Chemical and Biomolecular Engineering Theses and Dissertations
Permanent URI for this collectionhttp://hdl.handle.net/1903/2751
Browse
169 results
Search Results
Item ELECTROLYTE AND INTERPHASE DESIGN FOR HIGH-ENERGY AND LONG-LIFE LITHIUM/SULFURIZED POLYACRYLONITRILE (Li/SPAN) BATTERIES(2024) Phan, An Le Bao; Wang, Chunsheng; Chemical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Lithium/sulfurized polyacrylonitrile (Li/SPAN) recently emerged as a promising battery chemistry with theoretical energy density beyond traditional lithium-ion batteries, attributed to the high specific capacities of Li and SPAN. Compared to traditional sulfur-based cathodes, SPAN demonstrated superior sulfur activity/utilization and no polysulfide dissolution issue. Compared to batteries based on layered oxide cathodes, Li/SPAN shows two significant advantages: (1) high theoretical energy density (> 1000 Wh kg-1, compared to around 750 Wh kg-1 of Li/LiNi0.8Mn0.1Co0.1O2) and (2) transition-metal-free nature, which eliminates the shortcomings associated with transition metals, such as high cost, low abundance, uneven distribution on the earth and potential toxicity. The success of Li/SPAN chemistry with those two critical advantages would not only relief the range and cost anxiety persistently associated with electric vehicle (EV) applications, but also have great implications for the general energy storage market. However, current Li/SPAN batteries still fall far behind their true potential in terms of both energy density and cycle life. This dissertation aims to provide new insights into bridging the theory-practice gap of Li/SPAN batteries by appropriate interphase and comprehensive electrolyte designs. First, the effect of Li/SPAN cell design on energy density and cycle life was discussed using relevant in-house developed models. The concept of “sensitivity factor” was established and used to quantitatively analyze the influence of input parameters. It was found that the electrolyte, rather than SPAN and Li electrodes, represents the bottleneck in Li/SPAN development, which explains our motivation to focus on electrolyte study. Another remarkable finding is that although not well-perceived, electrolyte density has a great impact on Li/SPAN cell-level energy density. Second, design principles to achieve good electrode-electrolyte compatibility were explored. Novel approaches to promote the formation of more protective, inorganic-rich interphases (SEI or CEI) were proposed and validated with proper experiments, including electrochemical tests, material characterizations (such as SEM, XPS, NMR, IR, Raman), and their correlations. Finally, based on the principles discussed in previous chapters, we developed a new electrolyte that simultaneously offers good electrochemical performance (Li CE > 99.4%, Li-SPAN full-cells > 200 cycles), decent ionic conductivity (1.3 mS cm-1), low density (1.04 g mL-1), good processability (higher vapor pressure than conventional carbonates, b.p. > 140 °C), and good safety. Outlook and perspective will also be presented. Beyond Li/SPAN, we believe that our findings regarding cell design as well as electrolyte solvation structure, interphases chemistry, and their implications on electrochemical performance are also meaningful for the development of other high-energy battery chemistries.Item QUANTITATIVE UNDERSTANDING OF TEMPERATURE RISE AND SAFETY IN HIGH – ENERGY SOLID – STATE BATTERIES.(2024) Ogundipe, Taiwo Oladapo; Albertus, Paul; Chemical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)The rising demand for renewable energy and electric transportation has increased the need for advanced and safer battery technologies. Conventional lithium – ion batteries face limitations in energy density and safety risks due to the reaction of oxygen from the decomposed cathodes with other battery components, which can cause thermal runaway, leading to fires or explosions. Solid – state batteries, which use a solid electrolyte, offer a promising solution by potentially improving both energy density and safety. This study focuses on the thermal behavior and heat generation of anode – cathode – electrolyte (ACE) (Li / LPSCl / NMC811) solid-state batteries using differential scanning calorimetry (DSC). The results show significant heat generation, ranging from 4000 to 5400 J/g NMC811, with a corresponding adiabatic temperature rise of 1300 – 1750 ℃. When small amounts of liquid electrolyte are added, the onset temperature is lowered, and the heat release shifts to higher temperatures. However, the total heat generation remains within a similar range. These findings provide insights into the thermal stability of all – solid – state batteries , and solid – state batteries with small amount of liquid electrolyte, contributing to the development of safer and higher energy density energy storage systems.Item INTEGRATED PROCESS MODELING AND EXPERIMENTAL ANALYSIS FOR OPTIMIZING