Multiscale Analysis of Membrane-Assisted Integrated Reactors for CO₂ Hydrogenation to Dimethyl Ether

Abstract

The conceptual design and engineering of an integrated catalytic reactor requires a thorough understanding of the prevailing mechanisms and phenomena to ensure a safe operation while achieving desirable efficiency and product yields. The necessity and importance of these requirements are demonstrated in this investigation in the case of novel membrane-assisted reactors tailored for CO₂ hydrogenation. Firstly, a carbon molecular sieve membrane was developed for simultaneous separation of CO₂ from a hot post-combustion CO₂-rich stream, followed by directing it along a packed-bed of hybrid CuO-ZnO/ZSM5 catalysts to react with hydrogen and produce DiMethyl Ether (DME). The generated water is removed from the catalytic bed by permeation through the membrane which enables reaction equilibrium shift towards more CO₂-conversion. Extra process intensification was achieved using a membrane-assisted reactive distillation reactor, where similarly several such parallel membranes were erected inside a catalytic bed to form a reactive-distillation column. This provides the opportunity for a synchronized separation of CO₂ and water by a membrane, mixing the educts (i.e., hydrogen and CO₂) and controlling the reaction along the catalytic bed while distilling the products (i.e., methanol, water and DME) through the catalyst loaded column. The hybrid catalyst and carbon molecular sieve membrane were developed using the synthesis methods and proved experimentally to be among the most efficient compared to the state-of-the-art. In this context, selective permeation of the membrane and selective catalytic conversion of hybrid catalysts under the targeted operating temperature range of 200–260 °C and 10–20 bar pressure were studied. For the membrane, the obtained high flux of selective CO₂-permeation was beyond the Robeson upper bound. Moreover, in the hybrid catalytic structure, a combined methanol and DME yield of 15% was secured. Detailed results of catalyst and membrane synthesis and characterization along with catalyst test and membrane permeation tests are reported in this paper. The performance of various configurations of integrated catalytic and separation systems was studied through an experimentally supported simulation along with the systematic analysis of the conceptual design and operation of such reactive distillation. Focusing on the subnano-/micro-meter scale, the performance of sequential reactions while considering the interaction of the involved catalytic materials on the overall performance of the hybrid catalyst structure was studied. On the same scale, the mechanism of separation through membrane pores was analyzed. Moreover, looking at the micro-/milli-meter scale in the vicinity of the catalyst and membrane, the impacts of equilibrium shift and the in-situ separation of CO₂ and steam were analyzed, respectively. Finally, at the macro-scale separation of components, the impacts of established temperature, pressure and concentration profiles along the reactive distillation column were analyzed. The desired characteristics of the integrated membrane reactor at different scales could be identified in this manner.

1. Introduction

1.1. Motivation and Strategic View

For sustainably reducing global CO₂-emission it is crucial to target the CO₂-generating sectors with priority, where the highest impacts can be secured via synchronizing the targeted requirements relying on the available technological solutions [1,2]. Considering the significance of the overall emitted CO₂ from the industrial sector, in particular steel plants, developing an efficient technology for reducing their CO₂-emission can be considered as a prioritized choice from the strategic point of view [2,3]. Decarbonizing a steel plant by improving the energy efficiency of the production line seems to be a remote possibility. Instead, capturing the emitted CO₂ from the targeted off-gas stream (e.g., the exhaust of a blast furnace) and producing green-fuel for partially making-up of the required fossil-base fuel is technically feasible and conceptually an attractive solution since it contributes to the overall reduction in carbon dioxide emissions. Therefore, it is becoming a part of the strategy for sustainable design and retrofitting the steel plants with far-reaching environmental, economic and societal impacts. Nonetheless, from a technical point of view and considering the targeted operating pressure, temperature and concentration of the involved gaseous species, developing such technology for efficient local CO₂ capture-conversion and utilization of the produced fuel in the same combustion chamber is challenging. For the most part, however, this can be potentially addressed using an integrated membrane-assisted catalytic CO₂-hydrogenation to Dimethyl Ether (DME) process, since DME with its relatively high energy efficiency can serve as a make-up fuel in such large-scale furnaces [4]. Nevertheless, this process is still in development and its required characteristics should be carefully analyzed and adapted in different scales to become a large-scale globally accepted established technology for this application [2,5]. These include addressing the requirement of the separation and conversion sections in the catalyst, membrane, reactor and process scales. In doing so, the state-of-the-art and scientific progress in the catalyst, membrane, reactor, etc., are reviewed first, followed by a description of the proposed methodology for the design of a sustainable integrated process for this application.

1.2. State-of-the-Art and Potential Solutions in Different Scales

1.2.1. Catalyst for CO₂ Hydrogenation to Methanol and Its Dehydration to Dimethyl Ether

The reactions representing the targeted catalytic paths including the conversion of CO₂ (hydrogenation reaction for methanol production) and producing DME (methanol dehydration) are, respectively, listed as reaction 1 and reaction 2.

$${\mathrm{CO}}_2+{3\mathrm{H}}_2\leftrightarrow {\mathrm{CH}}_3\mathrm{OH}+{\mathrm{H}}_2\mathrm{O}\hspace{1em}\hspace{1em}∆{\mathrm{H}}^{°}=-49.4 \mathrm{kJ}·{\mathrm{mol}}^{-1}$$

(1)

$$2{\mathrm{CH}}_3\mathrm{OH}\leftrightarrow {\mathrm{CH}}_3{\mathrm{OCH}}_3+{\mathrm{H}}_2\mathrm{O}\hspace{1em}\hspace{1em}∆{\mathrm{H}}^{°}=-23.4 \mathrm{kJ}·{\mathrm{mol}}^{-1}$$

(2)

These catalytic reactions, respectively, occur on bimetallic (e.g., benchmark CuO-ZnO) and zeolite (e.g., benchmark ZSM5) catalysts [5,6]. Simultaneously, forward and reverse water gas shift (WGS) reactions 3 are also happening primarily along with reaction 1.

$${\mathrm{CO}}_2+{\mathrm{H}}_2\leftrightarrow \mathrm{CO}+{\mathrm{H}}_2\mathrm{O}\hspace{1em}\hspace{1em}\hspace{1em}∆{\mathrm{H}}^{°}=41.2 \mathrm{kJ}{·\mathrm{mol}}^{-1}$$

(3)

There are other side reactions involved, such as carbon monoxide hydrogenation and carbon oxide methanation, with a less-significant contribution in shaping the reactor output composition under the mild reaction conditions (pressure and temperature, respectively, up to 20 bar and 260 °C) targeted in this research. This range has been selected to ensure that efficient in-situ CO₂ separation and catalytic conversion [7,8] including hydrogenation and methanol dehydration reactions over a hybrid catalytic structure [6], can take place simultaneously. Without having such restrictions, a wide range of catalytic materials and operating conditions could be utilized for the hydrogenation reactor [7,9,10,11] and methanol dehydration [12] reactor separately. An extensive review of single-step CO₂-hydrogenation to DME, its requirements and restrictions can be found elsewhere [6,13].

After selecting the catalytic materials (benchmark CuO-ZnO and ZSM5), analyzing the phenomena in the scale of individual catalytic sites, and identifying the targeted range of operating parameters, the next step is to analyze the phenomena in the scale reflecting the structural characteristics of hybrid of two catalytic materials. In general, such relations between the catalytic behavior, structural characteristics and the preparation methods for this family of catalysts have not been thoroughly studied. Therefore, developing an active, selective and stable catalyst for CO₂ hydrogenation has been demonstrated to be a challenging task, but highly rewarding [13,14].

