Skip to main content

INnovaTive Enzymes and polyionic-liquids based membRAnes as CO2 Capture Technology

Final Report Summary - INTERACT (INnovaTive Enzymes and polyionic-liquids based membRAnes as CO2 Capture Technology)

Executive Summary:
INnovaTive Enzymes and polyionic-liquids based membRAnes as CO2 Capture Technology (INTERACT) was a collaborative research project aiming at the investigation and development of new means for post combustion CO2 capture. The complex task was addressed with a comprehensive multiscale methodology that started with the development of economic synthesis techniques for the production of innovative and highly selective poly(ionic liquid)s (PILs) and enzyme-solvent systems. The results of the sophisticated material design and characterization of the developed materials were further incorporated in highly effective gas separation membranes and absorption systems based on packed columns and membrane contactors. Extensive experimental investigations from lab-scale to technical and pilot-scale were performed to demonstrate the implementation of the newly developed technologies, while detailed performance models were developed for each technology in order to aid the experiments and determine optimal operating conditions. Finally a variety of process concepts was developed based on the newly developed technologies and a process-wide evaluation of the potential impact was performed, taking into account the integration with the power plant and techno-economic as well as environmental criteria for the evaluation. The technical work was structured in seven work packages, which can be clustered in two material design and characterization work packages, three technology-focused work packages and two process related work packages. While the three clusters started in a staggered fashion, in order to account for the necessary information transfer from bottom to top, there was also a severe overlap of the processing periods of the three clusters, in order to allow for feedback and recourse from top to bottom. Consequently, the complete development process was highly integrated. Within the material design work packages a multitude of different PILs were developed and effective production processes were designed to produce the most promising materials in a capacity of up to 100kg. The most selective porous and dense PILs proved to be competitive sorbents for the use in adsorption processes as well as highly selective active layers in thin film composite gas membrane processes respectively. A full characterization of enzyme-solvent systems allowed for the identification of highly efficient systems in which the enzyme significantly enhances mass transfer in absorption processes. This makes it possible to exploit solvents with significantly lower heat requirements for regeneration. Furthermore, the combined effort of the material design and technology work packages made it possible to design tailored membrane contactors, which exploit both a selective PIL coating in combination with an enzyme enhanced solvent system in order to maximize the synergetic benefits and improve mass transfer in absorption processes. For each of the newly developed technologies suitable process concepts were developed in order to exploit the potential benefits, while simultaneously taking into account the additional restrictions introduced by the novel materials. These concepts were evaluated in the scope of an industrial CO2 emission scenario, whereas detailed techno-economic and environmental evaluations showed that the newly developed technologies do not only provide the potential to significantly reduce the environmental impact of coal-fired power plants and other CO2 emitting processes, but also outperform the current state-of-the-art amine scrubbing process.
Project Context and Objectives:
In the last decades, CO2 capture gained significant attention to minimize the environmental impact of post-combustion processes, while conventional technologies like chemical scrubbing with amines face severe economic boundaries due to their high energy consumption. Novel materials, such as (poly(ionic liquid)s (PILs) and enzymes) implemented into innovative technologies (e.g. membrane contactors) can provide a promising solution, to reduce the associated energy penalty and the cost of carbon capture. Exploiting the synergies between the high CO2 capture affinity of the materials and the intensified technologies can facilitate the effort to make CO2 capture industrially feasible in the future.
General concept
The general aim of INTERACT was to investigate and introduce new means to capture CO2 in the post combustion process. Therefore developments of both, PILs and enzymes, as well as their integration into membrane and/or absorption systems was pursued. In-depth investigations of these separation technologies were performed in order to determine their performance with regard to separation effectiveness, durability, scale-up and finally, when embedded in suitable process concepts, techno-economic and environmental performance. Consequently, extensive experimental investigations were augmented with sophisticated and rigorous modelling, which was finally utilized to perform conceptual design studies for the development of optimized process concepts, which were to be rated by techno-economic as well as environmental criteria, making use of a detailed life cycle analysis. While the environmental benefit from CO2 capture technologies does not stand to question, the currently favored chemical solvent scrubbing based on activated amine solutions, suffers from a high energy requirement for solvent recovery, as well as solvent degradation, resulting in high capital costs as well as an energy penalty of about 30 – 40%. The general concept of this project was to open new pathways for the development of novel high-potential processes for post combustion CO2 capture based on new materials, in order to significantly reduce the energy penalty, as well as capital and operating costs associated with the CO2 capture process. Therefore, INTERACT combined two different classes of key materials integrated into three intensified technologies to enable synergy effects which provide promising new alternatives and “proof of concept” for CO2 capture. The innovative technologies were developed and tested to realize the “proof of concept” by demonstrating experimentally the feasibility and performance of the new CO2 capture prototype technologies and to show their economic benefits compared to conventional state-of-the-art processes by model-based analysis.
In the INTERACT project a variety of dense and highly porous PILs were synthesized and studied in a variety of CO2 capture strategies, designed e.g. as active separation layers in thin film composite membranes, that enables the possibility to integrate the CO2 affinity of the PIL in the membrane in order to enhance CO2 permeability and selectivity. With additional additives it was possible to synthesize material formulations that are located close to the upper bound on the Robeson plot. The other material-focused idea in INTERACT was to use the enzyme carbonic anhydrase, a nature originated bio-catalysts that handles CO2 in most mammals and other lifeforms, as an accelerant to support the CO2 absorption. This opens the opportunity to use solvents with a significantly lower desorption temperature, and thereby lower energy penalty of carbon capture. Several suitable enzyme-solvent systems have been identified and the improvements and restrictions of enzyme implementation have been quantified. Overall it was shown that a significant reduction of the energy penalty by several percent points should be feasible in a suitable process concept accounting for the restrictions of enzyme implementation.
The tailored gas separation membranes in INTERACT consist of a porous support layer, coated with a dense PIL layer. The PIL layer preferably adsorbs CO2, whereas the CO2 subsequently desorbs due to the partial pressure difference on the permeate side of the membrane and hence is transported through the membrane. Support layers were carefully designed from different materials, accounting for porosity, as well as chemical and mechanical stability and taking into account to produce the thin film composite membranes, including in-situ polymerization as well as dip or spray coating. The separation performance was further characterized and several configurations, including flow patterns and possible means for establishing the driving force have been evaluated in process design.
In INTERACT a combination of the classical chemically enhanced absorption process with new solvent systems exploiting the possible rate-enhancement by enzymes was investigated. Enzymes were either dissolved in the solvent systems, or immobilized in catalytic packings as well as other forms of immobilization. Application of enzymes in absorption columns was proven at different scales, from small lab-scale to large pilot scale set-ups, and the obtainable rate-enhancements have been demonstrated for solvent systems with dissolved enzyme as well as different forms of immobilized enzyme. Besides the application of classical columns furthermore the application of membrane contactors was considered for absorption. Membrane contactors can overcome some of the major drawbacks of columns, since they present a physical barrier between gas and liquid phase allowing a dispersion free contact, while providing an extremely high specific surface area. However, the membranes cause an additional mass transfer resistance and sensitivity towards fouling, resulting in a limited lifetime1, which has to be considered in a careful selection membrane materials. Besides the direct application of enzyme-activated solvents in non-selective membrane contactors a further integration of the innovative materials was considered in the INTERACT project, whereas a porous membrane was coated with a PIL layer in order to exploit the benefit of both materials, resulting in an enhanced driving force across the membrane for the hybrid membrane contactor. Since harsh desorption conditions may result in a deactivation of the enzymes, several desorption strategies were further carefully investigated, including the aforementioned avoidance of enzyme in the stripper by either immobilization of the enzyme in the absorber or recovery of the enzyme prior to the stripper by e.g. ultrafiltration membranes. A variety of different process concepts was developed, including the integration with the power plant and an evaluation of the options based on techno-economic as well as environmental performance.
The project structure allows for the exploitation of synergies at different levels (see Figure 1). At the technical level the addition of a PIL-based layer on a support is one of the key ideas to perform a targeted functionalization of gas membranes and membrane contactors for selective removal of CO2. The integration of enzymes into absorption columns allows for the exploitation of solvents which require less energy for the recovery in desorption without losing the efficiency of reactive solvents such as amines. Integrating enzymes into non-selective membrane contactors exploits the large surface area created by the membrane resulting in more compact equipment, while the implementation of a selective PIL coating combines the large surface area of the membrane and the selective and highly efficient enzyme systems with the driving force enhancement by the PIL layer. However, besides these technical advances, synergies on the knowledge level can be exploited due to the collaboration of experts from various research fields, bridging the gap between material designs, technology development and process engineering in order to determine an optimized solution for post combustion CO2 capture.

Figure 1: Conceptual approach of the INTERACT project
The major project objectives of the INTERACT project were the
1. Targeted enhancement of CO2 removal by the development of functionalized stable novel sorbent materials to aim for overall cost savings of 50% compared to a benchmark absorption column process using MEA as solvent.
2. Increased mass-transfer intensity in terms of increasing capacity of the solvent by up to 30% and enhancing mass transfer kinetics by at least a factor of 2 compared to the non-activated system by identifying novel stable biological activators in solvents based on reliable and robust datasets.
3. Development of highly efficient and stable recovery systems for activators and functionalized materials by immobilization or use of additional membranes.
4. Targeted creation of large phase contact areas.
These objectives were pursued by investigating innovative technologies, including gas separation membranes functionalized by a dense layer of PILs, absorption columns using enzyme activated solvents and immobilized enzymes and membrane contactors using enzyme activated solvents, as well as membrane contactors functionalized by a layer of poly(ionic liquid)s using conventional solvents as well as enzyme activated solvents
Only the most promising ones evaluated by simultaneous analysis of these technologies through calculations of energy penalty and economics were to be tested and validated as “proof of principle” by prototype testing in order to overcome the following technical and economic challenges:
1. Lack of reliable data on materials and their interaction with support materials in operation.
2. Lack of robust and reliable models, considering chemical reaction, fluid flow, mass transfer and energy transport
3. Lack of detailed understanding of key-technologies required for process design
4. Lack of strategies for effective integration of materials and scale-up of equipment
In order to overcome these challenges, the INTERACT project aimed at breakthroughs in the key activity areas of its integrated R&D-framework. More specifically, breakthroughs in the following scientific and technical domains were targeted:
1: New materials
Milestone: Enabling new materials for CO2 capture by the development of stable poly(ionic liquid) layers at lab-scale, the identification of stable combinations of IL-PIL-solvent-enzyme- membrane systems as the basis for the integration into INTERACT technologies: gas separation membranes, membrane contactors with and without enzymatic solvents and for testing at larger scales by the establishment of dedicated processing. After the identification of the ILs to PIL's (and blends of PIL's with IL and salts), the most suitable materials for CO2 absorption are to be tested as either highly porous materials or dense layers, integrated in gas separation membranes and membrane contactors. Different set of IL’s/PIL’s may be selected, while only the most promising are to be tested to account for different CO2 capture technology requirements.
2: New equipment for the technologies
Milestone: Development and testing of stable enzyme-activated new column internals at different scales, including laboratory scale-columns (Diameters < 25mm) and scalable pilot-columns (Diameter = 100mm). Enzymes will be investigated in solution as well as immobilized, taking into account surface coatings of packings and implementation of enzyme beads embedded into envelops into structured packings. The immobilization should allow for good contact between enzyme solvent and gas, while avoiding enzyme transfer to the desorber. Enzyme activity, loss and CO2 recovery rates are to be evaluated and the most promising enzyme supported CO2 absorption method in columns will be defined with regard to energy penalty and stability.
3: New technologies
Milestone: Development of new capture technologies exploiting synergies of enzymes and PILs with novel gas separation membranes and functionalized membrane contactors. Based on the newly developed materials, gas separation membranes as well as membrane contactors will be functionalized with PIL layers to enhance their performance. The complex interactions between the PIL layer and membrane support need to be fully understood to prevent leaching and other stability problems, which will result in the development of stable prototypes. A complete performance characterization and model development of each technology will be presented.
4: Multiphase, multiscale modelling
Milestone: Development of a computational model library for multiphase, multiscale models. The model library will comprise all models of different scale for each of the developed technologies including adsorption/desorption phenomena, kinetics and unit operation models and will be used for scale-up, integration as well as techno-economic analysis. All models will be validated by prototype testing of the new equipment and iteratively refined by more detailed reliable datasets at ideal and spiked process conditions. The fine-tuned final models are used for the evaluation of various process concepts based on the new technologies.
5: Assessment of new technologies in industrial application scenario
Milestone: Techno-economic analysis and LCA methodology set-up for investigation of INTERACT methodology. For quantitative decision-making at different stages of the INTERACT project, metrics and tools are used as support in order to provide feedback for technology research to focus on the most promising directions and to decide which technology or which combination of the most promising technologies can be integrated economically and under LCA metric into existing plants. The results will consequently help to guide current as well as future research.

