Facile, fine post-tuning of the longitudinal absorption wavelengths of pre-synthesized gold nanorods by introducing sulfide additives

Abstract

Because the specific surface plasmon resonance (SPR) wavelength of gold nanorods (GNRs) is required in most photothermal bio-applications to enhance therapeutic efficacy, the precise expectation of the SPR wavelength of GNRs in the synthesis is required. Here, we present a facile, fine post-tuning of the longitudinal absorption spectrum of pre-synthesized gold nanorods by introducing controlled amounts of sulfide additives. The longitudinal SPR wavelength was controlled to a longer wavelength (red-shifted) by increasing the concentration of sulfide. In addition, the SPR wavelength determined after sulfide treatment was not shifted and was fixed for a long period of time. This allows a powerful second chance to adjust the SPR wavelength of given GNRs by a simple post-treatment, which can circumvent the risk of experimental failure in a GNR synthesis with a specific wavelength. This method may be beneficial in the efficient preparation of GNRs for various applications requiring specific SPR wavelengths with high stability.

---

1. Introduction

Gold nanorods (GNRs) have been studied widely due to their strong near-infrared (NIR) absorption and scattering based on surface plasmon resonance (SPR).^1–3 The basis of this property of GNRs originates from the collective oscillation of conduction band electrons at the surface of GNRs.^4,5 There are two strong absorptions in GNRs, at ∼520 nm and at 700–1600 nm, the transverse and longitudinal bands, respectively. The position of longitudinal band can be controlled by changing the aspect ratio of GNRs.⁵ This tunable longitudinal SPR wavelength of GNRs to the NIR range and the transparency of tissue in this area allow diverse biomedical applications, including photothermal therapy of cancers via laser irradiation and biological imaging using scattered NIR light that penetrate tissues.^6–10 In addition, the surface of GNRs can be modified readily using thiol chemistry, enabling conjugation to biomolecules for specific targeting or delivery to diseased tissue.^11–14

As a specific SPR wavelength of the GNRs is demanded in most photothermal bio-applications, due to laser irradiation with a predetermined wavelength, for example, 808 nm, there is a precise expectation for a specific SPR wavelength in GNRs from the beginning of any synthesis.^15–17 This is important for efficient photothermal conversion, based on same amount of GNRs.⁵ A few nanometers of optical shift in the absorption wavelength of the GNRs could lead to a considerable decline in absorption intensity at the laser wavelength, resulting in diminished photothermal efficiency. There have been many reports on the tunability of the aspect ratio and geometry of GNRs to control the SPR absorption wavelength.^18–21 The El-Sayed group showed that the longitudinal band shifts to longer wavelengths as the aspect ratio increases. The aspect ratio is increased by increasing the amounts of the silver ions added in the growth solution.⁵ The Murphy group revealed that the aspect ratio of GNRs can be controlled readily by varying the concentrations of ascorbic acid and silver nitrate.¹⁸ However, it has also been reported that the plasmon resonance of the GNRs after synthesis drifts continuously towards shorter wavelengths (blue shift) over hours to days, up to ∼100 nm, which is probably due to slow, continued attachments of gold atoms on the side surface of the gold nanorods.²² This blue shift could be an obstacle to preparing GNRs with a fixed longitudinal SPR wavelength that is critical to preparing highly efficient photothermal agents. In 2005, Wei et al. reported that the blue shift in longitudinal peak of the gold nanorods after synthesis can be arrested by sulfide treatment and that the optical shift could be reduced by adding optimal millimolar concentrations of Na₂S, a molar ratio of sulfur to metal ions (gold, silver), 4 : 1.²² The sulfide addition was mostly conducted in relatively short time after seed addition to growth solution (e.g., 15 min). However, there has been no report on the systematic control of the SPR wavelengths of gold nanorods to tune the SPR wavelengths of pre-prepared GNRs, which may be useful in preparing a specific absorption wavelength using a stock of GNRs.

