Water-Mediated Assembly of Gold Nanoparticles into Aligned One

Jun 15, 2015 - Water-Mediated Assembly of Gold Nanoparticles into Aligned One-Dimensional ... for applications in electronics, determines their assemb...
1 downloads 0 Views 3MB Size
Download PDF
Article pubs.acs.org/Langmuir

Water-Mediated Assembly of Gold Nanoparticles into Aligned OneDimensional Superstructures Jean-Nicolas Tisserant,*,† Patrick A. Reissner,† Hannes Beyer,† Yuriy Fedoryshyn,‡ and Andreas Stemmer*,† †

Nanotechnology Group, ETH Zurich, Säumerstrasse 4, CH-8803 Rüschlikon, Switzerland Institute of Electromagnetic Fields, ETH Zurich, Gloriastrasse 35, CH-8092 Zurich, Switzerland

‡

S Supporting Information *

ABSTRACT: This Article shows that water in ethanol colloids of gold nanoparticles enhances the formation of linear clusters and, more important for applications in electronics, determines their assembly on surfaces. We show by dynamic light scattering that ethanol colloids contain mainly monomers and dimers and that wormlike superstructures are mostly absent, despite UV−vis evidence of aggregation. Water added to the colloid as a cosolvent was found to enhance the number of clusters as well as their average size, confirming its role in linear self-assembly, on the scale of a few particles. Water adsorbed from the atmosphere during coating was also found to be a powerful lever to tune self-assembly on surfaces. By varying the relative humidity, a sharp transition from branched to linear superstructures was observed, showing the importance of water as a cosolvent in the formation of cluster superstructures. We show that one-dimensional superstructures may form due to long-range mobility of precursor clusters on wet surfaces, allowing their rearrangement. The understanding of the phenomenon allows us to statistically align both clusters and resulting superstructures on patterned substrates, opening the way to rapid screening in molecular electronics.

■

INTRODUCTION The recent discovery of low dimension gold nanoparticle aggregates such as nanoparticle lines, rings, or bridges promised a myriad of applications ranging from optics and electronics to sensing.1−13 Many of these applications make use of novel optoelectronic properties arising from the coupling of nanoparticles. In particular, the electronic coupling between neighboring particles is responsible for an improvement of the efficiency of surface enhanced Raman scattering (SERS).14−17 While methods to fabricate two- or threedimensional (2D or 3D) arrays of gold particles are now common,18−23 One-dimensional (1D) structures were shown only recently due to difficulties in fabrication and the presumption that spherical particles are ill-suited for linear aggregation.12,24 The main routes toward linear aggregates of particles in the bulk are based on chemical functionalization and cross-linking of particles with thiols,25,26 or oligonucliotides.27 The destabilization of aqueous colloids by adding salt to trigger the formation of linear aggregates was also explored.28,29 The spontaneous self-assembly of gold nanoparticles in polar solvents such as ethanol to form linear aggregates was first reported by Liao et al.30 and then consistently observed.28,31−33 The mechanism behind the phenomenon is still under discussion. While most early studies insisted on the role of ethanol itself,28,30,34 Han et al. recently documented the importance of residual salt in the process as well as the reversible character of the aggregates in the bulk.11 Surprisingly, © 2015 American Chemical Society

the role of residual water was not clarified even if assemblies were typically observed in binary solvents, inherent from the synthesis or intentionally mixed. Furthermore, bulk assembly was explored in terms of UV−vis absorption mainly.11,30,31 Much less attention was paid to the size distribution of colloidal particles. While the shift in the plasmon resonance is a powerful tool to sense assembly, it can be misleading in terms of assembly efficiency due to the dramatic variation of the scattering cross-section with the number of particles per aggregate.35 In this study, we shed light on the particle distribution and the role of residual water in ethanol colloids using dynamic light scattering (DLS). Here we show that the vast majority of clusters is composed of monomers, dimers or trimers and that water enhances the formation of larger aggregates. Large linear aggregates observed on surfaces arise mainly from the assembly of clusters containing three or less particles upon solvent evaporation. The ability to control the assembly of gold nanoparticles on surfaces is expected to have a strong impact on their implementation in electronic devices. Self-assembly using dewetting is a promising method to form regular arrays of nanoparticles.36 For example by a combination of dewetting, coffee-stain effect37 and Marangoni instability,38 both lines39,40 or rings5,41 of nanoparticles could be manufactured. Similarly Received: March 27, 2015 Revised: June 11, 2015 Published: June 15, 2015 7220

DOI: 10.1021/acs.langmuir.5b01135 Langmuir 2015, 31, 7220−7227

Article

Langmuir

Zetasizer 7.10 software. Mie theory was used to compute volume distributions from the original intensity distributions shown in Figure S2 in the Supporting Information. Particle attenuance spectra were measured on a SmartSpec Plus spectrophotometer (Bio-Rad). Scanning electron microscopy was performed on a FEI NanoSEM instrument, at a typical voltage of 3 kV. The images were analyzed using ImageJ. For the fractal dimension analysis, the images were made binary by application of a threshold allowing isolated particles to be distinguished. The box counting method (ImageJ) was then applied on images of whole droplets. For assessing the alignment of aggregates, linear clusters were considered as separate objects.

