Home
Search
Collections
Journals
About
Contact us
My IOPscience
Deterministic nanoparticle assemblies: from substrate to solution
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2014 Nanotechnology 25 155302 (http://iopscience.iop.org/0957-4484/25/15/155302) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 134.117.10.200 This content was downloaded on 26/06/2014 at 11:05
Please note that terms and conditions apply.
Nanotechnology Nanotechnology 25 (2014) 155302 (6pp)
doi:10.1088/0957-4484/25/15/155302
Deterministic nanoparticle assemblies: from substrate to solution Steven J Barcelo1 , Ansoon Kim2 , Gary A Gibson1 , Kate J Norris3,4 , Mineo Yamakawa1 and Zhiyong Li1 1 2 3 4
Hewlett–Packard Laboratories, Palo Alto, CA 94043, USA Korea Research Institute of Standards and Science, Jung-gu, Daejeon, Korea University of California Santa Cruz, Baskin School of Engineering, Santa Cruz, CA 95064, USA NASA Ames Research Center, Moffett Field, CA 94035, USA
E-mail: [email protected] Received 29 October 2013, revised 7 February 2014 Accepted for publication 24 February 2014 Published 18 March 2014
Abstract
The deterministic assembly of metallic nanoparticles is an exciting field with many potential benefits. Many promising techniques have been developed, but challenges remain, particularly for the assembly of larger nanoparticles which often have more interesting plasmonic properties. Here we present a scalable process combining the strengths of top down and bottom up fabrication to generate deterministic 2D assemblies of metallic nanoparticles and demonstrate their stable transfer to solution. Scanning electron and high-resolution transmission electron microscopy studies of these assemblies suggested the formation of nanobridges between touching nanoparticles that hold them together so as to maintain the integrity of the assembly throughout the transfer process. The application of these nanoparticle assemblies as solution-based surface-enhanced Raman scattering (SERS) materials is demonstrated by trapping analyte molecules in the nanoparticle gaps during assembly, yielding uniformly high enhancement factors at all stages of the fabrication process. Keywords: nanoparticle, self-assembly, plasmonics, SERS, chemical sensing S Online supplementary data available from stacks.iop.org/Nano/25/155302/mmedia (Some figures may appear in colour only in the online journal)
1. Introduction
Colloidal metallic nanoparticles and assemblies are an important subset of this field due to their many potential applications, including novel sensors [6–8], customizable optical or magnetic nanomaterials [9, 10] and medical diagnostics and therapeutics [11, 12]. Colloidal SERS-active nanoparticles are particularly interesting because they open up the fields of Raman-based in vivo sensing and imaging [13–15], which compare favorably to fluorescence tagging approaches [16] due to their potential for high brightness and narrow signal leading to multiplexing [17]. However, standard approaches relying on random agglomeration of colloidal nanoparticles have a low yield of active hot spots [18]. Postprocessing techniques such as centrifugation can enhance the concentration of more desirable agglomeration states [19, 20], but direct control over the particle architecture cannot be achieved with
The study of nanostructured metals is a vast and rapidly growing field, largely due to the compelling features of localized surface-plasmon resonances which are readily manipulated by tuning the size, shape and positioning of nanoscale features. For example, the interaction of plasmon resonances in adjacent nanostructures with nanometer-scale gaps or crevices can generate a large electric field, E. The field of surface-enhanced Raman scattering (SERS) has grown around this fact, since the Raman scattering cross section is enhanced by a factor of E 4 [1, 2], allowing the detection of single molecules [3, 4]. Recent advances in modeling such as plasmon hybridization theory [5] have enabled the description of increasingly complex systems, but realizing them experimentally remains challenging. 0957-4484/14/155302+06$33.00
1
c 2014 IOP Publishing Ltd
Printed in the UK
Nanotechnology 25 (2014) 155302
S J Barcelo et al
this approach. Deterministic assembly is of particular interest due to the potential to provide uniform hot spots in an optimal configuration for consistent results. The unique and well-known properties of DNA have been used to generate a wide variety of assemblies, chains and arrays [21]. However, the monofunctionalization of large nanoparticles, as required for many plasmonic applications, remains challenging, limiting the high yield production of consistent geometric assemblies with the small gaps necessary for intense plasmonic coupling. Alternative binding molecules such as multivalent thiols [22] and Cucurbit[n]urils [23] still suffer from low yield. Template-based assembly methods can generate well-defined nanoparticle assemblies [24–26], but they are typically confined to a substrate, limiting their application. Alternatively, flexible polymer structures can be designed using top down lithography and assembly into structures with nanometer-scale features [27] or unique structures such as curved nanorods [28]. In this paper we discuss a method to capitalize on this unique combination of top down and bottom up approaches to generate deterministic colloidal nanoparticle assemblies.
