Confined Etching within 2D and 3D Colloidal Crystals for Tunable

Jan 17, 2017 - Due to a suppressed etching rate at the point of contact between the PS particles originating from their arrangement in 2D and 3D cryst...
4 downloads 15 Views 7MB Size
Research Article www.acsami.org

Confined Etching within 2D and 3D Colloidal Crystals for Tunable Nanostructured Templates: Local Environment Matters Fedja J. Wendisch, Richard Oberreiter, Miralem Salihovic, Michael S. Elsaesser, and Gilles R. Bourret* Department of Chemistry and Physics of Materials, University of Salzburg, Hellbrunner Straße 34/III, A-5020 Salzburg, Austria S Supporting Information *

ABSTRACT: We report the isotropic etching of 2D and 3D polystyrene (PS) nanosphere hcp arrays using a benchtop O2 radio frequency plasma cleaner. Unexpectedly, this slow isotropic etching allows tuning of both particle diameter and shape. Due to a suppressed etching rate at the point of contact between the PS particles originating from their arrangement in 2D and 3D crystals, the spherical PS templates are converted into polyhedral structures with well-defined hexagonal cross sections in directions parallel and normal to the crystal c-axis. Additionally, we found that particles located at the edge (surface) of the hcp 2D (3D) crystals showed increased etch rates compared to those of the particles within the crystals. This indicates that 2D and 3D order affect how nanostructures chemically interact with their surroundings. This work also shows that the morphology of nanostructures periodically arranged in 2D and 3D supercrystals can be modified via gas-phase etching and programmed by the superlattice symmetry. To show the potential applications of this approach, we demonstrate the lithographic transfer of the PS template hexagonal cross section into Si substrates to generate Si nanowires with well-defined hexagonal cross sections using a combination of nanosphere lithography and metal-assisted chemical etching. KEYWORDS: colloidal crystal, template, nanosphere lithography, metal-assisted chemical etching, Si nanowires

1. INTRODUCTION Nanostructure superlattices, characterized with 2D and 3D long-range order, have exceptional and tunable physicochemical properties.1−6 A strong modification of the materials’ optical properties is observed when the array periodicity becomes similar or smaller in size to the wavelength of the incident light.1,2,6−14 This can lead, for example, to characteristic photonic band gaps within 2D and 3D photonic crystals,1,2,6,12 and some unusual optical responses in the case of plasmonic crystals.7,11,13 Large-scale preparation of periodic architectures is not trivial and is difficult to achieve with standard lithography techniques. To prepare such materials, self-assembly methods have emerged as the most robust and simplest approaches.12,15−17 In this regard, the crystallization of 100− 1000 nm colloids within hexagonal close-packed (hcp) superlattices is particularly attractive because it is cheap, simple, and amenable to very large-scale syntheses.12,17−19 Due to the simplicity of the self-assembly process, colloidal crystals have been extensively used as versatile sacrificial templates to prepare 2D and 3D nanostructured arrays with well-defined porosities and periodicities.12,17 For example, large-scale plasmonic crystals can be prepared using polystyrene (PS) sphere arrays as templates,20 leading to periodic materials that are suitable for optical sensing,21−24 metallic gratings,25 plasmon-modulated fluorescence,26 and photocatalysis.27 While the periodicity of © XXXX American Chemical Society

such arrays mostly governs the optical far-field response, the electric near-field, which dictates the performance of most plasmonic applications (hot-electron injection, surface-enhanced Raman scattering, etc.), is dominated by the local geometry of the metal structure where, for example, the presence of sharp edges can lead to enhanced efficiencies and sensitivities.24,28,29 In this regard, the spherical shape of the PS templates clearly limits the possibilities offered by these colloidal crystals. Similarly, the use of PS sphere arrays as lithographic masks for preparing Si nanowire arrays via nanosphere lithography (NSL) and metal-assisted chemical etching (MACE) only allows the fabrication of nanowires with cylindrical cross sections.18,30−32 This is unfortunate because these arrays have outstanding optoelectronic properties that can show large modifications of light absorption over the 200−900 nm range that are very much dependent on nanowire diameter and morphology (cylindrical, rectangular, or hexagonal cross sections).33−37 Since nanostructuring of Si is difficult and usually involves the use of expensive microfabrication techniques that are limited to specialized research institutions (electron-beam lithography, reactive ion etching, etc.), the Received: November 7, 2016 Accepted: January 3, 2017

A

DOI: 10.1021/acsami.6b14226 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

