David R. Ely

Validity of the Bruggeman relation for porous electrodes

The ability to engineer electrode microstructures to increase power and energy densities is critical to the development of high-energy density lithium-ion batteries. Because high tortuosities in porous electrodes are linked to lower delivered energy and power densities, in this paper, we experimentally and computationally study tortuosity and consider possible approaches to decrease it. We investigate the effect of electrode processing on the tortuosity of in-house fabricated porous electrodes, using three-dimensionally reconstructed microstructures obtained by synchrotron x-ray tomography. Computer-generated electrodes are used to understand the experimental findings and assess the impact of particle size distribution and particle packing on tortuosity and reactive area density. We highlight the limitations and tradeoffs of reducing tortuosity and develop a practical set of guidelines for active material manufacture and electrode preparation.

Improvement of the dissolution rate of poorly soluble drugs by solid crystal suspensions.

We present a novel extrusion based approach where the dissolution rate of poorly soluble drugs (griseofulvin, phenytoin and spironolactone) is significantly accelerated. The drug and highly soluble mannitol are coprocessed in a hot melt extrusion operation. The obtained product is an intimate mixture of the crystalline drug and crystalline excipient, with up to 50% (w/w) drug load. The in vitro drug release from the obtained solid crystalline suspensions is over 2 orders of magnitude faster than that of the pure drug. Since the resulting product is crystalline, the accelerated dissolution rate does not bear the physical stability concerns inherent to amorphous formulations. This approach is useful in situations where the drug is not a good glass former or in cases where it is difficult to stabilize the amorphous drug. Being thermodynamically stable, the dissolution profile and the solid state properties of the product are maintained after storage at 40 °C, 75% RH for at least 90 days.

Residence time modeling of hot melt extrusion processes

The hot melt extrusion process is a widespread technique to mix viscous melts. The residence time of material in the process frequently determines the product properties. An experimental setup and a corresponding mathematical model were developed to evaluate residence time and residence time distribution in twin screw extrusion processes. The extrusion process was modeled as the convolution of a mass transport process described by a Gaussian probability function, and a mixing process represented by an exponential function. The residence time of the extrusion process was determined by introducing a tracer at the extruder inlet and measuring the tracer concentration at the die. These concentrations were fitted to the residence time model, and an adequate correlation was found. Different parameters were derived to characterize the extrusion process including the dead time, the apparent mixing volume, and a transport related axial mixing. A 2(3) design of experiments was performed to evaluate the effect of powder feed rate, screw speed, and melt viscosity of the material on the residence time. All three parameters affect the residence time of material in the extruder. In conclusion, a residence time model was developed to interpret experimental data and to get insights into the hot melt extrusion process.

From Process to Modules: End-to-End Modeling of CSS-Deposited CdTe Solar Cells

In this paper, we develop an end-to-end modeling framework to explore how various multiscale phenomena in solar cells translate from materials to module level. Specifically, the model captures the physics related to 1) the pressure-dependent grain growth of polycrystalline thin films (nanometers to micrometers), 2) averaging of the effects of grain-size distribution at the centimeter scale, and 3) effects of parasitic series and shunt resistance distributions on the efficiency of thin-film solar cell modules (centimeter to meter scale). As an idealized illustrative example, we consider a number of puzzling features that are associated with close space sublimated CdTe solar cells. The model explains both the increase in the grain size with deposition pressure, as well as the saturation of cell efficiency beyond a critical grain size. The analysis shows that grain geometry and grain-size distribution are unimportant for average grain sizes larger than 1 μm. The model attributes the significant efficiency loss at the module level to the series resistance and the operating point inhomogeneity caused by parasitic shunts. Overall, the model identifies opportunities for significant improvement at all length scales of thin-film solar cell technologies.

Determination of the scale of segregation of low dose tablets using hyperspectral imaging

In this work, near infrared (NIR) hyperspectral imaging was used to quantify the spatial distribution of drug in tablets containing tolmetin sodium dihydrate. Hyperspectral data cubes were generated by imaging the same spatial region of a sample while illuminated by a laser at a different wavelength for each image. Images were generated for wavelengths ranging from 1100 to 2200 nm. Ten tablets with concentrations ranging from 0.0 to 10.0% w/w tolmetin were imaged, and the scales of segregation were calculated for the tablets. Lactose anhydrous was used as the diluent, and all mixtures contained 0.5% magnesium stearate as a lubricant. This research has shown hyperspectral imaging to be viable tool for quantifying segregation of low dose drugs in tablets.

The Virtual Kinetics of Materials Laboratory

web interface to develop, modify, and execute FiPy, Gibbs, and other python-based applications

Progress towards modeling microstructure evolution in polycrystalline films for solar cell applications

Grain morphology has been long considered to be a major factor in the performance and efficiency of photovoltaic devices. Experimental work has demonstrated the effect of average grain size and surface roughness on the control of conversion efficiency. But its statistics and kinetics are not well understood. While much theoretical progress has been made for α-Si by various effective media and Monte Carlo models, the theory of polycrystalline material growth remains primitive. In this paper, a generalized 2D model was developed to predict cell microstructure under fabrication conditions. Growth of Cadmium Telluride was chosen due to its champion efficiency. The current model combines the numerical efficiency and elegance of the level set method to track the macroscopic interface and includes the simplicity of the phase field method to track the natural microstructure evolution for a non-conserved parameter.