1. Introduction

Light-induced collective oscillation of free elections of a nanometer-sized particle results in strong absorption and scattering of light. Gold and silver nanoparticles exhibit this phenomenon, localized surface plasmon resonance (LSPR), in the visible and near-infrared (NIR) regions. The optical properties of the plasmonic nanoparticles depend strongly on refractive index of the surrounding medium. In particular, the extinction ( = absorption + scattering) peak is redshifted as the refractive index increases. Therefore, plasmonic nanoparticles are used as LSPR sensors, which can be applied to affinity-based chemical sensing and biosensing [1–3]. For example, plasmonic nanoparticles modified with antibodies serve as LSPR-based immunosensors for corresponding antigens [1–3], because antigen-antibody binding raises refractive index in the proximity of the particles and gives rise to the extinction peak shift. The LSPR sensors are more cost-effective than other plasmonic sensors such as propagating surface plasmon resonance (SPR) sensors and are sensitive to refractive index changes exclusively in the immediate vicinity of the particles [3]. Among the plasmonic nanoparticles, gold nanoparticles, which are chemically stable and easily modifiable, have been widely utilized as LSPR sensors.

However, it is necessary to pay special attention when an LSPR sensor is applied to biological samples such as blood, which strongly absorbs visible light. Actually, a simple measurement of transition spectrum is difficult in the visible region when plasmonic nanoparticles are in contact with biological samples. We have addressed this issue by developing LSPR sensors that directly output electrical signals [4] on the basis of plasmon-induced charge separation (PICS) [5] or sensors based on plasmonic backward scattering [6]. On the other hand, there are two wavelength ranges in the NIR region in which light can penetrate biological samples; the first and second biological optical windows (650−950 and 1000−1350 nm, respectively) [7]. Therefore, nanoparticles that show LSPR in these optical windows can be exploited for LSPR-based biosensing. Although spherical gold nanoparticles exhibit LSPR in the visible region, anisotropic ones such as rod-shaped, branched, and bipyramidal nanoparticles resonate with NIR light [3,8,9]. The anisotropic particles have another advantage, namely high refractive index sensitivity. Miller and Lazarides reported that the LSPR peak shift per refractive index unit is improved with increasing LSPR wavelength [10]. However, the anisotropic nanoparticles of fcc metals such as gold are more difficult to prepare with high reproducibility in comparison with spherical ones. Although we demonstrated resonant wavelength selective dissolution of plasmonic nanoparticles based on PICS [11] and applied it to LSPR sensing on the basis of extinction dip shifts rather than peak shifts [12], deposition of nanorods is necessary for sensing in the NIR region [13]. Even for the potential-scanning LSPR sensor, which can detect refractive index changes at a single wavelength [14], anisotropic nanoparticles are required for operation in the NIR region.

Recently, it has been reported that compound nanoparticles such as tungsten oxide [15], tin-doped indium oxide [16], and copper sulfide [17] show LSPR in the NIR region. Their LSPR is based on light-induced collective oscillation of charge carriers in the compound nanoparticles (i.e. electrons in the conduction band or holes in the valence band). In general, compound nanoparticles have lower carrier density (N) and thereby lower plasma frequency (ω_p) in comparison with metal nanoparticles: ω_p = (Ne²/ε₀m)^1/2 where e, ε₀, and m are elementary charge, vacuum permittivity, and effective mass of electron or hole, respectively. Therefore, compound nanoparticles exhibit LSPR at longer wavelengths, typically in the NIR region. Most of the compound nanoparticles are cost-effective and more stable than silver and copper nanoparticles. Among the plasmonic compound nanoparticles, copper sulfide ones have attracted attention as an alternative material for noble metal nanoparticles because they consist of earth-abundant elements. Copper sulfide is classified into chalcocite (Cu₂S), djurleite (Cu_1.97S), digenite (Cu_1.8S), anilite (Cu_1.75S), cevellite (CuS), and others and shows various different chemical and optical properties depending on the chemical composition and crystal structure. There are increasing reports on synthetic procedures and control of shape, crystal structure, and carrier density of copper sulfide nanoparticles [17–20]. Although Swihart et al. mentioned that the LSPR peak of copper sulfide nanoparticles redshifts with increasing refractive index of the surrounding medium [18], there is no study on copper sulfide nanoparticles as LSPR sensors, to the best of our knowledge.

In this study, we focus on CuS nanoparticles that exhibit strong LSPR absorption in the NIR region [19] and systematically evaluate their LSPR properties and refractive index sensitivity. Here we demonstrate that LSPR sensing with the CuS nanoparticles is feasible and that the LSPR wavelength and refractive index sensitivity can be controlled within the second biological optical window.