inc., nominal diameter: 500 nm) or from lab-made PS (nominal diameter: 210 nm). The 500 nm PS sphere solution was prepared by centrifuging 4.0 g of the PS stock solution at 6000 rpm for 20 min. After centrifugation, 3.0 g of water was removed, leading to a solution with approximately 10% PS spheres. The concentrated dispersion was then used to prepare a 4:5 PS solution with methanol/Triton X-100 mixture (composed of 400:1 methanol/Triton X-100). Then, 100 μL of solution was dropped carefully onto the substrate surface, and spincoating was performed at 3500 rpm for 3 min on a Laurell WS-650SZ6NPP/LITE. These conditions avoided the formation of multilayers and yielded an approximate monolayer coverage >80% (i.e., with less than 20% of the substrate surface uncovered with PS spheres). The substrate coverage was determined via optical microscopy and SEM imaging. 2.5. 3D PS Sphere Arrays. The same PS sphere solution (i.e., mixed with methanol/Triton X-100) was drop-cast on top of cleaned Si substrate and allowed to evaporate at room temperature. 2.6. Plasma Etching. The PS particles were etched using a K1050X Plasma Asher (Quorum Emitech) equipped with an rf generator (13.6 MHz) and fed with O2. The etching power was set at 7 W, and the O2 flow rate was set at 2 mL/min. Under our experimental conditions, the hexagonal order of 2D arrays is lost when the particles become too small (i.e., slightly after the transition from hexagonal to circular cross section, around 250−300 nm diameter). A physical connection, in the form of a neck between the PS particles and the substrate, stabilizes the particles at the Si surface.39 Upon isotropic etching, this neck becomes thinner, up to a point where it is too thin to sustain good adhesion with the Si substrate. At this stage, the particles move, and the hexagonal order is lost. As reported by others, we found that annealing above the glass temperature of PS (∼95 °C) for a short time leads to a better adhesion, allowing for longer etching time. The faceting phenomena, however, is mostly undisturbed by this step: It occurs whether or not the PS particle monolayer is annealed, as long as the annealing temperature is ≤120 °C. 2.7. Nanosphere Lithography. Nanosphere lithography (NSL) is based on the use of the PS sphere hcp arrays as physical masks during physical evaporation of the metal film.20,22 Combined with plasma etching to reduce the PS sphere size, NSL can be used to produce metal films perforated with an array of nanoholes. The PS particle 2D arrays on Si were prepared, then used for NSL performed by first sputtering a very thin adhesion layer of ZnO doped with Al during 1s at 75 W using a Clustex 100 M sputtering system from Leybold Optics (base pressure before deposition around 1.10−6 mbar, Ar pressure during deposition: 3.10−3 mbar). The samples were then immediately sputtered with a 50 nm thick gold film using a Cressington Sputter Coater 108 auto (deposition at 10 mA during 1200 s). Finally, the spheres were removed using tape. The use of the ZnO adhesion layer was found to stabilize the gold film during the etching step (MACE). 2.8. Metal-Assisted Chemical Etching. Metal-assisted chemical etching (MACE) was used to prepare Si nanowires using the nanostructured gold film prepared via NSL in a HF/H2O2 mixture. Although the complete reaction pathway is still under debate, the general agreement is that H2O2 is reduced at the gold surface, injecting holes (h+) into the underlying silicon through the gold film. The injected holes oxidize Si that dissolves in the presence of HF as the soluble complex H2SiF6, generating H231

ability to modify Si nanowire morphology while using a simple, scalable, and cost-effective method will be beneficial and useful to researchers in this field. In this article, we report the use of isotropic plasma etching within a conventional inexpensive O2 radio frequency (rf, 13.6 MHz) plasma cleaner to control the morphology of hcp 2D and 3D PS colloidal crystals. Quite unexpectedly, the isotropy of the etching method used here leads to an anisotropic etching of the PS spheres. Indeed, due to the hexagonal symmetry of the nanosphere array, the etching rate varies depending on the location at the PS sphere surface: It is considerably slowed down at the point of contact between the spheres, leading to a hexagonal faceting of the polymeric sphere template. Our experimental measurements of local etching rates at the PS particle surface within 2D crystals indicate that 2D order directly affects the way nanostructures chemically interact with their surroundings: The hexagonal packing of the particles before etching leads to a hexagonal morphology after controlled etching. A similar phenomenon occurs within 3D superlattices. This demonstrates that the morphology of nanostructures ordered within crystalline 2D and 3D arrays can be tuned and programmed by the lattice symmetry of the array. We demonstrate the lithographic transfer of the modified PS particle arrays to generate 2D arrays of Si nanowires with welldefined hexagonal cross sections via NSL and MACE.

2. EXPERIMENTAL SECTION 2.1. Materials. Lab-grade acetone and ethanol, as well as 40% hydrofluoric acid (HF, AnalaR NORMAPUR analytical reagent), were obtained from VWR. Hydrogen peroxide (30% EMSURE ISO analytical reagent) was obtained from Merck Millipore. Styrene (≥99%), polyvinylpyrrolidone (PVP), potassium persulfate (KPS), and Triton X-100 were obtained from Sigma-Aldrich. An aqueous dispersion of 2.5% PS spheres was purchased from Polysciences Inc. (nominal diameter: 500 nm). All chemicals were used as received. 2.2. Scanning Electron Microscopy Imaging. Imaging was performed via field-emission gun scanning electron microscope (SEM) (Zeiss Ultra Plus 55) equipped with Gemini lenses. Image acquisition was performed on uncoated samples at 2 kV in the case of 2D arrays of PS spheres and 5 kV for the Au nanohole array and Si nanowires. The cross sections of the etched PS sphere 3D arrays were coated with a thin layer of gold (80 s, 20 mA with a Cressington Sputter Coater 108 auto) to limit charging and imaged at 1.5 kV. Image analysis was performed with the free software ImageJ. 2.3. Lab-Made 210 nm PS Spheres. Lab-made PS spheres were prepared according the procedure from Du et al.38 In a typical approach, 12.1 g of styrene was added to 100 mL of deionized water under low stirring at room temperature followed by the addition of 0.5 g of PVP. After 15 min of stirring at 300 rpm, 0.3 g of KPS, dissolved in 20 mL of deionized water, was added. This solution was deoxygenated under argon for 30 min under stirring. Subsequently, the solution was gradually heated up to 70 °C and stirred for 24 h. After centrifugation for 30 min at 20 000 rpm and three acetone washings, a white precipitate was obtained, which was further dispersed in 100 mL of deionized water. The average size of the spheres was 210 ± 6 nm, as measured via SEM imaging (>100 spheres measured). 2.4. 2D PS Sphere Arrays. The Si substrates were prepared by cutting n-type silicon wafers (resistivity 0.1−10 Ω cm, from Infineon) into 2 cm × 2 cm squares. The Si squares were then successively sonicated in acetone, ethanol, and ultraclean water for 4 min. After drying in air, the samples were treated with oxygen plasma (100 W, 4 min, O2 flow rate: 10 mL/min) to provide a clean and hydrophilic surface. To minimize possible contaminations, the samples were spincoated with the PS spheres directly after the oxygen plasma treatment. The PS sphere dispersion solution was prepared by concentrating the stock PS solution obtained from a commercial source (Polysciences

H 2O2 + 2H+ → 2H 2O + 2h+

Si + 6HF + n h+ → H 2SiF6 + nH+ +

(1)