1.2.2. Hybrid Catalysts for CO₂ Hydrogenation to Dimethyl Ether

The material-chemical-structural characteristics of the benchmark CuO-ZnO and ZSM5 catalytic materials have been analyzed extensively [7,9,10,11,12]. Having analyzed the reported data in the literature in this regard, different synthesis methods were applied in tailoring the material-structural-chemical characteristics of hybrid catalysts for one-step CO₂-hydrogenation to DME in the early stage of this research [6]. It has been demonstrated that by selecting the right synthesis method and material composition of hybrid catalysts, enabling a well-dispersion of the active components across the hybrid catalytic structure, DME-selectivity and yield could be improved [5,6,13].

After analyzing the characteristics of hybrid catalytic structures, the next step is to include the in-situ reactants dosing and product removal in the integrated reactor-scale reflecting the impacts of external transport phenomena on the performance of the catalytic reactor.

1.2.3. Membrane: CO₂-Dosing and Water Removal

There are three main approaches to addressing the limitations regarding thermodynamic equilibrium conversion of CO₂ to methanol [6,13,15] represented in reaction 1. The first approach comprises using a proper set of operating temperatures, pressure and reactant ratio (H₂/CO₂). In that context, the techno-economic requirements of such a hybrid catalytic system, in particular, the cost of the hydrogen and carbon dioxide and their downstream separation should be taken into consideration. As a result, the ranges of variation of these parameters have been fixed by defining the mild range of operating pressure and temperature and the overall H₂/CO₂ has been selected in this research to be stoichiometric. However, establishing a locally high H₂/CO₂ using a membrane CO₂-dosing along the bed can improve the carbon dioxide conversion, methanol selectivity and yield [8]. In fact, using a molecular sieve carbon membrane is a promising separation concept in this regard not only as it is compatible with the composition but also with the high temperature of the targeted CO₂-rich streams (e.g., steel mill gas). It has been demonstrated that such membranes with proper specifications [16] can be used for such tasks promising an energy-efficient separation process with a lower carbon footprint.

The second approach focuses on further converting methanol to other valuable products, which has been targeted to be DME in this research and materialized by developing a hybrid catalytic structure as described in Section 1.2.2. The third approach targets shifting the equilibrium by removing the products (e.g., water) through a perm-selective membrane from the catalytic bed [17,18].

In this research, all three approaches have been simultaneously implemented in designing a novel integrated membrane reactor. In such a membrane reactor, through a tailored carbon molecular sieve membrane, CO₂ is dosed from a CO₂-rich stream into the bed of hybrid catalyst to be converted to methanol and then DME, while the produced water permeates through the same membrane into the CO₂-rich hot gas stream. This enables an efficient process intensification for in-situ separation and conversion of CO₂ to DME. Several of such membrane reactor structures are assembled in the form of membrane-assisted reactive distillation for further reactor integration and process intensification as explained in the next Section 1.2.4.

This enables analyzing the impact of macro-scale distillation on tuning the temperature and concentration profile along the catalytic bed and thereby on the performance of the integrated catalytic-separation system.

1.2.4. Integrated Reactor and Process Intensification

Different levels of reactor integration and process intensification, which could be established for this catalytic process are reviewed in this section. From one side, a reactive distillation concept has been demonstrated [19] capable of efficiently integrating the methanol dehydration catalytic reactor and distillation-separation of the products (i.e., DME and water). Full conversion of methanol, handling the light non-condensable gaseous components (e.g., dissolved carbon oxides and H₂), and tuning the reaction intensity and temperature profile along the bed to secure high purity of the distilled DME, have been also among the capabilities demonstrated by such an integrated system [19].

On the other side, an improved methanol selectivity and yield have been secured using an integrated membrane reactor, where through an inorganic membrane CO₂ is separated from a CO₂-rich stream (e.g., off-gas of the blast furnace with 23% CO₂) and simultaneously distributed along the hydrogenation catalytic bed to be converted to methanol [8]. Selecting a proper membrane, which also enables simultaneous water removal from the catalytic bed, is known to be beneficial for the catalytic hydrogenation system [17,18,20]. This is particularly adapted in this research and applied in the form of a novel integrated membrane reactor, capable of simultaneous CO₂-dosing and water removal, for CO₂-hydrogenation to DME production.

2. Multi-Scale Analysis

In this section, after reviewing the general design-operation concept of the hybrid catalysts, membrane, and the integrated membrane reactor, through a multi-scale analysis-design approach, the know-hows and knowledge in different-scales required for the design of an efficient large-scale integrated reactor are explained. This is established in the form of an integrated reactive-distillation structure, promising significant process intensification potentials and improved energy efficiency in reference to the process consisting of a sequence of catalytic and separation modules. The detailed design and operating performance of the investigated process scenarios are explained in this section. In this context, the materials and methods utilized to analyze and tailor the characteristics in different scales of such configurations, shaping the performance of the phenomena in catalyst-scale, membrane and the integrated reactor scale and ultimately the process scale are reviewed here.

2.1. Catalyst

2.1.1. Material Selection and Synthesis

The applied multi-scale analysis in this research starts with systematically analyzing the involved phenomena and the material-chemical-structural characteristics of the catalysts and how to establish the desired ones in designing an efficient catalyst for CO₂-hydrogenation to DME. In the first step by focusing on CO₂-hydrogenation to methanol, it has been demonstrated that the well-dispersed active metal oxides (e.g., copper and zinc) across the CO₂-hydrogenation catalyst, usually being established in the form of smaller particles for instance through coprecipitation synthesis, result in higher catalytic conversion and methanol yield [7,10]. The relatively larger pore-volume and pore-size obtained through the coprecipitation synthesis method are also partially responsible for the relatively higher CO₂-conversion and methanol yield obtained in this case. Sequential precipitation synthesis of the CuO and ZnO in (50% CuO-50% ZnO) methanol-synthesis catalyst has shown a 2/3 reduction in the accessibility of the gaseous species to the catalytic hydrogenation surface area compared to the corresponding co-precipitated catalyst. These have been indicated through a comparative study of the measured BET surface area for these catalysts [6,21]. Such a comparative study included several catalysts synthesized via various methods also for hybrid catalysts. For instance, 94% further reduction in the surface of the catalytic hydrogenation area in the sequentially precipitated -SP- hybrid catalyst in comparison to the co-precipitated -CP- hybrid catalyst (33% CuO-33% ZnO-33% ZSM5) for CO₂-to-DME was observed. The CP hybrid catalyst in any case, however, reduces the access to the methanol-dehydration catalytic structure (e.g., ZSM5). As another general observation, impregnating the active metals (e.g., Cu and Zn) can result in less restriction on the accessibility of the gaseous species to the support structure. In particular, impregnating CuO and ZnO, either being simultaneously (Co-Impregnated -CI-) or in sequence (Sequentially Impregnated -SI-) introduced into the zeolite structure (constituting 2/3 of the hybrid catalyst structure), less significantly restricts the access of the reactants to the micropore area of the support. The catalytic hydrogenation materials do not provide an effective high surface area in this case. In fact, the impregnation method results in a less-distributed hydrogenation active material across the hybrid catalytic structure, and therefore, overall CO₂ conversion is reduced [21]. Hinting to a relatively less restricted access, the results of TPD analysis (

Table 1. The surface chemical-structural characteristics of the investigated hybrid catalysts.

Catalyst/Sample	Composition (wt.%)	SBET (m2·g−1) a	Vpore (cm3·g−1) b	Pore Size (nm) c	NH3uptake (µmol·g−1) d
SI	16.7% CuO: 16.7% ZnO: 66.6% H-ZSM-5	194.5	0.17	4.5	217.7
CP	33.3% CuO: 33.3% ZnO: 33.3% H-ZSM-5	150.5	0.33	9	156.5
CI	16.7% CuO: 16.7% ZnO: 66.6% H-ZSM-5	233.2	0.19	4.1	224.6
SP	33.3% CuO: 33.3% ZnO: 33.3% H-ZSM-5	141.8	0.19	5.9	97.7

^a BET surface area; ^b Pore volume; ^c Pore size calculated using BJH method; ^d Ammonia-TPD measurement indicating the surface acidity.

