Programming Nanoparticles in Multiscale: Optically Modulated

Feb 26, 2018 - Manipulating and tuning nanoparticles by means of optical field interactions is of key interest for nanoscience and applications in ele...
0 downloads 8 Views 10MB Size
www.acsnano.org

Programming Nanoparticles in Multiscale: Optically Modulated Assembly and Phase Switching of Silicon Nanoparticle Array Letian Wang,† Yoonsoo Rho,† Wan Shou,‡ Sukjoon Hong,†,§ Kimihiko Kato,∥ Matthew Eliceiri,† Meng Shi,†,⊥ Costas P. Grigoropoulos,*,† Heng Pan,‡ Carlo Carraro,# and Dongfeng Qi†,∇ †

Laser Thermal Laboratory, Department of Mechanical Engineering, University of California, Berkeley, California 94720-1740, United States ‡ Department of Mechanical and Aerospace Engineering, Missouri University of Science and Technology, Rolla, Missouri 65401, United States § Department of Mechanical Engineering, Hanyang University, 55 Hanyangdaehak-ro, 20 Sangnok-gu, Ansan, Gyeonggi-do 15588, Republic of Korea ∥ Department of Electrical Engineering and Information Systems, The University of Tokyo, Tokyo 113-0032, Japan ⊥ School of Energy and Power Engineering, Xi’an Jiaotong University, Xi’an 710049, People’s Republic of China # Department of Chemical and Biomolecular Engineering, University of California, Berkeley, California 94720-1462, United States ∇ Laboratory of Infrared Materials and Devices, The Research Institute of Advanced Technologies, Ningbo University, Ningbo, Zhejiang 315211, People’s Republic of China S Supporting Information *

ABSTRACT: Manipulating and tuning nanoparticles by means of optical field interactions is of key interest for nanoscience and applications in electronics and photonics. We report scalable, direct, and optically modulated writing of nanoparticle patterns (size, number, and location) of high precision using a pulsed nanosecond laser. The complex nanoparticle arrangement is modulated by the laser pulse energy and polarization with the particle size ranging from 60 to 330 nm. Furthermore, we report fast cooling-rate induced phase switching of crystalline Si nanoparticles to the amorphous state. Such phase switching has usually been observed in compound phase change materials like GeSbTe. The ensuing modification of atomic structure leads to dielectric constant switching. Based on these effects, a multiscale laser-assisted method of fabricating Mie resonator arrays is proposed. The number of Mie resonators, as well as the resonance peaks and dielectric constants of selected resonators, can be programmed. The programmable light-matter interaction serves as a mechanism to fabricate optical metasurfaces, structural color, and multidimensional optical storage devices. KEYWORDS: laser, modulated assembly, nanoparticle, Mie resonance, dewetting, crystallization manufacturing. On the other hand, pattern-guided laser10−12 and thermal13,14 dewetting, as well as liquid-assembly,15,16 can produce periodic metallic and dielectric nanostructures with a good combination of fidelity and low cost. However, although pre-patterning defines the arrangement of the building-blocks, it does not offer flexibility for complex and tunable patterning.

S

urface patterned periodic nanostructures are important building blocks to electronics1 and spintronics,2 chemical catalysts,3 plasmonic and photonic devices,4−6 as well as memory devices.7−9 Related sensing applications require high customization of specific optical and electrical signals, and therefore better control in the nanostructure pattern complexity and variation would be beneficial and desirable. Various efforts have been made to write nanostructures into well-defined configurations in an inexpensive way. Since high fidelity is usually associated with high-cost, e-beam lithography and spontaneous self-assembly represent two extremes in nano© XXXX American Chemical Society

Received: January 9, 2018 Accepted: February 20, 2018

A

DOI: 10.1021/acsnano.8b00198 ACS Nano XXXX, XXX, XXX−XXX

Article

Cite This: ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano

Figure 1. Process flow and schematic of the laser-induced modulated assembly (LiMA) and the laser-induced phase switching (LiPS). (A) Process flow of basic assembly. Schematics of (i) pulsed laser processing and (ii) assembled Si nanoparticle array. (iii−iv) SEM image of assembled nanoparticle array viewed from different tilted angles. False color was added in (iii) matching optical microscopy. (B) LiMA in micro and nanoscale. (i−iii) Schematics of the laser manipulation of the nanostructures using different parameters, including (i) polarization direction; (ii) laser fluence; and (iii) number of pulses. Representative SEM images of the fabricated Si nanoparticle array by tuning the polarization direction (iv,v), the laser fluence (vi,vii), and the number of pulses (ix,x), showing the capability of manipulating the nanoparticle size, number of nanoparticles, and particle symmetry. (C) LiPS of Si nanoparticle crystallinity and its effect on color appearance. A preassembled nanoparticle canvas (i) was irradiated with a 2nd pulse (iii) that changed the color from “green” to “red” through amorphization of crystalline Si. An arbitrary patterning of the “LTL” logo is demonstrated in dark-field microscopy (ii). The scale bars are 300 nm for A (iii−iv) and B (iv−iv) and 5 μm for C (ii).

Recent advancements in structural color,17−21 optical data storage,9,22,23 and active nanophotonic devices24−26 highlight the need of tuning and programming of the nano buildingblock patterns geometrical features and photonic properties. Tunable patterning of nanoparticles will shed light into building more complex nanostructures, as nanoparticles are the most basic nano building-blocks widely used functional components in aforementioned applications to date. Due to its noncontact and scalable nature, optical field patterning stands out as a promising and attractive option among various tuning methods.27−29 Printing Si resonators with tunable size was realized using the laser-induced forward transfer (LIFT) method.30 However, this method requires accurate alignment of the laser beam onto the donor substrate, as well as synchronization of the donor and receiver substrate positions for writing complex nanoparticle arrays with submicron unit size,31 which limits its scalability. Surface morphologies of Al17 and Ge islands18 were precisely tuned through laser pulse energy density and polarization, which successfully demonstrated the laser’s application in structural color printing. Therefore, a final milestone would be the development of a scalable process that can both pattern and tune the particle number, placement, and size. The availability of optically reconfigurable nanophotonics that are based on compound materials, like GeSbTe24−26 and VO2,32 suggests that a low-cost

optically tunable approach is needed for silicon based photonics and structural color.18 In this work, we showed that laser-induced modulated assembly (LiMA) could optically manipulate the nanostructure assembly process and simultaneously modulate the nanoparticle size, number, and placement. It is noteworthy that in typical dewetting the laser pulse acts simply as a transient heat source. In contrast, the LiMA process relies on the modulation of the local laser absorption due to the near-field optical energy coupling. Due to the near field interaction, LiMA does not require elaborate focusing of the laser beam and is easily scalable. Additionally, we reported laser-induced phase switching (LiPS), which offers the means to convert the crystalline Si nanoparticle to the amorphous state. Silicon nanostructure that is the most widely used photonic and semiconductor building block is optically switched to the amorphous state through an easily accessible and reversible transformation.30,33,34 To demonstrate the application, we further proposed a multiscale laser programmed fabrication of the dielectric photonic metasurfaces by combining LiMA and LiPS. Dielectric nanoparticles (NPs) offer low loss and compatibility with the conventional semiconductor processing and manufacturing.5,6 The present method can ultimately produce monoperiodic, biperiodic, and triperiodic patterns, hence, directly facilitating the application of Fano resonance35,36 and spectroscopy.37 B

DOI: 10.1021/acsnano.8b00198 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano

Figure 2. Polarization dependent near-field absorption and nanoparticle size modulation. (A, C, and F) SEM images of nanoparticle distributions over different polarization angles (i−iii) relative to the horizontal direction with a hole size of 250 nm (A) and 400 nm with high fluence (D) and low fluence (F). (B and E) FDTD simulated spatial distribution of the absorption cross-section for different polarization angles (i−iii) and a hole size of 250 nm (A) and 400 nm (D). Dashed lines are the boundaries of 4 quadrants, and the unit cell in FDTD simulation corresponds to the boxed area in A, D, and F. (C) 2D assembly pattern with increased fluence from left to right. Polarization directions are labeled with white arrows. Scale bars are 400 nm for A, D, and F and 800 nm for C, respectively.

