2788
Langmuir 2006, 22, 2788-2794
Nanoparticle Agglutination: Acceleration of Aggregation Rates and Broadening of the Analyte Concentration Range Using Mixtures of Various-Sized Nanoparticles Philip J. Costanzo,† Timothy E. Patten,*,† and Thomas A. P. Seery‡ Department of Chemistry, UniVersity of California at DaVis, One Shields AVenue, DaVis, California 95616-5295, and Institute of Materials Science, UniVersity of Connecticut, 97 North EagleVille Road, U-136, Storrs, Connecticut 06269-3136 ReceiVed August 22, 2005. In Final Form: NoVember 28, 2005 SiO2 particles of various sizes were prepared and surface modified with biotin-chain-end-functionalized poly(ethylene glycol). Dispersions of these particles were prepared, and their aggregation was induced upon the addition of avidin. The aggregate size and growth rate were monitored by DLS analysis, and SEM and TEM images of freeze-dried samples of the aggregate solutions were used to confirm the DLS data and to image the aggregate size and dimension. A linear correspondence between apparent diameter and time was observed, and both the 20 and 300 nm particles aggregated at slower rates than for the 40 nm particles. These observations were attributed to differences in the average functionality of the systems and the different initial concentrations of particles and avidin. The observed aggregation rates of binary combinations of the three particle sizes (i.e., 20 + 40 nm or 40 + 300 nm) were faster than those of the single-sized mixtures. These results were attributed to the faster flux of smaller particles toward larger particles in the mixture solutions as well as to the potentially larger number of productive collisions possible between mixtures of small particles and large particles versus only similarly sized particles. Combinations of the three sizes of particles were studied to find an empirical optimum formulation for generating large aggregates on a short time scale and over a wide range of analyte concentrations.
Introduction Many advances in the field of nanotechnology have involved analyte detection applications. There are a number of nanoparticlebased techniques being studied as possible sensor mechanisms. In one example, nanoparticle probes were used to amplify target analytes (e.g., prostate-specific antigen) and exhibited a detection limit of ∼3 aM, which is 6 orders of magnitude better than those found for clinically accepted conventional assays.1 Biologically modified quantum dots have been used for fluorescence imaging of cell components.2,3 Magnetic nanoparticles have also been functionalized with biologically active ligands for a variety of applications, including magnetic resonance imaging (MRI), cell separation, and tissue engineering.4 In prior work, it was shown that SiO2 nanoparticles with surfaces grafted with a water-soluble polymer (poly(ethylene glycol), i.e., PEG)) with chain ends containing a ligand for proteins (biotin) would aggregate in the presence of its complementary protein (avidin). This aggregation process is similar to the wellknown latex agglutination test used in a number of commercially available immunoassays and therefore is dubbed “nanoparticle agglutination.” Under optimal conditions, micrometer-sized aggregates formed within several minutes promoted by avidin concentrations as low as several nanomolar. A microfluidic device was developed that incorporated this aggregation-based sensing mechanism. The aggregates were selectively mobilized over individual nanoparticles using dielectrophoresis (DEP) and the presence of the aggregates; therefore, the analyte was detected * Corresponding author. E-mail:
[email protected]. † University of California at Davis. ‡ University of Connecticut. (1) Nam, J. M.; Thaxton, C. S.; Mirkin, C. A. Science 2003, 301, 1884. (2) Ballou, B.; Lagerholm, B. C.; Ernst, L. A.; Bruchez, M. P. Bioconjugate Chem. 2004, 15, 79. (3) Watson, A.; Wu, X.; Bruchez, M. P. BioTechniques 2003, 34, 296. (4) Berry, C. C. J. Mater. Chem. 2005, 15, 543.
using dielectrophoresis impedance measurements (DEPIM).5 It was also demonstrated that the range of analyte concentrations in which aggregation was induced was rather limited because of stoichiometry issues and intraparticle binding processes.6 In the presence of a large initial avidin concentration, the smaller protein would rapidly saturate the surface biotin sites on the particle and thus prevent particles from linking together. In the presence of small initial avidin concentrations, few interparticle cross links could form, and multiple biotin chain ends could bind to the same avidin, further preventing interparticle linking. Consequently, an important issue in nanoparticle agglutination is the broadening of the range over which aggregation will occur. Additionally, it is desirable to discover ways in which to accelerate the aggregation rate and lower the time for analyte detection. In this study, we investigated the effect of nanoparticle sizes on the rate of aggregate formation and the concentration range over which aggregation is observed.
