Machine Learning for Polymer Design to Enhance Pervaporation-Based Organic Recovery.

Pervaporation (PV) is an effective membrane separation process for organic dehydration, recovery, and upgrading. However, it is crucial to improve membrane materials beyond the current permeability-selectivity trade-off. In this research, we introduce machine learning (ML) models to identify high-potential polymers, greatly improving the efficiency and reducing cost compared to conventional trial-and-error approach. We utilized the largest PV data set to date and incorporated polymer fingerprints and features, including membrane structure, operating conditions, and solute properties. Dimensionality reduction, missing data treatment, seed randomness, and data leakage management were employed to ensure model robustness. The optimized LightGBM models achieved RMSE of 0.447 and 0.360 for separation factor and total flux, respectively (logarithmic scale). Screening approximately 1 million hypothetical polymers with ML models resulted in identifying polymers with a predicted permeation separation index >30 and synthetic accessibility score <3.7 for acetic acid extraction. This study demonstrates the promise of ML to accelerate tailored membrane designs.

Membrane Distillation-Crystallization for Sustainable Carbon Utilization and Storage.

Anthropogenic greenhouse gas emissions from power plants can be limited using postcombustion carbon dioxide capture by amine-based solvents. However, sustainable strategies for the simultaneous utilization and storage of carbon dioxide are limited. In this study, membrane distillation-crystallization is used to facilitate the controllable production of carbonate minerals directly from carbon dioxide-loaded amine solutions and waste materials such as fly ash residues and waste brines from desalination. To identify the most suitable conditions for carbon mineralization, we vary the membrane type, operating conditions, and system configuration. Feed solutions with 30 wt % monoethanolamine are loaded with 5-15% CO2 and heated to 40-50 °C before being dosed with 0.18 M Ca2+ and Mg2+. Membranes with lower surface energy and greater roughness are found to more rapidly promote mineralization due to up to 20% greater vapor flux. Lower operating temperature improves membrane wetting tolerance by 96.2% but simultaneously reduces crystal growth rate by 48.3%. Sweeping gas membrane distillation demonstrates a 71.6% reduction in the mineralization rate and a marginal improvement (37.5%) on membrane wetting tolerance. Mineral identity and growth characteristics are presented, and the analysis is extended to explore the potential improvements for carbon mineralization as well as the feasibility of future implementation.

Photothermally Triggered Nanoreactors with a Tunable Catalyst Location and Catalytic Activity.

Thermosensitive microgels based on poly(N-isopropylacrylamide) (PNIPAm) have been widely used to create nanoreactors with controlled catalytic activity through the immobilization of metal nanoparticles (NPs). However, traditional approaches with metal NPs located only in the polymer network rely on electric heating to initiate the reaction. In this study, we developed a photothermal-responsive yolk-shell nanoreactor with a tunable location of metal NPs. The catalytic performance of these nanoreactors can be controlled by both light irradiation and conventional heating, that is, electric heating. Interestingly, the location of the catalysts had a significant impact on the reduction kinetics of the nanoreactors; catalysts in the shell exhibited higher catalytic activity compared with those in the core, under conventional heating. When subjected to light irradiation, nanoreactors with catalysts loaded in the core demonstrated improved catalytic performance compared to direct heating, while nanoreactors with catalysts in the shell exhibited relatively similar activity. We attribute this enhancement in catalytic activity to the spatial distribution of the catalysts and the localized heating within the polydopamine cores of the nanoreactors. This research presents exciting prospects for the design of innovative smart nanoreactors and efficient photothermal-assisted catalysis.

Tough and Recyclable Phase-Separated Supramolecular Gels via a Dehydration-Hydration Cycle.

Hydrogels are compelling materials for emerging applications including soft robotics and autonomous sensing. Mechanical stability over an extensive range of environmental conditions and considerations of sustainability, both environmentally benign processing and end-of-life use, are enduring challenges. To make progress on these challenges, we designed a dehydration-hydration approach to transform soft and weak hydrogels into tough and recyclable supramolecular phase-separated gels (PSGs) using water as the only solvent. The dehydration-hydration approach led to phase separation and the formation of domains consisting of strong polymer-polymer interactions that are critical for forming PSGs. The phase-separated segments acted as robust, physical cross-links to strengthen PSGs, which exhibited enhanced toughness and stretchability in its fully swollen state. PSGs are not prone to overswelling or severe shrinkage in wet conditions and show environmental tolerance in harsh conditions, e.g., solutions with pH between 1 and 14. Finally, we demonstrate the use of PSGs as strain sensors in air and aqueous environments.

