XGBoost model as an efficient machine learning approach for PFAS removal: Effects of material characteristics and operation conditions

Understanding the Selective Removal of Perfluoroalkyl and Polyfluoroalkyl Substances via Fluorine-Fluorine Interactions: A Critical Review

As regulatory standards for per- and polyfluoroalkyl substances (PFAS) become increasingly stringent, innovative water treatment technologies are urgently demanded for effective PFAS removal. Reported sorbents often exhibit limited affinity for PFAS and are frequently hindered by competitive background substances. Recently, fluorinated sorbents (abbreviated as fluorosorbents) have emerged as a potent solution by leveraging fluorine-fluorine (F···F) interactions to enhance selectivity and efficiency in PFAS removal. This review delves into the designs and applications of fluorosorbents, emphasizing how F···F interactions improve PFAS binding affinity. Specifically, the existence of F···F interactions results in removal efficiencies orders of magnitude higher than other counterpart sorbents, particularly under competitive conditions. Furthermore, we provide a detailed analysis of the fundamental principles underlying F···F interactions and elucidate their synergistic effects with other sorption forces, which contribute to the enhanced efficacy and selectivity. Subsequently, we examine various fluorosorbents and their synthesis and fluorination techniques, underscore the importance of accurately characterizing F···F interactions through advanced analytical methods, and emphasize the significance of this interaction in developing selective sorbents. Finally, we discuss challenges and opportunities associated with employing advanced techniques to guide the design of selective sorbents and advocate for further research in the development of sustainable and cost-effective treatment technologies leveraging F···F interactions.

Prediction of g-C3N4-based photocatalysts in tetracycline degradation based on machine learning

Investigating the effects of g-C3N4-based photocatalysts on experimental parameters during tetracycline (TC) degradation can be helpful in discovering the optimal parameter combinations to improve the degradation efficiencies in general. Machine learning methods can avoid the problems of high cost, time-consuming and possible instrumental errors in experimental methods, which have been proven to be an effective alternative for evaluating the entire experimental process. Eight typical machine learning models were explored for their effectiveness in predicting the TC degradation efficiencies of g-C3N4 based photocatalysts. XGBoost (XGB) was the most reliable model with R2, RMSE and MAE values of 0.985, 4.167 and 2.900, respectively. In addition, XGB's feature importance and SHAP method were used to rank the importance of features to provide interpretability to the results. This study provided a new idea for developing g-C3N4-based photocatalysts for TC degradation and intelligent algorithms for predicting the photocatalytic activity of g-C3N4-based photocatalysts.

Removal of N-Nitrosodiphenylamine from contaminated water: A novel modeling framework using metaheuristic-based ensemble models

Investigating the complex interactions among physicochemical variables that influence the adsorptive removal of pollutants is a challenge for conventional one-variable-at-a-time (OVAT) batch methods. The adoption of machine learning-based chemometric prediction models is expected to be more accurate than the conventional method. This study proposed a novel modeling framework for predicting and optimizing the adsorptive removal of N-Nitrosodiphenylamine (NDPhA). Initially, models were trained by using OVAT data, with their hyperparameters subsequently fine-tuned through Bayesian optimization. In the second phase, the particle swarm optimization (PSO) technique was adopted to identify optimal parameters, specifically time, concentration, temperature, pH, and dose, to ensure the highest removal. The adopted analytical method enhances both prediction accuracy and removal efficiency. Utilizing OVAT data for NDPhA removal, the XGBoost regressor significantly outperformed other models. With a correlation coefficient of 0.9667 in the testing dataset, the XGBoost model exhibited its accuracy, emphasized by its low mean squared errors of 28.45 and mean absolute errors of 0.0982. Feature importance analysis consistently identified time and concentration as the most critical factors across all models.

A data-driven analysis to discover research hotspots and trends of technologies for PFAS removal

The frequent detection of persistent per- and polyfluoroalkyl substances (PFAS) in organisms and environment coupled with surging evidence for potential detrimental impacts, have attracted widespread attention throughout the world. In order to reveal research hotspots and trends of technologies for PFAS removal, herein, we performed a data-driven analysis of 3975 papers and 436 patents from Web of Science Core Collection and Derwent Innovation Index databases up to 2023. The results showed that China and the USA led the way in the research of PFAS removal with outstanding contributions to publications. The progression generally transitioned from accidental discovery of decomposition, to experimentation with removal effects and mechanisms of existing methods, and finally to enhanced defluorination and mechanism-driven design approaches. The keywords co-occurrence network and technology classification together revealed the main knowledge framework, which was constructed and correlated through contaminants, substrates, materials, processes and properties. Moreover, adsorption was demonstrated to be the dominant removal process among the current studies. Subsequently, we concluded the principles, advances and drawbacks of enrichment and separation, biological methods, advanced oxidation and reduction processes. Further exploration indicated the hotspots such as alternatives and precursors for PFAS ("genx": 1.258, "f-53b": 0.337), degradable mineralization technologies ("photocatalytic degrad": 0.529, "hydrated electron": 0.374), environment-friendly remediation technologies ("phytoremedi": 0.939, "constructed wetland": 0.462) and combination with novel materials ("metal-organic framework": 1.115, "layered double hydroxid": 0.559) as well as computer science ("molecular dynamics simul": 0.559, "machine learn"). Furthermore, the future direction of technological innovation might lie in high-performance processes that minimize secondary pollution, the development of recyclable and renewable treatment agents, and collaborative control strategies for multiple pollutants. Overall, this study offers comprehensive and objective review for researchers and industry professionals in this field, enabling rapid access to knowledge guidance and insights into research frontiers.