here, water in ethanol was found to induce aggregation of clusters into 1D superstructures via wetting transitions. The drawback of these methods is a poor control over the exact position of the nanoparticles due to the random nature of dewetting. Guided assembly methods on the contrary allow forming 1D assemblies with single particle resolution on different patterned substrates like poly(dimethylsiloxane)13,42 (PDMS) or silicon patterned with hydrogen silsesquioxane43 (HSQ) and to control the interparticle distance.44 To our knowledge, previous studies on guided assembly have concentrated on single particles in trenches where the confinement was close to the particle diameter. We show here that spin-coating particle clusters on patterned HSQ surfaces can guide their overall orientation even in patterns that do not confine single particles. These results may find an echo in the field of molecular electronics where a precise control over the orientation of 1D nanoparticle assemblies remains a technical bottleneck.

■

■

RESULTS AND DISCUSSION Threshold Effect of Water on Self-Assembly. Due to synthesis, atmosphere, or intentional addition, water is present as a cosolvent in all experiments shown in related literature. However, the role of traces of water in linear nanoparticle selfassembly was largely unexplored. In a first experiment, we show that the self-assembly of gold nanoparticles on Si/SiO2 substrates dramatically depends on the relative humidity (RH) of the atmosphere in which the film was coated. In these experiments, before adding ethanol, the nanoparticles were washed once in deionized water to remove salts and reactants from the synthesis. The concentration of nanoparticles was fixed to the concentration obtained directly after synthesis (C0). The nanoparticles were drop-cast from ethanol by a capillary method under controlled relative humidity at 23 °C (see the Experimental Section). Below 40% RH (Figure 1a), the surface is covered randomly with a distribution of separated branched linear aggregates as previously reported.31 Each wormlike aggregate is composed of several smaller chains

EXPERIMENTAL SECTION

Particle Synthesis. Citrate-protected gold nanoparticles (AuNPs) were synthesized following standard recipes.45 Briefly, 80 mL of DI water and 0.0125% (w/w) hydrogen tetrachloroaurate (III) trihydrate (Alpha Aesar 99.99%) were heated under stirring until boiling. A second solution containing 16 mL of water, 4 mL of 1% sodium citrate, and 60 μL of a 1% tannic acid (Sigma-Aldrich) in water was heated to 80 °C and quickly poured into the first solution. After 10 min of heating and stirring, the solution slowly cooled down to room temperature. Dynamic light scattering (DLS) showed that the particles had an average diameter of 15.0 nm with a polydispersity index (PDI) of 0.04 (see the Supporting Information). The colloids produced this way had a concentration C0 = 1.7 × 109 mL−1. The synthesis of particles with diameters of 35 and 55 nm is described in Figure S5 in the Supporting Information. Transfer to Ethanol. A volume of 1 mL of solution from the synthesis was precipitated (9.2 × 104 m/s2, 15 min), and the supernatant was removed. Then 1 mL of deionized water was added and the process was repeated to remove salts from the synthesis. After the supernatant was removed, 1 mL of a solution containing 0.0−25 vol % water in ethanol (99.8%, Fluka Analytical) was added. The colloid was sonicated for 30 s. In all cases, the colloid in ethanol was slightly violet (see the Supporting Information for UV−vis spectrum) and was used within the 15 h stability window (see the Supporting Information). Particle Deposition. The tip of a 20 μL filter tip (ErgoOne) was filled (1 μL) by capillarity in the colloid and brought in contact with the surface. The liquid wets Si/SiO2 with a 0° contact angle, and capillarity empties the tip. The relative humidity (RH) was monitored in situ in a homemade chamber. The measurement error for our setup is ca. 2% relative humidity. Spin-coating was done at 850 rpm, 30 s for all experiments. For successive ethanol rinsing, 300 μL of pure ethanol was spin-coated immediately on the pristine nanoparticle films. Pattern Formation. The nanostructures were fabricated on a Si wafer with a 50 nm thick negative tone electron-beam resist (HSQ002, Dow Corning). The wafer was cleaned in piranha solution and then dehydrated at 180 °C prior to resist deposition (spin-coating, 3000 rpm). The pattern was written in a Vistec EBPG5200 (100 kV electron beam) instrument and then developed in a 1:3 mixture of AZ 351B developer (AZ Electronic materials) and water. Characterization. Particle size distributions were measured by DLS on a Zetasizer Nano ZS (Malvern Instruments) apparatus equipped with a He−Ne Laser (633 nm). In all DLS experiments, the concentration was C0. Measurements were taken at 25 °C, at a scattering angle of 173°. Samples were left to settle 120 s before measurement in 1.5 mL disposable polystyrene cuvettes (Brand). For each sample, six measurements of 15 runs (10 s) each were taken and averaged. The data was interpreted with the associated Malvern

Figure 1. Threshold effect of relative humidity on AuNP self-assembly. Precursor clusters are marked with white arrows: (a) Branched wormlike assembly obtained at 21% relative humidity; the inset is a 5 μm zoom out. (b) Unbranched linear assembly obtained from the same solution at 43% relative humidity; the inset is a 40 μm zoom out. (c) Droplet diameter as a function of the relative humidity. Dashed lines show the average droplet size distribution. 7221