The project was planned for a period of 3.5 years. The members of the consortium were carefully selected to warrant the impact by the participation of highly innovative multinational industrial companies (NZDK/NZNA), research institutions (SINTEF, CID, ICE-PAS, PROD), universities (TUDO, DTU, KULEU) and KMUs (SOLV, SUPREN). The relevant scientific foundations for the development of PIL and enzyme-solvent systems (CID, SOLV, NZDK/NZNA, DTU), the development of gas membranes and membrane contactors (KULEU, SINTEF) and the investigation and characterization of the different technologies (TUDO, DTU, ICE-PAS) were thoroughly covered and aided by the experience in chemical, mechanical and computer aided process engineering (SUPREN).
Project Results:
WP1: Synthesis and characterization of poly(ionic liquids)
The main objective of WP1 was the development of tailored poly(ionic liquid)s (PILs), which were further applied as active materials in various CO2 capture technologies. Therefore, the main activities have been focused on the synthesis and characterization of the PILs that represent a subclass of polymers constituted by ionic liquid (IL) monomers (Figure 2). As salts comprised of ions with a wide range of poorly coordinated structures, ILs are liquid below 100°C, or as so-called room temperature ionic liquids (RTIL) even at room temperature. Since ILs show selective affinity towards CO2 over N2 or CH4 due to their polarity, they have been investigated widely in CO2 capture technologies. Since ILs are considered expensive and their recovery accordingly important, PILs, which retain the versatility of ILs, but they can be processed as thin films and used as membrane materials in gas separation or as solid sorbents present a huge potential, which is further fostered by their selective, fast and reversible sorption of CO2.

Figure 2: Schematic representation of a poly(ionic liquid).
In the scope of the project, two different types of PILs have been developed. On the one hand, dense (non-porous) PILs have been synthesized and used as active layer in gas separation membranes (WP3) and membrane contactors (WP5), providing a thin dense semi-permeable barrier that separates gas components based on differences in solubility and diffusivity. While most of these PILs have been synthesized by polymerization of ionic liquid (IL) monomers and post-functionalization of neutral polymers and electrolytes, others have been formulated as blends comprised of ILs, PILs and additives that were known to increase permeability. On the other hand, porous PILs have been developed and tested as CO2 sorbents.
Various IL monomers have been taken into account for the different approaches to prepare highly porous and dense PILs. The following types of monomers have been selected for the synthesis:
• Polymerizable ILs (carrying one polymerizable group) for the synthesis of dense PILs: a family of vinylimidazolium based ILs with pending glycol and aliphatic side-chains; methacrylate based ammonium based ILs.
• Cross-linker type ILs (carrying two polymerizable groups) for the synthesis of highly cross-linked porous PILs: imidazolium based dicationic ILs and ammonium based dicationic ILs.
• ILs with a functionalization-site for anchoring to high surface area materials: imidazoliun base IL with alkoxysilane pending group.
Porous PILs
Various strategies for the preparation of porous materials (or high surface area materials) were investigated, taking into account hard templating, complexation of PIL/polyelectrolytes, functionalization of alumina nanoparticles with ILs, and PIL nanoparticles. Some of the approaches turned out to be time- and energy-consuming and more importantly, difficult to scale up. Looking for a simple, fast, sustainable, and scalable strategy to prepare nanoporous materials based on PILs a new approach has been developed. The synthetic strategy relies on the radical polymerization of crosslinker-type IL monomers in the presence of an analogous IL, which acts as a porogenic solvent (Figure 3).

Figure 3: Schematic representation of the synthesis of porous PIL: cross-linker type IL (left) and IL (right).
This IL solvent can be extracted easily after polymerization and recycled for further use. The great advantages of this synthetic approach are the atom-efficiency and absence of waste. An illustration of this concept as well as a FSEM image of one of the produced PILs is shown in Figure 4. The effects of different monomer/porogen ratios on the specific surface area, porosity, and pore size have been investigated in the scope of the material development.

Figure 4: Schematic representation of the synthesis of porous PIL and FESEM photographs of a porous PIL (80 m2/g).
Although the obtained surface area values (up to 80 m2/g) were lower than those reported for materials obtained through the hard-templating pathway, for which specific surface areas of 150-220 m2/g were reported, the present methodology still remains valuable from the point of view of preparation efficiency and practical convenience. Finally, the potential of the materials as CO2 sorbents have been evaluated showing competitive performance and enhanced kinetics for the porous materials when compared with other sorbents reported in literature. The sorption capacity of the PILs can further be enhanced by modifying the chemical structure of the IL monomers, selecting cation, anion, and core structure with higher affinity towards CO2, for example, by increasing the ion polarity.
Dense PILs
Two main strategies to synthesize dense PILs were followed, which are a polymerization of IL monomers and a post-functionalization of existing polymers. The first approach entails the IL being initially synthesized and purified at the monomer level. Homo-polymers and combinations of co-polymers can be derived by varying the chemical content of IL monomers. However, the control of molecular weight can turn out to be a major challenge. The post-functionalization on the other hand allows for the conversion of commercial polymers into customized PILs and involves the post-synthetic modifications of neutral polymers through chemical quaternization (e.g. poly(vinylimidazole), poly(vinylbenzyl chloride), poly(diallyl methylamine)) and/or anion exchange reaction, such as poly(diallyldimethylammonium chloride) (P[DADMA][Cl]). During the project different families of PIL have been synthesized and evaluated for gas separation applications, here we present most representative families that have been tested as thin gas separation membranes: PIL/IL/additive blends, PILs derived from commercially available poly(vinylbenzylchloride) and PILs derived from renewable source cellulose.
In the project a new class of composites comprised of PIL/IL/additive blends have been prepared and characterized, starting from a commercially available polyelectrolyte through an anion exchange procedure, which is simple and straightforward. The incorporation of the additive to the PIL/IL mixture showed an improvement in both CO2 solubility and CO2/N2 selectivity. CO2 and N2 permeation coefficients were determined based on the relevant solubilities and diffusivities as measured gravimetrically at 20°C. Most importantly, it was found that an increase in the additive concentration leads to an increase in the CO2/N2 selectivity by a factor of 2.7 (compared with the original PIL/IL material), thus locating the new formulations close to the upper bound on the Robeson plot. A specific formulation, F9:1, showed favourable separation capabilities over a wide pressure range (0 – 18 bar), at temperatures below 40°C. Moreover, composite materials that contain the additives offer CO2 permeabilities and as high as 153 – 428 barrer and CO2/N2 selectivities as high as 27 – 82 (depending on temperature and pressure), providing a huge potential for removing CO2 from flue gas streams. F9:1 also showed the best performance in applications (WP3&5). Another PIL having the same polycation backbone but different anion, a simple carboxylate counterpart, PIL-D-14 is highly hydrophilic and showed higher selectivity (αCO2/N2) but considerably lower permeability, mechanical properties and overall performance.
A direct approach to prepare PILs by avoiding the polymerization of IL and the related inconveniences, such as molecular weight control, is the functionalization of neutral polymers. A wide range of commercial polymers can be derivatized into PILs. In the INTERACT project, the well available poly(vinylbenzylchloride) has been selected as starting material due to several advantages, having the poly(styrene) backbone, which enhances the CO2 separation performance in comparison with other PILs based on the CO2 and N2 sorption capacities , as the basis of material synthetic procedures, as well as a reactive site to react with amines to create ionic moieties. Three different poly(vinylbenzylchloride) analogues were prepared with IL-sites. An ammonium-derivatised polymer-analogue was prepared as a first PIL-type by polymerisation of an IL monomer, whereas two other PILs were synthesized from poly(vinylbenzylchloride) using a tertiary amine with a pending hydroxyl group and a cyclic tertiary amine. Introduction of fluorinated anions was accomplished by anion metathesis. A thorough characterisation of the material structure, composition, membrane morphology and gas separation properties analyzed in WP3 showed that that mixed-gas selectivity increases with the larger positive charge density of the cation species and that the presence of hydroxyl groups in the polycation enhanced the interaction with CO2 molecules. In general, the separation performance of the polymeric membranes is significantly affected by the process conditions, e.g. presence of water vapour in the feed stream. As a matter of fact, the experiments performed in humidified conditions revealed a doubled CO2 permeability in comparison to dry conditions.
Another material that was investigated for PIL synthesis is cellulose, which is not only one of the most abundant and sustainable resources available, but cellulose acetate (CA) membranes are already commercially used for large scale CO2 separation, such as natural gas sweetening. Nevertheless, the development of novel materials in the scope of the industrial use of cellulose and its derivatives is an active field of research (Figure 5). There are promising attempts showing that the functionalization of polymers, such as polybenzimidazoles, with ionic moieties leads to favourable synergies for CO2 separation. Hence, the synthesis and characterization of a novel cellulose derived PIL family as a CO2 separation membrane material was investigated in detail.

Figure 5: Schematic representation of various CA modified membranes.
The cellulose acetate derived PIL showed interesting preliminary gas separation properties (two-fold increased permeability compared to cellulose acetate) and features several advantages compared to other cellulose acetate derived materials that combine ionic liquids. Building on the synthesis that starts with the post-functionalization of a renewable, cheap and well-known industrial raw material (cellulose acetate) that comes without the need of controlled polymerization reactions the ionic groups are covalently grafted to the polymer matrix. In order to further investigate the potential of this material a family of related PILs was developed by changing the nature of the pending cations and incorporating chemical bonds that enable cross-linking. The effect of the cationic moiety chemical structure on the CO2 sorption capacity was investigated in gravimetric experiments and as TFC membranes for CO2 removal from flue gas (WP3). Those materials showed selectivities comparable to first PIL with stable CO2 permeabilities over the range of CO2 partial pressures in the experiments with different volumetric composition of feed gas and different applied pressures).
Improved synthetic strategies for economic, scalable and feasible industrial production of functionalized PILs
Throughout the project the selection of IL monomer and PILs was performed by carefully designing reaction pathways in view of using non-toxic and industrially available raw materials as far as possible (REACH, CMR products, safety issues are taken into account) and in consideration of a necessary scale-up of the production process. For the demonstration of the upscaling potential, the production of poly(diallyldimethyl ammonium) bis(trifluoromethanesulfonyl)imide), one of the promising PILs, was scaled up from 100 g batches up to 1kg batches. While bigger reactor sizes were not found to be efficient in reproducing the same quality than lab scale, the amount of PIL is sufficient for large-scale gas membrane applications, taking into account that only the active layer of the membrane has to be produced from the PIL. By taking into account new mixers and equipment options the scale-up was successfully performed and commercialization of this PIL has started. The developed equipment is flexible enough, such that other powders and additives can be scaled up, as well. Further improvement of the synthetic strategies for economic, scalable and feasible industrial production of functionalized PILs was carried out through a 1kg and 100kg production simulation for ten compounds.

