Natural Gas Decarbonation with Adsorption

In the adsorption process (ADS), CO, (the adsorbate) travels from the bulk of the CO,-rich NG phase layering up. Through selective binding of CO, on the surface of a solid or condensed liquid phase (substrate), ADS separates CO, with minimal CH4 loss (Al-Mamoori et al., 2017). CO, ADS may occur through physisorption or chemisorption links. In the former, weak Van der Waals interactions occur while, in the latter, chemical reactions like covalent bonding promote interaction (D'Allesandro et al., 2010).

Adsorbents are classified as high temperature or low temperature materials (Al-Mamoori et al.,

2017). High temperature materials consist of hydrotalcites, alkaline-earth oxides (calcium and magnesium oxides), alkali silicates, zirconates and double salts and are chemisorbeuts. Low temperature materials consist of zeolites, carbon-based materials (e.g., carbon nanotubes, activated carbon surfaces, carbon nauofibers and graphene), molecular sieves and advanced materials, such as covalent organic frameworks (COFs). porous polymer networks (PPNs), metal-organic frameworks (MOFs). Low temperature materials for CO, ADS are mostly pliysisorbents. Low temperature CO, chemiosorbents are rare in literature and consist of adsorbent materials with supported amines (amine-functionalized materials), in a hybrid adsorption and absorption mechanism (Al-Mamoori et al., 2017).

ADS is a cyclic process; after the adsorption phase, the active binding sites available are saturated with CO, and separation is no longer possible, characterizing CO, breakthrough (Sema-Guerrero et al., 2010). To regenerate the adsorption capacity, CO, desoiption is conducted, and another adsorption cycle occurs. Adsorbent regeneration may occur by; (i) pressure reduction—Pressure-Swing Adsorption (PSA); (ii) increasing temperature—Temperature Swing Adsorption (TSA); (iii) inducing an electrical field or passing an electrical current through the adsorbent—Electrical Swing Adsorption (ESA); or (iv) hybrid regeneration—Temperature and Pressure Swing Adsorption (TPSA). The efficiency of ADS depends on the regeneration procedure, process design, operational factors, and the adsorbent material. The main criteria an adsorbent must achieve are low cost, low regeneration requirements, high surface area, high CO, selectivity, fast kinetics and long-term stability (Al-Mamoori et al., 2017). The Technology Readiness Level (TRL) (Mankins, 2009) of ADS processes applied to NG upgrading is TRLrpSA< TRLESA< TRLjj,

<TRLPSA

Benefits and shortcomings

A great variety of adsorbent materials exist, such as zeolite 13X, HKUST-1 and Mg-MOF-74, making ADS one of the best technologies for trace CO, removal (D'Allesandro et al., 2010). High selectivity tradeoffs exist with total feed capacity and energy required for adsorbent regeneration—highly selective materials are costly, while commonly used adsorbents co-adsorb unwanted molecules, mainly water and hydrocarbons. To minimize co-adsorption, pre-treatment steps such as dehydration or desulfurization are required (Al-Mamoori et al., 2017). In parallel, surface treatments exist to improve selectivity— for example, the addition of supported functional groups (e.g., amines) and changing the adsorption interaction type.

The regeneration phase of the ADS consumes energy, e.g., PSA and TSA. PSA has attracted more attention than TSA for NG due to its simplicity, low cost, and lower energy requirements. TSA has higher heat requir ements and long downtimes between ADS cycles for system cooling, greatly increasing operating costs, but has better CO, recovery per cycle with less adsorbent losses. PSA is recommended for NG upgrading, while TSA suites flue gas CO, removal better (Al-Mamoori et al., 2017).

A shortcoming of ADS is that no adsorbent combined to CO, desorption (PSA/TSA/ESA) is yet cost effective (D'Allesandro et al., 2010). High costs are incurred due to adsorbent material and high heat consumption in regeneration. ADS requires multiple adsorption columns to phase adsorption and regeneration, assuring continuous CO, capture.

