4184
OPTICS LETTERS / Vol. 39, No. 14 / July 15, 2014
Si/ZnO nanocomb arrays decorated with Ag nanoparticles for highly efficient surface-enhanced Raman scattering Hong Jun Yin,1 Yu Fei Chan,1 Zheng Long Wu,2 and Hai Jun Xu1,* 1
School of Science, Beijing University of Chemical Technology, Beijing 100029, China 2
Analytical and Testing Center, Beijing Normal University, Beijing 100875, China *Corresponding author: [email protected] Received April 14, 2014; revised May 29, 2014; accepted June 10, 2014; posted June 12, 2014 (Doc. ID 209744); published July 10, 2014
High-density ZnO nanocombs were first grown on a nanoporous silicon pillar array, and pre-prepared 3D Si/ZnO/ Ag nanocomb arrays were employed as substrates for surface-enhanced Raman scattering (SERS). The finitedifference time-domain simulation result shows that two kinds of inter-Ag–NP nanogaps in the geometry create a large number of SERS “hot spots,” which contributes to the detection limits for rhodamine–6G as low as 10−12 M and the Raman enhancement factor as large as 109 . The linear dependence between the Raman peak intensities and the concentrations of thiram provides a new calibration method for rapid and quantitative detection of trace organic molecules. © 2014 Optical Society of America OCIS codes: (290.5860) Scattering, Raman; (300.6450) Spectroscopy, Raman; (160.4236) Nanomaterials; (250.5403) Plasmonics. http://dx.doi.org/10.1364/OL.39.004184
Surface-enhanced Raman scattering (SERS) has attracted much attention in the past several decades due to its high sensitivity and stability in trace chemical and biomolecular detection [1,2]. Two main mechanisms have been postulated for the SERS effect: electromagnetic enhancement (EM) and chemical enhancement (CM), with the former being mainly due to the EM resonant excitation of localized surface plasmons and the latter the interaction between the organic molecules and their proximal metallic structures [3,4]. It is widely accepted that the long-range EM mechanism (of the order 104 –106 ) plays a much greater role in SERS enhancement than the short-range CM mechanism (of the order 102 ) [4], particularly in most noble metal-based SERS systems [2–6]. Additionally, compared with the noble metal Au, Ag with an extremely strong and tunable surface plasmon resonance (SPR) from visible and near-infrared spectral regions, which could produce highly localized surface plasmons on its surface, has spurred efforts to fabricate more effective SERS-active substrates [7,8]. Therefore, trying to enhance the surface local electromagnetic field of Ag nanostructures is a most effective way to improve the enhancement performance of SERS [2,3,6]. To further improve the SERS effect, recent attempts have focused on three-dimensional (3D) nanostructures with nanogap-rich metal nanoparticles (NPs) that can generate a high-density “hot spot” within a detection volume [9,10]. For these reasons, a number of promising 3D Ag–NP-based SERS substrates have been proposed, such as Ag–NP-decorated Si nanowire/nanorod arrays [6,11], TiO2 nanowire arrays, and SnO2 nanowire arrays [12,13]. These substrates were focused to provide highly sensitive detection. However, considering that the 3D spatial gaps between the neighbor nanowires were too large to work efficiently as SERS “hot spots,” only Ag–NP nanogaps from the side surface of the same nanowire could act as efficient “hot spots” for SERS enhancement [12,13]. Compared with the nanowire structures, the Ag– NP-decorated nanocomb (NC) arrays comprise several 0146-9592/14/144184-04$15.00/0
nanowires arranged in a row, which confer the advantages of nanowires, as well as a close arrangement of NC teeth, which can establish the electromagnetic field link between NPs attached on two adjacent teeth. Thus, it is proposed that Ag–NP-decorated nanocomb (NC) arrays could achieve more sensitive SERS substrate loading, with high-density “hot spots” in 3D geometry [14]. In this Letter, we selected ZnO–NC arrays that were first grown on a nanoporous Si pillar array (NSPA) template to fabricate 3D Ag–NP-decorated SERS substrates. Compared with the substrates based on Ag–NP films, the NC nanostructures have a much higher specific surface area, thus allowing an increased loading of metal NPs and adsorption of a large number of target molecules, as well as the high possibility of forming 3D plasmon “hot spots” [15]. NSPA, which is a regular micron/nanometer Si pillar array, was prepared by hydrothermally etching (100) oriented p-type Si wafers (ρ ∼ 0.1 Ωcm) in a solution of hydrofluoric acid containing ferric nitrate at 160°C for 20 min [16,17]. ZnO–NCs grown on NSPA were synthetized by chemical vapor deposition (CVD) method [17]. By controlling the growth conditions, various ZnO–NCs with different morphologies could be obtained [18]. The large surface area of the formed ZnO–NCs was then coated with Ag–NPs for SERS application. First, the NSPA/ZnO–NC template was immersed into 0.01 M NaBH4 solution and shaken for 3 s. It was taken out and immersed into 0.01 M AgNO3 and HF (3%) solution and shaken for another 3 s. The previous processes were repeated three times. The growth process of Ag–NPs is under nonequilibrium conditions and can be explained by the self-assembled localized microscopic electrochemical cell model [19]. The final samples were copiously rinsed with deionized water and dried under a helium flow at room temperature. The Raman measurements were performed on the LabRAM ARAMIS Raman system using the He–Ne 633 nm line laser as excitation, which could decrease the fluorescent background of © 2014 Optical Society of America
