Interactions and Attachment Pathways between Functionalized Gold

Jan 24, 2017 - Nanoparticle (NP) self-assembly has been recognized as an important technological process for forming ordered nanostructures. However, ...
0 downloads 9 Views 6MB Size
Interactions and Attachment Pathways between Functionalized Gold Nanorods Shu Fen Tan,†,‡,§ Utkarsh Anand,†,‡,§,∥ and Utkur Mirsaidov*,†,‡,§,∥ †

Department of Physics, National University of Singapore, 117551 Singapore Centre for BioImaging Sciences, Department of Biological Sciences, National University of Singapore, 117557 Singapore § Centre for Advanced 2D Materials and Graphene Research Centre, National University of Singapore, 117546 Singapore ∥ NUSNNI-NanoCore, National University of Singapore, 117411 Singapore ‡

S Supporting Information *

ABSTRACT: Nanoparticle (NP) self-assembly has been recognized as an important technological process for forming ordered nanostructures. However, the detailed dynamics of the assembly processes remain poorly understood. Using in situ liquid cell transmission electron microscopy, we describe the assembly modes of gold (Au) nanorods (NRs) in solution mediated by hydrogen bonding between NR-bound cysteamine linker molecules. Our observations reveal that by tuning the linker concentration, two different NR assembly modes can be achieved. These assembly modes proceed via the (1) end-to-end and (2) side-toside attachment of NRs at low and high linker concentrations in solution, respectively. In addition, our time-resolved observations reveal that the side-to-side NR assemblies can occur through two different pathways: (i) prealigned attachment, where two Au NRs prealign to be parallel prior to assembly, and (ii) postattachment alignment, where two Au NRs first undergo end-to-end attachment and pivot around the attachment point to form the side-to-side assembly. We attributed the observed assembly modes to the distribution of linkers on the NR surfaces and the electrostatic interactions between the NRs. The intermediate steps in the assembly reported here reveal how the shape and surface functionalities of NPs drive their self-assembly, which is important for the rational design of hierarchical nanostructures. KEYWORDS: self-assembly, nanorods, in situ TEM, intermolecular forces, linker-mediated assembly and hydrophobic interactions.19,20 NP assemblies guided by external forces such as adsorption to liquid interfaces,21 electric and magnetic forces,22,23 capillary forces,24 and others are also known to produce very robust nanostructures. Our current understanding of NP self-assembly is based on imaging of the assembled structures using the “quench-andlook” approach,15,25 indirect dynamic spectroscopic methods,26,27 and modeling.8,28 The intermediate steps and the dynamics of self-assembly processes remain elusive because of experimental challenges associated with the direct visualization of nanoscale processes in solution. For example, the role of the intermolecular forces that act between the NPs and NP-bound surfactants during the self-assembly process is still not fully understood.29 Resolving these questions is important for the rational design of functional nanostructure assemblies for use in catalysis,30 optoelectronics,2 and biomedical applications.31 Single-crystal gold (Au) nanorods (NRs) are an important class of NPs that offer the flexibility to form one-, two-, and

T

op-down and bottom-up fabrication methods are two general approaches to engineering metallic nanostructures. Top-down techniques such as patterning through electron beam lithography are used to produce large arrays of metallic nanostructures of desired shapes with high yield and reproducibility.1 However, these nanostructures are usually polycrystalline, and their feature sizes are limited by the resolution of the lithography technique (∼10 nm). Bottom-up fabrication methods such as self-assembly, on the other hand, provide a robust method for the formation of nanostructures in a controllable manner from individual single crystalline nanoparticles (NPs). These NP assemblies exhibit distinct optical,2 mechanical,3 magnetic,4 and electrical5 properties that are drastically different from those of their isolated forms.6 Moreover, self-assembly is a relatively inexpensive and easily tunable process that can generate large-scale nanostructures.7 The self-assembly of NPs is driven by interparticle interactions and/or external forces.8,9 Interparticle interactions often utilize molecules to guide the NP organization by taking the advantage of intermolecular forces such as covalent bonding,10−13 van der Waals (vdW) forces, electrostatic interactions,14,15 π−π interactions,16 hydrogen bonding,17,18 © 2017 American Chemical Society

