Self-Assembly of Spherical Colloidal Photonic ... - ACS Publications

Article pubs.acs.org/crystal

Self-Assembly of Spherical Colloidal Photonic Crystals inside InkjetPrinted Droplets Enrico Sowade,*,†,⊥ Thomas Blaudeck,*,†,‡,⊥ and Reinhard R. Baumann*,†,§ †

Technische Universität Chemnitz, Digital Printing and Imaging Technology, 09126 Chemnitz, Germany Technische Universität Chemnitz, Center for Microtechnologies, 09107 Chemnitz, Germany § Fraunhofer Institute for Electronic Nanosystems, Department Printed Functionalities, Technologie-Campus 3, 09126 Chemnitz, Germany ‡

S Supporting Information *

ABSTRACT: The manufacturing of three-dimensional colloidal structures on solid substrates is an important topic of applied research, aiming for photonic components especially in photovoltaic and sensor applications. Whereas conventional techniques such as wet self-assembly are based on engineering of the substrate surface energy, alternative strategies envisage the independence of the interfacial conditions. We report on inkjet printing of colloidal suspensions of monodisperse silica or polystyrene nanoparticles or both and their self-assembly to spherical colloidal photonic crystals. The formation process of the colloidal nanoparticles into stable spherical colloidal assemblies (SCAs) is achieved by a self-assembly process inside tiny droplets of a stochastic mist generated intentionally instead of a jet of individual single droplets using inkjet printing. The mistjetted, shrinking droplets serve as confined geometries for the solidification of the nanoparticles during the evaporation; thus the particles are packed into stable ball-shaped assemblies. We show how fine-tuning of the jetting parameters allows the reliable generation and deposition of three-dimensional (3D) spherical colloidal assemblies of nanoparticles variable in size and with a high packing order. Microreflectance spectroscopy proves that the degree of order in the SCA is such that photonic stop bands occur inherent for photonic crystals.

■

INTRODUCTION

blies based on bottom-up approaches utilizing self-assembly are a current research path.1,22 According to Rastogi et al.,22 the techniques for preparing spherical colloidal assemblies (SCAs) from suspensions containing nanospheres can be classified into wet self-assembly (WSA) and dry self-assembly (DSA) methods. A similar classification for bottom-up self-assembly processes in general is done by Maenosono et al.31 The major drawback of the WSA methods is usually the long process time that has been addressed by several researchers.19,32 Although very fast consolidation times of colloidal particles inside droplet templates were reported recently by Gu et al.,33 the consolidation in WSA requires usually a few tens of minutes up to several hours.34 On the other hand, the manufacturing of SCAs with DSA methods is mostly very fast and flexible because there is no need for ulterior process steps such as demulsification. WSA methods require an expensive experimental setup, for example, emulsion-assisted technologies and sometimes additionally ultrasonic or microwave support to optimize and accelerate the particle consolidation. With respect to application in industry, DSA approaches are considered

Colloidal photonic crystals (PCs) received increasing attention during the past decade not only on account of the fundamental science of their formation and their artistic beauty but also with envisaging applications as waveguides and light manipulators in optoelectronic, photovoltaic, and sensoric applications.1−5 Colloidal PCs are artificially structured objects consisting of periodically arranged nanoparticles resulting in repeating regions of low and high dielectric constants. This structure allows a modulation of electromagnetic waves, which can be used, for example, to control the propagation of light. Within this topic, colloidal assemblies with a spherical macroscopic shape are a comparably new class of photonic materials. These spherical assemblies are called supraparticles,6−8 supraballs,2,6,9−11 colloidal (micro) clusters,6,7,10,12−14 spherical nanoparticle aggregates,15 microbeads or crystal beads,1,7,16−19 spherical colloidal (photonic) crystals,1,9,19−21 opal balls,22,23 or photonic balls.9,19,24−29 Recently, the manufacturing even of superstructures from supracolloids (e.g., crystal-like solids made from supraballs) and its application potential was reported.30 Spherical colloidal PCs combine a large number of highly ordered nanoparticles invoking unique photonic properties with a confinement of these to extensions of a few micrometers. Different manufacturing strategies for these spherical assem© XXXX American Chemical Society

Received: November 5, 2015 Revised: December 30, 2015

A

DOI: 10.1021/acs.cgd.5b01567 Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design

Article

Table 1. Characteristics of the Colloidal Suspensions Used for the Experiments manufacturer

Polysciences

BS-Partikel

Bangs Laboratories

custom-made binary mixture

abbreviation nanosphere material nanosphere diameter [nm] solids content [wt %] surface tension [mN/m] electrophoretic mobility [(μm/s)/(V/cm)] ζ potential (mV) pH value solids content used for printing [wt %]

PSC221 polystyrene (PS) 221 ± 17 2.6 44.1 ± 2.2 −2.4 ± 0.9 −34 ± 13 7.0 ± 0.2 2.6

BS305 polystyrene (PS) 305 ± 8 2 46.8 ± 0.8 −4.0 ± 1.2 −51 ± 15 7.0 ± 0.2 2

BL280 silica (SiO2) 280 ± 7 10 60.8 ± 1.3 −3.1 ± 1.5 −39 ± 19 7.0 ± 0.2 dilution: 4, 2, and 1

BS305 and BL280 polystyrene (PS) and silica (SiO2) 305 ± 8 (PS), 280 ± 7 (SiO2) 2 56.2 ± 1.2 −3.7 ± 1.1 −47 ± 13 7.0 ± 0.2 2

