Browsing by Author "Jerry L. Whitten, Committee Member"

Filter results by typing the first few letters
Now showing 1 - 5 of 5
  • No Thumbnail Available
    Ab Initio Electronic Structure Calculations For High-K Dielectrics
    (2005-01-04) Zhang, Yu; Harald Ade, Committee Member; Jerry L. Whitten, Committee Member; Dave Aspnes, Committee Member; Gerald Lucovsky, Committee Chair
    In current semiconductor industry, continuing improvement in the performance of MOSFET requires aggressive scaling down of the dimensions of CMOS devices. A better capacitance/unit area can be gained as gate oxide thickness decreases. An equivalent oxide thickness (EOT) less than 1.0nm is required according to the 2002 International Technology Roadmap for Semiconductor (ITRS). However, as gate oxide thickness scaling down, tunneling current will increase, which will lower the device performance. SiO2, as the widely used gate oxide material, has reached its scaling limit due to the high current leakage at this thickness. Non-crystalline alloys of i) group IIIB, IVB and VB TM oxides and ii) first row RE oxides with SiO2 and Al2O3 have been proposed as alternative high-k gate dielectrics for advanced Si devices. This dissertation addresses differences between the electronic structure of alternative high-k transition metal dielectrics and SiO2. Ab inito calculations, based on small clusters identify unique aspects of electronic structure that are associated with the TM atoms. The lowest conduction band states are derived from atomic d-states of the TM atoms, and are localized on these atoms. Excitations into these states i) from TM core states, ii) from oxygen K1, iii) from oxygen atom derived valence band states, are simulated by using ab inito calculations at self-consistant-field (SCF) Hartree-Fock and Configuration Interaction (CI) level. And these electronic structure calculations are used to interpret optical, ultra-violet (UV), X-ray and electron spectroscopies, including UV and X-ray photoemission (UPS and XPS, respectively), and Auger electron spectroscopy (AES), and also provide a basis for interpretation of electrical results and narrowing the field of possible replacement dielectrics for advanced semiconductor devices.
  • No Thumbnail Available
    Lipid-Protein Interactions by Advanced EPR Methods
    (2010-05-27) Chadwick, Thomas Gray; Tatyana I. Smirnova, Committee Chair; Alex I. Smirnov, Committee Member; Jerry L. Whitten, Committee Member; Alexander Deiters, Committee Member; Joseph C. Burns , Committee Member
    Biological membranes contain proteins that are responsible for many vital biological functions, and understanding the structure-function relationships between the two are fundamentally important. In this work we have utilized Electron Paramagnetic Resonance (EPR) spectroscopy to investigate the chemistry of the phospholipid biding protein Sec14p, and the insertion profiles of synthetic peptides corresponding to the E2 transmembrane domain of selected mutants of the Sindbis Virus. Sec14p is associated with the secretory pathway in the Gogli apparatus, and is thought to be involved in regulating membrane composition. It has been shown to bind phosphatidylcholine (PC) and phosphatidylinositol (PI) in vitro. We demonstrate that Sec14p binds spin labeled homologous of the PC lipid, and the corresponding EPR spectra provide information about the mobility of the bound lipid and the polarity within the binding pocket. Accessibility measurements also ascertain the orientation of the PC molecule within the binding pocket. The results show that the PC molecule is highly restricted inside the biding pocket; polarity and procticity decrease with distance from the polar head region, with increase polarity and procticity at the distal end of the sn-2 acyl chain. EPR data indicate that the molecule adopts a headgroup-out orientation and the polarity profile provides a hydrophobic matching necessary for Sec14p to extract a lipid from a membrane in a energy-independent mechanism. Sinbis virus infects both insects and mammal, and replication requires the incorporation of the host cell’s membrane into its structure. The physical properties of these membranes differ, and the transmembrane domains of the virus’s structural proteins must be able to assemble into the same structure in both membrane types. Virus mutants having truncated transmembrane domains exhibit differential ability to reproduce, where certain mutants favor the production of new virus in mammalian cells, while other mutants favor insect cells. We have investigated this differential infectivity of the Sindbis virus mutants by examining the insertion profiles of synthetic peptides STM16 and STM18 in insect and mammalian membrane mimics. These peptides correspond to the transmembrane segments of Sindbis virus mutants TM-16 and TM-18. The results indicate that while both peptides assume transmembrane orientation in the insect membrane mimic, addition of cholesterol affects peptide-membrane interactions in a cholesterol-concentration dependant manner.