2. Experimental

2.1 Materials

All reagents, namely copper(I) chloride, elemental sulfur, oleylamine (OAm), oleic acid (OAc), hexane, toluene, and ethanol were purchased from Wako Pure Chemical Industries.

2.2 Synthesis of CuS nanoparticles

Copper(I) chloride and elemental sulfur (12.5 mmol each) were suspended in a mixture of OAm and OAc (3.75 mL) and the suspension was heated at 140 °C for 5 min in a nitrogen atmosphere. After cooling to room temperature, the solution was mixed with ethanol (9 mL) and centrifuged at 5800 × g at 5 °C for 5 min. The obtained precipitate was dissolved in toluene (1 mL). After repeating the centrifugation cycle, CuS nanoparticles were obtained as a blackish brown powder. The volume ratio of OAm to OAc was changed to control the size of the resulting nanoparticles.

2.3 Characterization

Extinction spectra were measured with a JASCO V-670 spectrophotometer. Transmission electron microscopy (TEM) images were obtained on a JEOL JSM-2010F at an acceleration voltage at 200 kV. Scanning electron microscopy (SEM) and elemental analysis were performed on a JEOL JSM-7500FA equipped with an energy dispersive X-ray (EDX) spectrometer (JEOL EX-64195MU). X-ray diffraction patterns were recorded on a Rigaku RINT2000.

3. Results and discussion

3.1 Preparation and characterization of CuS nanoparticles

We synthesized CuS nanoparticles in mixtures of OAm and OAc with the OAm volume ratio of 15−100%. Selected extinction spectra of the synthesized CuS nanoparticles and the dependence of the extinction peak wavelength on the OAm volume ratio are shown in Fig. 1. Strong extinction peaks due to LSPR, which redshifted with increasing OAm ratio, were observed in the NIR region (1050−1300 nm). The extinction at <600 nm is ascribed to the electron excitation from the valence band to conduction band of CuS, as the onset of extinction is in good agreement with the band gap of covellite (2 eV, ca. 620 nm) [21].

 

Fig. 1 (a) Extinction spectra of toleune solutions of the CuS nanoparticles prepared in mixtures of OAm and OAc and (b) the relationship between the OAm volume ratio and the extinction peak wavelength.

Download Full Size | PDF

TEM images and photographs of the synthesized CuS nanoparticles are shown in Fig. 2. Hexagonal nanoplates with thickness of ca. 2−3 nm are observed in each sample. The mean in-plane diameters of the nanoplates prepared in mixtures with the OAm volume ratio of 15, 25, 55, 75, and 100 are 7.2 ± 0.9, 7.1 ± 1.4, 7.6 ± 1.2, 8.9 ± 1.1, and 12.7 ± 1.8 nm (mean ± standard deviation), respectively. In general, plasmonic nanoplates exhibit at least two distinctive LSPR modes, namely out-of-plane mode and in-plane mode [22,23]. The former based on out-of-plane electron oscillation shows a smaller extinction peak at a shorter wavelength, and the latter due to in-plane electron oscillation gives a larger extinction peak at a longer wavelength. The latter peak is known to redshift significantly as the aspect ratio (i.e. ratio of the in-plane diameter to thickness) increases [23]. A similar trend was also observed for previously reported [18] and the present CuS nanoplates (Fig. 1(b), Fig. 2).

 

Fig. 2 TEM images of the CuS nanoparticles prepared in mixtures of OAm and OAc. The OAm volume ratio is (a) 15, (b) 25, (c) 55, (d) 75, or (e) 100%. (f) Photograph of the CuS solution (OAm ratio = 100%).

Download Full Size | PDF

Figure 3 shows XRD patterns of the CuS nanoplates. Each pattern is assigned to covellite [24] regardless of the OAm ratio. The peak at 48.0° is assigned to the diffraction of (110) planes in covellite, which is perpendicular to the in-plane direction [19], and full width at half maximum of the peak becomes broader as the OAm ratio decreases. According to the Scherrer's equation, the in-plane diameters are calculated to be ca. 18, 23, and 26 nm for the nanoplates synthesized at the OAm volume ratios of 15, 75, and 100%. These values are higher than those determined from the TEM images (Fig. 2), probably because larger nanoplates give higher diffraction intensity. Even so, the same trend is observed in both cases; the diameter increases with the OAm ratio. As above, we successfully controlled the in-plane diameter and LSPR wavelength of the CuS nanoplates by changing the OAm/OAc ratio in the reaction suspension.