DOI: 10.1021/acs.langmuir.5b01135 Langmuir 2015, 31, 7220−7227

Article

Langmuir highlighted with white arrows in Figure 1a, which we call precursor clusters. They are typically linear assemblies of 2−20 nanoparticles. Above 43% RH, the nanoparticles are all confined in droplets with sizes in the range of some tens of micrometers, measured by scanning electron microscopy (SEM) (inset in Figure 1b) The rest of the surface is depleted of particles. Precursor clusters are also observed within the droplets, they are however not systematically linear and their superstructures typically do not show branching (Figure 1b). The formation of droplets above a threshold relative humidity can be attributed to the condensation of water on the evaporating ethanol film, and the consecutive Marangoni− Benard instability.38,46,47 Adding water to the colloid varies the size of the droplets but not their tendency of formation, or the assembly patterns within (see Figure S1, Supporting Information), suggesting a combined effect of water present in the bulk and in the atmosphere. The two different structures were described in the past as diffusion-limited colloid aggregation (DLCA) for the branched aggregates, and reaction-limited colloid aggregation (RLCA) for the more compact aggregates.48 In the first case, the electrostatic barrier does not prevent aggregation, which is solely diffusioncontrolled. In the second case, the attachment of the particles is electrostatically retarded and limits the attachment. It is clear from the result shown in Figure 1 that the spatial distribution of ethanol and water as a cosolvent around particles and clusters plays a major role on their linear self-assembly. This point will be discussed in the following, along with data on the size distribution of clusters. Size Distribution of Clusters. Depositing the particles under varying relative humidity revealed consistently the presence of precursor clusters and the role of water on their self-assembly. The colloids were characterized by DLS (i) to measure the size distribution of clusters in the suspension before coating, (ii) to assess the role of water in bulk assembly ,and (iii) to confirm their stability in the experimental time scale. DLS first shows that the colloids are stable for at least 15 hours after adding ethanol (see Figure S2, Supporting Information). To give a realistic image of the particle population, the volume distribution was chosen instead of the intensity distribution that is optically more sensitive to large aggregates compared to smaller ones.49 The volume distribution of particles in ethanol/water mixtures, normalized to the single particle diameter (R/R0, R0 = 15 nm) is given in Figure 2. In all samples, monomers (R/R0 = 1) and dimers (R/R0 = 2) represent over 70% in volume, trimers (R/R0 = 3) around 10% (Figure 2a). Larger clusters (R/R0 > 4) represent less than 20% in volume. Their distribution shows a minimum below R/R0 = 10 particles and a broad maximum that depends on the volume fraction of added water (Figure 2b). The transfer of particles from water to ethanol (see methods) induces water contents close to 0.1 vol %. When no water is added, the distribution is centered at 250 ± 100 nm, corresponding to R/R0 = 20 ± 8 and represents only 4% of the total volume. Adding water volume fractions between 0% and 1.0% affects the number of clusters but not the position of the maximum. The volume of aggregates larger than four particles reaches 11% in this interval. Between 5% and 25% of added water, the maximum gets significantly shifted toward larger particles and the total volume of aggregates reaches 15%. The distribution is now centered at 650 ± 250 nm corresponding to R/R0 = 50 ± 19. This distribution of particles containing a majority of monomers, dimers and a small fraction of larger oligomers was indeed

Figure 2. Volume distribution of colloidal clusters. (a) DLS volume distribution; insets show examples of spin-coated clusters, C0/10, 1.0 vol % added water. (b) Zoom in the cluster region. Red curves, no added water; gray curves, 0.1−1.0 vol % added water; cyan curves, 5− 25 vol % added water. No trend was observed within these windows, and the curves shown represent boundary values of multiple measurements.

observed by SEM on films spin-coated on SiO2 (insets of Figure 2). UV−vis studies usually presented in literature11,30 can be misleading in terms of cluster distribution as the scattering cross-section depends dramatically on the size of the cluster.35 As a comparison with previously published work, the UV−vis spectra of our particles dispersed in water and ethanol mixtures are shown in Figure S3 in the Supporting Information. A clear red shift is observed even if the sample contains a small volume of extended aggregates. UV−vis also confirms that the amount of aggregates increases with the volume of added water. Formation of Clusters. Linear clusters were typically not observed in pure solvents11,28,34 or in the absence of added salt.29 Our experiments demonstrate that water plays an important role both in the colloidal behavior and the morphology on surfaces. Because of the variety of solvent mixtures containing water in which the phenomenon was observed and the aggregation in the absence of salt in our case (see Figure S4, Supporting Information), we make the hypothesis that anisotropic aggregation may arise from an anisotropic distribution of two solvents present intentionally or by contamination. This hypothesis, sketched in Figure 3, is supported by recent results in the field of modeling. Molecular simulation showed that charged polymer nanoparticles in water−ethanol mixtures are covered by a layer of pure water (ca. 5 Å) and that a gradient develops over 2.5 nm outward.50 A similar behavior is expected for charged gold nanoparticles in the same binary solvent mixture. Once two nanoparticles meet upon collision, water forms a capillary bridge between them.51 This distribution of water generates a gradient in the dielectric constant (ε) from the middle to the apex of the dimer as εwater/ εethanol ≈3.3. This gradient results in an anisotropy of the Debye length, larger at the center than at the apexes. The probability for a monomer to attach to the dimer is higher where the Debye length is shorter, i.e. at the apexes. 7222

DOI: 10.1021/acs.langmuir.5b01135 Langmuir 2015, 31, 7220−7227

Article

Langmuir

Figure 3. Role of water distribution on self-assembly. Top: low water content induces an anisotropy in the Debye length forcing the monomers to attach at the apexes of dimers. Bottom: high water content reduces this anisotropy, leading to 2D clusters.

