Size Selective Adsorption of Gold Nanoparticles by ... - ACS Publications

Article pubs.acs.org/JPCC

Size Selective Adsorption of Gold Nanoparticles by Electrostatic Assembly Julian A. Lloyd,†,§ Soon Hock Ng,†,§ Timothy J. Davis,‡,§ Daniel E. Gómez,‡,§,⊥ and Udo Bach*,†,‡,§ †

Department of Materials Science and Engineering, Monash University, Clayton, VIC 3168, Australia Commonwealth Scientific and Industrial Research Organisation, Manufacturing, Research Way, Clayton, VIC 3168, Australia § Melbourne Centre for Nanofabrication, Wellington Road 151, Clayton 3168, Australia ‡

Downloaded via UNIV OF LOUISIANA AT LAFAYETTE on October 30, 2018 at 09:47:41 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

S Supporting Information *

ABSTRACT: In this study, we show that electrostatic interactions between charged substrates containing preattached nanoparticles and bidisperse nanoparticle colloids can be engineered to achieve size selective adsorption and dimer formation. Electrostatic interactions enable the assembly of the dimers with high yields due to the interplay between attractive and repulsive forces resulting from charges confined on the particles and substrate surfaces. We investigate in detail the effects of temperature, incubation time and particle mixing ratios of the bidisperse solution and benchmark the size-selectivity for different scenarios. Driving forces of the assembly process are explained using DLVO theory (Derjaguin, Landau, Verwey, and Overbeek).

■

INTRODUCTION

We have recently shown that electrostatic self-assembly can be used to form regular arrays of surface confined gold nanoparticle (AuNP) dimers in a self-limiting 2-step electrostatic assembly process.1 In this process, we carefully control the adsorption behavior of positively charged AuNPs (satellites) onto a substrate with negatively charged AuNPs (cores). The basic concept of the adsorption process is depicted in Figure 1. In a first step, negatively charged (core) particles are adsorbed onto a substrate that has a positive surface charge. In a second step, this substrate is immersed in a colloidal solution of positively charged (satellite) particles. The complex interplay between the electrostatic interparticle and particle-surface forces creates a “funneling” potential energy barrier (see Figure 2a). This funnel guides the positively charged satellite particles toward the negatively charged cores. In this study, we show that the potential energy barrier that the satellites have to overcome during their approach to the cores is strongly dependent on the size of the satellite particles. The size-dependence of the electrostatic barrier can be used to hinder the adsorption of bigger nanoparticles while facilitating the adsorption of smaller ones, thus providing a size-selective adsorption from a bidisperse satellite particle solution. To investigate the size selectivity of the electrostatic dimer-forming process, we calculate interaction energy maps and activation energy barriers for different satellite particle sizes and experimentally study the adsorption of satellite particles from bidisperse solutions. We present a statistical study of the adsorption yields and size-selectivity as a function of assembly temperature, time and mixing ratio of the bidisperse colloid and © 2017 American Chemical Society

Figure 1. Schematic illustration of the electrostatic, 2-step assembly process. Starting with a positively charged (green) SiO2/Si substrate and negatively charged (red) particles (cores) an array of uniformly distributed cores is created in step 1. In a second step, these substrates are exposed to a colloidal solution of positively charged particles (satellites) of two different sizes. The cores act as adsorption centers for these satellites, facilitating the dimer-forming process and favoring the adsorption of smaller particle over bigger ones.

explain the limits of the assembly process and the employed theoretical calculations. Received: October 9, 2016 Revised: December 8, 2016 Published: January 9, 2017 2437