RESULTS AND DISCUSSION The process flow is described in the Figure 1A for laser-induced modulated assembly of nanoparticles on pre-patterned amorphous silicon films. Amorphous silicon films are deposited on quartz substrates with a thickness of 20−50 nm. The film thickness was chosen so that dewetted nanoparticles fell into size range demonstrating the lowest mode of resonance. Followed by reactive ion etching (RIE), photolithography was applied to form nanohole array patterns. These pre-patterned holes, varying from 250 to 500 nm in diameter, are spaced with a period of 800 nm. Figure 1A, iii and iv, shows a 20 nm thick amorphous Si film with a 335 nm hole size irradiated by a single 7 ns laser pulse at 1.5 J/cm2 fluence. As a result of dewetting induced assembly (Figure 1A, iii and iv), a large Si nanoparticle (so-called “L” particle) was formed at the center of the four neighboring holes. Each hole is surrounded by symmetrically placed smaller sized particles (the “S” particles). The periodic nanoparticle array is different from the random nanoparticles generated on non-patterned silicon film (Figure S1). The assembled particles are of bottom-truncated spherical shape (Figure 1A, iii and iv). Theoretical studies have shown that nonspherical particles can be effectively modeled to predict and design resonators for metasurface applications. The L nanoparticle size fell at 230 ± 10 nm while the S particles were in the range of 140 ± 10 nm. An “Orange−Green” false color combination is added into the SEM image in Figure 1A, iii, corresponding to the actual bright-field micrograph (Figure 3E, i) that shows an “Orange+Green” alternating pattern. The color of the etched quartz is dark-brown and no scattering effect is observed (Figure S1). A laser-induced modulated assembly of nanoparticles in 2D and a laser-induced phase switching are schematically shown in Figure 1B,C. Modulations of the nanoparticle arrangements (including size, number, and symmetry) are achieved by

adjusting the laser beam polarization, amplitude, and the number of the pulse (Figure 1B, i−iii). Regarding the fabrication of Mie resonator metasurfaces, LiMA enables the deliberate placement of the resonance peak and the active selection of the resonator combination. Furthermore, LiPS can transform the nanoparticle’s crystallinity from crystalline to amorphous (Figure 1C). Figure 1C depicts the “LTL” logo written by subsequent nanosecond laser pulses onto a “Green− Green” canvas using a 50× objective lens (Figure 1C, iii). The “Green−Green” canvas is produced by the laser processing of a 25 nm thick film with a 500 nm hole size (Figure 1 C, i), yielding particle diameters of 280 and 130 nm, which both happen to yield green resonance peaks. The red “LTL” logo is composed of the laser amorphized “L” and “S” particles (Figure 1 C, ii). Selective dot-by-dot modification is possible with a 100× objective lens in conjunction with precise translation of the sample. Laser-Induced Modulated Assembly. Light polarization is the key LiMA mechanism that features near-field absorption and can be harnessed to tune the nanoparticle size. Using a sample rotational stage, the polarization was set at angles of 0, 45, and 90° with respect to the horizontal (+x) direction. Here, a 22 nm thick amorphous thin film with 250 nm hole pattern was irradiated by one 7 ns laser pulse. As illustrated in Figure 2 A, i−iii, the assembled S particles exhibited an interesting size variation against the polarization angle. In the case of light polarization in the horizontal direction (Figure 2A, i), smaller S particles (i.e., “SP” (70 ± 10 nm)) are produced in the polarization direction (P), while larger S particles (i.e., “SPP” (170 ± 5 nm)) are formed in the perpendicular direction of polarization (PP). On the other hand, laser irradiation polarized at a 45° angle produces just SPP particles (Figure 2A, ii). Detailed particle size statistics are included in Figure S2. The configuration shown in Figure 2A, iii, is identical to the one in Figure 2A, i, but rotated by 90°. The size of SP nanoparticles C

DOI: 10.1021/acsnano.8b00198 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano

Figure 3. Energy dependent assembly state transitions and the modulation of nanoparticle number. (A) SEM images and schematics of different nanoparticle arrangements irradiated by loosely focused laser pulses of increased fluence on a 400 nm hole patterned a-Si film. Nanoparticles assembled in PP and P directions are marked with blue and red colors, respectively. (B) Schematics of polarization and energy dependent modulated assembly in PP and P directions. The different assembled nanoparticle combinations are mapped on (C) the graph of the absorbed energy vs the position across a spot irradiated by a tightly focused Gaussian laser beam under increasing laser fluence from left to right. Dashed horizontal lines in B mark different energy levels for assembling S particles into different states (α, β, γ, and δ). Polarization directions in A−C are the same and labeled with white arrows with reference to the hole. (D−F) Modulation of “L” and “S” particle numbers above the ablation (i.e., nanoparticle removal) threshold. (D) SEM images of three representative arrangements (i−iii) under fluence of 1.5, 1.8, and 2.0 J/cm2 for a 335 nm hole pattern. (E) Bright-field optical microscopy corresponding to different arrangements in D. (F) Microscopy reflection spectra for the patterns shown in D and E. The scale bars are 1 μm for A, C, and D and 2 μm for E, respectively.

yielded the so-called SS particles, which are similar to those reported in the solidification of liquid metal stripes.11,39 Our observations highlight the crucial role of the polarization dependent heating, melting, and self-assembly induced by the near-field laser beam energy deposition. With a given laser wavelength, variation of hole sizes can define the magnitude of near-field enhancement and, therefore, tune the size range of nanoparticles. When the size of the hole is increased to 400 nm, the high fluence regime (Figure 2D) yields a similar pattern to Figure 2A, but the particle size difference between SP and SPP becomes much smaller; which is explained by the FDTD simulations, where a reduced near-field enhancement in PP direction was observed in Figure 2E. However, with a slightly lower fluence, we observed the particle in polarization direction split into 2 dots, whose placement also followed a polarization dependent pattern (Figure 2F). The formation of the 2-dot pattern is also related to laser amplitude, which will be explained in the following section. The second LiMA mechanism entails the programming of the number of nanoparticles via regulating the laser light amplitude (i.e., fluence). Two distinct regimes are identified for laser fluence below and above the ablation threshold. For laser fluence below the ablation threshold, LiMA offers control of the “S” nanoparticle numbers from 0 to 2 in two different directions (Figure 3A). The origin of such variation is based on the spatial absorption distribution and the energy dependent dewetting state transition. The schematic in Figure 3B illustrates how these two effects take place. With an increase of absorbed energy, four energy dependent “S” nanoparticle dewetting states are distinguished, i.e., α, β, γ, and δ. Here, the α

