ARTICLE pubs.acs.org/Langmuir
Anisotropic Magnetic Porous Assemblies of Oxide Nanoparticles Interconnected Via Silica Bridges for Catalytic Application Josias B. Wacker, Virendra K. Parashar, and Martin A. M. Gijs* Laboratory of Microsystems, Ecole Polytechnique Federale de Lausanne (EPFL), CH-1015 Lausanne, Switzerland
bS Supporting Information ABSTRACT: We report the microfluidic chip-based assembly of colloidal silanol-functionalized silica nanoparticles using monodisperse water-in-oil droplets as templates. The nanoparticles are linked via silica bridges, thereby forming superstructures that range from doublets to porous spherical or rod-like micro-objects. Adding magnetite nanoparticles to the colloid generates microobjects that can be magnetically manipulated. We functionalized such magnetic porous assemblies with horseradish peroxidase and demonstrate the catalytic binding of fluorescent dye-labeled tyramide over the complete effective surface of the superstructure. Such nanoparticle assemblies permit easy manipulation and recovery after a heterogeneous catalytic process while providing a large surface similar to that of the individual nanoparticles.
’ INTRODUCTION A hot field in synthetic assembly of inorganic materials is to build nanoparticle superstructures of predefined size, shape, and surface texture.1 One method for morphology control is confining an ensemble of nanoparticles in a single colloidal droplet, assembling these nanoparticles like building blocks, and subsequently consolidating them into a superstructure.2 6 While conventional emulsion techniques yield droplets with a wide size distribution, microfluidics allow one to generate monodisperse droplets.7 Several publications have shown the feasibility of using monodisperse microdroplets to form equally sized spherical colloidal assemblies.3,4,8 Sophisticated techniques have been developed to create tori,9 peanut-shaped, and hollow nonspherical colloidal assemblies,10,11 involving arrested coalescence of droplets, use of polymeric linkages, or formation of double emulsions. The cohesion between the nanoparticles in these assemblies is provided via van der Waals forces,2,12 hydrophobic interactions between functionalized nanoparticles,13 or embedding the nanoparticles in a (photo)polymer matrix.14,15 However, these joining techniques preclude generation of mechanically and chemically stable assemblies, a prerequisite for applying the latter as functional or catalytic supports. For most functionalized superstructures, stabilization via sintering,2,12 i.e., densifying by filling the pores between the particles at high temperatures, is no practicable option as functionality may be lost at elevated temperatures. ’ EXPERIMENTAL SECTION Materials. Tetraethyl orthosilicate (TEOS), aqueous ammonia, ethanol, and oleic acid were purchased from Sigma Aldrich (Switzerland). Poly-L-lysine grafted with polyethylene glycol side chains (PLL-g-PEG) and biotinylated PLL-g-PEG were obtained from Susos (Switzerland). Kit #42 for tyramide-based enzymatic signal amplification r 2011 American Chemical Society
with streptavidin-conjugated horseradish peroxidase (HRP) was bought from Invitrogen (Switzerland). Formation of Assemblies. The detailed synthesis procedure of silica nanoparticles from TEOS and ammonia is described in the Supporting Information. To interconnect the nanoparticles, we preactivated them with silanol groups by a high-pressure high-temperature procedure. In more detail, this involves autoclaving of the nanoparticles in deionized water at 120 °C for 10 min under indigenously built pressure in a high-pressure reaction vessel (type 4744, Parr Instrument Co., Moline, IL) at a slow stirring speed.16 After autoclaving, water is replaced with fresh deionized water to make a silica colloid of pH 6.2. After adjusting the concentration of nanoparticles in water, we injected them in a flow-focusing microfluidic droplet generator (see Figure 1) using microliter syringes (from ILS, St€utzerbach, Germany) and syringe pumps (Nemesys, Cetoni GmbH,Korbussen, Germany), thus forming monodisperse colloidal droplets in oleic acid. Water