Here, we show that the longitudinal SPR wavelength of pre-prepared GNRs can be precisely controlled by the simple addition of sulfide solution to the GNR solution. The longitudinal SPR wavelength was increased (red-shifted) by increasing the amounts of sulfide added and plateaued above a limiting concentration of sulfide. The SPR wavelength determined by sulfide treatment did not shift and was fixed over a long period of time. In addition, we showed that this approach could be applied to not only the fresh GNRs as used in the previous work²² but also GNRs aged for longer period time (e.g., 24 hours) after synthesis. This simple method could allow a powerful second chance to precisely adjust the SPR wavelength of given GNRs to a specific value by a simple post-treatment, which can circumvent the risk of experimental failure in GNR synthesis with a specific wavelength. Although there may be mistakes in the preparation of the ‘right’ SPR wavelength, the SPR peaks can be precisely modified to a target wavelength that does not then shift subsequently. This technique may be beneficial in the efficient preparation of GNRs for various applications requiring specific SPR wavelengths with high stability.

2. Experimental section

Materials

Gold chloride trihydrate (HAuCl₄·3H₂O), cetyltrimethylammonium bromide (CTAB), sodium sulfide nonahydrate (Na₂S·9H₂O), and 2,6-dihydroxybenzoic acid were purchased from Sigma-Aldrich (St. Louis, MO, USA). Silver nitrate (AgNO₃) and sodium borohydride (NaBH₄) were purchased from Samchun Chemicals (Kangnam-ku, Seoul, Korea). L-Ascorbic acid was purchased from Tokyo Chemical Industry (West Tambaram Chennai, Tamil Nadu 600047, India). All reagents were used without further purification.

Synthesis of gold nanorods

Gold nanorods were prepared by a slightly modified seed-mediated growth method.^18–20 First, 2,6-dihydroxybenzoic acid solution (0.067 M) was added into growth solution to improve the monodispersity and yield of gold nanorods.^23,24 The seed solution was prepared by sequential adding a HAuCl₄·3H₂O solution (0.01 M, 0.25 mL) and aqueous fresh ice-cold NaBH₄ solution (0.01 M, 0.6 mL) into a CTAB solution (0.1 M, 7.5 mL) under vigorous stirring for 2 min. The resulting gold seed solution was kept at 25 °C for 2 h. Next, the growth solution was prepared by sequential mixing of HAuCl₄·3H₂O (0.01 M, 0.5 mL), 2,6-dihydroxybenzoic acid (0.067 M, 1 mL), AgNO₃ (0.01 M, 0.1 mL), and L-ascorbic acid (0.1 M, 0.08 mL) into CTAB solution (0.1 M, 12 mL). To prepare gold nanorods, 0.1 mL of the seed solution was added into 13.68 mL of growth solution under stirring. After 30 s, the mixture solution was kept under static conditions to allow the growth of gold nanorods to proceed.

Optical shift control by sulfide introduction

To test the effects of the addition of sulfide solution on the absorption of GNRs, 13.78 mL of the GNR solution 40 min after synthesis was mixed with different volumes (25, 37, 50, 75, 100, and 150 μL) of 10 mM aqueous sodium sulfide solution, which corresponded to molar ratios of S to metal (Au and Ag) of 1 : 24, 1 : 18, 1 : 12, 1 : 9, 1 : 6, and 1 : 4.5, respectively. The resulting solutions were stirred for 30 s and kept under static conditions; absorption spectra were determined. To reveal the effects of the timing of sulfide addition, a determined sulfide solution (50 μL, 10 mM) was added at different time points (5, 10, 20, 30, and 40 min) after the synthesis of GNRs. To test the general controllability of GNR absorbance, GNRs with various longitudinal SPR wavelengths were first prepared by varying the volume of L-ascorbic acid (0.08, 0.27, 0.43, and 0.59 mL) in the GNR synthesis. Then, different volumes (25, 37, 50, 75, 100, 150 μL) of 10 mM sulfide solution were introduced into the pre-prepared gold nanorods solution. The final concentrations of sulfide in the gold nanorods solution was 18.3, 27.0, 36.5, 54.8, 73.0, and 110 μM, respectively.

Characterization

Absorption spectra of the samples were taken using a Multiskan GO Microplate Spectrophotometer (Thermofisher). Transmission electron microscopy (TEM) images were taken with a JEM 3010 ultrahigh resolution analytical electron microscope (Jeol). Elemental analysis of gold nanorods samples was obtained from the JEM 3010 ultrahigh resolution analytical electron microscope using energy dispersive X-ray spectroscopy (EDS).