July 15, 2014 / Vol. 39, No. 14 / OPTICS LETTERS
R6G. The diameter of the light spot area was ∼2 μm while the spectral resolution and the incident power were 1 cm−1 and 35 mW, respectively. The spectra were recorded with an accumulation time of 1 s and cycle twice. The same accumulation times and the laser power were used to obtain all the Raman spectra. From the tilted-view scanning electron microscopy (SEM) image of NSPA [Fig. 1(a)], the regular Si pillars are oriented perpendicular to the substrate surface with good uniformity and provide a 3D array substrate for ZnO–NC growth. After the CVD growth of ZnO–NCs, the initially smooth Si pillars branch out, forming flowerlike nanostructures. Figures 1(b)–1(d) show the different magnified SEM images of such structures. The great mass of ZnO–NCs stands on the Si pillar surfaces, and branch out to fill the spaces among the Si pillars. From the magnified SEM image [Fig. 1(d)], the average diameter and length of NC teeth are determined to be ∼45 nm and ∼1 μm, respectively, and the space between adjacent teeth is generally less than 50 nm. The above diameter, length, and space are controllable by adjusting the CVD growth time and temperature. The large aspect ratios of ZnO–NC arrays may lead to a large surface-to-volume ratio, which is of great significance for the adsorption of Ag–NPs. Figure 1(e) shows the representative and magnified SEM image of an individual ZnO–NC whose surface is decorated with Ag–NPs. Clearly, Ag–NPs are successfully deposited onto the whole surface of the
Fig. 1. (a) 45° tilted-view SEM image of NSPA. (b)–(d) Magnified SEM images of such structures. (e) Typical SEM and (f) TEM images of an individual ZnO–NC decorated with Ag–NPs. (g) HRTEM images of a single Ag–NP attached to ZnO–NC tooth and the inset the FFT pattern corresponding to the field given in (g). (h) XRD patterns of arrays.
4185
ZnO–NCs and a granule membrane is formed, which is further proved by the transmission electron microscope (TEM) result [Fig. 1(f)]. Judging from the TEM image, the average diameter of the Ag–NPs is nearly 10–20 nm. Figure 1(g) shows the high-resolution TEM (HRTEM) image of an individual ZnO–NC tooth decorated with one Ag–NP. The lattice fringe spacing of Ag is ∼0.24 nm, which compares well with the interplanar spacing of (111) plane of face-centered-cubic (fcc) Ag crystals. The spacing of ZnO is ∼0.28 nm, which corresponds to (100) plane of hexagonal ZnO crystals. To determine the phase of crystalline zones, electronic diffraction measurement with two-dimensional fast Fourier-transform (FFT) mode was carried out at the typical area shown in Fig. 1(g). The result, presented in the inset of Fig. 1(g), proves that the spots correspond to the diffraction from the (111) crystal planes of fcc Ag and the (100) crystal plane of hexagonal phase ZnO. As shown in Fig. 1(h), the XRD result also shows that Ag NPs can be determined to be the fcc structure and the crystal phase of ZnO–NC is confirmed to be a hexagonal wurtzite structure. In the experiment, rhodamine–6G (R6G) dye was used as a probe molecule to reveal the SERS sensitivity of NSPA/ZnO/Ag–NC arrays. As shown in Fig. 2(a), the strong Raman bands were clearly observed in R6G solutions with different concentrations from 10−3 to 10−11 M. The SERS enhancement factor (EF) could be estimated by comparing the measured intensities of R6G molecules on NSPA/ZnO/Ag and NSPA/ZnO, following the standard equation [11,20]: EF I SERS N Raman P Raman T Raman ∕ I Raman N SERS P SERS T SERS I SERS N Raman ∕I Raman N SERS , where I, N, P, and T represent the Raman intensity, the number of probe molecules, the laser power, and the acquisition time, respectively. Here, we carefully kept the experimental conditions all the same to make the EF estimation as reliable as possible. Choosing the Raman bands centered at 1510 cm−1 , and based on the average integrated peak intensity from 15 randomly selected spots over three different SERS substrates, the EF of
Fig. 2. (a) SERS spectra of R6G obtained at 10 different concentrations from 10−12 to 10−3 M using substrate. (b) Thirty-nine SERS curves of 10−5 M R6G molecules collected on two crossed lines shown in the inset of (b) (X and Y) and the corresponding RSD curve, and the inset the SERS mapping of substrate. (c) SERS spectra of thiram obtained at different concentrations from 10−12 to 10−4 M. (d) Linear relationship between logI of the band peaking at 1384 cm−1 as a function of logC, based on the SERS data of thiram from (c).
