E ﬃ cient Production of Clean Power and Hydrogen Through Synergistic Integration of Chemical Looping Combustion and Reforming

: Chemical looping combustion (CLC) technology generates power while capturing CO 2 inherently with no direct energy penalty. However, previous studies have shown signiﬁcant energy penalties due to low turbine inlet temperature (TIT) relative to a standard natural gas combined cycle plant. The low TIT is limited by the oxygen carrier material used in the CLC process. Therefore, in the current study, an additional combustor is included downstream of the CLC air reactor to raise the TIT. The e ﬃ cient production of clean hydrogen for ﬁring the added combustor is key to the success of this strategy. Therefore, the highly e ﬃ cient membrane-assisted chemical looping reforming (MA-CLR) technology was selected. Five di ﬀ erent integrations between CLC and MA-CLR were investigated, capitalizing on the steam in the CLC fuel reactor outlet stream to achieve highly e ﬃ cient reforming in MA-CLR. This integration reduced the energy penalty as low as 3.6%-points for power production only (case 2) and 1.9%-points for power and hydrogen co-production (case 4)—a large improvement over the 8%-point energy penalty typically imposed by post-combustion CO 2 capture or CLC without added ﬁring.


Introduction
Anthropogenic carbon dioxide (CO 2 ) emissions to the atmosphere have risen beyond 415 ppm, causing climate change [1]. Due to this, the Paris Climate Accord has vowed to limit the global temperature rise below 2 • C of the pre-industrial level [2]. The conventional power generation technologies such as natural gas power plants (NGCC) suffer a considerable energy penalty when integrated with a carbon capture facility. An amine-based capture system reduces plant efficiency by~8%-points [3] (all quoted efficiencies are LHV-based). Thus, the energy penalty is the primary cost driver for CO 2 capture technologies due to increased fuel costs and a greater amount of plant capital required to achieve a given electricity output. The increased fuel usage is also accompanied by increased emissions. Therefore, the development of novel energy conversion technologies with high CO 2 capture efficiency is essential.
Chemical looping combustion combined cycle power plants (CLC-CC) have the inherent capacity to capture CO 2 with a minimum penalty only required for compression [4]. The fuel and the oxidizer are treated separately using an oxygen carrier (OC), giving out a pure stream of CO 2 along with condensable water. The energy penalty is only for compressing the CO 2 to high pressures as well H2. Out of the five cases, two are developed based on the plant flexibility feature. The plant performances are estimated in terms of efficiencies such as electrical, H2 production, and CO2 capture and avoidance. Results will be compared with the NGCC combined cycle plant presented in our previous work [19].

Description of the Concept Working Principle
The principle behind the power and hydrogen production processes presented in this study is simply illustrated in Figure 1, where the red arrows indicate the two key points of integration between the CLC and CLR processes. As mentioned earlier, the reason for additional firing with hydrogen is to raise the temperature in the stream going to the gas turbine beyond that which is achievable in the CLC reactors, thus increasing power cycle efficiency. However, if SMR hydrogen production with conventional CO2 capture is used for this added firing, the gains from higher power cycle efficiency are canceled out completely by the low efficiency of hydrogen production with CO2 capture [19]. Therefore, the use of a highly efficient clean hydrogen production process is needed.
The primary energy penalty in hydrogen production with CLR-based processes is the provision of the steam required for NG reforming. When the produced hydrogen is combusted in a gas turbine, all the energy required to raise the steam in the hydrogen production process is lost because the condensation enthalpy of the steam resulting from hydrogen combustion cannot be converted to useful work. Nazir et al. [24] reported that this steam-related penalty accounts for about 5.8%-points of the 7.2%-point energy penalty of a combined cycle power plant fired by hydrogen from a gas switching reforming (GSR) process, illustrating the importance of this energy penalty. GSR works on the CLR principle, only keeping the oxygen carrier in a single reactor with switching valves to alternately expose it to different gases. The key novelty in the present study is the use of steam and heat in the CLC fuel reactor outlet stream to minimize this energy penalty. An important thermodynamic advantage of the CLC concept is that the steam resulting from the reduction reaction in the CLC fuel reactor along with the CO2 can be obtained at high pressures. The condensation enthalpy of this steam could be recovered at useful temperatures (150-200 °C), whereas it would be lost in the conventional NGCC plant where the water The primary energy penalty in hydrogen production with CLR-based processes is the provision of the steam required for NG reforming. When the produced hydrogen is combusted in a gas turbine, all the energy required to raise the steam in the hydrogen production process is lost because the condensation enthalpy of the steam resulting from hydrogen combustion cannot be converted to useful work. Nazir et al. [24] reported that this steam-related penalty accounts for about 5.8%-points of the 7.2%-point energy penalty of a combined cycle power plant fired by hydrogen from a gas switching reforming (GSR) process, illustrating the importance of this energy penalty. GSR works on the CLR principle, only keeping the oxygen carrier in a single reactor with switching valves to alternately expose it to different gases.
The key novelty in the present study is the use of steam and heat in the CLC fuel reactor outlet stream to minimize this energy penalty. An important thermodynamic advantage of the CLC concept is that the steam resulting from the reduction reaction in the CLC fuel reactor along with the CO 2 can be obtained at high pressures. The condensation enthalpy of this steam could be recovered at useful temperatures (150-200 • C), whereas it would be lost in the conventional NGCC plant where the water vapor in the flue gas condenses around room temperature. Thus, the integration of the CLC process presents the opportunity to recover and productively utilize this latent heat. Therefore, the main objective of the present work is to devise and compare different process integration options for efficiently utilizing the steam and heat in the CLC fuel reactor outlet gases for hydrogen production in the MA-CLR process. This principle can be applied to the integration of any NG reforming process where steam is required, but the MA-CLR process was selected for use in this study based on the promising techno-economic performance reported in Spallina et al. [22].