Received: November 3, 2016 Accepted: January 24, 2017 Published: January 24, 2017 1633

DOI: 10.1021/acsnano.6b07398 ACS Nano 2017, 11, 1633−1640

Article

www.acsnano.org

Article

ACS Nano three-dimensional architectures of different shapes because of their geometry and anisotropic affinity for surfactants.13,32,33 Cetyltrimethylammonium bromide (CTAB) is a commonly used capping agent in the synthesis of NRs, and CTAB-capped NRs are relatively easy to functionalize using different molecules to enable assembly through various interactions.25 For example, electrostatic interactions,15,34,35 covalent bonding,36,37 solvent-mediated interactions20,38 and host−guest molecular interactions39,40 can guide the assembly of chemically synthesized Au NRs. In addition, CTA+ ions have strong affinity for the {100} and {110} side facets of Au NRs compared to the {111} facets at the two ends of Au NRs.41 Therefore, the surfactant packing density on the sides of NRs is much higher than at its ends.36,42 The assembly of NRs can be induced by introducing linker molecules with an affinity for Au into the system.20,43,44 These linker molecules preferentially absorb onto the ends of the NRs because the ends are sparsely covered by surfactants. Such functionalization of NR ends with linker molecules connects the ends to form end-to-end assemblies.15,36,38,45 End-to-end and side-to-side NR assemblies can be induced by tuning the concentration of the linker molecules.42 However, the exact mechanism through which linkers mediate these end-to-end and side-to-side NR assemblies is unclear. In situ liquid cell (LC) transmission electron microscopy (TEM) serves as an ideal platform to study the dynamics of nanoscale processes in solution.46,47 This technique has been used to reveal the nucleation and growth mechanisms of NPs46,48−51 as well as self-assembly52−60 with nanometer resolution. For example, Sutter et al. demonstrated the use of LC TEM to study the self-assembly of octapod-shaped nanocrystals via vdW forces.55 Chen et al. used in situ TEM imaging to extract the interaction potentials of Au NRs from electron-beam-induced tip-to-tip assemblies.53 Welch et al. combined their study with MD and proposed that the preferential end-to-end attachment of silver (Ag) NRs is a result of weaker solvation forces occurring at the rod ends.54

Figure 1. Two assembly modes for gold (Au) nanorods (NRs) at two different concentrations of linker molecules (cysteamine). (A) Sulfur atoms of the linker molecules (cysteamine) bind to the Au surfaces and connect two NRs through hydrogen bonding. The expected length between two Au surfaces is ∼1.1 nm. Schematics and TEM images showing (B) end-to-end and (C) side-to-side NR assemblies in the presence of 150 and 500 μM cysteamine, respectively.

TEM images, see the Supporting Information, section 1). Note that we washed the Au NR solution before addition of cysteamine to remove unbound CTAB from the solution because the excess CTA+ ions will readily hinder the cysteamine molecules from displacing the CTA+ ions on the Au NRs (Supporting Information, section 6). Time-dependent UV−vis measurements (Supporting Information, section 5) show the appearance of an absorption band that signifies the presence of the end-to-end attachment mode of Au NRs at 150 μM, whereas a blue shift in the longitudinal plasmon peak indicates the existence of the side-to-side attachment mode of Au NRs at 500 μM. Our XPS measurements (Supporting Information, section 3) show that when cysteamine is added to a CTAB-coated Au surface, the sulfur signal associated with the Au-bonded thiol groups appears in the spectra. In addition, we observed a shift in the nitrogen peak that signifies a change in the oxidation state of nitrogen consistent with the replacement of CTA+ ions by cysteamine on the Au surface. To validate that the attachment of NRs is due to hydrogen bonding between linkers, we measured the interparticle distance between neighboring NRs for both assembly modes (Supporting Information, section 4). The sum of two linker lengths based on a ball-and-stick model (∼0.9 nm) and the hydrogen-bond length (∼0.2 nm)65 is ∼1.1 nm, which matches the average separation between neighboring Au NRs (Figure S4). The overall distribution of interparticle distances measured by ex situ and in situ TEM (Figure S4) did not change substantially, suggesting that the NR assembly is induced by the linker molecules and not by drying. In addition, when a longer linker molecule, 6-amino-1-hexanethiol hydrochloride, which has four extra −CH2− groups (each group is 1.3 Å long),66 was used, the average interparticle distance increased to 1.9 ± 0.1

