Full Surface Embedding of Gold Clusters on Silicon Nanowires for

Feb 24, 2012 - Dong Wook Kwak,. §,∥. Seong Yong Park,. †. Minseok Kim,. †. Jun-Ho Lee,. †. Hyouksoo Han,. †. Sung Heo,. †. Xiang Shu Li,...
4 downloads 0 Views 4MB Size
Letter pubs.acs.org/NanoLett

Full Surface Embedding of Gold Clusters on Silicon Nanowires for Efficient Capture and Photothermal Therapy of Circulating Tumor Cells Gyeong-Su Park,*,†,∥ Hyuksang Kwon,‡,∥ Dong Wook Kwak,§,∥ Seong Yong Park,† Minseok Kim,† Jun-Ho Lee,† Hyouksoo Han,† Sung Heo,† Xiang Shu Li,† Jae Hak Lee,† Young Hwan Kim,† Jeong-Gun Lee,† Woochul Yang,§ Hoon Young Cho,§ Seong Keun Kim,*,‡ and Kinam Kim† †

Samsung Advanced Institute of Technology, San 14−1, Nong-Seo Ri, Ki-Hung Eub, Yong-In Gun, Kyung-Ki Do 449-900, South Korea ‡ Department of Chemistry and WCU Department of Biophysics and Chemical Biology, Seoul National University, Seoul 151-747, South Korea § Department of Physics, Dongguk University, Seoul 100-715, South Korea S Supporting Information *

ABSTRACT: We report on rapid thermal chemical vapor deposition growth of silicon nanowires (Si NWs) that contain a high density of gold nanoclusters (Au NCs) with a uniform coverage over the entire length of the nanowire sidewalls. The Au NC-coated Si NWs with an antibody-coated surface obtain the unique capability to capture breast cancer cells at twice the highest efficiency currently achievable (∼88% at 40 min cell incubation time) from a nanostructured substrate. We also found that irradiation of breast cancer cells captured on Au NCcoated Si NWs with a near-infrared light resulted in a high mortality rate of these cancer cells, raising a fine prospect for simultaneous capture and plasmonic photothermal therapy for circulating tumor cells. KEYWORDS: Si nanowire, RTCVD, gold nanocluster, breast cancer cells, photothermal therapy

S

NCs, and the extent to which NCs couple plasmonically with one another.14,15 The modern understanding of cancer biology and the process of metastasis suggests that tumor cells circulating in patient’s bloodstream play a vital role in forming metastatic disease. Capturing and analyzing the circulating tumor cells (CTCs) in blood can identify malignant progression of cancer and cancer cell mutations to provide physicians with methods for improved cancer diagnosis, prognosis, and treatment. However, efficient capture of CTCs presents a tremendous technical challenge due to the extremely low number density of CTCs in the bloodstream. Recently, silicon-nanopillar arrays16 and carbon nanotube forests17 were reported to exhibit highly efficient capture of CTCs due to their enhanced local topographic interactions with nanoscale cellular surfaces. The capture sensitivity of these techniques still relies on the degree of enrichment of CTCs. Gold nanoparticles functionalized with antibodies were reported to facilitate strong binding with tumor cells,18 which raises the prospect that Si NWs coated with gold

emiconductor NWs are promising nanomaterials for applications in nanoscale optical devices,1,2 solar cells,3 and biosensors4 due to their unique feature of a large active surface area for high optical absorption across a broad spectrum. Of particular interest is deposition of metal nanoparticles on Si NWs5−7 that may attain novel plasmonic properties,8 but deposition of metal NCs of a narrow size distribution on Si NWs in high density and uniform coverage is difficult with conventional growth techniques. Metastable growth techniques such as chemical vapor deposition (CVD)9 and molecular beam epitaxy10 are known to extend the solubility limit of catalytic metals such as Au and lead to a highly nonequilibrium state of supersaturated Au on Si NWs. However, the Si NWs produced this way are often found in short and tapered wire morphologies11 and only locally covered with Au NCs on their sidewalls.12 Recently, Peng et al.13 have demonstrated that heteroepitaxial decoration of Ag nanoparticles on Si NWs can be successfully achieved using a galvanic surface reduction method. There are also other factors that need to be considered to enhance the local electromagnetic fields experienced by the surface metal NCs, such as the size and number density of NCs, the distance between adjacent © 2012 American Chemical Society