  • No Thumbnail Available
    Liquid-Phase Deoxygenation of Free Fatty Acids to Hydrocarbons Using Supported Palladium Catalysts
    (2010-04-30) Immer, Jeremy Glen; Jerry L. Whitten, Committee Member; Steven W. Peretti, Committee Member; Jan Genzer, Committee Member; H. Henry Lamb, Committee Chair
    Hydrocarbon biofuels that are drop-in replacements for traditional petroleum-derived liquid fuels can be produced from edible and inedible fats and oils (triglyceride sources) via thermocatalytic processes. Liquid-phase deoxygenation of stearic acid (SA) in dodecane at 300°C and 15 atm was employed to screen supported noble metal catalysts for decarboxylation of free fatty acids to hydrocarbons. Commercial samples of Pt/C, Pd/C (4), Pd/Al2O3, and Pd/SiO2 catalysts and an in-house prepared Pd/SiO2 catalyst (each containing 5 wt.% metal) were screened under flowing 0, 5, and 10% H2 (balance He). Under flowing He, most of the catalysts studied failed to achieve 100% SA conversion after 4 h under reaction conditions due to rapid deactivation. The exception was a uniformly impregnated Pd/C catalyst that gave >99% conversion in ~1 h with 99% CO2 selectivity. All of the catalysts were far more stable under H2 yielding nearly complete SA conversion after 4 h; however, they differed markedly in their CO2 selectivities. Pd/SiO2 and Pt/C catalysts were selective toward decarbonylation (CO production), and Pd/C and Pd/Al2O3 catalysts were selective toward decarboxylation. Even under H2, the uniformly impregnated Pd/C catalyst was the most active and selective for the hydrogen-neutral decarboxylation pathway. Semi-batch deoxygenation of SA employing this 5 wt.% Pd/C catalyst was investigated further using on-line quadrupole mass spectrometry. With fresh catalyst, SA deoxygenation under He occurred rapidly with very high CO2 selectivity; however, reuse of the catalyst showed an orders of magnitude loss of decarboxylation activity and high decarbonylation selectivity. Experiments employing smaller amounts of fresh catalyst evidenced that decarboxylation activity under He is limited to ~220 turnovers. Attempts to reactivate the used Pd/C catalyst by H2 treatment were only modestly effective. Increased catalyst lifetime (>2200 turnovers) was achieved by employing a H2-containing purge gas; however, the decarboxylation rate decreases with increasing H2 partial pressure resulting in lower CO2 selectivity. Increasing the initial SA concentration also inhibited decarboxylation, substantially prolonging the batch time and yielding lower overall CO2 selectivity. The origin of this effect was traced to catalyst poisoning by endogenous CO from the decarbonylation pathway. Catalyst poisoning experiments demonstrated that CO strongly inhibits the decarboxylation pathway and that the inhibitory effects of CO and H2 are additive. Under conditions of strong decarboxylation inhibition, the decarbonylation rate was unaffected, and we infer that decarboxylation occurs over different catalytic sites than decarbonylation. An elementary reaction sequence for Pd-catalyzed decarboxylation is proposed which accounts for our observations. Fed-batch deoxygenation of SA and oleic acid was demonstrated in a 50-mL stirred autoclave reactor with continuous feeding for run times up to 24 h. The maximum quasi-steady state decarboxylation rate observed under 5% H2 was 0.43 mmol/gcat•min (0.078 s-1 turnover frequency). When higher H2 partial pressures were employed, an abrupt switchover in product selectivity from CO2 to CO was observed. Higher CO selectivity leads to increased H2 consumption due to hydrogenation of heptadecene, the primary product of the decarbonylation pathway. The on-stream time at which this switchover occurs was found to increase with decreasing H2 pressure. We infer that the switchover phenomenon arises from H2 inhibition of the decarboxylation pathway resulting in SA accumulation. SA accumulation increases the decarbonylation rate leading to further inhibition of the decarboxylation pathway by endogenous CO. Parametric studies involving SA feed rate, H2 partial pressure and exogenous CO partial pressure support the proposed switchover mechanism. Inhibition of decarboxylation activity was reversible at least in the short term by lowering the H2 or CO partial pressure or stopping SA injection; however, if a catalyst was aged >10 h under reaction conditions favoring decarbonylation, decarboxylation activity could not be recovered.