Results and Discussion To study the effect of nanoparticle size on the aggregation of biotinylated nanoparticles, we selected two silica particle diameters, 20 and 300 nm, that bracketed the 40-nm-diameter particles used in a prior study.6 Some modifications to the methods for particle-surface functionalization using ethoxydimethylsilane end-functionalized poly(ethylene glycol)s (PEGs) were needed for the 20-nm-diameter particles. These silica particles were purchased as Ludox (aqueous suspension), which was observed to undergo flocculation when organic solvents were added. Consequently, sulfuric acid was added to the aqueous suspension, and then the acidic sol was added to 2-propanol/sulfuric acid without concomitant particle flocculation. Next, R-ethoxy(5) Costanzo, P. J.; Liang, E.; Patten, T. E.; Collins, S. D.; Smith, R. L. Lab Chip 2005, 5, 606-610. (6) Costanzo, P. J.; Patten, T. E.; Seery, T. A. P. Chem. Mater. 2004, 16, 1775.
10.1021/la0522909 CCC: $33.50 © 2006 American Chemical Society Published on Web 02/14/2006
Nanoparticle Agglutination
Langmuir, Vol. 22, No. 6, 2006 2789
Table 1. Nanoparticles Functionalized with PEG Containing Various Chain-End Moieties
sample 1 2 3 c
particle diameter (nm) 20 40 300
elemental analysis particle-N3 (% N) c
NA 0.29 0.06
N3 per particle a c
NA 2700 237 700
elemental analysis particle-NH2 (% N)
NH2 per particle a
biotin per particleb
biotin/nm2
0.08 0.11 0.03
280 3100 354 000
280 3000 300 000
0.22 0.60 1.06
a Determined by elemental analysis and IR spectroscopy. b Based on elemental analysis, IR spectroscopy, and FITC fluorescence experiments. Azide-terminated particles were not isolated. See text for further details.
dimethylsilyl-ω-azido PEG was added to this solution to graft the polymer chain to the particle surface. Attempts to isolate azide-terminated particles were unsuccessful because of partial azide reduction to amine groups in the acidic solution. Therefore, PPh3 was added for quantitative azide reduction. All of the particles containing PEG chains with amine termini could be biotinylated using methods previously reported.6 Elemental analysis of the biotinylated particles (Table 1) showed that the biotin content per unit area of the particles increased by an order of magnitude for each increase in particle diameter. Currently, we cannot offer a reason for this observed trend because steric crowding should get worse with decreasing surface curvature, but it might indicate surface structure reorganization for smaller particles. Various concentrations of avidin were added to independent aqueous suspensions of the 20 and 300 nm particles, and the average apparent hydrodynamic radius of the growing aggregates was monitored using 90° dynamic light scattering (DLS). The aggregation of the 40-nm-diameter particles in the presence of avidin was studied previously,6 and the data is reproduced here for comparison purposes. The angular dependence of the DLS relaxation time (plotted as the frequency, Γ, against the square of the scattering vector q) was determined for the 20 and 300 nm particles to verify that the particles displayed no unusual scattering behavior. Next, the aggregation process for a 20 nm particle experiment and a 300 nm particle experiment was stopped at various time intervals by flooding the reaction solution with biotin. Each of four timed interval samples was analyzed using multiangle DLS after quenching to confirm that the Γ versus q2 plots were linear so that time course data at 90° could be used for kinetic studies. To compare data between experiments monitoring the aggregation of solutions containing a single particle size, we divided the apparent hydrodynamic diameters by the initial particle diameter to yield a unitless, relative apparent diameter. The 20-nm-diameter particles do not suspend easily at high concentrations, so there was an upper limit to the initial concentration of these particles that could be used in the experiments. The 300-nm-diameter particles scatter light much more intensely than the 20 or 40 nm particles, so it was necessary to use low initial concentrations of particles in these experiments as well. Plots of relative apparent diameter versus time for a range of initial avidin concentrations are shown in Figure 1. Several key features of the plots were noted. First, there was an approximately linear correspondence between the apparent diameter and time. Second, both the 20 and 300 nm particles aggregated at slower rates than the 40 nm particles. To explain the first observation, we modeled the expected relationship between aggregate diameter and time using colloidal aggregation kinetics. For a fast aggregation system7 (model assumptions include spherical particles undergoing Brownian motion, square well potential so that particles adhere upon contact,
and rate constant of aggregation that is largely independent of particle size), the concentration of each aggregate of m particles, [Pm], at a given time interval, t/τ, can be calculated using eq 1, where [P]tot is the total number of aggregates.