Our DLS and UV−vis experiments show that, within the studied window (>0.1% water), adding water to the colloid influences the distribution of particles, generating more aggregates in a first step and larger aggregates in a second step. This suggests that water acts as a binding agent between charged nanoparticles. It is worth noting that the absolute water threshold triggering assembly in colloidal solvent mixtures is below the experimental minimum shown here. At C0 and considering a 2 nm water layer covering each particle, 35 ppb of water (≈10−6 %) would be enough to trigger the assembly. Our hypothesis is furthermore supported by the morphology of aggregates on solid substrates. Branched DLCA aggregates are observed when water is absent from the atmosphere, whereas unbranched RLCA aggregates are observed for higher water content within the Marangoni droplets. This suggests a more uniform electrostatic screening when more water is present.48 The hypothesis of anisotropic solvent distribution was furthermore tested by varying the size of the particles. 35 and 55 nm particles were synthesized by decreasing the concentration of sodium citrate (see Figure S5, Supporting Information). Similarly to the 15 nm particles, the ethanol colloids of 35 and 55 nm particles contain only a small fraction of clusters. Their fraction decreases with increasing diameter of the particles, as large clusters of 35 and 55 nm particles tend to precipitate (see Figure S6, Supporting Information). Figure 4 summarizes the morphologies obtained for different particles sizes and water contents at similar surface coverage. The structures obtained are self-similar for the three particle sizes. For 55 nm particles the resulting Marangoni droplets contain almost exclusively monomers and the RLCA superstructures are less developed than for smaller particles. However, extended DLCA superstructures were obtained when only residual water from the synthesis is present. A higher magnification of such a structure is shown in Figure S7 in the Supporting Information. These results confirm that the ratio between water and ethanol in the environment is decisive for the formation of 1D superstructures. For high water content, the water layer is not affected by the particle size and the RLCA morphologies are hindered by an increase in the net charge of the particle. For low water content, the capillary force between two bridged particles increases with the particle size,51 leading to extended DLCA morphologies. To summarize, water enhances aggregation when added to the ethanol colloid and shows a threshold effect when brought to the drying film from the atmosphere. This has severe

Figure 4. Self-similarity of morphologies with increasing particle size. Left column: clusters of 15, 35, and 55 nm particles observed within Marangoni droplets. Middle column: corresponding RLCA superstructures obtained in the presence of added water (RH > 50%). Right column: DLCA superstructures obtained in the absence of added water (RH < 30%).

consequences on the morphologies observed, for instance, by drop casting on electron microscopy grids without controlling the relative humidity. Table 1 summarizes the main experimental parameters explored in this study and their morphological outcome. Table 1. Summary of Relevant Experimental Parameters sample colloid dropcast dropcast spincast spincast

RH (%)

water (vol %)

concn (C/C0)

morphology

DLS/UV−vis SEM

43% allows placing 1D structures at the rims of micrometer-sized droplets as already shown in literature for other dewetting systems.36,41 By optimizing the nanoparticle concentration with respect to the droplet size and upon successive spin-coating steps, large 1D objects could be assembled at the rims of droplets (see Figure S12, Supporting Information). An example is given in Figure 6. At C0/10, the particles form extended, separated 1D monolayer networks reaching several micrometers in length and offering a small number of percolating pathways. Such assemblies at the rim of droplets are however ill-suited if one aims to electrically contact them in an optoelectronic device where the position of the aggregates should be better controlled. To reach this aim, AuNPs were coated from ethanol on topographically patterned HSQ templates. The 1D HSQ structures are composed of 50 nm high square ridges separated by gaps between 50 and 150 nm. After coating, no particles could be detected on the ridges, all were deposited between the ridges (Figure 7a−d). This again confirms that the adhesion of the precursor clusters is sufficiently low on SiO2 to allow their displacement. For aggregates covering several lines, the chains rupture upon contact with the ridges (see Figure S13 Supporting Information). A fundamental result is that the nanoparticle chains do not arrange randomly with respect to the direction of the 1D patterns on the HSQ substrate. Figure 7a−c shows chains self-assembled at 0°, 45°, and 90°, respectively. Because longer aggregates tend to form angles larger than 45°, a large portion bridges the gap between two parallel lines as shown in Figure 7c and d. To prove the statistical organization of aggregates in contact with a ridge, the orientation of N = 829 aggregates was measured, both in the RH > 43% and RH < 40% regimes, after one coating and one