  • No Thumbnail Available
    Nucleation and Growth of Defects in Nitrogen doped Silicon
    (2004-12-31) Karoui, Abdennaceur; Carl Koch, Committee Co-Chair; George A. Rozgonyi, Committee Chair; THOMAS BRENT GUNNOE, Committee Member; JAGANNADHAM KASICHAINULA, Committee Member; Jerry L. Whitten, Committee Member
    Ultra high purity silicon is advantageously modified by as low as 5x10¹⁴ cm⁻³ nitrogen. Such a doping level was proved to drastically impact grown-in and process-induced defects, enhances the denuded zone intended for making devices, improves impurity gettering, and increases the gate oxide integrity in metal oxide silicon devices. Interestingly, with such a low nitrogen level wafer toughness is significantly increased. However, nitrogen doping alters standard wafer heat treatment processes through the modification of the early stages of point defect clustering dynamics. In this thesis, the basic interactions of light element impurities, particularly N and O, with point defects and crystal defects in silicon are scrutinized in order to understand the mechanisms of extended defect nucleation and growth in N doped silicon. Experimental data are used with molecular dynamics and quantum mechanics calculations for modeling defect formation. Various thermal annealing have been utilized to produce diverse conditions for defect interactions. Defect type, size distribution, nanoscale and atomic structure, and composition have been determined with emphasis on the depth dependence. Nanoscale analysis of defects probed at different depths allowed to build models of point defect dynamics from the extended defect formation history. Defect nucleation during crystal growth was qualitatively discussed and defect precursors were mapped on the crystal hot zones showing point defect clustering stages during solidification. This was based on results from the atomistic modeling of atomic structure of chemical complex ground states, the thermodynamic parameters close to the melting point, and the adsorption/desorption of point defects by stable chemical complexes. It was found that N₂ is a stable mobile species, VN₂ is an active metastable complex, and V₂N₂ is an immobile stable nucleus for oxygen precipitation but not for vacancy clustering. The formation energy of VN₂ was found positive by DFT calculations, which negates the spontaneous formation of isolated complex. However, the formation energy is reduced to about k[subscript B]T/2 near the melting point by coupling to one oxygen atom, which activates the formation of VN₂, while weakly bound to the oxygen. The calculated thermal stability of a wide range of prominent chemical complexes was cross-checked with the signature of experimentally proven viable ones. Furthermore, IR absorption line intensities in annealed wafers were obtained as a function of depth by high spatial resolution synchrotron FTIR, which allowed having N-V-O depth profiles. These appeared in good agreement with that of the oxynitride precipitate profiles by OPP and SIMS. Such an agreement represents a strong support for both chemical complex spectroscopic identification and calculated thermodynamic parameters. At the macroscopic level, nitrogen appeared to slowly athermally segregate under compressive stresses to dislocations and wafer surface; the segregation is accelerated at high temperature. In addition, nitrogen was found to couple with oxygen to form oxynitride zones and it segregates to precipitate interfaces making N-rich shells. Finally, silicon mechanical properties measured by nanoindentation of a variety of substrates appeared to well correlate with the dislocation pinning by light elements such as N, O, and C. The locking mechanism was correlated to dislocation interaction with impurity atmospheres simulated using elasticity theory, the size effect model, and Tersoff inter-atomic potential.
  • No Thumbnail Available
    The Structure and Mechanism of Melt Crystallization of CZX-1, a Templated ZnCl2 Network Material
    (2009-07-10) Josey, Amanda Ann; Paul A. Maggard, Committee Member; Mike H. Whangbo, Committee Member; Jerry L. Whitten, Committee Member; James D. Martin, Committee Chair
    A glassforming templated derivative of ZnCl2, called CZX-1, is melt synthesized by combining stoichiometric amounts of ZnCl2, CuCl and (CH3)3NHCl in a 5:1:1 ratio. CZX-1 consists of templated sodalite cage structures with the empirical formula of CuZn5Cl121-. A cage-template charge interaction occurs where templates ((CH3)3NH1+) are positioned at one of four tetrahedrally distributed sites within a sodalite cage. Molten CZX-1 shows significant extended-range order including structural features up to 50 Å long. This network organization in molten CZX-1 does not allow for nanoparticle formation during melt crystallization. CZX-1 exhibits several temperature-induced phase transitions. Upon heating above 164 oC the templates disorder, forcing a phase transition of α-CZX-1 to β-CZX-1 at 170 oC. A peritectic point is observed at 175 oC where CZX-1 disproportionates into 20 % β-ZnCl2 and 80 % template- and copper(I)-rich liquid. β-ZnCl2 dissolves forming an isotropic liquid at 205 oC. Rapid quenching of the isotropic melt to below 164 oC produces α-CZX-1, while slow cooling produces β-ZnCl2 and the template- and copper(I)-rich liquid below 205 oC. The disproportionation of CZX-1 into β-ZnC2 and liquid is reversed if the material is rapidly cycled from below 164 oC to the isotropic melt. Time-resolved X-ray diffraction experiments are exploited to independently measure the rates of nucleation and crystal growth. These experiments clearly indicate that condensed-phase crystallization is not explained by Classical Nucleation Theory. Rather, melt crystallization kinetics fit a new model treating CZX-1 nucleation and growth separately as a serial process. Three types of nucleation are observed, (1) intrinsic: the inherent nucleation of CZX-1, (2) extrinsic: nucleation is influenced by outside parameters like defects or internal domains, and (3) growth-initiated: continuous nucleation that follows the rate of growth. The rate of growth increases with increasing temperature to a critical temperature of 164 oC where template disordering forces the rate of growth to drop dramatically - demonstrating that the respective structure of the melt and crystalline phases exhibit a strong controlling influence on the crystallization mechanism.