4186
OPTICS LETTERS / Vol. 39, No. 14 / July 15, 2014
NSPA/ZnO/Ag–NC arrays for 10−11 M R6G compared with 10−3 M R6G on NSPA/ZnO–NC arrays was finally estimated to be ∼1.36 × 109 , with a variation of 11.3%, and the same magnitude for EF was obtained with other fingerprint Raman peaks at 611, 774, 1361, and 1648 cm−1 , which is much higher than other 3D Ag–NPdecorated structures [21]. To test whether NSPA/ZnO/ Ag–NC arrays are able to give repeatable signals, the Raman mapping image of 10−5 M R6G signal integral intensity for the peak centered at ∼611 cm−1 is given in the inset of Fig. 2(b). Here, a 1 μm step size and 20 × 20 400 spectral lines in all are adopted, and the brightness of the grid is proportional to the signal integral intensity. Figure 2(b) is the 39 sectional views of two crossed lines X and Y in the inset of Fig. 2(b), which distinctly reveals the spectral intensity along two lines. The relative standard deviation (RSD) values of Raman peaks are noticed to be as low as 0.4 [Fig. 2(b)], and especially the values at the peaks were not higher than that at other points. These results showed the NSPA/ZnO/Ag–NC arrays own good uniformity and reproducibility across the entire area. Taking the advantages of high sensitivity and good reproducibility of the present substrate for SERS detection, we applied it to detect thiram, which is a typical pesticide and is usually used to deal with weeds, apple anthrax, and verticillium of alfalfa [21]. According to United States Food Standards, the maximum residue limit of thiram in most farm produce is limited to about 0.8 ppm. Therefore, a rational and convenient approach to detect thiram with high sensitivity is essential. The SERS spectra collected from the substrates after being immersed in thiram aqueous solutions with 10−4 to 10−12 M concentrations are shown in Fig. 2(c). Clearly, the SERS sensitivity of as-prepared 3D NC arrays for thiram is so high that the characteristic Raman bands of 10−12 M (2.4 × 10−7 ppm) could be clearly identified. Moreover, based on the measured Raman peaks with different thiram concentrations from 10−4 to 10−12 M, a linear dependence between the logarithmic integrated signal intensity (logI) of the band centered at 1384 cm−1 with the logarithmic concentration (logC) was found [Fig. 2(d), data were averaged over 15 randomly selected positions]. This linear curve provided a new method for rapid and quantitative detection of unknown (even trace) concentrations of thiram. Taken together, it is, therefore, reasonable to conclude that the NSPA/ZnO/Ag–NC arrays have good uniform and high SERS activity, with potential as effective SERS substrates for the rapid and quantitative detection of trace organic molecules. To understand the SERS response, the spatial distributions of the electromagnetic field intensity for the NC arrays were simulated with the 3D finite-difference timedomain (FDTD) method, using periodic boundary conditions [11]. Figure 3(a) presents the profile of the local model of ZnO/Ag–NC structure, where a 15 nm Ag–NPcoated 50 nm diameter ZnO nanowire and a 35 nm inter-wire spacing were adopted. A rectangular-shaped continuous wave laser with 633 nm wavelength propagating along the k direction was input into the structure with its polarization direction perpendicular to k. As shown in Figs. 3(b) and 3(c), which, respectively, exhibit the calculated spatial distributions of the electric field intensities for the x–y and y–z planes in Fig. 3(a), two
Fig. 3. FDTD modeling of NSPA/ZnO/Ag–NC array. (a) Shape of the model. (b) and (c) Calculated spatial istributions of the electric field intensity for the x–y and y–z planes in (a).