falls in the deep-blue scattering range. Note that the scattering captured under optical microscopy was overwhelmed by the response of the major resonator L particles. Finite-difference time-domain (FDTD) simulation revealed spatially varying absorption based on near-field light coupling to an amorphous nanopatterned Si film. The simulated absorption distribution in a unit cell is presented in Figure 2B, i−iii. Due to near-field optical effects, the local absorption enhancement is strongly dependent on the polarization angle. The local absorption distribution can be mapped into 4 quadrants; the left/right quadrants are labeled in the “x-direction” and the up/down quadrants are in the “y-direction”. As noted in Figure 2B and E, the distribution of the local absorption maxima (the so-called “hotspots”) is perpendicular to the polarization direction. Detailed absorption statistics can be found in Figure S2. We observed that the nonhomogeneous transient heating, melting, and dewetting triggered by near-field absorption should be the origin of the nanoparticle tuning mechanism. Figure 2C shows different stages of assembly induced by laser irradiation with fluence near the dewetting threshold. Nonuniform melting is observed at the initiation stage. The yquadrants melted and dewetted, exposing the substrate between the holes. At the transition stage, it appears that early melting in the y-direction introduced by concentrated hotspots led to the formation of SPP particles through direct dewetting. On the other hand, x-quadrants melted and merged into a continuous liquid stripe along the y-axis. Solidification arrested a Rayleigh− Plateau type of instability11,38 experienced by the liquid stripe. However, at increased fluence, this instability developed fully, breaking the stripe into large particles, while satellite droplets D

DOI: 10.1021/acsnano.8b00198 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano

Figure 4. Near-field enhanced selective particle removal and demonstration on anisotropic shape, scalability, and structural color. (A) First pulse-induced dewetting on 400 nm hole pattern. (i) is schematics and (ii) and (iii) are representative SEM images of dewetted patterns corresponding to Figure 3A, iii and vi. (B) Second pulse-induced selective removal of particles. (i) is schematics and (ii) and (iii) are the representative SEM images of pattern after the 2nd pulse corresponding to A, ii and iii. (C) SEM image of nanoparticle pattern corresponding to Figure 3A, vii, produced by (i) 1st pulse, (ii) selective (boxed region) removal, and (iii) complete removal of “S” particles (dashed region) following the second pulse irradiation. (D) FDTD simulation explains the laser selective removal of nanoparticles in the patterned array. (i, ii) Top view for simulated geometries for particles of different size (i) and same size (ii), where the unit cells are labeled with boxes. (iii, iv) Absorption profile for (i) and (ii), correspondingly. (E) With 45° of polarization and low fluence, the L particle is dewetted partially in a controlled way. (i) Rotated (45°) unit cell schematics indicating the quadrants of PP and P. (ii) Liquid stripes (dashed regions) are broken by the hole patterns. (iii) Large-scale controlled anisotropic nanostructure assembly. (F) Optical microscopy (i) and SEM (ii) images of the large-scale dewetted pattern. (G) A metasurface with various colors fabricated through a 7 ns laser pulse irradiation with a 1× objective lens on a 30 nm thick pre-patterned amorphous thin film. The scale bars are 10 μm for E (iii), F, and G and 1 μm for the rest.

state can occupy a wider region, allowing the formation of symmetrical and uniform assembled regions. A series of “S” nanoparticle combinations are obtained by applying loosely focused laser beams, and therefore shallower laser energy density gradients are shown in Figure 3A. Combinations of corresponding different states in the PP and P directions are mapped in Figure 3B. It is noteworthy that Figure 3A, ii, demonstrates dewetted states of “S” particles while the development of L particle has not been completed because the laser pulse energy level is relatively low. With hole sizes smaller than 400 nm, as the near-field enhancement is stronger, and therefore the transition will be faster, diminishing the chance to develop complex nanoparticle combinations (Figure S3). By applying laser fluence above the ablation threshold corresponding to the targeted particles, the number of “L” and “S” nanoparticles can be programmed (Figure 3D). The optical response is therefore also configured accordingly (Figure 3E,F). High fluence (1.8 J/cm2) subtracts L particles, thereby leaving a square 4-dot S particle array (Figure 3D, ii). An even higher fluence (2.0 J/cm2) will subtract 2 more S particles and generate 2-dot patterns (Figure 3D, iii). The preserved 2 S

state indicates dewetting initiation, where only a big patch of liquid started dewetting from the quadrant (refer to Figure 2C, initiation) with no S particle formation. Afterward, satellite particles are formed, from one small (β state), two small (γ state), to ultimately one large (δ state) S particle. The relative absorbed energy levels of these four states are labeled as dashed lines in Figure 3B. We then plotted two curves describing the absorbed energy against different locations across the laserirradiated region in Figure 3B (lower axis), where different locations on the substrate receive increasing laser fluence from left to right due to the Gaussian laser beam profile. The emergence of α, β, γ, and δ states on the substrate is linked to the absorbed energy curves in Figure 3B through arrows. Since the absorbed energy in the PP direction is higher than in the P direction, the PP line is higher than the P line in Figure 3B. As a consequence, we can see in Figure 3B that the transition of states always commences earlier in the PP direction than in the P direction, which is depicted Figure 3C. This difference leads to various combinations of particle numbers in two directions. It should be noted that the local laser energy density gradient on Figure 3C is steep, rendering the transition region too short. With shallower imparted energy density gradients, each energy E

DOI: 10.1021/acsnano.8b00198 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano

Figure 5. Laser-induced phase switching: characterization and explanation. (A) Schematics of nanoparticle array produced by (i) 1 pulse and (ii) 2 pulses. (B) Dark-field images of samples (i) before the 2nd laser pulse and (ii) after the 2nd laser pulse irradiation (the laser beam is 20 μm in diameter, indicated by the dashed line, and the regions marked by a, b, and c are the 2nd pulse irradiated area, transition area, and nonirradiated area, respectively). (C) (i) Raman-shift spectrum for locations within labeled area, and crystalline Si reference is also presented. Raman mapping of (i) peak position and (ii) peak fwhm for the sample after 2nd pulse. (D) Thermal and crystallization simulation on LiPS. (i) Schematics of simulated silicon nanocube on top of quartz substrate. (ii) Temperature history and (iii) cooling rate of the bottom point of nanocube (iv) simulated nanograin map of final silicon domain. All scale bars are 10 μm in length. The 2nd laser pulse has a 25 ns duration.

particles are located in the horizontal direction, which is also the hotspot location. This result revealed further evidence of the process dynamics following the laser-material interaction. The enhanced absorption leads to the dewetting of the “S” particle in the hotspot locations, after which the formation of the “L” particle and the “S” particle take place. The particle subtraction mechanism should happen in a later stage as the mass transfer driven by the spatially varying surface tension gradients could possibly provide enough force40−42 to remove the liquid in the colder domains. With the L particles being 0 or 4 and the S particles being 0, 2, 4, 6, or even 8, the combination of the nanoparticle arrangement is expanded to over 10 different layouts. Microscale reflection spectra were measured (Figure 3F), suggesting a promising application as programmed optical resonator arrays. The measurement setup is described in the Methods Section. For patterns with only “S” particles (Figure 3F, ii and iii), the evident green peaks around 550 nm are the overlapped electric dipole (ED) and magnetic dipole (MD) resonances (see Figure 5F) introduced by the “S” particles. The number of “S” particles can modify the intensity of the optical response without affecting the peak combination and position. For the pattern with both the “L” and “S” particles, additional ED and MD features in 800 and 630 nm are introduced by the “L” particles. The detailed simulation setup and analysis of dipole and quadrupole resonance can be found in Figures S4−7.

