The Potential of Gas Switching Partial Oxidation Using Advanced Oxygen Carriers for Efficient H₂ Production with Inherent CO₂ Capture

Abstract

The hydrogen economy has received resurging interest in recent years, as more countries commit to net-zero CO₂ emissions around the mid-century. “Blue” hydrogen from natural gas with CO₂ capture and storage (CCS) is one promising sustainable hydrogen supply option. Although conventional CO₂ capture imposes a large energy penalty, advanced process concepts using the chemical looping principle can produce blue hydrogen at efficiencies even exceeding the conventional steam methane reforming (SMR) process without CCS. One such configuration is gas switching reforming (GSR), which uses a Ni-based oxygen carrier material to catalyze the SMR reaction and efficiently supply the required process heat by combusting an off-gas fuel with integrated CO₂ capture. The present study investigates the potential of advanced La-Fe-based oxygen carrier materials to further increase this advantage using a gas switching partial oxidation (GSPOX) process. These materials can overcome the equilibrium limitations facing conventional catalytic SMR and achieve direct hydrogen production using a water-splitting reaction. Results showed that the GSPOX process can achieve mild efficiency improvements relative to GSR in the range of 0.6–4.1%-points, with the upper bound only achievable by large power and H₂ co-production plants employing a highly efficient power cycle. These performance gains and the avoidance of toxicity challenges posed by Ni-based oxygen carriers create a solid case for the further development of these advanced materials. If successful, results from this work indicate that GSPOX blue hydrogen plants can outperform an SMR benchmark with conventional CO₂ capture by more than 10%-points, both in terms of efficiency and CO₂ avoidance.

1. Introduction

Due to the growing concern of global warming caused by greenhouse gases (GHG) as highlighted in the IPCC Special Report [1], large-scale hydrogen production processes avoiding carbonaceous emissions have received increased attention. H₂ has a great potential to become the most attractive energy vector for decarbonizing several industry sectors, long-distance transport and gas heating networks, as discussed in the recent IEA report [2]. A comprehensive summary of current H₂ generation sources with present and projected end use to the year 2050 is presented by Voldsund et al. [3]. A detailed review of the different technological pathways for H₂ production has been carried out by Nazir et al. [4], while a study of the different applications and challenges of H₂ for the decarbonization of industry and energy sectors is provided in Nazir et al. [5].

Using an excess of renewable energy to produce H₂ from water electrolysis (“green hydrogen”) is a promising pathway avoiding CO₂ emissions, as suggested by Parra et al. [6]. Schnuelle et al. carried out a techno-economic assessment of an electrolyzer with direct renewable electricity supply achieving a minimum cost of 4.33 €/kg of H₂ when powered with the onshore wind [7]. However, 95% of the total H₂ produced is currently originated from fossil fuels and the traditional pathway through natural gas reforming with steam (SMR), is responsible for approximately 50% of the total H₂ production. Furthermore, it will remain more cost-effective when integrating CCS technologies (“blue hydrogen”) in several world regions [2]. For example, natural gas exporting regions like Norway can achieve an approximately 1 €/kg lower-levelized cost of H₂ with CCS than electrolysis [8]. In addition, this comparison does not account for the cost of intermittent hydrogen production from electrolyzers powered by wind and solar power, which adds substantial further costs, increasing the gap between blue and green hydrogen even for natural gas importers like Germany, as assessed by Cloete et al. [9]. If significant fossil fuel-based electricity generation is still on the grid, certification of green hydrogen production can be challenging. Some studies indicate the low development prospects of green hydrogen in the short term [10]. Nonetheless, blue hydrogen faces a relatively high energy penalty compared to the unabated process that results in large CO₂ avoidance costs (100 €/ton), well above the actual emissions tax, as reported by Spallina et al. [11]. Cormos et al. revealed that SMR with shifted syngas CO₂ absorption is the most cost-effective pathway for CO₂ abatement in H₂ production processes from natural gas with currently deployable technology [12]. Pre-combustion capture costs can be potentially reduced as membrane technology advances (with simpler fabrication methods from cheaper raw materials yielding enhanced performances, as studied by Miricioiu et al. [13,14]) by simplifying the separation processes (PSA and MDEA unit). Alternatively, the use of cheap adsorbents from waste-derived materials can increase the competitiveness of post-combustion capture [15], avoiding costly absorption systems.

On the other hand, Chemical Looping Combustion (CLC) first proposed by Ishida et al. offers the possibility of avoiding this large energy penalty by combusting a carbonaceous fuel with inherent CO₂ capture [16]. This is achieved by splitting the combustion reaction into two steps. CLC exposes a metallic oxygen carrier to oxidizing (e.g., air) and reducing (e.g., natural gas) streams in two separate reactors to avoid the mixing of the combustion products (CO₂/H₂O) with N₂. The oxygen carrier is hydrodynamically transported from one reactor to another, typically operating under fluidization conditions for efficient heat transfer. Alternative to full combustion of natural gas, Chemical Looping Reforming (CLR), developed by Rydén et al., obtains an H₂-rich syngas stream in the reduction outlet [17]. Subsequent water gas shift and CO₂ removal with MDEA is still required, as modeled by Nazir et al. [18]. Membrane-assisted chemical looping reforming (MA-CLR) has been proposed by Spallina et al. as an effective way to practically eliminate the energy penalty of CO₂ capture [11], but still requires several more development steps to de-risk and commercialize the complex system consisting of a membrane operating in a fluidized environment. Furthermore, a critical factor that hinders chemical looping applications is the slow scale-up under pressurized conditions required for efficient integration in power and chemical processes [19].

To overcome this challenge and avoid the operational difficulty of solids transportation between reactors, the gas switching (GS) technology has been suggested as a promising solution by Zaabout et al. [20]. Syngas production through the gas switching concept was experimentally demonstrated by Wassie et al. [21]. The oxygen carrier remains in the same reactor vessel while it is exposed alternatively to reducing, oxidizing and reforming steps by switching the feed and product streams by means of high-temperature valve mechanism. Gas Switching Reforming (GSR) appears as a highly attractive technological candidate for a step-change in H₂ production with CCS as shown by Nazir et al. [22], in which the reforming (H₂ production) and reduction (heat supply) steps are intrinsically decoupled, allowing for effective integration with the pressure swing adsorption unit (PSA), as suggested by Spallina et al. for packed bed dynamically operated reactors [23]. Very attractive prospects in terms of efficiency and emissions avoidance, with compelling economic assessment results carried out by Nazir et al. suggest that further research and development should be conducted [24].

The oxygen carrier is central to every chemical looping process, and its activity, selectivity and long-term stability are important to ensure high overall process efficiency. Many different types of oxygen carriers have been investigated experimentally for chemical looping reforming, most of them based on iron oxide. Zaabout et al. investigated iron oxide supported on gamma–alumina particles and found that the addition of nickel was beneficial to achieve a higher conversion of the methane (which was used as a surrogate for natural gas) than pure iron oxide at 800 °C [25,26]. Similar observations were made by Kang et al. when adding small amounts (~1 wt.%) of nickel or yttrium to the oxygen carrier [27,28]. In contrast, the partial substitution of iron oxide with ceria did not improve the performance significantly [26]. Other oxygen carrier compositions have been reviewed recently [29,30], and from these studies, it is apparent that mixed oxides composed of several redox-active oxides can have synergistic effects with regard to their activity and syngas selectivity. An important problem to overcome in the chemical looping-based reforming scheme is the deposition of carbon on the surface of the oxygen carrier due to the cracking reaction of methane that is catalyzed by metallic iron. Carbon deposits could potentially be combusted in the oxidation step of the process (see below) and therefore lower the CO₂ capture efficiency. Perovskite-based oxygen carriers (e.g., LaFeO₃ with various dopants partially substituting the cations La and Fe) have shown promise in reducing carbon deposition while possessing a high selectivity towards syngas and a high reactivity due to their high oxygen mobility [30]. Our recent experimental studies have demonstrated that some perovskite-based oxygen carrier formulations (e.g., La_0.85Sr_0.15FeO₃ or La_0.85Sr_0.15Fe_0.95Al_0.05O₃) possess excellent long-term stabilities in reactor systems of different scales.

With the purpose of delving further into the potential of gas switching chemical looping integration for H₂ production, this study aims to evaluate one of these experimentally proven oxygen carriers in an integrated plant model. Instead of requiring a catalytically-active component (e.g., Nickel), this material possesses thermodynamic properties that enable the partial oxidation of methane (or natural gas) to syngas at near 100% selectivity, as reported by Donat et al. [31].