4−n H2 2

(2)

Because the reduction of H2O2 at the metal surface is much faster than that at the Si surface, fast anisotropic etching of Si occurs in the regions covered by the metal film. As a result, the metal film sinks in the Si substrate, whereas the areas not covered by the metal are not etched.31,32 Caution: Appropriate safety precautions have to be observed when working with hydrof luoric acid (HF). HF is a contact poison. MACE was performed within a Teflon beaker using a lab-made Teflon sample holder perforated with small holes. The MACE solution was B

DOI: 10.1021/acsami.6b14226 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 1. Evolution of PS sphere morphology during slow isotropic O2 plasma etching. (a) Schematic of the etching process: the initially circular particle horizontal cross section (perpendicular to the crystal c-axis) becomes hexagonal after intermediate etching time, and reverts back to circular after longer etching. (b−i) SEM images of the PS spheres after O2 etching at 7 W and 2 mL/min O2 flow during 0 min (b, c), 15 min (d, e), 20 min (f, g) and 25 min (h, i). Scale bars for top images (b, d, f, and h): 500 nm; bottom images (c, e, g, and i): 1 μm.

diameter becomes too small.39 Unlike previous reports based on fairly rapid and/or low-temperature etching conditions,32,39,40 we found that slow isotropic etching at room temperature within an O2 plasma cleaner, where the ions generated in the plasma are not accelerated to the substrate, leads to a clear faceting of the spheres during the early stages of the etching process. While it is unclear if such a process does occur when other etching conditions are used, such a phenomenon has not yet been reported and studied. Within our experimental setup, the main chemical species responsible for the etching of the PS particles are expected to be neutral oxygen atoms (O). They are generated within the gas phase in large quantities during the formation of the rf plasma and are known to be strong oxidants toward organic polymers.41 The results shown and discussed in this manuscript were obtained using 500 nm PS spheres from Polysciences, Inc. Additional results based on the use of lab-made 210 nm PS spheres are shown in the Supporting Information and are discussed later in the text. As shown in Figure 1, after 15 min etching at 7 W (O2 flow set at 2 mL/min for all experiments), the PS particles lose their spherical shape and start to exhibit a horizontal hexagonal cross section. The presence of small bridges between the particles shows that etching occurs at different rates depending on the position at the sphere surface. It is obvious that this effect arises from the fact that the spheres are initially arranged in an hcp array: The particle−particle connections are located in the areas where the spheres initially touched each other (shown in red in Figure 1a). Such bridges between the PS particles can be seen in the electron microscopy images provided in previous works at short O2 etching time.39,40 However, to date and to our knowledge, they have been vastly ignored. Further O2 etching leads to the disconnection of the small bridges between the PS particles, yielding a non-close-packed array of PS particles with a

prepared fresh before the etching process and was composed of 10 mL of hydrofluoric acid, 10 mL of ultra clean water, and 2 mL of hydrogen peroxide. The sample was placed on the Teflon holder and immersed in the etching solution for the desired duration. Afterward, the sample was rinsed with deionized water four times, rinsed with ethanol once, and dried in air. 2.9. Morphology Factor. The diagonal d and the height h (edgeto-edge distance) of the horizontal cross sections of the Si nanowires and PS particles were measured using SEM imaging. Figure S4 provides a schematic depiction of d and h as a function of the crosssectional morphology. The morphology factor (MF) was defined as (d − h) 1 MF = d × 0.1340 × 100 and expressed as percent. In case of a circular cross section, d = h and MF = 0. In the case of a perfect (d − h)

(2 − 3 )

≅ 0.1340 and MF = 100%. hexagonal cross section, d = 2 In the case of hexagonal cross sections with elongated tips, (d − h) > 0.1340 and MF > 100%. d

3. RESULTS AND DISCUSSION Oxygen plasma etching is a common method to reduce the size of PS spheres organized within 2D hcp arrays. It is very often performed via reactive ion etching (RIE) that provides a directional etching. During RIE, the ions generated in the plasma are accelerated by an electric field toward the sample surface, leading to a homogeneous etching of the spheres from top to bottom resulting in their flattening. The sphere horizontal cross section (i.e., parallel to the substrate surface) stays circular during the whole etching process.32 The use of RIE usually provides access to non-close-packed hexagonal arrays with sphere sizes smaller than those of isotropic plasma etching performed in a plasma etcher where no bias is applied at the sample surface. Indeed, RIE limits etching of the substrate−particle connection that stabilizes the particle at the substrate surface, which is not the case for isotropic etching, where the hexagonal order might be lost when the sphere C

DOI: 10.1021/acsami.6b14226 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 2. Evolution of particle dimension and morphology as a function of etching time (7 W, O2 flow rate 2 mL/min). (a) Average particle width along two different directions as a function of etching time. The error bars represent the standard deviation of at least 40 different measurements. Blue circles: L1, particle width along the free particle surfaces (i.e., not protected by neighboring particles), shown in blue and detailed on the scheme in b). Black hexagons: L2, particle width along the surfaces that are initially protected by neighboring particles (contact points), detailed on the scheme in b. The blue and black dotted lines are guides for the eye. (b) Schematic showing the evolution of the particle morphology and dimensions as a function of etching. eL1 and eL2 are the etch rates along directions L1 and L2, respectively. We distinguished three types of surfaces having distinct local environments, which get etched at different rates. Surfaces that were not in contact with other particles at the beginning of the etching are shown in blue (constant etch rate eL1 ≈ 8 nm/min). Surfaces in contact with PS particles at the beginning of etching are shown in red (very slow etch rate: eL2 = 1 nm/min). Corners of the isolated hexagons are shown in green (very fast etch rate: eL2 = 17 nm/min). I−III correspond to the three different etching regimes with different eL2.