Anisotropic material depletion in epitaxial polymer crystallization.

The physical properties of a semicrystalline polymer thin film are intimately related to the morphology of its crystalline domains. While the mechanisms underlying crystallization of flat-on oriented polymer crystals are well known, similar mechanisms remain elusive for edge-on oriented thin films due to the propensity of substantially thin films to adopt flat-on orientations. Here, we employ an epitaxial polymer-substrate relationship to enforce edge-on crystallization in thin films. Using matrix-assisted pulsed laser evaporation (MAPLE), we deposit films in which crystal nucleation is spatially separated from subsequent epitaxial crystallization. These experiments, together with phase-field simulations, demonstrate a highly anisotropic and localized material depletion during edge-on crystallization. These results provide deeper insight into the physics of polymer crystallization under confinement and introduce a processing motif in the crystallization of ultrathin structured films.

Liquid-liquid phase separation within fibrillar networks.

Complex fibrillar networks mediate liquid-liquid phase separation of biomolecular condensates within the cell. Mechanical interactions between these condensates and the surrounding networks are increasingly implicated in the physiology of the condensates and yet, the physical principles underlying phase separation within intracellular media remain poorly understood. Here, we elucidate the dynamics and mechanics of liquid-liquid phase separation within fibrillar networks by condensing oil droplets within biopolymer gels. We find that condensates constrained within the network pore space grow in abrupt temporal bursts. The subsequent restructuring of condensates and concomitant network deformation is contingent on the fracture of network fibrils, which is determined by a competition between condensate capillarity and network strength. As a synthetic analog to intracellular phase separation, these results further our understanding of the mechanical interactions between biomolecular condensates and fibrillar networks in the cell.

Formulation and Scale-up of Delamanid Nanoparticles via Emulsification for Oral Tuberculosis Treatment.

Delamanid (DLM) is a hydrophobic small molecule therapeutic used to treat drug-resistant tuberculosis (DR-TB). Due to its hydrophobicity and resulting poor aqueous solubility, formulation strategies such as amorphous solid dispersions (ASDs) have been investigated to enhance its aqueous dissolution kinetics and thereby improve oral bioavailability. However, ASD formulations are susceptible to temperature- and humidity-induced phase separation and recrystallization under harsh storage conditions typically encountered in areas with high tuberculosis incidence. Nanoencapsulation represents an alternative formulation strategy to increase aqueous dissolution kinetics while remaining stable at elevated temperature and humidity. The stabilizer layer coating the nanoparticle drug core limits the formation of large drug domains by diffusion during storage, representing an advantage over ASDs. Initial attempts to form DLM-loaded nanoparticles via precipitation-driven self-assembly were unsuccessful, as the trifluoromethyl and nitro functional groups present on DLM were thought to interfere with surface stabilizer attachment. Therefore, in this work, we investigated the nanoencapsulation of DLM via emulsification, avoiding the formation of a solid drug core and instead keeping DLM dissolved in a dichloromethane dispersed phase during nanoparticle formation. Initial emulsion formulation screening by probe-tip ultrasonication revealed that a 1:1 mass ratio of lecithin and HPMC stabilizers formed 250 nm size-stable emulsion droplets with 40% DLM loading. Scale-up studies were performed to produce nearly identical droplet size distribution at larger scale using high-pressure homogenization, a continuous and industrially scalable technique. The resulting emulsions were spray-dried to form a dried powder, and in vitro dissolution studies showed dramatically enhanced dissolution kinetics compared to both as-received crystalline DLM and micronized crystalline DLM, owing to the increased specific surface area and partially amorphous character of the DLM-loaded nanoparticles. Solid-state NMR and dissolution studies showed good physical stability of the emulsion powders during accelerated stability testing (50 °C/75% RH, open vial).

Sequential Flash NanoPrecipitation for the scalable formulation of stable core-shell nanoparticles with core loadings up to 90.