Mathematical description of sorption/desorption phenomena of main components of flue gases on mesoporous PILs
A mathematical description of sorption/desorption phenomena of main components of flue gases on mesoporous PILs and its validation against experimental data concerning carbon dioxide sorption on the first set of mesoporous PILs has been carried out. The dual-mode model was developed which is widely used to describe gas solubility and transport in glassy polymers and applied also in the case of CO2 and other gases sorption into PILs. The model adopts the concept of sorption into two idealized environments: dissolved and “microvoids” explaining in that way a non-linearity observed in sorption isotherms of gases on PILs. The sorption in a dissolved environment is described by Henry’s law relation and the sorption in the “microvoids” is described as the Langmuir adsorption or “hole-filling” process. Apart from dual-mode model describing sorption equilibria the linear-driving force model was also applied in order to determine diffusivities from uptake curves accompanying each experimental point of sorption isotherms. The dual-mode model was validated against experimental data of CO2 sorption on first sets on mesoporous PILs based on imidazolium bis(trifluorosulfonyl)imide and ammonium bis(trifluorosulfonyl)imide and prepared using different monomer-to-porogen ratios. Coefficients of the dual-mode equation were determined and it was shown that the model describes experimental data with a very good accuracy. The physical meaning of the coefficients was also discussed as well as their usefulness in determining some other important parameters such as heat of sorption.

WP2: Characterization and functionalization of enzymes
The main focus of WP2 was on the development and characterization of enzyme-solvent systems for the implementation in enzymatically enhanced reactive absorption processes. A selection process was established to determine technological applications where rate enhancement can be practically and economically exploited by screening combinations of suitable enzyme-solvent systems, evaluating options for retaining/recycling the enzymes, and characterizing enzyme stability and kinetic performance under operating conditions.
Enzyme stability
Being considered as a major issue for the application of enzymes in industrial environments, the first investigations focussed on the investigation of enzyme stability, which depends on pH, temperature and the selected solvent system. Since the temperature of the solvent changes while passing through the different process units, the highest temperatures are encountered in the stripper, which in conventional CO2 absorption processes is operated well above 100 °C. In addition to temperature the pH of the solvent system changes as well, depending on the CO2 loading. Since CO2 is acidic, the highest pH is encountered at the lean loading of the solvent, after leaving the stripper and prior to entering the absorption (Figure 6).
The main results from the investigation of the performed enzyme stability tests are summarized in Figure 7 and Table 1, which illustrate the dependence of enzyme stability on pH and temperature, as well as the results from long-term stability trials for various investigated enzyme-solvent systems. It can be concluded that the enzyme retains more than 90% of its initial activity when operating conditions are confined to a range of pH between 7 and 11.

Figure 6: Illustration of temperature and pH greadient over the course of an absorbtion/desorption cycle.
Full retention of enzyme stability was experienced at temperatures up to 50°C, while minor deactivation up to a 75% activity resulted from a temperature increase to 60 °C. At higher temperatures the residual enzyme activity was dropping below the level of detection within a timeframe of 100hrs. Consequently high activity of the enzymes can be maintained when process conditions are confined to the determined stability ranges.

Figure 7: pH and temperature stability of CA.
The results of a long-term investigation (150 days) of six promising aqueous solvents (Monoethanolamine (MEA), 2-Amino-2-methyl-1-propanol (AMP), N-Methyldiethanolamine (MDEA), 2-aminoisobutyric acid (AIB), Potassium carbonate (K2CO3), and chlorideAmmonium chloride (AC)and the solvent 3-(methylamino)propylamine (MAPA) used together with MDEA as a mixed solvent) are presented in Tble 1. These solvents were chosen due to their reported positive results for CCS application. All solvents otherthan K2CO3 were used with a 3M concentration, wheras K2CO3 was applied at lower concentration due to solubility problems due to which the enzyme precipitated with the solvent resulting in a drastically reduced enzyme activity. Equal solvent concentrations were chosen to give a fair comparison based on the fact that CO2 is absorbed with a 1:1 ratio with these solvents. It should be noted that due to a different reaction mechanism, MEA absorbs 50% less CO2 per mol, than the other solvents. The pH’s used is the, investigating enzyme stability in the pH ranges between 8-10. The lowest pH represents the fully loaded solvent. The combined effect of temperature, solvent concentration and solvent type was tested.
It can be concluded that within the defined operating conditions the resulting enzyme stability shows great promise towards implementation, justifying the further investigations in WP4 and WP5 for application in columns and membrane contactors.
Table 1: Stability with different solvents after 5 and 150 days

Enzyme kinetics
While sufficient enzyme stability is a prerequisite for the application in an industrial environment the enzyme needs to add sufficient benefit to the process to justify the limited operating range. In order to quantify and model the obtainable benefit from enzyme application, the mass transfer of CO2 in different solvents was measured in a wetted wall column and the kinetics of the enzyme in the solution were determined taking into account the well-defined interfacial area in the experimental equipment. Based on the previous screening of enzyme-solvent systems the primary amine MEA, the tertiary amine MDEA and the sterically hindered amine AMP were investigated in aqueous solutions with 30 wt% solvent, while due to limited solubility the carbonate salt K2CO3 was investigated in an aqueous solution with 15 wt% solvent.

Figure 8: Mass transfer enhancement due to 2 g/L CA addition

CO2 mass transfer was quantified for the solvents with and without the addition of 2 g/L CA. In order to quantify the benefit of the enzyme addition the so-called catalytic effect (CE), which represents the ratio of the measured CO2 transfer with and without the added enzyme, has been determined for each enzyme-solvent system. The results are summarized in Figure 8. Obviously a significant enhancement of the mass transfer can be achieved by enzyme addition in the MDEA and K2CO3 solvent systems, whilst the improvement in AMP and MEA solvent systems is small or negligible under the conditions of these tests.
Based on the promising results achieved (including the stability results) and beneficial thermodynamic properties (i.e. high equilibrium CO2 loading), the solvent MDEA was chosen to be further investigated in packed columns at pilot scale in WP4. Building up on continuous experiments a mechanistic enzyme kinetic model was derived, based on a simplified Michaelis Menten type kinetic with product inhibition by the bicarbonate ion for CA, which can efficiently describe the effect of temperature, solvent concentration, enzyme concentration and solvent loading on the mass transfer. The quality of the derived model is illustrated by comparison to experimental results in Figure 9. This derived kinetic model was further implemented in a rate-based absorber column model in WP4.

Figure 9: Model for the liquid side mass transfer coefficient of enzyme enhanced MDEA

Enzyme immobilization
In order to protect the enzyme from the harsh conditions, especially in terms of temperature, that are usually experienced in the stripper, different concepts for immobilization of the enzyme in the absorption step were investigated. Free enzymes in solution either as soluble enzymes or immobilized on particles distributed in the solvent, will experience a variety of conditions throughout the process. Effective immobilization can provide increased stability of the enzyme by restraining the enzyme to preferred ranges of pH and temperature corresponding to the operating conditions of the absorber, resulting in increased lifetime of the enzyme. However, immobilization adds cost in developing the formula and producing the immobilized form of enzyme, and can cause additional mass transfer limitations. Finally, replenishing immobilized enzyme might introduce additional complexity in a capture process and require appropriate scheduling.
The first investigated immobilization strategy was a surface coating of an enzyme polymer mixture on the structured packing. Due to insufficient compatibility of the enzyme CA with the coating material this strategy was however not successful. Furthermore, another concept that was investigated was the immobilization of enzymes on magnetic nanoparticles, however this option was also discarded after initial experiments. A third concept the immobilization of enzymes on silica granules via spray coating, forming so-called enzyme beads, which can be filled into pockets of specific catalytic structured packings, such as Sulzer Katapak SP, presented a very promising and well working solution. Stability investigations for the immobilized enzyme were performed with a specially constructed test rig, in order to evaluate the mechanical stability of the particles as well as stability of the enzyme immobilization on the particle. These investigations were performed in a glass column equipped with catalytic packing operated in a closed loop setup with a constant solvent flow supplied to the top of the column, while counter-currently contacted with a CO2 containing gas stream and an aqueous K2CO3 solvent at temperatures between 20-40°C. Consequently, the particles experience comparable mechanical and chemical stress as in a later investigated pilot plant application. The total washout did not exceed 15 wt%, whereas no further washout was observed after a period of 8hrs. The application of this form of immobilization can therefore be recommended from the perspective of applicability and stability of the enzyme beads concerning mechanical and chemical stress. Nevertheless, a novel form of immobilized CA became available for testing through the strong extended network of INTERACT partners at the end of the project which provided even bigger prospect. This material, tested at the end of WP4, showed significant promise as a new way to provide highly active CA to the system for good absorption performance over long time periods (see WP4 for further information).
Enzyme recovery– modelling
Finally, the use of an additional separation step to recover dissolved enzyme prior to entering the stripper was modelled based on calculations adding an ultrafiltration (UF) membrane and investigating the effect of different enzyme recoveries at various operating temperatures for the stripper. Atheoretical investigation was performed, using experimentally determined enzyme deactivation rates from INTERACT, which were modelled by a first order reaction rate. The results of this study are exemplarily illustrated in Figure 10, showing that fast deactivation of the enzyme is expected at stripper temperatures above 70°C, requiring frequent exchange of the enzyme solvent-system unless the enzyme is immobilized or recovered by an additional UF separation.

Figure 10: Residual enzyme activity after one year of operation without enzyme recovery by a UF membrane at stripper temperatures of 60 °C (blue diamond), 70 °C (Orange squares), 80 °C (grey triangles), 90 °C (yellow circles) and 100 °C (blue dashes) [left] as well residual enzyme activity after one year of operation with enzyme recovery by a UF membrane at a stripper temperature of 90°C, assuming enzyme recovery of 90% (orange squares), 99% (grey triangles) and 99.9% (yellow circles) . Application without a UF membrane is shown as reference (blue diamond).

WP3: Gas separation membranes
The major objective of WP3 was to implement the dense PILs developed in WP1 as the selective layer of thin-film composite (TFC) membranes for the separation of CO2 from flue gas streams. An illustration of the concept is given in Figure 11. The PIL was applied as a film with a micrometre thickness on a porous support that provided mechanical resistance to the applied pressure difference. Different TFC membranes were developed using either in-house prepared or commercially available polymeric supports. Various factors were taken into account when selecting appropriate support materials, seeking negligible surface roughness to allow the formation of defect free PIL layer, and stability in the desired solvents during casting to avoid disintegration of the membrane morphology. Besides the separation performance of the selective PIL, physical and mechanical parameters of the layer were considered in terms of film-forming properties and integral stability.

Figure 11: Concept of PIL based composite membrane for CO2 separation from flue gas.

