|Home | About | Journals | Submit | Contact Us | Français|
A simple strategy for making nanoparticles by sonofragmentation of high-aspect-ratio 1D substrates is introduced. With common laboratory equipment, ultra-thin nanowires are fragmented into nanoparticles of size determined by the nanowire width, resulting within hours in monodisperse, crystalline nanoparticles of < 10 nm. This strategy is applicable to a diversity of semiconductor, oxide and metal nanowires.
Small (< 10 nm) nanoparticles (NPs) are important because of the unique physical and chemical properties that arise due to their small size and large surface area.1–9 A multitude of methods have been developed to produce such nanoparticles.10 Top-down synthesis methods that rely on breaking down bulk materials into smaller fragments can be scalably deployed.11–21 However, the method struggles with monodispersity and with percent yield for such small nanoparticles. Bottom-up synthesis methods can effectively assemble small molecule precursors into larger units to create small nanoparticles.22–35 However, both methods require harsh chemicals20, 34 or specialized equipment,13, 21, 27, 29, 35 out of reach from many end users. Ideally, one could obtain extremely monodisperse nanoparticles of small size and high yield on regular benchtop equipment on site.
We here propose a top-down method of nanoparticle synthesis that results in high-monodispersity nanoparticles. We hypothesized that nanowires of extreme aspect ratio could be ultrasonicated to generate nanoparticles. This hypothesis builds from recent studies36–40 that show ultrasonication can be used to break down nanowires into shorter nanowires, and nanotubes into shorter nanotubes. We hypothesized that by choosing nanowires of high aspect ratios, and then applying ultrasonication, it would be possible to perform top-down synthesis of many kinds of nanoparticle in effectively a single step. We note that the final yield of the nanoparticle synthesis would depend on the yield and supply of the starting materials, some of which require specialized equipment and precursors. With a constant supply of the nanowires, our method would enable scalable production of ultra-small nanoparticle production in large quantities. Such nanowire production could be realized by, for example, a catalyzed high-throughput gas phase synthesis with extremely high precursor efficiency and gram-scale yield.41
We set out to test our hypothesis using ultrathin Ge nanowires (Figure 1a). Up to date, synthesis of Ge NPs (< 10 nm) has been limited to gas-phase and liquid phase approaches that require expensive machines,42 and top-down approaches that do not yield monodisperse crystalline nanoparticles.12, 43 Therefore, a simple and inexpensive method for top-down synthesis of Ge nanoparticles would be potentially of both scientific and commercial interest.
We first dispersed ultra-thin Ge nanowires44 (diameters tapering from ~30 nm to ~2 nm) in DMF, and ultrasonicated the suspension with a bench-top ultrasonicator (40 kHz, 110 W). To track fragmentation of the nanowires, we imaged the ultrasonicated sample at different time points using scanning electron microscopy (SEM) (Figure 1b top row, Figure S1). We found that the nanowires readily fragmented into < 30 nm particles within 30 minutes of ultrasonication. During the subsequent long-term ultrasonication, the particle size further decreased with increasing ultrasonication time. For instance, the majority of the nanoparticles had diameters of < 10 nm with 18 hr ultrasonication. As comparison, we carried out the same ultrasonication using a non-1D Ge substrate (100~300 nm diameter nanopowder) (Figure 1b bottom row, Figure S2). Contrary to the nanowires, the nanopowder did not show a clear change in particle size with increasing ultrasonication time. For instance, after 18 hrs of ultrasonication, we observed a majority of ~100-300 nm particles, comparable to the size distribution of the starting material.
We analyzed the Ge NPs produced after 18 hrs of nanowire ultrasonication using transmission electron microscopy (TEM). The nanoparticles were resuspended in ethanol, filtered through a 0.2 µm filter to remove large debris and aggregates, and drop-casted and dried on a carbon-copper grid (Figure 2a). Analysis of the bright-field TEM images shows the nanoparticles had an average size of 3.58 nm and a standard deviation of 0.74 nm (n = 75 from a single TEM grid; Figure 2b), confirming generation of ultrasmall (< 10 nm) Ge NPs. Furthermore, high-resolution TEM (HRTEM) imaging of a typical Ge nanoparticle shows clear lattice fringes, indicating a minimal amorphization effect during the long-term ultrasonication (Figure 2a, inset). The ~0.20 nm spacing of lattice fringes corresponds to the spacing between (220) planes of Ge, consistent with the starting material of crystalline Ge nanowires.45 In addition to the 18 hrs ultrasonicated nanoparticles, we imaged Ge nanoparticles after 30 min and 1 hr of ultrasonication with TEM (Figure S3). The results show a nanoparticle size change consistent with the previous SEM experiments.
