Progress 10/01/23 to 09/30/24
Outputs
Target Audience:This project has provided the opportunities to train one graduate students, one post-doctor, and one undergraduate students in the interdisciplinary areas of membrane design, polymer chemistry, nanostructure design, sensor membrane, and machine learning driven membrane design. Changes/Problems:Nothing changes. What opportunities for training and professional development has the project provided?This project has provided the opportunities to train two graduate students, one post-doctor, and two undergraduate students in the interdisciplinary areas of membrane design, polymer chemistry, nanostructure design, sensor membrane, and machine learning driven membrane design. How have the results been disseminated to communities of interest?The results have been disseminated to communities of interest through the conference presentations. What do you plan to do during the next reporting period to accomplish the goals? Manipulate polyamine ligand structure to improve binding capacity We will select amine precursors based on their molecular weight and crosslinking structures, then synthesize and characterize amine polymer ligands with end groups designed for high-affinity metal chelation. Bayesian optimization-guided biocomposite membrane fabrication We will synthesize cellulose-based crosslinked biocomposite membranes, characterize their chemical and structural properties, and evaluate their performance in terms of adsorption, flux, fouling, and regeneration. Finally, we will apply Bayesian optimization to guide membrane fabrication and fine-tune performance. Evaluate triboelectric responses for metal ion detection and membrane regeneration actuation We will conduct laboratory experiments to characterize the triboelectric response of sensor membranes in single and multi-metal solutions. Following this, we will develop CNN-based machine learning models to differentiate metal ions, enabling optimized membrane regeneration actuation and sequencing. Evaluate triboelectric sensitivity for different heavy metal ions using Chitosan (CS) and Cellulose Nanofibrils (CNF) blend films We will prepare chitosan-cellulose nanofibrils blend films with different ratios (by weight) (20:80, 40:60, 60:40, 80:20) of Chitosan and CNF. We will perform FTIR and XRD characterizations on these films to test the functional groups presence and crystallinity. Roughness measurements will be done using a confocal microscope to evaluate the roughness and porosity calculations. The FTIR data shows the presence of the -NH2 stretching and carbonyl groups. Porosity controlled triboelectric sensitivity To bring out the difference in porosity levels of the films, we will perform different drying techniques, such as slow drying (under ambient conditions), air flow drying, and heated drying (at 70 C) to prepare the films. Porosity calculations will be performed on a confocal microscope to better explain the sensitivity changes. Differentiation between different heavy metal ions We will further use these films to test the detection of heavy metals using TENG. The films will be first placed on the Cu/PTFE substrate. Blank measurements will be taken to give a comparison with those of the heavy metal doped films. The sensitivity detection curve can then be used as a method to differentiate between multiple heavy metal ions.
Impacts
What was accomplished under these goals?