beginning of the oxidation reaction (regime I) is, although expected, quite remarkable: It pertains to the idea that the local environment around nanoparticles can modify the way they interact with their surroundings (i.e., participate in catalytic or chemical reactions). The close-proximity of neighboring particles results in the faceting of the PS spheres due to physical shielding at the point of contact between the spheres. Additionally, the fact that during regime II the particles lose their hexagonal cross sections and revert back to spheres due to faster etching at the corners than at the edges is in line with previous reports. For example, Mirkin et al. and others found that gold nanoprisms and nanohexagons can be converted into gold nanodisks due to the faster oxidation occurring at the exposed corners of the nanoparticles during their dissolution in aqueous solution.42,43 We believe that a simple geometric argument can explain this phenomenon: The polymer chains located at the hexagon corners have a higher access to reactive O atoms and are therefore etched faster than the chains located along the edge. This is in complete agreement with our experimental etch rate values in this etching regime (i.e., eL1 = 8 nm/min and eL2 = 17 nm/min). To further confirm that the presence of the 6 nearestneighbor particles around each PS particles before etching is, indeed, the reason for the formation of particles with hexagonal cross sections, we investigated the shape of particles located at the grain boundaries. Such particles have asymmetric environments: One half of the particle is bare, while the other half is protected with 3 or 4 particles. As expected, after a short etching (regime I), these particles have asymmetric cross sections. Figure 3 shows SEM images of PS particles located at the grain boundaries after 15 min etching: The side of the particle located in the grain boundary (i.e., that with no adjacent PS particle) has a semicircular cross section, while the side in contact with the PS particle array has a semihexagonal cross section. Such a drastic change in reaction rates between the particles located within the array and the particles located at

hexagonal morphology. Image analysis of the PS sphere SEM images after 20 min of etching (Figure 1f,g) yields an average angle between adjacent edges of 130 ± 5°, which is quite close to the 120° angle expected for perfect hexagons. Finally, longer etching leads to a rounding of the hexagon edges, producing more spherical PS particles. To gain a better understanding of the etching mechanism, we studied how the PS particle dimensions changed as a function of etching. We measured the PS particles along two different directions, schematically pictured in Figure 2: L1, corresponding to the particle width between unprotected surfaces (i.e., that were not originally in contact with other PS particles), and L2, corresponding to the particle width of the protected surfaces (i.e., that were initially in direct contact with other PS particles). Therefore, when the particles are spherical, L1 and L2 are equal to the particle diameter. When the particles have a hexagonal cross section, L1 is the hexagon height (edge-to-edge distance), and L2 is the hexagon diagonal. There are 3 different etching regimes, corresponding to dramatic changes in etch rates along L2, referred to as eL2, while etch rate along L1, eL1, is nearly constant at ∼8 nm/min. During regime I (0−15 min etching), etching along L2 is suppressed: eL2 ≈ 1 nm/min; therefore, eL1 ≫ eL2. As a result, the particles that are initially spherical (L1 = L2), are transformed into particles with hexagonal cross sections connected with each other via small bridges (L2 > L1). As etching proceeds further (regime II, 15− 25 min), the connections between the PS particles are quickly broken, leading to particles with well-defined hexagonal cross sections in the substrate plane (L2 > L1). At this stage, etching along L2 is significantly increased: eL2 = 17 nm/min, while eL1 stays constant at 8 nm/min. With eL2 > eL1, etching is faster at the hexagon corners than along the edges. At the end of regime II, the PS particles have a nearly spherical shape (L2 ≈ L1). Further etching (regime III, > 25 min) yields spherical particles, and etching becomes isotropic with eL1 = eL2 ≈ 8 nm/min. The fact that etching along L2 is dramatically suppressed at the D

DOI: 10.1021/acsami.6b14226 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 3. Isotropic O2 etching at the grain boundaries. (a) Schematic of the etching process before (top) and after (bottom) isotropic O2 etching. (b, c) SEM images of the grain boundary (shown by the white dashed−dotted line in b) after 15 min etch, showing the asymmetric morphology of the particles located at the grain boundary.

the periphery is striking. To our knowledge, there is only one previous report of a similar phenomenon: Pileni et al. previously reported differences in oxidation rate within 2D arrays of Co nanoparticles, where the nanoparticles located at the crystal edge (outermost layer) were more easily oxidized than the particles within the array.44 However, the use of such an effect to asymmetrically sculpt nanoparticles in threedimensions, has not been reported before, and as such, represent an interesting approach to generate well-defined Janus nanoparticles. We wanted to go further and studied how etching proceeds within 3D colloidal crystals. Multilayered hcp PS sphere crystals were produced via self-assembly at the surface of a Si substrate. Although both fcc and hcp lattices are expected to form19,45 under our experimental conditions, we almost exclusively observed hcp crystals. Within the hcp crystal, each PS sphere has the coordination number 12. As expected from the results in the 2D case, the nearest-neighbor particles locally suppress etching at the particle surface. This leads to the 3D faceting of the spheres after O2 etching, as shown in Figure 4. Due to the hcp lattice, the particle cross section parallel to the crystal c-axis (perpendicular to the substrate surface) becomes hexagonal. This is significantly different from previous results where RIE was used to sculpt multilayered colloidal crystals based on shadowing effects.46 In that case, the top particles acted as a standard physical mask protecting the particles underneath during the anisotropic dry etching step,46 very much like a photoresist during photolithographic and dry etching steps. This method can only modify the first crystal layers and is only applicable with anisotropic dry etching. Here, the active species present in the gas phase are not accelerated to the substrate and reach the PS spheres via diffusion, leading to a homogeneous isotropic etching of the particles. This suggests that the threedimensional sculpting reported here is applicable to other materials and reactions that can occur in both the gas and solution phases, unlike previous results based on the use of directional RIE,46 which cannot be reproduced in solution for example. Therefore, the concept and results presented here should be quite general. Interestingly, the top layer gets etched much faster than does the layers underneath. A gradual