DOI: 10.1021/acs.jpcc.6b10218 J. Phys. Chem. C 2017, 121, 2437−2443

Article

The Journal of Physical Chemistry C

incubating overnight in a Thermomixer (Eppendorf ThermoMixer C) at 20 °C, the particles were washed by centrifugation with Milli-Q water (see Supporting Information). Negatively Charged Particles. An initial volume of 1 mL of 30 nm AuNPs was concentrated to 0.1 mL and then mixed with 5 μL of a 2% TWEEN 20 (Sigma-Aldrich) solution, 30 μL of 0.1 M phosphate buffer (pH 7), 50 μL of 2 M NaCl solution, 20 μL of 100 nM thiolated DNA solution and 5 μL 100 mM bis(p-sulfonatophenyl)phenylphosphine dihydrate dipotassium (BSPP, Sigma-Aldrich). Following an overnight incubation in the Thermomixer at 20 °C, the mix was washed by centrifugation (see Supporting Information) and redispersed to 1 mL with Milli-Q. Substrates Functionalization. A silica coated (100 nm) silicon wafer cut into 4 by 6 mm pieces was used as substrate for the self-assembly. After a cleaning step with piranha solution, the substrates where incubated in a 95/3/2 vol % solution of 99.9% Ethanol, Milli-Q water and (3-aminopropyl)triethoxysilane (APTES, Sigma-Aldrich) for 1 h. The substrates were then washed with ethanol and baked at 110 °C for 10 min. Particle Assembly. The negative core particles were electrostatically assembled on the functionalized substrates by immersing a substrate in 200 μL of a core particle solution of optical density 0.1 at 65 °C for 2 h. The SiO2/Si samples were then washed in Milli-Q and dried under a nitrogen stream. The so prepared templates were then immersed in a 200 μL mix of 20 and 50 nm satellite particles with a density of 1.7 × 1011 particles/mL for a given time at a defined temperature and afterward washed in Milli-Q again. Particle concentrations of the initial solutions were calculated from UV−vis data of the colloidal solutions as specified by TedPella Inc. based on meanfree-path corrected Mie-theory calculations (Supporting Information, Figure S1).18 The particles were stable for several days before precipitation started to occur.

Figure 2. Interaction energy landscapes: (a) Calculated maps of the total interaction energy (Etot) of a positively charged AuNP (left, 50 nm; right, 20 nm) at position (d, z) and a negative 30 nm AuNP (indicated in yellow) sitting on a positive charged surface at (0, 0). Blue colors indicate attractive and red indicate repulsive interaction energies. The dotted lines mark where the Etot equals the mean thermal energy of the particles (at 20 °C). Inset: Schematic of the geometric setup for the calculations in parts a and b. (b) Dependence of Etot for different sized particles on z when approaching vertically toward the core particle (d = 0). The black dashed line indicates the thermal energy at 20 °C. (c) Maxwell−Boltzmann distribution (orange; see Supporting Information, theoretical modeling) at T = 20 °C. The black dashed line indicates the mean thermal energy of the particles, the blue and the green line mark the energies of the barriers for 50 and 20 nm particles (as shown in part b).

Controlled adsorption techniques like this can be used to selectively deposit particle species from a mixed colloid onto a substrate. This selectivity could be exploited, e.g., for targeted sensing and biomedical applications.2−6 Different from already successfully demonstrated magnetic separation methods,5,7−9 the proposed electrostatically driven process is independent of the nanoparticle material. The only requirement is a suitable surface chemistry to confine charged molecules on the nanoparticle surfaces. These chemistries are well established, as electrostatic self-assembly of nanoparticles already plays an important role in various bottom-up fabrication techniques.10−14 Applications range from decorating solar cells with gold nanoparticles (AuNPs)15 and photocatalysts16 to nanoparticle assemblies for sensing applications.17

■

THEORETICAL CONSIDERATIONS In order to calculate the barrier heights and the effect of the particle size and charge, the electrostatic interactions have been modeled by means of DLVO theory (Derjaguin, Landau, Verwey, and Overbeek).1,19−21 This theory accounts for the attractive interaction between the differently charged particles (Eatt), van der Waals interactions (EVdW) and repulsive electrostatic interaction between approaching particle and surface (Erep). The total interaction energy (Etot) can therefore be written as a linear combination: Etot = Eatt + Erep + EVdW

■

(1)

EVdW was calculated as in previous studies of similar systems (see Supporting Information, Theoretical modeling).1,22,23 The electrostatic interactions were taken into account as linear combinations of sphere-plane and sphere−sphere interactions:

EXPERIMENTAL METHODS Commercially available AuNPs (TedPella Inc.) of three different sizes (20, 30, and 50 nm) were used without further purification. Sizes were confirmed using transmission electron microscopy (TEM). Functionalization of the particles to confine charge at their surface was achieved using known protocols.1 Positively Charged Particles. An initial volume of 1 mL of 20 and 50 nm particles was washed once with 0.9 mL Milli-Q water by centrifugation and then concentrated to 0.1 mL. A 25 μL sample of a 24 mM N,N,N-trimethyl-(11mercaptoundecyl)ammonium chloride (TMAC, ProChimia) solution or 2.4 μL of a 12 mM TMAC solution were added, followed by addition of a cetyltrimethylammonium bromide (CTAB, GFS chemicals) solution (6 mM, 100 μL). After