1. We have prepared the adsorptive membrane by grafting hyperbranched polyethyleneimine (hPEI) on the commercially available and bio-based cellulose acetate (CA) membrane substrate for efficient adsorption of heavy metals. 2. We have investigated the properties of the adsorptive membranes using scanning electron microscopy (SEM), energy dispersive X-ray spectroscopy (EDS), Fourier-transform infrared spectroscopy (FTIR), X-ray photoelectron spectroscopy (XPS), and Kelvin probe force microscopy (KPFM). SEM images revealed that, compared to the pristine CA surface, the CA-PEI membrane exhibited a uniform hPEI coating and significantly smaller pore size. FTIR analysis confirmed the success of both the deacetylation and PEI grafting processes. EDS and XPS results further validated the grafting of PEI onto the membrane surface. 3. We have performed the adsorption test, and this result demonstrated how contact time affects Cu2+ adsorption capabilities. The adsorption kinetics are rapid in the initial stage, with equilibrium achieved after 30 minutes. The rapid adsorption in the initial stage is attributed to the high availability of surface-active sites. As the active sites become occupied by Cu2+ during the adsorption process, the rate of adsorption gets slower along with time. Furthermore, we found that our system closely follows the pseudo first order (PFO) kinetic model which suggests that the metal uptake rate is dependent on the concentration of Cu2+. 4. We have conducted an experiment to assess the impact of different cations on the adsorption capacity of CA-PEI membranes, single metal adsorption tests were conducted. Initially, salt solutions (Cu2+, Cd2+, Pb2+, Zn2+, Co2+, Ni2+, Mg2+, Sr2+, Mn2+, Ca2+, Ba2+, K+, Na+, Li+, Fe3+) with the same anion (Cl-) were prepared. The results indicate that the CA-PEI membrane exhibits the highest adsorption capacity for Cu2+ compared to other cations, demonstrating its high selectivity for different types of metal ions. The impact of different anions on the adsorption performance was evaluated with fixed cation (Cu2+) while the anion was varied (Cl-, SO42-, NO3-). The findings indicate that the adsorption capacities remain nearly identical, suggesting that the CA-PEI membranes' adsorption capacities for Cu2+ do not change with varying anions. 5. We have prepared the triboelectric membrane sensor (TEMS) that can detect and distinguish various heavy metals. After metal adsorption, TEMS exhibited a marked 74.2% increase in triboelectric effect, enabling the detection of heavy metal ions with high sensitivity. According to distinct triboelectric signals, TEMS can identify different metal ions. The system proposed in this work is superior to previously reported TENG sensors by its capability of detecting a much broader concentration range from 1 nM to 1 mM. Furthermore, we have combined mass balance equation, kinetic model, isotherm model and the adsorption-voltage model to predict the initial heavy metal concentration. 6. We have conducted desorption-adsorption experiments to confirm the reusability of TEMS. The system demonstrates excellent reusability, maintaining nearly the same adsorption capacity and triboelectric voltage over six adsorption-desorption cycles. The regeneration frequency of the adsorptive membrane can be optimized according to the voltage signal.
Publications
- Type:
Journal Articles
Status:
Published
Year Published:
2024
Citation:
R. Liu, Q. Li, L. Gan, Y. Chen, Z. Tong. Robust, Versatile, and Hemicellulose-derived Biocomposite Janus Membrane for Saline Wastewater Desalination. (2024) Journal of Membrane Science 697, 122566.
- Type:
Journal Articles
Status:
Published
Year Published:
2024
Citation:
C. Xiao, L. Yan, H. Gao, Z. Dou, X. Xie, Y. Chen. Selective Separation and Recovery of Co (II) and Ni (II) from Lithium-ion Battery Using Cyanex 272 Adsorptive Membrane. Environmental Science & Engineering, published 2024.
- Type:
Conference Papers and Presentations
Status:
Accepted
Year Published:
2023
Citation:
G.-S. Lin, H. Zou, L. Gan, Z. Tong, Y. Ding, Y Chen. Cellulose-based Triboelectric Membrane Sensor for Heavy Metal Ions Removal and Detection. 2023 PEERS/IBBC Conference, Atlanta, GA
|
Progress 10/01/22 to 09/30/23
Outputs