and oil flow rates were typically 0.05 0.1 and 0.001 0.01 μL/s, resulting in 0.3 0.4 nL colloidal droplets. The latter were transported via a capillary to the bottom of a container made of polydimethyl siloxane (PDMS) that was filled with oleic acid. After water evaporation at 160 °C, we obtained assemblies as shown in Figure 2a. To form pillar-like assemblies, we linearly oscillated the PDMS container with the colloidal droplets at a frequency of 0.5 Hz with an amplitude of 10 cm. A too high oscillation frequency would induce irreversible splitting of the colloidal droplets, resulting in a multitude of smaller spherical agglomerates, rather than forming elongated assemblies. Functionalization of Assemblies. To magnetically actuate the assemblies, we added 0.1% w/w Fe3O4 nanoparticles17 (7 nm in diameter, pI = 2, no surface modification) to the silica colloid and formed the assemblies as described before. The pH of the heterogeneous nanoparticle mixture was 3.7. Received: November 17, 2010 Revised: February 22, 2011 Published: March 21, 2011 4380
dx.doi.org/10.1021/la1045699 | Langmuir 2011, 27, 4380–4385
Langmuir To carry out experiments with HRP-functionalized assemblies, we prepared a 1% blocking solution (BS), as described in the manual provided with the TSA kit. Biotinylated PLL-g-PEG and PLL-g-PEG were diluted in BS and used at a concentration of 0.1 mg/mL. The magnetic assemblies were incubated with biotinylated PLL-g-PEG for 1 h, and the microflow cell was exposed to PLL-g-PEG. The microflow cell was assembled by gluing two strips of double sticky tape (Scotch, 3M) in parallel onto a microscope slide (Menzel-Gl€aser, Braunschweig, Germany), leaving a channel of ∼2 mm width and ∼10 mm length in between, and covering this channel with a coverslip (Menzel-Gl€aser, Braunschweig, Germany). After rinsing the microflow cell with BS, the assemblies were brought into the cell and held in place with a permanent magnet during rinsing with BS and subsequent incubation with different concentrations of HRP in BS. After 1 h, the cell containing the HRPfunctionalized nanoparticle assemblies was rinsed with BS, and fluorescently labeled tyramide in 0.0015% v/v H2O2 solution was allowed to enzymatically precipitate on the assemblies during 5 min. Then the microflow cell and the nanoparticle assemblies were rinsed again with BS, and the fluorescence signal was recorded under an upright microscope (Imager A1.m, Zeiss, Germany) with 5, 20, and 100
ARTICLE
objectives using the appropriate filter setup and a CCD camera (OrcaER C4742-80 from Hamamatsu, Japan). Characterization. The NMR spectrum was measured at ambient temperature on a Bruker Avance DRX 300wb spectrometer with a basic frequency of 59.622 MHz. The sample was placed in a zirconium dioxide rotor of 4 mm outer diameter, sealed with a Kel-F cap, and spun at a magic angle spinning rate of 12 kHz. The pulse duration was 4.5 μs, and the relaxation delay was 5 s. NMR spectra were recorded using 1200 scans. FTIR spectra were obtained on a Spectrum One FT-IR Spectrometer (Perkin-Elmer) equipped with a single-reflection attenuated total reflectance system from Specac. Measurements were performed with standard parameters (range 4000 400 cm 1; resolution 2 cm 1; 100 scans) using the supplied Perkin-Elmer software Spectrum v3.02. Mechanical testing of the assemblies was done by compressing them at a speed of 0.14 mm/min between two flat pieces of silicon using a DC motor coupled to a position controller (EPOS 24/1, Maxon, Sachseln, Switzerland). At the same time, we recorded the applied force with an extensometer with 10 nN minimum resolution (DD1 from HBM, Darmstadt, Germany). The resulting deformation was recorded using video microscopy.
’ RESULTS AND DISCUSSION
Figure 1. Experimental setup for the synthesis of silica nanoparticle assemblies inside colloidal microdroplets generated on-chip. Monodisperse colloidal droplets are formed by pinching off an aqueous colloidal stream with two side streams of oleic acid. To form rod-like assemblies, the droplets are rolled over the bottom of a PDMS container by imposing oscillatory motion to the latter.