(7) Stokes, R. J.; Evans, D. F. Fundamentals of Interfacial Engineering; WileyVCH: New York, 1997.
(8) Odian, G. Principles of Polymerization; John Wiley and Sons: Hoboken, NJ, 2004.
[Pm] ) [P]tot
(τt )
+m - 1
-m - 1
(1 + τt )
(1)
Assuming that aggregate volume scales with the fractal (2.5) power of the initial particle diameter (a number consistent with TEM images, vide infra) and with the number of particles, one can calculate an average number diameter (corresponding to a relative apparent diameter) and plot how the average number diameter would change with respect to time. Figure 2 shows that an approximately, though not exactly, linear correspondence between relative apparent aggregate diameter and time would be expected on short time scales. Considering the smaller number of data points for the experimental data and the subtle curvature of the plot in Figure 2, one can conclude that within experimental error the observed data is consistent with fast colloidal aggregation kinetics. The second observation was unexpected because the rate of particle aggregation should not depend on particle size in this model.7 Small particles will diffuse quickly but are small and collide less frequently with each other than larger particles, which will diffuse more slowly but collide more frequently because of their larger volumes, and both trends tend to counteract each other. A significant difference between the model assumptions and the present system is that the binding between particles is more complex than the assumption of simple spheres with anisotropic binding capability. Each particle in the present system has a finite number of biotin sites distributed over the surface, and they are linked together into an aggregate via avidin, which has the capacity to bind four biotin groups. Thus, the system is analogous to a step growth polymerization of multifunctional monomers, in which the ultimate size and how fast the aggregate size develops are related to the average functionality (fav) of the total system. The average functionality was calculated to be twice the total number of functional groups not in excess (i.e., avidin binding sites or surface-bound biotins) divided by the total number of species (avidin and particles).8 Shown in Table 2 are the initial concentrations of particles and avidin and the average functionality of the system for each experiment. For the 20 nm particle system, fav is lower for each experiment than fav for the 40 and 300 nm experiments (with one exception); therefore, the system should reach large aggregate sizes only at higher conversions, and consequently more slowly, than in the 40 and 300 nm systems. Additionally, because of low surface functionality the particles do not suspend at higher concentrations, so only about half the initial concentration of 20 nm particles relative to the 40 nm experiment could be used. This lower initial particle concentration also slowed the kinetics of aggregation. The fav values for the
2790 Langmuir, Vol. 22, No. 6, 2006
Costanzo et al.
