Bicomponent aggregation in finite systems

Themis Matsoukas, C. L. Marshall

Research output: Contribution to journalArticle

7 Citations (Scopus)

Abstract

The mixing state of a bicomponent population of clusters (granules) is characterized by the normalized variance of excess solute, χ, a parameter that measures the deviation of the composition of each granule from the overall mean. For certain initial conditions and types of kernels, this parameter is constant provided that the population is infinite in size. Here we examine the parameter χ in finite populations and we find that it always decreases. We then apply the analysis to Monte Carlo algorithms that use finite populations to simulate aggregation in infinite systems and find that, while increasing the number of Monte Carlo particles converges the solution to the result for the infinite system, this convergence depends on the degree of homogeneity. In particular, as the degree of homogeneity of non-gelling kernels approaches 1, the number of Monte Carlo particles required to maintain acceptable accuracy with respect to χ in the scaling limit may be prohibitively high.

Original languageEnglish (US)
Article number46007
JournalEurophysics Letters
Volume92
Issue number4
DOIs
StatePublished - Nov 1 2010

Fingerprint

homogeneity
solutes
deviation
scaling

All Science Journal Classification (ASJC) codes

  • Physics and Astronomy(all)

Cite this

@article{9d7a0a849a9e48f69701384baa97d105,
title = "Bicomponent aggregation in finite systems",
abstract = "The mixing state of a bicomponent population of clusters (granules) is characterized by the normalized variance of excess solute, χ, a parameter that measures the deviation of the composition of each granule from the overall mean. For certain initial conditions and types of kernels, this parameter is constant provided that the population is infinite in size. Here we examine the parameter χ in finite populations and we find that it always decreases. We then apply the analysis to Monte Carlo algorithms that use finite populations to simulate aggregation in infinite systems and find that, while increasing the number of Monte Carlo particles converges the solution to the result for the infinite system, this convergence depends on the degree of homogeneity. In particular, as the degree of homogeneity of non-gelling kernels approaches 1, the number of Monte Carlo particles required to maintain acceptable accuracy with respect to χ in the scaling limit may be prohibitively high.",
author = "Themis Matsoukas and Marshall, {C. L.}",
year = "2010",
month = "11",
day = "1",
doi = "10.1209/0295-5075/92/46007",
language = "English (US)",
volume = "92",
journal = "Europhysics Letters",
issn = "0295-5075",
publisher = "IOP Publishing Ltd.",
number = "4",

}

Bicomponent aggregation in finite systems. / Matsoukas, Themis; Marshall, C. L.

In: Europhysics Letters, Vol. 92, No. 4, 46007, 01.11.2010.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Bicomponent aggregation in finite systems

AU - Matsoukas, Themis

AU - Marshall, C. L.

PY - 2010/11/1

Y1 - 2010/11/1

N2 - The mixing state of a bicomponent population of clusters (granules) is characterized by the normalized variance of excess solute, χ, a parameter that measures the deviation of the composition of each granule from the overall mean. For certain initial conditions and types of kernels, this parameter is constant provided that the population is infinite in size. Here we examine the parameter χ in finite populations and we find that it always decreases. We then apply the analysis to Monte Carlo algorithms that use finite populations to simulate aggregation in infinite systems and find that, while increasing the number of Monte Carlo particles converges the solution to the result for the infinite system, this convergence depends on the degree of homogeneity. In particular, as the degree of homogeneity of non-gelling kernels approaches 1, the number of Monte Carlo particles required to maintain acceptable accuracy with respect to χ in the scaling limit may be prohibitively high.

AB - The mixing state of a bicomponent population of clusters (granules) is characterized by the normalized variance of excess solute, χ, a parameter that measures the deviation of the composition of each granule from the overall mean. For certain initial conditions and types of kernels, this parameter is constant provided that the population is infinite in size. Here we examine the parameter χ in finite populations and we find that it always decreases. We then apply the analysis to Monte Carlo algorithms that use finite populations to simulate aggregation in infinite systems and find that, while increasing the number of Monte Carlo particles converges the solution to the result for the infinite system, this convergence depends on the degree of homogeneity. In particular, as the degree of homogeneity of non-gelling kernels approaches 1, the number of Monte Carlo particles required to maintain acceptable accuracy with respect to χ in the scaling limit may be prohibitively high.

UR - http://www.scopus.com/inward/record.url?scp=78650391362&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=78650391362&partnerID=8YFLogxK

U2 - 10.1209/0295-5075/92/46007

DO - 10.1209/0295-5075/92/46007

M3 - Article

AN - SCOPUS:78650391362

VL - 92

JO - Europhysics Letters

JF - Europhysics Letters

SN - 0295-5075

IS - 4

M1 - 46007

ER -