Transport Behavior Of Endocrine Disrupting Chemicals In Sandy Sediment

Beside the process of biodegradation, sorption is fundamental in the control of transformation, fate, degradation, and biological activity of hydrophobic organic contaminants in the environment during the subsurface flow. Soil and sediments are a heterogeneous mixture of numerous components within which clays and organic matter have the ability to sorb organic species. Furthermore,the degradation of Organic compound in the subsurface zones differs according to the concerning area. In the saturated zone, a plume of organic pollutant may become more persistent than in the Vadoze zone due its lower oxygen content. So, the understanding of the reactivity of simple organic compounds on soil/sediment minerals surfaces can help prevent groundwater contamination.This research aims to investigate the transport behavior of endocrine disrupting chemicals (EDCs) in the sandy aquifer sediment of JiangHan plain, China. Natural typical sandy aquifer sediment was collected and used in this work to represent the sandy aquifer materials of JiangHan plain. Therefore, results from this research are transferable to the natural condition of the study area. This study then, provides parameters data for groundwater pollution modeling and for dealing with groundwater pollution and its control. The concerned EDCs in this research included phthalate esters (PAEs) such as di-methyl phthalate (DMP), diethyl phthalate (DEP), di-butyl phthalate (DBP) and di-isobutyl phthalate (DiBP), bisphenol-A (BPA), aniline (AN) and nitrobenzene (NB), which are from pesticides, dyestuffs, pharmaceuticals, petrochemicals, and other industries compounds. This research provides important parameters for the prediction and control of contaminant plume in the aquifer. Two concerns over such studies are the representation of describing the sorption in field where water is not stationary into the groundwater/sediment complex, but spatially dynamic with variable velocity. More, the concentration of pollutant is not constant but changes according to different local factors.Different laboratory miscible displacement experiments were conducted through columns of 4.8 cm length and 1.6 cm diameter for the PAEs and BPA, and 25 cm length with 2.6 cm diameter for that of AN and NB. Then, to obtain the best simulation parameters, the breakthrough curves (BTCs) data from the laboratory column experiments, which are the scatting of the relative concentration (C/Co) of the sorbate versus time, were fitted with inverses solution models of HYDRUS-1D simulation software. Several models included in HYDRUS-1D were used to compute the parameters from BTCs:linear, equilibrium (LE), linear, one-site non-equilibrium (LO), linear, two-site non-equilibrium (LFO), nonlinear, equilibrium (NE), nonlinear, one-site non-equilibrium (NO), and nonlinear, two-site non-equilibrium (NFO) models. Each model encompasses different parameters such as the distribution/partition coefficient (K/Kd), the fraction of instantaneous adsorption on "Type-1" sites (F), and the first-order sorption rate coefficient for the kinetic non-equilibrium (Type-2) sites (a), and the sorption isotherm coefficient (n). For each experiment, the best-fitting model was selected with the assistance of sum of square residual (SSR), coefficient of determination (r2) and calculated Akaike Information Criterion (AIC). According to the models r2 and AIC values, the optimized parameters from each models i.e. with higher r2 and lower SSR and AIC values were used to analyze the effect of different factors.The research was divided into three parts detailed in chapters 3 to 5:(1) The sorption and transport parameters of four phthalate esters in sandy aquifer were quantified using column experiments and at pore-water velocity of 37.8 m d-1. The transport curves of four individual PAEs were simulated with HYDRUS-1D through fitting linear and nonlinear equilibrium, linear and nonlinear, first order, one-site non-equilibrium, linear and nonlinear, first order, two-site non-equilibrium sorption models. The results showed that two-site models (LFO and NFO) displayed similar best fittings. The results from LFO model simulation showed that when water flowed 1 m in sandy aquifer, PAEs with shorter carbon chains (DMP and DEP) transport 0.0316 and 0.0222 m, respectively. Unexpectedly for the same water transport distance, PAEs with longer carbon chains (DBP and DiBP) transported 0.0402 and 0.0607 m, respectively, which were faster than DMP and DEP. The retardations were mainly caused by the sorption of PAEs on the time-dependent type-2 sites. DBP and DiBP exhibited higher mass transfer speed to and fro type-2 sites, but showed lower total sorption coefficient (K) due to the limited accessibility of sorption sites. Coexistence of PAEs and smaller sorbent particles increased total K values of DBP and DiBP due to synergic development of more sorption sites with DMP and DEP.(2) The sorption and transport behavior of BPA onto aquifer sediment were analyzed considering the effects of pore-water velocity (10-52 cm h-1) and initial concentration (2.5-40 mg l-1). The results showed that the BTCs of BPA fit the linear first-order non-equilibrium two-site model. The parameters such as partition coefficient (Kd), the fraction of instantaneous adsorption on "Type-1" sites (F), the first order sorption rate coefficient for the kinetic non-equilibrium (type-2) sites (a), the retardation coefficient (R), and sorption capacity (qcolumn) were computed. From the results, BPA transported 0.12 to 0.87 m with various pore water velocity in sandy sediment column when water flowed 1 m. The sorption of BPA was mainly caused by the instantaneous surface adsorption as F varied from 0.596 to 0.908. The transport velocity of BPA was affected by pore water velocity (v) and followed the linear equation 1/R= 0.0600+0.0110v (r2=0.9724). The parameter Kd were also closely related to v and followed the equation Ln Kd=1.0023-0.0482v (r2=0.9690). The sorption capacity was more related to the initial BPA concentration (C0) and followed the linear equation qcolumn=0.265+0.253 C0 (r2=0.9727). The parameter a was affected by both v and C0 whereas F was not dramatically affected by both.(3) The transport behavior of AN and NB in a sandy aquifer was quantified under varying pore-water velocities (v) (4.5 to 10 cm h-1) and pH values (5 to 8.5); which represents the groundwater pH range in the study area. The results showed that linear non-equilibrium two-site model fitted well the BTCs; therefore, it was selected as best model with assistance of correlation coefficient and Akaike Information Criterion. The parameter Kd of AN and NB showed positive correlations with both v and F for the binary-mixture experiments. However, NB Kd values correlated well with pH while AN did not. Generally, NB Kd values were greater than AN. Nevertheless, lower values of a, v and pH mostly promoted a higher sorption of NB onto the aquifer, which was NB selective in case of binary-mixture solution. No less than 60% of NB retardation and 20% of AN’s were caused by non-equilibrium sorption on Type-2 sites (1-F). Overall, when water flows 1000m down-gradient in a sandy aquifer, the transport distance was 405m for AN and 131m for NB, leading to their potential to pollute groundwater.The novelties are:(1) The transport behaviors of seven organic pollutants in sandy sediments have been systematically studied; (2) The important parameters and mechanism of sorption of seven organic pollutants in sandy sediments were gained through the simulation and inverse solution of HYDRUS-1D; (3) The effect of various factors such as pore water velocity, influent concentration, single/binary compounds and pH on different organic compounds sorption were analyzed in the sandy aquifer sediment.