Figure 4. Isotropic O2 etching within 3D hcp crystals. (a) Schematic of the 12-coordinated PS sphere (shown in red) within the hcp crystal (left) with the contact points to neighboring spheres (right). (b−e) Cross-sectional SEM images of 3D PS particle hcp crystals, (101̅0) plane, prepared on a Si substrate and etched at 7 W. (b) Bottom layer (in contact with the Si substrate) of a 4-layered crystal on Si, 20 min etch, and (c−e) multilayered crystal on Si, 25 min etch, where layer 1 corresponds to the top layer (crystal−air interface). (c, d) Layer 2. (e) Top 6 layers. Scale bars are 200 nm for all images. The small roughness at the surface of the spheres originates from the thin Au layer sputtered to limit charging during SEM imaging.

decrease in etch rate as one goes from the top to the bottom of the particle superlattice is observed: At the fifth layer beneath the surface, the etch rate is already reduced by more than 70% (Figures 5 and S1). This gas phase oxidation reaction can be used to probe chemical reactions in three dimensions within extended systems and shows that particles located below the surface are oxidized slower, probably due to limited mass transport through the pores of the hcp crystal (around 110 nm wide). This suggests that chemical reactions occur at different rates depending on the location within the colloidal crystal. Such an observation should be taken into account when considering nanoparticle superlattices for catalysis.3,5 Additionally, the preparation of templates with a controllable gradual increase in PS particle width (from top to bottom) would allow the fabrication of porous materials with interesting new geometries. This could be used, for example, to prepare plasmonic arrays composed of nanovoids with a range of diameters that could sustain a high number of localized surface plasmon resonances. Such materials would be of great interest for a range of applications such as light trapping surfaces, perfect light absorbers, or panchromatic plasmonic photocatalytic and photovoltaics systems.47−49 E

DOI: 10.1021/acsami.6b14226 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

hexagons (MF = 40%, diagonal: 320 ± 10 nm), respectively. Further O2 etching leads to typical cylindrical nanowires (MF ≈ 0, diameter: 300 ± 15 nm, Figure S5). As shown in Figure 6c, there is a good correlation between the MF of the PS template and resulting Si nanowires. We verified the generality and compatibility of this approach with PS spheres of different sizes and of different origins (obtained from commercial sources and lab-made) to synthesize hexagonal Si nanowires of different dimensions. Figure S6 shows results obtained using 210 nm PS templates that were made in our laboratories. A similar faceting phenomenon occurs, producing PS particles with a hexagonal morphology that can be transferred into a Si substrate to produce fairly well-defined hexagonal Si nanowires. We found that the hexagonal morphology of the Si nanowires produced after MACE is usually better defined when the wires are made using the large PS spheres (500 nm) rather than the smaller ones (210 nm) as templates. Overall, we found that we could reliably synthesize Si nanowires with acceptable hexagonal cross sections (40% < MF < 140%) with diagonals between ca. 65 and 90% of the original PS sphere diameter.

Figure 5. Local etch rates within 3D hcp crystals. (a) Schematic of the 3D hcp crystal. Green spheres correspond to the top layer (crystal−gas interface). (b) Relative etch rate decreases (in %) from layers 1 to 6 after 25 min O2 etching calculated as

(ER n − ER1) ER1

4. CONCLUSIONS We studied the isotropic etching of 2D and 3D PS sphere hcp arrays on Si substrates. We demonstrate the conversion of spherical PS templates into particles with a hexagonal cross section due a suppressed etching rate at the point of contact between the PS particles, originating from their arrangement in a hcp lattice. Interestingly, we measured a significant decrease of etching rate when the PS particles were located within the hcp crystal rather than at the crystal edge or surface. Successful transfer of the hexagonal shape via NSL and MACE to produce Si nanowires with well-defined hexagonal morphologies demonstrates the applicability of this etching approach for tuning the morphology of sacrificial templates. This work is significant because of the following: (i) It shows that nanoscale order can affect chemical reaction rates within colloidal crystals (modification of local reaction rates depending on the number of nearest-neighbors and location within the 2D or 3D crystals). (ii) It sets a precedent for predicting and controlling the shape of nanostructures arranged within 2D and 3D crystals based on lattice symmetry (here, a hexagonal order leads to a hexagonal morphology). This demonstrates that well-defined reaction conditions based on the use of specific ligands, temperature, or nanoparticle crystallinity are not always a prerequisite for the controlled modification of nanoparticle morphology.50−52 The present work shows an alternative etching pathway to tune nanoparticle shapes based on the nanoparticle local environment. While we have demonstrated our etching approach in the simple case of PS sphere hcp crystals, we believe that it should be compatible with a range of nanomaterial superlattices with different symmetries and compositions. We also expect that this concept is not limited to gas phase oxidation and should be applicable to other chemical reactions in the liquid phase. Recent developments in nanomaterial synthesis and processing have shown that the structure of such colloidal crystals can now be predicted, modified, and accessed by a range of approaches.15−17,19,53−55 This work might pave the way to a wealth of discoveries in the area of materials synthesis by design.

× 100 where ERn is

the etch rate at the layer n. The black dotted line is a guide to the eye.

In order to show the opportunities offered by our etching strategy, we used the faceted PS particles as 2D lithographic masks for nanosphere lithography (NSL; for details, see Experimental Section) to prepare gold films perforated with hexagonally shaped holes. The resulting nanostructured films were then used as catalytic masks during chemical etching in a HF/H2O2 mixture via MACE (Figures 6a and S2). Catalytic reduction of H2O2 at the metal surface generates holes that locally oxidizes Si, dissolving as H2SiF6 in the presence of HF.31,32 As a result, fast anisotropic etching of Si occurs in the regions covered by the metal film, resulting in the production of Si nanowires. This process is described in more detail in the Experimental Section. While this approach is a powerful way to chemically synthesize Si nanowire arrays with well-defined dimensions over very large areas, it is currently strictly limited to nanowires having cylindrical cross sections due to the use of spherical particles for the NSL step.18,30−32 We first performed NSL by sputtering a 50 nm gold film on PS particle arrays that were etched in the O2 plasma cleaner for various durations. At short etching time (15−20 min), an array of gold holes with a clear hexagonal shape results (Figure 6b), which after MACE produces nanowires with well-defined hexagonal cross sections over large areas (Figures 6d−m and S3). The bottom of the Si nanowires (region in contact with the Si substrate) has a clear hexagonal morphology (Figure 6d,g), supporting the fact that the hexagonal shape of the PS template is efficiently transferred during MACE. We quantified the morphology of the PS template and resulting Si nanowires with a morphology factor, MF (see Experimental Section and Figure S4 for details), where 0, 100, and >100% correspond to cross sections that are perfectly circular, hexagonal, and hexagonal with elongated corners, respectively. Using PS templates that were etched for 15, 20, and 25 min yields on average Si nanowires with cross sections that are elongated hexagons (MF = 129%, diagonal: 440 ± 20 nm), nearly perfect hexagons (MF = 98%, diagonal: 400 ± 15 nm), and rounded F