more promising.22 In these methods, the colloidal dispersions are dispensed on solid surfaces by using an evaporating droplet as a template. However, although there has been a significant progress due to the usage of microfluidics for the generation of such droplet templates, there is still a need for simple massproduction methods enabling bulk generation of the SCAs.1 In this work, we demonstrate the applicability of inkjet printing technology as a promising manufacturing method for the SCAs. As shown in our earlier work,15 the tuning of the control voltage signal of the inkjet printheads (known as “waveform”) allows a relocation of the evaporation-induced self-assembly process from the sessile droplet on a solid surface to the droplet-in-flight. As a consequence, the substrate does not affect the final morphology of the deposit anymore.15 In detail, the shape of the final deposit is prescribed by the shape of the equilibrated droplet-in-flight given by a spherical geometry (ball-shaped) due to the high surface tension of liquids. Thus, the tiny droplets ejected by inkjet printing act as a confined geometry for the nanosphere particles inside. Due to the solvent evaporation, the particles are self-organized and consolidate during the flight to SCAs. In 2009, Cho et al.16 exploited inkjet printing technology to eject toluene droplets containing polystyrene (PS) particles aiming to form SCAs. Inkjet printing was employed to drip the drops into an aqueous solution of emulsions stabilizer. This was followed by a demulsification process applying 100 °C for 1 h in order to self-assemble and consolidate the particles inside the toluene template. Finally, the aqueous solution containing the SCA was evaporated at room temperature to obtain spherical assemblies in a dry environment. Zhao et al. reported similar procedures using microfluidic devices as a preparation method for SCAs.1 Rastogi et al.22 and Marin et al.35 utilized an ultramicropipet to dispense colloidal suspension on superhydrophobic surfaces. Anselmann et al.36,37 applied a spray device in combination with a suitable, presumably hydrophobic surface for developing SCAs. Aiming for patterned photonic structures, Belgardt et al. proposed a combined approach comprising surface engineering by inkjet printing38 or laser writing,39 followed by a liquid deposition of the colloidal dispersion of nanospheres on the prestructured surfaces.39 Beside these approaches, a variety of more sophisticated technologies are available based on drop generators (ultrasonic nebulizer, electrospray sources), carrier gases, furnace, and particle-collecting parts to obtain colloidal microclusters.4,13,40 In contrast to other methods reported in literature,7,16,22 we could show that the SCAs can be manufactured by an in-flight inkjet printing process without employing a sophisticated emulsification process or sophisticated pretreatment of the substrate on a large area and with high throughput.15 In this work, inkjet printing turned out as a highly productive

technique for the patterned deposition of SCAs. Within a few seconds, several thousand SCAs were manufactured. The SCAs are developed independent of the surface energy of the substrate, in a dry environment, and in the absence of any heating devices. There is no need for a special pretreatment of surfaces to obtain superhydrophobicity or expensive downstream process steps such as demulsification or liquid evaporation by high temperature heating, which makes it promising for industrial applications.

■

MATERIALS AND METHODS

Inks and Substrates. Different commercially available colloidal suspensions were used as a base for inkjet inks. The commercial master suspensions contain (i) highly monodisperse organic PS nanosphere particles with anionic, hydrophobic surfaces or (ii) inorganic silica (SiO2) nanospheres with nonfunctionalized polar hydroxyl surface groups (Si−OH) suspended in an aqueous environment. The colloidal inks were obtained from Polysciences (Warrington, PA, USA; abbreviated PSC221), BS-Partikel GmbH (Wiesbaden, Germany; BS305), and Bangs Laboratories (Fishers, IN, USA; BL280). The colloidal ink BL280 containing 10 wt % SiO2 nanospheres was diluted with deionized water (16 MΩ·cm) to 4, 2, and 1 wt % solids content to allow a stable drop ejection via inkjet printing. The suspensions with polymer nanospheres were used as received. A binary suspension in which SiO2 and PS nanospheres were dispersed was prepared by mixing identical quantities of the 2 wt % BS305 PS suspension with the 2 wt % BL280 SiO2 suspension (56.2 ± 1.2 mN/m surface tension of the binary ink) under ultrasonic treatment. The suspension PSC221 was chosen due to its much lower diameter of the nanospheres compared with the ones of the other manufacturers. A summary of the used colloidal suspensions and their characteristics is provided in Table 1. The surface tension of the suspensions was determined with a DataPhysics OCA20 (DataPhysics Instruments GmbH, Filderstadt, Germany) system in pendant-drop mode. The dispersion stability of the suspensions BS305, BL280, and the binary mixture was investigated by determining the ζ potential (ζ) and the electrophoretic mobility (μe) of the particles. The investigations were carried out magnitudinal after dilution in 0.0001 M KCl(aq) with a Zetaview system from Particle Metrix (Particle Metrix GmbH, Meerbusch, Germany). Table 1 shows the measurement data. The relatively high negative charge of the particles in the different suspensions results in an effective repulsion and thus in a good dispersion stability qualified for inkjet printing. The high charging ensures that no particle agglomeration will take place in the ink. Because both kinds of particles, SiO2 and PS, exhibit a negative charge, the formulation of a stable binary suspension was possible. The ζ potential value of the binary suspension indicates nearly equal proportions of SiO2 and PS particles. All the inks prepared were treated ultrasonically for about 3−6 min before printing. Coverslip glasses (18 × 18 mm2 and 20 × 20 mm2; thickness 0.145 ± 0.015 mm, purchased from VWR Scientific, Dresden, Germany) served as substrates in the printing process. B



Article

The colloidal suspensions were printed using a Dimatix DMP 2831 laboratory drop-on-demand (DoD) inkjet printer (Fujifilm Dimatix Inc., Santa Clara, USA). Several piezoelectric inkjet printheads were employed with a nozzle diameter of 21.5 μm and a nominal drop volume of 10 pL. The DMP was applied in both single nozzle and multinozzle mode. The clear distance between the nozzle and the substrate was maintained at 1 mm during printing. All samples were printed at ambient conditions (laboratory conditions: 22.5 ± 0.8 °C and 22% ± 3% relative humidity). The DMP has a built-in stroboscopic drop watcher to allow for a determination and optimization of the droplet formation by adjusting the control signal applied to the lead zirconate titanate (PZT) piezoelectric transducer. When a voltage is applied to the piezoelectric transducer, a deformation takes place creating mechanical vibrations. These vibrations result in acoustic waves that force the liquid in the printhead out of the chamber through the nozzle. This repetitive control signal for the piezoelectric transducer is called waveform and is defined as voltage over a certain time. The design of a waveform with the used DMP 2831 is fully flexible due to the user-controlled variation of pulse shape, voltage, and frequency. Three waveforms were used for the deposition of the inks. They are shown in Figure 1 and denoted as