The Gas Switching Partial Oxidation (GSPOX) Process Concept

Relative to the GSR process, GSPOX enables the reforming step to be split into separate partial oxidation and water splitting steps. In addition, the Lanthanum-based oxidation carrier presents the opportunity to overcome the equilibrium limitations faced by catalytic natural gas reforming at higher pressures. The process configurations presented in this study aim to quantify the potential efficiency benefits of these oxygen carrier features.

Figure 1 shows the 3-step GSR operation and the two four-step GSPOX operating modes investigated. The first GSPOX configuration is designed to produce an H₂-rich syngas with minimal steam addition for maximum H₂ production efficiencies. In this case, the key GSPOX benefit is avoidance of equilibrium limitations in the reforming step, eliminating the need for high S/C ratios to achieve sufficient methane conversion at elevated pressures. The second GSPOX configuration targets combined H₂ and power production. In this case, the reforming step is split into partial oxidation and water splitting steps and the H₂ from the water-splitting step is used for low-carbon power production.

For the GSPOX H₂-mode simulations, the process flowsheet is similar to previous GSR studies [22]. A schematic of the GSPOX H₂-power concept is presented in Figure 2 The water-splitting step in the GSPOX process produces an H₂ stream with some N₂ and CO from undesired mixing between the different reactor steps. This stream does not meet H₂ export purity requirements, but it is well-suited for combustion with the hot depleted air stream from the oxidation step for highly efficient low-carbon power production. Meanwhile, the 2:1 H₂:CO stream from the partial oxidation step is sent to a PSA unit for separating out ultra-pure H₂ for export. The remaining PSA off-gas stream is sent to the reduction step where it is oxidized to yield a high-purity CO₂ stream after water condensation. This PSA off-gas stream contains enough heating value to supply heat for the endothermic reforming reaction and the pre-heating of surplus air in the oxidation step for power production.

Relative to the previous GSR studies, the GSPOX concept developed in this study aims to evaluate:

• An alternative oxygen carrier for H₂ production plants with comparable efficiency levels, mitigating the safety concerns related to Nickel. Furthermore, the influence of operating pressure on the performance of GSR and GSPOX designs is assessed.
• A process design that eliminates the potentially disadvantageous feature of electricity consumption in H₂ production plants with CCS, by efficiently integrating a power cycle with the cluster, making effective use of the low-grade heat sources throughout the plant for additional electricity production, instead of LP steam export. The GSPOX concepts present a significant electricity output as an additional revenue source.
• A simplification of the plant section dedicated to H₂ production, by eliminating the water gas shift conversion and associated heat recovery units, substantially reducing the Steam to Carbon (S/C) plant requirements for natural gas conversion, and therefore reducing capital costs.
• From a modeling perspective, this study provides an integrated plant model between the dynamic clusters (GSR and GSPOX) and the stationary flowsheet simulation, avoiding cumbersome information exchange interfaces between platforms, making the process synthesis and evaluation more agile.

This new process topology will be benchmarked against conventional unabated SMR plant, SMR with MDEA CO₂ capture and the previously studied GSR-H₂ plant, under consistent modeling assumptions. The following section provides a detailed description of the dynamic model development and its integration in the process flowsheet. The narrative for the GS-based H₂ and electricity production plants as well as the different benchmark processes is provided. Subsequently, the key performance indicators of the plants are presented based on pre-defined reference technologies, and a brief description of the modeling approach for the essential technological components is given. Results are shown attending to the efficiency definitions, energy breakdown across the plant and CO₂ emissions performance. Finally, the key results are summarized, and the main conclusions of this study are discussed.

2. Methodology

In this section, the modeling procedure and main assumptions taken for the representation of the transient GSR and GSPOX clusters are presented. Then, the stationary process flowsheet modeling techniques are detailed, and the coupling between interfaces is described.

2.1. Reactor Modelling

Two oxygen carriers were employed in the reactor models built in Scilab, which incorporated an in-house thermodynamic database (Patitug) for property calculations of the gas phase. The first reactor principle consists of gas switching reforming (GSR) using Nickel as an oxygen carrier, inspired in past studies from some of the authors [22]. The second reactor principle incorporates the novel oxygen carrier which achieves partial oxidation of methane, referred to henceforward as gas switching partial oxidation (GSPOX). For the solid property estimation, suitable data sources were employed [32,33]. The oxygen carrier used in the GSPOX schemes had the chemical composition La_0.85Sr_0.15Fe_0.95Al_0.05O₃. Generally, thermodynamic data for such compositions are scarce, therefore the specific heat capacity was estimated based on literature data for species that represent the components in the fully oxidized and fully reduced states of the oxygen carrier, respectively. For the oxidized oxygen carrier, the specific heat capacity reported for LaFeO₃ was used [34]. The composition of the reduced oxygen carrier consisted of 50.2 wt.% La₂O₃, 23.0 wt.%. Fe and 26.8 wt.% La_xSr_2-xFe_yAl_1-yO₄ [31]; the mixed oxide phase La_xSr_2-xFe_yAl_1-yO₄ was assumed to possess the same physical properties as LaFeO₃. The specific heat capacity of the individual components was thus averaged using published data and interpolated across the operating temperature range of the reactor [32].

The specific enthalpy of reduced or oxidized species at each reactor temperature was estimated based on the experimental value for the partial oxidation of methane with the oxidized species, and equal to 222.25 kJ/mol of CH₄ at 900 °C [31], assuming an enthalpy origin of the reduced species at 298.15 K. Further property details of this oxygen carrier and cluster modeling assumptions are provided in

Table A2. GS cluster modeling assumptions.

Item	Value	Unit
Maximum reactor temperature	1100	°C
Pressure drop	50	kPa
Thermal mass volume fraction	0.25	-
GSPOX Oxygen Carrier
XO molecular weight	233.6	g/mol
X molecular weight	212.0	g/mol
Density (constant)	2000	kg/m3
Inert materiel content (Al2O3)	20–40	wt.%

in Appendix A.

Each reactor has a 1 m diameter and an aspect ratio of 3, which resulted in fluidization velocities of approximately 0.6–0.8 m/s for operating temperatures between 900–1100 °C. A pressure drop of 0.5 bar in each reactor was assumed in all cases. With a similar approach as in [22], additional thermal mass in the form of steel rods was introduced in the reactor to mitigate the temperature variation across the GSR and GSPOX cycles. The steel rods were assumed to occupy 25% of the reactor volume for all simulations. The specific heat of the carbon steel rods was assumed from [35], constant in the range of reactor operating temperatures. The carbon steel rods did not actively participate in the redox reactions (such that gas would be converted or produced), but only acted as a medium for storing heat.

Both for GSR and GSPOX reactors, the dynamic molar species Equation (1) and energy Equation (2) balances are solved to determine the reactor profile (temperature and compositions) and outlet flows, assuming the fluidized bed behaves as a continuous stirred tank reactor (CSTR), in virtue of the high degree of mixing attained in industrial practice:

$$\frac{d{n}_k}{dt}=F_{in}{y}_{in,k}+F_{out}{y}_{out,k}+{\displaystyle \sum }_{r=1}^R{\upsilon }_{r,k}{\xi }_r$$

(1)

$$\frac{dT}{dt}=\frac{-F_{in}{\sum }_k{y}_{in,k}{\int }_{T_{in}}^T{c}_{p,k}dT+{\sum }_{r=1}^R{\xi }_r(-∆H_{r,T})}{{\sum }_k{n}_k{c}_{p,k}}.$$

(2)

The CSTR model assumes that both chemical and thermal equilibrium is reached. Thermal equilibrium is a realistic assumption in fluidized bed reactors and experimental work has shown that chemical equilibrium is also a reasonable approach for the highly reactive Ni-based oxygen carrier used in GSR [21]. The validity of this assumption is unknown for the GSPOX reactor, but, since the present study evaluates the potential of an ideal GSPOX oxygen carrier, it is assumed that this oxygen carrier also reacts fast enough to reach chemical equilibrium [31].

The energy balance is solved first and, making use of the ideal gas relationship, the instantaneous outlet flow $F_{out}$ Equation (3) is determined to subsequently solve the species balance. In this way, Equations (1) and (2) are decoupled. A stiff ordinary differential equation (ode) solver is used for this purpose.

$$F_{out}=F_{in}+{\displaystyle \sum }_R^{k\in gases}{\upsilon }_{k,r}{\xi }_r+\frac{P{V}_{gas}}{R{T}^2}\frac{dT}{dt}.$$

(3)

The instantaneous outlet flows, compositions and temperatures from each step in the GSR and GSPOX cycles are averaged to deliver constant values to the stationary simulation. This is done on the assumption that a sufficiently large cluster of GS reactors will be used so that multiple reactors will operate in each reactor step at any given time. For instance, in the GSR model for H₂ production the cluster was designed with 7 reactors operating in the reforming step, 3 in oxidation and 3 in reduction at any given instant. When the outlet streams from different reactors in each step were mixed, minimal transient fluctuations in flowrate, temperature and composition were achieved, in this way ensuring a stable operation of the downstream process units.