DOI: 10.1021/acsami.6b14226 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 6. Various Si nanowire morphologies produced by MACE using 500 nm PS templates on Si substrates. (a) Schematic of MACE. (b, d−m) SEM images. Top views: e, f, h, i, k and l; cross sections: d, g, j and m. (c) Morphology factor (MF) of the PS template and Si nanostructures produced as a function of plasma etching time of the PS templates. Magenta triangles: PS templates. Black squares: Si nanowires. The error bars represent the standard deviation of at least 40 different measurements. (b) Hexagonal gold hole array made after 15 min O2 etch at 7 W and gold deposition and lift off. (d−m) Si nanowires made after etching in HF/H2O2. (d−g) Hexagonal Si nanowires made from 15 min O2 NSL etch and 270 s MACE (MF = 129%, diagonal: 440 ± 20 nm). (h−j) Hexagonal Si nanowires made from 20 min O2 NSL etch and 270 s MACE (MF = 98%, diagonal: 400 ± 15 nm). (k−m) Rounded hexagonal Si nanowires made from 25 min O2 NSL etch and 180 s MACE (MF = 40%, diagonal: 320 ± 10 nm).



Notes

ASSOCIATED CONTENT

The authors declare no competing financial interest.

S Supporting Information *



The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsami.6b14226.



ACKNOWLEDGMENTS We gratefully acknowledge the Materials Chemistry group from the University of Salzburg for the use of their plasma oven, Dr. Nicola Hüsing for her helpful comments, and Dr. Nicolas Siedl (Infineon) for donating the Si wafers.

Additional figures of 3D crystal etching and MACE results (PDF)



AUTHOR INFORMATION

Corresponding Author

REFERENCES

(1) Arsenault, A. C.; Clark, T. J.; von Freymann, G. v.; Cademartiri, L.; Sapienza, R.; Bertolotti, J.; Vekris, E.; Wong, S.; Kitaev, V.; Manners, I.; Wang, R. Z.; John, S.; Wiersma, D. S.; Ozin, G. A. From colour fingerprinting to the control of photoluminescence in elastic photonic crystals. Nat. Mater. 2006, 5, 179−184.

*E-mail: [email protected]. ORCID

Gilles R. Bourret: 0000-0002-9774-1686 G

DOI: 10.1021/acsami.6b14226 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces (2) Gonzalez-Urbina, L.; Baert, K.; Kolaric, B.; Perez-Moreno, J.; Clays, K. Linear and Nonlinear Optical Properties of Colloidal Photonic Crystals. Chem. Rev. 2012, 112, 2268−2285. (3) Kang, Y.; Ye, X.; Chen, J.; Qi, L.; Diaz, R. E.; Doan-Nguyen, V.; Xing, G.; Kagan, C. R.; Li, J.; Gorte, R. J.; Stach, E. A.; Murray, C. B. Engineering Catalytic Contacts and Thermal Stability: Gold/Iron Oxide Binary Nanocrystal Superlattices for CO Oxidation. J. Am. Chem. Soc. 2013, 135, 1499−1505. (4) Baffou, G.; Berto, P.; Bermúdez Ureña, E.; Quidant, R.; Monneret, S.; Polleux, J.; Rigneault, H. Photoinduced Heating of Nanoparticle Arrays. ACS Nano 2013, 7, 6478−6488. (5) Auyeung, E.; Morris, W.; Mondloch, J. E.; Hupp, J. T.; Farha, O. K.; Mirkin, C. A. Controlling Structure and Porosity in Catalytic Nanoparticle Superlattices with DNA. J. Am. Chem. Soc. 2015, 137, 1658−1662. (6) Wong, S.; Kitaev, V.; Ozin, G. A. Colloidal Crystal Films: Advances in Universality and Perfection. J. Am. Chem. Soc. 2003, 125, 15589−15598. (7) Ebbesen, T. W.; Lezec, H. J.; Ghaemi, H. F.; Thio, T.; Wolff, P. A. Extraordinary optical transmission through sub-wavelength hole arrays. Nature 1998, 391, 667−669. (8) Zhao, L.; Kelly, K. L.; Schatz, G. C. The Extinction Spectra of Silver Nanoparticle Arrays: Influence of Array Structure on Plasmon Resonance Wavelength and Width. J. Phys. Chem. B 2003, 107, 7343− 7350. (9) Haynes, C. L.; McFarland, A. D.; Zhao, L.; van Duyne, R. P. V.; Schatz, G. C.; Gunnarsson, L.; Prikulis, J.; Kasemo, B.; Käll, M. Nanoparticle Optics: The Importance of Radiative Dipole Coupling in Two-Dimensional Nanoparticle Arrays. J. Phys. Chem. B 2003, 107, 7337−7342. (10) Ross, M. B.; Blaber, M. G.; Schatz, G. C. Using nanoscale and mesoscale anisotropy to engineer the optical response of threedimensional plasmonic metamaterials. Nat. Commun. 2014, 5, 4090. (11) Young, K. L.; Ross, M. B.; Blaber, M. G.; Rycenga, M.; Jones, M. R.; Zhang, C.; Senesi, A. J.; Lee, B.; Schatz, G. C.; Mirkin, C. A. Using DNA to Design Plasmonic Metamaterials with Tunable Optical Properties. Adv. Mater. 2014, 26, 653−659. (12) Phillips, K. R.; England, G. T.; Sunny, S.; Shirman, E.; Shirman, T.; Vogel, N.; Aizenberg, J. A colloidoscope of colloid-based porous materials and their uses. Chem. Soc. Rev. 2016, 45, 281−322. (13) Ross, M. B.; Mirkin, C. A.; Schatz, G. C. Optical Properties of One-, Two-, and Three-Dimensional Arrays of Plasmonic Nanostructures. J. Phys. Chem. C 2016, 120, 816−830. (14) Kanipe, K. N.; Chidester, P. P. F.; Stucky, G. D.; Moskovits, M. Large Format Surface-Enhanced Raman Spectroscopy Substrate Optimized for Enhancement and Uniformity. ACS Nano 2016, 10, 7566−7571. (15) Talapin, D. V.; Shevchenko, E. V.; Bodnarchuk, M. I.; Ye, X.; Chen, J.; Murray, C. B. Quasicrystalline order in self-assembled binary nanoparticle superlattices. Nature 2009, 461, 964−967. (16) MacFarlane, R. J.; Lee, B.; Jones, M. R.; Harris, N.; Schatz, G. C.; Mirkin, C. A. Nanoparticle Superlattice Engineering with DNA. Science 2011, 334, 204. (17) Li, F.; Josephson, D. P.; Stein, A. Colloidal Assembly: The Road from Particles to Colloidal Molecules and Crystals. Angew. Chem., Int. Ed. 2011, 50, 360−388. (18) Gao, P.; He, J.; Zhou, S.; Yang, X.; Li, S.; Sheng, J.; Wang, D.; Yu, T.; Ye, J.; Cui, Y. Large-Area Nanosphere Self-Assembly by a Micro-Propulsive Injection Method for High Throughput Periodic Surface Nanotexturing. Nano Lett. 2015, 15, 4591−4598. (19) Vogel, N.; Retsch, M.; Fustin, C.-A.; del Campo, A.; Jonas, U. Advances in Colloidal Assembly: The Design of Structure and Hierarchy in Two and Three Dimensions. Chem. Rev. 2015, 115, 6265−6311. (20) Haynes, C. L.; Van Duyne, R. P. Nanosphere lithography: a versatile nanofabrication tool for studies of size-dependent nanoparticle optics. J. Phys. Chem. B 2001, 105, 5599−5611.