For an initial investigation of the separation properties, the flat-sheet membrane geometry was considered as most suitable, as it also enables simplified membrane fabrication via physical layer deposition techniques, such as solvent casting and spray coating of the PIL directly on to the support. This efficient method for preparation and investigation of TFC membranes was used to determine the optimal selective layer materials with good characteristics both in terms of the physical processing (i.e. film formation, synthesis scale-up, expenses) and CO2 separation performance (i.e. permeability, and selectivity). A synthetic flue gas stream with a CO2/N2 molar ratio of 15/85, 1.2 bar feed pressure and a process temperature of 26 ºC has been chosen for the investigation of separation performance. The acquired data on membrane separation performance was further processed to derive and verify mathematical models for the gas permeation membranes. The most promising samples were finally tested under pre-defined real gas conditions, taking into account a humidified gas with realistic temperature. CO2 concentration, and traces of SO2 and O2 as would be expected in the flue gas of a coal fired power plant.

Development of the membrane prototype
Different membrane prototypes were prepared by application of active membrane layers consisting of PIL materials developed by CIDETEC and SOLVIONIC. While an extensive investigation of commercially available supports was conducted at SINTEF, lab made polymer supports that showed sufficient resistance towards solvent swelling, were considered at KU Leuven. In selecting commercially available supports the focus was placed on slightly hydrophobic supports in order to prevent pore penetration of the hydrophilic PILs inside the support. Pore penetration decreases the gas permeance and simultaneously may result in defects in the coating layer, which significantly reduced CO2/N2 selectivity. The selection of supports further addressed stability, ease of handling, cost, reduced surface roughness and good wettability. Membranes of various PIL materials were successfully fabricated using casting and ultrasonic spray coating procedures forming continuous films in the range of 0.5 – 20 µm thickness. The fabrication of TFC membranes using ultrasonic spray coating procedures to disperse the selective PIL layer on the membrane support is a relatively new concept with few reports of this in the literature. It is also a scalable technique and interesting from an manufacturing standpoint.

Screening in lab-scale experiments under dry and humid conditions
Mixed-gas selectivities and CO2 permeance were measured during permeation experiments with a gaseous mixture of 15% CO2 in N2 for PIL-based TFC membranes mounted in a custom-built high-throughput gas separation (HTGS) flat sheet membrane module at KU Leuven. The HTGS setup is shown in Figure 12. The screening was carried out in dry conditions with up to 5 bar feed pressure to facilitate the gas transfer across the separation barrier.

Figure 12: Overview of the HTGS set-up with compact GC (left) and detail of the 16 position membrane module (right).

Furthermore, permeation tests were performed in a custom designed test rig at SINTEF, using similar conditions to real flue gas conditions (15% CO2 in N2, 90% humidity). Although different membranes (different supports and membrane thickness as result of different preparation procedures) were tested, it became evident that compared with dry gas experiments the humified conditions influence the membrane separation performance. The key figures from two of the most promising membrane materials PIL_D_14, giving high selectivity, and F9:1, giving high permeability, as well as high selectivity, are highlighted in Table 2 below. The data sets comprise performance characteristics for both dry gas measurements, and humidified gas. While the selectivity of the membranes prepared with F9:1 seems independent of humidity and membrane thickness, there seems to be an increase (about a factor 2) in permeability at humified conditions. For PIL_D_14 membranes on the other hand, humified conditions seem to affect both the selectivity and permeability.

Table 2: Performance characteristics for the most promising PILs chosen for further investigations
PIL Partner δ a αCO2/N2 b PCO2 c Permeance d
[μm] [-] [Barrer] [m3(STP)/m2 bar h]
F9:1 KULEU 1.8 34 217 0.333
SINTEF 11.7 33 464 0.0130
PIL_D_14 KULEU 1.7 24 8 0.0130
SINTEF 2.0 58 247 0.034
(a = Thickness of the active layer, b = Mixed-gas selectivity, c = Permeability of carbon dioxide, d = Permeance of carbon dioxide)

Mathematical model of the gas permeation through the membrane under spiked process conditions
In order to evaluate the potential of the PIL-based gas membranes and develop suitable process design configurations in WP 6&7, a mathematical model of a gas membrane module was developed in gPROMS. The model allows for the consideration of various flow patterns of feed gas and permeate, as well as the evaluation of applied vacuum and sweep gas for generation of a suitable driving force for mass transfer and produces two streams, a purified gas (retentate) and a permeate stream (CO2-rich gas). The membrane model performance has been assessed using spiked gas conditions, taking into account a fed gas mixture consisting of CO2, O2 and N2 (with or without the presence of water vapour) whose composition and flow rate reflects those of the real reference flue gas. For a single membrane module with PIL (SOLVIONIC F9:1) active layer a single stage as well as multi-stage configurations were investigated, taking into account the separation of a 3-components CO2/O2/N2 mixture as dry gas and in the presence of water vapour. While for the first case permeability coefficients were determined from pure and dry gas experiments they were based on permeation tests using spiked conditions (15 % CO2 in N2, 90 % humidity) for the other cases. A permeate pressure of 0.25 bar was assumed for all these cases, while membrane area and feed gas pressure were varied in form of a sensitivity study.

Figure 13: CO2 enrichment and recovery vs. feed gas pressure for the membrane area of 3.25x106 m2. Solid symbols denote concentration and empty symbols denote recovery.
As a result, maps of CO2 purity, recovery and feed pressure were determined based on the simulation results for a given membrane area. An exemplary result of these calculations is given in Figure 13. From the results it can be concluded that the presence of H2O in the feed mixture reduces the CO2 concentration in the permeate, whereas it has a strong beneficial effect on CO2 recovery.

Figure 14: Simulation cases with CO2 recovery of 90%. Solid symbols denote concentration and empty symbols denote recovery.

Based on the relationship between CO2 purity, recovery, membrane area and feed pressure, determined in simulations, cases with CO2 recovery of 90 % were selected (see Figure 14). It can be concluded that the considered membrane can provide 90 % CO2 recovery in a permeate stream containing 36-60 vol% of CO2 (dry gas) and 34-61 vol% of CO2 (wet gas).

Performance under realistic flue gas conditions
In order to validate the obtained results for membrane performance obtained for idealized conditions and to verify the stability of the developed membrane materials further experimental investigations were performed under real process conditions. The most promising PIL (F9:1) developed by Solvionic was used as the selective material for membranes that were successfully prepared using ultrasonic spray coating, forming continuous films in the range of 5– 20 µm thickness on commercial polysulfone or lab-made polyimide support. Other membranes based on the materials from the same PIL family were further investigated as well and compared with PIL (F9:1). The influence of the process operation conditions was compared for vacuum and sweep gas driven mode in the experiments with dry gas. The results revealed that permeance was at least one order of magnitude higher in the vacuum driven process, while the selectivities were higher in the sweep gas driven mode. It was further observed that humidity positively affects the permeance of the membranes containing ionic liquids and salt additive. However, only small improvement in the CO2 selectivity could be achieved. The durability of the F9:1 membrane material upon exposure to a synthetic flue gas containing 15% CO2 and contaminating components, such as O2 and SO2 in humidified conditions, was investigated over 1500 h with 5% O2, including 500 h with 300 ppm SO2. There was no negative impact from these components on the membrane performance over this extended timeframe. Subsequently, long-term temperature stability was investigated up to an operating temperature of 75 °C. An increase of CO2 flux and decrease of CO2/N2 selectivity with increasing temperature was observed. The results from dry gas measurements, realistic gas measurements as well as durability tests, enabled to define a robust dataset for the performance characteristics including process and equipment parameters, as well as operating windows for the F9:1-based membranes.

WP4: Absorption in columns
Building up on the results obtained in WP2 the identified enzyme-solvent systems as well as developed immobilization strategies were further applied to packed columns for the investigation of their performance in reactive absorption of CO2 in WP4. While enzyme stability and rate improvements due to the presence of the biocatalyst are prerequisites for a feasible application it is important to evaluate the optimal choice of enzyme-solvent system and operating conditions in order to exploit the potential benefits and outperform currently applied solvent systems.
Experimental investigations of absorption with enzyme accelerated solvents
An illustration of the significant catalytic effects for these solvent systems was already presented in the context of WP2 in Figure 8. In general it can be concluded from the different solvent systems, which differ in the underlying reaction mechanism with CO2, that all solvents that form bicarbonate (all but MEA) showed a positive effect of CA addition on mass transfer. The catalytic effect of the enzyme, which is defined as ratio of the absorption rates between blank and enzyme accelerated solvents, showed a declining trend at higher temperatures. Experiments at different solvent concentrations revealed that higher MDEA and K2CO3 concentrations gave lower relative mass transfer rates benefits for the enzyme-solvent systems compared to solvent without enzyme. The optimum process conditions for high mass transfer in enzyme enhanced aqueous MDEA solutions were consequently identified to be low absorption temperature and low solvent concentration.
The most promising enzyme-solvent systems, which are based on aqueous solutions of K2CO3 and MDEA, were further investigated in technical scale absorption columns. At TUDO an absorption column with DN50 and a packing height of 2.3 m was equipped with Sulzer BX structured packing for the investigation of enzyme-solvent systems with dissolved enzyme. Comparable experiments showing the same trends and validating the obtained results were conducted in the technical scale absorber at DTU, which had a packing height of 8.2 m (filled with Sulzer Mellapak 250 Y) and a diameter of 0.1 m. Both experimental setups are presented in Figure 15.

Figure 15: Experimental setups of DTU (left) and TUDO (right)

Operating the column at a gas load of 0.93 Pa0.5 and three different liquid loads between 8 and 24 m3m-2h-1, the performance of the enzyme-solvent systems was first tested at common operating temperatures of 40 °C, referring to the temperature of the entering solvent.

Figure 16: Absorbed mole flow of CO2 in technical scale absorber column for most promising solvents (left) and influence of MDEA concentration on absorbed mole flow of CO2 (right) at 40 °C solvent inlet temperature.

Each aqueous solvent system was investigated without and with addition of 0.2 wt.-% of CA for a synthetic flue gas with a CO2 volume fraction of 15 vol.-%. From analysis of the CO2 content in the gas inlet and outlet with NDIR sensors the absorbed mole flow of CO2 was determined and compared as illustrated in Figure 16 (left). The results indicate that the previously determined catalytic effects can be obtained at higher liquid loads and that as indicated by the lab-scale experiments the MDEA-based solvent system outperforms the K2CO3-based solvent system in terms of absorbed mole flow rates. Furthermore, operational issues with K2CO3 occurred during the experiments because the enzyme was precipitating in the K2CO3 solution. Consequently, MDEA was selected as most promising solvent for more detailed studies in technical scale. In these studies the solvent concentration was increased to 50 wt.-% in order to investigate the effect of varying the composition of the solvent system. The results, which are illustrated in 16 (right), indicate that the increased amine concentration causes a reduction in absorbed mole flow and therefore 30wt% MDEA solutions or even lower chemical concentrations are to be preferred. This result is in good agreement with the previous observation of the importance of the bicarbonate formation for the rate enhancement by addition of the enzyme.

