Pin Ann Lin

Nucleation of graphene and its conversion to single-walled carbon nanotubes.

We use an environmental transmission electron microscope to record atomic-scale movies showing how carbon atoms assemble together on a catalyst nanoparticle to form a graphene sheet that progressively lifts-off to convert into a nanotube. Time-resolved observations combined with theoretical calculations confirm that some nanoparticle facets act like a vice-grip for graphene, offering anchoring sites, while other facets allow the graphene to lift-off, which is the essential step to convert into a nanotube.

Nanocatalyst shape and composition during nucleation of single-walled carbon nanotubes

The dynamic evolution of nanocatalyst particle shape and carbon composition during the initial stages of single-walled carbon nanotube growth by chemical vapor deposition synthesis is investigated. Classical reactive and ab initio molecular dynamics simulations are used, along with environmental transmission electron microscope video imaging analyses. A clear migration of carbon is detected from the nanocatalyst/substrate interface, leading to a carbon gradient showing enrichment of the nanocatalyst layers in the immediate vicinity of the contact layer. However, as the metal nanocatalyst particle becomes saturated with carbon, a dynamic equilibrium is established, with carbon precipitating on the surface and nucleating a carbon cap that is the precursor of nanotube growth. A carbon composition profile decreasing towards the nanoparticle top is clearly revealed by the computational and experimental results that show a negligible amount of carbon in the nanoparticle region in contact with the nucleating cap. The carbon composition profile inside the nanoparticle is accompanied by a well-defined shape evolution of the nanocatalyst driven by the various opposing forces acting upon it both from the substrate and from the nascent carbon nanostructure. This new understanding suggests that tuning the nanoparticle/substrate interaction would provide unique ways of controlling the nanotube synthesis.

Direct evidence of atomic-scale structural fluctuations in catalyst nanoparticles

Rational catalyst design requires an atomic scale mechanistic understanding of the chemical pathways involved in the catalytic process. A heterogeneous catalyst typically works by adsorbing reactants onto its surface, where the energies for specific bonds to dissociate and/or combine with other species (to form desired intermediate or final products) are lower. Here, using the catalytic growth of single-walled carbon nanotubes (SWCNTs) as a prototype reaction, we show that the chemical pathway may in-fact involve the entire catalyst particle, and can proceed via the fluctuations in the formation and decomposition of metastable phases in the particle interior. We record in situ and at atomic resolution, the dynamic phase transformations occurring in a Cobalt catalyst nanoparticle during SWCNT growth, using a state-of-the-art environmental transmission electron microscope (ETEM). The fluctuations in catalyst carbon content are quantified by the automated, atomic-scale structural analysis of the time-resolved ETEM images and correlated with the SWCNT growth rate. We find the fluctuations in the carbon concentration in the catalyst nanoparticle and the fluctuations in nanotube growth rates to be of complementary character. These findings are successfully explained by reactive molecular dynamics (RMD) simulations that track the spatial and temporal evolution of the distribution of carbon atoms within and on the surface of the catalyst particle. We anticipate that our approach combining real-time, atomic-resolution image analysis and molecular dynamics simulations will facilitate catalyst design, improving reaction efficiencies and selectivity towards the growth of desired structure.

Measurement of Local Specimen Temperature under Flowing Gas Ambient in the Environmental Scanning Transmission Electron Microscope (ESTEM) Using Diffraction

In situ heating experiments in the TEM have been widely used to study many materials phenomena. Among the factors limiting more quantitative analysis of in situ TEM data is a precise knowledge of the sample temperature during a reaction. Going back to the earliest days of in situ and gas-cell TEM experiments accurate measurement of the temperature at the specimen during heating has been an issue. The lack of information about the specimen temperature is particularly problematic for experiments with flowing gases which significantly cool the sample [1]. The use of a thermocouple near the sample cradle can estimate the temperature, but the newer MEMS-based heating holders do not have thermocouples and temperature calibration is performed ex situ in vacuum with a light pyrometer. It is desirable to have a technique for precise, local temperature measurement over a large temperature range that can be applied under flowing gas conditions relevant to ESTEM experiments.