) also indicate that the representative medium acidity strength of ZSM5 is reflected proportional to its weight percentage in the impregnated hybrid catalyst [6]. These data also indicate a slightly further reduction in the accessibility in the SI hybrid catalyst compared to the CI one. This is in line with the overall observed surface area measurements for these catalysts. The recorded overall surface area and ammonia-adsorption for the catalysts synthesized via coprecipitation (CP) and sequentially-precipitation (SP) methods with the same composition CuO-ZnO-ZSM5 (33%-33%-33%) showed a similar trend (See Table 1). These data also indicate a relatively more restricted access in the SP catalyst.

Having considered all these, the experimentally observed performances of the CP catalyst were selected to be used for the rest of the analysis reported in this paper because; (a) it can secure a significant CO₂-conversion even in the form of a hybrid catalyst, (b) it represents (at least theoretically) a homogeneously distributed catalytic materials for the consecutive reactions, which might secure even significant level of methanol-selectivity and DME yield, (c) considering the relatively larger pore-volume in such CP hybrid catalyst, the observed water condensation in its structure serves as a reference as how much this phenomena can be avoided in the scale of hybrid catalytic structure. The synthesis procedures of the single-coprecipitated catalyst for methanol synthesis and the hybrid coprecipitated catalyst for direct DME production are provided here and the detailed synthesis of the other catalysts mentioned above along with their characterization results are found elsewhere [6,21].

Coprecipitated catalyst for methanol production and hybrid catalyst for DME production:

The details of the coprecipitation synthesis recipe have been reported earlier [6,21]. As the first step of this synthesis procedure, separate aqueous solutions of copper nitrate Cu(NO₃)₂·3H₂O (purchased from VWR Chemicals, Radnor, PA, USA), zinc nitrate Zn(NO₃)₂·6H₂O and sodium carbonate Na₂CO₃ (both purchased from Sigma Aldrich, St. Louis, MO USA), each with a molarity of 0.2 M were prepared. The metal nitrate solutions were mixed to establish a light blue solution with a mole ratio of 50Cu:50Zn; 400 mL of this solution and 500–600 mL of the precipitant carbonate solution were dropwise added to 600 mL of De-Ionized (DI) water under vigorous stirring of 500 rpm. The metal nitrates and the precipitant solutions were added with controlled drop-rates for each solution while maintaining the pH of the solution around 7 ± 0.2 at the temperature range of 60–65 °C. After obtaining the precipitate, it was aged for 2 h at the same temperature and stirring conditions. Then, the precipitate was washed with hot DI water and filtered and then dried at 80 °C. After crushing and pelletizing the dried solid, it was sieved to obtain the desired size fraction (100–250 µm) and then calcined at 360 °C. The CP methanol catalyst is at this stage ready to be characterized and tested.

For the hybrid hydrogenation-dehydration CP DME catalyst, a similar procedure as explained above was followed with the only difference being that the 400 mL of the mixed metal nitrate (light blue) and 500–600 mL of the precipitant solutions were both added dropwise to 600 mL of demineralized water containing 3.2 g of H-ZSM5 (after calcination) under stirring at 500 rpm to obtain the targeted hybrid CuO-ZnO/ZSM5 catalyst.

2.1.2. Catalytic Performance Test for Various Hybrid Catalytic Bed Structures

Having considered the targeted process integration, the catalysts have been tested in an operating temperature range of 160–280 °C under a mild operating pressure of up to 20 bar [6,7].

In this section, starting with the performance of CP dual catalyst, the effect of operating parameters, namely temperature and residence time (reflected through Gas Hourly Space Velocity -GHSV-), on the catalytic performance of CO₂-hydrogenation to methanol and DME are reviewed. The catalysts were tested in a bench-scale fixed bed reactor with an inner diameter of 7.75 mm, equipped with a backpressure regulator and compact GC analyzer Interscience 4.0. More detailed specifications of the reactor setup and testing procedure can be found elsewhere [6,21,22]

Figure 1 represents a typical impact of varying operating temperatures on different performance indicators of the CP and CI hybrid catalysts [21]. Compared to the CI hybrid catalyst, the relatively better performance indicators of the CP hybrid catalyst are highlighted in this figure. The only recorded performance indicator that was higher for the CI hybrid catalyst was DME selectivity. This can be explained considering the recorded lower CO₂-conversion and the larger portion of zeolite content and less accessibility restriction of the gaseous species to the zeolite support in the hybrid CI catalyst compared to the CP hybrid catalyst.

This range of tested temperatures, namely 200–260 °C, not only covers the common temperature range, where CuO-ZnO and ZSM5 catalysts have been recommended to be utilized as a hybrid catalyst [6,7] but also reflects the temperature profile along the integrated reactor system investigated in this research. As seen in Figure 1, DME selectivity and methanol selectivity (for the targeted CP dual catalyst) are reduced by increasing the temperature, while CO-selectivity, CO₂-conversion and thereby the yield of desired products increase. Ensuring distributed catalytic materials for CO₂-hydrogenation across the whole catalytic structure from one side and less restriction in accessibility to the zeolite structure on the other side, were identified to be two important aspects in designing the characteristics in the scale of the internal meso-micro dimensions of a hybrid catalyst particle.

In the next scale, namely a micrometer-millimeter scale covering the interactions between the catalytic particles, the performance of various mixed-bed of hydrogenation and methanol dehydration catalysts was analyzed. In this manner, the performance of a bed made of CuO-ZnO catalyst followed by a bed of ZSM5 catalyst as a two-layer catalytic bed along with the performance of a 10-layer bed of a sequence of similar structure were experimentally tested. As a reference state representing a complete macro-scale mixed bed of these catalysts, a fully mixed bed of these catalysts’ particles “Physical Mixing” was also tested [6,21]. The comparative results of their performances indicated similar trends as observed for the hybrid catalysts (shown in Figure 1) in terms of CO₂-conversion and CO-selectivity as well as methanol and DME selectivity. It was demonstrated that the yield of desired products (i.e., methanol and DME) could be improved [6] using a well-mixed bed of these catalytic materials.

GHSV is another variable, which should be set to ultimately improve the catalytic performance according to the conditions in different sections of the macro-scale integrated reactor. It has been experimentally observed that in the studied range of variables in this research, reducing the contact time, through increasing the GHSV, both CO₂-conversion and the yield of the desired products (i.e., methanol and DME) are reduced. Methanol selectivity was found to be less affected by varying the contact time. The thermal operation of the reaction is significantly affected by the intensity of this exothermic reaction, which should be reflected in the interpretation of such observations.

Typical trends of the impact of varying H₂/CO₂ ratio on the catalytic performance of the CO₂-hydrogenation to methanol and CO₂-hydrogenation to DME have been investigated [7,21,22]. A higher H₂/CO₂ ratio than stoichiometric ratio 3 increases methanol and DME selectivity, while as expected it reduces the CO₂-conversion and the per-pass yield of the desired products. This has been demonstrated not only in a co-feed fixed-bed reactor, but also by establishing a local high H₂/CO₂ ratio in a membrane reactor [8]. This parameter and such impacts will be utilized in designing the integrated reactors in this research.