The results of the simulation led to mass balance relative errors in the order of around 0.1%. Due to the high degree of heat integration between the clusters and the stationary plant, besides the averaging of instantaneous outlets taking place, the process simulations were converged with a relative energy balance error below 0.5%.

2.1.1. Gas Switching Reforming

Analogously to previous work [22], full fuel conversion in the reduction was assumed and, to prevent kinetic limitations taking place at the end of this step, a minimum fraction (5%) of the oxygen carrier in its oxidized state is specified. The carrier is then fully reduced during the subsequent long reforming step. After the reforming, the degree of oxygen carrier oxidation with the air stream was set to 30%, while the total airflow rate was selected to avoid any accumulation of O₂ in the reactor. The heating value routed to the reduction inlet is controlled to ensure the autothermal operation by tuning the overall steam to carbon ratio at the reforming feed.

The oxygen carrier in the GSR is based on Nickel, with the formulation taken from Abad et al. [36]. As illustrated in Figure 1, the reactor operates in a cycle consisting of reforming (REF), oxidation (OX) and reduction (RED). The kinetic equations that determine the rate of reaction in each step are expressed in

Table 1. Reactions and kinetic expression for GSR.

Step	Reaction	Kinetics
REF	$C{H}_4+H_{2}O\leftrightarrow CO+3H_2$	$R_{SMR}=\frac{1}{\tau }\left(p_{C{H}_4}{p}_{H_{2}O}-\frac{p_{CO}{p}_{H_2}^3}{K_{SMR}}\right)$
	$CO+H_{2}O\leftrightarrow C{O}_2+H_2$	$R_{WGS}=\frac{1}{\tau }\left(p_{CO}{p}_{H_{2}O}-\frac{p_{C{O}_2}{p}_{H_{2}O}}{K_{WGS}}\right)$
OX	$O_2+2Ni\to 2NiO$	$R_{O_2}=\frac{1}{\tau }{n}_{O_2}{n}_{Ni}$
RED	$C{H}_4+4NiO\to 4Ni+C{O}_2+2H_{2}O$	$R_{C{H}_4}=\frac{1}{\tau }{n}_{C{H}_4}{n}_{NiO}$
	$H_2+NiO\to Ni+H_{2}O$	$R_{H_2}=\frac{1}{\tau }{n}_{H_2}{n}_{NiO}$
	$CO+NiO\to Ni+C{O}_2$	$R_{CO}=\frac{1}{\tau }{n}_{CO}{n}_{NiO}$

below. A value of 0.01 for the relaxation time constant τ was selected to reach full conversion (or equilibrium) of the reactants in the reactor volume.

An illustration of the GSR reactor profile (product stream composition and temperature) corresponding to one of the H₂ plant simulations is presented in Figure 3.

The equilibrium constants for the WGS and SMR reactions are a function of thermodynamic properties and dependent on temperature [37], as represented by Equations (4) and (5).

$$K_{SMR}=1.2\times {10}^{13}{e}^{\left(\frac{-\mathrm{223,080}}{RT}\right)}$$

(4)

$$K_{WGS}=0.0177e^{\left(\frac{\mathrm{36,850}}{RT}\right)}.$$

(5)

2.1.2. Gas Switching Partial Oxidation

The modeling of the La-Fe-based oxygen carrier for GSPOX is based on the promising experimental results obtained in [31]. Good performance has also been achieved in a fluidized bed, increasing confidence in the application to gas switching reactors [38]. The original oxygen carrier possesses an oxygen storage capacity of ~9 wt.% and thermodynamic properties such that ~96.5% of its redox-active lattice oxygen is selective for the partial oxidation of methane to syngas (in the reduction step) and ~95% of steam (or CO₂) can be converted to H₂ (or CO), in the oxidation step. The aim is to develop and assess the application of this oxygen carrier in an H₂ production plant, benchmarking it against the former GSR plant and the reference technology: SMR process. For this purpose, certain ideal performance assumptions have been taken, as further detailed below.

The chemical reactions and kinetic expressions considered for each step are presented in

Table 2. Reactions and kinetic expressions for GSPOX.

Step	Reaction	Kinetics
POX and REF/WS	$C{H}_4+0.74X{O}_{1.35}\to 0.74X+CO+2H_2$	$R_{C{H}_4}=\frac{1}{\tau }{n}_{C{H}_4}{n}_{XO}$
	$0.74X+C{O}_2\leftrightarrow 0.74X{O}_{1.35}+CO$	$R_{C{O}_2}=\frac{1}{\tau }\left(y_{C{O}_2}-y_{C{O}_2,eq}\right)n_{gas}{n}_X$
	$0.74X+H_{2}O\leftrightarrow 0.74X{O}_{1.35}+H_2$	$R_{H_{2}O}=\frac{1}{\tau }\left(y_{H_{2}O}-y_{H_{2}O,eq}\right)n_{gas}{n}_X$
OX	$1.48X+O_2\to 1.48X{O}_{1.35}$	$R_{O_2}=\frac{1}{\tau }{n}_{O_2}{n}_X$
RED	$0.74X{O}_{1.35}+H_2\to 0.74X+H_{2}O$	$R_{H_2}=\frac{1}{\tau }{n}_{H_2}{n}_{XO}$
	$0.74X{O}_{1.35}+CO\to 0.74X+C{O}_2$	$R_{CO}=\frac{1}{\tau }{n}_{CO}{n}_{XO}$
	$C{H}_4+2.96X{O}_{1.35}\to 2.96X+C{O}_2+2H_{2}O$	$R_{C{H}_4}=\frac{1}{\tau }{n}_{C{H}_4}{n}_{XO}$

for the GSPOX reactor, where the fully oxidized oxygen carrier is denoted by XO, while the reduced oxygen carrier is referred to as X.

The proposed cycle includes a two-step reduction. In the first step, the oxygen carrier shows full selectivity to CO₂ and H₂O and fully oxidizes all fuel gases fed to the reactor. In the second step, the oxygen carrier has 95% selectivity to H₂ and CO, by partially oxidizing CH₄ and reducing H₂O or CO₂ fed to the reactor. It is assumed that the transition point between these two steps (i.e., to what degree the oxygen carrier is reduced when the selectivity of the oxygen carrier changes from total to partial oxidation) can be tuned according to the requirements of the process, e.g., by decreasing the ratio of La:Sr in the oxygen carrier [31], or by adding other combustion oxygen carriers such as ilmenite to the formulation [39]. The modeling results reveal that approximately half of the lattice oxygen that is consumed during reduction (case dependent) must fully oxidize the fuel to achieve adequate step lengths. Since the overall degree of oxygen carrier reduction was limited in all cases below 70%, and due to the higher resistance of these materials to carbon deposition [40], it was neglected in the kinetic model.

In the partial oxidation step (POX) two moles of H₂ are present per mole of CO produced. Unlike in the reforming step in the GSR, a large flow of steam is not required to ensure high methane conversion. An oxygen carrier conversion of 30% in addition to that for the reduction step (where full combustion of the fuel takes place) is specified for the POX step. For GSPOX H₂ production plants, in the subsequent reforming step (REF), a methane-steam feed with a molar ratio of 1 is fed to the reactor, thereby attaining more H₂ in the outlet (ratio of 3:1). During the reforming step, a negligible carrier conversion is observed due to the equimolar feed of methane and water that cause forward and backward reactions to take place at identical rates. The following oxidation step is carried out similarly to the GSR reactor, with only sufficient air to fully re-oxidize the carrier (no O₂ slippage). Finally, in the reduction step (RED), the PSA off-gas fuel is completely combusted to CO₂ and H₂O.

In Figure 4, a case corresponding to the GSPOX model designed for H₂ production is presented, whereas in Figure 5, a case for H₂-power co-production is shown. In Figure 5, air is introduced in the oxidation step in excess, to allow subsequent extra firing to reach high gas turbine (GT) combustor outlet temperatures (COT) to maximize efficiency. The low-carbon fuel for the extra firing is generated by operating the water splitting (WS) step with a steam sweep (as opposed to the equimolar methane-steam feed to the reforming step), in order to obtain a relatively pure H₂ output. This latter strategy is not advisable for H₂ production only, since the H₂ purity of the product stream is approximately 90%, and therefore it would still require a cooldown and purification sequence, neglecting potential efficiency gains whilst increasing operational complexity. It was therefore decided to carry out the reforming as described earlier and mix the process stream outlets from the partial oxidation and reforming steps in Figure 4 before introducing them to the downstream WGS reactors.