Figure 17: Influence of solvent inlet temperature of the absorber column on absorbed mole flow of CO2

Furthermore, the liquid inlet temperature was reduced from 40 °C to 20 °C, which for chemical reactive absorption is usually resulting in a reduction in reaction rates because of an Arrhenius-like dependence of the reaction rate constants from temperature. The results of this investigation, which are illustrated in Figure 17, indicate however that the decreased liquid inlet temperature has no significant influence on the absorbed mole flow for the enzyme-solvent system. This unique feature offers a significant potential for enzyme accelerated solvents, since at lower temperatures higher CO2 loadings can be achieved, which enable operation at higher cyclic loads and therefore reduced solvent flow rates. Accordingly, an aqueous MDEA solution with MDEA concentration of 30 wt.-% that provides good CO2 loading capacity and a liquid inlet temperature as low as technically feasible (most likely 25 °C) is recommended as promising for further evaluation.
For this the performance of the enzyme enhanced solvent system was further benchmarked against the proven state-of-the art solvent system, taking into account a 30 wt.-% MEA solution and comparing the performance under similar conditions as the enzyme enhanced MDEA solvent. To highlight the comparability of the kinetic performance and the improvement by the addition of the enzyme the absorption rates for different solvents were evaluated in the 10m column at DTU. The results, which are indicated in Figure 18 (left), indicate the significantly increased absorption rates of the 30wt% MDEA solution by addition of 0.9 g/L CA and 3.5 g/L CA, as well as the improvement of mass transfer with increased liquid load. Furthermore, a direct comparison with the industrial standard 30 wt.-% MEA solution in 18 (right), highlights that without the enzyme the aqueous MDEA solution reaches only 20 % of the absorbed mole flow measured for a 30 wt.-% MEA solution. Addition of the enzyme improves the performance to 40-58 % and 68-76 % of the absorbed mole flow measured for a 30 wt.-% MEA solution for an enzyme concentration of 0.9 g/L CA and 3.5 g/L respectively.

Figure 18: Experimental results of the influence of liquid load on the total absorbed mol flow of CO2 at 10m column height and 28 °C inlet temperature (left) and relation to the 30wt% MEA solution (right)

Immobilized enzymes
In addition to the application of dissolved enzyme, different immobilization strategies for enzyme application were considered in WP2&4. The immobilization of enzyme beads in the bags of the catalytic packing Sulzer Katapak SP, which was considered as most suitable option in WP2, was successfully implemented in the technical scale absorber at TUDO. However, the enhancement of the absorbed mole flow of CO2 was not as significant as for the dissolved enzyme, which was mainly explained by the significantly lower concentration of enzyme that was approximately a factor of 50 smaller compared to the 0.2 wt.-% CA applied in the solution. Taking into account that the technical feasibility of this concept was proven, further improvements in the production of the enzyme beads would be likely to increase the enzyme accessibility significantly and this aspect is recommended for future development.

Figure 19: Influence of BDS on absorbed mole flow of CO2

Finally a new and highly innovative immobilization concept that was not considered in WP2 was investigated, the so-called biocatalyst delivery system (BDS), initially developed by Akermin Inc. .
In this concept the enzyme is encapsulated by polymeric microparticles, which can be fully mixed with the solvent when properly agitated, while the lower density and hydrophobicity of the material allows them to float on the solvent as soon as agitation is stopped. Therefore, the BDS can be applied as homogeneous solvent-BDS-mixture when sufficiently mixed and afterwards easily separated in a simple decanter, protecting the enzyme from the harsh conditions in the stripper. In contrast to the enzyme beads an effective CA concentration of 0.2 wt.-% was well feasible with the BDS-solvent-mixture. The results from investigations at technical scale at TUDO and a direct comparison with the dissolved enzymes-solvent system are presented in Figure 19. Especially for higher liquid loads the BDS system reaches a comparable absorption performance as the dissolved enzyme system. It is envisioned that the BDS based system could be implemented together with conventional absorber-desorber configurations with minimal retrofitting and allowing continued use of thermal regeneration without damaging the enzyme-based biocatalyst, which could accelerate commercial deployment. Thus, future research should focus on this way of enzyme application for CO2 capture processes.

Modelling of absorption with enzyme accelerated solvents
The most attractive strategy for enzyme application was to be evaluated based on the economic and environmental investigations in WP6&7. In order to allow for process design and evaluation, detailed rate-based models of the absorption and desorption columns have been developed on the basis of the experimental results at lab-scale and technical scale. Specifically for the desorption column dedicated experiments in a closed cycle plant at the Australian partner CSIRO were performed and evaluated. Finally, two alternative modelling approaches were developed by DTU and TUDO. While a simplified rate-based model was developed and implemented in the simulation environment Aspen Plus® by TUDO, which allowed a direct application into the process design workflow in WP6, a more detailed kinetic model incorporating enzyme inhibition by bicarbonate formation was established by DTU, allowing for validation of the results obtained by the simplified model. Both models were validated and showed sufficient accuracy, which is exemplarily depicted for the DTU model in a parity plot in Figure 20.

Figure 20: Parity plot between measured absorbed flux and simulation with DTUs CAPCO2 model with incorporated enzyme kinetics for pilot experiments performed by DTU.

Scale up of absorption with enzyme accelerated solvents
Considering industrial application of CO2 absorption with enzyme accelerated solvents, scalability is a crucial issue. To address the question of scalability and the reliability of the data determined at technical scale additional scale-up studies were successfully conducted from technical scale (diameter 110 mm) to demonstration scale (diameter 422 mm) resulting in equivalent results, experienced at comparable experimental process conditions. The results are exemplarily illustrated in Figure 21, whereas due to the differing dimensions of the used absorbers the measured absorbed mole flow of CO2 was normalized to the packing volume. Only for the lower F-Factor and low liquid loads a significant difference between the results obtained at the different scales was observed, such that scale up of applying enzyme accelerated solvents for CO2 capture based on the technical scale experiments and the developed models is justified. The validation of the transferability of the performance results between technical and demonstration scale confirms that smaller scale testing in the university technical scale environment can be reliably used to predict larger demonstration scale performance, allowing optimizations at technical scale to be considered immediately relevant for larger demonstration scale tests, giving significant advantage in reducing time and resources for developing demonstration scale applications.

Figure 21: Comparison of absorbed mole flow of CO2 between technical and demonstration scale.

WP5: Absorption in membrane contactors
In WP5 both newly developed PIL materials and enzyme-solvent systems were integrated into membrane contactors as innovative gas-liquid contacting equipment. The major benefits of the membrane contactor over columns is the provision of a well-defined and large (interfacial) surface area per volume in combination with a dispersion-free contact of the two phases (flue gas and absorption solution). The separation of the two phases by means of a porous membrane reduces solvent losses and allows for an additional flexibility concerning the operating pressure, flow rate and temperature of both, the incoming flue gas and the solvent solution. Of course there are also additional challenges that need to be resolved when applying membrane contactors, as material stability and surface properties need to be carefully selected and operating conditions need to be restricted in order to avoid breakthrough of one of the involved phases into the other.
Material systems selected and investigated
To maximize the merits of the membrane contactor, a proper selection of the combination of membrane material and liquid phase was made with regard to stability (resistance to both flue gas components and to solvent), mass transfer resistance, commercial availability, cost and environmental issues. The temperature stability limit of the enzyme of ~60°C, defined the upper temperature limit and allowed for the use of a variety of membrane materials. Hydrophobic membranes were selected to ensure a suitable gas-liquid contact interface at the permeate side of the membrane and the chemical resistance to the solvent systems determined in WP2 was a further selection criteria. Based on these criteria a variety of membranes was screened and experiments were subsequently performed with commercially available hollow fibre polypropylene membrane modules and flat sheet composite fluoro-polymer membranes on polypropylene backing, providing higher chemical stability, which is known to be an issue for plain polypropylene membranes in long-term applications. Initial experiments at NOVOZYMES proved the polypropylene hollow fibre membranes to be compatible with the enzyme. Besides the most promising enzyme enhanced MDEA solvent system, the MDEA solution was also investigated as a reference case.
Furthermore, the coating of a selective PIL layer was investigated for the flat sheet membranes in order to provide a proof-of-principle demonstration of a hybrid membrane contactor, the principle of which is depicted in Figure 22. The most promising blend of PILs for this application is PolyDDATFSI/PYR14TFSI/Zn(II)TFSI.

Figure 22: Hybrid membrane contactor system with added PIL selective layer (green) on the porous membrane (yellow) combined with enzyme activated absorption solvent.

Benchmarking of membrane contactor equipment
In order to compare the performance of flat sheet and hollow fibre membrane modules two membrane contactor rigs were designed, constructed and conditioned at TUDO and SINTEF as shown in Figure 23. Benchmarking of the membrane contactor equipment was performed at 25-35ºC using 30 wt% MEA solution without enzymes, applying hollow fibre membrane modules with active area of 1.045 m2 and flat sheet membranes of 12.6 cm2.
Standardized experimental procedures were established for the investigation of the membranes through critical pressure measurement of several candidate membranes. Due to the different nature of the membrane modules, the applied flow rates and pressure difference were determined specifically for each set-up. As expected, the more dense structure of the flat sheet composite fluoro-polymer membranes resulted in a lower area-specific molar flow of absorbed CO2 (0.35 to 0.88 mol·hr-1·m-2) as compared to that of the more open porous hollow fiber polypropylene membranes (11.0 to 22.9 mol·hr-1·m-2). However, the fluoro-polymer membrane provides greater long term durability to the amine based solvents and is less prone to biofouling by the enzyme. No comparable membrane was commercially available in the case of hollow fibers. The benchmarking test results were therefore mainly applied as baseline experiments for the further investigations on the individual test rigs.

Figure 23 Membrane contactor test rigs for flat sheet and hollow fibre membrane modules
Enzyme activated CO2 absorption with membrane contactors
Similar to the absorption in columns mass transfer was determined with and without the addition of the enzyme in MDEA solvent systems. However, the results indicated in Table 3 for both flat sheet membranes at 35ºC and hollow fiber membrane modules at 40ºC illustrate that the catalytic effect due to the addition of enzymes, while being comparable for both membrane configurations is significantly lower than experienced in columns. It seems reasonable that the hollow fiber membranes allow for higher flux of CO2due to a more open pore structure (larger pore size) than the investigated more dense flat sheet membranes. The obtained absorption rates were comparable to several reports from literature, e.g. from Hoff et al. (0.24 mol CO2/m2/hour @ 40ºC using PTFE membranes and 30%MDEA + 5%Piperazine) and Wang et al. (0,36 mol/m²/h @ 25°C using hollow fiber membrane contactor and 30% MDEA), being in the same range as the applied flat sheet membranes while the results for the hollow fiber membranes are an order of magnitude higher. It is assumed that the lower catalytic effect can be attributed to a steric hindrance limiting the enzymes approach to the gas-liquid interphase located in the pores of the membrane. The current results are further backed up by similar enhancements from introduction of enzymes in the solvent for absorption in membrane contactor reported by Cowan et al. . A closer fundamental study of enzyme-membrane interface behavior could lead to developments that improve the catalytic effect of CA in membrane-based systems.

Table 3: Absorption rates (= CO2 flux) for flat sheet and hollow fiber membrane modules from comparable experiments using 30 wt.% MDEA solution and feed gas of 15% CO2 in N2.
Membranes Solvent system Temperature CO2 flux (mol/m2/h) Catalytic effect
Flat sheet 30% MDEA 35ºC 0,11 1
Flat sheet 30% MDEA w/CA 35ºC 0,19 1.67
Hollow fiber 30% MDEA 40ºC 2,33 1
Hollow fiber 30% MDEA w/CA 40ºC 4,05 1,74

Proof-of-concept of CO2 absorption with hybrid membrane contactor
While the addition of a dense PIL layer to the membrane contactor provides additional mass transport resistance the high selectivity of the PIL as sorbent may result in improved performance especially considering potential synergies with the enzyme enhanced solvent. The concept of such a hybrid membrane contactor (cf. Figure 22) was therefore investigated for the optimal PILs developed in WP1 and investigated for gas membranes in WP3. After evaluating various configurations of the membrane contactor with the additional PIL layer a proof-of-principle for enhanced absorption rate in a hybrid membrane contactor was demonstrated using PIL material F9:1.

Figure 24: CO2 molar flows from absorption in membrane contactor using flat sheet membrane with and without selective PILs layer and 30% MDEA with and without enzyme. MEA reference added for comparison.