to poor adhesion and incomplete transfer, while a thick film can cause complete delamination of the nanofinger layer rather than the desired selective transfer. SEM images of a variety of nanoparticle assemblies on PMMA fabricated using this approach shown in figure 2 demonstrate its versatility, while the large array of transferred pentamer assemblies with few defects demonstrates its robustness. The PMMA film is then dissolved, leaving the nanoparticle assemblies dispersed in the solvent. 2.2. Characterization
To investigate the application of this process to the generation of SERS imaging solutions, a standard Raman marker, trans1,2-bis(4-pyridyl)ethylene (BPE), was included in the solution used to collapse the nanofingers. The BPE molecules are trapped in the junction between neighboring particles [31], known as ‘hot spots’ due to the intense electric field which can strongly enhance Raman scattering and fluorescence. Raman measurements for the liquid samples were performed on a DeltaNu Inspector Raman portable spectrometer with a liquid sample cell attachment using a 120 mW 785 nm light source. Raman spectra were recorded for solutions with and without nanoparticle assemblies, and the difference was reported to remove the effects of the background solution. Nanoparticle assembly solutions were generated by dissolving a 3 cm2 area containing approximately 4.5 × 108 transferred pentamer assemblies in 1.5 ml acetone, yielding 0.7 pM concentration. Reference colloidal solutions were received as 0.8 nM in water buffered with 10 mM 2-(N -morpholino)ethanesulfonic acid and were diluted to the reported concentrations in the buffer solution. All colloidal solutions were shaken by hand before measurement to ensure uniform dispersion. Further characterization by UV–vis spectroscopy was performed using a Cary 6000i UV–vis–NIR spectrophotometer. Dynamic light scattering measurements were performed using a Malvern Zetasizer Nano ZS. Scanning electron microscopy was performed on an FEI Sirion XL30 SFEG microscope with a 10 kV acceleration voltage using a through lens detector. Lacy carbon coated copper grids (300 mesh) were prepared for transmission electron microscopy by dipping in colloidal nanoparticle solution and blow drying. Measurements were taken in a Hitachi H-9500 with a 300 kV acceleration voltage.
2. Experimental methods 2.1. Fabrication process
Previously, we reported the fabrication of deterministic nanoparticle assemblies on flexible polymer ‘nanofingers’ by nanoimprint lithography [29, 30]. The nanofingers can be pulled together into designed geometries via exposure to a volatile liquid, in some cases trapping molecules in the nanometer-scale junction between metal particles [31]. Once formed, these nanoparticle assemblies can be transferred to a new substrate using chemical or metal–metal bonds [32]. Here we introduce a crucial extension to this technique, enabling the transfer of deterministic nanoparticle assemblies to solution with unprecedented control over a variety of parameters, including particle size, shape, material, number and geometry. The process for generating colloidal nanoparticle assembly is outlined in figure 1. First, groups of gold coated nanofingers with well-defined symmetry are fabricated on a silicon substrate. Upon exposure to a volatile liquid, the free ends of the nanofingers are pulled together by capillary forces as the liquid evaporates. As the nanofinger tips come in contact, deterministic nanoparticle assemblies are formed. Pentamer groupings were previously shown to have favorable SERS properties as compared to other geometries and will be the focus of this paper [30]. Next, 3% 950 K PMMA in anisole is spun on a new silicon substrate, generating a 200 nm thick film which is then pressed against the nanofinger template using a purpose-built imprinting tool under 150 psi and 100 ◦ C for 5 min. This heating causes reflow of the PMMA, embedding the tips of the nanoparticle assemblies in the PMMA film. The substrates are allowed to cool and then separated, trapping the nanoparticle assemblies on the PMMA surface. The film thickness is designed to enable complete immersion of the assemblies in PMMA without contacting the gold film underneath the nanofingers. Too little PMMA leads
3. Results and discussion 3.1. Steric hindrance due to PMMA film
A thin coating film can be observed in SEM images of colloidal nanoparticle assemblies dispensed on a substrate (figure S1 available at stacks.iop.org/Nano/25/155302/mme dia). Expansion of the film is observed after heating to 110 ◦ C for 10 min, just above the glass temperature of PMMA, indicating that this film likely consists of PMMA transferred to the nanoparticle assembly surface during the transfer process shown in figure 1(c). The same film thickness remains even after dilution of the initial solution by a factor of 1000, indicating that the PMMA is not coating the assembly surface 2
Nanotechnology 25 (2014) 155302
S J Barcelo et al
(a)
(b)
(c)
(d)
(e)
Figure 1. Process flow for deterministic formation of nanoparticle assemblies in solution: (a) nanofingers are fabricated via nanoimprint lithography; (b) exposure to a volatile liquid collapses the nanofingers forming uniform nanoparticle assemblies; (c) the nanofingers are imprinted onto a thin polymer layer on a rigid substrate; (d) nanoparticle assemblies are trapped in the polymer layer after the templates are separated. (e) The nanoparticle assemblies are transferred to solution by dissolving the polymer.