Figure 1. DLS analysis of aggregation experiments performed at various avidin concentrations at room temperature. (A) Sample 1: (20 nm particle, 280 biotin/particle) 2 mL of 0.12 mg/mL and 1 mL of avidin solution; (B) Sample 2: (40 nm particle, 3000 biotin/particle) 3 mL of 0.25 mg/mL, 1 mL of H2O and 1 mL of avidin solution; (C) Sample 3: (300 nm particle, 300 000 biotin/particle) 1 mL of 0.25 mg/mL, 3 mL of H2O, and 1 mL of avidin solution. Table 2. Initial Concentrations of Nanoparticles and Avidin for the Aggregation Experiments in Figure 1 as Well as the Average Functionality (fav) for Each Experiment [avidin]o (nM)
no. avidin
[particles]o (nM)
fav
no. biotin/ no. avidin
2.20 × 1013 2.20 × 1013 2.20 × 1013 2.20 × 1013 2.20 × 1013 2.20 × 1013
2.8 4.0 4.4 5.5 6.7 4.5
513 284 227 125 57 2
no. particles
20 nm Particles
Figure 2. Theoretical modeling of the aggregation of 40 nm particles induced by addition of avidin.
300 nm system are consistently large like those of the 40 nm experiments, but because of the scattering intensity of the 300 nm particles, only one-third of the initial concentrations of avidin and 300 nm particles relative to those in the 40 nm experiments could be used. Thus, the much slower observed aggregation rates were most likely a result of the lower initial concentrations. In addition, because of high surface biotin content of the 300 nm particles relative to that of the 40 nm system, there is a greater likelihood that multiple biotin chain ends can bind to an avidin molecule and inhibit cross linking.6 Next, different binary combinations of the three particle sizes were used in aggregation experiments. To provide for meaningful comparisons between experiments, the number of particles was kept constant from experiment to experiment, and only the ratio
6.7 12.0 15.0 27.3 60.3 1510
1.20 × 1013 2.17 × 1013 2.71 × 1013 4.94 × 1013 1.09 × 1014 2.73 × 1015
4.8 10.0 17.2 34.0 58.0 148
1.45 × 1013 3.01 × 1013 5.18 × 1013 1.02 × 1014 1.75 × 1014 4.46 × 1016
40 nm Particles 2.86 8.61 × 1012 2.86 8.61 × 1012 2.86 8.61 × 1012 2.86 8.61 × 1012 2.86 8.61 × 1012 2.86 8.61 × 1012
5.0 6.2 6.9 7.4 7.6 1.2
1781 858 499 252 148 0.58
0.45 0.90 1.80 4.24 9.00 180
1.35 × 1012 2.71 × 1012 5.42 × 1012 1.28 × 1013 2.71 × 1013 5.42 × 1014
300 nm Particles 2.26 6.80 × 109 2.26 6.80 × 109 2.26 6.80 × 109 2.26 6.80 × 109 2.26 6.80 × 109 2.26 6.80 × 109
8.0 8.0 8.0 8.0 8.0 7.5
1511 753 376 159 75 4
1.22 1.22 1.22 1.22 1.22 1.22
of particles of each size was varied. Additionally, the exact same initial concentration of avidin was used in each experiment. As a result fav was the same for each experiment. A few limitations to the range of experiments that could be performed were found.
Nanoparticle Agglutination
Langmuir, Vol. 22, No. 6, 2006 2791
in which small aggregates form and then cluster to form larger aggregates. With the above experimental data, one can rationalize the observed differences in aggregation rates. In experiments with two different particle sizes, the smaller particles can diffuse faster than the larger particles and will collide more frequently, all other variables being equal, with the larger particles. These facts lead to a larger rate constant of aggregation for the binary particle mixtures. Indeed, the rate constant of aggregation for mixtures of two particle sizes can be estimated using eq 2.