The SCAs were analyzed by scanning electron microscopy (SEM) using a Hitachi TM-1000 (Hitachi High-Technologies Cooperation, Tokyo, Japan), a Nova NanoSEM 200 (FEI, Hillsboro, USA), and a Zeiss Auriga 60 (Zeiss AG, Jena, Germany). To avoid the charging effect on the insulating nanospheres, the samples were coated with a Pt layer (about 18 nm thick) by sputtering at 40 mA for 120 s using a BAL-TEC SCD 050 (formerly BAL-TEC AG, Balzers, Liechtenstein) electron microscope preparation system. Optical microscopy analysis was carried out on a Leica DM 4000 M (Leica Microsystems CMS GmbH, Wetzlar, Germany) and a Zeiss Axio Imager M2m (Carl Zeiss Microscopy GmbH, Jena, Germany). Microreflectance spectroscopy was performed using the mentioned Zeiss microscope coupled with a Tidas S MSP 800 spectrometer (J&M Analytic AG, Esslingen, Germany). The light source of the microscope is exploited to irradiate the sample through the microscope objective lenses. The reflected light is collected by the same lenses and guided to the spectrometer by optic fibers. At the same time, the microscope is coupled to a CCD camera allowing one to precisely define the measurement position and to capture images. The measurement area for the spectroscopy can be adjusted with an aperture. A wavelength range of 200−980 nm is detectable. The angle of incidence was kept constant for all the measurements at 90°.

■

RESULTS AND DISCUSSION Influence of Printhead Control Signal (Waveform) on Deposit Morphology. In piezo inkjet printing, the waveform is one of the most important parameters to control the characteristics of the drop ejection and hence the shape of the material deposits.41 Usually, the waveform is optimized to enable the ejection of a single spherical, regular droplet on a trajectory perpendicular to the nozzle plate. Commonly, each pulse of the firing signal generates one defined drop in order to deposit it on a predetermined position of the substrate (exceptional case, gray scale inkjet printing42). Exactly this kind of droplet ejection is the case for waveform A: A single ball-shaped droplet is ejected on a trajectory perpendicular to the nozzle plate as shown in the scheme in Figure 2C (upper image) and the captured image of the drop ejection via drop watcher camera in Figure 2B (upper image) when the waveform A was applied with the different colloidal ink formulations (here, BS305 was used). The final deposit on the glass substrate is shown in the upper image of Figure 2D. The nanoparticles form a ring-like deposit as a result of the interaction between surface energy of the substrate and surface

Figure 1. Waveforms used to control the piezoelectric transducer of the inkjet printheads for the deposition of the colloidal suspensions. A, B and C. The main difference between the waveforms is the time duration of the voltage signal applied to the piezo-electric printhead and the duration of the maximum voltage plateau indicating the highest deformation of the piezoelectric actuators.

Figure 2. Influence of the waveform applied to the piezo inkjet printhead on the drop ejection and the resulting final deposits: (A) two classes of waveforms (both driven with a frequency of 3 kHz); (B) drop ejection captured with the DMP 2831 built-in drop watcher camera corresponding to the two waveforms; (C) scheme of the basic principle of the drop ejection based on waveforms A and B (waveform C shows similar drop ejection as waveform B); (D) SEM images showing the different deposit morphologies as a result of waveforms A and B. C



Article

Figure 3. Reliability of the developed method for the manufacturing of SCAs represented by (A) manufacturing yield and (B) number of manufactured spherical assemblies per area as a function of applied waveform and solids content of the ink formulation. The print pattern was defined as rectangle with an area of 0.35 mm2 at 5080 dpi.

stable ball-shaped droplet ejection.53−55 So far, the exploitation of usually unintended satellite drops or a mist of droplets in inkjet printing has not yet been reported. As a consequence, we studied the reliability of the developed process for the manufacturing of the SCAs. The generation of a mist of droplets instead of a ball-shaped single droplet with a trajectory orthogonally to the nozzle plate is presented in Figure 2B,C. Waveform B was designed with a slightly higher voltage but shorter pulse duration and waveform C with a lower voltage but a much longer pulse duration compared with waveform A (see Figure 1 for a comparison of all waveforms). The influence of pulse duration and applied voltage on the drop ejection are studied in detail and well explained by Wijshoff42 and Dong et al.53 In the case of waveforms A and B, no primary droplet can be identified using the built-in stroboscopic drop watcher camera of the DMP2831, but a large number of small droplets with similar size−all ejected by a single firing pulse. Phenomena such as satellite droplet formation in inkjet printing have been extensively investigated by Wijshoff.42 Wijshoff classifies satellite droplet formation into (i) mist of droplets, (ii) Rayleigh breakup, (iii) fast satellites, and (iv) slow satellites. However, our droplet formation process does not exactly match any of these types of satellites because Wijshoff considers satellite drops as additional droplets in close vicinity to a main droplet. Obviously, the image of the drop watcher camera in Figure 2B does not show a clearly distinguishable main droplet. A droplet mist characterized by a higher material concentration at the front of the droplet mist swath than at its back can be observed (see Figure 2B,C). Also Song et al.56 studied different phenomena in inkjet printing with focus on drop ejection. They describe one of the observed effects as ink spray that is very close to the effect that we can see in our case when applying waveform B or C. Song et al.56 assume partially clogged nozzles by (i) an emerging dried deposit buildup of ink adhering to the nozzle orifice or a partially clogged nozzle by (ii) particle agglomerations and larger particles as reasons for the obtained spray or mist of droplets in inkjet printing. However, partially clogged nozzles might be a reason for a mist of droplets ejected by inkjet printing that takes place suddenly, undesirably, and also uncontrollably from a formerly regular jetting and not clogged nozzle. But one can also induce a mist of droplets

tension of the ink formulation. Therefore, the properties of the substrate are a major influencing factor for the final structural architecture of the deposit.43−45 The reason for the ring-like deposits are capillary flows that transport the nanospheres from the center of the deposit toward the edge where they agglomerate. This phenomenon, well-known as coffee-ring effect, has been studied by many researchers.46−49 A completely different deposit morphology was obtained by application of waveform B as shown in Figure 2D (bottom image). While waveform A results in planar deposits with a morphology dependence on substrate properties,45 waveform B leads to 3D deposits with spherical morphology using the same substrate. This novel method for the manufacturing of SCAs is based on the DSA approach using inkjet printing for the generation of droplet templates. The nanoparticles are confined inside the inkjet-printed droplet. The ball shape of the ejected droplets caused by the high surface tension of the colloidal dispersions serves as a geometrical constraint for the resulting spherical assembly of the nanoparticles. Our method is a contrasting approach to the reported manufacturing of domeshaped colloidal assemblies using inkjet printing25,43,50−52 or SCAs using an ultramicropipet and superhydrophobic substrate surfaces.22 In these cases, the self-assembly process takes place on the substrate and the shape of the droplet template is a result of the surface energy of the substrate. Only a high contrast between surface energy of the substrate and surface tension of the droplet liquids will lead to dome shapes or spherical assemblies. In our case, the most important parameter for the development of the SCAs is the pulse shape of the driving voltage signal (waveform) applied to the piezoelectric transducer of the inkjet printhead. The waveform was configured contrary to the usual goal in graphical inkjet printing to generate a satellite-free single droplet with a controlled trajectory and shape. Satellite drops or droplet mists are generally not intended in graphical and functional inkjet printing due to droplet placement accuracy and thus quality aspects in printed patterns. Therefore, most of previous research efforts about drop formation in inkjet technologies were aimed at avoiding satellite drops and droplet mists and improving the drop generation process toward a single and D