3. Results and discussion

The GNRs were prepared by a seed-mediated growth method according to the previous reports.^18–21 To check the blue shift of the GNRs after synthesis over time, Vis-NIR absorption spectra were taken for every 3 min for 12 h after the seed solution was added to the growth solution (Fig. 1a). GNRs showed drastic increment of LSPR intensity till 15 min after the seed solution introduction, which is corresponding to the rapid growth of GNRs (Fig. S1†). In the following period, the blue shift of the longitudinal SPR wavelength occurred up to 47 nm during 12 h after synthesis, consistent with a previous report.²² This blue shift could be an obstacle to preparing GNRs with fixed longitudinal SPR wavelengths.

Fig. 1 (a) Change in longitudinal absorption peaks of GNRs over time. Time represents the duration since the seed solution was added into the growth solution. (b) Time-dependent change in the longitudinal absorption wavelength after injection of various amounts of sulfide solution into pre-prepared GNRs. Sulfide solution was introduced at 40 min after seed solution was added in growth solution. (c) Variation in longitudinal wavelengths before and after addition of sulfide solution. The control GNRs, with no addition of sulfide, showed a 30 nm blue shift. As the concentration of sulfide increased, the longitudinal wavelength increased gradually (red shift).

We next investigated whether the longitudinal absorption peaks could be controlled by the addition of sulfide solution and the longitudinal peaks could be fixed over time. At 40 min post-synthesis of the GNRs, different amounts of the sulfide solutions were added to the pre-prepared GNR-solution and the longitudinal SPR peaks were measured every 20 min for 2 h after addition of sulfide (Fig. 1b). The longitudinal absorption wavelength of the control GNR solution (sulfide-free) decreased gradually, from 836 nm to 822 nm, over time (Fig. 1b). In contrast, all conditions with the addition of sulfide resulted in red shifts of the longitudinal absorption wavelength compared with the control GNRs. Although the absorption intensity decreased around 10% after sulfide treatment (Fig. S2†), it was comparable to the intensity loss during the blue-shift of bare GNRs. The addition of 25 μL of sulfide solution led to maintaining the SPR wavelength of the GNRs at 40 min post-synthesis over time. For the addition of higher amounts of sulfide solution, there was a two-step wavelength change over time. First, a marked red shift was observed for 20 min after the addition of sulfide, dependent on the amount of sulfide added. Then, a subsequent plateau was observed with nearly constant wavelength until 2 h after the addition of the sulfide. The extent of the red shift was almost linearly proportional to the amount of sulfide when it was less than 100 μL (Fig. 1c), but apparently became saturated above a certain limiting concentration. There was no more notable red shift by the addition of 100 μL and 150 μL compared with the addition of 75 μL of sulfide solution. The maximum longitudinal SPR wavelength obtained from sulfide addition was 897 nm, which corresponds to ∼100 nm red shift compared with the control GNRs (806 nm wavelength). These results indicate that target longitudinal SPR wavelengths of up to 100 nm red-shift can be achieved by simple post treatment with sulfide and the resulting SPR wavelength are nearly fixed, in contrast to the original GNRs. Generally, a longer longitudinal wavelength of GNRs resulted from a higher aspect ratio.⁵

To investigate whether these red shifts in SPR wavelength were due to the aspect ratio of GNRs, the GNRs were further analyzed with transmission electron microscopy (TEM) and the aspect ratios of GNRs were measured (Fig. 2). There was no obvious change in the aspect ratio of GNRs obtained by addition of different amounts of sulfide, indicating that the red shift resulting from the addition of sulfide was not mainly due to a change in the aspect ratio alone.

Fig. 2 TEM images of GNRs obtained after addition of (a) 0 (control), (b) 25, (c) 37, (d) 50, (e) 75, (f) 100, and (g) 150 μL of 10 mM sulfide solution. The corresponding final sulfide concentrations in GNR solutions were 0, 18.3, 27.0, 36.5, 54.8, 73.0, and 110 μM, respectively. (h) Aspect ratios of GNRs measured from the TEM images.