For the H₂-power coproduction reactor profile (Figure 5), the maximum reactor temperature is attained in the oxidation step at the point where the carrier becomes fully oxidized, after which the reactor cools down and the O₂ concertation begins to increase. The mass flow rate of air is manipulated from the stationary simulation to reach autothermal operation (energy conservation across the cycle), while the initial reactor temperature was iteratively varied to reach the specified maximum temperature. On the other hand, the degree of oxygen carrier reoxidation during the WS step was manipulated in order to produce sufficient H₂ to reach the nominal COT value upon combustion with the O₂ depleted air stream from the oxidation step outlet.

The mole fraction at equilibrium of CO₂ and H₂O for the kinetic expressions in Table 2 is determined as a function of the specified selectivity to CO and H₂ (assumed conservatively to be 95%) and the actual mole fraction of each gas species in the reactor at any given time, as shown in Equations (6) and (7). Where $s_k$ indicates the selectivity towards a specific component $k$. Reactions taking place in reduction and oxidation steps proceed to completion, while only methane in the partial oxidation step is fully converted.

$$\frac{y_{H_2}}{y_{H_2}+y_{H_{2}O,eq}}=s_{H_2}$$

(6)

$$\frac{y_{co}}{y_{co}+y_{c{o}_2,eq}}=s_{co}.$$

(7)

2.2. Process Modelling

The averaged outlets from each cluster step are introduced in the stationary process simulator by means of a CAPE-OPEN unit operation. Similarly, the feed streams to the reactors are taken from the plant model. The process flowsheets were developed with consistent modeling assumptions to similar studies and are further detailed in

Table A1. H2 plants modeling assumptions.

Item	Value	Unit
Fired tubular reformer (FTR)
Reforming temperature	890	°C
Furnace gases temperature	1010	°C
Excess air	15	%
Air blower polytropic efficiency	80	%
Heat exchange
Temperature approach gas-gas	30	°C
Temperature approach steam-gas	20	°C
Temperature approach liquid-gas	10	°C
Temperature approach liquid-liquid	5	°C
Air side pressure drop	5	kPa
NG/syngas/reduction gases side pressure drop	20	kPa
Turbomachinery
Air compressor polytropic efficiency	92.5	%
Steam turbine efficiency (SMR)	80.0	%
N2 turbine polytropic efficiency	87.0	%
CO2 compression stage polytropic efficiency	80.0	%
H2 compression booster/stage polytropic efficiency	80.0	%
Intercooler pressure drop	20	kPa
Other assumptions
H2 delivery pressure	150	bar
CO2 delivery pressure	150	bar
Process streams cooled to	25	°C
LP steam conditions (sat.)	6	bar

in Appendix A. For the H₂-power GSPOX plants, two representative gas turbine technologies employed for base-load power production [41,42] were calibrated using natural gas fuel. The equipment efficiencies and combustor performance with representative blade cooling technology from each turbomachine were specified in the GT units of the H₂-power plants. The models were built in Unisim Design R451 from Honeywell, using the Peng Robinson equation of state for thermodynamic property calculations, and ASME steam tables for water and steam streams.

2.2.1. SMR H₂ Plant without CO₂ Capture

The SMR plant has been extensively reviewed in earlier research. The present work introduces an SMR concept based on the technological features described in [11,22,43], in order to provide a benchmark process developed with the same simulation tool as the novel H₂-power plant concepts. A natural gas heat input of 129 MW (LHV) is considered as a calculation basis for all the models developed.

A process diagram of the SMR plant is shown in Figure 6. The natural gas feed is introduced in a preheater with the reformer exhaust gas. After desulphurization and pre-reforming to eliminate C2+ species, the natural gas is heated to 620 °C fed to the fired tube reformer (FTR) with a steam-to-carbon ratio (S/C) of 2.7. Approximately 87.0% of methane conversion is achieved in the outcoming syngas at 890 °C, a limiting temperature to avoid overheating and failure of the catalyst-packed tubes exposed to the flame. After cooling in a series of heat exchangers producing HP steam at 92 bar and 485 °C, the CO present is converted CO₂ by steam shift, producing more H₂. Once the H₂-rich syngas is cooled down to ambient temperature, it is routed to the PSA unit where the H₂ product is retrieved, while the off-gas is used as fuel in the burner in the FTR furnace.

In order to provide sufficient heat for the endothermic reforming reaction to take place, a small portion of the original fuel after the desulphurizer is also combusted with air (introduced with a 15% O₂ excess) in the FTR. The exhaust gases at low pressure and high temperature (1010 °C) are used to preheat the incoming feed streams and to generate more HP superheated steam. A small air blower is needed to overcome the pressure losses of the heat recovery network.

2.2.2. SMR H₂ Plant with MDEA Absorption CO₂ Capture

For completeness of the benchmarking plants, a CO₂ capture SMR plant was developed using an absorption unit with MDEA prior to the PSA, in order to remove 95% of the CO₂ present in the shifted syngas stream. A process diagram is provided in Figure 7. The MDEA section was modeled with a dedicated thermodynamic model (DBR Amine Package: Kent-Eisenberg) facilitated by Honeywell to accurately represent the chemical absorption process. The use of a chemical solvent is suitable for the intermediate partial pressures of CO₂ at the inlet stream to the absorption unit. Furthermore, MDEA presents a low regeneration heat requirement amongst amine-based solvents: The Unisim model with partial flash solvent regeneration predicted a value of 1.07 MJ_th per kg of CO₂, very closely matching the values reported in [44]. The integration of conventional CO₂ capture through amine absorption has several implications: the S/C ratio was increased to 4,as in [11,43], to achieve maximum methane conversion in the tubular reformer (92.5%), while a second, low-temperature shift (LTS) unit is added to fully convert the CO present in the syngas produced. Furthermore, a portion of the purified H₂ stream after the MDEA unit must be employed in the burner to satisfy the energy balance without increasing CO₂ emissions. Lastly, since the steam demand is large, all the steam produced in the heat recovery systems is either employed to reach the required S/C ratio in the reformer or in the MDEA stripper column for solvent regeneration. A dedicated LP boiler after the LTS unit is used to fulfill the LP steam demand. The resulting decreased steam turbine output and the extra auxiliary consumption imposed by the 5-stage intercooled CO₂ compressor and downstream supercritical CO₂ pump results in negative net electricity production for this plant. Both H₂ and CO₂ products are delivered at a pressure of 150 bar. Given that the GSR and GSPOX H₂ plants are assumed to import a substantial amount of electricity, the SMR with MDEA CO₂ capture model was designed to rely as well on electricity imports to enable a fair comparison, and only sufficient steam is raised to satisfy the S/C requirements. Therefore, the steam turbine (ST) output will decrease considerably, but less pure H₂ product will be combusted in the reformer, ultimately leading to higher H₂ efficiencies and lower penalties than in the case where the steam turbine is operated to satisfy the plant energy demand.

2.2.3. GSR H₂ Plant

The GSR H₂ plant is modeled following the guidelines provided in Nazir et al. [22]. Due to the higher temperatures reached (up to 1100 °C at the end of the reduction and beginning of the reforming stage) in the flameless combustion attained with the chemical looping principle, a lower S/C ratio is needed to achieve high methane conversion relative to the SMR models. Precisely this parameter is tuned to obtain sufficient heating value in the PSA off-gas, which is routed to the reduction step after compression and heating to satisfy the cluster energy balance. Somewhat higher H₂ efficiencies are obtained in the present work relative to the original work [22] as a result of improved heat integration. Thus, a higher fraction of the fuel heating value is preserved as H₂ and not combusted to maintain reactor operating temperatures. The elements of which the GSR-H₂ plant consists are practically the same as the GSPOX-H₂ plant described in the next section and illustrated in Figure 8. The base case reactor operating pressure is the same as in the SMR plants (32.7 bar) and it was assumed that the air and PSA-off-gas stream pressurization was carried out in a single adiabatic compressor stage. The PSA separation performance was represented using the correlation derived by Nazir [22], assuming 100% purity of the H₂ product stream. As represented in Equation (A1) in Appendix A, the hydrogen recovery is directly related to the pressure relation between the adsorption and desorption stages, which is linked to the off-gas compressor auxiliary power for the plants integrating GS clusters. H₂ recoveries beyond 80% are achieved for all the plants investigated.