Figure 5. Morphological transition upon consecutive spin-coating. All samples were coated in air, RH = 52% at 850 rpm from AuNP suspensions in ethanol. (a) C0 pristine film; (b) one additional ethanol rinsing step; (c) three rinsing steps; (d) fractal dimension of droplets as a function of the normalized concentration (C/C0). Red, pristine film; gray, one additional rinsing step; blue, three rinsing steps.

number of rinsing steps, forming dispersed 1D superstructures (Figure 5c). During this process, the clusters are not dispersed on the substrate beyond two times the initial droplet radius and the coverage inside the droplets remains constant with successive coating (ca. 11% of the surface for films coated at C0; see Figure S8, Supporting Information). This suggests that the adhesion of small clusters on SiO2 prevents their resuspension but allows their transport and rearrangement on the surface. The transport of clusters occurs on length scales of some hundreds of nanometers (evaluated from SEM images), mostly inward, causing a slight decrease in the droplet radius (Figure 5a−c). The morphological transition was quantified by measuring the evolution of the fractal dimension (FD) of the droplets obtained for different nanoparticle concentrations upon multiple rinsing. FDs were calculated by box counting.52 Results are shown in Figure 5d. The fractal dimension of the droplets consistently decreases with the number of spin-coating steps, showing for each initial particle concentration a transition from a disclike shape (FD close to 2) to a coffee-stain regime (FD close to 1) to dispersed 1D objects (FD < 1). The superstructures formed this way retain the linear character of the clusters, as mobile clusters attach to the apexes or kinks of larger aggregates (Figure 6). To illustrate the phenomenon, examples of partially built linear superstructures and neighboring free building blocks are given in Figure S9 in the Supporting Information. Directed capillary flows within the colloidal film may also play an important role in limiting the formation of branched 2D superstructures inherent from a random aggregation process (see Figure S9, Supporting Information). The ability to displace clusters of particles by spin-coating of solvent while enhancing 1D superstructures is, to our knowledge, conceptually novel and may be an interesting post-treatment to modify the morphology of deposited particle films. The concept can be related to solvent annealing that induces crystallization and realignment of organic thin films.53 While rinsing on a spin-coater was applied in our experiments 7224

DOI: 10.1021/acs.langmuir.5b01135 Langmuir 2015, 31, 7220−7227

Article

Langmuir

Figure 7. Alignment of nanoparticle chains on HSQ templates. Preferred angles with respect to the patterns are spontaneously adopted by aggregates. Examples are given for (a) 0°; (b) 45°; (c) 90° bridging; (d) 45° bridging; (e) statistical distribution of the orientations of 829 aggregates. Gray columns, ratio between the number of aggregates at a given angle and the total aggregate count; white circles, aggregate counts; red diamonds, average number of particles per aggregate at 0°, 45°, and 90°.

previously aligned clusters. The presence of capillary bridges either between particles or between particles and the surface determine the optimal positioning of shorter chains with respect to larger, fixed ones. The population at 45° (19% of counted aggregates) may arise from a tilt of the 90° population until the aggregate feels the electrostatic repulsion of the HSQ ridge.

rinsing step. The results are summarized in Figure 7e. Three observations can be made from the data. First, over 77% of all aggregates touching a ridge do so at an angle of either 0° (38%), 45° (19%), or 90° (21%) with an uncertainty of ±2°. Second, smaller aggregates tend to align parallel to the patterned lines with an average number of particles n = 5. Larger aggregates align more often at 45° (n = 10) and 90° (n = 11). Third, small populations of aligned aggregates appear at 20° and 60°; they represent less than 4% of the total number of aggregates each. It is worth noting that the width between HSQ ridges is not sufficiently narrow to confine single particles, and the statistical alignment was observed for confinements in the range of 50−150 nm (see Figure S14, Supporting Information). We propose three attachment paths to explain the observed alignment of the aggregates and their respective occurrence. The first path is the attachment of small clusters along the HSQ lines, at 0° (38% of counted clusters). It occurs because mobile clusters can form twice as many particle-substrate bonds with a ridge compared to plain SiO2. This attachment is surfacemediated. The second path is a direct attachment form the colloid and follows the same mechanism as for the formation of the linear clusters. Considering that HSQ is a negatively charged surface,54 dimers and trimers preferably attach at 90° with respect to the HSQ lines due to an anisotropy in the Debye length around them (21% of counted aggregates). Similarly to the liquid bridging between two particles, a capillary bridge may form between the particles and the surface.51 The third path is flow-mediated. Clusters weakly bonded to the plain surface rearrange as described earlier. This leads to the growth of superstructures that anchor to the

■

CONCLUSION

We have shown by DLS and SEM that ethanol colloids of citrate-protected gold nanoparticles contain, in volume, mainly short linear clusters. While most studies have focused on the role of ethanol or added salt in the colloidal behavior of the aggregates, we found that water enhances the formation and growth of clusters, at low salt concentration. By way of relative humidity, water was also found to be a powerful trigger to tune morphologies of aggregates on solid substrates by inducing a sharp wetting transition above a critical limit of 43%. Our results strongly suggest the implication of an anisotropic water/ ethanol distribution in the linear assembly of clusters. It was furthermore found that the clusters may rearrange over hundreds of nanometers upon multiple solvent rinsing, acting as building blocks for the formation of linear superstructures. We believe that this cluster mobility enables novel postdeposition assembly strategies. Taking advantage of this effect, the linear clusters could be statistically aligned on patterned HSQ substrates. This work opens the way, through the robustness and simplicity of the exposed method, to rapid screening in the field of molecular and organic electronics. 7225