RESULTS AND DISCUSSION Here, using in situ LC TEM imaging, we followed the linkermediated self-assembly dynamics of Au NRs with cysteamine as the linker molecule. Cysteamine linker has two functional groups, amine and thiol moiety, and both can form bonds with Au. Bonding between the thiol moiety and Au is more prominent because of the strong Au−S covalent bond formation (∼209 kJ/mol or 84.5 kBT),61,62 which has been widely used in NR functionalization.15,36,42 Au-bonded cysteamine linkers can connect two adjacent Au NRs via hydrogen bonding (N−H···N), as illustrated in Figure 1A. We found that depending on the linker concentration in solution, NRs assemble through two different attachment modes: (1) endto-end attachment (Figure 1B) at low (150 μM) and (2) sideto-side (Figure 1C) attachment at high (500 μM) cysteamine concentrations. At low linker concentrations, the CTA+ ions that have less affinity for the {111} facets at the NR ends63,64 are easily displaced by thiol.42 This results in NRs with cysteamine-coated ends. The formation of hydrogen bonds between the cysteamine molecules of two NRs then leads to their end-to-end attachment. At high cysteamine concentrations, the excess thiol groups displace the CTA+ ions also on the sides ({110} and {100} facets) of the NRs. Hence, the formation of hydrogen bonds between the cysteamine linkers of two NRs now promotes side-to-side NR attachment (for more 1634

DOI: 10.1021/acsnano.6b07398 ACS Nano 2017, 11, 1633−1640

Article

ACS Nano

Figure 2. End-to-end attachment of NRs in solution. (A) TEM time series images showing the end-to-end attachment of Au NRs in the presence of 150 μM cysteamine (supporting video 1). (B) Schematics showing the orientation angle (β) prior to NR attachment, which is measured from the dashed lines connecting the ends of two NRs. The end-to-end distance (de−e) between NRs was defined as the distance between the ends of NRs that eventually connect. (C) Histogram of the orientation angles between NRs for 14 attachment events. Each bar color corresponds to an individual attachment event (Figure S11). (D) End-to-end distances between three different NR pairs (de−e) as a function of time.

Figure 3. Formation of Au NR chains in solution. TEM time series images showing the formation of Au NR chains in the presence of 150 μM cysteamine. (A) NR chains form through the sequential attachment of single NRs. (B) NR chains form through the attachment of two chain fragments, each consisting of two NRs (supporting video 2).

nm, consistent with the expected spacing of ∼2.0 nm (Supporting Information, section 9). To further confirm that the attachment of NRs is only due to the cysteamine linkers, we performed similar ex situ and in situ experiments in the absence of cysteamine linkers (Supporting Information, section 2, Figure S2). In both the ex situ and in situ experiments, the Au NRs were randomly dispersed and did not form any assemblies. These experiments validate that the assembly dynamics of Au NRs is due to the linker molecules and is not an artifact of the electron beam used for in situ imaging. To understand the assembly process of Au NRs, we investigated their attachment dynamics at different cysteamine concentrations using in situ LC TEM imaging. An aqueous solution of Au NRs with cysteamine was loaded into an LC and imaged by TEM. All of the experiments were performed at room temperature (T = 20 °C). For all image sequences, the time point at which the initial movements of the Au NRs were observed was defined as t = 0 s. The TEM time series images in