Flash NanoPrecipitation (FNP) is a scalable, single-step process that uses rapid mixing to prepare nanoparticles with a hydrophobic core and amphiphilic stabilizing shell. Because the two steps of particle self-assembly - (1) core nucleation and growth and (2) adsorption of a stabilizing polymer onto the growing core surface - occur simultaneously during FNP, nanoparticles formulated at core loadings above approximately 70% typically exhibit poor stability or do not form at all. Additionally, a fundamental limit on the concentration of total solids that can be introduced into the FNP process has been reported previously. These limits are believed to share a common mechanism: entrainment of the stabilizing polymer into the growing particle core, leading to destabilization and aggregation. Here, we demonstrate a variation of FNP which separates the nucleation and stabilization steps of particle formation into separate sequential mixers. This scheme allows the hydrophobic core to nucleate and grow in the first mixing chamber unimpeded by adsorption of the stabilizing polymer, which is later introduced to the growing nuclei in the second mixer. Using this Sequential Flash NanoPrecipitation (SNaP) technique, we formulate stable nanoparticles with up to 90% core loading by mass and at 6-fold higher total input solids concentrations than typically reported.

Predicting Extraction Selectivity of Acetic Acid in Pervaporation by Machine Learning Models with Data Leakage Management.

The extraction of acetic acid and other carboxylic acids from water is an emerging separation need as they are increasingly produced from waste organics and CO2 during carbon valorization. However, the traditional experimental approach can be slow and expensive, and machine learning (ML) may provide new insights and guidance in membrane development for organic acid extraction. In this study, we collected extensive literature data and developed the first ML models for predicting separation factors between acetic acid and water in pervaporation with polymers' properties, membrane morphology, fabrication parameters, and operating conditions. Importantly, we assessed seed randomness and data leakage problems during model development, which have been overlooked in ML studies but will result in over-optimistic results and misinterpreted variable importance. With proper data leakage management, we established a robust model and achieved a root-mean-square error of 0.515 using the CatBoost regression model. In addition, the prediction model was interpreted to elucidate the variables' importance, where the mass ratio was the topmost significant variable in predicting separation factors. In addition, polymers' concentration and membranes' effective area contributed to information leakage. These results demonstrate ML models' advances in membrane design and fabrication and the importance of vigorous model validation.

Quick-Release Antifouling Hydrogels for Solar-Driven Water Purification.

Hydrogels are promising soft materials for energy and environmental applications, including sustainable and off-grid water purification and harvesting. A current impediment to technology translation is the low water production rate well below daily human demand. To overcome this challenge, we designed a rapid-response, antifouling, loofah-inspired solar absorber gel (LSAG) capable of producing potable water from various contaminated sources at a rate of ∼26 kg m-2 h-1, which is sufficient to meet daily water demand. The LSAG-produced at room temperature via aqueous processing using an ethylene glycol (EG)-water mixture-uniquely integrates the attributes of poly(N-isopropylacrylamide) (PNIPAm), polydopamine (PDA), and poly(sulfobetaine methacrylate) (PSBMA) to enable off-grid water purification with enhanced photothermal response and the capacity to prevent oil fouling and biofouling. The use of the EG-water mixture was critical to forming the loofah-like structure with enhanced water transport. Remarkably, under sunlight irradiations of 1 and 0.5 sun, the LSAG required only 10 and 20 min to release ∼70% of its stored liquid water, respectively. Equally important, we demonstrate the ability of LSAG to purify water from various harmful sources, including those containing small molecules, oils, metals, and microplastics.

Effects of the polymer glass transition on the stability of nanoparticle dispersions.

In addition to the repulsive and attractive interaction forces described by Derjaguin-Landau-Verwey-Overbeek (DLVO) theory, many charged colloid systems are stabilized by non-DLVO contributions stemming from specific material attributes. Here, we investigate non-DLVO contributions to the stability of polymer colloids stemming from the intra-particle glass transition temperature (Tg). Flash nanoprecipitation is used to fabricate nanoparticles (NPs) from a library of polymers and dispersion stability is studied in the presence of both hydrophilic and hydrophobic salts. When adding KCl, stability undergoes a discontinuous decrease as Tg increases above room temperature, indicating greater stability of rubbery NPs over glassy NPs. Glassy NPs are also found to interact strongly with hydrophobic phosphonium cations (PR4+), yielding charge inversion and intermediate aggregation while rubbery NPs resist ion adsorption. Differences in the lifetime of ionic structuration within mobile surface layers is presented as a potential mechanism underlying the observed phenomenon.