We further carried out dynamic laser scattering (DLS) analysis of the ultrasonicated Ge nanowire sample. Consistent with the TEM analysis, we found that monodisperse (polydispersity (Pd) = 6.8%) Ge NPs of 2–5 nm diameters were generated after 18 hrs of ultrasonication, with no further purification (Figure 2c). We also carried out a temperature-controlled sonofragmentation experiments with two different temperature ranges of 10–20 °C and 60–65 °C (Figure S4). The results show that within the concerned range, temperature had minimal effect on the synthesized nanoparticle size distribution. In comparison to the Ge nanowire substrate, the Ge nanopowder substrate showed similar nanoparticle size range and distribution before (Pd = 16.2%) vs. after an ultrasonication time of 36 hrs (Pd = 19.1%) (Figure S5). This result further suggests the advantage of using an ultra-thin 1D substrate to produce monodisperse ultrasmall nanoparticles.
To investigate optical properties of the synthesized Ge NPs, we measured the absorbance of the ultrasonicated sample using a UV-vis spectrometer (Figure 2d, blue). We found the Ge NPs readily absorbed light with <400 nm wavelengths.46 Next, we measured the intrinsic photoluminescence (PL) of the Ge NPs under optical excitation using a UV-vis spectrometer. The sample showed a characteristic PL peak around 410 nm wavelengths, consistent with previous reports (Figure 2d, red).24 We note that the blue emission observed can possibly arise from surface oxidation and absorption of molecules. 24
To study the surface of the synthesized Ge NPs, we performed Fourier transform infrared (FTIR) spectroscopy on the nanoparticles (Figure 2e). The ultrasonicated Ge NPs were washed in chloroform three times and re-suspended in chloroform. The suspension was then drop-casted and air dried on the attenuated total reflectance (ATR) crystal before the FTIR measurements. The surface of the as-synthesized Ge NPs displayed both free hydroxyls (3334 cm-1) and DMFs, which are likely to be chemisorbed onto the surface through a C-O-Ge (1668 cm-1) bridge.47 We further studied the surface of the nanoparticles after solvent exchange and found that the DMF molecules were retained after the exchange to ethanol and water.
To assess whether our method can be applied to different types of 1D substrates, we explored synthesis of NPs using various commercially available nanowires, including oxide and metal nanowires. We first carried out ultrasonication of commercially available TiO2 nanowires (~10 nm diameter, Figure 3a) in water for 24 hrs. TEM analysis shows that the average and standard deviation of the nanoparticle size are 4.63 nm and 1.28 nm, respectively, confirming generation of nanoparticles of < 10 nm diameter (Figure 3b, c). HRTEM image of a typical TiO2 nanoparticle shows clear lattice fringes, indicating that the nanoparticles are crystalline (Figure 3b, inset). The ~0.28 nm spacing between fringes is consistent with the spacing between (100) planes of rutile TiO2.48 We also characterized the TiO2 NPs in the solvent and found the nanoparticles has a monodisperse size distribution (Pd = 11%) of ~3-6 nm diameter, a range consistent with the TEM analysis (Figure 3d).
Next we carried out ultrasonication of commercially available Ag nanowires (~20 nm diameter, Figure 3e) using the same sonofragmentation process as the TiO2 nanowires. TEM characterization shows the synthesized Ag NPs are crystalline and have average size and standard deviation of 3.46 nm and 0.75 nm, respectively (Figure 3f, g). The HRTEM image of a typical Ag NP shows a lattice fringe spacing of ~0.24 nm, consistent with the (111) plane spacing of Ag (Figure 3f, inset).49 We also measured the nanoparticle size in the solvent and the results show a monodisperse size distribution (Pd = 15 %) of ~2-7 nm diameter, consistent with the TEM results (Figure 3h).