Received: December 29, 2011 Revised: February 7, 2012 Published: February 24, 2012 1638

dx.doi.org/10.1021/nl2045759 | Nano Lett. 2012, 12, 1638−1642

Nano Letters

Letter

Figure 1. (a) SEM image of AuSi droplets on a Si substrate after annealing for 10 min at 973K. Inset, cross-sectional TEM image of the AuSi droplets that encroaches upon the Si surface. (b) SEM image of Si NWs on a Si substrate grown by RTCVD. (Inset) Si NW image showing the faceting at a distance of l from the tip. (c) Magnified SEM image of Si NW showing the subfacets and a high density of Au NCs. (d−f) HAADFSTEM images taken at the top (d), middle (e), and bottom (f) parts of the Si NW shown in the inset. The images indicate a uniform and highdensity coverage of Au NCs on Si NW. (g,h) TEM (g) and HAADF-STEM (h) images for a quasi-6-fold cross section of the Si NW, revealing the monolayer nature of the Au NC coverage. The section thickness with an ultramicrotome was about 65 nm.

Figure 2. (a) Typical HAADF-STEM image of the Si NW where Au is more brightly contrasted than Si. (b) Aberration-corrected STEM image of the two-phase boundary area (denoted A1 in a). Inset, intensity profiles along the three linearly marked regions A, B and C showing Au atoms (marked by arrows) in substitutional sites of the Si (440) dumbbell columns. (c) HRTEM image of the NW edge (A2 in a). (Inset) EDS spectrum showing the formation of SiO2 layer on the surface. (d) Low-loss EELS spectra (with a spatial resolution of 1 nm) taken at the Au tip (circle 1) and an Au NC (circle 2) shown in the inset. (e) Bright-field TEM image of a Si NW after thermal annealing at 1073 K. (f) Histogram for the size distribution of the Au nanoparticles after thermal annealing at 1073 K.

NWs grow under a low vacuum condition (see Supporting Information). By heating only the substrate rather than the gas or chamber walls,20 the RTCVD reactor has the advantage of reducing unwanted contaminants in Si NWs that act as a diffusion barrier for Au and thus facilitates the diffusion of Au onto the NW sidewalls during growth.12 The optimum thickness of the deposited Au layer for highdensity Si NW growth was 2−3 nm, because a thicker Au layer led to the short and tapered Si NW growth. When heated to 973 K, the Au layer forms AuSi droplets (Figure 1a), which encroach upon the Si surface and eventually form comparably large droplets upon agglomeration (Figure 1a, inset). Initial nucleation of Si NWs occurs via the AuSi droplets. A typical scanning electron microscopy (SEM) image of the Si NW array

particles may further enhance the efficiency of CTC capture. One may additionally take advantage of the high absorption cross section of near-infrared (NIR) light by gold nanoparticles 19 to use Au NC-coated Si NWs in in situ photothermal therapy for tumor cells on capture. The purpose of this work is to produce Si NWs whose entire surface is uniformly covered with a high-density of Au NCs for a large amount of antibody conjugation (for the epithelial cell adhesion antibody)16 and a strong surface plasmon activity to realize highly effective capture and plasmonic photothermal therapy of tumor cells. We employed rapid thermal chemical vapor deposition (RTCVD) to achieve simultaneous growth of long Si NWs and Au NCs on their surface through rapid diffusion of the liquid AuSi seeds on the Si substrate as the Si 1639