Recent focus

ADS is not yet in firll commercial scale for CO, applications in NG upgrading context, even though PSA processes are already available in other sectors (e.g., hydrogen purification) (D'Allesandro et al., 2010). The characteristic pressures and flow rates of the NG industry are much higher than in usual PSA applications (Grande et al., 2017; Riboldi and Bolland, 2017). Seeking cost effectiveness of CO, ADS, the bulk of academic research focus is in synthesizing and testing new or modified adsorbents to improve separation performance. Fu et al. (2017) modified MOF’s surface with amine groups to promote stronger binding interactions between CO, and the solid. The modified UiO-66/PEI had drastically reduced surface area but increased CO, loading to a maximum of 3.13 mol/kg of solid at 25 °C. Gil et al. (2017) used wastes from the aluminium industry to synthesize CoAl-MgAl and NiAl-hydrotalcite compounds, obtaining a loading of 5.26 mol CO,/kg of solid at 50 °C. Szczysniak et al. (2017) investigated hybrid graphene and MOF adsorbents and noticed that the new hybrid surface had a greatly increased binding area, leading to a loading of 9 mol CO,/kg of solid at 0 °C.

Although critical in the development of new materials, most of the present studies regarding new adsorbents ignore real application conditions (Al-Mamoori et al., 2017). ADS applications and its economic performance are reported through simulations of large-scale processes, employing laboratory- scale adsorption data gathered in the literature. Grande et al. (2017) simulated CO, separation from a gas composed of 83% methane and 10% CO, in a 12-tower adsorption process, concluding that huge methane losses would render ADS unfeasible for gas sweetening. Leperi et al. (2016) optimized a two-stage PSA process with varying degrees of dehydration, reporting zeolite 13X as the best performing material and concluding that the higher the water content in NG the greater the negative effect on ADS performance, with exponentially increasing costs.

New thermodynamic models to predict the behavior of ADS have been proposed, mostly of semi- empirical formulation fitted to experimental data. Elfving et al. (2017) modelled the equilibrium of CO, working capacity in PSA, TSA and TPSA, with good fit to experimental data. For TPSA, acceptable results occurred in high-purity CO, applications. Clark et al. (2013) simulated a fluidized bed ADS while Sema-Guerrero et al. (2010) used a semi-empirical model for amine-functionalized mesoporous silicas, obtaining equilibrium equations (Sema-Guerrero et al., 2010) and process kinetics (Sema-Guerrero etal., 2010).

Natural Gas Decarbonation with Gas Liquid Membrane Contactors

Gas Liquid Membrane Contactors (GLMC) combine absorption and membrane technologies (Norahim et al., 2017). In GLMC, a membrane provides an interfacial surface for gas-liquid contact, allowing CO, transfer from the NG into a solvent, where it is chemically absorbed (Figure 6), while avoiding phase dispersion (de Medeiros et al., 2013b).

CO, diffusion through the porous membrane interface is negligible when compared to the mass transfer rate into the absorbing solvent (de Medeiros et al., 2013b). However, if the membrane pores become filled by the liquid phase, the wetted membrane will show exponentially increased mass transfer resistance (Norahim et al., 2017). To prevent membrane wetting, the membrane should be highly hydrophobic, have high overall porosity (to minimize resistance) and exhibit high chemical resistance to withstand possible negative effects from the solvent. Therefore, non-porous (i.e., skin-dense) membranes commonly used in MP are not suitable for GLMC. Currently, the membranes developed for NG sweetening use polymeric materials due to their ease of synthesis in the desired porosity range and their hydrophobicity (Kang et al., 2017). The most common polymeric membrane materials in GLMC

Typical GLMC based process for upgrading CO,-nch NG

Figure 6. Typical GLMC based process for upgrading CO,-nch NG.

applications are polypropylene, polyethylene, polyethersulfone, polysulfone, polytetrafluoroethyleue and poly(vinylidene dufluoride) (P'DF). P'DF is the most promising material due to its high hydrophobicity, good chemical resistance, low cost and ease of fabrication (Norahim et al., 2017). GLMC is not a hybrid technology, as the membrane provides a physical barrier, not acting in the separation process.