Finally, we carried out ultrasonication of Si nanowires50, 51 (~30 nm diameter, Figure 3i) in DMF for 24 hrs. TEM analysis shows that the Si NPs are crystalline and the average and standard deviation of the nanoparticle size are 10.8 nm and 2.2 nm, respectively (Figure 3j, k). HRTEM image of a typical Si NP shows a ~0.27 nm spacing between the lattice fringes, which likely corresponds to the spacing between (200) planes of a diamond cubic lattice of silicon (Figure 3j, inset).52 We note that in this particular image, the commonly observable (111) fringes were not clearly resolved. To characterize the nanoparticle size in the solvent, we measured the nanoparticle size using DLS. The results show a monodisperse size distribution (Pd = 11.5 %) of ~10-12 nm diameter, a range consistent with the TEM results (Figure 3l). Furthermore, to characterize optical properties of the synthesized Si NPs, we measured the PL of suspension. The results show a violet-blue fluorescence peak at around 400 nm in wavelength, consistent with previous reports (Figure S6).53, 54
Based on previous theoretical and experimental studies of ultrasonication,36–40, 55–57 we think the effects of long-term and continuous sonofragmentation on ultra-thin nanowires are two-fold: physical and chemical. In a previous study that used a theoretical model to calculate the tensile stress applied by a cavitation bubble, the tensile stress on a 1D nanostructure is shown to be dependent on the ratio of its diameter to its length.33 The model suggests that thinner and longer nanowire and nanotube substrates can be more easily broken into fragments compared with substrates of low aspect ratio.36 In another mechanical study, it had been predicted and shown for the case of carbon nanotubes, that shorter nanofragments are produced with increasing sonication times.37 Our observation of nanoparticle generation from ultrasonication of high aspect ratio nanowires is consistent with these predictions and observations. Aside from mechanical fragmentation of nanowires, significant local heating up to a few thousand Kelvin near cavitation bubbles can be another cause of nanowire fragmentation.58 Previous studies have shown that metal and semiconductor nanowires, driven by the Plateau-Rayleigh instability, readily form a string of nanospheres when heated.59, 60 The thermal instability of ultra-thin nanowires could in principle therefore be another physical route for nanoparticle generation during ultrasonication.
From a chemical point of view, surface functionalization of the nanoparticles plays an important role in dispersing and stabilizing nanoparticles in solvents during the sonofragmentation.61, 62 For instance, the FTIR analysis of the ultrasonicated Ge NPs suggests that the surfaces of nanoparticles are terminated with DMF molecules with the CO groups coordinating to the Ge atoms. We suspect that these surface coordinated solvent molecules stabilize nanoparticles and prevent them from fast oxidation and decomposition. In addition, the partially positive charge on the nitrogen terminal is likely to prevent the Ge NPs from aggregating in polar solvents such as DMF and ethanol and thus keep the nanoparticles dispersed in these solvents.
Our time-evolution experiment of the Ge fragments further provides insight into possible mechanism of nanoparticle generation during sonofragmentation. During the initial phase of the ultrasonication, the Ge nanowires rapidly fragment into < 30 nm particles. This process is complete within ~30 minutes which is likely due to the high aspect ratio of the nanowire substrate. Increasing the ultrasonication time further reduces the size of these particles: with 18 hrs of ultrasonication, the size range decreases to 3-5 nm.
Future investigation waits on both theoretical and experimental fronts. Additional theoretical modeling for sonofragmentation of ultra-thin inorganic substrates of different materials, diameters and lengths as well as mechanical properties might be explored. In addition, it may be of great interest to test specific substrate-solvent combinations and observe the effect of different solvents and surfactants on different substrates and on the sonofragmentation process itself. Furthermore, a post-sonofragmentation surface modification could further tune and improve physical and chemical properties of the synthesized nanoparticles.
In this report, we presented a new method of nanoparticle synthesis based on sonofragmentation of ultra-thin 1D substrates. We discovered that short-term ultrasonication of high-aspect ratio 1D substrates could rapidly generate highly-monodisperse nanoparticles, and that the subsequent longer-term ultrasonication could result in ultrasmall nanoparticles. We believe our method opens up a new approach, which is implementable with a bench-top ultrasonicator, to synthesize nanoparticles of high purity, crystallinity and monodispersity. Thus, our methodology democratizes small nanoparticle production, potentially opening up doors in a variety of fields that would benefit from the use of small nanoparticles for their chemical and physical properties.