typical kinds of “hot spots” are formed: one existing between the adjacent Ag–NPs on the side surface of one single ZnO–NC tooth and the other stemming from Ag–NPs on two adjacent ZnO–NC teeth. The local electric fields for two kinds “hot spots” have arrived at the maxima of 95.2 V m−1 and 105.7 V m−1 , respectively. They give EFs of about 8.2 × 107 and 1.2 × 108 according to the relationship of the Raman enhancement scales, roughly as the fourth power of the local field [22]: GSERS ≈ jE loc ωexc ∕E inc ωexc j4 , where E loc ωexc and E inc ωexc are the E and E 0 in the FDTD calculations, respectively, which approaches the experimental value of ∼1.36 × 109 for 10−11 M R6G on NSPA/ZnO/Ag compared with 10−3 M R6G on NSPA/ZnO. The FDTD calculations clearly show that the SERS enhancement of the substrate is mainly due to the extremely strong electric fields at the inter-Ag−NP nanogaps and, therefore, can be explained by the EM mechanism. In conclusion, NSPA/ZnO–NC arrays decorated with Ag–NPs were investigated as potentially effective 3D SERS substrates, where the NC nanostructures generate two kinds of inter-Ag–NP nanogaps contributing to the high SERS sensitivity as “hot spots.” The NSPA/ZnO/ Ag–NC array substrate yielded extremely high SERS activity, with a calculated EF up to 109 for 10−11 M R6G compared with 10−3 M R6G on NSPA/ZnO–NC arrays, and pushed the detection limit down to 10−12 M for both R6G and thiram. The best enhancement is in good agreement with the results of FDTD simulations and, therefore, can be considered as an EM resonant excitation of localized surface plasmons. Furthermore, the linear dependence between the SERS signal intensities and the concentrations of thiram provided a new calibration method for rapid and quantitative detection of organic molecules. This will open up an opportunity to use NSPA/ZnO/Ag–NC array substrates for trace toxin and/ or molecule(s) detection in various fields, ranging from food safety and environmental health to criminology. This work was supported by the Natural Science Foundation of China (11104008), the Beijing Natural Science Foundation (4142040), and the Beijing Higher Education Young Elite Teacher Project. References 1. M. Mulvihill, A. Tao, K. Benjauthrit, J. Arnold, and P. Yang, Angew. Chem. 47, 6456 (2008). 2. K. Kneipp, H. Kneipp, I. Itzkan, R. R. Dasari, and M. S. Feld, J. Phys. Condens. Matter 14, R597 (2002).
July 15, 2014 / Vol. 39, No. 14 / OPTICS LETTERS 3. K. Kneipp, Y. Wang, H. Kneipp, L. T. Perelman, I. Itzkan, R. R. Dasari, and M. S. Feld, Phys. Rev. Lett. 78, 1667 (1997). 4. H. X. Xu, E. J. Bjerneld, M. Käll, and L. Börjesson, Phys. Rev. Lett. 83, 4357 (1999). 5. Z. G. Dai, X. H. Xiao, L. Liao, J. F. Zheng, F. Mei, W. Wu, J. J. Ying, F. Ren, and C. Z. Jiang, Appl. Phys. Lett. 103, 041903 (2013). 6. J. A. Huang, Y. Q. Zhao, X. J. Zhang, L. F. He, T. L. Wong, Y. S. Chui, W. J. Zhang, and S. T. Lee, Nano Lett. 13, 5039 (2013). 7. M. S. Schmidt, J. Hubner, and A. Boisen, Adv. Mater. 24, 11 (2012). 8. W. Song, X. Han, L. Chen, Y. Yang, B. Tang, W. Ji, W. Ruan, W. Xu, B. Zhao, and Y. Ozaki, J. Raman Spectrosc. 41, 907 (2010). 9. J. A. Dieringer, K. L. Wustholz, D. J. Masiello, J. P. Camden, S. L. Kleinman, G. C. Schatz, and R. P. Van Duyne, J. Am. Chem. Soc. 131, 849 (2009). 10. A. Musumeci, D. Gosztola, T. Schiller, D. N. Mimitrijevic, V. Mujica, D. Martin, and T. Rajh, J. Am. Chem. Soc. 131, 6040 (2009).
4187
11. C. X. Zhang, L. Su, Y. F. Chan, Z. L. Wu, Y. M. Zhao, H. J. Xu, and X. M. Sun, Nanotechnology 24, 335501 (2013). 12. Y. K. Lai, L. X. Lin, F. Pan, J. Y. Huang, R. Song, Y. X. Huang, C. J. Lin, H. Fuchs, and L. F. Chi, Small 9, 2945 (2013). 13. Y. F. Chan, H. J. Xu, L. Cao, Y. Tang, D. Y. Li, and X. M. Sun, J. Appl. Phys. 111, 033104 (2012). 14. Y. J. Liu, Z. Y. Zhang, Q. Zhao, R. A. Dluhy, and Y. P. Zhao, Appl. Phys. Lett. 94, 033103 (2009). 15. A. Reina, X. T. Jia, J. Ho, D. Nezich, H. Son, V. Bulovic, M. S. Dresselhaus, and J. Kong, Nano Lett. 9, 30 (2009). 16. H. J. Xu and X. J. Li, Opt. Express 16, 2933 (2008). 17. Y. F. Chan, W. Su, C. X. Zhang, Z. L. Wu, Y. Tang, X. Q. Sun, and H. J. Xu, Opt. Express 20, 24280 (2012). 18. T. L. Phan, Y. K. Sun, and R. Vincent, J. Korean Phys. Soc. 59, 60 (2011). 19. X. Wang, W. Shi, G. She, and L. Mu, J. Am. Chem. Soc. 133, 16518 (2011). 20. R. Liu, J. Liu, X. Zhou, M. Sun, and G. Jiang, Anal. Chem. 83, 9131 (2011). 21. J. Parisi, L. Su, and Y. Lei, Lab Chip 13, 1501 (2013). 22. F. J. Garcia Vidal and J. B. Pendry, Phys. Rev. Lett. 77, 1163 (1996).