⎛ k T ⎞2 Erep = 4πε0εr ⎜ B ⎟ r+Y+YAPTES exp( −κz) ⎝ e ⎠

Eatt =

4πε0εr

kBT 2 + − e

( )rr

z + r+ − r −

Y+Y − exp(−κ(z − 2r −))

(2)

(3)

In these equations, ε0 and εr are the permittivity of free space and the relative dielectric constant of water, respectively. The Boltzmann constant is given by kB, the electron charge by e. The radii r+, r‑, and z are geometric parameters as depicted in 2438


Article

The Journal of Physical Chemistry C the inset in Figure 2a. T represents the absolute temperature during the assembly process. κ is the Debye−Hückel parameter, which depends on the ionic strength of the electrolyte solution (see Supporting Information, theoretical modeling). The most important parameters are the scaled effective surface potentials Yx (where x = “+”, “−”, or “APTES”) as they provide a measure for the surface charge of the substrate and particles. They can be approximated according to the method of Ohshima (see Supporting Information, theoretical modeling) and ultimately depend on the surface ζ potentials.24,25 Thus, the variable experimental input parameters for the calculations are the ζ potentials of the particles and substrate along with the temperature and ionic strength of the solution. Calculating Etot at different coordinates (d, z) (Figure 2a, inset) yields the interaction energy maps shown in Figure 2a. The maps show the expected funnel ef fect:19 i.e. a gradient in the (repulsive) interaction energy guiding the approaching particle toward the adsorption site for two different sized satellite particles (left, 50 nm, right, 20 nm). It is also evident that there is a barrier present that particles have to overcome in order to reach the core particle (Figure 2b). At 20 °C the barrier for a 20 nm particle is 50 meV. This is only 0.5kBT (12 meV) higher than its mean thermal (kinetic) energy (Figure 2b, green line). For the 50 nm AuNPs, the barrier is in turn 4.8kBT (120 meV) higher than its mean thermal energy (Figure 2b, blue line). While for the 20 nm particles the barrier height is in a range where experiments have shown high yield electrostatic dimer assembly,1 for the 50 nm ones the barrier is in a range where dimer formation is not efficient.1 To roughly estimate the implications of this difference in barrier heights for the potential assembly of nanoparticles, we have calculate the Maxwell−Boltzmann kinetic energy distribution (Supporting Information, theoretical modeling) of the particles at 20 °C (Figure 2c). In Figure 2c, A20 (A20 = 0.266) and A50 (A50 = 0.007) represent the fraction of the population of the 20 or 50 nm particles (respectively) that have a thermal energy higher than the corresponding barrier. Using this very simple approach, the adsorption of the 20 nm AuNPs is favored (by a factor of 38:1) as there is a higher fraction of 20 nm AuNPs with an energy higher than their adsorption barrier. With these calculated population fractions, we can now compare the concentrations of the different sized particles in the area where the repulsive interaction reaches its maximum (the location of the barrier). The particles are exposed to a repulsive potential corresponding to their size. Therefore, only particles with an energy equal or higher than this potential are expected to be located in this area. Particles with lower energy are driven away to areas with lower Etot. We therefore expect a 38 times higher concentration of 20 nm particles at this point (assuming equal numbers of 20 and 50 nm particles in the initial solution). This, in turn, would lead to a higher probability for 20 nm satellite particles to adsorb on the cores and thus a size-discriminative adsorption process. The theoretical effect of the core and satellite particle ζ potentials on the size-dependence of the barrier was investigated as well. From these calculations (see Supporting Information, Figure S6), it could be seen that in order to increase the barrier height differences for 20 and 50 nm particles, a higher ζ potential for the satellite particles is beneficial and thus a maximization of these potentials was pursued experimentally. Besides these surface charge induced selectivity effect, we can further use the Stokes−Einstein−Sutherland (SES) equation to

estimate how the particle size affects their diffusion and whether that leads to an enhanced selectivity:26

D=

kBT 6πηr

(4)

with η as dynamic viscosity and r the radius of a spherical particle. Because of this inverse proportionality to the particle radius, we expect that the 20 nm particles diffuse 2.5 times faster than the 50 nm ones. This higher diffusion speed is another factor that contributes to the preferential adsorption of the 20 nm particles (assuming a kinetically driven process).