The third LiMA mechanism features selective nanoparticle subtraction based on the polarization and amplitude of second laser pulse irradiation. The method can effectively break the predefined “S” nanoparticle two-side symmetry (Figure 4A, ii and iii), leaving the single-side particle placement (Figure 4B, ii and iii, and C). The selective subtraction based on the additional laser pulse is related to the near-field enhanced absorption and ensuing rapid thermal expansion.43 Different S particle sizes lead to different subtraction mechanisms. From the results of Figure 4A,B, the larger S particles are effectively removed due to size-dependent resonance (Figure 4D, i and iii). However, when the S particles are of similar size (Figure 4C, i), the polarization dependent near-field enhancement becomes significant (Figure 4D, ii and iv), which was further validated in Figure 4C. At low fluence, only the particles in the P directions (Figure 4C, ii) are removed, while at higher fluence, the particles in the PP direction are also removed (Figure 4C, iii). It is also noted that the particle subtraction happens after the assembly process. We further demonstrate anisotropic patterning that features elliptical “L” particles and normal “S” particles (Figure 4E). The dewetting mechanism is illustrated in Figure 4E, i, with the PP and P quadrants for a 45° polarization. The PP direction shows an early dewetting feature as it absorbed the higher energy, which introduced 4 S particles in the 2 PP quadrants. In the P direction, the lagged dewetting formed liquid stripes diagonally between the holes (Figure 4E, ii) in a manner analogous to the F

DOI: 10.1021/acsnano.8b00198 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano

crystallization processes are fundamentally stochastic in nature,33 therefore yielding alternating polycrystalline and single crystalline Si upon successive pulses. However, the herein presented experimental evidence reveals that the application of the second ns laser pulse on a confined nanostructure can deterministically transform the particle from the poly crystalline to the complete amorphous state. To investigate the laser-induced melting and crystallization of the semiconductor NPs in the present experiment, we carried out a conductive heat transfer analysis coupled with nucleation statistics and crystal growth kinetics by utilizing the numerical modeling tool constructed to interpret the aforementioned TEM observations.33,34 A 25 ns laser irradiation at 532 nm wavelength is delivered at normal incidence to a periodic Si nanoparticle (Figure 5D, i). During the laser irradiation, the temperature increases (Figure 5D, ii) and once it reaches the melting point, the temperature ramp slows down as part of the laser energy contributes to the latent heat for phase change. Then the temperature started to cool down and the possibility of nucleation increases with the decrease of temperature. Once crystallization happens, the released latent heat raises the temperature increase again in a recalescence process. We captured the cooling rate before the crystallization and plotted it out in Figure 5D, iii. The predicted cooling rate of the NPs on a flat substrate (see Figure 5D, iv) is 5 × 1010 to 7 × 1010 K/ s, which is 5 times higher than that reported in ref 33. This difference is primarily due to the three-dimensional nature of the conductive heat transfer loss to the substrate (Figure 5D, i) in the present work, rather than the essentially one-dimensional heat loss along the narrow pillar in the above referenced study. A faster cooling rate leads to a deeper supercooling and a transition from the quasi steady state (QSS) nucleation to nonQSS nucleation and then to athermal nucleation.48,50 For fast quenched liquid silicon thin film, the critical cooling rate for “complete amorphization” is estimated experimentally above 1010 K/s51 or theoretically at 6.65 × 1010 K/s,50 in agreement with our observations. The silicon domain grain map in Figure 5D, iv, shows numerous close-packed nanocrystals are generated upon solidification, which to the best extent resembles the amorphization phenomena. Nanoparticle amorphization by nanosecond laser heating provides an intriguing phase transformation mechanism that can contribute to the fundamental investigation of nanoscale crystallization. To further investigate the modification of the optical characteristics, we carried out microreflection spectral measurement in conjunction with FDTD optical simulation (Figure 6B). The bright-field microscopy image (Figure 6A, i) shows the pattern before and after an additional laser pulse (Figure 6A, ii, dashed region is laser beam). We can clearly see the color change from the second pulse irradiation, i.e., the green “2-dot” (a) and “4-dot” (c) region was transformed to yellow color (b and d) after the second laser pulse. The color response in the bright-field somewhat differs from the dark-field image (Figure 5B, ii), as the reflected component from the substrate is strong and serves as a large background color on top of scattered color. In the FDTD simulation, we compared the reflection spectra modeled using single crystal silicon properties with predictions using amorphous silicon data.52 Details of the simulation can be found in the Methods Section. The reflection spectra match the simulation results well, especially regarding the peak locations. In both the single pulse and the double pulse cases, the magnetic dipole (MD) and the electric dipole (ED) moments contribute to the detected green and red peaks,

transition depicted in Figure 2C. However, since the liquid stripes are separated by the periodic hole array, the delayed dewetting formed ellipsoidal features (Figure 4E, iii). The method shows good scalability, uniformity, and potential applications in gradient resonator array and optical structural color surfaces. On the current experimental apparatus, the metasurface fabricated by a single Gaussian laser pulse is ∼25 μm in size (Figure 4F, i,ii). A uniform light field induced by the flat-top beam profile can be easily applied to fabricate larger scale patterns. A full visible color spectrum from blue to red is demonstrated with the single laser pulse irradiation in Figure 4G by adjusting the laser fluence. The modulation of the nanoparticle combination and crystallinity via the applied laser fluence and the respective variation of the color response will be discussed in the subsequent section. The emergence of nanoparticles having blue and green spectra is achieved in the high fluence regime, where L particles have been removed (to the very right of Figure 4G). Laser-Induced Phase Switching. Besides LiMA, we report laser-induced phase switching (LiPS) on silicon nanoparticles. Due to the Gaussian energy profile of the laser beam (Figure 5A, i), three different configuration patterns (Figure 3E, i−iii) are obtained on one spot (Figure 5B, i) irradiated by a 7 ns laser pulse. The nanoparticle arrays were irradiated again by a 25 ns laser pulse through a higher magnification objective lens (10×) at the fluence of 2.5 J/ cm2(Figure 5A, ii). A red shift of scattered light after the additional laser pulse was observed in the optical dark-field image in Figure 5B, i and ii, which can be attributed to the crystallinity decrease predicted from the optical resonance calculation.30 Figure S8 further confirms that oxidation is not the cause for this red-shift. Upon an additional laser pulse, the Raman spectrum point measurement (Figure 5C, i, curves a−c) shows a clear shift of the peak position and broadening of the full width half maximum (fwhm) compared to the crystalline Si, that are indicative of smaller crystalline grains (curve b) and amorphous state (curve a).44 As silicon Raman scattering is a well-studied topic, it is noteworthy that the size effect45 and the strain effect46 will only reflect on the peak broadening and shift of the Raman signal. Only the amorphous silicon structure will remove the complete peak.44 The Raman spectra has been normalized to each maximum intensity. The Raman signal of the amorphous nanoparticle presents noisy features because it is very weak. The SiO2 Raman peak is located in the 400 nm,47 which is far from the probed results, and it eliminated the effects of oxidation. It is further validated that the geometries of nanoparticles do not change upon the second laser irradiation (Figure S9). Raman mapping of the peak position and fwhm (Figure 5C, ii, iii) further prove that the second laser pulse amorphized the region with “L+S”, 4-dot S, and 2-dot S nanoparticles (Figure 3E, i−iii). Thermal analysis shows that amorphization induced by the nanosecond laser pulse is driven by the fast cooling rate experienced by a fully melted domain and the ensuing athermal nucleation mechanism. Athermal nucleation implies that clusters become supercritical as a consequence of dramatic reduction of required critical size, rather than the growth of clusters observed in classical thermal nucleation.48 Single or multiple nanosecond laser pulses are typically applied in the display industry to anneal the amorphous silicon thin films in order to enhance their crystallinity.49 The in-situ transmission electron microscopy (TEM) observation of a nanoconfined Si precursor (200 nm) showed that the nucleation and G

