli-Q water and subsequently aliquots of 5 mM NiCl2 were added to achieve a desired Ni(II) concentration. Resulting mixtures were incubated for 3 min before measuring either their extinction spectra or SERS intensity.

Nanoparticle Characterization: SERS Measurements: SERS spectra were recorded on a Leica DM/LM microscope equipped with an Olympus 20x/0.4 long-working distance objective to collect 180° backscattered light from a microcuvette. The spectrometer system was a Renishaw Ramascope System 2000 with the 785 nm line of a diode laser as the excitation source. The unfocussed power output was measured to be approximately 140 mW at the sample. The spectra were typically acquired with an exposure time of 2 × 10s. Dielectric edge filters were used to reject the Rayleigh scattered light.

Nanoparticle Characterization: UV–Visible Spectroscopy: UV–vis spectra were recorded on a Cary 300 Bio UV–vis spectrophotometer using 1 cm path length quartz cells.

Nanoparticle Characterization: Scanning Electron Microscopy (SEM): SEM investigations were carried out by preparing poly(diallyldimethylammonium) (PDDA) coated silicon wafers. Silicon wafers (Agar Scientific) were first cleaned with methanol and oxygen plasma (Diener electronic femto oxygen plasma cleaner, 72 cm3 per min gas flow). They were then coated with a 10 mg mL−1 PDDA solution in 1 mM NaCl for 30 min. After this time the wafers were rinsed with milli-Q water and dried under N2 flow. 10 μL of the colloidal solutions were deposited on individual wafers and allowed to rest for 10 min, the samples were then removed and the wafer washed with milli-Q water. Imaging was carried out on a Sirion 200 Schottky field-emission electron microscope (FEI) operating at an accelerating voltage of 5 kV. The samples did not require additional metallic coating before imaging. Image analysis was carried out using Image J, v1.43u.

Supporting Information

Supporting Information is available from the Wiley Online Library or from the author.

Acknowledgements

ŽK and LG contributed equally to this work. The authors acknowledge support from the EPSRC through grants EP/D062861/1 and EP/E032715/1 to ŽK and LG and wish to acknowledge the Royal Society for Wolfson Research merit award to DG. IAL thanks the EU 7th Framework Programme, NMP-2008-1.1-1; SMD-229375 for funding.