Article

Figure 4. Tilted view by SEM of SCAs based on BS305 with changing number of constituent particles: (A) N ≈ 140, (B) N ≈ 1900, and (C) N ≈ 31800; (D, E) variable sizes of the assemblies obtained in a line arrangement as well as randomly distributed; (F) line-up of particles with similar size.

intentionally using a fine-tuned waveform in a controlled manner as we demonstrate with the designed waveforms B and C. Reliability of the Manufacturing Process of SCAs Based on Inkjet Printing. Colloidal inks based on BL280 were used with varying solids contents (1, 2, and 4 wt %) to study the influence of the solids content on the inkjet process and the development of SCAs. We obtained SCAs with all solids contents when using waveform B or C. But no SCAs were obtained with waveform A. We found that 4 wt % particle content represents an upper limit for the inkjet printing process. The relatively high concentration in conjunction with the size of the nanoparticles results in limited jetting performance. The ejection of the droplet mist was interrupted frequently due to clogged nozzles so that a cleaning was required. Therefore, 4 wt % solid content was not considered for the detailed reliability study. Figure 3 shows the reliability of the manufacturing process represented by the process yield of the SCAs as a function of the applied waveform and solids content of the ink formulation. The yield was determined by printing 10 rectangular areas of 0.35 mm2 with a print resolution of 5080 dots-per-inch (dpi) per ink formulation and per waveform on cleaned glass substrates and counting the number of samples on which SCAs were obtained. The number of deposited SCAs was determined using light microscopy and calculating the average of the counted assemblies per ink formulation and applied waveform. The higher the solids content was, the higher the yield of the manufacturing process and also the higher the number of SCAs. In total, 10 rectangles were printed with waveform A. However, no SCAs were found in any of the 10 rectangular areas. As a consequence, the manufacturing yield is 0%. All the deposited droplets have a morphology similar to a flat disc with a pronounced coffee-ring effect as shown in the upper image of Figure 2D. As show in Figure 3A, SCAs were achieved with a similar yield for waveforms B and C. The yield depends obviously on the solids content since 1 wt % results in about 70% manufacturing yield and 2 wt % in about 90% manufacturing yield. The number of manufactured SCAs per 0.35 mm2 was about five times higher using the 2 wt % solids content ink

formulation compared with the 1 wt % ink. Rectangles deposited with waveform B show a higher number of SCAs but also a much higher standard deviation than those with waveform C. The number of assemblies for waveform C was about 510 ± 60 for an area of 0.35 mm2 with a print resolution of 5080 dpi. This implies a stable and reproducible manufacturing process. We could not see any influence of the applied waveform on the size of the SCAs. But the solids content has a very strong influence on the size. The average diameter of the assemblies with 1 wt % solids content was found to be about 5.0 ± 0.3 μm and the average diameter with 2 wt % was about 8.2 ± 0.6 μm. Thus, a higher solids content results in larger SCAs since the number constituents per ejected droplet increases with the solids content. Classification of the Inkjet-Printed SCAs. Referring to literature,7,16 SCAs can be divided into three categories depending on the number of constituent particles: (i) highorder clusters (N < 100), (ii) supraparticles (N > 100), and (iii) supraballs (N → ∞). Figure 4 shows the size variety of the SCAs manufactured by inkjet printing using the colloidal suspension BS305. They have a very regular hexagonal ordering of the nanosphere particles at the surface comparable to the surface of SCAs reported in the literature4,7,9,10,16,24,35,37,57−64 based on other manufacturing methods. However, there are also some point as well as line defects such as grain boundary scars20,65 visible at the surface of the spheres. They are intrinsic for spherical shapes due to the surface curvature.1,66 We could not derive any influence of printing parameters on the formation of the mentioned defects. A reduction of defects might be obtained by two approaches: (i) usage of smaller nanosphere particles and their assembly to larger SCAs and (ii) the application of specially designed nanosphere particles with soft shells facilitating a crystal structure formation with low defect density in a very short time duration as reported by McGrath et al.67 and Cui et al.68 As shown in Figure 4, some of the SCAs are surrounded by planar assemblies of microspheres. This indicates, that usually both types, planar as well as 3D assemblies, exist in parallel due to the manufacturing process based on inkjet mist. E



Article

The number of constituent particles (N) was calculated with eq 1 assuming a hexagonal or face-centered cubic (fcc) order of particles with a packing density of about 74% (given by π ).16 3 2

The fcc order is considered as the most energetically stable structure for colloidal crystals.69 Thus, we can assume that the SCAs exhibit the fcc structure or a mixture of fcc and hexagonal order. The quantity dS represents the diameter of the SCA, and dP represents the nanoparticle diameter. ⎛ dS ⎞3 N = ⎜ ⎟ 0.74 ⎝ dP ⎠

(1)