OPTICS LETTERS / Vol. 39, No. 14 / July 15, 2014
Si/ZnO nanocomb arrays decorated with Ag nanoparticles for highly efficient surface-enhanced Raman scattering Hong Jun Yin,1 Yu Fei Chan,1 Zheng Long Wu,2 and Hai Jun Xu1,* 1
School of Science, Beijing University of Chemical Technology, Beijing 100029, China 2
Analytical and Testing Center, Beijing Normal University, Beijing 100875, China *Corresponding author: [email protected] Received April 14, 2014; revised May 29, 2014; accepted June 10, 2014; posted June 12, 2014 (Doc. ID 209744); published July 10, 2014
High-density ZnO nanocombs were first grown on a nanoporous silicon pillar array, and pre-prepared 3D Si/ZnO/ Ag nanocomb arrays were employed as substrates for surface-enhanced Raman scattering (SERS). The finitedifference time-domain simulation result shows that two kinds of inter-Ag–NP nanogaps in the geometry create a large number of SERS “hot spots,” which contributes to the detection limits for rhodamine–6G as low as 10−12 M and the Raman enhancement factor as large as 109 . The linear dependence between the Raman peak intensities and the concentrations of thiram provides a new calibration method for rapid and quantitative detection of trace organic molecules. © 2014 Optical Society of America OCIS codes: (290.5860) Scattering, Raman; (300.6450) Spectroscopy, Raman; (160.4236) Nanomaterials; (250.5403) Plasmonics. http://dx.doi.org/10.1364/OL.39.004184
Surface-enhanced Raman scattering (SERS) has attracted much attention in the past several decades due to its high sensitivity and stability in trace chemical and biomolecular detection [1,2]. Two main mechanisms have been postulated for the SERS effect: electromagnetic enhancement (EM) and chemical enhancement (CM), with the former being mainly due to the EM resonant excitation of localized surface plasmons and the latter the interaction between the organic molecules and their proximal metallic structures [3,4]. It is widely accepted that the long-range EM mechanism (of the order 104 –106 ) plays a much greater role in SERS enhancement than the short-range CM mechanism (of the order 102 ) [4], particularly in most noble metal-based SERS systems [2–6]. Additionally, compared with the noble metal Au, Ag with an extremely strong and tunable surface plasmon resonance (SPR) from visible and near-infrared spectral regions, which could produce highly localized surface plasmons on its surface, has spurred efforts to fabricate more effective SERS-active substrates [7,8]. Therefore, trying to enhance the surface local electromagnetic field of Ag nanostructures is a most effective way to improve the enhancement performance of SERS [2,3,6]. To further improve the SERS effect, recent attempts have focused on three-dimensional (3D) nanostructures with nanogap-rich metal nanoparticles (NPs) that can generate a high-density “hot spot” within a detection volume [9,10]. For these reasons, a number of promising 3D Ag–NP-based SERS substrates have been proposed, such as Ag–NP-decorated Si nanowire/nanorod arrays [6,11], TiO2 nanowire arrays, and SnO2 nanowire arrays [12,13]. These substrates were focused to provide highly sensitive detection. However, considering that the 3D spatial gaps between the neighbor nanowires were too large to work efficiently as SERS “hot spots,” only Ag–NP nanogaps from the side surface of the same nanowire could act as efficient “hot spots” for SERS enhancement [12,13]. Compared with the nanowire structures, the Ag– NP-decorated nanocomb (NC) arrays comprise several 0146-9592/14/144184-04$15.00/0
nanowires arranged in a row, which confer the advantages of nanowires, as well as a close arrangement of NC teeth, which can establish the electromagnetic field link between NPs attached on two adjacent teeth. Thus, it is proposed that Ag–NP-decorated nanocomb (NC) arrays could achieve more sensitive SERS substrate loading, with high-density “hot spots” in 3D geometry [14]. In this Letter, we selected ZnO–NC arrays that were first grown on a nanoporous Si pillar array (NSPA) template to fabricate 3D Ag–NP-decorated SERS substrates. Compared with the substrates based on Ag–NP films, the NC nanostructures have a much higher specific surface area, thus allowing an increased loading of metal NPs and adsorption of a large number of target molecules, as well as the high possibility of forming 3D plasmon “hot spots” [15]. NSPA, which is a regular micron/nanometer Si pillar array, was prepared by hydrothermally etching (100) oriented p-type Si wafers (ρ ∼ 0.1 Ωcm) in a solution of hydrofluoric acid containing ferric nitrate at 160°C for 20 min [16,17]. ZnO–NCs grown on NSPA were synthetized by chemical vapor deposition (CVD) method [17]. By controlling the growth conditions, various ZnO–NCs with different morphologies could be obtained [18]. The large surface area of the formed ZnO–NCs was then coated with Ag–NPs for SERS application. First, the NSPA/ZnO–NC template was immersed into 0.01 M NaBH4 solution and shaken for 3 s. It was taken out and immersed into 0.01 M AgNO3 and HF (3%) solution and shaken for another 3 s. The previous processes were repeated three times. The growth process of Ag–NPs is under nonequilibrium conditions and can be explained by the self-assembled localized microscopic electrochemical cell model [19]. The final samples were copiously rinsed with deionized water and dried under a helium flow at room temperature. The Raman measurements were performed on the LabRAM ARAMIS Raman system using the He–Ne 633 nm line laser as excitation, which could decrease the fluorescent background of © 2014 Optical Society of America