As shown in Figure 24, an enhancement of the absorption rate of 36% was found when adding the selective PIL layer and using 30% MDEA solution as solvent, while the combination of enzyme activated MDEA solution and a PIL layer resulted in more than twice the CO2 absorption compared to the MDEA baseline.

CO2 desorption with membrane contactor
Membrane contactor allows for two modes of operation for stripping, one is the application of sweep gas (N2 or air), the other is the application of slight vacuum on the permeate side. Desorption with sweep gas was performed for both flat sheet and hollow fiber membranes. Despite some differences in the operation conditions concerning mainly the pressure difference the measurement with CO2 loaded 30% MDEA solvent at 60ºC gave comparable results. Application of a slight vacuum (0.80-0.98 bara) was tested in the case of flat sheet membranes which resulted in considerably higher desorption rates. Key figures of the performed experiments are presented in Table 4. While experimentally more challenging, the work with a slight permeate vacuum demonstrates the potential for applying the membrane contactor technology also for CO2 desorption, however this may not be realistic with the large volumes of flue gas to be processed at a typical coal fired power plant It may be more realistic to consider the use of the membrane contactor for absorption and carry out desorption using alternative methods.

Table 4: Desorption rates obtained with membrane contactor from CO2 loaded 30% MDEA solvent at 60ºC.

Membranes Solvent system Desorption mode CO2 flux (mol/m2/h) Catalytic effect
Flat sheet 30% MDEA Sweep, Pgas - Pliq ≈ 0.15 bar 0,08 – 0,14 -
Hollow fiber 30% MDEA Sweep, Pliq - Pgas ≈ 0.15-0.40 bar 0,15 – 0,19 -
Hollow fiber 30% MDEA w/CA Sweep, Pliq - Pgas ≈ 0.15-0.40 bar 0,33 – 0,40 2,13
Flat sheet 30% MDEA Vacuum, Pliq - Pgas ≈ 0.40-0.60 bar 1,4 – 5,4 -

Modelling and simulation of transport phenomena and mechanisms
In order to allow for process design and evaluation also a model of the membrane contactor for CO2 absorption was developed. The model, which is structured in three hierarchical levels as shown in Figure 25 (left), was developed using Aspen Custom Modeler (ACM) in order to describe mass transfer of CO2 absorption/desorption in membrane contactors and allow for easy integration within the process design approach undertaken in WP6. The model considers an adjusted gas-liquid equilibrium model that accounts for the effect of liquid loading, as well as energy balances in order to give a more accurate prediction of the absorption rates and temperature profiles over the contactor length. An experimental validation of the model was performed for the ternary system MEA-H2O-CO2, where the accuracy of the model is indicated by the parity plot illustrated in Figure 25 (right), in which experimental results are compared with the predictions from the model. The model is capable of accurately predicting the absorbed molar flow of CO2

Figure 25: Structure of the membrane contactor model with the different hierarchical levels (left) and parity plot of the absorbed mol flow of CO2 in membrane contactor (right)

Effect of real flue gas contaminants
In order to evaluate the applicability of the developed membrane contactor under realistic conditions, including the use of flue gas components like SO2 and NOx, the tolerance of the individual components i.e. membrane materials and enzyme-solvent systems towards such components were tested. The flat sheet membranes applied for absorption/desorption and the hybrid membrane contactor materials have been tested as a regular gas separation membrane using more realistic conditions of 15% CO2 concentrations, 5% O2, 100% relative humidity and eventually 300 ppm of SO2. The performance of the hybrid membrane contactor material was stable over a time frame of >1500 hours. In addition the activity of the enzyme in the solvent system upon exposure to SO2 and NOx was tested. Judging from esterase and CO2 assays the enzyme carbonic anhydrase showed no degradation of activity for hydration of CO2 or hydrolysis of p-nitrophenyl acetate (pnp) during short term exposure of up to 4 hours to SO2 (20-100 ppm) and NOx (180-360 ppm). The enzyme remains active also after extended exposure tests up to 45 hours with 100 ppm SO2, although a significant drop of activity towards hydration of CO2 was observed.

WP6: Process development

The activities of process development aimed at the design of novel and competitive process concepts for post-combustion carbon capture based on the highly innovative technologies and materials developed in the INTERACT project. In the course of those activities an integrated approach building on the experimental results obtained in WP3-5 and simulation-based process analysis (WP6) was applied to generate a fundamental understanding of the technologies considered.
Exploiting these deep insights promising process concepts on INTERACT post combustion technologies were identified. To tap their full potentials process simulation was conducted facilitating analysis, debottlenecking and optimization of these concepts from an energy and/or economic perspective. Mathematical models developed for each individual INTERACT technology were applied. The most promising process concepts, which were evaluated as such in WP7 under industrial conditions, are further presented in some detail.

Absorption/desorption using enzyme activated MDEA-solvents
As a result of work of WP 2 and 4 it became obvious that the chemical absorption of carbon dioxide using solvent mixtures that store CO2 in form of bicarbonates prospers from the addition of a catalytically active enzyme. The presence of the enzyme accelerates the formation and decomposition of bicarbonates; even at reduced temperatures equilibrium loading conditions are virtually obtained during absorption and regeneration which allows for larger cyclic loadings and thereby for an economically attractive reduction of the amount of the solvent utilized.
While the presence of the enzyme facilitates the operation at more favorable, hence lower, absorption temperatures it adds new constraints to the process. Among others, the most essential constraints to be addressed by process development are related to the temperature-sensitivity of the enzyme as well as to the ability to cool down gas and liquid feeds of the absorber at industrial conditions to temperatures that cannot be reached using cooling water in conventional set-ups.
In a structured manner various process concepts were developed to tackle the aforementioned constraints while exploiting the benefits of the innovative technology. Figure 26 provides an overview on these process concepts.

Figure 26: Overview on types of enzyme activated process concepts investigated.

Initially starting from the application of the conventional process concept known from MEA-based carbon capture (MDEA_ENZ), modifications were performed to boost the performance of the capture process: The process alternative MDEA_COOL makes use of a chilled utility, generated from absorption or compression cooling, to reduce the temperature of the absorber feed streams. The potential benefits of this set-up are found from the trade-off between energy savings due to exploitation of larger cyclic CO2-solvent loads and the energy demand of refrigerant generation.
In contrast, the process MDEA_ISC prevents from the use of chilled utilities. Efficient cooling is conducted by application an external inter-stage cooler attached to the absorber. By this non-conventional arrangement heat rejection is intensified as thermal energy originating from absorption in the upper section of the column becomes removable by cooling water. Both concepts are illustrated in Figure 27.

Figure 27: Process concept of the process MDEA_COOL (a) and MDEA_ISC (b).

To protect the sensitive catalyst from thermal damage in the hot section of the process, hence inside the stripper, the immobilization or reclamation of the enzyme inside the low temperature process section was investigated. Figure 28 illustrates the process configuration MDEA_IMB which uses immobilized enzymes as catalyst in the absorber. Moreover, the configuration MDEA_UF applies ultrafiltration to restrain the free enzymes in the ”cold” section of the process.

Figure 28: Process concept of the process MDEA_IMB (left) and MDEA_UF (right).

In addition to the exploitation of conventional columns for decarbonization and solvent regeneration, the usage of more sophisticated technologies such as membrane contactors were investigated. The membrane contactor was applied to replace the conventional absorber, the desorber or both of them, respectively.

Gas separation membrane processes
The application of INTERACT-gas separation membrane technology allows for the selective removal of a target component, CO2, from a feed stream. Three alternative process concepts were developed distinguishing between processes with and without sweep gas use: A single stage- membrane process were designed, whereas driving forces for transmembrane separation are created by feed gas pressurization and vacuum conditions on the permeate side. No sweep gas was used. Since the simultaneous realization of numerous targets such as high CO2 capture rates and stringent purities of the product streams is challenging using a single stage process only, a 2-stage membrane process was considered as well. The subsequent membrane operations supplement each other, whereat the upstream membrane focuses on the decarbonization of the flue gas to a sufficient extend, while the second membrane stage purifies the pre-concentrated CO2 rich permeate stream from the first stage to a composition that allows long-term CO2-storage. The concepts are illustrated by Figure 29 (a) and (b).

a) c)


Figure 29: Process concept of the process single stage membrane process (a) the 2-stage membrane process (b) and the MTR-membrane concept.

Different from the two aforementioned membrane process alternatives, the MTR-membrane process concept applies exhaust gas recirculation. The process, using a sweep gas to support the off gas decarbonization, is illustrated by Figure 29 (c). The process utilizes an assembly of three membrane operations to generate a CO2 rich stream at CCS conditions and is applicable for the decarbonization of power plants. Its concept is rather simple: the flue gas of the power plant is compressed and conditioned. By an upstream membrane separation carbon dioxide is recovered from the flue gas along with other components, mainly nitrogen. The resulting permeate stream is fed to a CO2-compression train; at high pressure carbon dioxide is condensed, separated from the gas steam and prepared for pipeline transport.
Naturally, the removal of CO2 by condensation from the compressed gas stream is incomplete. Thus, for efficiency reasons, the remaining gas fraction is processed by a second membrane operation that aims at the recovery of compressed CO2. While the permeating gases are recycled into the compression train, the retentate stream is mixed into the upstream flue gas stream.
In contrast to the other membrane processes reported, no deep flue gas decarbonization is targeted by the operation of the upstream membrane (“Decarbonization membrane”). Accordingly, a significant amount of CO2 is found in its retentate stream. To comply with the envisioned capture rate this product stream is directed to a successive membrane unit. Here, the remainder of CO2 is majorly removed from the flue gas; carbon capture at an aimed rate is realized. The operation is supported by the presence of a sweep gas. Using the combustion air of the boiler as diluting agent provides numerous benefits since neither extra sweep gas generation nor downstream separation of the CO2 from the diluting gas stream is required. Via the off gas of the boiler the collected carbon dioxide is recycled to the capture facility and becomes available for removal.

Adsorption using the solid, PIL-coated beads
In addition to the absorption/desorption- and gas separation membrane processes a further process concept based on adsorption was developed. This process exploits the PIL-based sorbents as solid inventory that were developed in WP1. The new material exhibits a reasonable adsorption capacity of carbon dioxide and shows a good CO2/N2-selectivity. To exploit this technology for post combustion carbon capture a promising pressure swing adsorption (PSA) process was designed. The process concept that is illustrated by Figure 30 comprises of two adsorption stages. The first stage decarbonizes the flue gas. The decarbonized gas is directed to the stack for release. By pressure reduction a pre-concentrated stream of CO2 is generated. After the pressurization of this stream, gas treatment is conducted by a second PSA stage, where CO2 is selectively adsorbed. Since the removal of carbon dioxide is incomplete the breakthrough gases are recycled upstream. By the regeneration of the loaded sorbent of the second stage a CO2 rich stream at desired concentration is produced by pressure reduction. This stream is compressed to pipeline pressure.
To facilitate a continuous operation and to improve the energy efficiency of the process each PSA stage operates multiple columns in parallel. Benefiting from a sophisticated schedule of pressurizing and depressurizing operations the dynamic behavior is smoothed out and the process operates close to steady-state conditions.

Figure 30: Process concept of the PSA capture process.