CONTINUOUS MANUFACTURE OF DRUG SUBSTANCE CARBAMAZEPINE(2024) Kraus, Harrison; Choi, Kyu Yong; Chemical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)This dissertation presents a comprehensive study on the continuous manufacture (CM) of the drug substance (DS) carbamazepine (CBZ), a widely used anti-epileptic medication, aimed at enhancing process efficiency and product quality. The research progresses through a series of investigations, beginning with the development of kinetic models for CBZ synthesis from iminostilbene via two different synthetic routes using urea and potassium cyanate across various reactor setups, including batch and continuous flow systems. Discrepancies between batch and continuous models, particularly in yield prediction and impurity formation, are thoroughly examined and addressed through adjustments in reactant addition methods and system designs. This demonstrates the value of mechanistic modeling, a tool that has been undervalued in recent research particularly for its ability to compare between batch and continuous systems. Subsequently, the research delves into the crystallization processes, employing a population balance model (PBM) to study CBZ polymorph form III crystal formation, highlighting the influence of seed crystal size distribution on product crystal quality. It also provides novel strategies for modeling the evolution of crystal size distribution (CSD) due to nucleation and growth and evaluates the robustness of these strategies as seed CSD varies. Lastly, the scope is expanded to a holistic view of the integrated synthesis and crystallization process presenting one of the first studies of a complete DS CM system and emphasizing the development of a robust Quality-by-Control (QbC) framework. This includes the implementation of in-line Raman spectroscopy for real-time concentration monitoring, an active feedback level control system, dynamic modeling of impurity partitioning for enhancing disturbance mitigation across the CM process, and a retrograde design strategy that optimizes the upstream synthesis based on downstream purification capabilities/limitations. Through all these contributions, the dissertation aims to advance the modernization of continuous manufacturing practices in the pharmaceutical industry and promotes a shift towards more adaptive and controlled production environments.Item Sutureless Anastomosis: Electroadhering a Hydrogel Sleeve Over Cut Pieces of Tubular Tissue(2024) Grasso, Samantha Marie; Raghavan, Srinivasa R; Chemical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Recently, our lab demonstrated that cationic gels could be adhered to animal tissues by applying an electric field (10 V DC, for ~ 20 s). This phenomenon, termed electroadhesion (EA), could potentially be used to repair injured tissues without sutures. An extreme injury is when a tube in the body (e.g., a blood vessel or an intestine) is cut into two segments. The surgical process of joining the segments is termed anastomosis, and thus far has only been done clinically with sutures. Here, we explore the use of EA for performing sutureless anastomoses in vitro with bovine aorta and chicken intestine. For this purpose, we make a strong and stretchable cationic gel in the form of a sleeve (i.e., a hollow tube). By using a custom plastic mold, we control both the sleeve diameter and wall thickness. A sleeve with a diameter matching that of the tubular tissue is slipped over the cut segments of the tube, followed by application of the DC electric field. Thereby, the sleeve becomes strongly adhered by EA to the underlying tube. Water or blood is then flowed through the repaired tube, and we record the burst pressure Pburst of the tube. We find that Pburst is > 80 mm Hg and close to the Pburst of an intact (uncut) tube. In comparison to the sleeve, a long strip of the gel attached around the cut tubular pieces allows a much lower Pburst. Thus, our study shows that gel-sleeves adhered by EA could enable anastomoses to be performed in the clinic without the need for sutures.Item Slow and Sustained Release of Hydrophilic Solutes from Capsules(2024) Cho, Hannah Yeonhee; Raghavan, Srinivasa R; Chemical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Capsules can be used to deliver solutes encapsulated in their aqueous core, including drugs. However, small, hydrophilic solutes tend to leak out of typical capsules in a matter of minutes. Solute release can be completely prevented if the capsule shell is made of a hydrophobic solid like paraffin wax. The goal of this study is to achieve release profiles between these extremes 3⁄4 i.e., release of solutes in a slow and sustained manner over a period of days. For this, we modify the wax-shell design above and include a nonionic surfactant (from the Brij series) along with the wax. Most of our studies have been done with Brij-C10, which has a polyethylene glycol (PEG) head and a hexadecyl (C16) tail. We show that capsules with a shell of 80/20 wax/Brij can release model dyes from the aqueous core over 20+ days. Such slow release has not been reported previously. The release rate can be tuned via the concentration of Brij in the shell (the higher the Brij, the faster the release) as well as the chemistry of the Brij (i.e., the size of the surfactant head or tail). The sustained release occurs because the Brij-bearing shell has microchannels through which the solute can permeate. Our approach can be used to slowly release many solutes, including dyes, drugs, and reactive reagents (such as H2O2). The simplicity and versatility of this approach make it highly suitable for controlled release applications across the pharmaceutical, agrochemical, and cosmetics industries.Item Chemically Fueled Transient Porous Hydrogels as Autonomous Soft Actuators(2024) Battumur, Sarangua; Woehl, Taylor J.; Chemical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Biomimetic materials take inspiration from biological systems to develop transformativereconfigurable synthetic materials. One example of a biological system is the contraction of muscle fibers driven by biochemical reactions. Actin filaments are crucial to cellular functions, facilitating movement, division, and structural integrity through ATP-driven polymerization and depolymerization. This ability to self-assemble and disassemble in response to biochemical signals provides a model for creating materials that mimic the sophisticated control found in biological systems. This thesis describes the use of a chemical reaction network to rapidly and autonomouslyreconfigure the size of a porous polymer hydrogels within tens of minutes. This hydrogel, composed of acrylic acid and acrylamide monomers, exhibits exceptional microporosity and an ability to expand approximately 150 times its original size within 15 seconds when exposed to water. The chemical reaction network utilizes a carbodiimide molecule, which transiently converts carboxylic acid moieties to anhydride bonds, causing transient shrinking of the hydrogel. The hydrogel spontaneously swelled due to hydrolysis of the anhydride bonds. We enhance the swelling rate of the porous hydrogels by adding polymer beads loaded with formaldehyde and sulfite buffer to the reaction vessel, which generate a time delayed pH increase that accelerates the hydrolysis of the anhydride bonds. This multi-reaction network forms porous hydrogels that contract and reswell autonomously in less than 10 minutes,compared to about 1 hour without the delayed pH change. This approach not only enables a faster reconfiguration of the porous gel induced by EDC through the hydrolysis of the anhydride due to a transient pH change, but also integrates this external stimulus into a single-step, autonomous process. This work also deepens our understanding of natural and synthetic actuation systems and how to couple different reactions to enhance their response. By harnessing the principles of bioinspired actuation, our work bridges the gap between the orchestrated movements of biological systems and the engineered behavior of synthetic materials, infusing our constructs with a level of adaptability and responsiveness that mirrors the natural world.Item HISTATIN 5 MODIFICATIONS IMPACT PROTEOLYTIC STABILITY IN THE PRESENCE OF FUNGAL AND SALIVARY PROTEASES(2024) Makambi, Wright Kingi; Karlsson, Amy J; Chemical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Candida albicans, found in the oral cavities of 30-50% of the global population, can lead to oral candidiasis, particularly in immunocompromised individuals like those with HIV or diabetes. The current treatments, small-molecule antifungals, often fall short due to drug resistance and toxicity. To address these challenges, histatin 5 (Hst5), a 24-amino-acid peptide naturally present in human saliva, has been studied as a potential antifungal therapy. Hst5, however, is susceptible to degradation by secreted aspartyl proteases (Saps) produced by C. albicans and salivary enzymes, limiting its potential efficacy as a therapeutic. We have engineered Hst5 variants utilizing rational design in order to understand the interactions with Saps and Saliva. We have also made advancements in developing a novel screening method utilizing the directed evolution technique yeast surface display. Our study employed rational design to modify Hst5, at its lysine residues (K5, K11, K13, and K17), substituting them with leucine or arginine to examine their influence on interactions with Saps (Sap1, Sap2, Sap3, Sap5, Sap6, Sap9, and Sap10). Sap5, Sap6, and Sap10 did not degrade Hst5 at the tested conditions, while Sap1, Sap2, Sap3, and Sap9 did. Some modifications, such as K13L, are particularly susceptible to proteolysis by Sap1, Sap2, Sap3, and Sap9. In contrast, K17L substantially increases the stability and antifungal activity of Hst5 in the presence of Saps. Additionally, although