Agent-Based Modeling of Nanoparticles at Gas-Liquid Interface with Emphasis on Environmental Systems

Abstract

The main objective of this thesis is to investigate the effects of nanoparticles on surface tension in the presence of SDS at gas-liquid interface. This thesis also aims to model the behavior of same-charged and uncharged nanoparticles as well as SDS molecules with agent-based modeling approaches to simulate their effects on surface tension. To investigate the surface tension and its dependency on the nanoparticles and SDS, various pendant drop surface tension experiments conducted for different types of nanoparticles, SDS concentrations, and solvents at Hacettepe University Advanced Technologies Application and Research Center. The experimental results revealed that both nanoparticles and SDS have the ability to decrease the surface tension of solvents. These experimental results used as reference for the agent-based models to computationally mimic the behavior of surface tension for selected SDS concentrations varying between 0.1×CMC to 10×CMC. At the model development stage, agent schema developed at first and it has been used to create the agent-based models to picture the behavior of surface tension at gas-liquid interface with the aid of Waterfall Software Development Methodology. All agents defined as sets of attributes consisting real life counterparts of nanoparticles, SDS, and interface. Agent-based models designed to simulate the agents to reside in 12 different solvents with density values varying between 0.49 g/ml to 1.45 g/ml. Agent-based models designed to allocate the micelle formation under certain conditions. Throughout the thesis, agent-based models implemented in NetLogo to run on both Linux and Windows based systems. Models have been subjected to validation, verification, sensitivity, and load tests prior to their deployment. Each test executed twice for 5×103, 7.5×103, 10×103, 12.5×103, 15×103, 50×103, and 75×103 same-charged and uncharged nanoparticles as well as SDS molecules. Load tests exposed that the computer system used in this thesis is capable of implementing models with less than 200×103 agents. The analysis of experimental results from HUATARC and agent-based models showed that agent-based models are capable of predicting the steady state of the surface tension. The relation between as average # nanoparticles at the interface and measured surface tension values are non-linear and this condition cannot be defined by normal distribution. Spearman and Kendall correlation analysis applied and their coefficients for the mentioned variables have been calculated as -0.999 and -0.995, respectively. Time correlation analysis for surface tension experiments and agent-based models has been done for the selected models. The results showed that each tick corresponds to 2.49 ms. In addition to above studies, this thesis intended to analyze and identify the effects of nanoparticles and SDS to surface tension by developing a simple, practical, and low cost measurement technique. The experimental results of the technique intended to be used in the implementation of artificial neural network (ANN) models for simulating the behavior of surface tension under certain conditions. Since, the total number of data from the experiments found to be insufficient for ANN models, this part of the study put on hold.