Some considerations will be made regarding the heat integration for the benchmark and GS-based plants. For the fired tubular reforming SMR plants, on the combustion gases side, the low pressure of the stack gases will result in a small heat transfer coefficient, which leads to large temperature differences with respect to the incoming natural gas feed, achieving a relatively low preheating temperature. On the syngas side, the steam generation network is arranged in such a way as to avoid steam superheating with the high-temperature reformer outlet syngas, preventing an environment that would cause exchanger tube failure due to metal dusting [45]. On the other hand, the temperature approaches considered in the GS plants allow for a considerably higher heat recovery performance. The pressurized conditions of the streams found in the GSR and GSPOX clusters maximize the heat transfer coefficient relative to the reference plants, enabling closer temperature approaches. Special precautions may be required for the exchangers which preheat the mixed steam and gas streams to the GS clusters with the high-temperature syngas product, despite the fact that the steam pressures (25–40 bar) are notably below the design values of the conventional SMR processes (>100 bar). The PSA off-gas recuperator with the GS reduction outlet is an item that will also require advanced heat exchanger design [46]. The potential to achieve higher preheating temperatures inherent to the GS cluster operation represents an important feature to accomplish the high H₂ efficiency conversions attained in the concepts integrating this technology.

2.2.4. GSPOX H₂ Plant

The process configuration for the GSPOX-H₂ concept is depicted in Figure 8 It is assumed that the pre-reformer catalyst can cope with a decreased steam to carbon ratio, and it was specified to achieve a 99.9% equilibrium conversion of C2+ species. It is noteworthy to mention that the GSPOX oxygen carrier is likely to be able to handle higher molecular weight hydrocarbons, but in order to avoid greater modeling complexity in the Scilab code (additional kinetic expressions and increased n° of differential equations), the C2+ components were converted prior to the GSPOX. The outlets of the partial oxidation and reforming steps are mixed and cooled down in Hex1, Hex2 and Hex3, prior to a water gas shift conversion in two stages. Intermediate Pressure (IP) steam is generated in boilers with the heat of reaction; it is employed in the pre-reformer and the reforming steps of the GSPOX and also added to the feed of the High-Temperature Shift (HTS) bed. After cooling to ambient temperature, the shifted syngas is routed to the PSA unit, where a pure H₂ stream is compressed to 150 bar and delivered as a product. The PSA off-gas at low pressure is recompressed and sent to the reduction step feed. By tuning the steam input to the HTS bed the CO conversion is controlled. In this way, the amount of heating value remaining in the PSA off-gas is manipulated to operate the GSPOX reactors autothermally. The reduction gases outlet from the GSPOX are used to preheat the incoming PSA off-gas, and further cooled producing more IP steam. The off-gas temperature inlet to the reduction step was manipulated to satisfy the steam demand balance, in such a way that all process models are self-sufficient.

The oxidation section presents a simple power unit, in which the compressor delivers air to the oxidation step of the GSPOX and a low turbine inlet temperature (TIT) of around 1000 °C retrieves some useful electricity for the plant. No blade cooling was considered for this case. Due to the large pressure ratio of the turbine and low TIT, the turbine outlet temperature (TOT) is relatively low. An IP evaporator and LP evaporator produce steam with the remaining heat of the air stream. The IP steam is used within the plant, but the LP steam must be exported as a product.

2.2.5. GSPOX H₂-Power Plant

As described earlier, the GSPOX H₂-power plant can simultaneously produce H₂ and electricity, by operating the reforming step of the cluster with a water sweep. A similar option is possible with the GSR, withdrawing some H₂ product for extra firing, but this hydrogen would need to be separated from the syngas stream via larger WGS and PSA units instead of being directly produced via water splitting as in the GSPOX process. In addition, if the oxidation step is operated with excess air, the reactor cool-down taking place after carrier full oxidation will reduce the subsequent reforming step temperature, thereby increasing the steam requirement for comparable methane conversion.

The basic layout of this concept is represented in Figure 9. The essential differences with respect to the H₂ only production plant are (1) the use of a water sweep in the reforming step to produce a semi-pure H₂ stream for extra firing, (2) the avoidance of a WGS step after the partial oxidation and heat recovery sections, and (3) a more complex power cycle in the oxidation section. By removing the WGS units, a lower amount of the heating value originally present in the natural gas is withdrawn as H₂ in the PSA unit. Consequently, a larger flow rate of air is heated in the GSPOX reactor, producing more electricity upon expansion. From an efficiency perspective, removal of the WGS reactors also avoids degrading a significant fraction of the fuel heating value to relatively low-grade heat via the exothermic WGS reaction. Since the PSA off-gas is now much richer in volatile components (CO, H₂) relative to the H₂ plant (with a large presence of CO₂ and H₂O from the shift), the PSA-off-gas compressor was split into two intercooled stages, to reduce the safety concerns of an excessive adiabatic temperature rise.

The power cycle indicated in Figure 9 is specifically integrated with the rest of the plant. It integrates hot water/LP steam produced from low-temperature heat recovery units of the reduction and partial oxidation streams to effectively use them for extra electricity generation. Moreover, the cycle is designed to satisfy the internal steam demand, delivering IP steam to the reforming step and pre-reformer as required. In this way, the potential disadvantage in the case of no suitable end-use for LP steam production in the H₂-only plants is entirely avoided. Given the uncertainty of the scale of the devised concepts using GS technology, several power cycles may prove competitive, depending on different factors such as plant location, natural gas heat input, price of electricity and H₂ etc. However, for this study, only a combined power cycle is considered to evaluate the potential benefits of the H₂-power GSPOX configuration relative to the H₂-only configuration.

In such a combined power cycle, the gas turbine integrated with the GSPOX oxidation and water splitting steps delivers a high-temperature exhaust gas stream to a steam cycle, to produce additional electricity. This configuration is depicted in Figure 10, where only the oxidation and water splitting sections of the GSPOX cluster unit are shown and process equipment within the dotted lines refer to the units appearing in the block “power cycle” in Figure 9.

The gas turbine elements associated with the GSPOX H₂-power plants are representative of two industrial turbomachines employed for power production in combined cycles. It is noted that the power systems integrated into these processes are not specifically the actual GTs, but rather incorporate several defining technological features to represent that technology level. Therefore, the case incorporating the technological features of the H-class turbines is henceforward denoted as “Advanced”, while the case with the F-class characteristics is referred to as “Current”, to potentially avoid confusion with the actual machines. A simplified gas turbine model as depicted in Figure A1 in Appendix A using reference values for combustor outlet temperature (COT), turbine inlet temperature (TIT) and turbine outlet temperature (TOT) is calibrated, specifying the compression and expansion polytropic efficiencies to reach the known reference open cycle efficiency.

Table A3. Gas turbine modeling assumptions.

Reference Values
Item	F-Class (Current)	H-Class (Advanced)	Unit
COT	1440	1648	°C
TIT	1360	1550	°C
TOT	603	640.1	°C
Pressure ratio	18.1	23.7	-
Simple cycle efficiency	39.0	43.0	%
NG-fired rated power	280.0	520.0	MW
Calibrated Parameters *
Item	F-class (Current)	H-class (Advanced)	Unit
Compressor polytropic efficiency	92.5	94.0	%
Expander polytropic efficiency	87.0	88.0	%
Stator cooling constant bs	0.0616	0.0534	-
Rotor cooling constant br	0.0674	0.0791	-
Other Parameters
Item	F-class (Current)	H-class (Advanced)	Unit
Air filter pressure loss	1	1	%
Combustor pressure drop	3	3	%
Exhaust pressure loss (simple cycle)	1	1	kPa
Electromechanical efficiency	98.6	98.6	%
pressure drop constant k	0.07	0.07	-
Max. stator blade temperature	850	900	°C
Max. rotor blade temperature	825	875	°C

Air at 15 °C and 60% relative humidity is used for the calibration *.

in Appendix A shows the results of the calibration for the two types of turbomachines employed, as well as the reference values for each GT open cycle.

