Search

Organic micropollutants (OMPs) in climate change affected natural environment such as wetlands, and engineered systems have brought serious concerns for water security and public health. These issues have increased the demand for better managing water resources and developing effective technologies for aqueous micropollutants removal. This thesis investigated these subjects through the following five sub-research projects. First, boreal peatland was used as a case study for understanding how peatland fires and droughts impacts peatland resilience. Laboratory results suggested that heating and moisture condition, coupled with peat organic hydrophobic transformations, influence peat soil hydrophobicity and the resultant water-extractable pollutant leaching, which potentially threatens peatland downstream receiving waters such as potable waters by high organic loads. Further, post-fire peat chemistry and their mechanistic relationships to leached pollutants (total organic carbon (TOC), nutrients and phenols) were elucidated through a laboratory leaching study. Increased contaminant loading was observed in post-heated peat leachates, suggesting negative effects to water treatment efficiency and an increase of treatment costs to surface waters as potable water source. Next, peat soils damaged from extreme fires and droughts were upcycled for producing high surface area, value-added porous carbons based on a rapid, facile chemical activation approach. This application had the simultaneous benefit of peatland ecological restoration, protecting downstream communities from heavy run-off, and using the sustainable damaged peats for effective environmental remediation though adsorption. Moreover, a critical review of nano-enabled composite membranes for OMP removal (size-exclusion, adsorption, charge interaction, and photo- and electro-catalysis) and their respective benefits and limitations were discussed. This work brought new perspectives for next-generation nanocomposite membranes for OMP removal. Finally, a novel, hyperbranched polyethylenimine (HPEI) crosslinked iron doped reduced graphene oxide (rGO) nanocomposite membrane was synthesized for process-intensified flow-through separation of phenolic micropollutants. Mechanisms and separation performance to phenolic micropollutant and azo dyes were investigated. Thesis Doctor of Philosophy (PhD) Climate change, industrial manufacture, and population growth have been exacerbating the global water stress. Organic micropollutants (OMPs) are potentially toxic, persistent and can exist even at trace levels, which have been increasingly discovered and identified in natural and built systems. In this research, environmental chemistry of climate-change impacted boreal peatland soils, and their mechanistic relationships to peat soil hydrophobicity, organic substance transformations, micropollutant leaching, and impacts to downstream potable water quality was investigated as a case study. Two different innovative water treatment strategies were developed for restoring peatland resilience and enhancing water resource sustainability including treating peatland phenolic micropollutants. The first approach converted shallow layer wildfire- and drought- damaged peats into value-added porous carbons for adsorption. The second approach synthesized a novel reduced graphene oxide (rGO) nanocomposite membrane for process-intensified flow-through separation. These solutions provide novel insights for source water protection and wastewater treatment in adaption to climate change.

Polymers of intrinsic microporosity (PIMs) are a new type of polymer material with unique microporous structure and have been widely applied in gas separation. In recent years, PIMs have also exper...

Here we study the effects of differentiated embryonic chondrocyte gene 1(DEC1) deficiency on midbrain dopaminergic(DA) neurons in the substantia nigra pars compacta(SNpc) through behavioral, histological and molecular analysis. We have found that compared to the age-matched WT mice, DEC1 deficient mice show a decrease in locomotor activity and motor coordination, which shows the main features of Parkinson’s disease(PD). But there is no significant difference in spatial learning and memory skills between WT and DEC1 KO mice. Compared to the age-matched WT mice, DEC1 deficient mice exhibit the loss of DA neurons in the SNpc and reduction of dopamine and its metabolites in the striatum. The activated caspase-3 and TH/TUNEL+ cells increase in the SNpc of 6- and 12-month-old DEC1 KO mice compared to those of the age-matched WT mice. But we haven't found any NeuN/TUNEL+ cell increase in the hippocampus of the above two types of mice at the age of 6 months. Furthermore, DEC1 deficiency leads to a significant inhibition of PI3K/Akt/GSK3β signaling pathway. Additionally, LiCl could rescue the DA neuron loss of midbrain in the 6-month-old DEC1 KO mice. Taken together, the loss of DA neurons in the DEC1 deficient mice potentially involves the downregulation of PI3K/Akt/GSK3β signaling.

Relation Extraction (RE), as an essential task of Natural Language Processing (NLP), aims to extract potential relations between two entities in a sentence. It is a crucial step in information extraction from unstructured data and building a Knowledge Graph (KG). The performance of deep learning methods for RE, like Recurrent Neural Network (RNN) and Convolutional Neural Network (CNN), heavily depends on the quality and scale of the training set. Recently, pre-training models like BERT and ERNIE, have achieved State-Of-The-Art (SOTA) results in many NLP tasks, because they can obtain the prior semantic knowledge during the procedure of pre-training. Therefore, it is interesting to know whether the performance of RE can be improved utilizing the pre-training models. In this paper, we propose a method of RE using two kinds of pre-training models: BERT and ERNIE. First, in the input sequence, unique symbols are appended around the entities. RE is then regarded as a text classification task, and the prior semantic knowledge obtained by pre-training models is used to improve the performance. Experiments are carried on the SemEval 2010 Task 8 dataset. Results demonstrate that the method we proposed improves the performance of RE compared with previous approaches.