■

EXPERIMENTS In order to investigate the system described above, silica coated silicon (SiO2/Si) substrates and colloidal AuNPs were functionalized with charge carrying molecules (thiolated DNA for negative charge, N,N,N-trimethyl-(11-mercaptoundecyl)ammonium chloride for positively charged AuNPs and (3aminopropyl)triethoxysilane for the substrate). Negatively charged 30 nm AuNPs were then assembled onto the positively charged substrates to provide evenly distributed adsorption centers for positive particles as described by Zheng et al.1 Positively charged 20 and 50 nm AuNPs were mixed to form a bidisperse colloid (satellite particle solution). For the sizeselective adsorption step, dried substrates with the core assembly were immersed in the satellite particle solution for a fixed amount of time at a constant temperature. Assembly performance was assessed using SEM imaging (FEI Nova NanoSEM 430 or FEI Magellan 400 SEM) and UV−vis spectroscopy (for glass substrates, Agilent Technologies Cary 60 UV−vis spectrometer). As a measure of the effectiveness of the size-selective assembly process, we have defined a selectivity factor (SF): SF =

(N20: N50)captured (N20: N50)solution

(5)

Here N is the number of particles, either captured on the substrate or initially present in solution. The selectivity factor is a measure of how much the 20 to 50 nm particle ratio has increased on the substrate when compared to the solution. A SF = 1 means that the substrate is not selective, SF < 1 indicates preferential adsorption of the 50 nm particles. SF values greater than one indicate a selectivity toward the 20 nm particles (e.g., SF = 2 means that the 20 to 50 nm particle ratio doubles on the surface compared to in solution). On the basis of the electrostatic analysis shown above (and neglecting any diffusion effects) one would expect an ideal selectivity factor of up to 75.

■

RESULTS In the following, particle statistics and exemplary SEM images show the adsorption statistics and patterns for different assembly conditions. Adsorption statistics are derived from SEM images. We investigated the effect of the solution temperature, the incubation time and the mixing ratio on the size-selectivity of the electrostatic self-assembly. An effect of the overall particle concentration in the assembly solution on the size-selectivity of the process was not observed, as long as colloidal stability was maintained (Supporting Information). Temperature. In order to vary the mean energy of the particles (dashed black line in Figure 2) and the particle diffusion speed, we varied the temperature at which the 2439


Article


the influence of the incubation time in step 2 of Figure 1 on the fractions of adsorbed 20 and 50 nm satellite particles. This is especially interesting as the process seems to be more selective but slower at lower temperatures. Therefore, 5 and 22 h incubation times (at 14 °C, see Figure 3, parts c and d) were compared, and it was found that the longer time leads to a full coverage of the adsorption sites. At the same time, the amount of adsorbed 50 nm particles increased by 8% with increased time. This reduction of the selectivity might seem contradictory to the theoretical model at first. However, a closer look at the SEM images (Figure 3d) reveals a high yield of multimers (i.e., core particles with more than one satellite particles attached) as opposed to mainly dimers found for the shorter incubation time. That means that the dimer forming model, its barrier heights and the calculated selectivity factor are not applicable anymore. An interaction energy map (similar to the one presented in Figure 2) for a 20 nm particle approaching an already formed core-20 nm dimer shows that the barrier has now greatly increased to 7.3kBT (184 meV). For this barrier, the Maxwell−Boltzmann distribution yields a fraction of population of particles with higher energy of only 0.2%. Furthermore, the model does not account for attractive or repulsive effects resulting from additional core particles in the close proximity of the investigated core. We thus conclude that due to the given reasons, the considerations based on the dimer forming model reflect the experimental situation best until high dimer coverage is achieved and sizeselectivity is reduced after such a state is reached. An additional experiment was undertaken to address the question whether the adsorption process is thermodynamically or kinetically driven. In a thermodynamic equilibrium state, already adsorbed particles would also desorb again and thereby render a core particle available for a new adsorption to take place. In a kinetically driven process, satellite particles would not desorb again once captured by a core particle. Two samples were run though the adsorption process with an incubation time of 5 h. After the incubation, one sample (sample A) was washed as per the protocol we have described above. The second sample (sample B) was immersed in 600 mL of Milli-Q water, being in solution at all times (see Supporting Information, Figure S4, for schematic workflow). For the case of a reversible process (thermodynamically driven) one would expect less satellite particles to be adsorbed on the cores of sample B, due to reversible adsorption−desorption events. The desorption would be driven by the at least five times lower particle concentration in the desorption bath compared to the assembly solution (Supporting Information, Figure S4). Additionally, one would expect a significant difference in the 20 to 50 nm satellite ratio on the surface if a desorption process was size dependent (e.g., due to different van der Waals interaction energies for 20 and 50 nm particles). However, the 20−50 nm particle ratio did not differ significantly for both samples (sample A, 6.1; sample B, 7.7; standard deviation, 1.5). The same is true for the overall yield of adsorbed particles (65% and 68% with a standard deviation of 10%; sample size >1000 counts for each substrate). We thus concluded that no significant desorption of satellite particles occurs after adsorption to a core which indicates a diffusion controlled process rather than a thermodynamically controlled one. This also explains why the considerations presented in Figure 2 are more accurate for shorter incubation times where many unoccupied core particles are available for adsorption of satellite particles.