Machine learning vs. statistical model for prediction modeling and experimental validation: Application in groundwater permeable reactive barrier width design

Permeable reactive barrier (PRB) is an effective in-situ technology for groundwater remediation. The important factors in PRB design are the width and reactive material. In this study, the beaded coal mine drainage sludge (BCMDS) was employed as the filling material to adsorb arsenic pollutants in groundwater, aiming to design the width of PRB. The design methods involving traditional continue column experiments and empirical formulas, as well as machine learning (ML) predictions and statistical methods, which are compared with each other. Traditional methods are determined based on breakthrough curves under several conditions. ML method has advantages in predicting the width of mass transfer zone (WMTZ), which simultaneously consider the characteristics of material, pollutant, and environmental conditions, with data collected from articles. After data preprocessing and model optimizing, selected the XGBoost algorithm based on the high accuracy, which shows good prediction for WMTZ (R2 = 0.97, RMSE = 0.15). The experimentally derived WMTZ values were also used to validate the predictions, demonstrating the ML low error rate of 7.04 % and the feasibility. Subsequent statistical analysis of multiple linear regression (MLR) showed the error rate of 39.43 %, interpret superiority of ML due to the complexity of influencing factors and the insufficient precision of math regression. Compared to traditional width design methods, ML can improve design efficiency and save experimental time and manpower. Further expansion of the dataset and optimization of algorithms could enhance the accuracy of ML, overcoming existing limitations and gaining broader applications.

Toward the development of a molecular toolkit for the microbial remediation of per-and polyfluoroalkyl substances

Per- and polyfluoroalkyl substances (PFAS) are highly fluorinated synthetic organic compounds that have been used extensively in various industries owing to their unique properties. The PFAS family encompasses diverse classes, with only a fraction being commercially relevant. These substances are found in the environment, including in water sources, soil, and wildlife, leading to human exposure and fueling concerns about potential human health impacts. Although PFAS degradation is challenging, biodegradation offers a promising, eco-friendly solution. Biodegradation has been effective for a variety of organic contaminants but is yet to be successful for PFAS due to a paucity of identified microbial species capable of transforming these compounds. Recent studies have investigated PFAS biotransformation and fluoride release; however, the number of specific microorganisms and enzymes with demonstrable activity with PFAS remains limited. This review discusses enzymes that could be used in PFAS metabolism, including haloacid dehalogenases, reductive dehalogenases, cytochromes P450, alkane and butane monooxygenases, peroxidases, laccases, desulfonases, and the mechanisms of microbial resistance to intracellular fluoride. Finally, we emphasize the potential of enzyme and microbial engineering to advance PFAS degradation strategies and provide insights for future research in this field.

An artificial intelligence platform for automated PFAS subgroup classification: A discovery tool for PFAS screening

Since structural analyses and toxicity assessments have not been able to keep up with the discovery of unknown per- and polyfluoroalkyl substances (PFAS), there is an urgent need for effective categorization and grouping of PFAS. In this study, we presented PFAS-Atlas, an artificial intelligence-based platform containing a rule-based automatic classification system and a machine learning-based grouping model. Compared with previously developed classification software, the platform's classification system follows the latest Organization for Economic Co-operation and Development (OECD) definition of PFAS and reduces the number of uncategorized PFAS. In addition, the platform incorporates deep unsupervised learning models to visualize the chemical space of PFAS by clustering similar structures and linking related classes. Through real-world use cases, we demonstrate that PFAS-Atlas can rapidly screen for relationships between chemical structure and persistence, bioaccumulation, or toxicity data for PFAS. The platform can also guide the planning of the PFAS testing strategy by showing which PFAS classes urgently require further attention. Ultimately, the release of PFAS-Atlas will benefit both the PFAS research and regulation communities.

