Session P34: Focus Session: Nano III: New Nanoscale Fabrication and Sensing

Computational nanomaterials for novel desalination membrane design: Nanoporous graphene

We describe a novel approach for desalination based on nanoporous graphene. Our molecular dynamics calculations show that freestanding graphene patterned with nanometer-sized pores can act as an ultra-thin filtration membrane. Due to size exclusion and chemical interactions with the confining pores, salt ions can be blocked from permeating the membrane at sufficiently small pore diameters. Notably, the pore diameter and the chemical interactions at the water-membrane interface are most important criteria for this system's desalination performance. We will share insights from Molecular Dynamics calculations regarding the theoretical performance of this membrane system and the effects of chemical passivation of the graphene pores on the filtration dynamics. Although the narrow range of acceptable pore sizes suggests that further design innovations will be necessary at the molecular scale before large-scale applications are possible, our existing results predict that pressure requirements for this system can be made roughly competitive with commercial Reverse Osmosis.   

Structural and electronic properties of bare and capped CdnSen/CdnTen nanoparticles (n = 6, 9)

Relationships between structures and properties (energy gaps, vertical ionization potentials (IP$_{v})$, vertical electron affinities (EA$_{v})$, and ligand binding energies) in small capped CdSe/CdTe nanoparticles (NPs) are poorly understood. We have performed the first systematic density functional theory study of the structures and electronic properties of Cd$_{n}$Se$_{n}$/Cd$_{n}$Te$_{n}$ NPs (n = 6, 9), both bare and capped with NH$_{3}$-, SCH$_{3}$, and OPH$_{3}$-ligands. NH$_{3}$- and OPH$_{3}$-ligands cause HOMO/LUMO energy \textit{destabilization} in capped NPs, more pronounced for the LUMOs than for the HOMOs. Orbital destabilization drastically reduces both the IP$_{v}$ and EA$_{v}$ of the NPs compared with the bare NPs. For SCH$_{3}$-capped Cd$_{6}$X$_{6}$ NPs, formation of expanded structures was found to be preferable to crystal-like structures. SCH$_{3}$-groups cause \textit{destabilization} of the HOMOs of the capped NPs and \textit{stabilization} of their LUMOs, which indicates a reduction of the IP$_{v}$ of the capped NPs compared with the bare NPs. For the Cd$_{9}$X$_{9}$ NPs, similar trends in stabilization/destabilization of frontier orbitals were observed.   

New Directions in Plasmonics: Pushing the Space-Time Limit

The lecture will begin with the discussion of our efforts to provide a robust existence proof for SMSERS (Single Molecule Surface Enhanced Raman Spectroscopy). Further, fundamental questions such as: (1) what is the largest possible enhancement factor (EF) and (2) what nanostructure produces the largest EF, will be addressed. Our approach to answering these questions involved the development of new tools such as single nanoparticle SERS and single nanoparticle LSPR spectroscopy spatially correlated with high resolution transmission electron microscopy (HRTEM). Recent results using LSPR biosensors to detect molecular binding events and conformation changes will be presented, including discussions of: (1) pushing the sensitivity of plasmonic biosensors towards the single-molecule detection limit, (2) combining LSPR with complementary molecular identification techniques such as matrix assisted laser desorption ionization mass spectrometry (MALDI-MS), and (3) the development of new instrumentation for high throughput plasmonic biosensing, and gas sensing with plasmonic nanosensors. Finally, recent developments showing that for the first time, the revolutionary techniques of surface enhanced Raman spectroscopy and femtosecond stimulated Raman spectroscopy (FSRS) can be combined and substantial progress in tip-enhanced Raman spectroscopy (TERS) will be presented. A UHV-TERS instrument has been constructed with atomic resolution of the surface and sub-molecular resolution of the adsorbate, as illustrated with the copper phthalocyanine (CuPc)/Ag(111) system. We can now foresee the day when it will be possible to combine UHV-TERS and surface enhanced FSRS to enable single-molecule spectroscopy with simultaneous nanometer spatial resolution and femtosecond time resolution.   

Charge Distributions in Polar Semiconductor Nanorods explored with Linear-Scaling DFT Calculations

Binary polar semiconductors in the wurtzite structure can be grown as nanorods along $\pm$[0001]. In such structures, large dipole moments have been observed. We have studied the distribution of charge in GaAs and ZnO nanorods to elucidate the origin of the dipole moments. To make contact with realistic experiments, rods containing thousands of atoms are simulated using Linear-Scaling DFT calculations with ONETEP [1]. From our calculations we show that both the direction and magnitude of the dipole moment of a nanorod, and its electric field, depend sensitively on how its surfaces are terminated, not on the spontaneous polarization of the underlying lattice. Furthermore, we observe that the Fermi level for an isolated nanorod always coincides with significant density of electronic surface states on its polar surfaces (either mid-gap states or band-edge states). These states pin the Fermi level, and therefore fix the potential difference along the rod. We provide evidence that this effect has a determining influence on the polarity of nanorods, with consequences for the response to changes in surface chemistry, scaling of dipole moment with size, and dependence of polarity on composition.\\[4pt] [1] C. Skylaris et al, JCP 122, 084119 (2005).\\[0pt] [2] P. Avraam et al, PRB 83 241402(R) (2011).   