adsorption process took place, which consequently changes the difference between the Etot of the barrier and the mean thermal energy of the particles (1.5kBT) and the speed of the process (as particle diffusion is slowed down at lower temperatures). Comparing the adsorption behavior of incubations at 21 and 14 °C after 5 h while keeping all other parameters fixed leads to two observations (Figure 3): First, the overall coverage of the

Figure 3. Effect of temperature and incubation time on the assemblies when having equal amounts of 20 and 50 nm particles in solution. (a) Top: A comparison between the percentage of 50 nm particles in the initial assembly solution (orange) and adsorbed on the surface (blue) for different assembly temperatures and incubation times. Bottom: The overall adsorption yield (i.e., core particles with a satellite attached) for the different assembly conditions. (b−d) SEM images showing representative results for the 3 different assembly conditions (20 and 50 nm particles indicate in green and blue, respectively): (b) Incubation at 21 °C for 5 h. (c) Incubation at 14 °C for 5 h. (d) Incubation at 14 °C for 22 h. (scale bars: 250 nm).

adsorption sites is much higher at 21 °C compared to the lower temperature. This is to be expected, as the thermal energy of the particles is lower in the second case, which slows the process down (as at this temperature a lower fraction of the particles has a thermal energy higher than the barrier). The much lower yield also indicates that the particle diffusion toward the surface slows down, which is in accordance with the SES equation for the diffusion constant (see eq 4). Second, the selectivity factor for the lower temperature is SF = 9.8, more than twice as high as for the 21 °C incubation and represents an almost 10-fold increase of the 20 to 50 nm particle ratio when compared to solution. A higher selectivity at lower temperatures is expected, as the Maxwell−Boltzmann distribution is temperature dependent. That leads to a higher theoretical selectivity factor as the fractions of population (Figure 2c, A20 and A50) change for the different sized particles by a different amount (see Supporting Information, Figure S2). The slower process dynamics mentioned above contribute as well to the higher selectivity at lower temperatures as we will see in the following. Incubation Time. The adsorption mechanism is a process where particles must overcome an energy barrier higher than their mean thermal energy. Therefore, it is of interest to investigate the time dependence of the process and compare 2440


Article

The Journal of Physical Chemistry C Mixing Ratio. The dependence of the selectivity on the initial mixing ratios of the two different sized satellite species in the bidisperse colloid was investigated. The results shown here were achieved with an incubation time of 3 h at 20 °C. A summary of the results is shown in Figure 4, where the initial mixing ratios between 20 and 50 nm AuNPs of the different batches (no. 1 to no. 4) were 19:1, 2:1, 1:1, and 1:2 (respectively).

the selectivity of the adsorption process decreases significantly when increasing the 50 nm content further. These observed variations of the selectivity factor show strong dependence of the size-selective adsorption process on the assembly conditions. While the process shows the predicted preference for the adsorption of smaller particles (which cannot be explained based on a simple diffusion model), the observed variations of the selectivity factor cannot be described by the considerations summarized in Figure 2 alone. As discussed earlier, the electrostatic interaction model describes the formation of a dimer, neglecting multiparticle effects (e.g., neighboring core particles, multimer formation or satellite particle interactions in the solution). A possible approach to further enhance the selectivity of the size-selective adsorption process at higher initial mixing ratios is to increase the ζ potential on the TMAC-AuNPS (Supporting Information, Figure S6) which may require the use of different surface-capping ligands. This increase is expected to increase the difference in the adsorption barriers of the different sized particles and therefore also for the adsorption yields for the different species.

■

Figure 4. Effects of the mixing ratio in solution on the selectivity of the adsorption process. Top: SEM-based adsorption statistics showing the amount of adsorbed 50 nm particles on the surface (blue) when using different 20 to 50 nm particle mixing ratios in the initial assembly solution (orange: percentage of 50 nm particles in solution). The batch numbers are a reference for the different mixing ratios (batch 1:19:1, batch 2:2:1, batch 3:1:1, batch 4:1:2). Bottom: The selectivity factor for different mixing ratios calculated from the adsorption statistics according to eq 5.