(21) Gao, D.; Chen, W.; Mulchandani, A.; Schultz, J. S. Detection of tumor markers based on extinction spectra of visible light passing through gold nanoholes. Appl. Phys. Lett. 2007, 90, 073901. (22) Lee, S. H.; Bantz, K. C.; Lindquist, N. C.; Oh, S.-H.; Haynes, C. L. Self-Assembled Plasmonic Nanohole Arrays. Langmuir 2009, 25, 13685. (23) Sannomiya, T.; Scholder, O.; Jefimovs, K.; Hafner, C.; Dahlin, A. B. Investigation of plasmon resonances in metal films with nanohole arrays for biosensing applications. Small 2011, 7, 1653−1663. (24) Sigle, D. O.; Perkins, E.; Baumberg, J. J.; Mahajan, S. Reproducible Deep-UV SERRS on Aluminum Nanovoids. J. Phys. Chem. Lett. 2013, 4, 1449−1452. (25) Hall, A. S.; Friesen, S. A.; Mallouk, T. E. Wafer-Scale Fabrication of Plasmonic Crystals from Patterned Silicon Templates Prepared by Nanosphere Lithography. Nano Lett. 2013, 13, 2623−2627. (26) Ding, B.; Hrelescu, C.; Arnold, N.; Isic, G.; Klar, T. A. Spectral and Directional Reshaping of Fluorescence in Large Area SelfAssembled Plasmonic−Photonic Crystals. Nano Lett. 2013, 13, 378− 386. (27) Li, J.; Cushing, S. K.; Zheng, P.; Meng, F.; Chu, D.; Wu, N. Plasmon-induced photonic and energy-transfer enhancement of solar water splitting by a hematite nanorod array. Nat. Commun. 2013, 4, 2651. (28) Rycenga, M.; Langille, M. R.; Personick, M. L.; Ozel, T.; Mirkin, C. A. Chemically isolating hot spots on concave nanocubes. Nano Lett. 2012, 12, 6218−6222. (29) Brongersma, M. L.; Halas, N. J.; Nordlander, P. Plasmoninduced hot carrier science technology. Nat. Nanotechnol. 2015, 10, 25−34. (30) Huang, Z.; Fang, H.; Zhu, J. Fabrication of Silicon Nanowire Arrays with Controlled Diameter, Length, and Density. Adv. Mater. 2007, 19, 744−748. (31) Huang, Z.; Geyer, N.; Werner, P.; de Boor, J. d.; Gösele, U. Metal-Assisted Chemical Etching of Silicon: A Review. Adv. Mater. 2011, 23, 285−308. (32) Yeom, J.; Ratchford, D.; Field, C. R.; Brintlinger, T. H.; Pehrsson, P. E. Decoupling Diameter and Pitch in Silicon Nanowire Arrays Made by Metal-Assisted Chemical Etching. Adv. Funct. Mater. 2014, 24, 106−116. (33) Cao, L.; Fan, P.; Vasudev, A. P.; White, J. S.; Yu, Z.; Cai, W.; Schuller, J. A.; Fan, S.; Brongersma, M. L. Semiconductor Nanowire Optical Antenna Solar Absorbers. Nano Lett. 2010, 10, 439−445. (34) Kim, S.-K.; Day, R. W.; Cahoon, J. F.; Kempa, T. J.; Song, K.-D.; Park, H.-G.; Lieber, C. M. Tuning Light Absorption in Core/Shell Silicon Nanowire Photovoltaic Devices through Morphological Design. Nano Lett. 2012, 12, 4971−4976. (35) Kim, S.-K.; Song, K.-D.; Kempa, T. J.; Day, R. W.; Lieber, C. M.; Park, H.-G. Design of Nanowire Optical Cavities as Efficient Photon Absorbers. ACS Nano 2014, 8, 3707−3714. (36) Abujetas, D. R.; Paniagua-Domínguez, R.; Sánchez-Gil, J. A. Unraveling the Janus Role of Mie Resonances and Leaky/Guided Modes in Semiconductor Nanowire Absorption for Enhanced Light Harvesting. ACS Photonics 2015, 2, 921−929. (37) Ko, M.; Baek, S.-H.; Song, B.; Kang, J.-W.; Kim, S.-A.; Cho, C.H. Periodically Diameter-Modulated Semiconductor Nanowires for Enhanced Optical Absorption. Adv. Mater. 2016, 28, 2504−2510. (38) Du, X.; He, J. Facile Size-Controllable Syntheses of Highly Monodisperse Polystyrene Nano- and Microspheres by Polyvinylpyrrolidone-Mediated Emulsifier-Free Emulsion Polymerization. J. Appl. Polym. Sci. 2008, 108, 1755. (39) Plettl, A.; Enderle, F.; Saitner, M.; Manzke, A.; Pfahler, C.; Wiedemann, S.; Ziemann, P. Non-Close-Packed Crystals from SelfAssembled Polystyrene Spheres by Isotropic Plasma Etching: Adding Flexibility to Colloid Lithography. Adv. Funct. Mater. 2009, 19, 3279− 3284. (40) Vogel, N.; Goerres, S.; Landfester, K.; Weiss, C. K. A Convenient Method to Produce Close- and Non-close-Packed Monolayers using Direct Assembly at the Air-Water Interface and H