Automated Image Processing Scheme to Measure Atomic-Scale Structural Fluctuations

The interaction of gases with a solid catalyst nanoparticle during catalysis is a non-equilibrium process that requires high spatial and temporal resolution measurements to elucidate underlying mechanisms. State-of-the-art environmental transmission electron microscopy (ETEM) enables in situ measurements of the dynamic changes occurring under reaction conditions [1,2]. These changes usually take place rapidly at the nanometer scale. Recently, direct electron detection cameras, have enabled us to record atomic-resolution images with ns time resolution, but generate videos with large amount of data (≈ GB s). It is laborious to manually analyze such large-size videos, frame by frame, to extract the events of interest at the required time resolution. Automated analysis would be preferable, but is complicated by (a) noise in individual frames due to rapid readout times and (b) sample drift that occurs in a single video recording period. In order to overcome these issues, we have developed an automated image processing scheme (AIPS), to obtain structural information from the images extracted from videos. AIPS uses a combination of algorithms publically available and developed at NIST that perform noise reduction, drift correction, template matching, atom-position location, and triangulation to accurately determine the positions of atomic columns. We tested our method by quantitatively relating the crystal structure fluctuations in a catalyst nanoparticle to the growth of single-walled carbon nanotube (SWCNT) as a function of time.

Dynamic structural changes in a single catalyst particle during single walled carbon nanotube growth

The interaction of gases with a solid catalyst nanoparticle during catalysis is a non-equilibrium process. For example, during single-walled carbon nanotube (SWCNT) growth, carbon atoms diffuse in and out of the catalyst particle, causing variations in the chemical potential, and possibly structure, of the particle. We have employed an environmental scanning transmission electron microscope (ESTEM), with an aberration corrector, operated at 300 kV to record real-time, atomic-resolution videos (6 frames s) with ≈ 1000 frames of SWCNT growth from a Co-Mo/MgO system in a C2H2 gaseous environment at synthesis temperatures. The time-resolved videos, generating large data sets, are used to identify individual reaction steps. We have developed methods to analyze these large data sets accurately and efficiently.

Measuring Gas Adsorption on Individual Facets of a Nanoparticle by a Surface Plasmon Nanoprobe

Gas adsorption on metal nanoparticles is a fundamental step that controls a number gas-solid reaction processes, important in applications ranging from catalysis to gas sensing. Gas absorption depends on the nature of the gas molecules and atomic arrangement of the surface facets. These factors control the gas binding energies, which have not, so far, been directly measured. Here we show that electron-density changes, induced by gas adsorption on the active metal nanoparticle facet, cause shifts in surface plasmon (SP) energies that can be measured by electron energy loss spectroscopy (EELS). We employ a 1 nm-diameter electron beam from a monochromated 80kV electron source to locally (< 2 nm) excite SPs and measure their energies with 100 meV energy resolution. This high spatial and energy resolution allows us to resolve SP energy shifts in the range of a few meV, localized to individual nanoparticle facets. Using this technique in an environmental scanning transmission electron microscope (ESTEM), we are able to map in situ SP responses on different facets (i.e. corners and sides) of individual Au nanoparticles in vacuum, CO, and H2 at various gas pressures. Our results were further confirmed by finite-difference-time-domain (FDTD) simulations for the spatial localization of the electron beam excited SP nanoprobe. Figure 1a shows the STEM dark-field image of a triangular Au nanoparticle on a TiO2 support and the local SP excitation at the corner location A and side location B. Energy loss maps in the Au nanoparticle, obtained by FDTD simulations, reveal the highly-localized excitation volumes near the edge of the particle in both locations (Fig. 1b and c). SP-EELS spectra at both locations were collected in vacuum, and over a range of gas pressures. In the CO environment, the SP energy was observed to shift to higher energies and the magnitude of the energy shift was larger when probing location B than location A. (Fig. 2a). In contrast, in the H2 environment, the SP energy shifted to lower energies and the magnitude of the energy shift was larger when probing location A than location B. To confirm that energy shifts measured with such localized excitation are sensitive only to the electron density variations of the probed location, two models of electron density variations in a 1 nm skin depth at the corner and the side, respectively, were applied in our FDTD simulations. When the electron density varies at the corner, the SP energy varies only for probing location A (Fig.3a) but not when probing location B (Fig.3b). Similarly a variations in the electron density at the side yielded energy shifts in location B (Fig.3c), whereas no changes in location A (Fig. 3d). Therefore, we can conclude that the difference in magnitudes of gas-induced SP energy shifts reports a local phenomenon, and indicates crystallographically-specific absorption of the different gases on the Au nanoparticle facets. A detailed discussion of the SP energies measured for different facet-gas interactions and quantitative measurements of binding energies and number density of molecules adsorbed for different gasses on various facets of Au nanoparticles will be presented. Paper No. 1025 2053 doi:10.1017/S1431927615011046