A lumped blade cooling flow model where the coolant is taken from compressor discharge and injected in the hot gas path at two points is developed, similar to what is proposed in [47]. Coolant addition after the combustor represents the refrigeration requirements of the 1st stator (non-chargeable cooling flow), while the rotor and subsequent rows (chargeable cooling flow) is added at 1/3 of the expansion path. For calibration, each coolant flow is varied to reach the specific TIT and TOT of the GT. In this way, the $b$ values for stator and rotor shown in Equations (A2) and (A3) respectively in Appendix A, given a certain maximum blade temperature, are determined for each GT. When integrating an analogous technological cycle in the H₂-power plants, the cooling flows are predicted with a reasonable level of accuracy, assuming constant $b$ values and open loop convective cooling as heat transfer model. Given that the COT values are the same as those of the reference machines, the cooling flow fractions do not deviate largely from the NG fired model. A pressure drop after the mixing of coolant and gas streams is considered to represent efficiency losses associated with this mixing, proportional to the coolant to hot gas ratio $\phi$ as suggested in [48], and shown in Equation (A4) in the Appendix A. A more detailed derivation of the heat exchange effects between coolant and hot gas path in the blade as well as different cooling mechanisms is described by Sanjay et al. [49]. A precise prediction of coolant flows with injection at several stages of the expansion path is provided by Chiesa and Macchi for deployable natural gas combined cycles [50]. For the conceptual design of H₂-power processes at a low technology readiness level (TLR) and the current state of development, requiring tailor-made turbomachinery components to reach attractive efficiencies, the simplified approach is considered reasonable. Indeed, the reference gas turbines are taken from combined cycle applications [41,51], whose heat rate is substantially above the natural gas heat input to the plants presented in this study.

The air compressor is therefore operated at the reference pressure ratio of each specific GT. The H₂ rich stream produced in the GSPOX water splitting stage is cooled down in a recuperator and then compressed in a booster and reheated. This is done to allow for an overpressure of 10 bar for the H₂-rich fuel stream prior to extra firing in the combustor chamber. The fuel must be delivered at sufficient pressure to the combustor by means of multiple fuel injectors to achieve a turbulent mixing between the rich H₂ fuel and hot depleted air from the GSPOX, leading to a swirling flow, which limits the flame temperature (to reduce NOx formation to acceptable limits), while avoiding the risk of flashback. An embodiment of the concept assumed here has been presented and experimentally tested in [52]. Small fluctuations in H₂ flow during valve switch in the WS stage that would affect combustion performance can be mitigated by pulsating water/steam addition at the recuperator inlet, although this was not accounted for in the plant simulation as the transient model delivers averaged values. Due to the larger number of reactors in the oxidation stage, the dimensionless flow rate variations of the blended oxidation outlet resulted in values below 1%, ensuring safe turbine operation [53].

Finally, the heat recovery steam generator unit (HSRG) of the steam cycle is the typical efficient design with three pressure levels and intermediate reheat previously modeled by the authors [54]. Further modeling assumptions for this section are shown in

Table A4. Steam cycle modelling assumptions.

Item	Value	Units
Condenser pressure	0.04	bar
Auxiliaries for heat rejection	0.008	kW/kWth
Water pumps isentropic efficiency	80	%
LP/IP Pinch point	10	°C
LP/IP Approach point	9	°C
LP ∆P/P eco+eva	25	%
IP ∆P/P eco+eva	15	%
∆P/P superheaters	8	%
HP pinch (once through)	9	°C
Exhaust air side pressure drop	3	kPa
Pressure level in drum/eva (HP/IP/LP)	185/43/6	bar
LP superheat	300	°C
LP Stage isentropic efficiency	87.7	%
IP Stage isentropic efficiency	92.0	%
HP Stage isentropic efficiency	90.3	%
Electromechanical Efficiency	98.3	%
Maximum steam temperature	565/600	°C

in Appendix A.

2.3. Plant Performance Indicators

To assess the performance in terms of efficiency and emissions of the different H₂-power production processes, several indicators are defined in this section. The H₂-power production plant can be envisaged as represented in Figure 11, attending to the products and by-products derived from it.

The H₂ efficiency ${\eta }_{H_2}$ in Equation (8) is determined as the relation between the heating value output of the H₂ stream and the heating value of the natural gas feed. However, this plant metric falls short to evaluate the electricity and steam demand or production that may take place in each of the plants. To fully account for these items, the equivalent natural gas feed ${\dot{m}}_{NG,eq}$ is calculated by subtracting (or adding) to the original feed the corresponding amount that would be required to generate the same quantity of steam or electricity as the current plant is exporting or importing, given a predefined benchmark efficiency for such production, as represented in Equation (9). In this way, a homogenous or equivalent efficiency ${\eta }_{H_2,eq}$ is defined in Equation (10). In previous works [22], it was shown that GSR H₂ concepts are reliant on electricity imports. On the other hand, it is frequent that substantial LP steam is produced as well. On some occasions, this steam can be utilized by other processes, and is accounted for in the equivalent efficiency calculation. However, it may be the case that there is no useful application for this steam, therefore the equivalent efficiency is recalculated accounting only for the electricity by-product (or import thereof), and denoted by a superscript in Equations (11) and (12).

$${\eta }_{H_2}=\frac{{\dot{m}}_{H_2}LH{V}_{H_2}}{{\dot{m}}_{NG}LH{V}_{NG}}$$

(8)

$${\dot{m}}_{NG,eq}=\left({\dot{m}}_{NG}LH{V}_{NG}-\frac{{\dot{W}}_{el}}{{\eta }_{el}}-\frac{{\dot{Q}}_{th}}{{\eta }_{th}}\right)/LH{V}_{NG }$$

(9)

$${\eta }_{H_2,eq}=\frac{{\dot{m}}_{H_2}LH{V}_{H_2}}{{\dot{m}}_{NG,eq}LH{V}_{NG}}$$

(10)

$$\dot{m}{’}_{NG,eq}=\left({\dot{m}}_{NG}LH{V}_{NG}-\frac{{\dot{W}}_e}{{\eta }_e}\right)/LH{V}_{NG }$$

(11)

$$\eta {’}_{H_2,eq}=\frac{{\dot{m}}_{H_2}LH{V}_{H_2}}{\dot{m}{’}_{NG,eq}LH{V}_{NG}}.$$

(12)

In terms of CO₂ emissions performance, a similar approach is followed. The specific emissions $E_{C{O}_2}$ calculated by Equation (13) are given in g of CO₂ per MJ of H₂ product. Alongside this, the capture ratio $CC$ is the relation of the mass flow rate of CO₂ that is captured divided by the CO₂ intensity of the fuel $E_{NG}$ as shown in Equation (16). When considering equivalent emissions $E_{C{O}_2,eq}$, the corresponding intensity of emissions avoided, derived from an alternative benchmark process for steam $E_{th}$ and electricity production $E_{el}$, are accounted for and presented in Equation (14). The CO₂ avoidance is consequently defined by Equations (17) and (18) when steam is or is not considered as a useful product, respectively. Finally, the Specific Primary Energy Consumption of CO₂ Avoided ($SPECCA$) of every concept with CO₂ capture is calculated using the equivalent efficiency and emissions of the unabated SMR process as reference plant, as shown in Equation (19). This parameter gives an indication of the original fuel heating value that must be invested due to the integration of a CO₂ capture technology in the plant.

$$E_{C{O}_2}=\frac{{\dot{m}}_{C{O}_2,emit.}}{{\dot{m}}_{H_2}LH{V}_{H_2}}$$

(13)

$$E_{C{O}_2,eq}=\frac{{\dot{m}}_{C{O}_{2,}emit.}-E_{th}{\dot{Q}}_{th}-E_{el}{\dot{W}}_{el}}{{\dot{m}}_{H_2}LH{V}_{H_2}}$$

(14)

$$E{’}_{C{O}_2,eq}=\frac{{\dot{m}}_{C{O}_2,emit.}-E_{el}{\dot{W}}_{el}}{{\dot{m}}_{H_2}LH{V}_{H_2}}$$

(15)

$$CC=\frac{{\dot{m}}_{C{O}_2, capt.}}{{\dot{m}}_{NG}LH{V}_{NG}{E}_{NG}}$$

(16)

$$CA=\frac{{\dot{m}}_{C{O}_2,capt.}}{E_{NG}{\dot{m}}_{NG}LH{V}_{NG}-E_{th}{\dot{Q}}_{th}-E_{el}{\dot{W}}_{el}}$$

(17)

$${CA}^{\prime }=\frac{{\dot{m}}_{C{O}_2, capt.}}{E_{NG}{\dot{m}}_{NG}LH{V}_{NG}-E_{el}{\dot{W}}_{el}}$$

(18)

$$SPECCA=\frac{\frac{1}{{\eta }_{H_2,eq}}-\frac{1}{{\eta }_{H_2,eq ref}}}{E_{C{O}_2,eq,ref}-E_{C{O}_2,eq}}$$

(19)

$$SPECC{A}^{\prime }=\frac{\frac{1}{{{\eta }^{\prime }}_{H_2,eq}}-\frac{1}{{{\eta }^{\prime }}_{H_2,eq ref}}}{{E^{\prime }}_{C{O}_2,eq,ref}-{E^{\prime }}_{C{O}_2,eq}}.$$

(20)

The benchmark for unabated power production with and without CO₂ capture is taken from [42], i.e., advanced combined cycle configurations using H-class GTs. When steam production is considered as a valuable output, a conventional boiler and an associated emissions intensity is taken into account, as detailed in

Table A5. Efficiency and emissions intensity for natural gas and reference steam and electricity plants.