WP7: Assessment of new technologies in industrial application scenarios

To allow for a significant quantification of the performances of the novel post combustion carbon capture technologies developed under the umbrella of the INTERACT project a system-integration into an industrially relevant CO2 emission scenario was performed. Mass and energy balancing were carried out and the each integrated process concept was optimized. The framework of the investigation has been derived from the guidelines of the European Benchmark Task Force (EBTF) and was applied likewise to all process alternatives under consideration.
The resulting energy and mass balances formed the bases of the techno-economic as well as environmental analysis of the process performances. For a fair comparison, the outcome is benchmarked against the Base Case, which is a capture process that uses state-of-the-art technology.
Following the recommendations of the EBTF core document “European best practice guidelines for assessment of CO2 capture technologies” an Advanced Supercritical Pulverized Coal-fired power plant acted as such emission scenario. The MEA-based absorption/desorption process served as the Base Case to benchmark the performances of the innovative technologies.

Whole process simulation
Under the umbrella of the aforementioned techno-economic framework all innovative as well as standard post combustion capture technologies were simulated using the commercially available simulation environment Aspen One V8.8. For each technology system-integration into the power plant scenario took place; mass and energy integration was performed. All process alternatives were investigated at a similar level of detail and were optimized with respect to technical feasibility, robustness, energy efficiency and costs.

Techno-economical assessment
Based on the energy and mass balances obtained from the simulation activities the techno-economic performances of the various technologies were quantitatively assessed. Table 5 summarizes the outcome of this assessment with regard to the most promising representatives of each INTERACT technology under investigation. Along with that the corresponding calculation results of the Base Case capture process are presented.
Most encouraging results are found: for each INTERACT technology option at least a single process concept was identified that outperforms today’s standard technology (Base Case) from an efficiency perspective. But even more so, taking into account the possible reduction of the energy penalty by ~3.5% points and the reduction of CO2 avoidance cost by ~15% the INTERACT technologies present a breakthrough in post combustion carbon capture in terms of net power output, energy penalty and costs of electricity production.

Table 5: Overview on the techno-economic results obtained on the best performing representatives of innovative INTERACT CO2 capture techniques. In addition information on the standard technology (Base Case) is provided.

While the results of the technical evaluation are apparent and easy to interpret, the findings obtained by the economic assessment require a more detailed breakdown due to its diversity; in particular with regard to the capital investment, the costs of electricity production and CO2 avoidance costs. The solvent based processes exploiting conventional columns to conduct CO2 capture and recovery require capital investments in the range of the Base Case. Significant reductions in electricity generation costs and CO2 avoidance costs are estimated. Savings up to 4 €/MWhe (costs of electricity production) or 6 €/tonCO2 (avoidance costs) seem realistic. In contrast to the solvent based processes, the assessments of the membrane-based and adsorption-based processes indicate elevated investment costs. This increase is inflicted by the predicted size of membrane area and the estimated amount of solid inventory as both are consumed in large quantities and acquired at comparably high specific purchase prices. In addition, concerning these technologies, periodically follow-up investments are required due to the limited lifetime of the core materials. As a consequence, the costs of electricity generation and CO2 avoidance are found in the range of the Base Case.
Among all technologies investigated, membrane contactors in combination with enzyme-activated MDEA-solvents stand out since electricity is generated at a significantly higher price. In this case CO2 avoidance is more costly than in MEA-based technology. This difference results from the large share of investment costs allocated to purchase tremendous amounts of the required membrane area. For the time being it is theorized that the astonishingly high demand on membrane area is attributed to an adverse interaction between the enzyme that intensifies the absorption of CO2 and the selected membrane material. In that sense it needs to be understood that the findings represent the performance of a unique combination of materials and technologies and therefore do not allow drawing general conclusions on the performances of membrane contactors in the field of carbon capture.

Life cycle assessment
Based on the characterization of the benchmark process and the techno-economic optimization of the different process concepts a life cycle assessment (LCA) was performed to evaluate the environmental impact of the different processes. A summary of main steps of the LCA application to INTERACT CO2 capture alternative processes is given in Table 6.

Table 6: Summary of main steps, selections and assumptions for LCA application to INTERACT CO2 capture alternative processes

LCA step Description
Full scenario definition and description A reference case was selected and characterized, an Advanced Super Critical (ASC) pulverized fuel bituminous power with a generation capacity of 819,8 MWe (gross), a net output of 757,1MWe and a net cycle efficiency of 45,2%.
Functional unit definition Definition of the functional unit as 1 MWhe of electricity delivered by the plant to the grid.
Scope definition The LCA is performed following a Cradle-to-Gate approach; the system boundaries of the LCA include the extraction and processing of the coal and other raw materials, the production of electricity, the capture of carbon dioxide and the treatment of the present wastes.
Data sources definition The data used to perform the LCA analysis has been taken from the documentation regarding the Common Framework delivered by EBTF and the simulation reports of the different process delivered by SUPREN under the scope of the INTERACT project.
Key assumptions All the assumptions regarding the consideration of workforce, capital equipment, relevant flows, storage or use for captured CO2 etc., are primary based on the recommendations given by the EBTF in the Common Framework Definition Document
LCA software selection The calculation regarding the LCA tasks within the INTERACT project were conducted using the software SimaPro developed by Pré Consultants, the leading LCA software chosen by industry, research institutes, and consultants world-wide
Lifecycle inventory In the course of data collection the database ECOINVENT 3.2 was extensively used for characterizing impacts and footprints of substances present in inlet and outlet streams of the processes.
Lifecycle impact assessment Midpoint and endpoint impact categories were selected for impact assessment, allowing the interpretation of the final results of the selected alternatives.

The results of the LCA were evaluated based on the calculated midpoint and endpoint indicators, which are illustrated for the Base Case and the most promising process concepts in Table 7 and 8. From the interpretation of the estimated midpoint indicators it becomes clear that carbon capture prospers from the application of the novel INTERACT process technologies; in particular with respect to environmental impact categories such as “climate change”, “freshwater eutrophication”, “human toxicity”, “photochem. oxidant formation” (with the exception of PSA + PIL process) and “fossil depletion”.

Table 7: Midpoint categories results for environmentally best performing INTERACT CO2 capture process alternatives (90% C02 capture rate)

To ease the understanding of the outcome of the LCA endpoint indicators are used which are simpler to interpret. These endpoint indicators represent a sum of the mid-point categories weighted according to their significance. The resulting end-point indicators imply that for each INTERACT CO2 capture technology essential improvements regarding the environmental impact are identified, outperforming the benchmark process. The main midpoint contributors to the endpoints are fossil depletion, climate change, human toxicity and particle matter formation. To provide a distinct quantitative example, for the process concept MDEA + free enzymes more than 98 % of the total end-points can be attributed to the mid-point categories fossil depletion (≈57%), climate change (≈22%), human toxicity (≈11%) and particle matter formation (≈8%). This leaves a remaining 2% contribution for the rest of the mid-point categories. Therefore, the large relative increases in marine eutrophication and water depletion, which is listed for some of the process concepts in Table 7, has only modest influence on the environmental impact characterized by the endpoint categories.

Table 8: Endpoint categories results for environmentally best performing INTERACT CO2 capture process alternatives (90% C02 capture rate)

For a better visualization Figure 31 presents the relative environmental improvements compared to the Base Case. Obviously each of the process concepts presents an improvement concerning the environmental impact, whereas the estimated improvements are maximized for the enzyme activated reactive absorption processes with dissolved enzyme.

Figure 31: Impact reduction for environmentally best performing INTERACT CO2 capture process alternatives over benchmark process

Conclusion: Concise assessment of the process concepts from techno-economic and environmental perspective
On the basis of the aforementioned findings, it can be concluded that among the investigated technologies the absorption/desorption process exploiting enzyme activated MDEA-solvents provide the biggest potential for competitive and environmental friendly carbon capture; overcoming the considered standard capture technology, i.e. MEA-based CO2-absorption, by far A significant reduction on the energy penalty of the CO2 capture process (≈-3,5%) has been presented, associated with a decrease in the electricity production costs (≈-4,5%) as well as the costs of CO2 avoidance (≈-13,5%) and a low environmental profile (≈-7,5%, endpoint indicators). Taking into account that the process concept can be implemented with established gas-liquid contacting equipment, for which successful scale-up studies were performed within WP4, provides further prospect towards successful implementation from a technical feasibility and robustness point of view.
Gas membrane separation, enhanced by a PIL coating, has been found to be less competitive in economic terms (~+7% electricity production costs, ~+18% costs of CO2 avoidance). However its promising environmental impact (~-5%) and the opportunity for future improvement of this technology becoming more mature show a clear potential for a successful application. In particular, applications aiming at reduced CO2 capture rates, as in such as CO2/CH4 for biogas upgrading or natural gas refinery, seem promising.
Despite the potential reduction of the energy penalty and potential improvements in terms of environmental potential, the developed pressure swing adsorption process based on PIL-impregnated beads, and the membrane contactor based enzyme-activated MDEA reactive absorption/desorption process require further development to compensate economic disadvantages when compared to the base case and commercially available alternative materials. Further development might however foster further exploitation of the potential and result in additional improvements, taking into account further optimized PILs and improvements in terms of the concept of hybrid membrane contactors, which was developed as a proof-of-concept in the scope of the INTERACT project and provides further potential for optimization.

Potential Impact:
Improved materials and production techniques for a targeted functionalization of tailored gas separation membranes
In the INTERACT project various synthetic strategies have been pursued for the development of a range of dense ILs to be used as active material in gas separation, membranes. The materials have been screened for CO2/N2(main flue gas components) separation, at lab scale but also for other aspects, such as the scalability and economic feasibility, that have major impact for the industrial production. Poly(diallyldimethylammonium) bis(trifluoromethanesulfonyl)imide (polyDDATFSI), one specific PIL material, combined with ionic liquid and salt additive excelled in the preliminary tests at lab scale, where special attention was paid to the synthetic scalability, defect-free membrane formation capability, mixed gas selectivity (αCO2/N2 = 34), and permeability of CO2 (P=217-464 Barrer).