Figure 2A show the end-to-end attachment of two Au NRs in the presence of 150 μM cysteamine (supporting video 1). Two NRs approached each other in the solution (t = 2.56−3.68 s) and underwent end-to-end attachment (t = 4.76 s). We further characterized the assembly of NRs by analyzing the orientation angle (β) of each NR pair before attachment. We defined β as the angle between the major axes of two interacting NRs as shown in Figure 2B. The histogram shown in Figure 2C, which plots the distribution of angles between projected images of 14 interacting NR pairs in each frame until they attached and formed a connected pair, reveals that their orientation was preferentially aligned for linear (end-to-end) attachment (i.e., β > 90°). In Figure 2C, we omitted six events in which the β remained unchanged for more than 10 s (i.e., NRs were stuck to the surface for prolonged periods of time) to obtain reliable orientation angles for freely moving NRs. Importantly, even when all of the NR pairs were included, the overall distribution of the orientation angles did not change substantially (Supporting Information, section 10). The end-to-end distance 1635

DOI: 10.1021/acsnano.6b07398 ACS Nano 2017, 11, 1633−1640

Article

ACS Nano

Figure 4. Side-to-side assembly of NRs in solution. Schematics and TEM time series images show the side-to-side assemblies of Au NRs in the presence of 500 μM cysteamine linkers. Side-to-side assembly occurs through two different pathways: (A) prealigned side-to-side attachment and (B) postattachment alignment of an end-to-end assembly into a side-to-side assembly (see Figure S7 for more side-to-side assembly events).

Figure 5. Side-to-side assembly through the postattachment alignment pathway. TEM time series images showing (A) the side-to-side assemblies of Au NRs in the presence of 500 μM cysteamine (supporting video 3). (B) End-to-end distance between two NRs (de−e) as a function of time. (C) Orientation angles (β) between NRs as a function of time after the end-to-end attachment of NRs. The green and red curves correspond to the events shown in Figures 4B and 5A, respectively.

(de−e), defined as the distance between the ends of two NRs as shown in Figure 2B, reveals that two NRs approach each other at different speeds (Figure 2D) as they align for end-to-end attachment (Figure 2A). This suggests that long-range repulsive electrostatic interactions between NRs, which aid in their alignment for end-to-end configuration, are not strong enough to prevent their attachment. After attachment, we rarely observed NRs detaching from the assembly, indicating that the linker-mediated hydrogen bonding between NRs is strong enough to stabilize the assemblies (Supporting Information, section 11).

Our observations of NR assembly at a low linker concentration (150 μM cysteamine) reveal that NR chains form either through the attachment of individual NRs (Figure 3A) or the attachment of smaller NR chain fragments (Figure 3B, supporting video 2). Note that these NR chains were not strongly adsorbed onto the surface and underwent both rotational and translational movements (t = 1.9−37.2 s in Figure 3A and t = 1.4−2.4 s in Figure 3B). We also observed the formation of longer NR chains via a combination of the two pathways described above (Supporting Information, section 8). Next, we examined how the assembly dynamics change when the linker concentration is increased. At a linker concentration 1636