Plant Configurations
A conventional NGCC plant developed in our previous work [19] is used as a benchmark for comparing the results obtained in this study. The NGCC plant and all the other models are developed based on the recommendations of the European benchmarking task force (EBTF) report [25]. In an NGCC plant, fuel is preheated and the air is compressed before supplying to the combustor. The hot flue gases are expanded to produce power and sent to steam cycle for heat recovery. The F-class gas turbine system GE 9371FB with a COT of 1416 • C has been considered due to the robust design and fuel flexibility [25]. A triple pressure single reheat heat recovery steam generator (HRSG) is considered for steam generation from recovered heat. The steam is expanded in steam turbine (ST) and cooled down in condenser (COND) and cooling tower (CT). The entire steam cycle is simulated in Thermoflex [26] while the other equipment are modeled in Aspen Plus [27]. The main assumptions are taken from the EBTF report [25] and are given in Table 1. Figure 2 shows a schematic of a CLC combined cycle plant with an additional combustor and the arrangements for combustor fuel supply, steam injection, and exhaust gas recirculation for NOx control. The plant specifications are similar to the conventional CLC cycle without added firing. A nickel-based OC (NiO) supported on NiAl 2 O 4 is supplied to the fuel reactor (FR) where the endothermic reduction reaction produces CO 2 and H 2 O. The reduced OC is re-oxidized in the air reactor (AR) in an exothermic reaction with air. The O 2 depleted air from the AR is passed through an expander to produce power and sent to HRSG for heat recovery. The CO 2 stream is also expanded and is further used to preheat the incoming fuel followed by condensation of steam and CO 2 compression to high pressures (supercritical state). The operating conditions of this plant are taken from the work of Naqvi and Bolland [7]. More details about this plant, including stream information, can be found in our previous work [19].
This conventional CLC plant is fitted with an added combustor to raise the COT to match the reference NGCC plant. The H 2 is compressed to about 1.5 times the oxidizer (O 2 depleted air) pressure and injected into the combustor. The O 2 -depleted air from the AR is considered as the oxidizer in the combustor where its temperature is raised from 1160 • C to 1416 • C. The Gibbs energy minimization concept is considered in solving the heat and mass balances. As a result, the air reactor exhaust temperature of 1160 • C is obtained for the same inlet operating conditions. Such a simplified reactor modeling assumption is merited by the high reactivity of the Ni-based oxygen carrier employed. Given the limitations to the thermal stability of the oxygen carrier material, the temperature obtained in the air reactor is more conservative. It should be noted that higher CLC operating temperatures pose problems in solid handling such as sintering and agglomeration. The plant main assumptions are given in Table 1 and are also taken from the EBTF report [25].
The source of H 2 in Figure 2 is the MA-CLR plant represented in Figure 3. The NG is preheated to 324 • C before subjecting to desulfurization to convert any sulfur compounds. Then the steam is mixed with the NG to result in a steam/carbon ratio of 2, which is slightly higher than the ratio of 1.75 used by Spallina et al. [22] to ensure good methane conversion and no carbon deposition. The mixture is then pre-reformed using a nickel-based catalyst to convert higher hydrocarbons at 490 • C for minimizing coke formation in the MA-CLR reactor. Then the pre-reformed gas is further heated before entering the fuel reactor of the MA-CLR plant. As shown in Figure 3, the FR contains Pd-membranes for H 2 extraction. In this process, the same OC (NiO-NiAl 2 O 4 ) is used, which also acts as the catalyst for reforming and water-gas shift reactions. The retentate stream (CO 2 and H 2 O) comes out of the reactor from the top while the permeate stream consisting of pure H 2 is extracted through the membranes. The plant specifications are listed in Table 1 and more details, includeing stream information, can be found in our previous work [28]. In the current study, five cases with different degrees of integration between the CLC plant shown in Figure 2 and the MA-CLR plant shown in Figure 3 are investigated ( Table 2). All cases feature added combustion of hydrogen after the CLC reactors and the cases are arranged in order of increasing integration between the CLC and MA-CLR processes. Case 1 is the direct integration of the CLC and MA-CLR process by connecting the H 2 supply line to the combustor.