The effect of varying operating pressure was experimentally studied in the range of 10–20 bar and demonstrated to be less-straightforward [6,7,21,22]. On one hand, increasing pressure will result in an increase in the CO₂-conversion. On the other hand, its impact on the selectivity towards methanol and DME depends on the structural characteristics of the catalysts as well as the set of operating conditions, under which the effect of pressure is being investigated. Generally, higher pressure is in favor of higher methanol production, but not necessarily higher DME production. It should be also highlighted that selecting the operating pressure in the targeted integrated reactor in this research is strongly related also to the performance of the membrane, which is studied in the next section. On the other hand, using a higher H₂/CO₂ ratio improves the CO₂-conversion and methanol selectivity. Therefore, the membrane-assisted reactor engineering scenario, where dosing CO₂ with a lower partial pressure and a local low H₂/CO₂ ratio along the catalytic bed can address these potentials and would be conceptually attractive. At the end of this section, using lower temperatures not only from a thermodynamic point of view is expected to shift the equilibrium in favor of selective methanol and DME production, but also kinetically been demonstrated to be favorable in that regard. CO₂-conversion, however, would suffer by decreasing the temperature. The impact of temperature in the form of establishing a desired temperature profile along the catalytic bed should be primarily addressed through feeding strategy, establishing the desired reaction intensity profile along the bed and tuning the gas composition along the bed for instance via bulk-separation mechanism such as distillation.

2.2. Membrane and Material Processing

This section provides an overview of the specifications and the performance of the membrane tailored for the applications investigated in this research, namely for integrating catalytic CO₂-hydrogenation with steel mill gas treatment in the form of an integrated membrane reactor and membrane-assisted reactive distillation systems.

2.2.1. Membrane Synthesis and Potentials

A carbon Molecular Sieve Membrane (CMSM) was used in this research for the separation of CO₂ from steel mill off-gas with an assimilated composition of 4% H₂, 22.4 CO₂, and 73.6% N₂. In fact, steel mill gas contains approximately 22% mole-fraction of CO₂ and similar content of CO as well as a small portion of methane (about 4%), which were not included in the experimental analysis primarily due to the safety restriction of the testing-facility. Their portion was considered to act like the nitrogen content (originally around 50%). The impact of the water content of the steel mill gas, however, was qualitatively considered by running a controlled-saturated and water-saturated gas feed in the experimentations.

In particular, it is targeted in this research to integrate an in-situ selective CO₂-separation through the membrane with the adjacent catalytic CO₂-conversion, where CO₂ is simultaneously distributed along the catalytic bed while being ideally hydrogenated to methanol [8] and DME. The general characteristics of this CMSM, including its capability to operate at high temperature and pressure [16] along with the available potential in tuning the synthesis parameters to tailor its selective separation performance, are the main advantages of this type of membrane for the targeted application in this research.

An asymmetric porous alumina tube with an inner diameter (ID) of 7 mm and outer diameter (OD) of 14 mm with an external layer of 100 nm average pore size, supplied by Inopore GmbH, was used as the support for each membrane preparation-batch of the experimental demonstration/study. At each batch, three membranes were prepared using similar materials, preparation conditions and synthesis procedures, indicating a reproducible performance. At first, the tubular alumina support was connected to a dense alumina tube, on one side to form the permeate side outlet, while the other side of the support was blocked via glass sealings prepared at 950 °C with 10 min curing-time. The details of the general assembly can be found in our earlier publication [16].

Firstly, 69 g (0.73 mol) of phenol (99%, CAS No. 108-95-2) was melted at 60 °C in a four neck round bottom glass flask, which was equipped with a reflux condenser. Then, 1.5 g of oxalic acid (98%, CAS No. 144-62-7) was added to the solution, the temperature was raised to 90 °C and 52 g (0.64 mol) of formaldehyde solution (37% VWR chemicals, CAS No. 50-00-0) was added dropwise to the solution. The reaction was carried out for 8 h, the solution was centrifuged at 20 °C and was rinsed three times using denoised water with a speed of 4000 rpm for 15 min in each step. The oligomer was collected in a porcelain dish and was vacuum dried at 30 °C for 24 h. In the next step, 28 g of the resulted oligomer was dissolved in 92 g of N-methyl-2-pyrrolidone (NMP, 99.5%, CAS No. 872-50-4) in a high shear mixer (Thinky ARE-250) with 1800 rpm speed for a duration of 30 min in two-stages of mixing. Then, 0.3 g of oxalic acid was added to the solution and mixed for 20 min at 2000 rpm. In the next step, 1.8 g of formaldehyde solution was added to the solution and mixed for two cycles of 30 min at 1500 rpm and 1.2 weight percentage (wt.%) of ethylenediamine (99.9%, CAS No. 107-15-3) was added to the solution and mixed for 30 min at 1000 rpm. The supported membranes were established by dip-coating the support in a dipping solution and drying it in accordance with the thermal-treatment and carbonization methods. All the chemicals were added to the membrane preparation procedure without further purification.

The synthesized membrane, with the characteristics reported in

Table 2. The characteristics of the CMSM used for the permselectivity tests.

Type	Support Pore Size	# of Layers	Diameter	Length	Carbonization Temperature	Polymerization Temperature	Ethylenediamine in Dipping Solution
CMSM	100 nm	2	10 mm	180 mm	600 °C	90 °C	1.2 wt.%

, was selected for the experimental study in this research. The superior selective performance of CMSM for CO₂-separation from steel mill gas has been demonstrated in the form of CO₂/N₂ ideal selectivity Robeson upper bound [16]. General information about the synthesis, characterization and typical separation performance of CMSM, in particular, for similar high temperature-pressure permeation tasks, can be found in earlier publications [16,23,24].

After carefully identifying the proper CO₂-selective carbon membrane, its permeation towards CO₂, H₂O, H₂ and other involved components was experimentally measured.

2.2.2. Membrane Separation Performance: Permeation Study and Screening the Conditions

Performance indicators of the selected CMSM around the targeted range of operating conditions, namely 150–260 °C and 5–30 bar, assimilating its separation behavior in the investigated integrated reactors in this research, in particular, in the membrane-assisted reactive distillation system, are reviewed in this section.

For assimilating the operation of the integrated reactors under different operating conditions, we tuned and set the temperature around the membrane and the pressure in the retentate and permeate sides of the high-pressure permeation test setup at different levels. The flowrate and the gas compositions of the outlet gas streams leaving the retentate and permeate sections in these setups were, respectively, measured using Horiba VP1.3 and Agilent Micro GC Avian 490. More information about the permeation test setup can be found in our previous publications [16].

The CO₂-permeation of the CMSM was studied under different operating conditions and states of the membrane, namely dry, atmospheric saturated and saturated at 5 bar pressure of the permeate-stream. This is represented quantitatively by the values of the CO₂-concentration permeated across the CMSM, which can be tracked while changing the operating temperature and pressure as shown in Figure 2.

Figure 2a shows that securing a greater mass transfer driving force across the membrane, by establishing a smaller pressure on the permeate side, increases the CO₂-permeation as indicated by the recorded higher concentration of CO₂ on the permeate side. In fact, the catalytic bed is positioned on the CO₂-permeated side in the membrane-assisted integrated reactors studied in this research. Therefore, the value of pressure on the permeate side along with the values of other parameters are determined in accordance with their effects on both the catalytic and separation performances. Figure 2b shows that setting the lowest pressure (atmospheric pressure) on the permeate side can even further facilitate the permeation, especially at lower temperatures. However, the catalytic performance under such low pressure is significantly hampered.

By increasing the temperature, the absolute value of permeance (mol·m⁻²·s⁻¹·Pa⁻¹) increases for each component in a single-component permeation [16]. However, the permeation of CO₂, H₂ and N₂ could show different trends and intensities in competition with each other while varying temperatures. For instance, the CO₂ adsorption-solution is the dominant governing mass transfer mechanism across the membrane in the presence of water. Therefore, a varying temperature in the presence or absence of water affects the CO₂ permeation for the dry and saturated states of the membrane differently (Figure 2b). In this manner, as seen in Figure 2a,b, the increase in temperature can result in a decrease in relative CO₂ permeate percentage. This is due to the fact that by increasing temperature; (a) some of the other components (e.g., hydrogen) permeate faster than CO₂, and (b) the adsorption diffusion as the main CO₂ transport mechanism is weakened. This is indicated by the recorded increasing trend of permeate-percentages with the rising temperature under most sets of operating conditions (Figure 2c,d), in particular, for the saturated membranes. The observed trend of CO₂-permeate while changing the temperature can be explained by analyzing its impact on the CO₂-adsorption diffusion mechanism, and by tracking the reverse but yet complementary trend of the recorded hydrogen-permeate (mol%). In this manner, the inhibiting effect of temperature on CO₂ separation boosts the competitive hydrogen diffusion. This explains why hydrogen permeation across dry membranes is higher than the saturated ones as shown in Figure 2d.