According to the results shown in Figure 4, a broad size variety of SCAs can be manufactured depending on (i) the volume of the ejected droplets in the droplet mist and (ii) the number of particles contained in a single ejected droplet. The droplet volume is correlated with the waveform and frequency applied to the printhead, while the number of particles is mainly a function of ink formulation as explained before. However, SCAs with changing number of constituents also appear sometimes without any alterations of the waveform signal or ink formulation based on a single driving pulse as depicted in Figure 4D. The droplet mist can contain droplets of different volume depending on the drop formation dynamics. Reasons for the different sizes could be also build-up of dried deposits of nanoparticles adhering to the nozzle tip or particle agglomerations that temporarily clog the nozzle as observed by Song et al.56 Although similar SCAs have been demonstrated using other preparation methods, the development of monodisperse, microscale, small, and uniform SCAs as depicted in Figure 4F is still a challenge.24,64 The highly monodisperse SCAs can be obtained by the repetition of the voltage signal to the piezoelectric transducer resulting is the formation of similar droplet mists with droplets having similar volumes and thus a similar number of constituent particles (see Supporting Information, Figure S1). Other colloidal suspensions, e.g. BL280, were tested as well and show very similar behavior concerning the formation of SCAs. Also suspensions with smaller particle sizes using PSC221 were printed, and SCAs of similar shape were obtained (see Supporting Information, Figure S2). Preparation of Inkjet-Printed Binary SCAs. Aside from manufacturing of SCAs consisting of only one specific nanoparticle material, one can also manufacture SCAs based on two nanoparticle materials. A mixture of BS305 and BL280 was formulated to a binary suspension with 2 wt % solids content and inkjet-printed to develop binary SCAs composed of organic and inorganic nanoparticles. Waveform A was applied first, aiming to obtain planar deposits of the ink formulation. The samples were analyzed by SEM using the BSE (backscattered electron) mode allowing the visual indication of material properties since materials with a high atomic number will appear brighter than materials with low atomic number. PS particles have a lower density compared with the SiO2 particles. A section of the planar deposit imaged by SEM in BSE mode is shown in Figure 5A. Dark and bright particles can be easily distinguished. Energy-dispersive X-ray (EDX) analysis was performed on the same position to confirm that the grayscale (dark and bright) difference is a result of the nature of the material. Figure 5B shows the result. The blue color refers to SiO2 particles and red color to PS particles. The two kinds of particles are randomly ordered without any preferential

Figure 5. (A) Deposits of inkjet-printed binary ink formulation consisting of PS (BS305) and SiO2 (BL280) imaged by SEM in BSE mode based on waveform A and (B) EDX mapping of the same image position; (C, D) top views of a SCA obtained with the binary ink formulation by applying waveform B. Panel C shows the original SEM image in BSE mode, and panel D shows a contrast-intensified version by image processing of panel C to indicate the distribution of PS and SiO2 nanospheres.

alignment. Waveforms B and C were employed to manufacture SCAs using the binary ink formulation. Figure 5C depicts a topview image of a spherical assembly with a diameter of about 6 μm obtained by SEM in BSE mode. It shows a close-packed spherical assembly containing SiO2 and PS particles with some point and line defects but of similar quality as with only PS or SiO2 particles (compare with Figure 4). Figure 5D is a contrastintensified version of Figure 5C by image processing indicating the distribution of the particles on the surface of the SCA. There is also no preferential order visible. EDX analysis of the SCA was performed but did not give results since the assemblies were displaced a few tens of micrometers due to the intense electron beam. Preparation of Skeletons of the Inkjet-Printed SCAs. Some samples of the binary SCAs were post-treated after printing by heating in a furnace at 300 °C for 5 h (Thermo Scientific VT 6060). A depolymerization process of the PS is initiated resulting in styrene and finally thermal decomposition of the organic materials. Afterward, the samples were sputtered with Pt and characterized by SEM to analyze the particle arrangement in detail. A scheme of the preparation concept is displayed in Figure 6A. The images in Figure 6B−D show the skeleton of the remaining SiO2 particles after removal of the PS particles from the SCA. The sizes of the depicted SCAs correspond to the classification of higher-order clusters, supraparticles, and supraballs. The stability of the assemblies is remarkably, although the PS particles are removed and only the skeleton of SiO2 particles is remaining. This observation implies a well-mixed architectural structure of the microclusters. For binary clusters of roughly equimolar concentrations, upon removal of the PS particles, the porosity increased dramatically, which could qualify these structures for applications such as drug delivery. Mechanical Stability and Internal Structure of the Inkjet-Printed SCAs. The adhesion of the SCAs to the substrate was tested by two abrasive micromechanical methods. Rinsing with DI (moderate rinsing with standard narrow neck dash bottle) and purging with compressed air (1 bar) were used to invoke a detachment of the SCAs from the substrates. Both F



Article

Figure 6. (A) Scheme of the selective removal of PS particles from the binary SCAs by heating in a furnace; (B−D) spherical skeletons of SiO2 particles obtained by the removal of PS from the inkjet-printed SCAs with different numbers of constituents representing (B) higher-order clusters, (C) supraparticles, and (D) supraballs.

Figure 7. SCA based on on BS305: (A) broken SCA upon mechanical force from the top; (B) internal structure of broken SCA with a well-filled fcc structure; (C) SCA that was cut by FIB aiming to reveal the (D) cross-sectional area of the resulting hemisphere by the FIB cut; (E) a magnified image of the center area of panel D.

such as calcination8,17,73 were carried out in this study. The high stability of the SCAs is interesting for biomedical applications such as drug delivery or biomolecular screening and detection.8,70,73 Some of the SCAs were cut to obtain a cross-sectional image of the resulting hemisphere. The cutting was performed using a focused ion beam (FIB) of a gallium source within an evacuated SEM chamber. Pt was deposited locally by means of electron beam and assisted by ion beam as protection layer. The angle of the SCAs of nanoparticles to the electron and ion beam source was varied during the Pt deposition to ensure proper embedding of the spheres. One of the spherical assemblies before the FIB process was carried out is depicted in Figure 7C. The diameter of the SCA was determined to approximately 10 μm. The FIB cut was performed in a plane close to the center of the spherical

methods failed, which can be explained by the fact that the SCAs are sorbed to the glass substrate via strong van der Waals forces.1,70 The internal stability of the SCAs was probed by a microforce apparatus consisting of two glass coverslips, one of which contained the printed SCAs on the inside. These plates were supplied with a manual force directed exactly vertical with the intention to break up the SCAs. It was demonstrated that the clamped SCAs were deformed or broke in a few big parts as shown in Figure 7A,B but were not completely destroyed by disordering all constituent particles. These observations assume strong van der Waals interaction forces among the particles and among the particles and the substrate, potentially enhanced by residual traces of the water in the menisci between adjacent particles as well as particles and the substrates.71,72 No further processes aiming for an increase of the mechanical strength G



Article

Figure 8. (A) Microreflectance spectra of SCA based on BL280 and PSC221; light microscopy images of the SCA of BL280 are shown using (B) bright field microscopy and (C) dark field microscopy. Panel D is a dark field microscopy image indicating the difference in appearance between SCA and planar particle assemblies with coffee-ring shape.