Reduction of graphene oxide to graphene, A study of changes in the atomic structure

An economic method for large scale production of graphene is based on exfoliation of graphite into 1-atom thick sheets by oxidation, creating graphene oxide (GO) and subsequent reduction of GO into graphene. Reduced GO sheets approach the highly desired properties of graphene, such as electrical conductivity and mechanical strength, to various degrees, but not completely. To understand why, we must understand the nanostructure of the sheets. Different methods of reduction result in products that are similar to graphene, but these products retain some oxidized areas or contain regions with sp$^{3}$ bonded carbon. The concentration and distribution of these defects on the reduced GO sheet affect the properties of the 2D material. Here, we have characterized the atomic structure of GO and reduced GO via high resolution transmission electron microscopy, electron diffraction, and electron energy loss spectroscopy. Spectroscopic data taken during thermal reduction of GO shows changes in the fine structure of carbon K-edge as the carbon changes from an oxidized form to elemental amorphous carbon to graphite like form, clearly delineating the process of reduction of GO to graphene. Products of several other reduction methods are also characterized revealing information on electronic environment surrounding carbon atoms, distribution of crystalline areas, and oxygen removal from GO.   

Atomic-scale mapping of cerium valence in ceria-zirconia-supported Pd model planar catalysts

Cerium-based oxides have long been regarded as an important class of catalyst support materials. It is also recognized that the interaction between precious metal and ceria-based support material enhances the reducibility of the ceria. The combination of scanning transmission electron microscopy (STEM) and electron energy loss spectroscopy (EELS) can provide an atomic-scale picture of the interaction between precious metal particles and their support material. In our work, aberration corrected STEM-EELS is used to study the valence of cerium in the vicinity of palladium nanoparticles supported on a ceria-zirconia (CZO) thin film. A monolayer-equivalent of Pd was deposited onto a 50nm-thick CZO thin film, which was then subjected to different thermal treatments. The EELS spectra extracted from the top several atomic layers of the CZO film exhibit typical 3+ character following a low-temperature reduction treatment, indicating the formation of oxygen vacancies. A variety of control experiments have also been performed to exclude possible artifacts caused by electron beam irradiation.   

Correlated Percolation Model of Graphene Hydrogenation

Hydrogenation of graphene by exposure to an atomic hydrogen plasma is a random process. However, the presence of hydrogen already attached to the plane increases the sticking probability of incoming adatoms. We simulate this process as a correlated percolation model where the hydrogen occupation probability of a carbon site is increased or decreased depending on the hydrogenation of the nearest neighboring carbon atoms. This enhancement modifies the cluster distribution on the surface and consequently the electronic structure of the system. We study these effects with a tight binding model and find that, although the density of states at the Fermi level is greatly increased by hydrogenation, the inverse participation ratio shows that not all of these states will contribute to conduction. In fact, for hydrogenation levels of greater than 40\% of the lattice, localized states begin to dominate at the Fermi level. A realistic set of values for the sticking probabilities is determined by analysis of the STM images of this system. Through the modeling of this mesoscopic process, we gain a better understanding of how chemically modified graphene is produced and what its transport properties are.   

Effect of charged impurities and morphology on oxidation reactivity of graphene

Chemical reactivity of single layer graphene supported on a substrate is observed to be enhanced over thicker graphene. Possible mechanisms for the enhancement are Fermi level fluctuations due to ionized impurities on the substrate, and structural deformation of graphene induced by coupling to the substrate geometry. Here, we study the substrate-dependent oxidation reactivity of graphene, employing various substrates such as SiO$_{2}$, mica, SiO$_{2}$ nanoparticle thin film, and hexagonal boron nitride, which exhibit different charged impurity concentrations and surface roughness. Graphene is prepared on each substrate via mechanical exfoliation and oxidized in Ar/O$_{2}$ mixture at temperatures from 400-600 $^{\circ}$C. After oxidation, the Raman spectrum of graphene is measured, and the Raman D to G peak ratio is used to quantify the density of point defects introduced by oxidation. We will discuss the correlations among the defect density in oxidized graphene, substrate charge inhomogeneity, substrate corrugations, and graphene layer thickness. This work has been supported by the University of Maryland NSF-MRSEC under Grant No. DMR 05-20471 with supplemental funding from NRI, and NSF-DMR 08-04976.   

Numerical Study of Carrier Multiplication in Nanocrystalline form of PbSe and PbS and the bulk

Using previously developed Exciton Scattering Model, we report on systematic numerical study of the carrier multiplication (CM) dynamics in spherically symmetric PbSe and PbS nanocrystals (NCs) and bulk. The quantum efficiency (QE) associated with the photogeneration and population relaxation processes are calculated. It is found that the photogeneration event provides small, about $5\%$, contribution to the total QE compared to the contribution from the population relaxation process. The analysis shows that the impact ionization dynamics is the main mechanism responsible for the CM during {\em both} the photogeneration and the population relaxation events. Furthermore, the calculated photogeneration and total QEs for various size NCs are found never to exceed the calculated limiting values for bulk. This observation is explained in terms of the quantum-confinement induced weak effective Coulomb enhancement whose contribution to the impact ionization rate is fully suppressed by the reduction in the biexciton density of states. We also find weak dependence of the total QE on the transform limited pump pulse duration. Comparison of the calculated QEs to published experimental data shows that our calculations well reproduce the experimentally observed trends.