DOI: 10.1021/acs.langmuir.5b01135 Langmuir 2015, 31, 7220−7227

Article

Langmuir

■

(9) Tam, J. M.; Murthy, A. K.; Ingram, D. R.; Nguyen, R.; Sokolov, K. V.; Johnston, K. P. Kinetic Assembly of near-IR-Active Gold Nanoclusters Using Weakly Adsorbing Polymers to Control the Size. Langmuir 2010, 26, 8988−8999. (10) Wang, M.-H.; Li, Y.-J.; Xie, Z.-X.; Liu, C.; Yeung, E. S. Fabrication of Large-Scale One-Dimensional Au Nanochain and Nanowire Networks by Interfacial Self-Assembly. Mater. Chem. Phys. 2010, 119, 153−157. (11) Han, X.; Goebl, J.; Lu, Z.; Yin, Y. Role of Salt in the Spontaneous Assembly of Charged Gold Nanoparticles in Ethanol. Langmuir 2011, 27, 5282−5289. (12) Tang, Z.; Kotov, N. A. One-Dimensional Assemblies of Nanoparticles: Preparation, Properties, and Promise. Adv. Mater. 2005, 17, 951−962. (13) Rey, A.; Billardon, G.; Lörtscher, E.; Moth-Poulsen, K.; StuhrHansen, N.; Wolf, H.; Bjørnholm, T.; Stemmer, A.; Riel, H. Deterministic Assembly of Linear Gold Nanorod Chains as a Platform for Nanoscale Applications. Nanoscale 2013, 5, 8680−8688. (14) Porter, M. D.; Lipert, R. J.; Siperko, L. M.; Wang, G.; Narayanan, R. SERS as a Bioassay Platform: Fundamentals, Design, and Applications. Chem. Soc. Rev. 2008, 37, 1001−1011. (15) Camden, J. P.; Dieringer, J. A.; Wang, Y.; Masiello, D. J.; Marks, L. D.; Schatz, G. C.; Van Duyne, R. P. Probing the Structure of SingleMolecule Surface-Enhanced Raman Scattering Hot Spots. J. Am. Chem. Soc. 2008, 130, 12616−12617. (16) Li, T.; Liu, D.; Wang, Z. Microarray-Based Raman Spectroscopic Assay for Kinase Inhibition by Gold Nanoparticle Probes. Biosens. Bioelectron. 2009, 24, 3335−3339. (17) Maher, R. C.; Maier, S. A.; Cohen, L. F.; Koh, L.; Laromaine, A.; Dick, J. A. G.; Stevens, M. M. Exploiting SERS Hot Spots for DiseaseSpecific Enzyme Detection. J. Phys. Chem. C 2010, 114, 7231−7235. (18) Kotov, N. A.; Meldrum, F. C.; Wu, C.; Fendler, J. H. Monoparticulate Layer and Langmuir-Blodgett-Type Multiparticulate Layers of Size-Quantized Cadmium Sulfide Clusters: A ColloidChemical Approach to Superlattice Construction. J. Phys. Chem. 1994, 98, 2735−2738. (19) Collier, C. P.; Vossmeyer, T.; Heath, J. R. Nanocrystal Superlattices. Annu. Rev. Phys. Chem. 1998, 49, 371−404. (20) Murray, C. B.; Kagan, C. R.; Bawendi, M. G. Synthesis and Characterization of Monodisperse Nanocrystals and Close-Packed Nanocrystal Assemblies. Annu. Rev. Mater. Sci. 2000, 30, 545−610. (21) Han, X.; Li, Y.; Wu, S.; Deng, Z. A General Strategy toward pHControlled Aggregation-Dispersion of Gold Nanoparticles and SingleWalled Carbon Nanotubes. Small 2008, 4, 326−329. (22) Kuznetsov, A. I.; Kiyan, R.; Chichkov, B. N. Laser Fabrication of 2D and 3D Metal Nanoparticle Structures and Arrays. Opt. Express 2010, 18, 21198−21203. (23) Fernandez, C. A.; Wai, C. W. A Simple and Rapid Method of Making 2D and 3D Arrays of Gold Nanoparticles. J. Nanosci. Nanotechnol. 2006, 6, 669−674. (24) Pramod, P.; Thomas, K. G. Plasmon Coupling in Dimers of Au Nanorods. Adv. Mater. 2008, 20, 4300−4305. (25) Lin, S.; Li, M.; Dujardin, E.; Girard, C.; Mann, S. OneDimensional Plasmon Coupling by Facile Self-Assembly of Gold Nanoparticles into Branched Chain Networks. Adv. Mater. 2005, 17, 2553−2559. (26) Hussain, I.; Brust, M.; Barauskas, J.; Cooper, A. I. Controlled Step Growth of Molecularly Linked Gold Nanoparticles: From Metallic Monomers to Dimers to Polymeric Nanoparticle Chains. Langmuir 2009, 25, 1934−1939. (27) Coomber, D.; Bartczak, D.; Gerrard, S. R.; Tyas, S.; Kanaras, A. G.; Stulz, E. Programmed Assembly of Peptide-Functionalized Gold Nanoparticles on DNA Templates. Langmuir 2010, 26, 13760−13762. (28) Zhang, H.; Wang, D. Controlling the Growth of ChargedNanoparticle Chains through Interparticle Electrostatic Repulsion. Angew. Chem., Int. Ed. 2008, 47, 3984−3987. (29) Yang, M.; Chen, G.; Zhao, Y.; Silber, G.; Wang, Y.; Xing, S.; Han, Y.; Chen, H. Mechanistic Investigation into the Spontaneous