Figure 32: successful commercialisation of polyDDATFSI

It was observed that the gas separation properties of PIL-based membrane materials can be tuned by the incorporation of free ionic liquid and metal additives. Based on the positive results of the scale-up studies, successful commercialization of this selected PIL, polyDDATFSI, has started. The product is now in SOLVIONIC portfolio and is available up to the kg scale. As shown on Figure 32, the WP1 upscaling work was linked to the volume of sales of polyDDATFSI over the last few years.
Interestingly the developed PIL material has raised most interest so far in the frame of electrochemical energy storage systems (metal-ion battery applications, e.g. Li-ion batteries) as a component of gel/polymer based electrolytes rather than CO2 capture processes, for which the material was initially developed in the project.
A range of this type of composition was marketed to the attention of the new generation of battery producers. Polymer electrolytes (PE) are membranes composed of a dissolution of salts in a polymer matrix with high molecular weight, whereas gel type electrolytes contain ionic liquid to enhance the ionic conductivity. Ionic liquids (ILs) and PILs represent an important class of materials as they can exceed the stability and performance of electrochemical devices made with conventional, organic solvent-based electrolytes . The main advantages that these materials offer are excellent physicochemical properties, i.e. thermal and electrochemical stability, high ionic conductivity, non-volatility and a high viscosity to suppress leakage. Consequently the developed materials provide additional prospects for other applications, which in this context can further foster the development of a holistic solution to a sustainable power production including gas treatment for fossil fired power plants and energy storage for renewable energy production.
An upscaling work has also been carried out on the salt additive, Zn(II)TFSI, which has led to increased production capacity, linked with substantial sales increase. In any case, the results obtained in the project provide a framework for the synthesis of PIL/IL/salt additive formulations, following the same strategy but using other PIL materials developed in the project (e.g. PILs prepared starting from commercially available polymer materials, such as cellulose acetate and poly(vinylbenzylchloride)), that can be applied for other applications. Additionally, the testing of PIL-based membrane technology for extended periods under “real-world” conditions has given valuable information on the real potential of this technology, similar to the other leading membrane technologies that are being evaluated. The INTERACT project has contributed in this sense, since literature data is scarce, and we have provided preliminary testing data. The membrane based on this PIL was selected for durability tests under realistic flue gas conditions (synthetic flue gas containing 15% CO2 as well as minor contaminating components, such as O2 and SO2 in humidified conditions). As there was no negative impact observed on the membrane performance over this extended timeframe, the long-term temperature stability was investigated up to an operating temperature of 75 °C. An increase of CO2 flux and decrease of CO2/N2 selectivity with increasing temperature was observed. Although the INTERACT project was focused on flue gas separation from coal fired power plants, the developed PIL-based membranes also proved their potential for other light gas separation, such as CO2/CH4 for biogas upgrading or natural gas refinery. PIL research has a very strong application focus driven by their functions which are not limited to separation applications. Some other emerging applications that have attracted the interest of the scientific community are: energy harvesting and storage (e.g. electrolyte materials for batteries) and catalysis (e.g. direct synthesis of cyclic carbonates via the coupling of CO2 with epoxides).
Innovative concepts for enzyme application in gas-liquid contacting systems
In the scope of the INTERACT project a full investigation of the suitable operating ranges in terms of pH, temperature and solvent concentrations have been performed for the investigated enzyme carbonic anhydrase, including dedicated long-term stability studies for various solvents. While the focus was placed on the application of an enzyme-solvent system for CO2 separation in the context of a coal-fired power plant the results can further be used to evaluate the application of the enzyme-solvent systems in other gas cleaning applications, based on the characterized operating ranges for each solvent system.
The characterization of the identified enzyme-solvent systems as well as several means for enzyme immobilization further demonstrated the potential improvement of the mass transfer rates towards an effective and energy-efficient CO2 capture process, by means of enzymatic reactive absorption. In the course of the extensive experimental studies an improvement of the measured CO2 transfer by addition of the enzyme by a factor of up to 9 was determined for application in aqueous MDEA which is a well-known solvent in the gas scrubbing industry. Therefore, adding enzyme directly overcomes the poor reaction kinetics of commercially important MDEA solvent which allows for harvesting the beneficial properties of this kinetically limited tertiary amine, including lower heat of reaction and evaporation as well as a reduced corrosiveness when compared to primary amines such as MEA . The investigations furthermore revealed that the enzyme is capable of efficiently catalysing the reactive absorption even at reduced temperature, extending the operating window towards higher gas solubility and therefore higher cyclic loadings, resulting in further potential for a reduction in solvent requirements. The transferability of the results obtained at lab scale and demonstration scale was further demonstrated by experimental investigations at pilot scale. The developed insight on enzyme accelerated solvent systems can further be transferred to innovative solvents offering even better thermodynamic properties than MDEA, taking into account blends of different molecules and computer aided molecular design tools.
The investigated means for enzyme immobilization, taking into account surface coating of magnetic nanoparticles or catalytic packings, the entrapment of enzyme beads inside the mesh-bags of catalytic packings and the innovative biocatalyst delivery system (BDS) provide a portfolio of different means for the implementation of enzymes in fluid separation processes in general and the results can be transferred to other applications, such as enzymatic reactive distillation for the production of bulk as well as fine chemicals and even chiral compounds . Immobilization of the enzyme provides the possibility to control the conditions that are experienced by the enzyme and therefore adhere to the operating ranges for a stable enzyme application and extend the application ranges to further systems. In this respect, BDS particles worked very well, nearly matching dissolved enzyme performance. Other immobilization approaches resulted in a potential performance decrease that might be caused by a mere reduction of catalyst mass per volume, or could result from additional mass transfer limitation introduced by the solid form of immobilized enzymes, i.e. gas-liquid-solid interactions.
The results obtained for the application of enzyme activated solvents in membrane contactors additionally provide valuable insights that can be exploited for further evaluation of the technology in various CO2 capture processes. Within the INTERACT project an improvement of a ~36% increase in absorption rate was demonstrated by addition of PILs to the membrane surface exposed to the gas side in the contactor. Such a hybrid membrane contactor concept provides additional prospect and triggers interest for further development of the idea using other routes in order to generate low cost, high flux membrane surfaces with high CO2 affinity in order to increase the mass transport. Such membrane improvements would reduce the required amount of membrane surface area, making the technology cost-competitive and allowing for further investigations in forthcoming demonstration to pilot scale projects. The hybrid membrane contactor design concept is interesting not only for CO2 capture but for other separations such as sulfur dioxide from various gas streams, methane / CO2 and paraffin's/olefins in which the delivery of the desired product in a solvent stream is beneficial. It also opens up a further application for tailored PILs. In addition, other materials, which provide facilitated transport, could be considered for use in the hybrid membrane contactor.
Innovative energy and resource efficient process concepts for post combustion carbon capture
The strong decrease or elimination of CO2 emissions from fossil fuelled power plants is a major target in the present and future environmental and political landscape. The necessity to reduce greenhouse gas emissions such as CO2 by 80 to 95% before 2050 compared to 1990 has been proposed by the EU and reconfirmed by the European Council in February 2011. In order to succeed in this ambitious objective novel production processes and routes are required that allow the generation of targeted products and services with significantly reduced environmental footprints. Eventually being even more important, existing technologies need to be upgraded with technologies that limit the release of climate relevant emissions. With regard to post combustion capture the INTERACT project represents a breakthrough in the development of novel materials and techniques for the decarbonization of industrial CO2 rich flue gases. Applying the materials and concepts developed on the application towards the flue gas stream of fossil fuelled power plants - which represent one of the main emitters of carbon dioxide - post combustion capture is performed in an energy- and cost-efficient way. Compared to today’s standard MEA-based capture technology a reduction of the expected energy penalty of the power plant of 29 % is predicated at substantially lower energy production costs (- 4 €/MWe) and CO2-avoidance costs (- 6 €/tonCO2) as well as being more sustainable form a LCA perspective. In that sense the outcome of the INTERACT project can be considered a milestone concerning the European and global efforts of reducing carbon emissions up to 2050. Effective new materials embedded in novel process schemes have been designed. Exploiting these in industrial scenarios, economically attractive solutions for decarbonization of CO2 rich flue gas streams have been identified that are technically feasible and robust to move to the next scale of demonstration. These solutions show the potential to pioneer efficient ways of environmentally friendly carbon dioxide sequestration and therefore can provide a crucial contribution to the future CCS efforts required to maintain a high quality of life.
The developed models, process concepts and methods provide the necessary foundation to efficiently evaluate the application of the technologies to other gas cleaning processes, as e.g. in NGCC, IGCC, natural gas or biogas upgrading, or cement production. The transfer of the developed methodologically approach to these additional applications requires merely an adaption of the problem specific parameters, accounting for feed and product specifications, cost and LCA data. Especially for the treatment of biogas and other scenarios with lower CO2 capture rates process concepts such as the PIL-based gas membrane concepts, which were outperformed by the enzymatically catalysed reactive absorption process, provide a huge prospect. The developed methods present the necessary tools required to evaluate and optimize the various process concepts for these and other alternative applications. A consecutive use developed conceptual design and process evaluation methods and further development of the tools is already pursued in the context of a subsequent project named "NANOMEMC2-Projekt" ("NanoMaterials Enhanced Membranes for Carbon Capture", ( which is implemented in the context of Horizon 2020.
Impact on global market
The aim of the project was the development of new technologies for the CO2 capture in various industries, for example in power plants. Although market share of coal, oil and gas fired power plants in Europe is decreasing, the number of fossil fuel powered power plants is increasing on the world scale. In China alone there are about 693 coal fired power plants active , while the US hold for another 380 coal fired power plants at the present time. Fossil fuel power plants are responsible for 45% of the worldwide emissions of greenhouse gases thus saving potential in this field is huge. Based on state-of-the art technology of packing vendors one 500 MW power plants requires absorption columns for CO2 capture with a packing volume of ca. 10.000 m3, the unit price for packings used at the present time in ca. 200 €m-3, thus investment only in the packing is 2 million euro. With the use of novel enzymatic enhancers pressure drop can be reduced and capacity and mass transfer will be increased – this leads to smaller dimensions of the columns. The enhancement of mass transfer of 20% leads to savings only in packing investment of 400.000 €, not accounting for additional savings in smaller apparatus and construction. This leads to significant savings for the operators of the plant.
Energy requirements significantly benefit the use of bioreactive absorption over other technologies. According to Life Cycle studies, bioreactive absorption uses up to 250 kWh per ton recovered CO2 compared to significantly higher values for pressure swing adsorption (1600-180 kWh), cryogenic distillation (600-800kWh) or amine absorption (330-340 kWh). Through the replacement of chemical absorption through enzyme enhanced CO2 capture in which carbonic anhydrase (the enzyme responsible for CO2 balance in the human body) is used we expect substantial reduction of inventory of toxic species in CO2 capture processes
Technological impacts and risks
Despite the carved out large potential of the developed process concepts, especially the developed concept for an enzymatic-catalyzed reactive absorption process, further investigations are necessary to address remaining challenges for large-scale applications. Most of all further scale up demonstration of the technologies under commercially relevant conditions is required to address the transferability of the results in an industrial environment and to prove long-term stability and validate the cost-effectiveness of the developed process concepts. The research in the current project was devoted to the development of the new technologies. While the consortium tried to bridge the gap between material development, technology development and characterization and process design and evaluation, the core of the project was focused on the proof-of-concept of the developed technologies. By developing production processes that allows for an economic production of the developed materials (PILs and enzymes) and characterizing the developed technologies at lab- and demonstration scale the basis for further development to demonstration scale is set, and the developed process concepts already took into account issues related to further scale-up at a conceptual design level. The INTERACT consortium formed the ideal mix of partners for the pursued developments. Further progress in the direction of an industrial application requires appropriate industrial partners to implement and test the relevant technologies in an industrial environment and further develop the technology to suit the industrial requirements for a broad application.
European Transnational Approach
In order to develop the innovative materials and technologies for an economic and sustainable application of CO2 capture in an industrial application scenario we need to strengthen the cooperation between European industry and research institutions. A very important impact of the INTERACT project was the close interactive nature of the consortium and frequent cross-work meaning that many researchers on the project gained exposure to new skills in applying the advanced technologies evaluated. This demystifies and mitigatesrisks further technology developments and builds up the labor force of, especially, young researchers, who are employable by industry and academia to further develop and implement the technologies. As many different factors play an important role in energy production and power output of single power plants is introduced into a highly complex and strongly integrated multinational network, not only the cooperation between different countries, but also the integration into a holistic approach, considering carbon capture and storage or utilization for continuously operated fossil fired power plants as well as the discontinuously producing renewable energy production and the related energy storage, need consideration. This complex nationwide optimization problem can only be tackled efficiently if the overall efforts are strategically coordinated and the different technologies are put to the best possible use in an integrated and cooperative approach. Consequently, a European transnational approach provides the optimum platform for such coordinated developments to provide a more sustainable future throughout the European Union.
The realization of the current project has brought an improvement in terms of innovative and new materials, aiding the toolbox of different measures to provide more sustainable and economic solutions to reduce CO2 emissions in the scope of power production but also other large-scale emitters, such as e.g. cement production. The knowledge accumulation in the involved research institutions will eventually result in new activities improving the developments in the different sectors and with time have a socially positive impact on the European Community. The involved industrial partner and SMEs in this project already benefitted from the novel developments and related markets that can improve their short and long term profits.

List of Websites:


Coordinator: Andrzej Górak

Scientific Manager: Mirko Skiborowski

Administrative Management: Petra Marciniak