assembly. The cross section of the resulting hemisphere with the buildup of the protection layer at its outside is shown in Figure 7D. Figure 7E is a zoom-in of Figure 7D. The surface texture at the cut plane is depicted well, but the inner structure is indicated only blurry due to charging effects. It can be seen that the FIB cuts the nanosphere particles at different positions so that cross sections of the nanosphere particles of different area appear. Indeed, it is difficult to derive any order from the cross-sectional images due to the nature of the FIB method. However, in combination with the cracked SCA of Figure 7A,B it can be concluded that all the manufactured assemblies were completely filled without any obvious domain formation and that the nanospheres are most probably arranged in a mixture of close-packed fcc and hexagonal structure (see Figure 7B). Considering the convective assembly hypothesis of Norris et al.69 (although the hypothesis is based on planar opal growth), we assume that the fcc order outweigh the hexagonal ordering in the manufactured SCAs − at least for the inner part of the SCAs. The theoretical interplanar spacing D111 for the (111) planes in the fcc lattice (representing the lattice constant) of the SCAs can be calculated with following equation: D111 =

2 dP 3

bright and shiny due to the diffraction of light. Thus, one can distinguish easily between planar layers and SCAs. The spectra of the different SCAs in Figure 8A show specular reflectivity peaks that can be assigned to photonic stop bands. These peaks correspond to the Bragg reflection resulting mainly from the (111) planes of the SCAs. The SCAs printed with BL280 and PSC221 have one characteristic stop band in the considered wavelength regimes at 618 and 526 nm, respectively. The wavelength gap or relative stop bandwidth, Δλ/λ, with λ corresponding to the maximum peak wavelength and Δλ corresponding to the full-width at the half-maximum of the peak wavelength, was calculated.74 The SCAs with PSC221 and BL280 have a relative stop bandwidth of about 30% and 20%, respectively. These comparable large relative band widths confirm the interaction of the SCAs with light, and they can be considered as indicator for the well-ordered structure of the particles within the SCAs. The position of the photonic band gap can be explained in a more quantitative way. A theoretical approximation for the wavelength λ of the reflection peak of a photonic crystal can be made for light incidence normal to the (111) planes based on Braggs law using the equation2,3,75,76

(2)

Thus, we expect an interplanar spacing for (111) planes of the SCAs in Figure 7 of about 249 nm. Optical Properties of the Inkjet-Printed SCAs. Microreflectance spectroscopy was performed to investigate the optical properties of the SCAs. The ordered packing of the particles within the assemblies as determined by SEM imaging and forced breakup (see Figure 7) results in periodic refractive index variation indicating photonic properties. Figure 8A shows the UV−vis reflectance spectra of a SCA based on BL280 (SiO2 nanoparticles, diameter 280 nm) and PSC221 (PS nanoparticles, diameter 221 nm). Figure 8B,C shows light microscopy images of the SCAs of BL280. Figure 8C,D was obtained with incident light in dark field mode. In bright-field mode, the SCAs are difficult to identify as shown in Figure 8B, but the bright red-colored circular deposits indicate the position of the assemblies. On the contrary, the dark-field mode allows a clear visualization of the SCAs. Figure 8D is an image depicting both SCAs and planar layer deposits with coffee-ring shape with one or only a few nanoparticle layers stacked. The SCAs appear

λ=

⎛ 8 ⎞1/2 ⎜ ⎟ dP(nP 2 0.74 + nA 2 0.26)1/2 ⎝3⎠

(3)

where nP is the refractive index of the nanospheres, nA is the refractive index of the voids (i.e., ambient air) between the nanospheres, 0.74 and 0.26 are the assumed volume fraction values for the nanospheres and the voids, and dP is the diameter of the nanospheres. Usually, with increasing constituent nanoparticle size, the position of the reflectance peak shifts to longer wavelengths implying that the photonic band gap shifts to lower energies. In our case, this fact comes clearly true for SCAs of PSC221 with nanoparticles of 221 nm and BL280 with nanoparticles of 280 nm in diameter. The measured peak positions are 618 and 526 nm as indicated in Figure 8A, respectively. According to eq 3, the calculated reflection peaks for SCAs of BL280 and PSC221 are 614 and 526 nm, respectively. Thus, the calculated reflection peaks are in perfect agreement with the measurement data. H



■

Article

edged for providing the measurements of the ζ potential and electrophoretic mobility. Sabine Riemenschneider and Tobias Seifert are acknowledged for support in inkjet printing the colloidal suspensions and evaluation of the deposition of the planar layers. Rebecca Wagner and Frank Cichos (University of Leipzig, Molecular Nanophotonics) as well as Matthew Jorgensen (Leibniz IFW, Dresden, Rolled-up Photonics) are acknowledged for preliminary angle-resolved optical measurements of the samples and for fruitful discussion. The silica nanospheres (Bangs Laboratories) were thankfully provided by Jolke Perelaer and Ulrich S. Schubert (FSU Jena, Institute for Organic and Macromolecular Chemistry). The authors acknowledge partial financial support from the EU-FP7 NoE PolyNet (Grant Agreement 214006) and ICT TDK4PE (Grant Agreement 287682).

SUMMARY AND CONCLUSION We demonstrated a simple and unique approach for the manufacturing of SCAs based on inkjet printing. The control signal for the piezoelectric printhead was found to be the most important parameter influencing the morphology of the deposits. Using a sophisticated signal allows to eject a mist of tiny droplets instead of a single main droplet and thus to define whether the obtained deposits are planar layers of nanospheres or 3D SCAs. The usually undesired effects of drop ejection in inkjet printing such as a mist of droplets were exploited to develop the presented SCAs based on the in-flight self-assembly approach. Finite, highly ordered spherical assemblies of nanospheres were obtained in a reproducible manner with variable diameters and thus different number of constituents using SiO2 or PS particles. Also binary SCAs were obtained consisting of PS and SiO2 nanospheres by printing a mixture of both suspensions. Selective removal of the organic nanospheres results in stable skeletons of the SiO2 nanospheres. After the deposition, the assemblies are located on solid surfaces in a dry environment and exhibit high stability. There is no need for further expensive process steps such as demulsification or hightemperature heating, which is usually required for WSA processes. This approach is less time-consuming and more convenient and practical compared with those previously reported. The SCAs have a well-ordered packing of nanospheres qualified for photonic properties. We could confirm the stop band by microreflectance measurements of the SCAs that are in perfect agreement with the theoretical calculations. Aside from applications in optoelectronics, photovoltaics, and sensorics, the manufactured SCAs can be interesting for displays, as carriers for drug delivery, as light-diffusion pigments for innovative coatings, and others. Further, metals and other nanomaterials such as carbon nanotubes could be deposited in a similar manner in advanced geometrical shapes, allowing the study of self-assembly and solidification phenomena in condensed-matter physics and its exploitation for nanopatterning purposes.