The permeation trends of reference nitrogen while changing the temperature have been also shown in Figure 2e,f, indicating that molecular sieving is the dominant mass transfer mechanism for nitrogen. In particular, the ascending nitrogen concentration on the permeate side while using a dry membrane is an indication of the relatively straightforward relation between the nitrogen permeation and operating temperature in this case. The recorded real selectivity CO₂/N₂ representing a normalized qualitative indicator of the CMSM in terms of CO₂-separation performance under different operating temperatures and pressures is also shown in Figure 3.

This relationship can be also tracked in Figure 3a,b. As seen in Figure 3b for instance, the absolute values of relative selectivity of CO₂/N₂ are higher than in Figure 3a, indicating that the relative CO₂-separation is intensified through the adsorption diffusion mechanism using a saturated membrane. Tracking the represented trends of the relative selectivity of CO₂/N₂ for different operating pressures and temperatures for the dry membrane (Figure 3a) is even more straightforward. Figure 3a clearly shows that by increasing pressure and decreasing temperature, the relative selectivity of CO₂/N₂ increases. The trends shown in Figure 3b should be analyzed considering the effect of varying operating pressure and temperature through impacting the mechanism governing the CO₂-separation across the membrane. The impact of temperature in this regard can be clearly seen in Figure 3c, where using dry membrane real selectivity CO₂/N₂ decreases monotonically by increasing temperature. This reflects the straightforward contribution of the molecular sieving mechanism in the absence of water. Again, here (in Figure 3c) it can be seen that the maximum pressure difference across the membrane has resulted in higher permeation.

On one hand, since the presence and the intensity of the interaction of water with CO₂ under different operating temperatures do not follow a monotonic pattern, such a straightforward trend was not expected and observed for the saturated membrane.

On the other hand, since molecular sieving is the governing mass transfer mechanism for both hydrogen and nitrogen [16], varying pressure similarly affects their overall separation performance trends. Therefore, no clear trend is observed in Figure 3d with regard to the impact of pressure on H₂/N₂ selective permeation. Having considered the tight competition of H₂ and CO₂ in permeating through the dry membrane, similar behavior regarding the impact of temperature on the H₂/N₂ and CO₂/N₂ selective permeation was observed, especially in the low operating temperature range. The saturated membrane in this range, however, showed a relatively higher (by comparing Figure 3b,e) and descending CO₂/N₂ selectivity permeation profile, while increasing the temperature (Figure 3b) reduces the CO₂-permeation through weakening the adsorption diffusion mechanism. This can be tracked by comparing the recorded trends of CO₂/N₂ and H₂/N₂ real selectivities for the dry membrane in Figure 3c,f in the absence of water and, in particular, at higher temperature ranges. Comparing the trends for the saturated membranes on the other side indicates the strong contribution of the adsorption-diffusion mechanism in facilitating the CO₂-permeation. The dominant effect of such phenomena is also responsible for the recorded real H₂/CO₂ selectivity to be lower than 1 for the saturated membranes, while it is higher than 1 for the dry membrane as shown in Figure 3g,h. As shown in Figure 3h, the observed ascending real H₂/CO₂ permeation selectivity of saturated membranes with rising temperature, where CO₂-adsorption diffusion is weekend, is also in line with this analysis. For the dry membrane in the absence of water, the competition between CO₂ and H₂-permeation is tighter favoring the latter at higher temperatures, where molecular sieving becomes dominant. In this manner, as seen in Figure 3g, except for the low temperatures, the real H₂/CO₂ selectivity is not significantly affected by the set pressure on the permeate side. This is especially the case at the higher temperatures corresponding to the operating temperature of the integrated reactor in this research. This along with the impact of the interaction of gaseous species with water on the quantity and qualitative selective separation performance of the membrane needs to be carefully accounted for in the design of the integrated reactors. Possible interaction with methanol should be also taken into consideration as it and water are the side products of the targeted reactions in the studied integrated reactors in this research. This water content is much higher than the water content of the steel mill gas.

2.2.3. Membrane Separation Performance: In-Situ Separation in Integrated Reactors

In this section, the observed separation performance of the membrane while CO₂, H₂, N₂ and H₂O compete to permeate through it are assimilated with its separation performance in the proposed membrane reactor and membrane-assisted reactive distillation in this research.

A hydrophilic membrane is preferred for the membrane-assisted integrated reactors in this application as it potentially facilitates water and even possibly methanol-removal out of the reactive atmosphere. CO₂ permeates from CO₂-rich streams by the driving force of a sharp CO₂ partial-pressure gradient across the membrane into the catalytic bed, where its concentration due to rapid consumption can secure a high H₂/CO₂ ratio. In fact, only establishing a CO₂-dosing along such membrane-assisted integrated reactors enables an improved removal of methanol and water in comparison to the co-feeding of H₂ and CO₂. The reason for this is that the partial pressure of CO₂ and H₂ in the catalytic bed of the later one is high, and therefore, they (CO₂ and possibly hydrogen) can permeate out along with the targeted water and/or methanol, which makes such a design implausible. The next key-design aspect, besides ensuring the CO₂-dosing and in-situ removal of water, is to enable selective CO₂ permeation to be facilitated over hydrogen through the established permeation mechanism under a targeted set of conditions. This should be materialized by selecting the synthesis procedure and tailoring the chemical-material-structural characteristics of the membrane support and its synthesized layers.

The tested synthesized CMSM in this research showed a high hydrophilicity and permeation selectivity towards carbon dioxide in the presence of water in competition with the reactions’ reactants and products under targeted operating ranges of temperatures and pressure, identified for the catalytic reactions thoroughly studied at Section 2.1. This includes studying, in particular, the impact of water presence in the steel mill gas as well as the water generated during the CO₂-hydrogenation and methanol dehydration reactions. By injecting different levels of steam (i.e., 10% and 20%) and applying the targeted range of temperature (150–260 °C), permeation of carbon dioxide, hydrogen and nitrogen in a mixture of all components (i.e., set I [30% CO₂, 50% H₂, 20% H₂O] and set II [30% CO₂, 60% H₂, 10% H₂O]) through the membrane was tested. It was observed that when the membrane pores are filled with water when the feed gas contains a significant portion of steam, assimilating the water content generated through catalytic hydrogenation and dehydration reaction in the integrated reactor, loss of hydrogen will be reduced and separation of CO₂ is facilitated. The selected values of mole fractions of carbon dioxide and hydrogen in this experimental study represent the average value of each of these components in the retentate and permeate sides. Figure 4 depicts the recorded typical trends.

The observed trend of permeation of all three components (i.e., H₂, CO₂, H₂O) in the form of tracking their mole fraction in the permeate side while changing temperature is shown in Figure 4a. The contribution of the adsorption-diffusion mechanism in facilitating CO₂ permeation along with water is pronounced at low temperature ranges. At a higher temperature range, the wall of the membrane pores becomes drier and directly interacts with all three components rather than staying wet. Therefore, by reducing the water permeation at high temperatures, the contribution of interactive CO₂-permeation becomes lager reflecting its relatively higher content (30% CO₂ and 20% water). Hydrogen permeation governed by molecular sieving is facilitated at higher temperatures, indicated by its recorded rising trend of mole fraction of hydrogen on the permeate side as shown in Figure 4a. This is particularly seen in Figure 4b, where the H₂O/H₂ selectivity reduces by increasing the temperature. This is inline with the ascending trend of single-component hydrogen permeation by increasing temperature.