DOI: 10.1021/acsnano.8b00198 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano

second pulse irradiation on the bits displayed in Figure 6A, iii. All the laser processing is done by translating an automated x− y-stage. If the crystalline and the amorphous pixels are assigned binary values of 0 and 1, the digitized bit information can be recorded in the silicon nanoparticles. Even though the amorphization is demonstrated on the LiMA-produced nanoparticles, a different nanostructure is expected to have similar amorphization effects as long as the cooling rate requirement is met. In addition, since LiMA can modulate the physical geometry and establish direct correlation of the optical properties with respect to the applied optical intensity and polarization, it is possible to record information through the transient optical field. These encoding effects offer a pathway toward CMOS compatible multiplexed optical data storage.9 Further investigations on material properties and characteristic length scale in LiMA and LiPS are needed to establish a more elaborate programming schemes. LiMA involves sequentially light absorption, melting, dewetting, and solidification. These four fundamental physical processes are regulated by the light-material interaction, the coupled heat and mass transport, phase-transformation, and nucleation phenomena. Therefore, the absorption, thermal diffusivity, surface tension, and crystallographic properties can all contribute to the process. We note that the formation of the additional S particles is markedly different than the production of particles in a holepatterned Au film dewetted by the electron beam,57 but shares some similarity to Ni film dewetting upon laser pulse irradiation.11 The characteristic length scales of the hole size and periodicity will affect not only the optical near-field, localized energy deposition, but will also the competing mechanisms of cumulative mass transport and Raleigh−Plateau type of instability.38 As a consequence, the ability to induce modulated/directed assembly via the controlled dewetting process and the respective limiting cases should be further studied. In order to further understand LiPS, a systematic investigation of laser parameter and geometries effects on the cooling rate should be carried out. Advanced computational58,59 and in situ experimental60 techniques can be helpful for revealing the underlying fundamental physical mechanisms.

Figure 6. Optical property characterization and potential application for LiPS. (A) Color modification and data storage demonstration based on LiPS. (i, ii) Bright-field reflection microscope image of a selected area before (i) and after (ii) additional laser pulse irradiation (laser beam is 20 μm in diameter). The dashed line marks the area irradiated by the 2nd laser pulse. Areas with labels of a, b, c, and d correspond to the regions with a different numbers of “S” particles and are linked with the curves in (F). (iii, iv) Dark-field microscope image illustrating on-demand site-selective phase switching (i) equally spaced laser dewetted regions (bits) and (ii) the 2nd pulse switched bits. (B) Reflection spectrum and simulated spectrum of assembled (1 pulse) and phase-switched (2 pulse) “S” nanoparticles. Scale bars are 10 μm in A and 1 μm in B insets. The 2nd laser pulse has a 25 ns duration.

marking the change of crystallinity. Additionally, the magnetic quadrupole (MQ), as well as the electrical quadrupole (EQ), are also detected in agreement with the reported sequence.53,54 Based on LiMA we developed a monoperiodic (Figures 4C, iii, and 3D, ii and iii), biperiodic (Figure 4A, ii), and triperiodic (Figure 4A, iii) compatible nanoparticle patterning method with the particle size ranging from 60 to 350 nm. The spectrum generated by the S particles can cover the range of “Blue” to “Yellow”, while the L particles can yield “Green” to “Red” spectral responses. The combination of the L and S particles provides additional tunability of the photonic response of the fabricated dielectric metasurface. Even though only orthogonal arrays are demonstrated in the present study, hexagonal and other irregular arrays can also be designed.55,56 LiPS of silicon nanoparticles can be applied to metasurface writing and optical data storage. Since the dielectric function of the amorphous Si differs appreciably from the crystalline Si, the photonic properties can be switched optically. Such crystallinity modification expands the metasurface fabrication capability, providing a user-defined laser writing mechanism. Apart from writing on the large area metasurfaces demonstrated in Figure 1C, ii, we illustrated optical recording/writing on the defined bits. Figure 6A, iii, presents the bits formed by the single laser pulses. Figure 6A ,iv, shows the encoded bits amorphized by the

CONCLUSION In summary, we proposed a laser programming method that can tune the nanoparticle arrays’ size, unit cell composition, symmetry, and crystallinity. The laser-induced modulated assembly (LiMA) utilizes nanosecond laser pulse irradiation as a source modulating the assembly of a pre-patterned amorphous Si film to a periodic nanoparticle array. We demonstrated polarization dependent particle size modulation, fluence dependent particle number modulation, and selective particle subtraction, which all leverage the optical near-field interaction during the fabrication process. We further reported laser-induced phase switching (LiPS), where the fast cooling rate upon nanosecond irradiation leads to the amorphization of the crystalline Si nanoparticles. LiMA and LiPS help program the silicon Mie resonator combination, as well as the resonance peak position, and the dielectric constant of selected resonators in three different scales. Hence, the present laser-based integrated approach offers an effective way for programming dielectric metasurfaces and would be potentially applied in structural color and multidimensional optical data storage. H

DOI: 10.1021/acsnano.8b00198 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano

grained crystals were not sufficient to demonstrate amorphization, as the simulator was based on the quasi-steady state nucleation. With considering the athermal nucleation flux, a complete amorphization was therefore anticipated.