Spherical Assemblies. The method presented here allows producing low-disperse spherical nanoparticle assemblies that are chemically and mechanically stable in which silica nanoparticles are linked via silica bridges that are formed far below the normal sintering temperature of silica (∼1000 °C). The main parameters for assembly control are surface preactivation state and size of the individual nanoparticles, the concentration of the latter in the colloid, and the diameter of the colloid droplets. While surface preactivation is essential for the cohesion of the assembly, the diameter of the colloidal droplet is mainly influencing the assembly size. The size and concentration of the nanoparticles determine the morphology of a spherical assembly, as is clear from the scanning electron microscopy (SEM) images in Figure 2a. A colloid having small (50 nm) nanoparticles generates assemblies with a wrinkled surface. Bigger nanoparticles (350 nm) result in assemblies with disordered yet smooth surfaces, while the most regular
Figure 2. SEM and TEM analysis of spherical assemblies. (a) SEM micrograph of assemblies made of nanoparticles with various sizes d and colloid weight concentrations c. The relative variance (polydispersity index) for each type of assembly is given in the upper right corner of the micrograph. The combinations marked with a cross indicate the blocking of the channels on the microchip. Scale bars are 10 μm. (b) A magnified view of the assembly prepared from 350 nm nanoparticles at 5% concentration. The scale bar is 500 nm. (c) TEM image of two nanoparticles linked via a silica bridge. The scale bar is 200 nm. 4381
dx.doi.org/10.1021/la1045699 |Langmuir 2011, 27, 4380–4385
Langmuir
Figure 3. Solid state 29Si NMR and ATR-FTIR spectra of preactivated silica nanoparticles. (a) NMR peak at 108 ppm indicates silicon atoms in a Q4 configuration. While the signal at 102 ppm represents silicon atoms that are linked to one silanol group, the shoulder at 92 ppm in the NMR spectrum suggests the presence of two silanol groups per silicon atom. (b) The prominent ATR-FTIR peak at 1050 cm 1 is due to an asymmetric Si O Si stretch; the signals at 960 and 3350 cm 1 indicate the presence of silanol groups and O H groups, respectively.
assemblies are created using 500 nm nanoparticles. We propose the following hypothesis for this finding: During heating, water vapor is released from the shrinking droplet through the pores between aggregating nanoparticles. While the pores of bigger nanoparticle assemblies may provide orifices that are large enough for vapor to exit without significantly changing the arrangement of the nanoparticles, the smaller nanoparticle arrangement will be more easily perturbed, leading to assemblies with more surface roughness and less regularity. Another determining factor for the surface disorder could be in the origin of the arrangement, which is provided by formation of a percolating gel layer linking the individual nanoparticles of the arrangement; the induced irregular displacement related to the gelling phenomenon may be relatively more important for the smaller nanoparticles. We did not find any evidence of diffusion of nanoparticles to the oil phase, probably due to the highly hydrophilic surface of the hydroxylized nanoparticles. While a concentration of 1 2% w/w silica nanoparticles in water permits the assembly of spherical structures with all presented nanoparticle sizes, the use of higher colloidal concentrations can block the microchannels on the chip or lead to uncontrolled coalescence of droplets, thus deteriorating the
ARTICLE
Figure 4. Compression experiments of a silica nanoparticle assembly. (a) Compression force curve of an assembly made of preactivated silica nanoparticles. Initially, the curve follows Hertz’ law of elastic contact pressure. The breakage point is characterized by a sudden loss in force and the appearance of a transversal crack in the assembly. (Inset) Video still image obtained for a compressive displacement of 16 μm. (b) Optical micrograph of the loose arrangement of silica nanoparticles that were not preactivated and lack silanol surface groups. The assembly collapsed under its proper weight. Scale bars are 50 μm.
uniformity of the assemblies. For smaller concentrations, size variance for a given set of process parameters is below 15% for a representative sample of ∼20 nominally identical assemblies, as indicated by the polydispersity index in the respective micrographs. Figure 2b is a high-magnification SEM image of assembled nanoparticles that were consolidated by sintering for 12 h at 400 °C. During this heat treatment, interparticle bonds are progressively converted into strong siloxane linkages, resulting in a mechanically stable assembly. The droplet contained 5% w/w 350 nm nanoparticles that were preactivated with silanol groups. Figure 2c is a transmission electron microscopy (TEM) image of the completely fused interface between two nanoparticles. We confirmed formation of a silanol gel layer on the preactivated nanoparticles using solid state 29Si nuclear magnetic resonance (NMR) and Fourier transform infrared resonance (FTIR) spectroscopy coupled with attenuated total reflectance (ATR): silicon in untreated silica particles exclusively adopts a Q4 configuration, which is reflected by a single NMR peak at around 110 ppm. Preactivated silica nanoparticles show a 29Si NMR signal at 102 ppm, in addition to the Q4 signal at 108 ppm (see curve a of Figure 3). The signal peak at 102 ppm indicates the presence of silicon atoms in a Q3 environment, corresponding to one silanol group per silicon atom. A 29Si NMR feature like the merged shoulder at 92 ppm suggests the existence of Q2 silicon species, representing two silanol groups per silicon atom, as often found in silica gel.18,19 The ATR-FTIR spectrum of high-pressure high-temperaturetreated nanoparticles, presented by curve b in Figure 3, shows a 4382
dx.doi.org/10.1021/la1045699 |Langmuir 2011, 27, 4380–4385
Langmuir
Figure 5. Enzymatic bonding of fluorescently labeled tyramide on HRP-functionalized magnetic porous assemblies of oxide nanoparticles. (a) Fluorescent and (b) optical micrograph of a single HRP-functionalized and two nonfunctionalized assemblies after flowing the substrate tyramide solution through the microflow cell. (c) Fluorescent response (I) curve as a function of HRP concentration. (Inset) Fluorescent micrograph of an assembly that was fractured after the catalytic process, showing fluorescence developed over the bulk of the assembly.