We next hypothesized that there would be a critical effect of timing of the sulfide introduction on the final longitudinal SPR wavelength. To investigate this, the same volume (50 μL) of 10 mM sulfide solution, corresponding to 36.5 μM of final sulfide concentration, was introduced to the pre-prepared GNR-solution at 5, 10, 20, 30, and 40 min after the start of GNR growth by mixing seed nanoparticles and growth solution. The longitudinal SPR wavelength was measured over time (Fig. 3). The sulfide-free control GNRs (red line in Fig. 3a) showed a steep increase of SPR wavelength until 10 min, due to the growth of GNRs from seed nanoparticles, and a continued blue shift in wavelength after 10 min was observed (as in Fig. 1a). When the sulfide solution was added at 20, 30, or 40 min after the start of GNR growth, immediate red shifts of the longitudinal SPR wavelength to longer wavelengths were observed and the red-shifted wavelengths were maintained afterwards. In contrast, when the sulfide solution was added at earlier time points, 5 or 10 min post synthesis, the absorption wavelengths were immediately blue-shifted to shorter wavelengths than that of the sulfide-free control GNRs. The minimum wavelength obtained in this experiment was 805 nm, resulting from the addition of sulfide at 5 min post synthesis, which is a 13 nm blue shift from the control GNRs. These observations indicate that there is a critical timing to determine the red shift of GNRs by addition of sulfide and the introduction timing of the sulfide in addition to the amounts of the sulfide can be also a controlling parameter to determine the SPR wavelength of GNRs.

Fig. 3 (a) Time-dependent change in the longitudinal wavelengths of GNRs after addition of the same amount of sulfide solution to pre-prepared GNR solutions at different time points (5, 10, 20, 30, and 40 min after GNR synthesis). (b) TEM images of GNRs without addition of sulfide (left) and GNRs with additions of sulfide at 5 min (middle) and 40 min (right) after GNR synthesis. The corresponding aspect ratios measured from the TEM images were 3.80 ± 0.51, 3.64 ± 0.52, and 3.64 ± 0.49, respectively.

As the sulfides added to GNRs can bind and scavenge the gold atoms on the surface of GNRs,²² the addition of sulfides might decrease the aspect ratio of GNRs by inhibiting the growth of GNRs. However, TEM images of control GNRs, GNRs obtained by addition of sulfide at 5 and 40 min post synthesis showed that although there was a small decrease of aspect ratio by addition of sulfide, the difference of aspect ratios between GNRs obtained by addition of sulfide at 5 min and 40 min was negligible (Fig. 3b). Thus, there may be other factors resulting in the shorter or longer longitudinal SPR wavelength of the resulting GNRs.

Based on these results, we hypothesized that there might be a predictable relationship between sulfide addition and the change in longitudinal SPR wavelength of the GNRs. From a control point of view with sulfide concentration, as starting materials, we used GNRs with stable SPR peaks obtained from aging for 24 h to enhance the accuracy of the final wavelength. Although the SPR wavelength of GNRs can be modified by introduction of sulfide to GNR-solution at earlier time points than 2 h post synthesis (Fig. 1b and 3a), the SPR peaks of GNRs are still under blue-shifting at that early period of the synthesis, which may cause a mismatch in the final wavelength from the targeted wavelength. To find a more precise method to control the SPR wavelength of GNRs by post-treatment with sulfides, four GNRs with different aspect ratios were first prepared by changing the concentration of ascorbic acid in the synthesis and used as starting materials for post-control of the longitudinal SPR wavelength^18,19 (Fig. 4a). The SPR wavelengths of the prepared GNRs were 730, 748, 770, and 797 nm, respectively. The different volumes of sulfide solutions were introduced to the four types of 24 h-aged, pre-prepared GNRs solutions, respectively. Fig. 4a shows the wavelength change of GNRs with 797 nm of longitudinal wavelength. All samples showed a similar trend: a higher red shift with the addition of more sulfide solution, as shown in the result of GNRs of 797 nm (Fig. 4a). The addition of sulfide led to a marked red shift within 10 min after the addition of sulfide followed by a constant wavelength up to 1 h, as expected from the previous observations.

Fig. 4 (a) Change in longitudinal absorption wavelengths of GNRs with 797 nm of LSPR wavelength after addition of different amounts of sulfide solutions. The GNRs tested was prepared and aged for 24 h before the addition of sulfide solutions. (b) Two-stage linear relationship between the concentration of sulfide in GNR solutions and the red shift after sulfide addition. GNRs with different longitudinal wavelengths (730, 748, 770, and 797 nm) were tested in the experiments. (c) The dielectric constants and the aspect ratios of GNRs obtained from different sulfide concentrations. The aspect ratios of GNRs were measured from TEM images of GNRs. The dielectric constants were calculated from the aspect ratios and longitudinal absorption wavelengths.

The variation in absorption wavelengths of the four types of GNRs were analyzed together as a function of sulfide concentration (Fig. 4b). There was a clear two-stage linear relationship between sulfide concentration and the red shift of the resulting GNRS. This relationship may support the potential application of the post-treatment of GNRs with sulfide to the precise post-control of the SPR wavelength of GNRs.