METHODS Nanofabrication. Amorphous Si (a-Si, 20−50 nm) was deposited on a 600 μm-fused silica substrate via low pressure chemical vapor deposition (LPCVD) at 500 °C. The deposited film was then patterned with standard lithography and reactive ion etching (Cl2, HBr, and O2). During the process of etching, a 50 nm deep overetched into the oxide substrate was carried out. The over-etching ensured the melted film would not agglomerate. As for the base substrate, similar results were also obtained with a-Si deposited on a film of 2 μm low temperature oxide (LTO) on silicon wafer. Laser Processing. A high power Nd:YAG pulsed laser (New Wave Solo II Laser, 532 nm, 7 ns) was employed to fabricate large area arrays for characterization. The YAG laser was focused using a Mitutoyo APO 2X objective lens with a beam diameter at 40 μm. A high repetition rate Nd:YVO4 laser (Spectra Physics Navigator, 532 nm, 25 ns) was employed to study the second pulse effects on the tuning of the Si nanoparticle property. The laser was focused with a Mitutoyo APO 2X objective lens, creating a beam size of 15 μm in diameter. Characterization. To characterize the optical response, we utilized a dark-field compatible optical microscope (Olympus BX60). The images were taken with a CMOS camera (AmScope U300). The morphology characterization was carried-out in SEM (Quanta FEI). Raman mapping was performed with a Renishaw inVia Micro Raman System with a 488 nm laser as the excitation source. The mapping stage was capable of 100 nm resolution motion. The mapped spectrum was then processed with built-in WiRE 3.3 software provided by Renishaw, Inc. The spectra were fitted with combined Gaussian and Lorentzian profile peaks ranging at 480−525 cm−1. The reflection spectrum measurement was obtained by integrating a commercial reflection probe (Thorlabs RP28) with a 100× objective lens (Olympus MS Plan) and was measured with a spectrometer (Acton SpectraPro 2300i). The spectrometer was illuminated with a stabilized light source (Thorlabs SLS201) and the reflection spectra was obtained by normalizing the reflected signal over the illuminating signal. Simulation of the Optical Response. A commercial-grade simulator (Lumerical FDTD61) based on the finite-difference timedomain method was used to perform the calculations. For “S” particle, a 140 nm in diameter nanoparticle was placed on top of a 5 μm quartz with 10 nm overlapping to depict the truncated geometry. The detailed process of optimizing the size and overlap is described in Figures S4 and S5. For “L” particles, a 220 nm nanoparticle with 20 nm overlapping was adapted to describe the truncated geometry. The plane-wave source was at a 2 μm distance on top of quartz and the FDTD domain contained the source and a 2 μm quartz domain. Crystalline Si (c-Si) properties were selected from the built-in model while amorphous Si (a-Si) properties were from ref 52. A reflection monitor with a spectrum resolution of 10 nm/point was placed at a height matching the numerical aperture of the objective lens we used. A 5 nm grid in the xy-plane and a 2 nm grid in the z-direction with a 1000 fs simulation time were applied. The method to determine the resonance mode is described in Figures S6 and S7. Simulation of Heat Transfer and Crystallization. The simulation was carried out in MATLAB based on the simulation method reported earlier33,34 and was validated again in Figure S10. The Si dot was placed on quartz with ambient set to be vacuum. The nanoparticle was modeled as a rectangular solid with dimensions from the “S” particle measurements. The boundaries in the x- and ydirections were considered adiabatic to reflect the periodic structure. The thickness of quartz was chosen to be 2 times the extent of the heat affected zone. To reduce the uncertainty of laser fluence estimation to the cooling rate prediction, we modeled the fluence at the threshold of full-melting, which was smaller than the experiment value. The threshold fluence gave a cooling rate close to the predicted critical transition rate 66.5 × 109 K/s. As illustrated in Figure 5D, iii, higher fluence introduced a higher cooling rate, the experimental cooling rate was, therefore, considered beyond the estimated transition rate. Despite the valid prediction of the cooling rate, the simulated fine-

ASSOCIATED CONTENT S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsnano.8b00198. Bare silicon thin-film dewetting and ablation, polarization dependent nanoparticle size distribution and FDTD absorption simulation, fluence dependent nanoparticle assembly with smaller hole size, SEM measurement and FDTD simulation on single crystalline “S” nanoparticle, SEM, optical color mapping and FDTD simulation of “L +S” nanoparticle pattern, S nanoparticle E and H contour field and vector field at resonance peaks, “L” and “S” particles E and H contour field at resonance peaks, FDTD simulated reflection spectra for additional oxidation and oxide removal effects, geometry comparison with or without second pulse laser irradiation, and validation of crystallization simulation against references (PDF)

AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected]. ORCID

Costas P. Grigoropoulos: 0000-0002-8505-4037 Author Contributions

L.W. designed the experiment; carried out nanofabrication, laser processing, and characterization; as well as analyzed the results. Y.R. contributed in carrying out Raman analysis. W.S. and S.H. offered critical suggestions on the result summary of this work. K.K. helped and guided the nanofabrication. M.E assisted in demonstrating “LTL” patterns. D.Q. built the processing setup. C.C. assisted the Raman analysis. M.S. and H.P. offered helpful discussions on thermal and crystallization analysis. C.G. supervised and supported the work. Funding

The work is supported by the collaborative NSF Grant CMMI1363392 to C.G. and H.P. Notes

The authors declare no competing financial interest.

ACKNOWLEDGMENTS The authors are thankful to Prof. Junqiao Wu of UC Berkeley Materials Science and Engineering for providing access to Raman mapping system. We also thank Dr. Andrew King from Renishaw for helpful discussions about Raman mapping functionality. The nanofabrication and SEM were carried out at the Marvell Nanofabrication Laboratory and the California Institute of Quantitative Bioscience (QB3) of UC Berkeley. We would like to acknowledge Allan Huang’s participation in the fabrication of “LTL” logo. Lastly, we appreciate Dr. Jung Bin In’s contribution in developing the simulator. REFERENCES (1) Shipway, A. N.; Katz, E.; Willner, I. Nanoparticle Arrays on Surfaces for Electronic, Optical, and Sensor Applications. ChemPhysChem 2000, 1, 18−52. I

DOI: 10.1021/acsnano.8b00198 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano

(24) Wang, Q.; Rogers, E. T. F.; Gholipour, B.; Wang, C.-M.; Yuan, G.; Teng, J.; Zheludev, N. I. Optically Reconfigurable Metasurfaces and Photonic Devices Based on Phase Change Materials. Nat. Photonics 2016, 10, 60−65. (25) Tittl, A.; Michel, A. K. U.; Schäferling, M.; Yin, X.; Gholipour, B.; Cui, L.; Wuttig, M.; Taubner, T.; Neubrech, F.; Giessen, H. A Switchable Mid-Infrared Plasmonic Perfect Absorber with Multispectral Thermal Imaging Capability. Adv. Mater. 2015, 27, 4597− 4603. (26) Michel, A.-K. U.; Chigrin, D. N.; Maß, T. W. W.; Schönauer, K.; Salinga, M.; Wuttig, M.; Taubner, T. Using Low-Loss Phase-Change Materials for Mid-Infrared Antenna Resonance Tuning. Nano Lett. 2013, 13, 3470−3475. (27) Hong, S.; Lee, H.; Yeo, J.; Ko, S. H. Digital Selective Laser Methods for Nanomaterials: From Synthesis to Processing. Nano Today 2016, 11, 547−564. (28) Sygletou, M.; Petridis, C.; Kymakis, E.; Stratakis, E. Advanced Photonic Processes for Photovoltaic and Energy Storage Systems. Adv. Mater. 2017, 29, 1700335. (29) Makarov, S. V.; Zalogina, A. S.; Tajik, M.; Zuev, D. A.; Rybin, M. V.; Kuchmizhak, A. A.; Juodkazis, S.; Kivshar, Y. Light-Induced Tuning and Reconfiguration of Nanophotonic Structures. Laser Photonics Rev. 2017, 11, 1700108. (30) Zywietz, U.; Evlyukhin, A. B.; Reinhardt, C.; Chichkov, B. N. Laser Printing of Silicon Nanoparticles with Resonant Optical Electric and Magnetic Responses. Nat. Commun. 2014, 5, 3402. (31) Zywietz, U.; Schmidt, M. K.; Evlyukhin, A. B.; Reinhardt, C.; Aizpurua, J.; Chichkov, B. N. Electromagnetic Resonances of Silicon Nanoparticle Dimers in the Visible. ACS Photonics 2015, 2, 913. (32) Dong, K.; Hong, S.; Deng, Y.; Ma, H.; Li, J.; Wang, X.; Yeo, J.; Wang, L.; Lou, S.; Tom, K. B.; Liu, K.; You, Z.; Wei, Y.; Grigoropoulos, C. P.; Yao, J.; Wu, J. A Lithography-Free and FieldProgrammable Photonic Metacanvas. Adv. Mater. 2018, 30, 1703878. (33) Xiang, B.; Hwang, D. J.; In, J. B.; Ryu, S.-G.; Yoo, J.-H.; Dubon, O.; Minor, A. M.; Grigoropoulos, C. P. In Situ TEM near-Field Optical Probing of Nanoscale Silicon Crystallization. Nano Lett. 2012, 12, 2524−2529. (34) In, J. B.; Xiang, B.; Hwang, D. J.; Ryu, S.-G.; Kim, E.; Yoo, J.-H.; Dubon, O.; Minor, A. M.; Grigoropoulos, C. P. Generation of SingleCrystalline Domain in Nano-Scale Silicon Pillars by near-Field Short Pulsed Laser. Appl. Phys. A: Mater. Sci. Process. 2014, 114, 277−285. (35) Zhao, W.; Jiang, H.; Liu, B.; Jiang, Y.; Tang, C.; Li, J. Fano Resonance Based Optical Modulator Reaching 85% Modulation Depth. Appl. Phys. Lett. 2015, 107, 171109. (36) Zhao, W.; Leng, X.; Jiang, Y. Fano Resonance in All-Dielectric Binary Nanodisk Array Realizing Optical Filter with Efficient Linewidth Tuning. Opt. Express 2015, 23, 6858. (37) Huang, Z.; Wang, J.; Liu, Z.; Xu, G.; Fan, Y.; Zhong, H.; Cao, B.; Wang, C.; Xu, K. Strong-Field-Enhanced Spectroscopy in Silicon Nanoparticle Electric and Magnetic Dipole Resonance near a Metal Surface. J. Phys. Chem. C 2015, 119, 28127−28135. (38) Wu, Y.; Fowlkes, J. D.; Rack, P. D.; Diez, J. A.; Kondic, L. On the Breakup of Patterned Nanoscale Copper Rings into Droplets via Pulsed-Laser-Induced Dewetting: Competing Liquid-Phase Instability and Transport Mechanisms. Langmuir 2010, 26, 11972−11979. (39) González, A. G.; Diez, J. A.; Gratton, R.; Gomba, J. M. Rupture of a Fluid Strip under Partial Wetting Conditions. Europhys. Lett. 2007, 77, 44001. (40) Habenicht, A. Jumping Nanodroplets. Science 2005, 309, 2043− 2045. (41) Afkhami, S.; Kondic, L. Numerical Simulation of Ejected Molten Metal Nanoparticles Liquified by Laser Irradiation: Interplay of Geometry and Dewetting. Phys. Rev. Lett. 2013, 111, 1−5. (42) Qi, D.; Paeng, D.; Yeo, J.; Kim, E.; Wang, L.; Chen, S.; Grigoropoulos, C. P. Time-Resolved Analysis of Thickness-Dependent Dewetting and Ablation of Silver Films upon Nanosecond Laser Irradiation. Appl. Phys. Lett. 2016, 108, 211602.

(2) Cavallini, M.; Biscarini, F.; Gomez-Segura, J.; Ruiz, D.; Veciana, J. Multiple Length Scale Patterning of Single-Molecule Magnets. Nano Lett. 2003, 3, 1527−1530. (3) Somorjai, G. A.; York, R. L.; Butcher, D.; Park, J. Y. The Evolution of Model Catalytic Systems; Studies of Structure, Bonding and Dynamics from Single Crystal Metal Surfaces to Nanoparticles, and from Low Pressure (10 −3 Torr) to Liquid Interfaces. Phys. Chem. Chem. Phys. 2007, 9, 3500−3513. (4) Stewart, M. E.; Anderton, C. R.; Thompson, L. B.; Maria, J.; Gray, S. K.; Rogers, J. A.; Nuzzo, R. G. Nanostructured Plasmonic Sensors. Chem. Rev. 2008, 108, 494−521. (5) Kuznetsov, A. I.; Miroshnichenko, A. E.; Brongersma, M. L.; Kivshar, Y. S.; Luk’yanchuk, B. Optically Resonant Dielectric Nanostructures. Science 2016, 354.aag247210.1126/science.aag2472 (6) Jahani, S.; Jacob, Z. All-Dielectric Metamaterials. Nat. Nanotechnol. 2016, 11, 23−36. (7) Cui, P.; Seo, S.; Lee, J.; Wang, L.; Lee, E.; Min, M.; Lee, H. Nonvolatile Memory Device Using Gold Nanoparticles Covalently Bound to Reduced Graphene Oxide. ACS Nano 2011, 5, 6826−6833. (8) Tsoukalas, D. From Silicon to Organic Nanoparticle Memory Devices. Philos. Trans. R. Soc., A 2009, 367, 4169−4179. (9) Zijlstra, P.; Chon, J. W. M.; Gu, M. Five-Dimensional Optical Recording Mediated by Surface Plasmons in Gold Nanorods. Nature 2009, 459, 410−413. (10) Wu, H.; Chan, G.; Choi, J. W.; Ryu, I.; Yao, Y.; McDowell, M. T.; Lee, S. W.; Jackson, A.; Yang, Y.; Hu, L.; Cui, Y. Stable Cycling of Double-Walled Silicon Nanotube Battery Anodes through SolidElectrolyte Interphase Control. Nat. Nanotechnol. 2012, 7, 310−315. (11) Fowlkes, J. D.; Roberts, N. A.; Wu, Y.; Diez, J. A.; González, A. G.; Hartnett, C.; Mahady, K.; Afkhami, S.; Kondic, L.; Rack, P. D. Hierarchical Nanoparticle Ensembles Synthesized by Liquid Phase Directed Self-Assembly. Nano Lett. 2014, 14, 774−782. (12) Fowlkes, J. D.; Kondic, L.; Diez, J.; Wu, Y.; Rack, P. D. SelfAssembly versus Directed Assembly of Nanoparticles via Pulsed Laser Induced Dewetting of Patterned Metal Films. Nano Lett. 2011, 11, 2478−2485. (13) Wood, T.; Naffouti, M.; Berthelot, J.; David, T.; Claude, J. B.; Métayer, L.; Delobbe, A.; Favre, L.; Ronda, A.; Berbezier, I.; Bonod, N.; Abbarchi, M. All-Dielectric Color Filters Using SiGe-Based Mie Resonator Arrays. ACS Photonics 2017, 4, 873−883. (14) Ye, J.; Thompson, C. V. Templated Solid-State Dewetting to Controllably Produce Complex Patterns. Adv. Mater. 2011, 23, 1567− 1571. (15) Cui, Y.; Björk, M. T.; Liddle, J. A.; Sönnichsen, C.; Boussert, B.; Alivisatos, A. P. Integration of Colloidal Nanocrystals into Lithographically Patterned Devices. Nano Lett. 2004, 4, 1093−1098. (16) Flauraud, V.; Mastrangeli, M.; Bernasconi, G. D.; Butet, J.; Alexander, D. T. L.; Shahrabi, E.; Martin, O. J. F.; Brugger, J. Nanoscale Topographical Control of Capillary Assembly of Nanoparticles. Nat. Nanotechnol. 2016, 12, 73−80. (17) Zhu, X.; Vannahme, C.; Højlund-Nielsen, E.; Mortensen, N. A.; Kristensen, A. Plasmonic Colour Laser Printing. Nat. Nanotechnol. 2015, 11, 325−329. (18) Zhu, X.; Yan, W.; Levy, U.; Mortensen, N. A.; Kristensen, A. Resonant Laser Printing of Structural Colors on High-Index Dielectric Metasurfaces. Sci. Adv. 2017, 3, e1602487. (19) Chen, Y.; Duan, X.; Matuschek, M.; Zhou, Y.; Neubrech, F.; Duan, H.; Liu, N. Dynamic Color Displays Using Stepwise Cavity Resonators. Nano Lett. 2017, 17, 5555−5560. (20) Duan, X.; Kamin, S.; Liu, N. Dynamic Plasmonic Colour Display. Nat. Commun. 2017, 8, 14606. (21) Wang, X.; Kuchmizhak, A.; Storozhenko, D.; Makarov, S. V.; Juodkazis, S. Single-Step Laser Plasmonic Coloration of Metal Films. ACS Appl. Mater. Interfaces 2018, 10 (1), 1422. (22) Wuttig, M.; Yamada, N. Phase-Change Materials for Rewriteable Data Storage. Nat. Mater. 2007, 6, 824−832. (23) Gu, M.; Li, X. The Road to Multi-DimensionalBit-by-Bit Optical Data Storage. Opt. Photonics News 2010, 21, 28−33. J