MOF-GRU: A MOFid-Aided Deep Learning Model for Predicting the Gas Separation Performance of Metal-Organic Frameworks

The remarkable versatility of metal-organic frameworks (MOFs) stems from their rich chemical information, leading to numerous successful applications. However, identifying optimal MOFs for specific tasks necessitates a thorough assessment of their chemical attributes. Conventional machine learning approaches for MOF prediction have relied on intricate chemical and structural details, hampering rapid evaluations. Drawing inspiration from recent advancements exemplified by Snurr et al., wherein a text string was used to represent a MOF (MOFid), we introduce a MOFid-aided deep learning model, named the MOF-GRU model. This model, founded on natural language processing principles and utilizing the gated recurrent unit architecture, leverages the serialized text string representation of metal-organic frameworks (MOFs) to forecast gas separation performance. Through a focused study on CH4/N2 separation, we substantiate the efficacy of this approach. Comparative assessments against traditional machine learning techniques underscore our model's superior predictive accuracy and its capacity to handle extensive data sets adeptly. The MOF-GRU model remarkably uncovers latent structure-performance relationships with only MOF sequences, obviating the necessity for intricate three-dimensional (3D) structural information. Overall, this model's judicious design empowers efficient data utilization, thereby hastening the discovery of high-performance materials tailored for gas separation applications.

Experiments and machine learning-based modeling for haloacetic acids rejection by nanofiltration: Influence of solute properties and operating conditions

Because of potential risks to public health, the presence of haloacetic acids (HAAs) in drinking water is a major concern. Nanofiltration (NF) has shown potential for HAAs rejection, and several factors, namely, membrane properties, solute properties, and operating conditions, have been revealed key roles. However, knowledge of NF separation mechanism by quantifying these factors is limited. This study investigated and modeled NF performance on HAAs rejection. NF performance was experimentally investigated under various transmembrane pressure (TMP), cross-flow velocity (CV), temperature, pH, ionic strength (IS), and HAAs initial feed concentration (C_in). We used machine learning (ML) to understand the mechanism from the perspective of HAAs properties and operating conditions. Multiple linear regression (MLR), support vector machine (SVM), multsilayer perceptron (MLP), extreme gradient boosting (XGBoost), and random forest (RF) models were used. The MLP, XGBoost and RF models achieved significant performance with high R² (0.970, 0.973, and 0.980) and low RMSE (4.71, 4.41, and 3.84). These three models were analyzed using the Shapley Additive explanation (SHAP) to quantify relative contributions of HAAs properties and operating conditions. XGBoost-SHAP produced the most logical results and was the best-performing model for selecting optimal input variables combinations. The results showed that Stokes radius (r_s), logarithmic octanol-water partitioning coefficient (logK_ow), molecular weight (MW), pH, TMP, and temperature are key variables for interpreting NF process. The effects of HAAs properties were ranked as r_s > logK_ow > MW, suggesting significance of size exclusion and hydrophobic interaction. The impact of the operational conditions followed the order pH > TMP > temperature, illustrating that pH was the major influencing operating condition. This study demonstrated significant capacity of ML, which reduced amount of experimental work. In addition, the main operating conditions can be evaluated in terms of their contributions, making ML an efficient tool for risk management and process optimization.

A juxtaposed review on adsorptive removal of PFAS by metal-organic frameworks (MOFs) with carbon-based materials, ion exchange resins, and polymer adsorbents

The removal of poly- and perfluoroalkyl substances (PFAS) from the aquatic environment is a universal concern due to the adverse effects of these substances on both the environment and public health. Different adsorbents, including carbon-based materials, ion exchange resins, biomaterials, and polymers, have been used for the removal of short-chain (C < 6) and long-chain (C > 7) PFAS from water with varying performance. Metal-organic frameworks (MOFs), as a new generation of adsorbents, have also been recently used to remove PFAS from water. MOFs provide unique properties such as significantly enhanced surface area, structural tunability, and improved selectivity compared to conventional adsorbents. However, due to various types of MOFs, their complex chemistry and morphology, different PFAS compounds, lack of standard adsorption test, and different testing conditions, there are inconclusive and contradictory findings in the literature. Therefore, this review aims to provide critical analysis of the performance of different types of MOFs in the removal of long-chain (C > 7), short-chain (C < 6), and ultra-short-chain (C < 3) PFAS and comprehensively study the efficiency of MOFs for PFAS removal in comparison with other adsorbents. In addition, the adsorption mechanisms and kinetics of PFAS components on different MOFs, including Materials of Institute Lavoisier (MIL), Universiteit of Oslo (UiO), Zeolitic imidazolate frameworks (ZIFs), Hong Kong University of Science and Technology (HKUST), and other hybrid types of MOF were discussed. The study also discussed the effect of environmental factors such as pH and ionic strength on the adsorption of PFAS on MOFs. In addition to the adsorption process, the reusability and regeneration of MOFs in the PFAS removal process are discussed. Finally, challenges and future outlooks of the utility of MOFs for PFAS removal were discussed to inspire future critical research efforts in removing PFAS.