July 15, 2014 / Vol. 39, No. 14 / OPTICS LETTERS
R6G. The diameter of the light spot area was ∼2 μm while the spectral resolution and the incident power were 1 cm−1 and 35 mW, respectively. The spectra were recorded with an accumulation time of 1 s and cycle twice. The same accumulation times and the laser power were used to obtain all the Raman spectra. From the tilted-view scanning electron microscopy (SEM) image of NSPA [Fig. 1(a)], the regular Si pillars are oriented perpendicular to the substrate surface with good uniformity and provide a 3D array substrate for ZnO–NC growth. After the CVD growth of ZnO–NCs, the initially smooth Si pillars branch out, forming flowerlike nanostructures. Figures 1(b)–1(d) show the different magnified SEM images of such structures. The great mass of ZnO–NCs stands on the Si pillar surfaces, and branch out to fill the spaces among the Si pillars. From the magnified SEM image [Fig. 1(d)], the average diameter and length of NC teeth are determined to be ∼45 nm and ∼1 μm, respectively, and the space between adjacent teeth is generally less than 50 nm. The above diameter, length, and space are controllable by adjusting the CVD growth time and temperature. The large aspect ratios of ZnO–NC arrays may lead to a large surface-to-volume ratio, which is of great significance for the adsorption of Ag–NPs. Figure 1(e) shows the representative and magnified SEM image of an individual ZnO–NC whose surface is decorated with Ag–NPs. Clearly, Ag–NPs are successfully deposited onto the whole surface of the
Fig. 1. (a) 45° tilted-view SEM image of NSPA. (b)–(d) Magnified SEM images of such structures. (e) Typical SEM and (f) TEM images of an individual ZnO–NC decorated with Ag–NPs. (g) HRTEM images of a single Ag–NP attached to ZnO–NC tooth and the inset the FFT pattern corresponding to the field given in (g). (h) XRD patterns of arrays.
4185
ZnO–NCs and a granule membrane is formed, which is further proved by the transmission electron microscope (TEM) result [Fig. 1(f)]. Judging from the TEM image, the average diameter of the Ag–NPs is nearly 10–20 nm. Figure 1(g) shows the high-resolution TEM (HRTEM) image of an individual ZnO–NC tooth decorated with one Ag–NP. The lattice fringe spacing of Ag is ∼0.24 nm, which compares well with the interplanar spacing of (111) plane of face-centered-cubic (fcc) Ag crystals. The spacing of ZnO is ∼0.28 nm, which corresponds to (100) plane of hexagonal ZnO crystals. To determine the phase of crystalline zones, electronic diffraction measurement with two-dimensional fast Fourier-transform (FFT) mode was carried out at the typical area shown in Fig. 1(g). The result, presented in the inset of Fig. 1(g), proves that the spots correspond to the diffraction from the (111) crystal planes of fcc Ag and the (100) crystal plane of hexagonal phase ZnO. As shown in Fig. 1(h), the XRD result also shows that Ag NPs can be determined to be the fcc structure and the crystal phase of ZnO–NC is confirmed to be a hexagonal wurtzite structure. In the experiment, rhodamine–6G (R6G) dye was used as a probe molecule to reveal the SERS sensitivity of NSPA/ZnO/Ag–NC arrays. As shown in Fig. 2(a), the strong Raman bands were clearly observed in R6G solutions with different concentrations from 10−3 to 10−11 M. The SERS enhancement factor (EF) could be estimated by comparing the measured intensities of R6G molecules on NSPA/ZnO/Ag and NSPA/ZnO, following the standard equation [11,20]: EF I SERS N Raman P Raman T Raman ∕ I Raman N SERS P SERS T SERS I SERS N Raman ∕I Raman N SERS , where I, N, P, and T represent the Raman intensity, the number of probe molecules, the laser power, and the acquisition time, respectively. Here, we carefully kept the experimental conditions all the same to make the EF estimation as reliable as possible. Choosing the Raman bands centered at 1510 cm−1 , and based on the average integrated peak intensity from 15 randomly selected spots over three different SERS substrates, the EF of
Fig. 2. (a) SERS spectra of R6G obtained at 10 different concentrations from 10−12 to 10−3 M using substrate. (b) Thirty-nine SERS curves of 10−5 M R6G molecules collected on two crossed lines shown in the inset of (b) (X and Y) and the corresponding RSD curve, and the inset the SERS mapping of substrate. (c) SERS spectra of thiram obtained at different concentrations from 10−12 to 10−4 M. (d) Linear relationship between logI of the band peaking at 1384 cm−1 as a function of logC, based on the SERS data of thiram from (c).