Rapid Precipitation of Ionomers for Stabilization of Polymeric Colloids.

Polymeric colloids have shown potential as "building blocks" in applications ranging from formulations of Pickering emulsions and drug delivery systems to advanced materials, including colloidal crystals and composites. However, for applications requiring tunable properties of charged colloids, obstacles in fabrication can arise through limitations in process scalability and chemical versatility. In this work, the capabilities of flash nanoprecipitation (FNP), a scalable nanoparticle (NP) fabrication technology, are expanded to produce charged polystyrene colloids using sulfonated polystyrene ionomers as a new class of NP stabilizers. Through experimental exploration of formulation parameters, increases in the ionomer content are shown to reduce the particle size, mitigating a significant trade-off between the final particle size and inlet concentration; thus, expanding the processable material throughput of FNP. Further, the degree of sulfonation is found to impact stabilization with optimal performance achieved by selecting ionomers with intermediate (2.45-5.2 mol %) sulfonation. Simulations of single ionomer chains and their arrangement in multicomponent NPs provide molecular insights into the assembly and structure of NPs wherein the partitioning of ionomers to the particle surface depends on the polymer molecular weight and degree of sulfonation. By combining the insights from simulations with diffusion-limited growth kinetics and parametric fits to experimental data, a simple design formulation relation is proposed and validated. This work highlights the potential of ionomer-based stabilizers for controllably producing charged NP dispersions in a scalable manner.

Decoupling of Glassy Dynamics from Viscosity in Thin Supported Poly(n-butyl methacrylate) Films.

We utilized fast scanning calorimetry to characterize the glass transition temperature (T g) and intrinsic molecular mobility of low-molecular-weight poly(n-butyl methacrylate) thin films of varying thicknesses. We found that the T g and intrinsic molecular mobility were coupled, showing no film thickness-dependent variation. We further employed a unique noncontact capillary nanoshearing technique to directly probe layer-resolved gradients in the rheological response of these films. We found that layer-resolved shear mobility was enhanced with a reduction in film thickness, whereas the effective viscosity decreased. Our results highlight the importance of polymer-substrate attractive interactions and free surface-promoted enhanced mobility, establishing a competitive nanoconfinement effect in poly(n-butyl methacrylate) thin films. Moreover, the findings indicate a decoupling in the thickness-dependent variation of T g and intrinsic molecular mobility with the mechanical responses (shear mobility and effective viscosity).

Tuning Morphologies and Reactivities of Hybrid Organic-Inorganic Nanoparticles.

Hybrid nanoparticles (hNPs), or nanoparticles composed of both organic and inorganic components, hold promise for diverse energy and environmental applications due to their ability to stabilize reactive nanomaterials against aggregation, enhancing their ability to pervade tortuous spaces and travel long distances to degrade contaminants in situ. Past studies have investigated the use of polymer or surfactant coatings to stabilize nanomaterials against aggregation. However, fabrication of these materials often requires multiple steps and lacks specificity in the control of their morphologies and reactivities. Here, we demonstrated a method of producing stable hNPs with tunable morphologies by incubating polystyrene nanoparticles formed via Flash NanoPrecipitation with citrate-stabilized gold nanocatalysts. Using this simple fabrication technique, we found that gold adsorption to polystyrene nanoparticles was enabled by the presence of a good solvent for polystyrene. Furthermore, changing process parameters, such as gold incubation time, and molecular parameters, such as polymer molecular weight and end-group functionality, provided control over the resultant nanocatalyst loading and dispersal atop hNPs. We classified these morphologies into three distinct regimes─aggregated, dispersed, or internalized─and we showed that the emergence of these regimes has key implications for controlling reaction rates in applications such as heterogeneous catalysis or groundwater remediation. Specifically, we found that hNPs with gold nanocatalysts embedded below the surfaces of polystyrene nanoparticles exhibited slower bulk catalytic reduction capacity than their disperse, surface-decorated counterparts. Taken together, our work demonstrates a simple way by which hNPs can be fabricated and presents a method to control catalytic reactions using reactive nanomaterials.