Plant Product/Feed	Efficiency (%)	CO2 Emissions Intensity (kgCO2/MJ)
Electricity (Unabated)	62.1	0.0923
Electricity (with CO2 capture)	54.0	0.0106
Steam	90.0	0.0633
Natural gas	-	0.0568

in Appendix A, in order to determine equivalent natural gas flow and emissions.

3. Results

In this section, the process simulation results for each plant design are shown, attending to the key performance indicators defined in Section 2.3. Results of the SMR benchmark plants, the GSR and GSPOX designed for H₂ production only and the GSPOX hybrid concept with H₂ and electricity co-production, are provided, using the unabated NGCC efficiency and CO₂ intensity as reference electricity production plant. The effect of CO₂ capture integration in the reference electricity plant is presented in a subsequent section. A detailed summary of the results from each process model is shown in the Plant Results Summary in the Appendix A.4.

3.1. Plant Model Results

The efficiency results and CO₂ capture performance of the benchmark SMR plants without CO₂ abatement and with CO₂ capture employing MDEA absorption, as well as the GSR and GSPOX H₂ production plants and the GSPOX H₂-power co-production schemes, are shown in Figure 12. Efficiency results are indicated attending to the definitions presented in Section 2.3. It can be seen that when steam exports are neglected in the calculation, the equivalent efficiency reduction is greatest for the unabated SMR plant, while it remains unchanged for the GSPOX H₂-power plants (since there are no steam exports).

The equivalent efficiency penalty for CO₂ capture with MDEA absorption amounts to approximately 7.9%-points (when accounting for steam exports). As discussed, this penalty originates from the fact that a fraction of the H₂ produced must be invested to provide heat to the reformer. Since the operation is carried out with a high S/C ratio of 4, the steam turbine low-pressure stage output is comparatively reduced. Furthermore, LP steam must be sent to the MDEA stripper reboiler for solvent regeneration, neglecting the efficiency benefit of steam export attained in the unabated SMR plant. When the energy from exported steam is neglected as a useful by-product, the energy penalty of the MDEA process reduces to only 4.5%-points: the LP steam produced is fully utilized for solvent regeneration in the CO₂ stripper reboiler. Simulation results in this work yield a significantly higher hydrogen efficiency for the SMR with MDEA capture process relative to what has been presented by Spallina et al. [11]. Reformer inlet and outlet conditions are identical, yet apparently, the fuel demand to satisfy the reformer balance is notably smaller, yielding a higher H₂ output. Equivalent efficiency on the other hand is close to the value reported by Martínez et al. [43], albeit the fact that in the present work no LP steam exports are considered for this model and the H₂ is compressed to 150 bar. Given the higher steam turbine output in these works, it is concluded that in these earlier works a larger portion of the final H₂ product is combusted to generate more steam for the turbine in the reformer heat recovery network in order to minimize electricity imports to the plant, therefore attaining a comparatively lower amount of H₂ product. Since the GS-based plants are modeled assuming substantial electricity imports, the SMR model with MDEA CO₂ capture is designed to generate only sufficient steam for the reformer and stripper column demands.

Regarding the GS-based plants, as predicted by Nazir et al. [22], the integration of GS clusters for H₂ production results in a substantial increase (in this work up to 11.6%-points) of the H₂ efficiency relative to conventional SMR. As discussed earlier, this is a consequence of the higher reforming temperature that shifts the SMR reaction to the products side (consequently reducing the S/C ratio), as well as a more effective heat recovery network, minimizing the heat losses in the plant. On the other hand, for the GS plants with electricity imports (cases GSR-H₂ and GSPOX-H₂), similar results are obtained for the plants operating at a reactor pressure of 32.7 bar. Approximately the same amount of steam must be raised for the GSPOX prior to the WGS unit, leading to a similar H₂ production (around 0.5%-points higher for the GSPOX).

A sensitivity study regarding the cluster operating pressure is carried out and shown in Figure 13, to assess the effect on the efficiency of the reduced equilibrium conversion resulting from higher pressures in the GSR process. In this sensitivity study, it was assumed, in order not to distort the efficiency trends, that all turbomachinery components (air and PSA off-gas compressors) operate in a single adiabatic stage.

Mild benefits are reached with the GSPOX configuration: The S/C clearly declines for higher pressures, as it is not required to convert as much CO in the shift units due to the higher recovery of H₂ in the PSA, leading to better heat recuperation of the streams fed to the cluster and increased H₂ efficiency. More low-grade heat is converted to steam, resulting in a consistent improvement of the equivalent efficiency, as shown in Figure 13. For the GSR, higher pressures increase the required S/C to reach the desired methane conversion in the cluster, and consequently, the H₂ efficiency is curtailed. It is concluded that if the turbomachinery can reach the specified pressures cost-effectively (or alternatively adding intercooled stages), the GSPOX carrier will only result in small benefits relative to the GSR. However, this advantage can be expected to continue increasing for even higher pressures, although practical challenges will constrain the maximum achievable system pressure. If, for instance, higher pressure ratios are attained by means of intercooled compression, the lower air temperature entering the oxidation step will result in lower H₂ efficiencies relative to the single adiabatic compression mode of operation because more fuel is combusted to reach the desired reactor temperature. Simultaneously, higher pressures will reduce the number of H₂ and CO₂ compression stages required to reach delivery pressures, reducing the capital costs of these items. Thus, this parameter is closely interlinked with the remaining plant elements. However, it should also be noted that higher operating pressures in the reactor increase undesired mixing of the outlet streams from the different steps of the reactor cycle, reducing the CO₂ avoidance. In the range of pressures studied here, the CO₂ avoidance was reduced by 1.0 and 1.5%-points for GSR and GSPOX plants when increasing the pressure from 25 to 40 bar.

As the advantages of GSPOX over GSR for H₂ production are relatively small, the present assessment is extended by integrating a combined power cycle with the GSPOX to simultaneously produce H₂ and power. The key operational feature from the GSPOX that allows this with reduced energy penalty is the possibility of producing a relatively pure H₂ stream which can be used for extra firing if the air is introduced in the cluster with an excess of O₂. As expected, the Advanced GT combined cycle configuration yields the most attractive results, approximately 3.4%-points of additional equivalent efficiency points compared to the model employing Current GT technology. Still, the equivalent efficiency is around 1.0%-points below the unabated SMR benchmark. Given the pressure ratio of each GT, the H₂ recovery and consequently the H₂ efficiency vary slightly. Indeed, the H₂ recovery drops in the PSA unit at lower inlet pressure, falling from 85.6% to 82.0%, when shifting from Advanced to Current GT technology.

For the GSPOX power generating plants, given the fixed natural gas input of 129 MW (LHV basis), the net electricity production resulted in around 25 MW as shown in Figure 14 (~20% of the heat input) whereas the H₂ output (LHV basis) decreased by approximately 30% relative to the unabated SMR process. An effective operational handle to trade-off H₂ production for additional electricity would be to operate the PSA with a higher off-gas desorption pressure (such that the size of the associated compression unit is reduced), but this is limited to the extent that H₂ production in water splitting step of the GSPOX is sufficient to reach COT GT values in the extra firing chamber (unless the H₂ product is also employed for extra firing, which would imply additional penalties and process variations). Heat rejected comparatively increases relative to the SMR H₂ production plant without CO₂ capture due to the inherent thermodynamic inefficiencies related to a larger fuel degradation for power generation, reaching a degree of overall losses comparable to the SMR MDEA capture process. The effective use of low-grade heat across the plant in the H₂-power configurations for additional electricity production prevents the need to export any LP steam. Interestingly, both GSPOX H₂-power plants deliver a similar electricity output, underlying the higher efficiency of the Advanced GT power production scheme, which accomplishes the same amount of electricity production with a lower fraction of the heating value (given the fact that the H₂ production is somewhat higher due to the increased recovery in the PSA).