■

■

(1) Zhao, Y.; Shang, L.; Cheng, Y.; Gu, Z. Acc. Chem. Res. 2014, 47, 3632. (2) Yin, S.-N.; Wang, C.-F.; Liu, S.-S.; Chen, S. J. Mater. Chem. C 2013, 1, 4685. (3) Sato, O.; Kubo, S.; Gu, Z.-Z. Acc. Chem. Res. 2009, 42, 1. (4) Anselmann, R. J. Nanopart. Res. 2001, 3, 329. (5) Marlow, F.; Muldarisnur; Sharifi, P.; Brinkmann, R.; Mendive, C. Angew. Chem., Int. Ed. 2009, 48, 6212. (6) Cho, Y. S.; Kim, S. H.; Yi, G. R.; Yang, S. M. Colloids Surf., A 2009, 345, 237. (7) Manoharan, V. N. Solid State Commun. 2006, 139, 557. (8) Wang, Y.; Wise, A. K.; Tan, J.; Maina, J. W.; Shepherd, R. K.; Caruso, F. Small 2014, 4244. (9) Kim, S. H.; Lee, S. Y.; Yi, G. R.; Pine, D. J.; Yang, S. M. J. Am. Chem. Soc. 2006, 128, 10897. (10) Cho, Y. S.; Yi, G. R.; Kim, S. H.; Pine, D. J.; Yang, S. M. Chem. Mater. 2005, 17, 5006. (11) Yu, Z.; Chen, L.; Chen, S. J. Mater. Chem. 2010, 20, 6182. (12) Edwards, E. W.; Wang, D.; Möhwald, H. Macromol. Chem. Phys. 2007, 208, 439. (13) Cho, Y. S.; Yi, G. R.; Chung, Y. S.; Bin Park, S.; Yang, S. M. Langmuir 2007, 23, 12079. (14) Yi, G.-R.; Manoharan, V. N.; Michel, E.; Elsesser, M. T.; Yang, S.-M.; Pine, D. J. Adv. Mater. 2004, 16, 1204. (15) Sowade, E.; Hammerschmidt, J.; Blaudeck, T.; Baumann, R. R. Adv. Eng. Mater. 2012, 14, 98. (16) Cho, Y. S.; Kim, S. H.; Yi, G. R.; Yang, S. M. Colloids Surf., A 2009, 345, 237. (17) Zhao, Y.; Zhao, X.; Sun, C.; Li, J.; Zhu, R.; Gu, Z. Anal. Chem. 2008, 80, 1598. (18) Zhao, Y.; Zhao, X.; Pei, X.; Hu, J.; Zhao, W.; Chen, B.; Gu, Z. Anal. Chim. Acta 2009, 633, 103. (19) Xu, K.; Xu, J. H.; Lu, Y. C.; Luo, G. S. Cryst. Growth Des. 2013, 13, 926. (20) Ling, X. S. Nat. Mater. 2005, 4, 360. (21) Zhang, J.; Sun, Z.; Yang, B. Curr. Opin. Colloid Interface Sci. 2009, 14, 103. (22) Rastogi, V.; Melle, S.; Calderón, O. G.; García, A. A.; Marquez, M.; Velev, O. D. Adv. Mater. 2008, 20, 4263. (23) Ge, J.; Lee, H.; He, L.; Kim, J.; Lu, Z.; Kim, H.; Goebl, J.; Kwon, S.; Yin, Y. J. Am. Chem. Soc. 2009, 131, 15687. (24) Yi, G.; et al. Synth. Met. 2003, 139, 803. (25) Kim, S. H.; Lim, J. M.; Jeong, W. C.; Choi, D. G.; Yang, S. M. Adv. Mater. 2008, 20, 3211. (26) Kim, S. H.; Jeon, S. J.; Yi, G. R.; Heo, C. J.; Choi, J. H.; Yang, S. M. Adv. Mater. 2008, 20, 1649. (27) Moon, J. H.; Yi, G.-R.; Yang, S.-M.; Pine, D. J.; Park, S. B. Adv. Mater. 2004, 16, 605. (28) Kim, S.-H.; Lim, J.-M.; Lee, S.-K.; Heo, C.-J.; Yang, S.-M. Soft Matter 2010, 6, 1092.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.cgd.5b01567. Additional images of SCAs based on BS305 and PSC221. (PDF)

■

REFERENCES

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. *E-mail: [email protected]. Author Contributions ⊥

E.S. and T.B. contributed equally to the presented work.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors thank Torsten Jagemann, Steffen Schulze, and Michael Hietschold (TU Chemnitz, Department of Solid Surfaces Analysis) for experimental support with SEM and EDX mapping. Steffen Hemeltjen and Werner Goedel (TU Chemnitz, Department of Physical Chemistry) are acknowlI