the K11RK17L variant was degraded more than the K17L variant, their antifungal activities were largely similar. The proteolysis products of were also identified by mass spectrometry identifying the [4-24], [1-17], and [14-24] Sap proteolysis products. We also evaluated the proteolytic stability of these variants in saliva. Both K17L and K5R showed improved stability; however, the enhancements were modest, suggesting that further engineering is required to achieve significant improvements. Further experiments evaluated how additional amino acid substitutions at K13 and K17 affect the peptide’s proteolytic stability in the presence of Saps (with and without zinc). Our findings suggest that the positive charge at K13 is important for the proteolytic stability of Hst5, as all other variants tested except K13R reduce overall proteolytic stability. Furthermore, many substitutions at K17, including tryptophan, significantly enhance proteolytic resistance and antifungal activity following incubation with Saps. The K17W variant showed improved stability and antifungal efficacy, maintaining its function even in the presence of zinc and exhibiting stronger antibiofilm activity than the parent Hst5. In addition to the rational design work, we have advanced the development of a directed evolution yeast surface display platform for screening peptides for proteolytic stability. This would allow for the expression of large peptide libraries on the surface of Saccharomyces cerevisiae. Through optimization of expression and display conditions, we determined an induction media at 30°C with a pH of 3.5 and devoid of glucose improved the expression and display of Hst5 peptides on the surface of S. cerevisiae. We also optimized the degradation conditions for Sap2 37°C, a pH not exceeding 7.4, and a Sap2 concentration of 0.78 µg/mL led to the best discrepancy between proteolytically stable variants. Additionally, we found that a 40 amino acid linker between the peptide and the yeast surface provided the best observing proteolytic degradation. Using the optimized system, we showed that yeast surface display can be used to discriminate between peptide variants with different levels of proteolytic stability. This lays the foundation for future work to screen large libraries of peptides for proteolytic stability. From these results, we have gained a deeper understanding of the interactions between Hst5 and Saps, showing that modification at different lysine residues greatly impacts the proteolytic stability of Hst5. Furthermore, we have shown that the yeast surface display platform can be used to screen the proteolytic stability of peptides. Looking forward, this peptide should be engineered for proteolytic stability in saliva. Furthermore, mock screens should be made before screening a library of peptides using the yeast surface display platform.Item Study of Membrane Binding Proteins and Related Signaling Molecules(2023) Allsopp, Robert James; Klauda, Jeffery B; Chemical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)The membrane contact site theory is a critical theory to understanding lipid transport. The Osh protein is a yeast lipid transport protein theorized to form membrane contact sites. We investigated the contact site theory by identifying a second binding domain and studying the Osh Amphipathic Lipid Packing Sensor (ALPS) to explain better why each protein might target different organelles. The α6- α7 domain appears more charged and prefers lipids with oppositely charged inositol sugars, making it ideal for binding to the Trans Golgi Network (TGN) and the plasma membrane. The ALPS peptide is another dedicated binding domain bound in several membrane types with varied Phosphatidylcholines (PC) tails to vary the lipid packing. If the force field was valid, the results indicate that Osh4 ALPS prefers the loose packing of POPC, and Osh5 ALPS prefers the tighter packing of DMPC. More input from the wet lab is needed before researchers can make predictions from the force field. Another vital area of research is antimicrobial peptides (AMPs) that disrupt the membrane. Part of the dissertation focused on determining the dual placement of the AMPs on the surface and inserted into the membrane. For the first time, the membrane properties of bilayers with AMPs were studied, using the combination of all-atom simulation informed by x-ray scattering. The surface tension was a critical parameter that enabled us to compare the simulation to the wet lab results and became vital in allowing the peptide to be inserted into the membrane and remain stable. The 5-HT3A project simulated predicted structures of toxins with computational tools. Our work simulated these toxins for the first time, and we observed the unbiased binding of σ-GVIIIA conotoxin to the allosteric binding pocket. In the first trajectories, the ion channel pore remained closed, similar enough to the native apo crystal