4186
OPTICS LETTERS / Vol. 39, No. 14 / July 15, 2014
NSPA/ZnO/Ag–NC arrays for 10−11 M R6G compared with 10−3 M R6G on NSPA/ZnO–NC arrays was finally estimated to be ∼1.36 × 109 , with a variation of 11.3%, and the same magnitude for EF was obtained with other fingerprint Raman peaks at 611, 774, 1361, and 1648 cm−1 , which is much higher than other 3D Ag–NPdecorated structures [21]. To test whether NSPA/ZnO/ Ag–NC arrays are able to give repeatable signals, the Raman mapping image of 10−5 M R6G signal integral intensity for the peak centered at ∼611 cm−1 is given in the inset of Fig. 2(b). Here, a 1 μm step size and 20 × 20 400 spectral lines in all are adopted, and the brightness of the grid is proportional to the signal integral intensity. Figure 2(b) is the 39 sectional views of two crossed lines X and Y in the inset of Fig. 2(b), which distinctly reveals the spectral intensity along two lines. The relative standard deviation (RSD) values of Raman peaks are noticed to be as low as 0.4 [Fig. 2(b)], and especially the values at the peaks were not higher than that at other points. These results showed the NSPA/ZnO/Ag–NC arrays own good uniformity and reproducibility across the entire area. Taking the advantages of high sensitivity and good reproducibility of the present substrate for SERS detection, we applied it to detect thiram, which is a typical pesticide and is usually used to deal with weeds, apple anthrax, and verticillium of alfalfa [21]. According to United States Food Standards, the maximum residue limit of thiram in most farm produce is limited to about 0.8 ppm. Therefore, a rational and convenient approach to detect thiram with high sensitivity is essential. The SERS spectra collected from the substrates after being immersed in thiram aqueous solutions with 10−4 to 10−12 M concentrations are shown in Fig. 2(c). Clearly, the SERS sensitivity of as-prepared 3D NC arrays for thiram is so high that the characteristic Raman bands of 10−12 M (2.4 × 10−7 ppm) could be clearly identified. Moreover, based on the measured Raman peaks with different thiram concentrations from 10−4 to 10−12 M, a linear dependence between the logarithmic integrated signal intensity (logI) of the band centered at 1384 cm−1 with the logarithmic concentration (logC) was found [Fig. 2(d), data were averaged over 15 randomly selected positions]. This linear curve provided a new method for rapid and quantitative detection of unknown (even trace) concentrations of thiram. Taken together, it is, therefore, reasonable to conclude that the NSPA/ZnO/Ag–NC arrays have good uniform and high SERS activity, with potential as effective SERS substrates for the rapid and quantitative detection of trace organic molecules. To understand the SERS response, the spatial distributions of the electromagnetic field intensity for the NC arrays were simulated with the 3D finite-difference timedomain (FDTD) method, using periodic boundary conditions [11]. Figure 3(a) presents the profile of the local model of ZnO/Ag–NC structure, where a 15 nm Ag–NPcoated 50 nm diameter ZnO nanowire and a 35 nm inter-wire spacing were adopted. A rectangular-shaped continuous wave laser with 633 nm wavelength propagating along the k direction was input into the structure with its polarization direction perpendicular to k. As shown in Figs. 3(b) and 3(c), which, respectively, exhibit the calculated spatial distributions of the electric field intensities for the x–y and y–z planes in Fig. 3(a), two
Fig. 3. FDTD modeling of NSPA/ZnO/Ag–NC array. (a) Shape of the model. (b) and (c) Calculated spatial istributions of the electric field intensity for the x–y and y–z planes in (a).