ASSOCIATED CONTENT

S Supporting Information *

SEM study showing the role of water as a cosolvent on the formation of droplets, DLS and UV−vis characterization of pristine colloids in water/ethanol mixtures, photography showing the destabilization of aqueous colloids upon multiple washing, UV−vis characterization of 35 and 55 nm particles and their colloidal behavior, SEM image showing 1D DLCA morphology for 55 nm particles, SEM and SPM images tracking a given set of clusters experiencing drying between successive coating, time-dependent rinsing experiments, SEM images showing micrometer-sized assemblies obtained upon multiple spin-coating, the rupture of large particle chains on HSQ ridges and SEM images of chains oriented at 90° and 45° under different degrees of confinement. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.langmuir.5b01135.

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. 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 The authors thank C. Ruiz-Vargas, T. Wagner, and K. Javor for fruitful discussions and D. Webb for SEM support. We gratefully acknowledge technical support from the BRNC cleanroom staff.

■

REFERENCES

(1) Klein, D. L.; Rotht, R.; Lim, A. K. L.; Alivisatosti, A. P.; McEuen, P. L. A Single-Electron Transistor Made from a Cadmium Selenide Nanocrystal. Nature 1997, 389, 699−701. (2) Liu, X.; He, X.; Jiu, T.; Yuan, M.; Xu, J.; Lv, J.; Liu, H.; Li, Y. Controlled Aggregation of Functionalized Gold Nanoparticles with a Novel Conjugated Oligomer. ChemPhysChem 2007, 8, 906−912. (3) Wang, X.; Li, G.; Chen, T.; Yang, M.; Zhang, Z.; Wu, T.; Chen, H. Polymer-Encapsulated Gold-Nanoparticle Dimers: Facile Preparation and Catalytical Application in Guided Growth of Dimeric ZnONanowires. Nano Lett. 2008, 8, 2643−2647. (4) Sidhaye, D. S.; Prasad, B. L. V. Linear Assembly of Hexadecanethiol Coated Gold Nanoparticles. Chem. Phys. Lett. 2008, 454, 345−349. (5) Chen, J.; Liao, W.-S.; Chen, X.; Yang, T.; Wark, S. E.; Son, D. H.; Batteas, J. D.; Cremer, P. S. Evaporation-Induced Assembly of Quantum Dots into Nanorings. ACS Nano 2009, 3, 173−180. (6) Xing, S.; Tan, L. H.; Yang, M.; Pan, M.; Lv, Y.; Tang, Q.; Yang, Y.; Chen, H. Highly Controlled Core/shell Structures: Tunable Conductive Polymer Shells on Gold Nanoparticles and Nanochains. J. Mater. Chem. 2009, 19, 3286. (7) Chen, G.; Wang, Y.; Yang, M.; Xu, J.; Goh, S. J.; Pan, M.; Chen, H. Measuring Ensemble-Averaged Surface-Enhanced Raman Scattering in the Hotspots of Colloidal Nanoparticle Dimers and Trimers. J. Am. Chem. Soc. 2010, 132, 3644−3645. (8) Ren, S.; Lim, S.-K.; Gradecak, S. Synthesis and Thermal Responsiveness of Self-Assembled Gold Nanoclusters. Chem. Commun. (Cambridge, U. K.) 2010, 46, 6246−6248. 7226

