Search

It is traditionally assumed that the pair energies of atoms on the surface are site independent and equal to the pair energy in the bulk of the crystal (constant bond energy approach). We have assumed that the internal energy of an atom, ψ, depends nonlinearly on the number of nearest neighbors n in the form ψ(n)=-(an+bn2), where a and b are material constants that can be evaluated from experimentally determined heats of sublimation and energies of bulk vacancy formation. This simple approximation of the actual many-body interactions accommodates site-dependent pair energies. Based on this phenomenological ‘‘variable bond model’’ and the Bragg–Williams approximation, occupation fractions of an equilibrium surface are obtained in two-level and three-level models. These provide a criterion of surface roughness for specific orientations. Application of the criterion to copper, lead, and zinc yields good agreement with experimental observations of anisotropic melting, while the traditional constant bond approach fails. It is shown that in addition to these systems many other metals possess negative b values, i.e., exhibit ‘‘bond strengthening’’ on surfaces. Similar conditions are shown to exist in numerous ionic, covalent, and van der Waals solids. Hence, one must expect that constant bond models underestimate the surface roughness for a large group of materials. [ABSTRACT FROM AUTHOR]

Background: Soil organic carbon (SOC) and carbon (C) functional groups in different particle-size fractions are important indicators of microbial activity and soil decomposition stages under wildfire disturbances. This research investigated a natural Tsuga forest and a nearby fire-induced grassland along a sampling transect in Central Taiwan with the aim to better understand the effect of forest wildfires on the change of SOC in different soil particle scales. Soil samples were separated into six particle sizes and SOC was characterized by solid-state C nuclear magnetic resonance spectroscopy in each fraction. Results: The SOC content was higher in forest than grassland soil in the particle-size fraction samples. The O-alkyl-C content (carbohydrate-derived structures) was higher in the grassland than the forest soils, but the alkyl-C content (recalcitrant substances) was higher in forest than grassland soils, for a higher humification degree (alkyl-C/O-alkyl-C ratio) in forest soils for all the soil particle-size fractions. Conclusions: High humification degree was found in forest soils. The similar aromaticity between forest and grassland soils might be attributed to the fire-induced aromatic-C content in the grassland that offsets the original difference between the forest and grassland. High alkyl-C content and humification degree and low C/N ratios in the fine particle-size fractions implied that undecomposed recalcitrant substances tended to accumulate in the fine fractions of soils. [ABSTRACT FROM AUTHOR]

A new general equation has been developed for the lag time for a dilute diffusant permeating through a one-dimensional membrane in which both diffusivity D and partition coefficient K are dependent on position. The lag time is given by tL=∫h0K(x)[∫x0 R(y)dy][∫hxR(y)dy] ×dx/ ∫h0R(x)dx, where h is the membrane thickness and R(x)=1/[D(x)K(x)]. Formulas derived from this equation include those for: lag time in the presence of an electrical potential gradient; lag time in a multilaminar periodic membrane; and an upper bound for lag time given by tL <(h2/ 4Deff), where Deff is the effective diffusivity as deduced from steady-state partitioning and permeability measurements. The lag time is shown to be independent of the direction of diffusion. A computational technique is presented in which the equation for tL is rearranged to a system of two differential equations, which are readily solvable by numerical integration, even when the system has discontinuities in D and K which would increase the difficulty of evaluating the double integral directly. [ABSTRACT FROM AUTHOR]