prominent peak at 1050 cm 1 due to an asymmetric Si O Si stretch and shoulders at 960 cm 1 due to silanol groups and 3350 cm 1 due to the presence of O H groups.20 To further substantiate the silanol-forming effect of the pretreatment, we performed thermogravimetric analysis (TGA) and differential thermogravimetric analysis (DTG) on untreated and preactivated silica nanoparticles. For untreated silica nanoparticles we found a weight loss of 0.70% between 200 and 900 °C. For pretreated nanoparticles this value is 2.50% (see Supporting Information). This bigger relative weight loss, in addition to a local minimum in the DTG curve at 600 °C, which is only present in the case of pretreated nanoparticles, is attributed to the higher density of surface silanol groups of pretreated nanoparticles.21 To assess the high mechanical stability of the consolidated assemblies, we performed compression experiments. Figure 4a is the compressive displacement force curve for a 180 μm size consolidated assembly made of 350 nm particles. As suggested by others,22 we fitted the initial part of the curve with Hertz’ law for the elastic compression of spherical objects: F = 2/3E 3 (R 3 s3)1/2, where F is the applied force, E the effective modulus of elasticity, R the radius of the assembly, and s the absolute compressive displacement of the assembly. Using this approach, we could determine an effective modulus of elasticity of 104 MPa. We could not perform comparable experiments with assemblies composed of nanoparticles that were not preactivated (and hence did not have silanol surface groups). Rather than forming
ARTICLE
Figure 6. SEM analysis of 1 100 μm size nanoparticle assemblies. (a) Dumbbell-shaped and (b) triangular assemblies are obtained from emulsions with ∼0.5 nL droplets containing 0.01% w/w nanoparticles in water. (c) Micrometer-sized spherical nanoparticle assemblies are generated by decreasing the droplet volume to ∼0.1 nL and increasing the concentration to 0.1% w/w. Rolling ∼0.5 nL droplets of 5% w/w aqueous colloid over the bottom of a PDMS container produces nonspherical assemblies. Assemblies were subjected to increasing rolling times (d f).
silica bridges, such nanoparticles are loosely arranged in assemblies that collapse under their proper weight, thereby assuming a dome-like shape (see Figure 4b). Functionalization of Assemblies. By simply adding iron oxide nanoparticles to the initial aqueous colloid, we formed magnetic assemblies that can be easily manipulated, as shown in the video in the Supporting Information. We further confirmed the presence of iron in the assembly via energy-dispersive X-ray (EDX) analysis (see Supporting Information). To demonstrate the utility of these magnetic assemblies as catalytic supports, we functionalized them with horseradish peroxidase (HRP), an enzyme frequently used in immunoassays as a signal amplifier.23 While previous publications relied on dyes to prove the functionality of nanoparticle assemblies,5,24 we here show a full enzymatic reaction on silica nanoparticle superstructures. Typically, 10 magnetic assemblies were injected into the microflow cell where they were held in place using a permanent magnet while subsequently flowing through the different reagents and washing solutions (∼100 μL/min). The porous assemblies were first functionalized in the microflow cell by exposing them to a variable concentration of HRP (0.1 10 μg/mL). After rinsing with buffer solution, flowing the substrate-tyramide solution through the superstructures lead to enzymatic precipitation of the fluorescent tyramide. Figure 5a and 5b shows fluorescent and optical micrographs, respectively, of three porous assemblies positioned in the microflow cell. The two assemblies on the right 4383
dx.doi.org/10.1021/la1045699 |Langmuir 2011, 27, 4380–4385