structure that water could form a partially water-filled channel for a few microseconds. In one example, the 5-HT3A had serotonin in all of the binding pockets for close to 1 µs. The long simulation of the conotoxin showed that the extracellular domain (ECD) was deformed by more than a nanometer compared to a control. This deformation was the first indication that such a conformation is possible and might be related to the presence of the toxin. Finally, traumatic brain injury was studied by identifying new molecules that activate fibroblast growth factor (FGF) and toll-like receptor (TLR) proteins. The focus on FGF resulted in identifying a critical conformational change and potential new binding sites (previously unknown) that activate FGF without activating damaging inflammatory TLR responses.Item LITHIUM ANODE INTERFACE DESIGN FOR ALL-SOLID-STATE LITHIUM-METAL BATTERIES(2023) Wang, Zeyi; Wang, Chunsheng; Chemical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)All-solid-state lithium-metal batteries (ASSLBs) have attracted intense interest as the next generation of energy storage devices due to their high energy density and safety. However, the Li dendrite growth and high interface resistance remain challenges due to the lack of understanding of the mechanism. Inserting a solid-electrolyte interlayer/interphase (SEI) with high lithiophobicity, high ionic conductivity, and low electronic conductivity at the Li/SSE interface can solve these problems. However, how lithiophobicity, ionic conductivity, and electronic conductivity of the interlayers affect the lithium dendrite suppression capability of the SEI has not been systematically investigated yet but is critical for ASSLBs. The main goal of this dissertation is to propose a comprehensive interface design principle/frame by considering the impacts of interlayer lithiophobicity, electronic/ionic conductivity, and porosity to Li striping/plating behavior. A combination of modeling and experiments was used to validate the design principle. The developed principle could help to resolve the electrolyte reduction and void formation issues in all-solid-state batteries. The design principle can be applied to different solid electrolytes that have different reactivity against Li, which was presented in the 3rd-6th Chapters for detail. The interlayer design principle opens opportunities to develop safe and high-energy ASSLBs.In the 3rd chapter, we investigated the correlation among ionic and electronic conductivities, lithiophobicity, and Li plating stability in the Li7N2I-Carbon Nanotube (LNI-CNT) interlayer. LNI solid electrolyte has a high ionic conductivity of 3.1 × 10–4 S cm–1 and a low electronic conductivity, high lithiophobicity, and high electrochemical stability against Li, while CNT has a high lithiophobicity, high electronic conductivity, and low tap density. Therefore, mixing LNI with CNT at different ratios can form porous lithiophobic interlayers with variable ionic and electronic conductivity. The 90 μm LNI-5% CNT interlayer enabled Li to plate on the Li/LNI-CNT interface (rather than the SSE/LNI-CNT interface) and then reversibly penetrate into/extract from the porous LNI-CNT interlayer during Li plating/stripping. The 3-dimensional Li/LNI-5% CNT interlayer contact achieved by well-controlled Li nucleation and growth enabled Li/LNI/Li cell to charge/discharge at a high current density of 4.0 mA cm-2 and a high capacity of 4.0 mAh cm-2 for > 600 hours. We also reported that a stable Li plating/stripping cycle can be achieved if the Li nucleation region in the interlayer is smaller or equal to the Li growth region in the interlayer (from the Li anode). This study represents a comprehensive interlayer design for ASSLBs with a significantly improved dendrite suppression capability and reversibility. In the 4th chapter, we develop an LNI-Mg interlayer to increase the Li dendrite suppression capability of Li//Li cells with Li6PS5Cl solid electrolyte. LNI-25%Mg interlayer can form gradient electronic conductivity inside the interlayer due to Mg migrating from the interlayer to the Li anode during activation, which can reduce the interlayer thickness and enhance the Li dendrite suppression capability. The migration of Mg was attributed to the formation of LiMg solid solution. It was found that the gradient electronic conductive LNI-Mg interlayer has better Li dendrite suppression capability than the homogeneous electronic conductive LNI-CNT interlayer due to more constrained Li plating region and mitigated electrolyte reduction. As a result, 18.5 µm LNI-25%Mg interlayer enables Li4SiO4@NMC811/LPSC/Li full cells with an areal capacity of 2.2 mAh cm-2 to be charged/discharged for 350 cycles at 60 oC with capacity retention of 82.4%. This study promotes the development of ASSLBs with higher energy density. In the 5th chapter, we combined experimental techniques and simulation methods to investigate the relationship