Scale is another important consideration in such combined power and hydrogen plants. The plant with Advanced GT technology will need to be very large (around 20× the heat input of the plants presented in this work), given the size of the heavy-duty gas turbines required to maximize firing temperature and efficiency. Such plants can only be considered for the longer-term future when a large hydrogen market has been established. Efficiencies representative of the Current GT technology can be achieved by much smaller industrial gas turbines [55], only requiring an increase in plant size of ~3×. Such a plant could supply an industrial cluster with power and hydrogen, avoiding any dependency on a broader integrated hydrogen market.

For the GSR and GSPOX H₂ plants, a small export of LP steam is needed in both cases. Furthermore, electricity imports of approximately 5% of the natural gas heat input are required. However, the heat rejected is reduced to around 16% of the heat input, in effect around 4%-points and 11%-points below the unabated SMR plant and SMR plant with MDEA CO₂ capture, respectively.

In terms of emissions, MDEA absorption integration achieves a CO₂ avoidance of approximately 85% (Figure 12), similar to earlier published works. CO₂ emissions are not curtailed entirely due to a small remaining fraction of CO, CH₄ and CO₂ present in the PSA off-gas used as fuel in the reformer. The CO₂ avoidance for the GSR and GSPOX is approximately 5%-points above the MDEA benchmark, due to the large capture rate (>95%) achieved with inherent CO₂ capture in the GS concepts. However, CO₂ avoidance is considerably less than the CO₂ capture rate due to the associated emissions resulting from the imported electricity generation. On the other hand, for the H₂-power co-generating concepts, since electricity is produced with inherent CO₂ capture, the CO₂ avoidance is largely above the capture ratio, reaching values above 140%.

In terms of equivalent specific emissions, reflected in Figure 15, the value decreases for the SMR plant without capture due to a small export of steam (and associated CO₂ avoidance of a boiler), whereas for the plant with MDEA capture it increases, due to the associated emissions of electricity imports. A similar trend is seen for the GSR and GSPOX H₂ production plants. For the concepts that produce additional electricity with CO₂ capture, the equivalent specific emissions become negative. Finally, regarding the SPECCA indexes, it can be seen that the value obtained with the SMR with MDEA CO₂ capture benchmark is substantially reduced when implementing GS-based CO₂ capture in the remaining models. In fact, when the efficiency of a specific plant is above the reference unabated SMR benchmark, the SPECCA index turns out to be negative. Indeed, this is a consequence of a negative energy penalty when integrating inherent CO₂ capture technology. If the SPECCA index is evaluated using equivalent efficiency values which do not account for the steam exports as a beneficial outcome from the plant (both for the specific plant with CO₂ capture and the reference plant), the SPECCA index becomes even smaller, as the associated CO₂ intensive steam generation in boilers is not considered.

Lastly, it should be noted that due to the substantial mixing occurring in the GS clusters, the purity of the CO₂ stream obtained after cooling and water knockout resulted in >90 mol% of CO₂. It was assumed that the larger proportion of impurities (primarily N₂) was acceptable for transport and storage conditions and that a higher compression duty resulting from these lower boiling point gases was considered acceptable. However, some sources recommend CO₂ purities >96 mol%, for which a purification unit may be required [56]. The overall power consumption will be minimally affected, with a marginal increase of specific emissions (due to purging in the purification unit), although the addition of cryogenic exchangers and vessels will lead to a higher overall plant cost [57].

3.2. Effect of Electricity Utilities

The critical parameter affecting the equivalent efficiency and emissions performance of the novel GSPOX plants presented in this study is the efficiency and CO₂ emissions intensity of the reference power plants. As explained at the beginning of the Results section, the results presented earlier are determined considering an efficient NGCC power plant using H-class GTs without CO₂ capture as a reference electricity generation plant (a completely unabated value chain). In this section, the previous results are compared to the case where electricity production with CO₂ capture is used as a reference plant, using the performance values shown in Table A5 in Appendix A.

It is noted that the calculations reflected in this section account for steam exports. Similar effects as shown in the previous section would be observed if steam exports are neglected. Given the negligible electricity production in the unabated SMR plant, sensitivity to the electricity utilities is not presented for this case.

Figure 16 shows the equivalent efficiency (left axis) and CO₂ avoidance (right axis) for each model. When considering electricity generation with CO₂ capture, the equivalent H₂ production efficiency of the plants that import electricity decreases, whereas the plants exporting electricity benefit from a less efficient reference.

On the other hand, when looking at the CO₂ avoidance, this parameter slightly increases for the importing electricity plants (H₂ production), when it is provided by a power plant with CO₂ capture. On the other hand, the CO₂ avoidance of the GSPOX CC plants decreases to around 100% because low-carbon electricity exports now displace electricity from a low-carbon reference plant. This is further emphasized in the equivalent CO₂ emissions values shown in Figure 17: For reference electricity generation with CO₂ capture, they slightly decrease for power importing plants, while in the H₂-power co-production schemes they increase from an originally negative value to slightly above zero.

The influence in the SPECCA index of these combined effects is also reflected in Figure 17. For the SMR plant with MDEA CO₂ capture, very small variations are observed, given the overall small electricity imports that result in very similar equivalent efficiencies. For the H₂ producing plants, electricity imports with CO₂ capture tend to increase the SPECCA index (mainly due to the lower equivalent H₂ efficiency). For the power-producing GSPOX plants, the large increase in equivalent H₂ efficiency (due to a less efficient reference electricity production with CO₂ capture) leads to negative SPECCA values in both cases.

4. Summary and Conclusions

This study investigated the potential of a novel La-Fe-based oxygen carrier with thermodynamic properties that enable the partial oxidation of natural gas with integrated CO₂ capture. Such an oxygen carrier can avoid the equilibrium limitations restricting conventional catalytic methane reforming and facilitates direct hydrogen production via the water-splitting reaction. In addition, it can alleviate safety concerns related to the toxicity of Ni-based reforming catalysts. The benefits facilitated by these features were quantified and compared to three benchmarks: conventional steam methane reforming (SMR), SMR with conventional CO₂ capture (pre-combustion MDEA absorption), and the gas switching reforming (GSR) process using a Ni-based oxygen carrier.

In terms of equivalent H₂ production efficiencies, it was shown that the novel gas switching partial oxidization (GSPOX) process achieved mild benefits relative to the GSR concept (of around 0.6%-points for the case with reactor operating pressure of 32.7 bar). Even though GSPOX could convert methane with a much lower steam requirement, a similar amount of steam as in GSR is needed in the subsequent water-gas shift reactors to maximize hydrogen production. Next, the possibility of designing the GSPOX plant for H₂ and power co-production was investigated. A semi-pure H₂ fuel stream can be produced for extra firing; to heat the depleted air stream from the GSPOX oxidation step outlet temperature to the gas turbine firing temperature for high-efficiency power production. Thus, the cluster was integrated into combined power cycles with two different firing temperatures, representative of two GT technology levels. In terms of equivalent H₂ efficiencies, the Current and Advanced GT configurations were 4.5 and 0.9%-points below the unabated SMR benchmark, respectively, but both plants achieved equivalent CO₂ avoidance of 140% because of considerable low-carbon electricity exports. When an NGCC plant with CO₂ capture was taken as reference for imported/exported electricity, CO₂ avoidance was approximately 100% with equivalent H₂ production efficiencies 0.7 and 4.8%-points higher than the SMR benchmark.

In comparison to the GSR technology with Ni-based oxygen carriers, the GSPOX process in H₂ mode shows mild improvements that increase with pressure. When operated at 40 bar, the GSPOX-H₂ plant achieves 1.5%-points higher equivalent H₂ production efficiency than GSR with similar CO₂ avoidance. The combined power and H₂ GSPOX configuration with the Advanced gas turbine can increase this benefit to 4.1%-points when exported electricity is assumed to displace NGCC power production with CO₂ capture. However, such an efficient power cycle requires a large gas turbine that will lead to a plant with a very high hydrogen output, which will only be feasible when a large hydrogen market is established. When a smaller and less efficient gas turbine is used, the GSPOX co-production plant offers no efficiency benefits relative to GSR. However, such a smaller plant will be a good option for independent clean power and hydrogen supply to an industrial cluster.

To conclude, the successful development of a GSPOX oxygen carrier can further increase the efficiency gains promised by GSR, producing clean “blue” hydrogen at a higher efficiency than carbon-intensive “grey” hydrogen from steam methane reforming. These gains increase when the system can be operated at very high pressures or market conditions allow for the construction of large, combined power and hydrogen plants. When factoring in the avoidance of toxicity concerns related to Ni-based oxygen carriers needed in GSR to catalyze the reforming reaction, further development of these novel oxygen carriers can be safely recommended.