Person:

Aziz, Michael

A review of the literature has found a factor of 4 spread in the estimated values of the energy penalty for post-combustion capture and storage of CO2 from pulverized-coal (PC) fired power plants. We elucidate the cause of that spread by deriving an analytic relationship for the energy penalty from thermodynamic principles and by identifying which variables are most difficult to constrain. We define the energy penalty for CCS to be the fraction of fuel that must be dedicated to CCS for a fixed quantity of work output. That penalty can manifest itself as either the additional fuel required to maintain a power plant's output or the loss of output for a constant fuel input. Of the 17 parameters that constitute the energy penalty, only the fraction of available waste heat that is recovered for use and the 2nd-law separation efficiency are poorly constrained. We provide an absolute lower bound for the energy penalty of ~11%, and we demonstrate to what degree increasing the fraction of available-waste-heat recovery can reduce the energy penalty from the higher values reported. It is further argued that an energy penalty of ~40% will be easily achieved while one of ~29% represents a decent target value. Furthermore, we analyze the distribution of PC plants in the U.S. and calculate a distribution for the additional fuel required to operate all these plants with CO2 capture and storage (CCS).

During rapid solidification of intermetallic compounds, the atoms may not have time to find the lowest-energy sites in the crystal, resulting in the growth of a solid with partially or completely suppressed chemical order. A kinetic model has been developed for “disorder trapping” during rapid solidification, which predicts the long range order parameter, composition and temperature at the interface of a chemically ordered phase as functions of interface velocity and liquid composition. The model predicts that as the solidification velocity increases, the long range order parameter decreases. Beyond a critical velocity the order parameter is zero. The predicted transition to solidification of a disordered solid is discontinuous for thermodynamically first-order cases, and continuous for thermodynamically second-order cases.

Ni2TiAl (L21) and Ni3AI (L12), which in equilibrium are ordered to their melting points, have been kinetically disordered by rapid solidification following pulsed laser melting. The order-disorder transitions are second-order and first-order, respectively. The interface velocity and order parameter can be followed with nanosecond resolution by monitoring the reflectivity and lateral resistance of a thin film sample during and immediately after solidification. X-ray diffraction and Transmission Electron Microscopy (TEM) indicate that metastable fcc Ni3Al was retained by quenching. In rapidly solidified bulk Ni2TiAl TEM reveals a fine array of antiphase boundaries, indicating that the material formed from the melt with the unstable B2 structure and subsequently transformed to the equilibrium Heusler structure during cooling to room temperature. In rapidly solidified Ni2TiAl thin films, electron diffraction indicates that the unstable bcc structure was formed directly from the melt and retained upon cooling.

The velocity dependence of the partition coefficient was measured for rapid solidification of polycrystalline Si-4.5 at% As and Si-9 at% As alloys induced by pulsed laser melting. The results constitute the first test of partitioning models both for the high velocity regime and for non-dilute alloys. The continuous growth model (CGM) of Aziz and Kaplan fits the data well, but with an unusually low diffusive speed of 0.46 m/s. The data show negligible dependence of partitioning on concentration, also consistent with the CGM. The predictions of the Hillert-Sundman model are inconsistent with partitioning results. Using the aperiodic stepwise growth model (ASGM) of Goldman and Aziz, an average over crystallographic orientations with parameters from independent single-crystal experiments is shown to be reasonably consistent with these polycrystalline partitioning results. The results, combined with others, indicate that the CGM without solute drag and its extension to lateral ledge motion, the ASGM, are the only models that fit the data for both solute partioning and kinetic undercooling interface response functions. No current solute drag models can match both partitioning and undercooling measurements.

Measurements of liquid-phase diffusion coefficients for dilute tungsten and molybdenum in molten nickel were made using a pulsed ion-beam melting technique. A high-intensity beam of nitrogen ions is focused on the surface of a nickel substrate that was implanted with known concentration profiles of W and Mo. Melting of the surface to a depth of 1 m allows broadening of the implant profiles while molten. Solute concentration-depth profiles were determined before and after melting using Rutherford backscattering spectrometry. Using a series of numerical simulations to estimate the melt history and diffusional broadening for mean liquid temperatures in the range 1755 to 2022 K, an effective diffusion coefficient is determined in each case by comparison to the measured depth profiles. This is found to be (2.4 0.2) 10^-5 cm²/s for W and (1.6 0.4) 10^-5 cm2/s for Mo, with an additional systematic uncertainty of 0.5 10 5 due to instrumental and surface effects.

We have conducted the first experiments under identical thermal, background, and surface preparation conditions to compare metal-on-insulator growth morphology in Pulsed Laser Deposition (PLD) and Physical Vapor Deposition (PVD). Such films deposited from a thermal vapor are known to exhibit a characteristic morphological progression beginning with isolated three-dimensional islands and ending with a percolating, continuous film that conducts electrically. We have studied this progression for Pulsed Laser Deposition, a technique that differs from PVD in that depositing species arrive in short bursts (<10 microseconds) with kinetic energy typically of 10-100 eV. Kinetic Monte Carlo (KMC) simulations that take into account only the pulsed nature of the flux predict that PLD films should advance to percolation with relatively less deposition compared with PVD under otherwise identical conditions. Our experiments, with PLD and PVD performed in the same chamber, reveal that PLD films actually require more deposition to reach percolation. We conclude that energetic effects are important in determining morphology evolution.

The effects of externally applied stress and rate-enhancing dopants on interfacial roughness during the solid phase epitaxial growth of ion-implantation-doped Si are investigated using cross-sectional transmission electron microscopy and time-resolved reflectivity. We find long-wavelength roughness in the absence of an applied stress that arises solely from the dopant-gradient. With the addition of a compressive stress, the interface roughens further with an enhanced magnitude and a dramatically reduced wavelength. We discuss the experimental results in the context of a simulation that includes our current understanding of stress, dopant-gradient, and interface anisotropy effects. We find a rich interplay between these effects in determining growth morphology evolution, and demonstrate the successes and current limitations of the model.

During rapid solidification, kinetically suppressed solute partitioning at the crystal/melt interface, as well as kinetic interfacial undercooling, become important. Both of these effects have significant stabilizing influences on a planar interface during rapid solidification. We review experimental tests we have performed of models for the transition from planar to cellular growth, and for the velocity-undercooling function of the dendrite tip, in the velocity regime where nonequilibrium interface kinetics are important.

We present a model of the temperature dependence of point-defect-mediated luminescence in silicon derived from basic kinetics and semiconductor physics and based on the kinetics of bound exciton formation. The model provides a good fit to data for W line electroluminescence and G line photoluminescence in silicon. Strategies are discussed for extending luminescence to room temperature.

No abstract

Experimental results relevant to models for nonequilibrium interface kinetics during rapid solidification are reviewed. Models are examined critically in light of these experiments. The kinetic Ising model is shown to compare unfavorably with experiment. The Continuous Growth Model without solute drag and its extension to non-(001) interfaces, the Aperiodic Stepwise Growth Model, account well for all relevant experimental results.