CONCLUSION In conclusion, we have studied the size selective adsorption process of bidisperse satellite particles on core particles surfaceconfined to a charged substrate. We found the process to be kinetically controlled. The highest selectivity toward adsorption of small particles from a solution containing a bimodal size distribution of Au nanoparticles was achieved for an initial (number) mixing ratio of 1:1. For this case, the small particles were adsorbed preferentially by number ratio (selectivity factor) of 28. Using DLVO theory we explain the driving force of the adsorption process as well as its size-selectivity. Interestingly, the selectivity factor depends strongly on the assembly conditions, which was not reflected by the initial considerations based on the electrostatic dimer forming process. Our calculations also show the effect of the ζ potential of the particles on the barrier height. They suggest that by tuning the ζ potential of otherwise identical particles, a sorting mechanism similar to the here presented one should be achievable. Possible applications of the size-selective adsorption process can be found in the fields of microfluidics and targeted sensing. Unlike other separation methods such as ultracentrifugation or filtration,9,29 the technique presented here does not require a full separation of the different species in the colloid. Additionally, no further preparation (like drop casting) is required for surface deposition for subsequent investigations (e.g., spectroscopy of captured molecules). Once captured, the satellite and core form asymmetric nanodimers providing ideal conditions for a spectroscopy method like surface enhanced Raman spectroscopy, due to the strong plasmonic enhancement of the E-field in the interparticle gap.30−33

Because of the superior number of 20 nm particles in batch 1, it is not surprising to have mainly 20 nm particles adsorbing on the core particles. The number of the small satellites adsorbed on the surface did indeed increase to 98% as can be seen in Figure 4, batch 1, blue bar (indicating a 3-fold increase of the 20 nm particle concentration at the surface). This initial mixing ratio is approximately equal to an optical density ratio of 1:1. Therefore, it was used (at a highly diluted absolute particle concentration) to evidence the preferential adsorption by comparing the UV−vis spectra before and after the adsorption (see Supporting Information, Figure S3). The AuNP colloids show a pronounced plasmon resonance peak in the UV−vis absorbance spectrum, which is at a different wavelength for 20 and 50 nm AuNPs.27,28 The measured spectra for a bidisperse colloid are therefore a sum of those two spectra. Depending on the mixing ratio, the peak position of this sum is closer to the absorption peak of the particle species with the higher optical density (∼particle concentration) in the mix. We could thereby detect a drop of 20 nm particle concentration in the solution during the adsorption experiments, confirming the trend observed by statistical evaluation of the SEM micrographs. When increasing the fraction of the 50 nm particles in the bidisperse colloid further, we did also observe a significant increase in the selectivity factor to a maximum value of 28 which is reached with equal particle ratios in solution (batch no. 3). This result shows that the electrostatic self-assembly method is suitable to increase the fraction of the 20 nm AuNPs to over 96% when starting with a 1:1 mix of 20 and 50 nm AuNPs. An SEM image of the resulting adsorption pattern is depicted in the Supporting Information, Figure S7. Increasing the 50 nm particle fraction further to 67% leads to an increased fraction of adsorbed 50 nm particles of 23.1% indicating that

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpcc.6b10218. More details on theoretical modeling of the assembly process, temperature dependence of the selectivity, UV− vis measurements, schematic of desorption test, schematic of experimental workflow, particle size and ζ 2441