kij )
(
)
2kT 1 1 + (R + Rj) 3η Ri Rj i
(2)
Because of the limited dispersibility of the 20 nm particles and the scattering intensity of 300 nm particles, there was an upper limit to the number of each of these particle types that could be suspended in the starting solutions. As seen in Figure 3, the aggregation rates for binary mixtures of the particles (i.e., 20 + 40 nm or 40 + 300 nm) were faster than those of the single-sized mixtures. Also, on a per particle basis the 300 nm particles appeared to have the largest effect on the overall aggregation rate. TEM and SEM analyses of the 40 + 300 nm experiment (representative images located in Figure 4) revealed an aggregate morphology in which the 300 nm particles were nearly always surrounded by the 40 nm particles, suggesting that the 40 nm particles accumulated about the 300 nm particles. Additional insight into a reason for the faster binary particle aggregation came from replotting the data from Figure 3 as the number of particles per aggregate per time (Figure 5). The number average of particles per aggregate was estimated using 〈nn〉 )(〈Rh〉/Ro)df, where 〈nn〉 is the number average of particles per aggregate, 〈Rh〉 is the mean hydrodynamic radius of the growing aggregate, Ro is the initial hydrodynamic radius of the particle, and df is the fractal dimension of the aggregate.9 Although this model was initially designed for aggregates composed of a single-sized particle, it is reasonable to extrapolate to aggregates composed of various particle sizes. The initial number-averaged particle radius for mixtures of particles of various sizes was calculated using (R1N1 + R2N2 + RnNn)/(N1 + N2 + Nn), where R is the particle diameter and N is the number of particles of a particular diameter in solution. A fractal dimension of 2.5 was estimated for the aggregates via analysis of the TEM images. Figure 5 shows that the number of particles per aggregate increased in an approximately linear fashion for the binary particle experiments, suggestive of particles rapidly accumulating about a seed aggregate or particle. In contrast, the single-sized aggregation experiments showed aggregate size increases with time that were slower initially and then accelerated, which suggests a slower formation of seed aggregates or a process
For the single-particle experiments (20-, 40-, or 300-nmdiameter particles), the rate constant is calculated to be 0.54 × 10-17 m3 particle-1 s-1. Because of the righthand terms containing particle radii, the rate constants for the binary mixtures will be larger: for the 20-40 nm system, kij ) 2.43 × 10-17 m3 particle-1 s-1, and for the 40-300 nm system, kij ) 5.20 × 10-17 m3 particle-1 s-1. Thus, in the binary systems there is a faster initial flux of the smaller particles toward the larger particles relative to other similarly sized particles, which accelerates the observed rate of aggregation. In addition, for aggregation driven by cross linking of polymer brush layers on the particle surfaces, there is an inherent advantage to aggregating smaller particles on the surfaces of larger particle, as we have observed previously.10 For two particles of similar size, the number of coordination sites per particle is low because the area on one particle required for the binding of the other particle is a significant fraction of the surface area of the particle. In contrast, for particles of very different diameters there are many more coordination sites on the larger particle to which the smaller particle can bind. Consequently, for the system with very different particle sizes there are more potentially productive binding collisions that can occur relative to the number for the system with similarly sized particles. For sensing applications utilizing nanoparticle agglutination, an optimal system would generate the largest aggregates in the shortest period of time. Thus, we performed a series of experiments with varying ratios of the 20, 40, and 300 nm particles to determine, empirically, which systems would yield the fastest aggregation rates. In all experiments, the total number of particles present in solution and the amount of avidin added to each solution remained constant to compare effectively how particles of different sizes and formulations would interact. The key data are shown in Figure 6, and the fastest aggregation rates occurred with the formulations of 40 and 300 nm and 20, 40, and 300 nm particles. The binary system had a 10 000:1 ratio of 40 to 300 nm particles, whereas the ternary system had a 8000:10 000:1 ratio of 20 to 40 to 300 nm particles. These two formulations were then used in a series of experiments in which the initial avidin concentration was varied. The apparent diameters of growing aggregates were divided by the initial diameter data point and plotted against time for mixtures of 40 and 300 nm and 20, 40, and 300 nm particles (Figure 7A and B, respectively). A comparison of aggregation rate versus avidin concentration curves is shown in Figure 7C. Compared to the curve observed for the 40-nm-particle-only system, which showed a measurable aggregation rate at avidin concentrations bracketing 100 nM, the binary (40 + 300 nm) system showed appreciable aggregation rates spanning a larger range of avidin concentrations (including those as low as 10 nM).