DOI: 10.1021/acsami.6b14226 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces Subsequent Plasma-Induced Size Reduction. Macromol. Chem. Phys. 2011, 212, 1719−1734. (41) Jeong, J. Y.; Park, J.; Henins, I.; Babayan, S. E.; Tu, V. J.; Selwyn, G. S.; Ding, G.; Hicks, R. F. Reaction Chemistry in the Afterglow of an Oxygen-Helium, Atmospheric-Pressure Plasma. J. Phys. Chem. A 2000, 104, 8027−8032. (42) Hong, S.; Shuford, K. L.; Park, S. Shape Transformation of Gold Nanoplates and their Surface Plasmon Characterization: Triangular to Hexagonal Nanoplates. Chem. Mater. 2011, 23, 2011−2013. (43) O’Brien, M. N.; Jones, M. R.; Kohlstedt, K. L.; Schatz, G. C.; Mirkin, C. A. Uniform Circular Disks With Synthetically Tailorable Diameters: Two-Dimensional Nanoparticles for Plasmonics. Nano Lett. 2015, 15, 1012−1017. (44) Lisiecki, I.; Turner, S.; Bals, S.; Pileni, M. P.; Tendeloo, G. V. The Remarkable and Intriguing Resistance to Oxidation of 2D Ordered hcp Co Nanocrystals. A New Intrinsic Property. Chem. Mater. 2009, 21, 2335−2338. (45) Nho, H. W.; Kalegowda, Y.; Shin, H.-J.; Yoon, T. H. Nanoscale characterization of local structures and defects in photonic crystals using synchrotron-based transmission soft X-ray microscopy. Sci. Rep. 2016, 6, 24488. (46) Choi, D.-G.; Yu, H. K.; Jang, S. G.; Yang, S.-M. Colloidal Lithographic Nanopatterning via Reactive Ion Etching. J. Am. Chem. Soc. 2004, 126, 7019−7025. (47) Li, W.; Valentine, J. Metamaterial Perfect Absorber Based Hot Electron Photodetection. Nano Lett. 2014, 14, 3510−3514. (48) Mubeen, S.; Lee, J.; Liu, D.; Stucky, G. D.; Moskovits, M. Panchromatic Photoproduction of H2 with Surface Plasmons. Nano Lett. 2015, 15, 2132−2136. (49) Hägglund, C.; Apell, S. P. Plasmonic Near-Field Absorbers for Ultrathin Solar Cells. J. Phys. Chem. Lett. 2012, 3, 1275−1285. (50) Xia, Y.; Xiong, Y.; Lim, B.; Skrabalak, S. E. Shape-Controlled Synthesis of Metal Nanocrystals: Simple Chemistry Meets Complex Physics? Angew. Chem., Int. Ed. 2009, 48, 60−103. (51) Personick, M. L.; Mirkin, C. A. Making Sense of the Mayhem behind Shape Control in the Synthesis of Gold Nanoparticles. J. Am. Chem. Soc. 2013, 135, 18238−18247. (52) O’Brien, M. N.; Jones, M. R.; Brown, K. A.; Mirkin, C. A. Universal Noble Metal Nanoparticle Seeds Realized Through Iterative Reductive Growth and Oxidative Dissolution Reactions. J. Am. Chem. Soc. 2014, 136, 7603−7606. (53) Henzie, J.; Grünwald, M.; Widmer-Cooper, A.; Geissler, P. L.; Yang, P. Self-assembly of uniform polyhedral silver nanocrystals into densest packings and exotic superlattices. Nat. Mater. 2011, 11, 131− 137. (54) Tan, R.; Zhu, H.; Cao, C.; Chen, O. Multi-component superstructures self-assembled from nanocrystal building blocks. Nanoscale 2016, 8, 9944−9961. (55) Kim, H.; Kim, T.; Kim, D.; Shim, W. Wave-Tunable Lattice Equivalents toward Micro- and Nanomanipulation. Nano Lett. 2016, 16, 6472−6479.

I

DOI: 10.1021/acsami.6b14226 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX