Dynamic behaviour of integrated chemical looping process with pressure swing adsorption in small scale on-site H₂ and pure CO₂ production

Highlights

• The modelling and integration of dynamically operated chemical looping and PSA is presented.

• 8 beds PSA and 3 beds CLR are required for 130 Nm3/h H2 with > 98% CO2 capture.

• Chemical looping performance are not affected by the dynamic behavior of the PSA.

• Heat losses in chemical looping are mitigated by changing the heat management.

Abstract

The design of a fully integrated Chemical looping reforming (CLR), single adiabatic water gas shift reactor (WGSR) and Pressure swing adsorption (PSA) operated under dynamic conditions for small scale H₂ generation with inherent pure CO₂ production is carried out. The dynamically operated packed bed reactors taking part in the chemical looping process have been modelled, designed and simulated to operate with transient feeds from an integrated PSA unit used for the production of 130 Nm³/h of pure H₂ (99.9999% purity). As by-product, 51 Nm³/h of pure CO₂ (>98.8% purity) is also produced. A rapid cycle 8-bed configuration increases the H₂ recovery by 4% whilst reducing the tail gas buffer tank volume requirement by 44%. The effect of the PSA dynamic tail gas composition used as fuel for the CLR reduction reactor stage was found negligible regarding the continuity of the process and the performance of the plant, as it affected only the reduction outlet gas composition profile but had little effect on the reactor bed temperature profile. With respect to the design of the chemical looping reactor beds, an analysis has been performed on the effect of heat losses showing that at higher heat transfer coefficient (U = 5.0 W∙m⁻²∙K⁻¹) CH₄ conversion decreased significantly (≈90% compared to adiabatic operation), therefore a different strategy was implemented. The overall study demonstrates the process design feasibility for producing blue H₂ or renewable H₂ from methane/bio-methane in decentralised and modular units.

Introduction

The route to full scale green hydrogen production using water electrolysis powered by renewable energy has attracted a lot of attention from research, industry and policy makers [1] and it represents the long term solution to enable clean H₂ at different scales. Despite the technology being under development currently, H₂ from electrolysers represent only 0.1% of the global production [2].

Today, H₂ manufacture mostly comes from steam reforming of natural gas (50% of the global production [3]). The predominant driving force for this production route is process economics at relevant scale. Hydrogen production with integrated carbon capture, utilisation and storage (CCUS) also referred to as blue H₂, is a valid mid-long term alternative to avoid greenhouse gas emissions to atmosphere and in the meantime develop the required infrastructures for H₂ distribution and utilisation as an energy carrier for chemicals production [4], fuel cells and ultimately heat decarbonisation [5]. Currently, only a few large demonstration projects have been proposed in the UK, Europe [6], [7], USA [8], Australia [9] and Japan [10] to establish the benchmark and assess the feasibility for large scale implementation. However, there is no existing infrastructure for supply of hydrogen in areas of small demand such as small industries, residential use or refuelling stations for transportation. Alternatively, the use of different carbon-based feedstock such as biogas, waste gaseous fuel from chemical processes, flare gases, typically available at different quantities and the possibility to inject H₂ in the natural gas pipelines [9] could unlock other opportunities for small scale H₂ production with effective waste recovery. Small-scale production with conventional steam reforming techniques is very costly and CO₂ capture is very expensive as well making the implementation not economically viable yet [11]. Among different options, inorganic membrane reactors have been proposed as a solution to convert methane or alcohol to H₂ for small scale with high technical and economic performance [12], [13], [14]. However marginal CO₂ reduction is obtained from improved reforming efficiency, unless entirely fuelled with renewable feedstock or including a CO₂ capture unit.

Chemical looping technology is an interesting alternative for power and H₂ production combined with CO₂ capture because it presents higher efficiency than conventional CO₂ capture technologies [15], [16]. Chemical looping combustion (CLC) is a standard chemical looping process to produce high temperature heat for power generation. It is comprised of two reactors, an air reactor (AR) and a fuel reactor (FR), where combustion is happening in two stages in the presence of a metal oxygen carrier (OC) [17], [18], [20]. In the FR, a fuel is fed to reduce the OC and the main products are CO₂ and H₂O, from which pure CO₂ can be easily captured with condensation, avoiding any extra downstream separation process. In the AR, air is used to oxidised the OC and the main products is O₂-depleted air (N₂, Ar and traces of CO₂, H₂O). The heat stream generated from the oxidation of the OC can be used as a thermal source for power generation. Based on the same principles as in the CLC, chemical looping can be deployed for hydrogen or syngas production with one of the most common applications being Chemical looping reforming (CLR). In CLR, heat generated from the redox reactions is used for the endothermic steam/dry reforming of CH₄ or other light hydrocarbons which are present in the natural gas (e.g. C₂H₆, C₃H₈, nC₄H₁₀) to produce syngas. A lot of research is currently focused on hydrogen production chemical looping technologies and a list of the existing pilot-scale plants is presented in Table 1.

The development of oxygen carrier materials such as the conventional metals (Ni, Fe, Cu, Mn, Co), perovskites and core-shell structures [21] is a key aspect to enable the technology. Some examples of recent advances in material development are high-Fe₃O₄ content magnetite with ultrasonic treatment to increase syngas production and selectivity [22], perovskite-based LaNiO₃ on montmorillonite support to minimize LaNiO₃ sintering and reduce carbon deposition [23], Ca₂Fe₂O₅ for high CH₄ conversion and syngas yield in moving-bed – fluidized-bed reactors [24] and titania supported iron (Fe₂O₃/Y_0.20Ti_0.15Zr_0.65O_1.90) for higher reactivity [25].

The concept of performing chemical looping reforming for the production of syngas in packed bed reactors (CLR-PB), was first introduced by Spallina et al. [33]. In CLR-PB, oxidation, reduction and reforming are performed in three different stages, presented in Fig. 1. In this concept, the solid material is inside the reactor and the gases are sequentially fed to the reactor by means of switching valves. Firstly, the bed gets oxidized with air and the temperature rises across the reacting bed, due to the highly exothermic reaction (e.g., ΔH^ο_298Κ = −479.4 kJ mol⁻¹ for Ni oxidation) with the main product being N₂. The second step is to reduce the bed with a low-grade fuel (off-gas) obtained from a downstream process such as the tail gas from a PSA unit (usually a low grade mixture of CO, H₂ and CO₂) which is converted into H₂O and CO₂. The reduction reactions are mostly heat neutral (e.g., ΔH^ο_298Κ = −43.26 kJ mol⁻¹_CO and ΔH^ο_298Κ = −2.13 kJ mol⁻¹_H2 for Ni reduction), therefore the solid temperature slightly changes and only part of the bed is cooled due to the heat front moving from the inlet across the reactor. The CO₂ product can be sent to storage, or alternatively used as a reactant for other processes such as urea manufacture [34]. At the start of the third stage, the reforming, most of the heat originating from the oxidation stage is still present inside the bed. In fact, in gas-solid reactions (reduction and oxidation) the reaction front velocity is much faster than the heat front velocity and therefore, while the solid is being converted, the heat generated is not vented off, resulting in a heat storage [35]. This heat is used to enable the endothermic (steam and dry) methane reforming. The high temperature conditions along with the frequent valve switching are the primary technical challenges for this process.

Another way of producing hydrogen with gas-solid reactions is sorption enhanced reforming. In this configuration, CaO is used to capture CO₂ generated during the reforming/water gas shift and converted into CaCO₃ which is later regenerated to CaCO₃ [36], [37], [38], [39]. The endothermic reaction of CaCO₃ regeneration, also called calcination, occurs at high temperature by means of an oxy-combustion or by adding Cu to CaO and using the reduction of CuO to Cu (exothermic) [40], [41], [42], [43], [44].