(9) Tirado-Miranda, M. S.; Callejas-Fernandez, J.; Fernandez-Barbero, A. Langmuir 1999, 15, 3437.
(10) Costanzo, P. J.; Balko, S. M.; Moore, N.; Patten, T. E.; Kuhl, T. L. J. Phys. Chem. B, in press, 2006.
Figure 3. DLS analysis of aggregation experiments performed with particles of various sizes. Conditions: total number of particles dispersed in solution, 1.63 × 1013, except for 20 nm particles/total number of particles dispersed in solution, 2.91 × 1013, and 300 nm particles/total number of particles dispersed in solution, 9 × 109. Total volume, 3 mL. [Avidin], 84 nM. Ratio for experiment containing 20 and 40 nm particles, 4:5. Ratio for experiment containing 40 and 300 nm particles, 10 000:1.
2792 Langmuir, Vol. 22, No. 6, 2006
Costanzo et al.
Figure 4. SEM/TEM analysis of aggregates composed of 40 and 300 nm particles freeze dried from an 84 nM avidin solution after 2 h. Conditions: samples 2 and 3 in a particle ratio of 10 000:1 (total number of particles in solution, 1.63 × 1013).
Figure 5. Data from Figure 3 replotted to compare the number of particles per aggregates against time.
In summary, the aggregation of nanoparticles of a single size obeyed fast colloidal aggregation kinetics, and the rate of aggregation was dependent upon the fav of the system and the initial concentration of the particles. The 40 nm particles showed the fastest aggregation, followed by the 20 nm particles and then the 300 nm particles. The aggregation of binary mixtures of particles occurred at faster rates than that of the single-sized experiments. This difference was attributed to a faster flux of the smaller particles toward the larger particles relative to other similarly sized particles and to the greater coordination number of the larger-sized nanoparticles.
Conclusions SiO2 particles of various sizes were prepared and surface modified with biotin-chain-end-functionalized poly(ethylene
Figure 6. DLS analysis of the aggregation of formulations containing particles of various sizes. Conditions: total number of particles dispersed in solution, 1.63 × 1013. Total volume, 3 mL. [Avidin], 84 nM. Ratio of particles for sample listed by 20:40:300 nm: (A) 5000:10 000:2; (B) 5000:10 000:1; (C) 5000:15 000:1; (D) 8000: 10 000:1; (E) 8000:10 000:0; (F) 0:10 000:1.
glycol). Individual dispersions of these particles were prepared, and their aggregation was induced upon addition of avidin. In these systems, a linear correspondence between apparent diameter and time was observed, and both the 20 and 300 nm particles aggregated at slower rates than the 40 nm particles. These observations were attributed to differences in the average functionality of the systems and the different initial concentrations of particles and avidin. Next, different binary combinations of the three particle sizes were used in aggregation experiments. The observed aggregation rates for binary mixtures of the particles (i.e., 20 + 40 nm or 40 + 300 nm) were faster than those of the
Nanoparticle Agglutination
Langmuir, Vol. 22, No. 6, 2006 2793
Figure 7. DLS analysis of aggregation experiments of (A) 20, 40, and 300 nm particles and (B) 40 and 300 nm particles performed at various avidin concentrations. (C) Dependence of initial aggregation rate upon different formulations of particles of varying diameter at various avidin concentrations. Conditions: total number of particles dispersed in solution, 1.63 × 1013; total volume, 3 mL. Ratio of particles for sample listed by 20:40:300 nm: (A) 8000:10 000:1; (B) 0:10 000:1.