It should be highlighted that usually, the hydrogen content of steel mill-gas (about 10%) is higher than its considered value in this research. The higher hydrogen content of such CO₂-rich gas acts as extra assurance that the predicted separation and reaction performance of the integrated reactors investigated in this research could be obtained in real applications.

The reported permeation behaviors in this section hint that in the membrane-assisted integrated reactor (a) selecting the pressure in the CO₂-permeate side (the catalytic bed) is determined primarily by the kinetic requirements and then because of relatively less significant impact on the separation performance, (b) using a hydrophilic membrane under water saturation is expected to be the desired scenario, which favours higher CO₂-separation and lower hydrogen permeation-loss, (c) the temperature profile along the membrane and the catalytic bed should be tailored considering the requirements for both the separation and reaction. These will be addressed in the design of integrated reactors in the next sections.

2.3. Evolutionary Analysis of the Integrated Reactor-Separation Systems

In the previous section, the impacts of operating conditions including temperature, pressure and gas compositions in both the retentate and permeate sites of the membranes on selective CO₂-separation were reviewed using experimental data. In this section, the conceptual design and the considerations regarding the desired operation and the technical characteristics of the catalysts and membranes are reviewed. In the reactor-scale, utilizing a proper reactant feeding strategy could improve both the micro-scale and macro-scale operation of the reactor. This, along with other design aspects and operating parameters, should be synchronized to secure an improved performance of the whole integrated system through shifting the equilibrium in different scales in favor of the energy-efficient high selective CO₂-conversion to DME.

2.3.1. Integrated Reactors: Membrane Reactor

In the proposed integrated reactors in this research, overcoming the thermodynamic equilibrium is materialized in three different scales, namely catalyst-scale (phenomena on the nano-micro scale), membrane-scale (phenomena on the nano-macro scale) and the reactor-scale (phenomena on the micro-macro scale). In the catalyst-scale using a hybrid catalytic structure, the produced methanol through hydrogenation is quickly converted to DME. The potentials and limitations in that regard were explained in Section 2.1. In the micro-scale of reactor operation, through a membrane, higher local H₂/CO₂ could be established and thereby the selective CO₂ conversion could be improved [8]. In the same micro-scale reactor operation, it has been demonstrated that removal of methanol and water would be also beneficial in displacing the equilibrium in favor of securing higher selective CO₂-conversion [17,18,20,25,26,27]. In order to utilize these potentials, a novel integrated membrane reactor is proposed schematically represented in Figure 5. The operating concept of this membrane reactor can be summarized as the simultaneous removal of CO₂ from a CO₂-rich stream (e.g., steel mill gas stream) and its permeation into the hydrogenation catalytic bed, where it reacts with hydrogen, while excess generated water is removed through the membrane into the CO₂-rich hot gas stream.

The catalytic bed can be considered made of the best hybrid CuO/ZnO-ZSM5 catalyst demonstrated in Section 2.1. The operating temperature (160–260 °C) and pressure (10–20 bar) are selected accordingly targeting to maximize the selective conversion of CO₂ to the desired products, primarily DME, and simultaneously secure a high selective CO₂-permeation and water removal through the membrane as explained in Section 2.2.

The composition of the feed stream on the retentate side is the composition of the targeted CO₂-rich gas stream entering the tube side of the membrane-reactor. The composition and the flow of the feed stream on the permeate side, which is the catalytic bed, is the set composition of the targeted H₂-rich gas stream entering the catalytic bed. This is selected based on the observed performance of the catalytic bed under the targeted H₂/CO₂ ratio. By providing enough membrane area (using several parallel membranes as shown in Figure 5c) and securing enough CO₂-permeation along the reactor, the overall accumulated mole of the permeated CO₂ divided by the inlet mole of H₂ reaches the targeted H₂/CO₂ ratio. This ratio in the bulk reaction atmosphere can be locally tuned in a multiple-pass sequence of such configuration, where the bulk gas composition could be altered between the sections as shown in Figure 5d. The detailed operating concept of such a multi-tubular multi-pass membrane-assisted integrated reactor will be explained in Section 2.3.2. In such a configuration, the hybrid catalyst is placed between the tubular-membranes. Therefore, water generated inside the catalytic bed increases the water content of the gas stream outside the membrane, which will be reflected through a relatively higher fugacity of water outside the membrane in reference to the water fugacity inside the membrane.

Regarding the performance of the membrane in such a configuration, setting the pressure inside the membranes (e.g., 30 bar on the retentate side) significantly higher than the pressure outside the membranes (e.g., 10 bar on the permeate side) will increase the CO₂-permeation and prevents the hydrogen escape from the catalytic section. Fast reactions inside the catalytic bed also intensify these desired phenomena.

2.3.2. Integrated Reactors: Membrane-Assisted Reactive Distillation

In Section 2.3.1 the operating concept of the membrane reactor and multi-tubular multi-pass reactor was explained as depicted in Figure 5, where the gas-feed (10–20 bar, 160–260 °C) to the catalytic hydrogenation chamber enters from one side and leaves the chamber as a bulk gas stream containing the unreacted reactants and remaining products from the other side. The flow rate inside (CO₂-rich stream) and between the tube (H₂-rich stream) and the length and diameter of the membranes are determined primarily based on the targeted CO₂-permeation (e.g., 150 mol·m⁻²·s⁻¹·Pa⁻¹). In a multi-tubular multi-pass reactor (Figure 5c,d), a distinguished macro-scale feature, namely tuning the bulk gas compositions via feed-splitting-injection along the reactor, has been established. In this manner, while CO₂ is gradually distributed and H₂-conversion is progressing along the bed, refreshing feed portions are injected between the pass-chambers to enable keeping a H₂/CO₂ ratio in the macro-scale high and further improve the CO₂-conversion in the sequential conversion-chambers in three passes. Depending on the performance of the catalyst and the membrane, the number of passes could be two, three or more. Considering these and the flow rate and dimensions of the reactor system, three passes have been used in the conceptual design so that the effective overall H₂/CO₂ ratio could be established. The cool feeds also enable cooling-down the reactive-zone and controlling the temperature along the reactor. Moreover, the desired residence/contact time in different pass-chambers can be optimized by choosing the numbers, lengths, and diameters of the membranes in each pass-chamber. The liquid and vapor phases coexist in the catalytic bed, constructed with enough voidage, for instance in the form of small bale-packings [19]. The catalytic bed in this manner, not only contributes to converting the reactants but also for the thermodynamic separation of species. Constructing the sequence of pass-chambers vertically and enabling the co-existence of gas and liquid phases inside the catalytic bed in the form of a distillation column, facilitates controlling the temperature and establishing the desired temperature profile along this integrated reactor. This introduces an extra distinguished feature for affecting the macro-scale operation of this system through the bulk separation of the products and unreacted reactants inside the catalytic bed via distillation and selective permeation through the membrane. These, for instance, result in the separation of the components with high-boiling temperature (e.g., water) and low-boiling temperature (e.g., DME), respectively, from the bottom and top of the catalytic bed inside the column. Carbon dioxide, water, and even methanol can be separated through the membrane and either converted on the other side of the membrane or taken out of the integrated reactor. These features are established in the form of a membrane-assisted reactive distillation system as shown in Figure 6a. The unreacted light-components such as hydrogen and CO could be easily separated from methanol in a partial condenser before being sent back as a reflux stream into the catalytic bed.

The energy efficiency of such an integrated system could show an improvement over the energy efficiency of a sequence of separation and catalytic units [28] and even with the view of the energy density of the DME product [29]. The conceptual representation of the design and operation of these alternative systems are shown schematically in Figure 6.