typical kinds of “hot spots” are formed: one existing between the adjacent Ag–NPs on the side surface of one single ZnO–NC tooth and the other stemming from Ag–NPs on two adjacent ZnO–NC teeth. The local electric fields for two kinds “hot spots” have arrived at the maxima of 95.2 V m−1 and 105.7 V m−1 , respectively. They give EFs of about 8.2 × 107 and 1.2 × 108 according to the relationship of the Raman enhancement scales, roughly as the fourth power of the local field [22]: GSERS ≈ jE loc ωexc ∕E inc ωexc j4 , where E loc ωexc and E inc ωexc are the E and E 0 in the FDTD calculations, respectively, which approaches the experimental value of ∼1.36 × 109 for 10−11 M R6G on NSPA/ZnO/Ag compared with 10−3 M R6G on NSPA/ZnO. The FDTD calculations clearly show that the SERS enhancement of the substrate is mainly due to the extremely strong electric fields at the inter-Ag−NP nanogaps and, therefore, can be explained by the EM mechanism. In conclusion, NSPA/ZnO–NC arrays decorated with Ag–NPs were investigated as potentially effective 3D SERS substrates, where the NC nanostructures generate two kinds of inter-Ag–NP nanogaps contributing to the high SERS sensitivity as “hot spots.” The NSPA/ZnO/ Ag–NC array substrate yielded extremely high SERS activity, with a calculated EF up to 109 for 10−11 M R6G compared with 10−3 M R6G on NSPA/ZnO–NC arrays, and pushed the detection limit down to 10−12 M for both R6G and thiram. The best enhancement is in good agreement with the results of FDTD simulations and, therefore, can be considered as an EM resonant excitation of localized surface plasmons. Furthermore, the linear dependence between the SERS signal intensities and the concentrations of thiram provided a new calibration method for rapid and quantitative detection of organic molecules. This will open up an opportunity to use NSPA/ZnO/Ag–NC array substrates for trace toxin and/ or molecule(s) detection in various fields, ranging from food safety and environmental health to criminology. This work was supported by the Natural Science Foundation of China (11104008), the Beijing Natural Science Foundation (4142040), and the Beijing Higher Education Young Elite Teacher Project. References 1. M. Mulvihill, A. Tao, K. Benjauthrit, J. Arnold, and P. Yang, Angew. Chem. 47, 6456 (2008). 2. K. Kneipp, H. Kneipp, I. Itzkan, R. R. Dasari, and M. S. Feld, J. Phys. Condens. Matter 14, R597 (2002).
July 15, 2014 / Vol. 39, No. 14 / OPTICS LETTERS 3. K. Kneipp, Y. Wang, H. Kneipp, L. T. Perelman, I. Itzkan, R. R. Dasari, and M. S. Feld, Phys. Rev. Lett. 78, 1667 (1997). 4. H. X. Xu, E. J. Bjerneld, M. Käll, and L. Börjesson, Phys. Rev. Lett. 83, 4357 (1999). 5. Z. G. Dai, X. H. Xiao, L. Liao, J. F. Zheng, F. Mei, W. Wu, J. J. Ying, F. Ren, and C. Z. Jiang, Appl. Phys. Lett. 103, 041903 (2013). 6. J. A. Huang, Y. Q. Zhao, X. J. Zhang, L. F. He, T. L. Wong, Y. S. Chui, W. J. Zhang, and S. T. Lee, Nano Lett. 13, 5039 (2013). 7. M. S. Schmidt, J. Hubner, and A. Boisen, Adv. Mater. 24, 11 (2012). 8. W. Song, X. Han, L. Chen, Y. Yang, B. Tang, W. Ji, W. Ruan, W. Xu, B. Zhao, and Y. Ozaki, J. Raman Spectrosc. 41, 907 (2010). 9. J. A. Dieringer, K. L. Wustholz, D. J. Masiello, J. P. Camden, S. L. Kleinman, G. C. Schatz, and R. P. Van Duyne, J. Am. Chem. Soc. 131, 849 (2009). 10. A. Musumeci, D. Gosztola, T. Schiller, D. N. Mimitrijevic, V. Mujica, D. Martin, and T. Rajh, J. Am. Chem. Soc. 131, 6040 (2009).
4187
11. C. X. Zhang, L. Su, Y. F. Chan, Z. L. Wu, Y. M. Zhao, H. J. Xu, and X. M. Sun, Nanotechnology 24, 335501 (2013). 12. Y. K. Lai, L. X. Lin, F. Pan, J. Y. Huang, R. Song, Y. X. Huang, C. J. Lin, H. Fuchs, and L. F. Chi, Small 9, 2945 (2013). 13. Y. F. Chan, H. J. Xu, L. Cao, Y. Tang, D. Y. Li, and X. M. Sun, J. Appl. Phys. 111, 033104 (2012). 14. Y. J. Liu, Z. Y. Zhang, Q. Zhao, R. A. Dluhy, and Y. P. Zhao, Appl. Phys. Lett. 94, 033103 (2009). 15. A. Reina, X. T. Jia, J. Ho, D. Nezich, H. Son, V. Bulovic, M. S. Dresselhaus, and J. Kong, Nano Lett. 9, 30 (2009). 16. H. J. Xu and X. J. Li, Opt. Express 16, 2933 (2008). 17. Y. F. Chan, W. Su, C. X. Zhang, Z. L. Wu, Y. Tang, X. Q. Sun, and H. J. Xu, Opt. Express 20, 24280 (2012). 18. T. L. Phan, Y. K. Sun, and R. Vincent, J. Korean Phys. Soc. 59, 60 (2011). 19. X. Wang, W. Shi, G. She, and L. Mu, J. Am. Chem. Soc. 133, 16518 (2011). 20. R. Liu, J. Liu, X. Zhou, M. Sun, and G. Jiang, Anal. Chem. 83, 9131 (2011). 21. J. Parisi, L. Su, and Y. Lei, Lab Chip 13, 1501 (2013). 22. F. J. Garcia Vidal and J. B. Pendry, Phys. Rev. Lett. 77, 1163 (1996).