Article


■

(9) Kowalczyk, B.; Lagzi, I.; Grzybowski, B. A. Nanoseparations: Strategies for Size And/or Shape-Selective Purification of Nanoparticles. Curr. Opin. Colloid Interface Sci. 2011, 16 (2), 135−148. (10) Sastry, M.; Rao, M.; Ganesh, K. N. Electrostatic Assembly of Nanoparticles and Biomacromolecules. Acc. Chem. Res. 2002, 35 (10), 847−855. (11) Chapel, J.-P.; Berret, J.-F. Versatile Electrostatic Assembly of Nanoparticles and Polyelectrolytes: Coating, Clustering and Layer-byLayer Processes. Curr. Opin. Colloid Interface Sci. 2012, 17 (2), 97− 105. (12) Nie, Z.; Petukhova, A.; Kumacheva, E. Properties and Emerging Applications of Self-Assembled Structures Made from Inorganic Nanoparticles. Nat. Nanotechnol. 2010, 5 (1), 15−25. (13) Kalsin, A. M.; Fialkowski, M.; Paszewski, M.; Smoukov, S. K.; Bishop, K. J. M.; Grzybowski, B. A. Electrostatic Self-Assembly of Binary Nanoparticle Crystals with a Diamond-Like Lattice. Science (Washington, DC, U. S.) 2006, 312 (5772), 420−424. (14) Bishop, K. J. M.; Wilmer, C. E.; Soh, S.; Grzybowski, B. A. Nanoscale Forces and Their Uses in Self-Assembly. Small 2009, 5 (14), 1600−1630. (15) Reineck, P.; Lee, G. P.; Brick, D.; Karg, M.; Mulvaney, P.; Bach, U. A Solid-State Plasmonic Solar Cell via Metal Nanoparticle SelfAssembly. Adv. Mater. 2012, 24 (35), 4750−4755. (16) Liu, G.; Wang, D.; Zhou, F.; Liu, W. Electrostatic Self-Assembly of Au Nanoparticles onto Thermosensitive Magnetic Core-Shell Microgels for Thermally Tunable and Magnetically Recyclable Catalysis. Small 2015, 11 (23), 2807−2816. (17) Zheng, Y.; Thai, T.; Reineck, P.; Qiu, L.; Guo, Y.; Bach, U. DNA-Directed Self-Assembly of Core-Satellite Plasmonic Nanostructures: A Highly Sensitive and Reproducible Near-IR SERS Sensor. Adv. Funct. Mater. 2013, 23 (12), 1519−1526. (18) Ng, S. H. DNA Directed Self-Assembly of Gold Nanoparticle Structures Using Templated Substrates; Monash University: 2016. (19) Ma, L.-C.; Subramanian, R.; Huang, H.-W.; Ray, V.; Kim, C.-U.; Koh, S. J. Electrostatic Funneling for Precise Nanoparticle Placement: A Route to Wafer-Scale Integration. Nano Lett. 2007, 7 (2), 439−445. (20) Derjaguin, B.; Landau, L. Theory of the Stability of Strongly Charged Lyophobic Sols and of the Adhesion of Strongly Charged Particles in Solutions of Electrolytes. Prog. Surf. Sci. 1993, 43 (1−4), 30−59. (21) Verwey, E. J. W.; Overbeek, J. T. G.; van Nes, K. Theory of the Stability of Lyophobic Colloids: The Interaction of Sol Particles Having an Electric Double Layer; Elsevier Publishing Company: 1948. (22) Tadmor, R. The London-van Der Waals Interaction Energy between Objects of Various Geometries. J. Phys.: Condens. Matter 2001, 13 (9), L195−L202. (23) Kumagai, S.; Yoshii, S.; Yamada, K.; Matsukawa, N.; Fujiwara, I.; Iwahori, K.; Yamashita, I. Electrostatic Placement of Single Ferritin Molecules. Appl. Phys. Lett. 2006, 88 (15), 153103. (24) Ohshima, H. Surface Charge Density/Surface Potential Relationship for a Spherical Colloidal Particle in a Solution of General Electrolytes. J. Colloid Interface Sci. 1995, 171 (2), 525−527. (25) Ohshima, H. Effective Surface Potential and Double-Layer Interaction of Colloidal Particles. J. Colloid Interface Sci. 1995, 174 (1), 45−52. (26) Einstein, A. Ü ber Die von Der Molekularkinetischen Theorie Der Wärme Geforderte Bewegung von in Ruhenden Flüssigkeiten Suspendierten Teilchen. Ann. Phys. 1905, 322 (8), 549−560. (27) Jain, P. K.; Lee, K. S.; El-Sayed, I. H.; El-Sayed, M. A. Calculated Absorption and Scattering Properties of Gold Nanoparticles of Different Size, Shape, and Composition: Applications in Biological Imaging and Biomedicine. J. Phys. Chem. B 2006, 110 (14), 7238− 7248. (28) Sönnichsen, C.; Franzl, T.; Wilk, T.; von Plessen, G.; Feldmann, J. Plasmon Resonances in Large Noble-Metal Clusters. New J. Phys. 2002, 4, 93. (29) Akthakul, A.; Hochbaum, A. I.; Stellacci, F.; Mayes, A. M. Size Fractionation of Metal Nanoparticles by Membrane Filtration. Adv. Mater. 2005, 17 (5), 532−535.

potential dependence of the adsorption barrier height, and sample SEM images used for statistical analysis (PDF)

AUTHOR INFORMATION

Corresponding Author

*(U.B.) E-mail: [email protected]. ORCID

Udo Bach: 0000-0003-2922-4959 Present Address ⊥

School of Applied Science, RMIT University, Melbourne VIC 3000, Australia. Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This study has been supported by CSIRO through the OCE Science Leader program and the Australian research Council through an Australian Research Fellow grant (DP110105312) to U.B. and a Future Fellowship to D.E.G. (FT140100514). This work was performed in part at the Melbourne Centre for Nanofabrication, the Victorian Node of the Australian National Fabrication Facility, a company established under the National Collaborative Research Infrastructure Strategy to provide nanoand microfabrication facilities for Australia’s researchers. The authors acknowledge use of facilities within the Monash Centre for Electron Microscopy. This research used equipment funded by Australian Research Council Grant LE140100104.