Sonofragmentation of 1D substrates were carried out using a bench-top bath ultrasonicator (40 kHz, max sonication power 110 W, Bransonic Ultrasonic Baths, Thomas Scientific). Starting materials in powder or suspended form (TiO2 nanowires, Sigma-Aldrich; Ag nanowires, Novarials Corp.; Ge nanopowder, SkySpring Nanomaterials, Inc.) were added directly to an amber glass vial (4 ml, Sigma-Aldrich) with the solvents for the ultrasonication. Starting materials attached to a wafer substrate were first gently ultrasonicated in the solvents for 2 min and then the supernatant was transferred to another amber glass vial for the subsequent ultrasonication. The bath temperature of the ultrasonicator was not actively controlled unless otherwise noted. The temperature typically increased from about 25 °C to about 60 °C for the18 hr ultrasonication. Active control of temperature was achieved by using a chiller (RC2 Basic, IKA) and the internal heating system of the ultrasonicator for the temperature range of 10–20 °C, and 60–65 °C, respectively.
TEM characterization of the nanoparticles (NPs) was carried out using a JEM-2100 TEM (JEOL). The as-synthesized nanoparticles were (re)suspended in ethanol (for Ge, TiO2 and Si NPs) or water (for Ag NPs) before being filtered through a 0.2 µm filter to remove large aggregates and debris. The suspension was then drop-casted on a carbon-copper grid (Ted Pella, Inc.), and dried in a vacuum desiccator for 20 min. The imaging was carried out at 200 keV under bright-field illumination. SEM characterization of the nanowires and fragments was carried out using an UltraPlus FE-SEM (Zeiss) with an inlens detector.
DLS characterization of the nanoparticles was carried out with a dynamic light scattering instrument (DynaPro NanoStar, Wyatt Technology Corp.). About 100 uL of the sample was transferred to a disposable cuvette (Wyatte Technology Corp.) for the DLS measurement. The final histogram of nanoparticle size distribution was generated from 10 measurements for each sample.
PL characterization of the nanoparticles was carried out using a fluorescence spectrometer (Cary Eclipse, Agilent). About 40 ul of the sample was transferred to a quartz cuvette (Sigma-Aldrich) for the fluorescence measurement. UV-vis spectra of the nanoparticles were measured using a bench-top UV-vis spectrometer (NanoDrop 2000, ThermoFisher).
FTIR characterization of the Ge NPs was carried out using an FTIR spectrometer (SpectrumOne, Perkin Elmer). After 18 hrs of ultrasonication in DMF, the nanoparticles were dried under vacuum and resuspended in chloroform for three times to completely remove the DMF. The nanoparticle suspension was then drop-casted onto the attenuated total reflection (ATR) crystal of the FTIR spectrometer and air-dried for 15 min before the measurement. The FTIR measurement was carried out for 3 min and the baseline was automatically corrected.
Ge and Si nanowires were synthesized with vapor-liquid-solid (VLS) growth mechanism using published protocols.44, 50, 51 Briefly, Ge nanowires were grown with 2 nm gold nanocatalyst for 150 min using GeH4 (2 sccm) and H2 (18 sccm) at total pressure of 400 torr and temperature of 270 °C. Si nanowires were grown for 60 min with 30 nm gold nanocatalyst using SiH4 (2.5 sccm) and H2 (60 sccm) at total pressure of 40 torr and temperature of 450 °C.
E. S. Boyden acknowledges support from the MIT Media Lab, the New York Stem Cell Foundation-Robertson Investigator Award, NIH Director’s Pioneer Award 1DP1NS087724, the DOD MURI program, the Open Philanthropy Project, and NSF INSPIRE Award CBET 1344219. The authors thank Prof. C. M. Lieber and Lieber group members for insightful discussion. This work was performed in part at the Harvard University Center for Nanoscale Systems (CNS), a member of the National Nanotechnology Coordinated Infrastructure Network (NNCI), which is supported by the National Science Foundation under NSF award no. 1541959. This work also made use of the MRSEC Shared Experimental Facilities at MIT, supported by the National Science Foundation under award number DMR-1419807.
Dr Ruixuan Gao, Media Lab, Massachusetts Institute of Technology, Cambridge, MA 02139.
Ishan Gupta, Media Lab, Massachusetts Institute of Technology, Cambridge, MA 02139. Department of Biological Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139.
Edward S. Boyden, Media Lab, Massachusetts Institute of Technology, Cambridge, MA 02139. Department of Biological Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139. McGovern Institute for Brain Research, Massachusetts Institute of Technology, Cambridge, MA 02139.