Gas-metallic nanoparticle surface interaction characterized with in-situ electron energy loss spectroscopy

We use an environmental scanning transmission electron microscope (ESTEM) equipped with electron energy loss spectroscopy (EELS) and a monochromated electron source to perform energy loss measurements on metallic nanoparticles (NPs) exposed to local gaseous environments at varying pressures. In particular, we characterize the effect of exposure to CO or H2 on the surface plasmon resonance of a gold NP. By addressing various sites around the perimeter of a triangular NP (edge length ~20 nm) with the electron beam in STEM mode, the energy loss spectrum resulting from site-specific excitation of surface plasmon resonance is probed with a spatial resolution of ~1 nm and energy resolution of ~100 meV. Local gas adsorption is evidenced by peak shifts in the energy loss spectrum, which are found to be positive for CO and negative for H2. Strong site selectivity is evident, with CO and H2 adsorbing preferentially at the edge and corner sites, respectively. To characterize the sign and magnitude of the energy shifts, finite-difference time-domain (FDTD) simulations of electron-beam excitation of the NP are performed using a specialized model in which the local electron concentration is allowed to vary spatially over the particle volume. This is a result of both the inhomogeneous spatial distribution of the adsorbate and its degree of electronegativity.

Atomic Resolution Single Walled Carbon Nanotube Nucleation Steps on Faceted Catalyst Particle Reveal Potential for Chirality Control

Single-walled carbon nanotubes (SWCNTs) continue to be one of the most desirable materials for numerous nanotechnology applications due to their unique combination of quantum confinement and large surface area-to-volume ratio. Despite this, only limited numbers of SWCNT-based technologies have come to the market because of the difficulty in generating sufficient quantities of SWNCTs with uniform electronic properties. This is because the electronic properties of SWCNTs vary with their diameter, defect density and, most importantly, with their chirality. Although a few reports indicate that a degree of chirality control is possible with specific nanoparticle (NP) catalyst-support system [1,2], there is a fundamental lack of knowledge about what processes determine SWCNT chirality, particularly concerning the NT nucleation process on the catalyst, which precludes the design of efficient experimental approaches and rational catalyst design. We have employed an Environmental Scanning Transmission Electron Microscope (ESTEM) [3], operated at 300 kV with an aberration corrector, which permits us to record real-time atomic resolution videos (6 frames s -1 ) to capture nanotube cap nucleation, lift-off, and growth in C2 H2 flowing over the Co-Mo/MgO system at synthesis temperatures. Detailed observations at this relatively slow frame rate are made possible by using lowpressure (0.005 Pa of C2H2) growth conditions (at 650 °C) to slow down the NT nucleation and growth rate.