DOI: 10.1021/acsnano.8b00198 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano (43) Kelley, J. D.; Hovis, F. E. A Thermal Detachment Mechanism for Particle Removal from Surfaces by Pulsed Laser Irradiation. Microelectron. Eng. 1993, 20, 159−170. (44) Voutsas, A. T.; Hatalis, M. K.; Boyce, J.; Chiang, A. Raman Spectroscopy of Amorphous and Microcrystalline Silicon Films Deposited by Low-Pressure Chemical Vapor Deposition. J. Appl. Phys. 1995, 78, 6999−7006. (45) Paillard, V.; Puech, P.; Laguna, M. A.; Carles, R.; Kohn, B.; Huisken, F. Improved One-Phonon Confinement Model for an Accurate Size Determination of Silicon Nanocrystals. J. Appl. Phys. 1999, 86, 1921−1924. (46) Ureña, F.; Raskin, J.; Olsen, S. H. Raman Measurements of Uniaxial Strain in Silicon Nanostructures. J. Appl. Phys. 2013, 114, 144507. (47) Li, B.; Yu, D.; Zhang, S.-L. Raman Spectral Study of Silicon Nanowires. Phys. Rev. B: Condens. Matter Mater. Phys. 1999, 59, 1645− 1648. (48) Im, J. S.; Gupta, V. V.; Crowder, M. A. On Determining the Relevance of Athermal Nucleation in Rapidly Quenched Liquids. Appl. Phys. Lett. 1998, 72, 662−664. (49) Ryu, S.-G.; Gruber, I.; Grigoropoulos, C. P.; Poulikakos, D.; Moon, S.-J. Large Area Crystallization of Amorphous Si with Overlapping High Repetition Rate Laser Pulses. Thin Solid Films 2012, 520, 6724−6729. (50) Shneidman, V. A. Interplay of Latent Heat and Time-dependent Nucleation Effects Following Pulsed-laser Melting of a Thin Silicon Film. J. Appl. Phys. 1996, 80, 803−811. (51) Stiffler, S. R.; Thompson, M. O.; Peercy, P. S. Transient Nucleation Following Pulsed-Laser Melting of Thin Silicon Films. Phys. Rev. B: Condens. Matter Mater. Phys. 1991, 43, 9851−9855. (52) Demichelis, F.; Minetti-Mezzetti, E.; Tagliaferro, A.; Tresso, E.; Rava, P.; Ravindra, N. M. Optical Properties of Hydrogenated Amorphous Silicon. J. Appl. Phys. 1986, 59, 611. (53) Kuznetsov, A. I.; Miroshnichenko, A. E.; Fu, Y. H.; Zhang, J.; Luk’yanchuk, B. Magnetic Light. Sci. Rep. 2012, 2, 492. (54) Evlyukhin, A. B.; Novikov, S. M.; Zywietz, U.; Eriksen, R. L.; Reinhardt, C.; Bozhevolnyi, S. I.; Chichkov, B. N. Demonstration of Magnetic Dipole Resonances of Dielectric Nanospheres in the Visible Region. Nano Lett. 2012, 12, 3749−3755. (55) Wu, Y.; Dong, N.; Fu, S.; Fowlkes, J. D.; Kondic, L.; Vincenti, M. A.; De Ceglia, D.; Rack, P. D. Directed Liquid Phase Assembly of Highly Ordered Metallic Nanoparticle Arrays. ACS Appl. Mater. Interfaces 2014, 6, 5835−5843. (56) Naffouti, M.; Backofen, R.; Salvalaglio, M.; Bottein, T.; Lodari, M.; Voigt, A.; David, T.; Benkouider, A.; Fraj, I.; Favre, L.; Ronda, A.; Berbezier, I.; Grosso, D.; Abbarchi, M.; Bollani, M. Complex Dewetting Scenarios of Ultrathin Silicon Films for Large-Scale Nanoarchitectures. Sci. Adv. 2017, 3, eaao1472. (57) Kojima, Y.; Kato, T. Nanoparticle Formation in Au Thin Films by Electron-Beam-Induced Dewetting. Nanotechnology 2008, 19, 255605. (58) Pan, H.; Grigoropoulos, C. Crystallization in Nano-Confinement Seeded by a nanocrystalA Molecular Dynamics Study. J. Appl. Phys. 2014, 115, 104307. (59) Pan, H.; Shou, W. Single Crystal Formation in Micro/nanoConfined Domains by Melt-Mediated Crystallization without Seeds. J. Phys. D: Appl. Phys. 2015, 48, 225302. (60) Hatano, M.; Moon, S.; Lee, M.; Suzuki, K.; Grigoropoulos, C. P. Excimer Laser-Induced Temperature Field in Melting and Resolidification of Silicon Thin Films. J. Appl. Phys. 2000, 87, 36. (61) Lumerical Solutions, Inc., https://www.lumerical.com/tcadproducts/fdtd/.

K

DOI: 10.1021/acsnano.8b00198 ACS Nano XXXX, XXX, XXX−XXX