The role of sulfide in controlling the SPR wavelengths of GNRs was investigated further. Although GNRs showed red shifts in the absorption wavelength, the aspect ratios of GNRs measured from TEM analysis was not increased (see Fig. 2h and 3b). We hypothesized that the absorption of sulfide on the surface of the GNRs may significantly affect the dielectric function of the resulting GNRs and thus lead to a red shift in the SPR wavelength. The energy dispersive spectroscopy (EDS) analysis of GNRs in Fig. 4a indicated that there are S atoms on the surface of GNRs treated with the sulfide solution (Table 1), showing that added sulfides were adsorbed on the surface of GNRs, as reported before.²² As the sulfur–gold bonds formation is known as reversible, the surface modification of sulfide-adsorbed GNRs via gold–thiol chemistry is still applicable.^25–27 In addition to the aspect ratio of GNRs, another important parameter affecting the SPR wavelength is the dielectric constant of the medium surrounding the GNRs. Link et al. proposed that there are functional relationships between the optical absorption spectra, the aspect ratio, and dielectric constant of the medium.^28,29 Based on Mie–Gans theory, a parameterized equations, as follows, was proposed:

λ_max = (53.71 × R − 42.29) × ε_m + 495.14

where λ_max is the longitudinal absorption wavelength with maximum intensity of GNRs, R is the aspect ratio of GNRs and ε_m is the effective dielectric constant of the surrounding media.

Table 1 Atomic compositions of GNRs obtained from energy dispersive spectroscopy during TEM analysis of GNRs without (Fig. 2a) and after (Fig. 2e) addition of sulfide solution. n = 3 for each sample

Element	No sulfide added	Sulfide added
S	0	12.1 ± 1.38
Br	3.66 ± 0.64	6.82 ± 2.57
Ag	6.22 ± 0.36	9.24 ± 1.32
Au	90.1 ± 0.28	71.9 ± 2.68

---

To see the effect of sulfide addition on the dielectric constants and subsequent red shift of wavelengths of GNRs, the absorption wavelengths and aspect ratios of GNRs after additions of differing amounts of sulfide solutions were measured. Although GNR showed red shifts in the absorption wavelength (Fig. 4a), the aspect ratios of GNRs were not increased and, instead, were decreased slightly by the introduction of larger amounts of sulfide (blue line in Fig. 4c). Based on the aspect ratio and wavelength of GNRs, the effective dielectric constants around GNRs were calculated from the above equation (red line in Fig. 4c), showing that the effects on dielectric constants were increased at higher sulfide concentrations. This indicates that the increased dielectric constants of the GNR surroundings due to the sulfide introduction contribute dominantly to the increase in SPR absorption wavelength rather than the aspect ratio. The surface gold atoms may be saturated above the certain concentration of sulfide, which might lead to the negligible additional red shift observed with higher sulfide concentrations (Fig. 4b).

4. Conclusions

In summary, we demonstrated precise control over the longitudinal SPR wavelength of gold nanorods by introducing sulfide solution, a noble metal capping agent. Gold nanorods with different LSPRs were synthesized by adding appropriate amounts of sulfide solutions from a one-pot synthesized stock gold nanorods solution. In addition, the same experiments were carried out for GNRs having other LSPR wavelengths. There was a clear two-stage linear relationship between sulfide concentrations and LSPR wavelength variation. Although the aspect ratio of GNRs did not changed significantly, the adsorbed sulfur on the GNRs contributed to the enhanced dielectric constants in the surroundings of the GNRs, which resulted in the red shifts with the addition of sulfide solutions. This facile, post-tuning method may provide an enhanced precise tunability of LSPRs of pre-synthesized gold nanorods as well as a valuable means for diagnostic and therapeutic applications.

Acknowledgements

This work was supported by a grant funded by the National Research Foundation under the Ministry of Science, ICT & Future Planning, Republic of Korea (grant number: 2010-0027955) and a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (grant number: HI14C0211).