Figure 2. (a) Schemes of a variety of particle assembly designs, (b) SEM images, all with a scale bar of 100 nm, showing Au assemblies transferred to PMMA and (c) demonstration of large area sample transfer with few defects over 10 s of µm2 , with the assembly structure shown in the inset.
during precipitation from solution but rather remains bound to the assembly in solution. The PMMA film provides steric hindrance that limits agglomeration as indicated by the stable Raman signal and nanoparticle assembly size over a period of weeks (figure S2 available at stacks.iop.org/Nano/25/155302/ mmedia). This is similar to recent results showing that adding PMMA to a solvent during nanoparticle generation leads to significantly higher colloidal stability for both solution-based synthesis [33, 34] or physical process such as pulsed laser ablation in liquid [35, 36].
require high pressure or clean surfaces under UHV conditions, but later demonstrated with low pressure and dirty surfaces on elastomeric supports [37]. Nanoscale cold welding has recently been demonstrated for applied stress
Search
Collections
Journals
About
Contact us
My IOPscience
Deterministic nanoparticle assemblies: from substrate to solution
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2014 Nanotechnology 25 155302 (http://iopscience.iop.org/0957-4484/25/15/155302) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 134.117.10.200 This content was downloaded on 26/06/2014 at 11:05
Please note that terms and conditions apply.
Nanotechnology Nanotechnology 25 (2014) 155302 (6pp)
doi:10.1088/0957-4484/25/15/155302
Deterministic nanoparticle assemblies: from substrate to solution Steven J Barcelo1 , Ansoon Kim2 , Gary A Gibson1 , Kate J Norris3,4 , Mineo Yamakawa1 and Zhiyong Li1 1 2 3 4
Hewlett–Packard Laboratories, Palo Alto, CA 94043, USA Korea Research Institute of Standards and Science, Jung-gu, Daejeon, Korea University of California Santa Cruz, Baskin School of Engineering, Santa Cruz, CA 95064, USA NASA Ames Research Center, Moffett Field, CA 94035, USA
E-mail: [email protected] Received 29 October 2013, revised 7 February 2014 Accepted for publication 24 February 2014 Published 18 March 2014
Abstract
The deterministic assembly of metallic nanoparticles is an exciting field with many potential benefits. Many promising techniques have been developed, but challenges remain, particularly for the assembly of larger nanoparticles which often have more interesting plasmonic properties. Here we present a scalable process combining the strengths of top down and bottom up fabrication to generate deterministic 2D assemblies of metallic nanoparticles and demonstrate their stable transfer to solution. Scanning electron and high-resolution transmission electron microscopy studies of these assemblies suggested the formation of nanobridges between touching nanoparticles that hold them together so as to maintain the integrity of the assembly throughout the transfer process. The application of these nanoparticle assemblies as solution-based surface-enhanced Raman scattering (SERS) materials is demonstrated by trapping analyte molecules in the nanoparticle gaps during assembly, yielding uniformly high enhancement factors at all stages of the fabrication process. Keywords: nanoparticle, self-assembly, plasmonics, SERS, chemical sensing S Online supplementary data available from stacks.iop.org/Nano/25/155302/mmedia (Some figures may appear in colour only in the online journal)
1. Introduction
Colloidal metallic nanoparticles and assemblies are an important subset of this field due to their many potential applications, including novel sensors [6–8], customizable optical or magnetic nanomaterials [9, 10] and medical diagnostics and therapeutics [11, 12]. Colloidal SERS-active nanoparticles are particularly interesting because they open up the fields of Raman-based in vivo sensing and imaging [13–15], which compare favorably to fluorescence tagging approaches [16] due to their potential for high brightness and narrow signal leading to multiplexing [17]. However, standard approaches relying on random agglomeration of colloidal nanoparticles have a low yield of active hot spots [18]. Postprocessing techniques such as centrifugation can enhance the concentration of more desirable agglomeration states [19, 20], but direct control over the particle architecture cannot be achieved with