Article

(29) Zhao, Y.; Xie, Z.; Gu, H.; Zhu, C.; Gu, Z. Chem. Soc. Rev. 2012, 41, 3297. (30) Lu, F.; Yager, K. G.; Zhang, Y.; Xin, H.; Gang, O. Nat. Commun. 2015, 6, 6912. (31) Maenosono, S.; Okubo, T.; Yamaguchi, Y. J. Nanopart. Res. 2003, 5, 5. (32) Hu, Y.; Wang, J.; Wang, H.; Wang, Q.; Zhu, J.; Yang, Y. Langmuir 2012, 28, 17186. (33) Gu, H.; Rong, F.; Tang, B.; Zhao, Y.; Fu, D.; Gu, Z. Langmuir 2013, 29, 7576. (34) Sun, C.; Zhao, X. W.; Zhao, Y. J.; Zhu, R.; Gu, Z. Z. Small 2008, 4, 592. (35) Marin, A. G.; Gelderblom, H.; Susarrey-Arce, A.; van Houselt, A.; Lefferts, L.; Gardeniers, J. G. E.; Lohse, D.; Snoeijer, J. H. Proc. Natl. Acad. Sci. U. S. A. 2012, 109, 16455. (36) Kursawe, M.; Anselmann, R.; Hilarius, V.; Pfaff, G. J. Sol-Gel Sci. Technol. 2005, 33, 71. (37) Anselmann, R.; Winkler, H. Adv. Eng. Mater. 2003, 5, 560. (38) Belgardt, C.; Sowade, E.; Blaudeck, T.; Baumgärtel, T.; Graaf, H.; von Borczyskowski, C.; Baumann, R. R. Phys. Chem. Chem. Phys. 2013, 15, 7494. (39) Belgardt, C.; Blaudeck, T.; von Borczyskowski, C.; Graaf, H. Adv. Eng. Mater. 2014, 16, 1090. (40) Lenggoro, I. W.; Xia, B.; Okuyama, K.; De la Mora, J. F. Langmuir 2002, 18, 4584. (41) Gan, H. Y.; Shan, X.; Eriksson, T.; Lok, B. K.; Lam, Y. C. J. Micromech. Microeng. 2009, 19, 055010. (42) Wijshoff, H. Phys. Rep. 2010, 491, 77. (43) Ko, H. Y.; Park, J.; Shin, H.; Moon, J. Chem. Mater. 2004, 16, 4212. (44) Uno, K.; Hayashi, K.; Hayashi, T.; Ito, K.; Kitano, H. Colloid Polym. Sci. 1998, 276, 810. (45) Sowade, E.; Blaudeck, T.; Baumann, R. R. Nanoscale Res. Lett. 2015, 10, 362. (46) Deegan, R. D.; Bakajin, O.; Dupont, T. F.; Huber, G.; Nagel, S. R.; Witten, T. a Nature 1997, 389, 827. (47) Soltman, D.; Subramanian, V. Langmuir 2008, 24, 2224. (48) Layani, M.; Gruchko, M.; Milo, O.; Balberg, I.; Azulay, D.; Magdassi, S. ACS Nano 2009, 3, 3537. (49) Perelaer, J.; Smith, P. J.; Hendriks, C. E.; van den Berg, A. M. J.; Schubert, U. S. Soft Matter 2008, 4, 1072. (50) Kuang, M.; Wang, J.; Bao, B.; Li, F.; Wang, L.; Jiang, L.; Song, Y. Adv. Opt. Mater. 2014, 2, 34. (51) Park, J.; Moon, J.; Shin, H.; Wang, D.; Park, M. J. Colloid Interface Sci. 2006, 298, 713. (52) Park, J.; Moon, J. Langmuir 2006, 22, 3506. (53) Dong, H.; Carr, W. W.; Morris, J. F. Phys. Fluids 2006, 18, 072102. (54) Li, R.; Ashgriz, N.; Chandra, S. Exp. Therm. Fluid Sci. 2008, 32, 1679. (55) Basaran, O. A. AIChE J. 2002, 48, 1842. (56) Song, J.; Nur, H. M. J. Mater. Process. Technol. 2004, 155−156, 1286. (57) Kim, S. H.; Cho, Y. S.; Jeon, S. J.; Eun, T. H.; Yi, G. R.; Yang, S. M. Adv. Mater. 2008, 20, 3268. (58) Masuda, Y. KONA 2007, 25, 244. (59) Zhao, X.; Cao, Y.; Ito, F.; Chen, H.-H.; Nagai, K.; Zhao, Y.-H.; Gu, Z.-Z. Angew. Chem. 2006, 118, 6989. (60) Mouaziz, H.; Lacki, K.; Larsson, A.; Sherrington, D. C. J. Mater. Chem. 2004, 14, 2421. (61) Shchukin, D. G.; Caruso, R. A. Chem. Mater. 2004, 16, 2287. (62) Kuncicky, D. M.; Velev, O. D. Langmuir 2008, 24, 1371. (63) Gokmen, M. T.; Du Prez, F. E. Prog. Polym. Sci. 2012, 37, 365. (64) Yi, G.-R.; Manoharan, V. N.; Klein, S.; Brzezinska, K. R.; Pine, D. J.; Lange, F. F.; Yang, S.-M. Adv. Mater. 2002, 14, 1137. (65) Lipowsky, P.; Bowick, M. J.; Meinke, J. H.; Nelson, D. R.; Bausch, A. R. Nat. Mater. 2005, 4, 407. (66) Zhao, Y.; Xie, Z.; Gu, H.; Zhu, C.; Gu, Z. Chem. Soc. Rev. 2012, 41, 3297.

(67) McGrath, J. G.; Bock, R. D.; Cathcart, J. M.; Lyon, L. A. Chem. Mater. 2007, 19, 1584. (68) Cui, L.; Li, Y.; Wang, J.; Tian, E.; Zhang, X.; Zhang, Y.; Song, Y.; Jiang, L. J. Mater. Chem. 2009, 19, 5499. (69) Norris, D. J.; Arlinghaus, E. G.; Meng, L.; Heiny, R.; Scriven, L. E. Adv. Mater. 2004, 16, 1393. (70) Zhao, X.; Cao, Y.; Ito, F.; Chen, H.-H.; Nagai, K.; Zhao, Y.-H.; Gu, Z.-Z. Angew. Chem. 2006, 118, 6989. (71) Schiffer, P. Nat. Phys. 2005, 1, 21. (72) Scheel, M.; Seemann, R.; Brinkmann, M.; Di Michiel, M.; Sheppard, A.; Breidenbach, B.; Herminghaus, S. Nat. Mater. 2008, 7, 189. (73) Zhao, Y.; Zhao, X.; Hu, J.; Xu, M.; Zhao, W.; Sun, L.; Zhu, C.; Xu, H.; Gu, Z. Adv. Mater. 2009, 21, 569. (74) Kim, S.-H.; Lee, S. Y.; Yang, S.-M.; Yi, G.-R. NPG Asia Mater. 2011, 3, 25. (75) Yu, Z.; Chen, L.; Chen, S. J. Mater. Chem. 2010, 20, 6182. (76) Parker, A. R.; Welch, V. L.; Driver, D.; Martini, N. Nature 2003, 426, 786.

J


Self-Assembly of Spherical Colloidal Photonic ... - ACS Publications

Recommend Documents