To convert the syngas produced by a CLR-PB unit into high-purity H₂, one or more water–gas-shift stages are used followed by removal of impurities such as N₂, CO, CO₂ and CH₄ [45], [46]. In terms of gas separation, PSA is known as a highly well-stablished and versatile technology for hydrogen purification due to its low energy consumption and high product purity. Hydrogen purities from 98% to 99.9999% are reported for the PSA process [47], [48], [49]. The performance of a PSA unit is often defined by product purity, recovery and adsorbent productivity (i.e., the amount of H₂ produced per unit mass or volume of adsorbent). In order to achieve high recovery and purity, different adsorbents have been proposed and commercialised and process parameters can be varied. Activated carbons and zeolites are commonly used in H₂ PSA units [48], [50], [51], [52] and typically adsorption columns are packed with at least two layers of adsorbent to obtain high purity hydrogen. The feed end is commonly packed with activated carbon to selectively adsorb CO₂ and hydrocarbons such as CH₄. This is complemented with a zeolite layer that has an enhanced capacity for capturing N₂ and CO [45], [52], [53]. Zeolites adsorb CO₂ and water vapour strongly and these components cannot be readily desorbed by decreasing the pressure. Therefore, CO₂ and water vapour must be prevented from reaching the zeolite layer by adequately sizing the activated carbon layer and potentially adding a short desiccant layer at the inlet of silica gel or activated alumina [54].

PSA units can be used to supply the full spectrum of capacities from small plants producing a few hundred Nm³/h to large-scale plants that produce>400,000 Nm³/h. On the biggest plants, PSA units are normally operated with 10 or more columns with three or more pressure equalisations stages resulting in higher recovery [55], [56], [57]. There is though a growing requirement for hydrogen production facilities in small H₂ demand industries. PSA units at this scale are normally 4–6 bed systems with one equalisation stage and this makes it challenging to obtain a high H₂ recovery. The recovery values are found to be at or below ca. 75%, mainly due to the lower operating pressures reducing adsorbent capacities and increasing the amount of H₂ needed for purging. They can be designed with more than one equalisation to improve the recovery loss, but the cost of a PSA unit increases with more pressure equalisation steps due to the additional vessels required [57], [58].

Both CLR-PB and PSA processes are pseudo-continuous meaning the outlet stream conditions vary with time, but the process is overall operated continuously. Since the PSA process is downstream of the syngas generation process, the variation of feed concentration, temperature and flowrate might change the performance of the PSA unit and subsequently the composition of the off-gas. This could subsequently affect the performance of the reduction step and hence the CLR-PB process. Therefore, the effect of temperature, feed composition and flowrate variation need to be addressed during the design of the PSA and CLR-PB units. Little research has been done on the effect of feed condition variations in these systems. Ahn et al. [59] studied the effects of feed composition on a layered bed H₂ PSA process, where they looked at three feed compositions based on different N₂ contents. Li et al. [60] also investigated the effect of 24 different feed compositions on purity and recovery of H₂. However, none of these studies looked at the real case, where the composition of gases in the feed varies dynamically over time. Additionally, these studies did not investigated the effect of the time dependent changed tail gas composition and flowrate as this stream is returned to the chemical looping system and results in further changes to the inlet feed to the H₂ PSA system.

In this work, we have simulated the performance of H₂ PSAs based on the gas coming out of a chemical looping reactor on the reforming step from a dynamically operated CLR-PB. The reformate syngas from the CLR-PB is fed to a PSA after one-stage of WGS at high temperature and the PSA off-gas is send back to the CLR-PB for the reduction step. Since both are dynamic processes and mutually dependent, the operation strategy and heat management are essential to identify the main challenges. This is more relevant for small scale on-site H₂ generation where the overall volume should be limited to reduce the cost and make the system compact. In all the cases, the feed to the H₂ PSA has been modelled using the actual dynamics in composition and flow coming out of a simulation of the CLR. In this study, a layered activated carbon/zeolite LiX process is considered for the H₂ PSA unit. Both CLR-PB and PSA processes have been modelled and simulated using an advanced mono-dimensional, non-isothermal model with integrated kinetics to account for the gas-solid and heterogeneous reactions (in case of CLR) and adsorption kinetics (in case of PSA).