Langmuir were injected into the cell after the HRP functionalization step and before application of the substrate tyramide solution. Therefore, they did not exhibit fluorescence. Figure 5c shows the fluorescent response as a function of HRP concentration. The fluorescence signal grows with increasing amount of HRP, in agreement with standard immunoassay experiments.25 The insert shows a fluorescent micrograph of an assembly which was fractured after performing the enzymatic reaction and illustrates that fluorescence indeed developed all over the interior of the assembly. By exposing a comparable effective surface to the surrounding liquid while offering straightforward separation after catalytic use, these porous assemblies therefore present a tremendous advantage as catalytic reaction support.26,27 Nonspherical Assemblies. In addition to spherical assemblies, the presented droplet-based method allowed us to form nonspherical assemblies of nanoparticles of different sizes (see Figure 6). Reducing the droplet volume to ∼0.5 nL and the nanoparticle concentration to 0.01% w/w produced, for example, the nanoparticle doublets and triplets shown in Figure 6a and 6b, respectively. For lower concentrations, the number of nanoparticles per droplet is significantly smaller than expected. We attribute this finding to the observation that a certain amount of nanoparticles is deposited in the inlet channel and therefore not incorporated in the droplets. The number of nanoparticles per droplet is reported to follow a Poisson distribution,8 causing a big shape variety in assemblies containing only a few nanoparticles. Droplets with more nanoparticles lead to formation of spherical assemblies containing ∼50 nanoparticles (see Figure 6c). Rolling the colloidal droplets on the bottom of the PDMS container prior to the 160 °C heating step produced nonspherical assemblies, as shown in the micrographs of Figure 6d f. To form these pillar-like assemblies, we linearly oscillated the PDMS container at a frequency of 0.5 Hz with an amplitude of 10 cm. A too high oscillation frequency would induce irreversible splitting of the colloidal droplets, resulting in a multitude of smaller spherical agglomerates, rather than forming elongated assemblies. The structures of Figure 6d f, made of 350 nm nanoparticles at 5% w/w, were obtained after progressive rolling during 2 15 min, longer times corresponding to more elongated shapes. This morphology is independent of whether magnetic Fe3O4 nanoparticles were incorporated in the colloidal solution or not.
’ CONCLUSIONS We assembled on a microfluidic chip magnetic iron oxide and preactivated silica nanoparticles into consolidated chemically stable porous micro-objects with a large effective surface. The droplet-based assembly of colloidal nanoparticles resulted in microspheres of small size variance; rolling the droplets over a hydrophobic surface allowed us to synthesize rod-like or ellipsoidal assemblies. The work presented here is, to the best of our knowledge, the first successful attempt to form durable rodshaped assemblies of inorganic nanoparticles. Owing to their magnetic anisotropy, such micro-objects allow convenient magnetic manipulation, in particular rotation, implying a high interaction with the surrounding fluid and easy recovery from a matrix after heterogeneous catalysis while offering an effective surface comparable to that of the individual nanoparticles. To illustrate their potential, we functionalized magnetic assemblies with HRP and demonstrated a catalytic precipitation reaction. We expect that they can be used in many chemical and pharmaceutical
ARTICLE
applications, which would benefit from efficient and reusable catalysis supports. For example, they can be applied during the development of drugs and chemical reagents or in combinatorial chemistry, multiplex chemical sensing, or drug delivery.
’ ASSOCIATED CONTENT
bS
Supporting Information. Video showing the motion of ellipsoidal magnetic porous assemblies in the field of a rotating permanent magnet, detailed information on chip fabrication, nanoparticle synthesis, graphs from TGA/TGD, and an EDX spectrum of a magnetic assembly. This material is available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION Corresponding Author
*E-mail: martin.gijs@epfl.ch.