 

Fig. 3 XRD patterns of the CuS nanoplates prepared in mixtures of OAm and OAc. Reference pattern of covellite (data from Ref 24) is also shown.

Download Full Size | PDF

3.2 Refractive index sensitivity of CuS nanoplates

Extinction spectra of the CuS nanoplates were measured in mixed solvents of toluene (refractive index n = 1.496) and hexane (n = 1.375) in order to evaluate their refractive index sensitivity. When the volume ratios of toluene to hexane are 25:75, 50:50, 75:25, and 100:0, refractive indices of the solvents are 1.405, 1.437, 1.466, and 1.496, respectively. Extinction spectra of CuS nanoplates prepared at the OAm volume ratio of 55% are shown in Fig. 4(a). The extinction peak is redshifted with increasing refractive index of the solvent as in the case of most plasmonic nanoparticles. Figure 4(b) shows the relationship between the refractive index of the solvent and the extinction peak wavelength. The plot is almost linear and the refractive index sensitivity of the nanoplates is determined to be 223 nm RIU⁻¹ (RIU: refractive index unit) from the slope of the approximate straight line.

 

Fig. 4 (a) Extinction spectra of the CuS nanoplates (the OAm volume ratio = 55%) in mixed solvents of toluene and hexane and (b) the relationship between the refractive index of the solvent and the extinction peak wavelength.

Download Full Size | PDF

The sensitivities of the other CuS nanoplates shown in Fig. 1(b) were also evaluated in the same way and plotted against the peak wavelength in toluene (Fig. 5). Corresponding data for previously reported gold nanoparticles with different shapes are also shown in Fig. 5 for comparison [3,8,9]. The sensitivity of CuS nanoplates varies in the range 160−600 nm RIU⁻¹ and tends to increase linearly with the peak wavelength in toluene. These results indicate that the sensitivity of CuS nanoplates can be controlled by changing the particle size and predicted from the extinction peak wavelength. It is known that the refractive index sensitivity of gold nanoparticles is almost linearly dependent on the peak wavelength in a certain solvent (e.g. water) [10]. Such a relationship for plasmonic compound nanoparticles has not yet been reported, to the best of our knowledge. Because the sensitivities of CuS nanoplates are lower than those of gold nanoparticles in the range in Fig. 5, they are more suitable for sensing in a wide dynamic range. The figure of merit (FOM) for the CuS nanoplates were evaluated by dividing the sensitivity by the full width at half maximum of the extinction peak to be 0.4−1.0, being comparable to those for gold nanospheres and nanobranches [3,8].

 

Fig. 5 Relationships between the extinction peak wavelength and refractive index sensitivity of the present CuS nanoplates and the reported gold nanoparticles (data from Refs 3, 8, and 9). The peak wavelengths were determined in toluene and water for CuS and gold, respectively. The first and second optical windows for biological tissues are also shown.

Download Full Size | PDF

The first and second optical windows for biological tissues [7] are also shown in Fig. 5. Because gold nanoparticles and CuS nanoplates almost cover the first and second windows, respectively, both of them are suitable for biosensing in terms of the operation wavelength. However, as described in Introduction, the peak wavelength of spherical gold nanoparticles, which are easy to prepare and available at low cost, is outside the window (ca. 520 nm). Although some gold nanoparticles with anisotropic shape exhibit LSPR within the optical windows [3,8,9], their synthetic procedures are somewhat complicated. The present CuS nanoplates have advantages over the anisotropic gold nanoparticles in terms of easy synthesis and controllability; CuS nanoplates can be prepared simply by heating the reactants and their LSPR properties can be controlled only by changing the OAm/OAc ratio. In addition, CuS is much less expensive than gold. CuS nanoplates can easily be modified with thiols, amines, or carboxylic acids for introduction of receptors for biosensing, and the modification would improve chemical stability of the nanoplates as well.

4. Summary

Plasmonic CuS (covellite) nanoplates with different sizes were prepared and characterized in detail. The in-plane diameter of the nanoplates was controlled by changing the composition of the reaction suspension. Refractive index sensitivity of the CuS nanoplates measured in mixed solvents of toluene and hexane increased from 160 to 600 nm RIU⁻¹ as the particle diameter and the peak wavelength increased, indicating that the sensitivity and operation wavelength of the CuS-based LSPR sensor can be controlled by changing the synthesis conditions. Because CuS nanoplates show LSPR in the second biological optical window and are much more cost-effective than noble metals, they are suitable for LSPR-based biosensing.