dx.doi.org/10.1021/nl2045759 | Nano Lett. 2012, 12, 1638−1642

Nano Letters

Letter

Figure 3. Analysis for cell-capture performance. (a) Numbers of captured MCF-7 cells at an incubation time of 5 min with four different substrates. (b) Fluorescence image of cells captured on the Au-SiNW2 substrate with live/dead staining. Cell viability was quantified by using a live/dead viability/cytotoxicity kit (Invitrogen, OR). The green dots indicate Calcein-AM staining of viable cells while the red dots (marked by arrows) indicate ethidium homodimer-1 staining of nonviable cells. (c−e) Fluorescence images of MCF-7 cells captured on pure Si NW (c), Au-SiNW1 (d), and Au-SiNW2 (e) substrates. (f−h) Magnified SEM images of a single cell captured on pure Si NW (f), Au-SiNW1 (g), and Au-SiNW2 (h) substrates at an incubation time of 5 min. To observe the morphology of captured cells by SEM, they were fixed with 4% paraformaldehyde and dehydrated. (i) Increase of cell-capture yield on different substrates as the incubation time is varied. Each error bar represents the standard error of the mean.

cover the NW sidewalls at regular intervals are of one monolayer thickness. The data strongly supports the previous observation that the liquid AuSi spreads from the seed at the NW tip along the sidewalls during NW growth.23 We note that all Si NWs, regardless of the wire diameter or length, exhibit a uniform coverage of Au NCs on the NW sidewalls. In order to determine the number density of Au NCs on the Si-NW sidewalls, we reconstructed a three-dimensional image by HAADF-STEM tomography24,25 (see Supporting Information Movie 1). The number density of Au NCs on the Si NW sidewalls is at least 3 × 1012 cm−2. If we assume that the Au NCs have a spherical shape with an average diameter of 4 nm, we estimate that approximately 37% of the entire Si NW surface is covered with Au NCs. Figure 2a shows a typical HAADF-STEM image of a Si NW on the ⟨112̅⟩ zone axis. The elemental identity of Au in the bright dot region was confirmed using energy dispersive X-ray spectroscopy (EDS) and low-loss electron energy loss spectroscopy (EELS) analyses (Supporting Information, Figure S2). Interestingly, the low-loss EELS peak shows a shift of 2.5 eV from 25 eV at the Au tip (P3 in Supporting Information Figure S2), which corresponds to the bulk plasmon excitation of Au,26 to 22.5 eV at the Au NC (P2 in Supporting Information Figure S2). Figure 2b shows a high-resolution HAADF-STEM image taken from the two-phase area (marked A1 in Figure 2a) where

on a p-type Si(111) wafer shows a high-density of Si NWs grown perpendicularly to the Si surface (Figure 1b). Such perpendicular growth of Si NWs is explained by the formation of a lowest-free-energy state at the liquid AuSi droplet/Si NW interface that is parallel to the Si(111) plane.21 We found that the diameter (50−160 nm) and length (5−10 μm) of the Si NWs could be controlled by the droplet size and partial pressure of silane at relatively low growth temperatures (673− 823 K). It is to be noted that the Si NW sidewall exhibits periodic faceting except22 within a length l of ∼350 nm from the Au tip (Figure 1b, inset). The magnified SEM image taken in the middle part of a Si NW clearly shows the pronounced faceting of the NW surface and the abundant presence of NCs on it (Figure 1c). From transmission electron microscopy (TEM) and selected-area electron diffraction (SAED), we found that these Si NWs grew along the ⟨111⟩-direction and had a hexagonal cross section with the {110}-facets (Supporting Information, Figure S1). High-angle annular dark field scanning transmission electron microscopy (HAADFSTEM) images taken from the top to the bottom of a Si NW (Figure 1d−f) clearly show a high-density and uniform coverage of Au NCs (bright dots) with a size of 3−5 nm on the entire surface of the Si NW. The cross-sectional TEM (Figure 1g) and HAADF-STEM (Figure 1h) images of the faceted section of a Si NW further reveal that the Au NCs that 1640

dx.doi.org/10.1021/nl2045759 | Nano Lett. 2012, 12, 1638−1642

Nano Letters

Letter

Figure 4. Plasmonic photothermal therapy of captured cells on three NW-based substrates. (a−c) Fluorescence images of MCF-7 cells captured on pure Si NW (a), Au-SiNW1 (b), and Au-SiNW2 (c) substrates before and after exposure to NIR radiation of a continuous-wave laser (laser power, 3 W; irradiation area, 2 × 5 mm2) . All samples were stained using a live/dead viability/cytotoxicity kit to check cell viability. The green dots indicate viable cells while the red dots indicate nonviable cells. White arrows in each figure indicate the same tumor cells for comparison. (d) SEM images of MCF-7 cells captured on Au-SiNW2 substrates before and after the NIR radiation. The morphology of MCF-7 cells was observed after chemical fixation with 4% paraformaldehyde and dehydration.