The study of nanostructured metals is a vast and rapidly growing field, largely due to the compelling features of localized surface-plasmon resonances which are readily manipulated by tuning the size, shape and positioning of nanoscale features. For example, the interaction of plasmon resonances in adjacent nanostructures with nanometer-scale gaps or crevices can generate a large electric field, E. The field of surface-enhanced Raman scattering (SERS) has grown around this fact, since the Raman scattering cross section is enhanced by a factor of E 4 [1, 2], allowing the detection of single molecules [3, 4]. Recent advances in modeling such as plasmon hybridization theory [5] have enabled the description of increasingly complex systems, but realizing them experimentally remains challenging. 0957-4484/14/155302+06$33.00
1
c 2014 IOP Publishing Ltd
Printed in the UK
Nanotechnology 25 (2014) 155302
S J Barcelo et al
this approach. Deterministic assembly is of particular interest due to the potential to provide uniform hot spots in an optimal configuration for consistent results. The unique and well-known properties of DNA have been used to generate a wide variety of assemblies, chains and arrays [21]. However, the monofunctionalization of large nanoparticles, as required for many plasmonic applications, remains challenging, limiting the high yield production of consistent geometric assemblies with the small gaps necessary for intense plasmonic coupling. Alternative binding molecules such as multivalent thiols [22] and Cucurbit[n]urils [23] still suffer from low yield. Template-based assembly methods can generate well-defined nanoparticle assemblies [24–26], but they are typically confined to a substrate, limiting their application. Alternatively, flexible polymer structures can be designed using top down lithography and assembly into structures with nanometer-scale features [27] or unique structures such as curved nanorods [28]. In this paper we discuss a method to capitalize on this unique combination of top down and bottom up approaches to generate deterministic colloidal nanoparticle assemblies.
to poor adhesion and incomplete transfer, while a thick film can cause complete delamination of the nanofinger layer rather than the desired selective transfer. SEM images of a variety of nanoparticle assemblies on PMMA fabricated using this approach shown in figure 2 demonstrate its versatility, while the large array of transferred pentamer assemblies with few defects demonstrates its robustness. The PMMA film is then dissolved, leaving the nanoparticle assemblies dispersed in the solvent. 2.2. Characterization
To investigate the application of this process to the generation of SERS imaging solutions, a standard Raman marker, trans1,2-bis(4-pyridyl)ethylene (BPE), was included in the solution used to collapse the nanofingers. The BPE molecules are trapped in the junction between neighboring particles [31], known as ‘hot spots’ due to the intense electric field which can strongly enhance Raman scattering and fluorescence. Raman measurements for the liquid samples were performed on a DeltaNu Inspector Raman portable spectrometer with a liquid sample cell attachment using a 120 mW 785 nm light source. Raman spectra were recorded for solutions with and without nanoparticle assemblies, and the difference was reported to remove the effects of the background solution. Nanoparticle assembly solutions were generated by dissolving a 3 cm2 area containing approximately 4.5 × 108 transferred pentamer assemblies in 1.5 ml acetone, yielding 0.7 pM concentration. Reference colloidal solutions were received as 0.8 nM in water buffered with 10 mM 2-(N -morpholino)ethanesulfonic acid and were diluted to the reported concentrations in the buffer solution. All colloidal solutions were shaken by hand before measurement to ensure uniform dispersion. Further characterization by UV–vis spectroscopy was performed using a Cary 6000i UV–vis–NIR spectrophotometer. Dynamic light scattering measurements were performed using a Malvern Zetasizer Nano ZS. Scanning electron microscopy was performed on an FEI Sirion XL30 SFEG microscope with a 10 kV acceleration voltage using a through lens detector. Lacy carbon coated copper grids (300 mesh) were prepared for transmission electron microscopy by dipping in colloidal nanoparticle solution and blow drying. Measurements were taken in a Hitachi H-9500 with a 300 kV acceleration voltage.