DOI: 10.1021/acs.langmuir.5b01135 Langmuir 2015, 31, 7220−7227

Article

Langmuir Linear Assembly of Gold Nanospheres. Phys. Chem. Chem. Phys. 2010, 12, 11850−11860. (30) Liao, J.; Zhang, Y.; Yu, W.; Xu, L.; Ge, C.; Liu, J.; Gu, N. Linear Aggregation of Gold Nanoparticles in Ethanol. Colloids Surf., A 2003, 223, 177−183. (31) Liao, J. H.; Chen, K. J.; Xu, L. N.; Ge, C. W.; Wang, J.; Huang, L.; Gu, N. Self-Assembly of Length-Tunable Gold Nanoparticle Chains in Organic Solvents. Appl. Phys. A: Mater. Sci. Process. 2003, 76, 541−543. (32) Hussain, I.; Wang, Z.; Cooper, A. I.; Brust, M. Formation of Spherical Nanostructures by the Controlled Aggregation of Gold Colloids. Langmuir 2006, 22, 2938−2941. (33) Wang, H.; Schaefer, K.; Moeller, M. In Situ Immobilization of Gold Nanoparticle Dimers in Silica Nanoshell by Microemulsion Coalescence. J. Phys. Chem. C 2008, 112, 3175−3178. (34) Shipway, A. N.; Lahav, M.; Gabai, R.; Willner, I. Investigations into the Electrostatically Induced Aggregation of Au Nanoparticles. Langmuir 2000, 16, 8789−8795. (35) Ghosh, S. K.; Pal, T. Interparticle Coupling Effect on the Surface Plasmon Resonance of Gold Nanoparticles: From Theory to Applications. Chem. Rev. 2007, 107, 4797−4862. (36) Stannard, A. Dewetting-Mediated Pattern Formation in Nanoparticle Assemblies. J. Phys.: Condens. Matter 2011, 23, 083001. (37) Deegan, R. D.; Bakajin, O.; Dupont, T. F.; Huber, G.; Nagel, S. R.; Witten, T. A. Capillary Flow as the Cause of Ring Stains from Dried Liquid Drops. Nature 1997, 389, 827−829. (38) Maillard, M.; Motte, L.; Ngo, A. T.; Pileni, M. P. Rings and Hexagons Made of Nanocrystals: A Marangoni Effect. J. Phys. Chem. B 2000, 104, 11871−11877. (39) Xu, J.; Xia, J.; Lin, Z. Evaporation-Induced Self-Assembly of Nanoparticles from a Sphere-on-Flat Geometry. Angew. Chem., Int. Ed. Engl. 2007, 46, 1860−1863. (40) Su, B.; Zhang, C.; Chen, S.; Zhang, X.; Chen, L.; Wu, Y.; Nie, Y.; Kan, X.; Song, Y.; Jiang, L. A General Strategy for Assembling Nanoparticles in One Dimension. Adv. Mater. 2014, 26, 2501−2507. (41) Suematsu, N. J.; Ogawa, Y.; Yamamoto, Y.; Yamaguchi, T. Dewetting Self-Assembly of Nanoparticles into Hexagonal Array of Nanorings. J. Colloid Interface Sci. 2007, 310, 648−652. (42) Kraus, T.; Malaquin, L.; Schmid, H.; Riess, W.; Spencer, N. D.; Wolf, H. Nanoparticle Printing with Single-Particle Resolution. Nat. Nanotechnol 2007, 2, 570−576. (43) Asbahi, M.; Mehraeen, S.; Lim, K. T. P.; Wang, F.; Cao, J.; Tan, M. C.; Yang, J. K. W. Template-Induced Structure Transition in Sub10 nm Self-Assembling Nanoparticles. Nano Lett. 2014, 14, 2642− 2646. (44) Jiang, L.; Wang, W.; Fuchs, H.; Chi, L. One-Dimensional Arrangement of Gold Nanoparticles with Tunable Interparticle Distance. Small 2009, 5, 2819−2822. (45) Slot, J. W.; Geuze, H. J. A New Method of Preparing Gold Probes for Multiple-Labeling Cytochemistry. Eur. J. Cell Biol. 1985, 38, 87−93. (46) Maillard, M.; Motte, L.; Pileni, M.-P. Rings and Hexagons Made of Nanocrystals. Adv. Mater. 2001, 13, 200−204. (47) Khanal, B. P.; Zubarev, E. R. Rings of Nanorods. Angew. Chem., Int. Ed. Engl. 2007, 46, 2195−2198. (48) Lin, M. Y.; Lindsay, H. M.; Weitz, D. A.; Ball, R. C.; Klein, R.; Meakin, P. Universality in Colloid Aggregation. Nature 1989, 339, 360−362. (49) Khlebtsov, B. N.; Khlebtsov, N. G. On the Measurement of Gold Nanoparticle Sizes by the Dynamic Light Scattering Method. Colloid J. 2011, 73, 118−127. (50) Srikant, S.; Muralidharan, S. S.; Natarajan, U. Behaviour of Hydrogen Bonding and Structure of Poly(acrylic Acid) in Water− Ethanol Solution Investigated by Explicit Ion Molecular Dynamics Simulations. Mol. Simul. 2013, 39, 145−153. (51) Dörmann, M.; Schmid, H.-J. Simulation of Capillary Bridges between Nanoscale Particles. Langmuir 2014, 30, 1055−1062.

(52) Smith, T. G.; Lange, G. D.; Marks, W. B. Fractal Methods and Results in Cellular MorphologyDimensions, Lacunarity and Multifractals. J. Neurosci. Methods 1996, 69, 123−136. (53) Tisserant, J.-N.; Wicht, G.; Göbel, O. F.; Bocek, E.; Bona, G.-L.; Geiger, T.; Hany, R.; Mezzenga, R.; Partel, S.; Schmid, P.; et al. Growth and Alignment of Thin Film Organic Single Crystals from Dewetting Patterns. ACS Nano 2013, 7, 5506−5513. (54) Kang, Y.; Fukuoka, T.; Takahashi, R.; Utsumi, Y.; Haruyama, Y.; Matsui, S. Fabrication of Gold Nanoparticle Pattern Using Imprinted Hydrogen Silsesquioxane Pattern for Surface-Enhanced Raman Scattering. Microsyst. Technol. 2013, 20, 1993−2000.

7227

DOI: 10.1021/acs.langmuir.5b01135 Langmuir 2015, 31, 7220−7227