Figure 2e is a bright-field TEM image that show that Au nanoparticles in a Si NW increase rapidly in size and dispersal at the high-annealing temperature, possibly due to atomic Au diffusion, coalescence, and Ostwald ripening.29 The size range of Au nanoparticles annealed for 20 min at 1073 K is 3−14 nm (Figure 2f). We next examined its application to cell capture and photothermal therapy using EpCAM-positive breast cancer cells (MCF-7) grown on a cell culture medium (see Supporting Information) as a model for CTCs. A 25 μL suspension (105 cells mL−1) of the MCF-7 cell was introduced onto various Si NW-based substrates (1 × 1 cm) that were anti-EpCAM antibody coated. Figure 3a shows the number of captured MCF-7 cells on four different substrates at an incubation time of 5 min: flat Au (“AuP”), pure Si NW (“SiNW”), Au NCcoated Si NW (“Au-SiNW1”), and Au NC-coated Si NW annealed at 1073 K (“Au-SiNW2”). The cell-capture yield is only 2−4% on flat Au or pure Si NW substrate but increases to 15% on Au NC-coated Si NW and becomes ∼40% on its annealed form. In parallel with recent studies,17 the viability of MCF-7 cells on the Au NC-coated Si NW substrate was higher than 95% (Figure 3b). In order to identify and enumerate MCF-7 cells captured on the above four substrates, we employed fluorescent confocal laser scanning microscopy (Figure 3c−e) and SEM (Figure 3f−h) analyses. Notably, the cells captured on the Au-SiNW1 and Au-SiNW2 substrates showed extensive cell protrusions in contrast to the smooth round cell morphology on the pure Si NW substrate. This finding suggests a significant increase in local interaction between the cellular surfaces of MCF-7 cells (microvilli)30 and the antibodies on the surface of Au nanoparticles. Figure 3i shows how the cell-capture yield varies on different substrates as the incubation time is increased. Of particular importance is the fact that the cell-capture yield can reach ∼88% on the AuSiNW2 substrate upon incubation for 40 min, which is nearly twice the highest yield currently available from a 3Dnanostructured substrate.16 Integration of the Au-SiNW2 substrate into microfluidic systems is expected to achieve an

Au atoms are distributed throughout the Si lattice. Individual Au (Z = 79) and Si (Z = 14) atoms are easily distinguished by their intensity in the HAADF-STEM image. The intensity profiles in the inset of Figure 2b along the three linearly marked regions (denoted A, B, and C) indicate that Au atoms are generally substituted in one of the Si (440) dumbbell columns, although the site occupation itself by Au atoms (denoted by arrows) is random at the phase boundary. A previous study using density functional theory calculations found that the energetically stable site for an Au atom inside a Si lattice was the substitutional site rather than the interstitial one.11 It also appears that the 2.5 eV shift in the EELS peak of the Au NC (Supporting Information Figure S2) results from the supersaturated Au atoms at the substitutional Si sites. Figure 2c shows a high-resolution TEM (HRTEM) image obtained at the edge of the NW (A2 in Figure 2a) and its EDS spectrum (inset), which strongly indicates that the Si NW surface is covered with a SiO2 thin layer that could have been formed during the cooling stage of the Si NW formation.9 The native oxide layer is amorphous and 1−1.3 nm in thickness. Notably, the lattice-resolved TEM image reveals planar defects (twin faults) at the intersection between the surface facets. Au nanoparticles possess localized surface plasmon activity that imparts different optical properties from those of bulk material.27 To investigate the surface plasmon mode of Au NCs, we made use of the high spatial (