References

1. C. J. Murphy, T. K. Sau, A. M. Gole, C. J. Orendorff, J. Gao, L. Gou, S. E. Hunyadi and T. Li, J. Phys. Chem. B, 2005, 109, 13857–13870.
2. C. Burda, X. Chen, R. Narayanan and M. A. El-Sayed, Chem. Rev., 2005, 105, 1025–1102.
3. P. K. Jain, X. Huang, I. H. El-Sayed and M. A. El-Sayed, Acc. Chem. Res., 2008, 41, 1578–1586.
4. C. J. Murphy, L. B. Thompson, A. M. Alkilany, P. N. Sisco, S. P. Boulos, S. T. Sivapalan, J. A. Yang, D. J. Chernak and J. Huang, J. Phys. Chem. Lett., 2010, 1, 2867–2875.
5. X. Huang, S. Neretina and M. A. El-Sayed, Adv. Mater., 2009, 21, 4880–4910.
6. L. Tong, Q. Wei, A. Wei and J. X. Cheng, Photochem. Photobiol., 2009, 85, 21–32.
7. A. M. Alkilany, L. B. Thompson, S. P. Boulos, P. N. Sisco and C. J. Murphy, Adv. Drug Delivery Rev., 2012, 64, 190–199.
8. X. Huang, I. H. El-Sayed, W. Qian and M. A. El-Sayed, J. Am. Chem. Soc., 2006, 128, 2115–2120.
9. H. Liao and J. H. Hafner, Chem. Mater., 2005, 17, 4636–4641.
10. M. A. Mackey, M. R. Ali, L. A. Austin, R. D. Near and M. A. El-Sayed, J. Phys. Chem. B, 2014, 118, 1319–1326.
11. E. Ruoslahti, S. N. Bhatia and M. J. Sailor, J. Cell Biol., 2010, 188, 759–768.
12. V. Shanmugam, Y. H. Chien, Y. S. Cheng, T. Y. Liu, C. C. Huang, C. H. Su, Y. S. Chen, U. Kumar, H. F. Hsu and C. S. Yeh, ACS Appl. Mater. Interfaces, 2014, 6, 4382–4393.
13. E. C. Dreaden, A. M. Alkilany, X. Huang, C. J. Murphy and M. A. El-Sayed, Chem. Soc. Rev., 2012, 41, 2740–2779.
14. T. Niidome, M. Yamagata, Y. Okamoto, Y. Akiyama, H. Takahashi, T. Kawano, Y. Katayama and Y. Niidome, J. Controlled Release, 2006, 114, 343–347.
15. V. Ntziachristos, A. G. Yodh, M. Schnall and B. Chance, Proc. Natl. Acad. Sci. U. S. A., 2000, 97, 2767–2772.
16. R. Weissleder, Nat. Biotechnol., 2001, 19, 316.
17. L. Vigderman, B. P. Khanal and E. R. Zubarev, Adv. Mater., 2012, 24, 4811–4841.
18. T. K. Sau and C. J. Murphy, Langmuir, 2004, 20, 6414–6420.
19. S. E. Lohse and C. J. Murphy, Chem. Mater., 2013, 25, 1250–1261.
20. B. Nikoobakht and M. A. El-Sayed, Chem. Mater., 2003, 15, 1957–1962.
21. C. J. Johnson, E. Dujardin, S. A. Davis, C. J. Murphy and S. Mann, J. Mater. Chem., 2002, 12, 1765–1770.
22. D. A. Zweifel and A. Wei, Chem. Mater., 2005, 17, 4256–4261.
23. L. Scarabelli, M. Grzelczak and L. M. Liz-Marzán, Chem. Mater., 2013, 25, 4232–4238.
24. X. Ye, L. Jin, H. Caglayan, J. Chen, G. Xing, C. Zheng, V. Doan-Nguyen, Y. Kang, N. Engheta, C. R. Kagan and C. B. Murray, ACS Nano, 2012, 6, 2804–2817.
25. L. Tong, Y. Zhao, T. B. Huff, M. N. Hansen, A. Wei and J.-X. Cheng, Adv. Mater., 2007, 19, 3136–3141.
26. A. L. Oldenburg, M. N. Hansen, T. S. Ralston and A. Wei, J. Mater. Chem., 2009, 19, 6407–6411.
27. H. Wang, T. B. Huff, D. A. Zweifel, W. He, P. S. Low and A. Wei, Proc. Natl. Acad. Sci. U. S. A., 2005, 102, 15752–15756.
28. S. Link, M. B. Mohamed and M. A. El-Sayed, J. Phys. Chem. B, 1999, 103, 3073–3077.
29. S. Link, M. B. Mohamed and M. A. El-Sayed, J. Phys. Chem. B, 2005, 109, 10531–10532.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra05979h

---