Physical aging of hydroxypropyl methylcellulose acetate succinate via enthalpy recovery.

Amorphous solid dispersions (ASDs) utilize the kinetic stability of the amorphous state to stabilize drug molecules within a glassy polymer matrix. Therefore, understanding the glassy-state stability of the polymer excipient is critical to ASD design and performance. Here, we investigated the physical aging of hydroxypropyl methylcellulose acetate succinate (HPMCAS), a commonly used polymer in ASD formulations. We found that HPMCAS exhibited conventional physical aging behavior when annealed near the glass transition temperature (Tg). In this scenario, structural recovery was facilitated by α-relaxation dynamics. However, when annealed well below Tg, a sub-α-relaxation process facilitated low-temperature physical aging in HPMCAS. Nevertheless, the physical aging rate exhibited no significant change up to 40 K below Tg, below which it exhibited a near monotonic decrease with decreasing temperature. Finally, infrared spectroscopy was employed to assess any effect of physical aging on the chemical structure of HPMCAS, which is known to be susceptible to degradation at temperatures 30 K above its Tg. Our results provide critical insights necessary to understand better the link between the stability of ASDs and physical aging of the glassy polymer matrix.

Hysteresis in the thermally induced phase transition of cellulose ethers.

Functionalized cellulosics have shown promise as naturally derived thermoresponsive gelling agents. However, the dynamics of thermally induced phase transitions of these polymers at the lower critical solution temperature (LCST) are not fully understood. Here, with experiments and theoretical considerations, we address how molecular architecture dictates the mechanisms and dynamics of phase transitions for cellulose ethers. Above the LCST, we show that hydroxypropyl substituents favor the spontaneous formation of liquid droplets, whereas methyl substituents induce fibril formation through diffusive growth. In celluloses which contain both methyl and hydroxypropyl substituents, fibrillation initiates after liquid droplet formation, suppressing the fibril growth to a sub-diffusive rate. Unlike for liquid droplets, the dissolution of fibrils back into the solvated state occurs with significant thermal hysteresis. We tune this hysteresis by altering the content of substituted hydroxypropyl moieties. This work provides a systematic study to decouple competing mechanisms during the phase transition of multi-functionalized macromolecules.

Development of an In Vitro Release Assay for Low-Density Cannabidiol Nanoparticles Prepared by Flash NanoPrecipitation.

Nanoparticle encapsulation is an attractive approach to improve the oral bioavailability of hydrophobic therapeutics. The high specific surface area of nanoparticle formulations, combined with the thermodynamically driven increased solubility of an amorphous drug core, promotes rapid drug dissolution. However, the physicochemical properties of the hydrophobic therapeutic can present obstacles to in vitro characterization of nanoparticle formulations. Namely, drugs with low density and high membrane binding affinity frustrate traditional analytical methods to monitor release kinetics from nanoparticles. In this work, cannabidiol (CBD) was encapsulated into nanoparticles with low polydispersity and high drug loading via Flash NanoPrecipitation (FNP), a scalable self-assembly process. Hydroxypropyl methylcellulose acetate succinate (HPMCAS) and lecithin were employed as amphiphilic particle stabilizers during the FNP process. However, the low density and high membrane binding affinity of the amorphous CBD nanoparticle core prevented the characterization of in vitro release kinetics by conventional methods. Released CBD could not be separated from intact nanoparticles by filtration or centrifugation. To address this challenge, an alternative approach is described to coencapsulate 6 nm hydrophobic Fe3O4 colloids with CBD during FNP. The Fe3O4 colloids were added at 33% by mass (approximately 20% by volume) to increase the density of the nanoparticles, resulting in particles with an average diameter of 160 nm (CBD-lecithin-Fe3O4) or 280 nm (CBD-HPMCAS-Fe3O4). This densification enabled the centrifugal separation of dissolved (released) CBD from unreleased CBD during the in vitro assay while avoiding the losses associated with a filtration step. The resulting nanoparticle formulations provided more rapid and complete in vitro dissolution kinetics than bulk CBD, representing a 6-fold improvement in dissolution compared to crystalline CBD. The coencapsulation of high-density Fe3O4 colloids to enable the separation of nanoparticles from release media is a novel approach to measuring in vitro release kinetics of nanoencapsulated low-density, hydrophobic drug molecules.