DOI: 10.1021/acsnano.6b07398 ACS Nano 2017, 11, 1633−1640

Article

ACS Nano of 500 μM, the Au NRs form side-to-side assemblies through two different pathways: (i) prealigned attachment (Figure 4A) or (ii) postattachment alignment (Figure 4B). In the prealigned attachment pathway (Figure 4A), two Au NRs first align parallel to each other and then undergo side-to-side attachment to form an assembly. In the postattachment alignment pathway shown in Figure 4B, two randomly aligned NRs first undergo end-to-end attachment (Figure 4B: t = 0.0−58.9 s), similar to that observed at a low linker concentration (Figure 2). These NRs then rotate and zip to form the side-to-side assembly (Figure 4B: t = 67.3−93.8 s). When imaging assemblies in the presence of a high linker concentration, we often found residues displayed as gray spots (Figures 4B and 5A). These residues could originate from the CTAB surfactant that is displaced from the surface of the Au NRs by the cysteamine linkers. Similar observations were made when we examined the effect of an increased CTAB concentration on NR assembly (Supporting Information, section 6). A detailed examination of the postattachment alignment pathway (supporting video 3) revealed that once the NR ends make contact (Figure 5A,B), the angle (β) between them starts to steadily change until they make full sideways contact (Figure 5A,C) and β ≈ 0°. The force that drives this rotational alignment of NRs from end-to-end assembly toward side-toside assembly can be attributed to the development of a hydrogen-bonding network between the cysteamine molecules in NR pairs. The estimated binding strengths between NRs due to hydrogen bonding can be as high as ∼2 × 10−17 J (5000 kBT) and ∼9 × 10−17 J (22000 kBT) for the end-to-end and side-to-side configurations, respectively. This estimate is based on the assumption that a pair of nearby cysteamine molecules, one in each NR, contributes to one hydrogen bond (N−H···N) (Supporting Information, section 11). Since the net strength of the hydrogen bonding linking the NRs is stronger (i.e., more hydrogen bonds) for the side-to-side configuration than for end-to-end configuration, we expect the end-to-end assemblies to transition to side-to-side assemblies for NRs that are densely capped by linkers (i.e., at a higher linker concentration). The hydrogen bonding between the linkers that connect NRs can occur only when the NRs are separated by two linker lengths; beyond these distances, vdW and electrostatic forces regulate the pairwise interaction between the NRs. Therefore, the effect of the charge screening by divalent cysteamine linkers in solution, which reduces the electrostatic repulsion between NRs, also needs to be considered. At a low cysteamine concentration, the electrostatic repulsion in the end-to-end configuration of Au NRs is substantially lower (Supporting Information, section 12). Therefore, at the low linker concentration, we did not observe any pivoting of end-to-end NR assemblies into side-to-side assemblies, even after extended imaging (Supporting Information, section 16). When the concentration of linkers is high, the excess cysteamine molecules provide the divalent counterions67 that screen the NR charge and hence reduce the Coulombic repulsion56 (Supporting Information, section 12). Consequently, at a high linker concentration, NRs can approach and contact each other via both their ends and sides, which is why two different pathways exist for the side-to-side assembly mode.

NR surfaces. Here, the attachment of NRs is driven by the hydrogen bonding between NR-bound linkers. The end-to-end assembly mode occurs at low linker concentrations when only the ends of NRs are capped with linkers. At high linker concentrations, the NRs are fully capped with linkers and form side-to-side assemblies. The side-to-side assembly mode proceeds through two different pathways: (i) attachment of NRs aligned sideways or (ii) end-to-end attachment of NRs, followed by their rotation into a side-to-side assembly. Our study highlights the importance of direct dynamic observations of individual interactions between NPs in resolving assembly pathways, which is important in the design of hierarchical nanostructures from nanoscale building blocks.

MATERIALS AND METHODS Sample Preparation. We used a solution of CTAB-stabilized Au NRs (Cat. No. NR-10-750-50, NanoSeedz Ltd., Shatin, N. T., Hong Kong), which we measured to be 50−60 nm in length and 10−15 nm in diameter. Cysteamine (Cat. No. M9768-5G, Sigma-Aldrich, St Louis, MO) and 6-amino-1-hexanethiol hydrochloride (Cat. No. 733679-500MG, Sigma-Aldrich, St Louis, MO) served as the linkers and were used as received. First, we transferred 200 μL of a solution of Au NRs from a stock solution into a 1.5 mL centrifuge tube. We centrifuged the NR solution at 10000 rpm for 5 min and resuspended it in 200 μL of deionized water (18.2 MΩ·cm) to reduce the overall concentration of unbound CTAB in the NR suspension to