between the interlayer’s ionic/electronic conductivity ratio, lithiophobicity, and Li plating/striping behavior in carbon-based interlayers. Firstly, we screen the carbon materials based on their ionic/electronic conductivity ratio and lithiophobicity. Li stripping/plating mechanisms were identified in different carbon materials from simulations. Secondly, we predict the critical current density of the interlayer based on the boundary condition of avoiding Li nucleation during Li plating and void formation during striping. Finally, guided by the theoretical prediction, we optimized the ionic/electronic conductivity and lithiophobicity of the carbon-based interlayer by dopping with CuO. The CuO-CNF-M (M= Mg or Ag) interlayer in situ converts to Cu-Li2O-CNF SEI/LiM structure during Li plating. The optimized SEI with ionic conductivity of 0.41 S/m and electronic conductivity of 3.3×10-3 S/m coupling with LiM anode (in-situ formed during Li plating) enables lithium-free NMC811||Cu cell to achieve long cycle life. This work represents a valuable attempt to promote the development of high-performance Li anode interlayer with a joint effort of simulations and experiments. In the 6th chapter, we design a P and I rich SEI for halide electrolytes. Halide electrolytes have the advantage of matching with high-voltage cathodes due to the high thermodynamic oxidation potential. However, they are unstable against Li anode due to their strong reactivity with Li and the formation of electronic conductive metal. In this chapter, we propose and verify critical overpotential as a criterion for Li dendrite growth. By tuning the composition of the SEI, we reduce the overpotential to lower than critical overpotential using P and I containing SEI. The P and I containing SEI with a high ionic/electronic conductivity ratio of the SEI enable Li/LYbC/Li cells to cycle at the current density of 0.1 mA cm-2 with a capacity of 0.05 mA cm-2 for more than 220 hours without a short circuit. This work represents a valuable attempt to achieve Li-stable halide electrolyte.Item HYDROGEN SEPARATION AND CARBON CAPTURE BY CARBON MOLECULAR SIEVE MEMBRANES DERIVED FROM INTERFACIALLY POLYMERIZED POLYARAMIDS(2023) IYER, GAURAV MURALI; Zhang, Chen; Chemical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Due to its high energy density and zero-emission combustion, hydrogen (H2) has emerged as a clean fuel for energy generation and transportation. Also, H2 is an important chemical used in petrochemical refining, metal production, and fertilizer manufacture. In the United States, more than 10 million metric tons of H2 is produced each year by steam methane reforming, which gives 100 million metric tons of carbon dioxide (CO2) by-product. Downstream H2/CO2 separation is therefore needed to produce high-purity H2 product while simultaneously capturing the CO2 by-product. State-of-the-art separation technologies such as pressure-swing adsorption (PSA) and amine absorption are energy intensive with large footprints. Membrane-based H2/CO2 separation provides an energy-efficient alternative with smaller footprints. Commercial implementation of membrane-based H2/CO2 separation requires scalable membranes with high H2/CO2 selectivity to produce high-purity H2 product.The overarching goal of this PhD dissertation is to understand the formation and pore structure-transport property relationships in novel carbon molecular sieve (CMS) membranes derived from interfacially polymerized aromatic polyamides (polyaramids) for H2/CO2 separation. Polyaramid precursor hollow fiber membranes were fabricated by solution spinning of an uncrosslinked polyaramid precursor synthesized by stirred interfacial polymerization, which gave polyaramid-derived CMS membranes following pyrolysis. The formation, pore structure, and transport properties of the novel polyaramid-derived CMS membranes were systematically investigated. The polyaramid-derived CMS membrane pyrolyzed at 925 °C showed unprecedented H2/CO2 separation performance under single-gas permeation. Further increasing the pyrolysis temperature to 1050 °C dramatically enhanced the mixed-gas H2/CO2 separation factor to more than one order of magnitude higher than the most selective CMS membrane reported in literature. Modeling further demonstrates the attractiveness of the polyaramid-derived CMS membrane for enrichment of highly-pure H2 from the reaction product of steam methane reforming. Finally, the effect of precursor amide moiety on CMS membrane pore structure and transport properties was studied by comparing the polyaramid-derived CMS membrane with CMS membranes derived from a polyimide precursor and a polyamide-imide copolymer precursor under identical pyrolysis conditions. The results show that introducing precursor amide moiety is a powerful tool to tailor the H2/CO2 transport properties of CMS membranes via controlling the precursor hydrogen bonding and CMS membrane pore structure.