Search

We present calculations of the nucleation barrier during crystallization in binary hard sphere mixtures under moderate degrees of supercooling using Monte Carlo simulations in the isothermal-isobaric semigrand ensemble in conjunction with an umbrella sampling technique. We study both additive and negatively nonadditive binary hard sphere systems. The solid-fluid phase diagrams of such systems show a rich variety of behavior, ranging from simple spindle shapes to the appearance of azeotropes and eutectics to the appearance of substitutionally ordered solid phase compounds. We investigate the effect of these types of phase behavior upon the nucleation barrier and the structure of the critical nucleus. We find that the underlying phase diagram has a significant effect on the mechanism of crystal nucleation. Our calculations indicate that fractionation of the species upon crystallization increases the difficulty of crystallization of fluid mixtures and in the absence of fractionation (azeotropic conditions) the nucleation barrier is comparable to pure fluids. We also calculate the barrier to nucleation of a substitutionally ordered compound solid. In such systems, which also show solid-solid phase separation, we find that the phase that nucleates is the one whose equilibrium composition is closer to the composition of the fluid phase. [ABSTRACT FROM AUTHOR]

Objective: Heatlhcare institutions are establishing frameworks to govern and promote the implementation of accurate, actionable, and reliable machine learning models that integrate with clinical workflow. Such governance frameworks require an accompanying technical framework to deploy models in a resource efficient, safe and high-quality manner. Here we present DEPLOYR, a technical framework for enabling real-time deployment and monitoring of researcher-created models into a widely used electronic medical record system., Materials and Methods: We discuss core functionality and design decisions, including mechanisms to trigger inference based on actions within electronic medical record software, modules that collect real-time data to make inferences, mechanisms that close-the-loop by displaying inferences back to end-users within their workflow, monitoring modules that track performance of deployed models over time, silent deployment capabilities, and mechanisms to prospectively evaluate a deployed model's impact., Results: We demonstrate the use of DEPLOYR by silently deploying and prospectively evaluating 12 machine learning models trained using electronic medical record data that predict laboratory diagnostic results, triggered by clinician button-clicks in Stanford Health Care's electronic medical record., Discussion: Our study highlights the need and feasibility for such silent deployment, because prospectively measured performance varies from retrospective estimates. When possible, we recommend using prospectively estimated performance measures during silent trials to make final go decisions for model deployment., Conclusion: Machine learning applications in healthcare are extensively researched, but successful translations to the bedside are rare. By describing DEPLOYR, we aim to inform machine learning deployment best practices and help bridge the model implementation gap., (© The Author(s) 2023. Published by Oxford University Press on behalf of the American Medical Informatics Association.)

The adsorption of CO2 and CH4 in a mixed-ligand metal-organic framework (MOF) Zn 2(NDC) 2(DPNI) [NDC = 2,6-naphthalenedicarboxylate, DPNI = N, N'-di-(4-pyridyl)-1,4,5,8-naphthalene tetracarboxydiimide] was investigated using volumetric adsorption measurements and grand canonical Monte Carlo (GCMC) simulations. The MOF was synthesized by two routes: first at 80 degrees C for two days with conventional heating, and second at 120 degrees C for 1 h using microwave heating. The two as-synthesized samples exhibit very similar powder X-ray diffraction patterns, but the evacuated samples show differences in nitrogen uptake. From the single-component CO2 and CH4 isotherms, mixture adsorption was predicted using the ideal adsorbed solution theory (IAST). The microwave sample shows a selectivity of approximately 30 for CO2 over CH4, which is among the highest selectivities reported for this separation. The applicability of IAST to this system was demonstrated by performing GCMC simulations for both single-component and mixture adsorption.

We present density-functional theory (DFT) calculations for critical cavities inside model superheated liquids with varying intermolecular potentials. Our calculations show that the radius of the critical cavity and the ratio of the work of formation of the critical cavity to the work of formation of the critical bubble as predicted by the classical nucleation theory exhibit universal scaling across similar intermolecular potentials. We then utilize this observed scaling behavior by proposing two new criteria for the kinetic spinodal of superheated liquids. These criteria are based on various properties of the critical cavity as obtained from our DFT studies of the superheated Lenanrd-Jones liquid. Our predictions of the kinetic spinodal compare favorably with experimental data of the limits of superheating for various organic liquids., ((c) 2004 American Institute of Physics.)

A semiempirical approach, based on both the scaled particle theory of hard particle fluids and free-energy perturbation methods, that predicts the work of formation of a cavity inside a model fluid is presented. The method is tested using the pure component Lennard-Jones fluid. A good agreement between values obtained via the theory and molecular simulation is observed even as the size of the cavity becomes larger than the effective diameter of Lennard-Jones particles. The method also yields reasonably accurate estimates when the pressure of the liquid is quite low, which is the case for liquids close to the coexistence line, or when the pressure is negative, which is the case when the liquid is metastable.