■ ■

ABBREVIATIONS AuNP, gold nanoparticle; SF, selectivity factor REFERENCES

(1) Zheng, Y.; Rosa, L.; Thai, T.; Ng, S. H.; Gómez, D. E.; Ohshima, H.; Bach, U. Asymmetric Gold Nanodimer Arrays: Electrostatic SelfAssembly and SERS Activity. J. Mater. Chem. A 2015, 3 (1), 240−249. (2) Guo, L.; Chen, L.; Hong, S.; Kim, D.-H. Single Plasmonic Nanoparticles for Ultrasensitive DNA Sensing: From Invisible to Visible. Biosens. Bioelectron. 2016, 79, 266−272. (3) Wen, J.; Chen, J.; Zhuang, L.; Zhou, S. Designed Diblock Hairpin Probes for the Nonenzymatic and Label-Free Detection of Nucleic Acid. Biosens. Bioelectron. 2016, 79, 656−660. (4) Tong, L.; Righini, M.; Gonzalez, M. U.; Quidant, R.; Käll, M. Optical Aggregation of Metal Nanoparticles in a Microfluidic Channel for Surface-Enhanced Raman Scattering Analysis. Lab Chip 2009, 9 (2), 193−195. (5) Munir, A.; Zhu, Z.; Wang, J.; Zhou, H. S. FEM Analysis of Magnetic Agitation for Tagging Biomolecules with Magnetic Nanoparticles in a Microfluidic System. Sens. Actuators, B 2014, 197, 1−12. (6) Heuer-Jungemann, A.; Harimech, P. K.; Brown, T.; Kanaras, A. G. Gold Nanoparticles and Fluorescently-Labelled DNA as a Platform for Biological Sensing. Nanoscale 2013, 5 (20), 9503−9510. (7) Lee, S. H. S.; Hatton, T. A.; Khan, S. A. Microfluidic Continuous Magnetophoretic Protein Separation Using Nanoparticle Aggregates. Microfluid. Nanofluid. 2011, 11 (4), 429−438. (8) Bu, M.; Christensen, T. B.; Smistrup, K.; Wolff, A.; Hansen, M. F. Characterization of a Microfluidic Magnetic Bead Separator for HighThroughput Applications. Sens. Actuators, A 2008, 145−146 (1−2), 430−436. 2442


Article

The Journal of Physical Chemistry C (30) Yang, A.; Huntington, M. D.; Cardinal, M. F.; Masango, S. S.; Van Duyne, R. P.; Odom, T. W. Hetero-Oligomer Nanoparticle Arrays for Plasmon-Enhanced Hydrogen Sensing. ACS Nano 2014, 8 (8), 7639−7647. (31) Punj, D.; Regmi, R.; Devilez, A.; Plauchu, R.; Moparthi, S. B.; Stout, B.; Bonod, N.; Rigneault, H.; Wenger, J. Self-Assembled Nanoparticle Dimer Antennas for Plasmonic-Enhanced SingleMolecule Fluorescence Detection at Micromolar Concentrations. ACS Photonics 2015, 2 (8), 1099−1107. (32) Biswas, S.; Liu, X.; Jarrett, J. W.; Brown, D.; Pustovit, V.; Urbas, A.; Knappenberger, K. L.; Nealey, P. F.; Vaia, R. a. Nonlinear ChiroOptical Amplification by Plasmonic Nanolens Arrays Formed via Directed Assembly of Gold Nanoparticles. Nano Lett. 2015, 15 (3), 1836−1842. (33) Yan, B.; Thubagere, A.; Premasiri, W. R.; Ziegler, L. D.; Dal Negro, L.; Reinhard, B. M. Engineered SERS Substrates with Multiscale Signal Enhancement: Nanoparticle Cluster Arrays. ACS Nano 2009, 3 (5), 1190−1202.

2443


Size Selective Adsorption of Gold Nanoparticles by ... - ACS Publications

Recommend Documents