Date of Graduation

5-2016

Document Type

Dissertation

Degree Name

Doctor of Philosophy in Engineering (PhD)

Degree Level

Graduate

Department

Civil Engineering

Advisor

Kevin D. Hall

Committee Member

S Ranil Wickramasinghe

Second Committee Member

Findlay G. Edwards

Keywords

Applied sciences; CECs; DPR; Nanofiltration; Potable reuse; Recalcitrant cecs

Abstract

As reuse of municipal water resource recovery facility (WRRF) effluent becomes vital to augment diminishing fresh drinking water resources, concern exists that conventional barriers may prove deficient and the upcycling of contaminants of emerging concern (CECs) could prove harmful to human health and aquatic species if more effective and robust treatment barriers are not in place.

There are no federal Safe Drinking Water Act (SDWA) regulations in place specifically for direct potable reuse (DPR) of WRRF effluent. Out of necessity, some states are developing their own DPR reuse regulations. Currently, reverse osmosis (RO) is the default full advanced treatment (FAT) barrier for CEC control. However, the potential exists for tight thin-film composite (TFC) nanofiltration (NF) membranes to provide acceptable CEC rejection efficacies for less capital, operations and maintenance (O&M), energy, and waste generated.

Recognizing the inherent complexity of CEC rejection by membranes, this research program was designed to elucidate the vital predictive variables influencing the rejection of 96 CECs found in municipal WRRF effluents. Each of the CECs was cataloged by their intended use and quantitative structure activity relationship (QSAR) properties, and measured in secondary effluent samples from WRRFs in Texas and Oklahoma. These secondary effluent samples were then processed in bench-scale, stirred, dead-end pressure cells with water treatment industry-specified TFC NF and RO membranes.

A multi-level, multi-variable model was developed to predict the probable rejection coefficients of CECs with the studied NF membrane. The model was developed from variables selected for their association with known membrane rejection mechanisms, CEC-specific QSAR properties, and characteristics of the actual solute matrix. R statistics software version 3.1.3 was utilized for property collinearity analysis, outlier analysis, and regression modeling. The Pearson correlation method was utilized for selection of the most vital predictor variables for modeling. The resulting Quantitative Molecular Properties Model (QMPM) predicted the NF rejection CECs based on size, ionic charge, and hydrophobicity. Furthermore, the QMPM was verified against a CEC rejection dataset published by an independent study for a similar commercially available TFC NF membrane.

Share

COinS