The CO₂-concentration inside the catalytic bed of this integrated reactor is effectively kept low, which favors the selective hydrogenation and enables establishing a sharp concentration profile across the membrane and thereby higher CO₂-permeation. Depending on the flow rate of the H₂-rich stream, the CO₂ mole fraction can be even lower than 10% on average at the shell cross-section, which is equivalent to a H₂/CO₂ ratio of 9 or higher. The operating conditions inside the catalytic bed are designed, for instance, the temperature profile in the range 200–260 °C along the reactor from top to bottom, to get as close as possible to the optimal operation needed for the hybrid catalytic system. The overall H₂/CO₂ ratio is also controlled by a split-feed mechanism, through which not only the gradual declining H₂/CO₂ ratio in the upward bulk-flow stream before entering the next pass-chamber is compensated to be set higher, but also the temperature and the flow rate at each section can be tuned. The latter one along with defining the number and length of the membranes and cross-sectional areas of the catalytic section determines the effective GHSV and indirectly the amount of CO₂ to be converted in that section. The hybrid catalyst has been tested in the GSHV range of 200–400 h⁻¹. Mixing the product stream from the previous section with the relatively cold fresh feed portion (Feed 2 and Feed 3 in Figure 5) compensates for the declining effective reactants H₂/CO₂ ratio; overall the residence time increases without significantly increasing the risk of undesired non-selective reactions. In fact, at the top section of the reactive distillation column, the improved selectivity of methanol production due to lower operating temperature can up to some extent compensate for the decline in selectivity due to the lower effective H₂/CO₂ ration there, as explained in Section 2.1.2. In this configuration, light species and, in particular, hydrogen tend to rise to the top, where the selectivity by introducing fresh H₂-rich feed effective H₂/CO₂ ratio can be improved adjacent to the CO₂-rich stream inside the membranes. The temperature on the top (e.g., 200 °C) is lower than at the bottom (e.g., 260 °C), and therefore, the selective conversion is favored there. The combination of two factors, namely the intensely generated water and the low temperature on the top section enables reducing hydrogen permeation-loss. DME is volatile and tends to rise up from the top and escape the reaction chamber as soon as it is generated. Methanol is a heavier (relatively less volatile) component and tends to join the reflux ratio on the top and fall-down to the bottom, where it gets evaporated and converted to DME under higher temperatures over the hybrid catalyst. Water is continually removed through the membrane. The removal of water through the membrane is particularly beneficial in shifting the CO₂-conversion equilibrium towards more methanol and DME production.

In this manner, the unreacted reactants and the products will be also separated along the bed. The water-rich stream leaves the bottom of the column from the permeate side and is reboiled to evaporate its methanol content and get it back into the catalytic bed. The DME-rich stream leaves the top of the column from the retentate side.

In a similar configuration (as shown in Figure A1) designed for methanol dehydration to DME, the unreacted reactants and light components are separated in a partial condenser at the top of a reactive distillation column. The generated methanol, water and DME are separated along the column via a distillation mechanism enabling the establishment of their descending or ascending concertation profiles. Water is removed from the bottom of the column while the methanol-rich stream is feed-split into the catalytic bed and converted to DME and water has been previously demonstrated to be efficiently feasible [19]. In the proposed membrane-assisted reactive distillation in this research, additional advantages through the contribution of membrane and separation of components through membranes as well as integrating the CO₂-hydrogenation and methanol-dehydration over hybrid catalytic structure, have been secured. For instance, in the partial condenser, the separated gases can be boosted and sent back to the feed gas stream to be further processed. In this manner, a high level of process intensification and significant energy efficiency could be secured, which is essential for such technologies [30]. Figure 6a depicts the configuration and operating concept of the proposed membrane-assisted reactive distillation.

The direct impacts of the micro and macro-scale mechanisms on the equilibrium-limited conversion of CO₂ inside the hybrid catalyst structure as well as in the vicinity of the membrane and along the multi-pass integrated reactors were analyzed. Now, their indirect impacts on the catalytic performance being felt via altering the thermal behavior of the system are reviewed to complete the analysis. The interactive impact of temperature and other parameters is particularly investigated in this regard. For instance, the high temperature of 260 °C and the highest-water concentration in the bottom of the membrane-assisted reactive distillation column affect the selective-permeation efficiency across the membrane differently. In this manner, the relatively smaller CO₂-permeation due to the highest operating temperature (e.g., 260 °C) can be expected, although the CO₂-adsorption diffusion in the presence of the highest water content in that section could be intensified. It should be also taken into account that water can escape the membrane pores in such a high-temperature range (e.g., 260 °C) and thereby its facilitating impact on selective CO₂-permeation can be neutralized. On the top of this integrated reactor primarily, the relatively smaller methanol-content (e.g., less than 5%) is dehydrated to DME. Nevertheless, the intensity of reactions is distributed along the catalytic bed not only through the direct impact of temperature on the reaction rate and its indirect effect on the local membrane permeation but also through bulk-distributing the reactants along the bed.

Following is the review and summary of the background design concepts for the proposed integrated reactors: (a) distribution of reactant through membranes along the catalytic bed improves the local selective reaction (micro-scale), (b) cross-sectionally distribution of feed across the multi-tubular catalytic beds improves the thermal control of the system, and (c) sequentially splitting of feed along the multi-pass reaction chambers improves the thermal-reaction performance of the reactor (macro-scale). Another macro-scale engineering design parameter, GHSV can be also efficiently controlled close to its desired values from the reaction-kinetic point, along the reactor by designing the tube diameters in each tube-pass. In this manner, using smaller-diameter tubes in the later stages, results in decreasing the residence time needed to secure the selective catalytic conversion. Overall, tuning the micro and macro-scale reaction intensity along and across the reactor in such reactor configuration enables establishing the desired temperature profile and securing a high-selectivity while converting enough limiting reactant and securing the yield of the desired product.

Figure 6b represents a classic process flowsheet for the same application under 10 bar pressure simulated in Aspen Plus, in which separate reactions (in the range > 200 °C) and separation units (in the range > 200 °C) have been utilized for the production of DME (4.5–6.5 kg·h⁻¹) with 99.3% purity. The reactor is filled with the hybrid catalyst (3.5 kg) under the reaction conditions identified in Section 2.1 and its performance has been simulated using the kinetic model reported by Iliuta et al. [25]. The simulation specifications of this flowsheet along with the specifications and the simulated performance of distillation columns used for separation of light components, DME and methanol (as shown in Figure 6b) have been reported in Appendix A. The performance of the reactive distillation has been also simulated using the RATEFRAC module in Aspen Plus under comparable conditions.

Simulating the performance of reactive distillation containing a hybrid catalyst for CO₂-hydrogenation to DME showed that the energy consumption and DME production, in this case, have been improved compared to the conventional indirect DME production method involving a sequence of reactor and distillation units. In this manner, a 13% higher DME production rate could be secured.

The potential for process intensification using such integrated reactors can be highlighted by the number of unit operations that can be spared depending on the design and operation of the process. For instance, a mixture of methanol and DME can be taken to a partial condenser on the top of the integrated reactor, where non-condensable gases can be separated to be recycled back to the bottom of the reactor in the shell side, while methanol can be condensed and recycled back to the top of this membrane-assisted reactive distillation column as a reflux stream. Detailed information supporting these findings is available through the model-based analysis of the integrated process for CO₂-hydrogenation reported in Appendix A and elsewhere [28].

The design-operating concept of such integrated reactors can boost the separation and catalytic performance as well as the energy efficiency of CO₂-separation and conversion from different perspectives. From one perspective, high-energy efficiency, high temperature, and high-pressure in-situ CO₂ separation are materialized [16]. From another perspective, the exothermic nature of the catalytic hydrogenation also conceptually matches this design [8] and all together significantly reduces the required operating and capital costs of this process and improves the potential of such CO₂-hydrogenation reactor to be integrated with the upstream processes to process their generated CO₂ emission.