Target Audience:The research conducted for this project contributes to the training ofone post-doctor who has an exciting opportunity to be involved in interdisciplinary research including membrane design, polymer functionalization, and nanomaterials. The target audience we will target includes graduate students or post-doctors, undergraduate students, and the broader academic communities through conference presentations, and published journals. Changes/Problems:There are no major changes/problems. We have finely tuned the objective to add more advanced technology for adsorption membrane design such as machine-learning-based adsorption membrane design, including triboelectric sensor and nanocluster concept to the membrane design, etc. What opportunities for training and professional development has the project provided?This project has provided the opportunities to train two graduate students, one post-doctor, and two undergraduate students in the interdisciplinary areas of membrane design, polymer chemistry, nanostructure design, sensor membrane, and machine learning driven membrane design. How have the results been disseminated to communities of interest?The results have been disseminated to communities of interest through the conference presentations. What do you plan to do during the next reporting period to accomplish the goals?Triboelectric Membrane Sensor for Heavy Metal Ions Removal and Detection We will propose adsorption kinetics and isotherm models to investigate the adsorption behavior of the as-prepared adsorptive membranes. We will explore the underlying mechanism of how the adsorbed heavy metal ions change the triboelectric surface potential of the adsorptive membranes. Nanocluster based sensor for Heavy Metal Ions 1.We would do some structure and property characterizations using the MNCs/CNCs composites, such as SEM, TEM, XPS, XRD, TGA, UV-vis spectra, excitation and emission spectra, and so on., aiming to understand the relation between structure and property of MNCs/CNCs composites. 2. We will make the obtained MNCs/CNCs composites into membranes or other shapes according to needs to serve as fluorescence probes/sensors in different work conditions. 3. We are going to test the sensitivity of sensors made from MNCs/CNCs composites to figure out the best strategy for the high-performance and sensitive sensors. Machine Learning-Guided Design of Adsorptive Membranes for Heavy Metal Removal We will introduce multiple model evaluation methods and fine-tune the parameters of the model to improve its prediction performance. Based on model-selected substructures, we will construct the five most promising adsorbent structures. We will compare their similarity to available polymers to identify the best-performing polymers. With the model prediction results, we will conduct heavy metal removal experiments to evaluate the performance of selected polymers and validate the model.
Impacts
What was accomplished under these goals?
Based on previous studies to synthesize robust adsorption membrane for multiple heavy metal adsorption, we have made the following progresses on three aspects: 1) We coordinate a triboelectric sensor with an adsorption membrane for heavy metal removal and detection; 2) we designed metal nanocluster sensor for precisely detecting heavy metal ions; 3) we designed a new machine learning model to guide adsorption membrane design. Triboelectric Membrane Sensor for Heavy Metal Ions Removal and Detection We have successfully grafted hyper-branched polyethylenimine (PEI) onto the cellulose acetate (CA) membrane (CA-PEI). We have investigated the properties of adsorptive membranes by using scanning electron microscopy (SEM), energy-dispersive X-ray spectroscopy (EDS), Fourier-transform infrared spectroscopy (FTIR), and X-ray photo-electron spectroscopy (XPS). According to SEM image observation, there is no observable morphological change on the membrane surface before and after PEI grafting. FTIR measurement results indicate that the deacetylation and the PEI grafting procedures were successful. EDS and XPS results further confirm that PEI was grafted onto the membrane surface. We have performed the adsorption test, and this result demonstrated how contact time affects Cu2+ adsorption capabilities. Adsorption kinetics were quick in the beginning, and saturation was reached after 30 minutes. Afterward, no discernible change was observed with the increasing contact time. Rapid adsorption in the initial stage is attributed to the higher availability of surface-active sites, which are occupied by Cu2+ during the adsorption process. After Cu2+ adsorption, the membrane was dried and paired with Cu/polytetrafluoroethylene (PTFE) for triboelectricity generation. We tested the relationship between metal adsorption capacity and triboelectric voltage difference of the CA-PEI membrane with an initial Cu2+ concentration of 60 ppm and under different timeslots. We found that as the adsorption capacity increases, the voltage value increases monotonically. This can be explained by the adsorption of Cu2+ increasing the tendency of the surface to gain electrons. This relationship can be further