Case 2 has additional integration as shown in Figure 4. In this configuration, some part of the fuel reactor outlet stream of the CLC system is mixed with the NG required in the MA-CLR system maintaining steam-to-carbon ratio 2. This has the advantage of replacing the steam that normally needs to be raised for hydrogen production in the MA-CLR process with steam resulting from the combustion of NG. Just enough CLC fuel reactor flue gases are used to supply the steam required to produce enough hydrogen for the added combustor. The outlet stream from the AR of MA-CLR is expanded in the turbine followed by preheating the feed water. The retentate stream is used to produce saturated steam and then sent for compression whereas the permeate stream is used to superheat the steam, which is then sent to the HRSG. The permeate is then used to further preheat the feed water before it is compressed and sent to the combustor. Since the CO 2 stream is at 20 bar pressure, it is introduced in the CO 2 compression process after the second stage.
control. The plant specifications are similar to the conventional CLC cycle without added firing. A nickel-based OC (NiO) supported on NiAl2O4 is supplied to the fuel reactor (FR) where the endothermic reduction reaction produces CO2 and H2O. The reduced OC is re-oxidized in the air reactor (AR) in an exothermic reaction with air. The O2 depleted air from the AR is passed through an expander to produce power and sent to HRSG for heat recovery. The CO2 stream is also expanded and is further used to preheat the incoming fuel followed by condensation of steam and CO2 compression to high pressures (supercritical state). The operating conditions of this plant are taken from the work of Naqvi and Bolland [7]. More details about this plant, including stream information, can be found in our previous work [19]. This conventional CLC plant is fitted with an added combustor to raise the COT to match the reference NGCC plant. The H2 is compressed to about 1.5 times the oxidizer (O2 depleted air) pressure and injected into the combustor. The O2-depleted air from the AR is considered as the oxidizer in the combustor where its temperature is raised from 1160 °C to 1416 °C. The Gibbs energy minimization concept is considered in solving the heat and mass balances. As a result, the air reactor exhaust temperature of 1160 °C is obtained for the same inlet operating conditions. Such a simplified reactor modeling assumption is merited by the high reactivity of the Ni-based oxygen carrier employed. Given the limitations to the thermal stability of the oxygen carrier material, the temperature obtained in the air reactor is more conservative. It should be noted that higher CLC operating temperatures pose problems in solid handling such as sintering and agglomeration. The plant main assumptions are given in Table 1 and are also taken from the EBTF report [25]. The source of H2 in Figure 2 is the MA-CLR plant represented in Figure 3. The NG is preheated to 324 °C before subjecting to desulfurization to convert any sulfur compounds. Then the steam is mixed with the NG to result in a steam/carbon ratio of 2, which is slightly higher than the ratio of 1.75 used by Spallina et al. [22] to ensure good methane conversion and no carbon deposition. The mixture is then pre-reformed using a nickel-based catalyst to convert higher hydrocarbons at 490 °C for minimizing coke formation in the MA-CLR reactor. Then the pre-reformed gas is further heated before entering the fuel reactor of the MA-CLR plant. As shown in Figure 3, the FR contains Pdmembranes for H2 extraction. In this process, the same OC (NiO-NiAl2O4) is used, which also acts as the catalyst for reforming and water-gas shift reactions. The retentate stream (CO2 and H2O) comes out of the reactor from the top while the permeate stream consisting of pure H2 is extracted through the membranes. The plant specifications are listed in Table 1 and more details, includeing stream information, can be found in our previous work [28]. In the current study, five cases with different degrees of integration between the CLC plant shown in Figure 2 and the MA-CLR plant shown in Figure 3 are investigated ( Table 2). All cases feature added combustion of hydrogen after the CLC reactors and the cases are arranged in order of increasing integration between the CLC and MA-CLR processes. Case 1 is the direct integration of the CLC and MA-CLR process by connecting the H2 supply line to the combustor. Table 2. Summary of the CLC and MA-CLC integrations investigated in this study.