Benefits and shortcomings

GLMC uses a chemical absorption mechanism spatially confined by a membrane, potentially allowing the system to share benefits of the individual technologies (MP and CA) while avoiding their major drawbacks. For instance, GLMC eliminates flooding and entrainment, common operational problems in absorption packing columns (Norahim et ah, 2017). The higher packing density and surface area of GLMC results in faster separation than CA or PA. Furthermore, GLMC allows for independent control of gas and liquid flows without gravitational effects. Lastly, its modularity provides easy scaling-up (and down) and flexibility' in a wide variety of scenarios (de Medeiros et ah, 2013b).

The technology is released from the permeability-selectivity tradeoff found in MP technologies. Since the separation occurs through the absorption of CO, into a chemically selective solvent, developments of new membranes turn the focus from selectivity to permeability', aiming at increased rates of CO, transfer. However, selectivity is not the main operational parameter; the pressure differential through the membrane can be considerably reduced compared to the MP process, minimizing power consumption. Although GLMC has lower energy intensity, solvent stripping occurs at low-pressure, posing an energy penalty due to compression of the captured CO, (de Medeiros et ah. 2013).

The main shortcoming in GLMC is membrane wetting over time. GLMC systems have a breakthrough pressure, i.e., the pressure at which liquid breaks the membrane barrier and diffuses into the membrane pore. The larger the pore, the lower the breakthrough pressure, so smaller pores are preferred. In the GLMC process life-span, physical aging of the membrane or effects like plasticization occur and the membrane barrier loses resistance, lowering the breakthrough pressure until membrane wetting occurs (Norahim et ah, 2017). Although not yet cost effective, a thin layer of a highly hydrophobic material on the liquid side of the membrane reduces wetting (Gugliuzza and Drioli, 2007).

Another relevant aspect is that mass transfer rate is directly related to solvent viscosity (de Medeiros et ah, 2013b). The higher the viscosity, the lower the mass transfer rate, highly jeopardizing the performance of GLMC. It is noteworthy that GLMC long-term cost effectiveness, when compared to other CO, capture technologies, has yet to be proven.

Recent focus

The main literature body in GLMC seeks enhancements in membrane materials and solvents. Rezaei et ah (2015) synthesized an МММ of PYDF and Cloisite 15A using distilled water as a solvent and, studying the effects of filler content, noting that surface area and permeation were greatly improved. Ghasem et ah

(2012) added o-xylene into polyethersulfone membranes to increase hydrophobicity, obtaining increased CO, recovery.

A rigorous thermodynamic model was proposed by de Medeiros et ah (2013b) in order to reproduce behaviors of industrial GLMC systems, such as (i) temperature increase along GLMC length due to exothermal reactive absorption of CO, into aqueous MEAMDEA solutions and (ii) equilibrium shifts in the solvent side when inert components (e.g., CH4) penetrate through the membrane causing stripping action in the solvent. The model considers vapor-liquid reactive equilibrium (RYLE) flow in the solvent side due to CO,-CH4 penetration and rigorous compressible flow modelling for both gas and (two-phase) solvent sides. Quek et ah (2018) proposed a model related to membrane pore size for high pressure applications and Zolfaghari et ah (2018) modelled membrane wetting-critical points in GLMC operation. Fougerit et ah (2017) studied trans-membrane pressure in wetting, specifically in breakthrough pressure of liquids, showing that membrane wetting is independent of trans-membrane pressure. Hashemifard et ah (2015) evaluated partial wetting of membrane pores, correlating pore size with wetting potential. An overlooked aspect is GLMC’s economic performance, contrasting and comparing it to more widespread technologies, and determining allowable wetting ratios.

 
Source
< Prev   CONTENTS   Source   Next >