analyzed by linear regression (R2=0.91), which implies that the triboelectric signals correspondingly represent different membrane metal adsorption capacities. The sensing performance of triboelectric sensors for detecting different heavy metals was evaluated. Results show that the triboelectric signals differentiate different heavy metal ions. During the active sensing process, we also observed that the energy produced could also be collected, which was sufficient to power light-emitting diodes (LEDs), forming the "TENG" word. To test the sensitivity of the adsorptive membrane, seven different concentrations of Cu2+ solutions ranging from 1 nM (about 0.134 ppm) to 1 mM (about 134.45 ppm) were prepared. We then submerged adsorptive membranes in these solutions for 10 seconds. After the membranes had dried, we measured their corresponding triboelectric current signals. Results indicate that the triboelectric short circuit current grows linearly (R2 = 0.97) with increasing copper ion concentration. This finding suggests that our membrane sensor has exceptional sensitivity to various Cu2+ solution concentrations, even when the concentration is less than 1 ppm. Metal Nanocluster Based Sensor for Heavy Metal Ions 1. We have prepared some high fluorescence quantity yield metal nanoclusters (MNCs), e.g., Au9Ag2, Au13Ag12, Au11Cu5, etc., and revealed their crystal structures at atomic level via various characterization technologies including electron spray ionization mass spectra, X-ray photoelectron spectra, single crystal x-ray diffraction and nuclear magnetic resonance. 2. We further studied the optical properties like UV-vis light absorption and fluorescence emission, which are related to their unique electron structures. 3. We have prepared uniform cellulose nanocrystals (CNCs), and revealed their special size distribution, surface structures, and properties, which is a very cheap and ideal supporter to prevent the aggregation of nanoclusters. 4, We have figured out the way to form MNCs/CNCs composites by combining obtained MNCs and CNCs together in the water system, which is ready for the following sensing capability test. Machine Learning-Guided Design of Adsorptive Membranes for Heavy Metal Removal We have identified the relevant features for the predictive model, including the structure of the active adsorbents, fabrication method of the adsorptive membrane, operation conditions, and properties of the ions to remove. We have constructed the dataset with 230 data points and converted the structure of the active adsorbent into molecular fingerprints to represent the substructure. The substructures are prepared for the sake of studying the contribution of each atom group and further screening for promising adsorbents. We have screened through the combination of 12 candidate machine learning models, 8 encoders for encoding categorical features, and 5 scalers for translating numerical features. The most promising combination consists of CatBoost regressor, M-Estimate encoder and MaxAbsScaler. Currently, the predictive model has R2 for the training data and the test data at 0.98 and 0.66, respectively. We have quantitively evaluated the contribution of each feature by calculating the SHAP value. Targeting membrane rejection as the model output, the most influential ones are pH, and metal concentration followed by the substructures that we will carry over to the adsorbent material selection phase.
Publications
- Type:
Conference Papers and Presentations
Status:
Accepted
Year Published:
2022
Citation:
Lin, G.-S., Zou, H., Gan, L., Ding, Y.*, Chen, Y.*, and Tong Z.* (2022) Simultaneous Heavy Metal Ion Capture and Detection via Bio-Based Triboelectric Sensor. 2022 AIChE Annual Meeting, 13-18 November 2022, Phoenix, AZ.
- Type:
Conference Papers and Presentations
Status:
Accepted
Year Published:
2023
Citation:
Lin, G.-S., Zou, H., Gan, L., Ding, Y.*, Chen, Y.*, and Tong Z.* (2023) Cellulose-based Triboelectric Membrane Sensor for Heavy Metal Ions Removal and Detection. RBI Spring Workshop on Packaging Innovation & the Circular Economy, 11-12 May 2023, Atlanta, GA.
- Type:
Conference Papers and Presentations
Status:
Accepted
Year Published:
2023
Citation:
Gan, L., Pei, X., Tong., Z.*, Gao, H., Meng, S., Zhang, W., Wang, P., Chen, Y.* (2023) Robust Bio-Sourced Composite Adsorptive Membrane for Heavy Metal Capture. 2023 AEESP Research and Education Conference, 20-23 June 2023, Boston, MA.
- Type:
Conference Papers and Presentations
Status:
Accepted
Year Published:
2023
Citation:
Gan, L., Zhong, S., Chen, Y.* (2023) Data-driven Method Guided Design of Adsorptive Membrane for Heavy Metal Removal. ACS Fall 2023, 13-17 August 2023, San Francisco, CA & Hybrid.
|