’ ACKNOWLEDGMENT The authors acknowledge the Swiss National Science Foundation for financial support (Grant No. 200020-119788) and also acknowledge Prof. Heinrich Hofmann for providing the iron oxide nanoparticles and help in TGA/DTG, the help of Dr. Csilla Miko in performing the compression tests, Dr. Frederic Lacharme’s advice in the HRP experiments, and Dr. Jalil Sayah’s support in powder blasting. ’ REFERENCES (1) Nie, Z.; Petukhova, A.; Kumacheva, E. Nat. Nano 2010, 5, 15–25. (2) Dinsmore, A.; Hsu, M.; Nikolaides, M.; Marquez, M.; Bausch, A.; Weitz, D. Science 2002, 298, 1006–1009. (3) Nie, Z.; Park, J. I.; Li, W.; Bon, S. A. F.; Kumacheva, E. J. Am. Chem. Soc. 2008, 130, 16508–16509. (4) Shah, R. K.; Kim, J.-W.; Weitz, D. A. Langmuir 2009, 26, 1561–1565. (5) Samanta, B.; Patra, D.; Subramani, C.; Ofir, Y.; Yesilbag, G.; Sanyal, A.; Rotello, V. M. Small 2009, 5, 685–688. (6) Zerrouki, D.; Baudry, J.; Pine, D.; Chaikin, P.; Bibette, J. Nature 2008, 455, 380–382. (7) Anna, S. L.; Bontoux, N.; Stone, H. A. Appl. Phys. Lett. 2003, 82, 364–366. (8) Yi, G. R.; Thorsen, T.; Manoharan, V.; Hwang, M. J.; Jeon, S. J.; Pine, D.; Quake, S.; Yang, S. M. Adv. Mater. 2003, 15, 1300–1304. (9) Velev, O. D.; Lenhoff, A. M.; Kaler, E. W. Science 2000, 287, 2240–2243. (10) Studart, A. R.; Shum, H. C.; Weitz, D. A. J. Phys. Chem. B 2009, 113, 3914–3919. (11) Lee, D.; Weitz, D. A. Small 2009, 5, 1932–1935. (12) Hsu, M. F.; Nikolaides, M. G.; Dinsmore, A. D.; Bausch, A. R.; Gordon, V. D.; Chen, X.; Hutchinson, J. W.; Weitz, D. A.; Marquez, M. Langmuir 2005, 21, 2963–2970. (13) Lee, M. H.; Prasad, V.; Lee, D. Langmuir 2009, 26, 2227–2230. (14) Kim, S. H.; Shim, J. W.; Lim, J. M.; Lee, S. Y.; Yang, S. M. New J. Phys. 2009, 11, 17. (15) Kim, S.-H.; Heo, C.-J.; Lee, S. Y.; Yi, G.-R.; Yang, S.-M. Chem. Mater. 2007, 19, 4751–4760. (16) Parashar, V. K.; Orhan, J.-B.; Sayah, A.; Cantoni, M.; Gijs, M. A. M. Nat. Nano 2008, 3, 589–594. (17) Chastellain, M.; Petri, A.; Hofmann, H. J. Colloid Interface Sci. 2004, 278, 353–360. (18) Glaser, R. H.; Wilkes, G. L.; Bronnimann, C. E. J. Non-Cryst. Solids 1989, 113, 73–87. 4384
dx.doi.org/10.1021/la1045699 |Langmuir 2011, 27, 4380–4385
Langmuir
ARTICLE
(19) Magi, M.; Lippmaa, E.; Samoson, A.; Engelhardt, G.; Grimmer, A. R. J. Phys. Chem. 1984, 88, 1518–1522. (20) Wood, D. L.; Rabinovich, E. M. Appl. Spectrosc. 1989, 43, 263–267. (21) Odlyha, M.; Scott, R.; Simpson, C. J. Therm. Anal. Calorim. 1993, 40, 1197–1212. (22) Antonyuk, S.; Tomas, J.; Heinrich, S.; M€orl, L. Chem. Eng. Sci. 2005, 60, 4031–4044. (23) vonWasielewski, R.; Mengel, M.; Gignac, S.; Wilkens, L.; Werner, M.; Georgii, A. J. Histochem. Cytochem. 1997, 45, 1455–1459. (24) Lee, D.; Weitz, D. A. Adv. Mater. 2008, 20, 3498–3503. (25) Wild, D. The Immunoassay Handbook, 3rd ed.; Elsevier: New York, 2005. (26) Choi, M.; Cho, H. S.; Srivastava, R.; Venkatesan, C.; Choi, D. H.; Ryoo, R. Nat. Mater. 2006, 5, 718–723. (27) Liu, Q. Y.; Guo, X. H.; Li, Y.; Shen, W. J. Langmuir 2009, 25, 6425–6430.
4385
dx.doi.org/10.1021/la1045699 |Langmuir 2011, 27, 4380–4385