single-sized mixtures, and on a per particle basis, the 300 nm particles appeared to yield a larger aggregation rate enhancement relative to the 20 nm particles. The rate enhancements of the binary mixtures were attributed to the faster flux of smaller particles toward larger particles in the mixture solutions as well as to the potentially larger number of productive collisions possible between mixtures of small particles and large particles versus only similarly sized particles. Finally, combinations of the three sizes of particles were studied to find an empirical optimum formulation for generating large aggregates on a short time scale and over a wide range of analyte concentrations. Experimental Section Materials. All materials were purchased from commercially available sources and used without further purification. Three hundred nanometer SiO2 particles were prepared using the Sto¨ber method.11,1211-12 Forty and three hundred nanometer SiO2 particles were biotinylated according to a previously reported method.6 Instrumentation. Elemental analysis was conducted by Midwest Microlabs. Dynamic light scattering (DLS) analysis was conducted using a Brookhart Coherent DPSS 532 laser with an IEM 9863 detector (UC Davis) and a BI-200SM goniometer (Brookhaven Instruments) using a 2.5 W Ar laser light source (model Innova 70-3 from Coherent) operating at 514.5 nm and a BI-9000 AT digital correlator (Brookhaven Instruments) (U Conn). Transmission electron (11) Bogush, G. H.; Tracy, M. A.; Zukowski IV, C. F. J. Noncryst. Solids 1988, 104, 95. (12) Philipse, A. P.; Vrij, A. J. J. Colloid Interface Sci. 1989, 128, 121.
microscopy (TEM) was conducted on a Phillips CM-12 TEM. Scanning electron microscopy (SEM) was conducted on an FEI XL30-SFEG. All microscopy samples were prepared by dipping carbon-coated copper grids into the appropriate solutions and then removing the water by freeze drying. Synthesis of 40 nm SiO2. Forty nanometer SiO2 particles were synthesized via a modification of a previously reported method.13 Igepal CO-520 (40.0 g, 38.4 mL), cyclohexane (100 mL), and a stir bar were placed into a 500 mL round-bottomed flask. The reagents were stirred until a clear microemulsion was present. Next, NH4OH (8.13 mL, 3.46 M, 28.1 mmol) was added, and the reaction was stirred until clear. Tetraorthosilicate (23.7 mL, 106 mmol) was added, and the reaction was stirred for 16.5 h. Particles were isolated by a series of centrifuge/decant/resuspend cycles (5×) with CH3OH/ hexane/THF (1:1:3) and then stored in THF. Preparation of 20 nm SiO2-PEG-NH2. An aqueous solution of SiO2 nanoparticles (2.5 mL, 40 wt %) and a stir bar were placed into a 50 mL round-bottomed flask to which 1 mL of concentrated H2SO4 was quickly added. In a second 50 mL round-bottomed flask, a stir bar, 2-propanol (15 mL), and concentrated H2SO4 (1 mL) were added and stirred. The contents of flask 1 and R-dimethylethoxysiloxy-ω-azide poly(ethylene glycol) (2.6 g, 5.2 mmol) were then added to the second flask and heated at reflux for 18 h. The flask was cooled to room temperature, and PPh3 (1.6 g, 6 mmol) was then added to the flask and allowed to stir at room temperature for 24 h. Excess solvent was removed by rotary evaporation, and the particles were isolated by a series of centrifuge/decant/resuspend cycles (5×) with MeOH/hexane/THF (1:1:3). Particles were dried under vacuum and analyzed by elemental analysis (Table 1). (13) Farmer, S. C.; Patten, T. E. Chem. Mater. 2001, 13, 3920.
2794 Langmuir, Vol. 22, No. 6, 2006
Costanzo et al.
Preparation of 20 nm SiO2-PEG-Biotin. Biotin-succinimide ester was generated and then transferred to a 100 mL round-bottomed flask containing SiO2-PEG-NH2 suspended in DMF (45 mL) and stirred for 24 h. Particles were isolated by a series of centrifuge/ decant/resuspend cycles (5×) with MeOH/hexane/THF (1:1:3) and dried under vacuum.
(U. Maine, Department of Electrical Engineering) and Professor Scott Collins (U. Maine, Department of Chemistry) for helpful discussions. This work was supported by NSF (DMR-0306055) and NEAT-IGERT (DGE-9972741).
Acknowledgment. We thank Mike Dunlap for SEM instrumentation, and we recognize Professor Rosemary Smith
LA0522909