2. Experimental methods 2.1. Fabrication process
Previously, we reported the fabrication of deterministic nanoparticle assemblies on flexible polymer ‘nanofingers’ by nanoimprint lithography [29, 30]. The nanofingers can be pulled together into designed geometries via exposure to a volatile liquid, in some cases trapping molecules in the nanometer-scale junction between metal particles [31]. Once formed, these nanoparticle assemblies can be transferred to a new substrate using chemical or metal–metal bonds [32]. Here we introduce a crucial extension to this technique, enabling the transfer of deterministic nanoparticle assemblies to solution with unprecedented control over a variety of parameters, including particle size, shape, material, number and geometry. The process for generating colloidal nanoparticle assembly is outlined in figure 1. First, groups of gold coated nanofingers with well-defined symmetry are fabricated on a silicon substrate. Upon exposure to a volatile liquid, the free ends of the nanofingers are pulled together by capillary forces as the liquid evaporates. As the nanofinger tips come in contact, deterministic nanoparticle assemblies are formed. Pentamer groupings were previously shown to have favorable SERS properties as compared to other geometries and will be the focus of this paper [30]. Next, 3% 950 K PMMA in anisole is spun on a new silicon substrate, generating a 200 nm thick film which is then pressed against the nanofinger template using a purpose-built imprinting tool under 150 psi and 100 ◦ C for 5 min. This heating causes reflow of the PMMA, embedding the tips of the nanoparticle assemblies in the PMMA film. The substrates are allowed to cool and then separated, trapping the nanoparticle assemblies on the PMMA surface. The film thickness is designed to enable complete immersion of the assemblies in PMMA without contacting the gold film underneath the nanofingers. Too little PMMA leads
3. Results and discussion 3.1. Steric hindrance due to PMMA film
A thin coating film can be observed in SEM images of colloidal nanoparticle assemblies dispensed on a substrate (figure S1 available at stacks.iop.org/Nano/25/155302/mme dia). Expansion of the film is observed after heating to 110 ◦ C for 10 min, just above the glass temperature of PMMA, indicating that this film likely consists of PMMA transferred to the nanoparticle assembly surface during the transfer process shown in figure 1(c). The same film thickness remains even after dilution of the initial solution by a factor of 1000, indicating that the PMMA is not coating the assembly surface 2
Nanotechnology 25 (2014) 155302
S J Barcelo et al
(a)
(b)
(c)
(d)
(e)
Figure 1. Process flow for deterministic formation of nanoparticle assemblies in solution: (a) nanofingers are fabricated via nanoimprint lithography; (b) exposure to a volatile liquid collapses the nanofingers forming uniform nanoparticle assemblies; (c) the nanofingers are imprinted onto a thin polymer layer on a rigid substrate; (d) nanoparticle assemblies are trapped in the polymer layer after the templates are separated. (e) The nanoparticle assemblies are transferred to solution by dissolving the polymer.
Figure 2. (a) Schemes of a variety of particle assembly designs, (b) SEM images, all with a scale bar of 100 nm, showing Au assemblies transferred to PMMA and (c) demonstration of large area sample transfer with few defects over 10 s of µm2 , with the assembly structure shown in the inset.
during precipitation from solution but rather remains bound to the assembly in solution. The PMMA film provides steric hindrance that limits agglomeration as indicated by the stable Raman signal and nanoparticle assembly size over a period of weeks (figure S2 available at stacks.iop.org/Nano/25/155302/ mmedia). This is similar to recent results showing that adding PMMA to a solvent during nanoparticle generation leads to significantly higher colloidal stability for both solution-based synthesis [33, 34] or physical process such as pulsed laser ablation in liquid [35, 36].
require high pressure or clean surfaces under UHV conditions, but later demonstrated with low pressure and dirty surfaces on elastomeric supports [37]. Nanoscale cold welding has recently been demonstrated for applied stress