After reviewing the conformity of the design with the conceptual requirements of this application and performing the conceptual design and feasibility study, some practical aspects regarding the tailoring and fabrication of such integrated structures are also analyzed in the next section.

2.3.3. Practical Aspects in Fabricating and Operating the Integrated Systems in Large-Scale

In this section, the main practically relevant concerns for implementing these integrated reactors in industrial-scale operations are addressed.

The catalytic material in this research (CuO-ZnO/ZSM5) was chosen also with regard to its stability at high temperatures, and for its resilience under high water content, as discussed by Álvarez et al. [5]. The stability of CuO-ZnO catalyst synthesized through different methods has been extensively studied, even after being exposed to oxygen, showing a fast and complete recovery of activity [7]. Further details along with the specifications of the catalytic tests conducted at bench-scale fixed bed reactors can be found elsewhere [6,7].

On the catalyst scale, 3D-printed thermally-conductive catalytic structures could be preferably used to address the exothermicity of this catalytic system and thermal engineering its impact in these integrated reactors, in particular, in the membrane-assisted reactive distillation. For instance, the feasibility and efficiency of fabricating and utilizing 3D-printed hybrid catalysts for this application via micro-extrusion of the catalytic material paste through robocasting procedure have been demonstrated [31]. This enables establishing a tailored structure, which could provide the desired structural characteristics in macro-scale (e.g., surface and geometry) and micro-scale (e.g., porosity) affecting the heat and mass transfer efficiency; 3D-printed Titanium or Zirconium could be used to prepare the catalytic support for this application after being processed, for instance using the Plasma-Electrolytic-Oxidation technique [32]. This can improve the energy efficiency as well as the thermal-reaction controllability of the system.

On the membrane scale, the CMSM used in this research has 0.5 to 0.8 nm pores representing the adsorption diffusion mechanism. Water, hydrogen, carbon dioxide, carbon monoxide, methanol and DME, respectively, show the smallest to largest molecular sieving size components. Since water is the smallest and the most polar molecule in this system, it is selectively adsorbed on the hydrophilic groups on the surface of the pores and affects the permeation of other species. Competitive permeation study through membrane reactor for this system has been studied, with a focus on the impact of water removal [17,33,34]. Considering the range of reaction temperature in this application, in order to find a balance between the interactive impacts of water and carbon dioxide in the pores and their permeations, the pore size and the pore distribution of CMSM were tuned. In doing so, different sets of carbonization parameters, including carbonization temperature, environment and soaking time, can be applied. Modifying the surface characteristics of the membrane (e.g., via immobilizing it with functional groups) and balancing the partial pressure of the components in the permeate and retentate side of the membrane enabled establishing the desired level of CO₂ and water permeation as well as preventing the diffusion transfer of other gaseous species in line with the designed operating concept of this integrated membrane reactor. In fact, the membranes with specifically tailored structural-chemical characteristics required for each chamber-pass could be synthesized and utilized in this integrated system. Details of the synthesis of the membrane as well as a systematic permeation study for such a system and an experimental separation performance study are available elsewhere [16].

Feed-splitting enables increasing the effective reactants (H₂/CO₂) ratio and thereby the overall methanol and DME selectivity and yield. Such a multi-tubular multi-pass reactor with feed-splitting can be also designed and operated for even impermeable tubes, where the temperature inside the tubes is controlled using a proper media in the shell side.

Regarding the stability and life-time of the membrane material, CMSMs are known to be mechanically and chemically stable under operational pressures and temperatures of up to 150 bar and 500 °C, except in environments containing oxygen. Therefore, the potential of the membrane being exposed to oxygen also needs to be considered in the design phase of such an integrated reactor, which can eventually cause an unstable separation performance. Apart from this, the lifetime of the CMSMs for packed bad membrane reactor applications, where no significant attrition is expected, could be in the scale of years. The relatively easier fabrication and sealing with higher reproducibility, homogenous permeation profile (defect-free), higher chemical and swelling resistances as well as thermal-mechanical resistances of CMSMs under high temperature and pressure disturbances can address the serious challenges of current membrane technologies for being implemented in an industrial-scale plant for this application. In case the membrane has the characteristics that a significant amount of methanol can pass through it, a methanol-dehydration catalyst can be placed on the other side of the membrane to quickly convert it to DME.

Similar to many other integrated systems, the best catalytic and separation performance, which could be achieved in the configuration composed of a sequence of catalytic and separation units, might not be attainable using the investigated integrated systems in this research. In a reactive distillation, the operating windows for reaction and separation must overlap [35], posing a significant design challenge. Integrating membranes into an RD column can further constrict this operational window, potentially leading to less-than-ideal conditions for one or more of the components. Such constraints could cause issues like challenging separations, decelerated reaction rates, or underwhelming membrane performance if overlooked during the design phase. A system operating within such a narrow window inherently possesses less flexibility than a setup that involves multiple units. However, the advantages of such an integrated system manifested in the form of improved selective catalytic conversion compared to the sequence of unit operations indicated significant potential for the applied integration. In addition, a significant level of process intensification with positive implications on the overall operating and capital costs could be expected using these integrated systems [28].

3. Conclusions

This research offers a comprehensive multi-scale approach to track and tailor the impacts of the nano-/micro-/macro-scales phenomena on the performance of novel membrane-assisted catalytic reactors for converting carbon dioxide to DME. Specifically, the performance of a membrane reactor and a membrane-assisted reactive distillation were analyzed. In-situ removal of carbon dioxide and water along the catalytic bed was demonstrated to be capable of improving the thermal-reaction performance as well as the energy efficiency of the whole integrated process.

The tailored hybrid catalyst enables efficient shifting of the equilibrium-limitation of the consecutive catalytic hydrogenation and dehydration reactions by establishing the right intensity and dynamic of activation and conversion of CO₂ to methanol and dehydrating it to DME over the surface of the catalyst. Establishing an efficient hybrid catalyst, in which the active catalytic materials are accessible and homogeneously distributed, is crucial.

In the scale of the reactor and to secure a selective reactor performance, the local ratio of the reactants and the reaction intensity across the catalytic bed and along the reactor were fine-tuned. Moreover, as a distinguished potential of using CMSM for this application, it was demonstrated that CO₂ and H₂O not only do not compete to permeate through CMSM but also CO₂ permeation is facilitated in the presence of water. The high flux of selective CO₂ permeation, beyond the Robeson upper bound, was obtained using the synthesized membrane. The desired and undesired impacts of water condensation, respectively, in the membrane for facilitating CO₂-permeation and in the catalyst to minimize the reaction-hindering effect due to water condensation in the microporous zeolite, could be synchronized in the design of such a macro-scale integrated reactor.

In addition, the conceptual and detailed design of these integrated reactors addresses the need for securing the desired residence time and temperature profile at each catalytic section as well as ensuring the practicality of the construction in terms of feeding, sealing, etc. Ensuring the fast conversion of methanol to DME and removal of water improved the CO₂-conversion and selective DME production by shifting the reaction equilibrium, at the micro and meso-scales.

In the macro-scale process intensification, lowering the capital and operating costs via heat integration and loss-prevention as well as the high capability of being integrated with the upstream process to reduce the overall CO₂ emission were demonstrated to be feasible using the developed integrated system as a result of this research.

Concurrent analysis and design of different aspects of the membrane-assisted reactive distillation could address a) shifting the equilibrium limited CO₂-conversion at the scale of catalyst, reactor and the intensified process, b) improve the selectivity via tuning the distributed intensity of reactions and local and overall ration of the reactants across the catalytic structure, c) being able to separate and handle the components, including the reactants, main and side products such as water and carbon monoxide. These features identify the proposed design as a promising and fine-tailored one.