Case # Description
Case 1 A simple integration where H2 from MA-CLR is fed to the added combustor after the CLC reactors.

Case 2
Case 1 with an additional coupling by using part of the CLC fuel reactor flue gas as a steam source for reforming in MA-CLR.
Case 3 Using all the CLC fuel reactor flue gas as a steam source for reforming in MA-CLR and producing excess H2 for export.

Case 4
Combining CLC and MA-CLR into a single reactor unit and using a 2-phase flow heat exchanger to  Table 2. Summary of the CLC and MA-CLC integrations investigated in this study.

Case # Description
Case 1 A simple integration where H 2 from MA-CLR is fed to the added combustor after the CLC reactors.

Case 2
Case 1 with an additional coupling by using part of the CLC fuel reactor flue gas as a steam source for reforming in MA-CLR. Case 3 Using all the CLC fuel reactor flue gas as a steam source for reforming in MA-CLR and producing excess H 2 for export.

Case 4
Combining CLC and MA-CLR into a single reactor unit and using a 2-phase flow heat exchanger to raise steam from the steam condensation enthalpy in the fuel reactor outlet stream.
Case 5 Case 4 produces excess H 2 and Case 5 was formulated to produce only power by sweeping the membranes with additional steam to extract more heat from the reactor.
Energies 2020, 13, 3443 7 of 19 produce enough hydrogen for the added combustor. The outlet stream from the AR of MA-CLR is expanded in the turbine followed by preheating the feed water. The retentate stream is used to produce saturated steam and then sent for compression whereas the permeate stream is used to superheat the steam, which is then sent to the HRSG. The permeate is then used to further preheat the feed water before it is compressed and sent to the combustor. Since the CO2 stream is at 20 bar pressure, it is introduced in the CO2 compression process after the second stage.
. Case 3 is shown in Figure 5. In this case, the whole fuel reactor outlet stream of the CLC plant is mixed with NG and injected into the fuel reactor of MA-CLR. Hence, the CO2 expander is not included in this configuration. A large amount of steam from the CLC fuel reactor means that more H2 is produced than that required in the combustor. Additional H2 is compressed to high pressure (HP) and exported. Both the air reactor outlet streams of CLC and MA-CLR are mixed and sent to the combustor. This is done to further simplify the process by eliminating the expander used in the MA-CLR process. The AR outlet stream from the MA-CLR plant is no longer used for steam generation purpose, rather it is fed to the combustor where more H2 can be burnt to produce more power (in absolute terms) and subsequently recover more heat in the steam cycle. The retentate stream is used to produce saturated steam and preheats part of feed water whereas the permeate stream is used to superheat the steam followed by preheating the rest of the feed water.
Cases 4 and 5 combine the CLC and MA-CLR reactors into a single unit for greater process intensification by increasing the air flowrate and temperature of the MA-CLR air reactor to CLC levels. Since a large fraction of the methane fed to the membrane reactor must be combusted in this configuration, a large amount of fuel must slip past the membranes to reduce the oxygen carrier, allowing a smaller membrane surface area concentrated in the lower reactor regions where H2 partial pressures are high. Relative to cases 2 and 3, cases 4 and 5 also avoid the feed of CO2 in the CLC fuel reactor stream to the membrane reactor, thereby shifting the equilibrium reactions further towards greater H2 production, allowing additional reductions in the required membrane surface area. A Case 3 is shown in Figure 5. In this case, the whole fuel reactor outlet stream of the CLC plant is mixed with NG and injected into the fuel reactor of MA-CLR. Hence, the CO 2 expander is not included in this configuration. A large amount of steam from the CLC fuel reactor means that more H 2 is produced than that required in the combustor. Additional H 2 is compressed to high pressure (HP) and exported. Both the air reactor outlet streams of CLC and MA-CLR are mixed and sent to the combustor. This is done to further simplify the process by eliminating the expander used in the MA-CLR process. The AR outlet stream from the MA-CLR plant is no longer used for steam generation purpose, rather it is fed to the combustor where more H 2 can be burnt to produce more power (in absolute terms) and subsequently recover more heat in the steam cycle. The retentate stream is used to produce saturated steam and preheats part of feed water whereas the permeate stream is used to superheat the steam followed by preheating the rest of the feed water.
Cases 4 and 5 combine the CLC and MA-CLR reactors into a single unit for greater process intensification by increasing the air flowrate and temperature of the MA-CLR air reactor to CLC levels. Since a large fraction of the methane fed to the membrane reactor must be combusted in this configuration, a large amount of fuel must slip past the membranes to reduce the oxygen carrier, allowing a smaller membrane surface area concentrated in the lower reactor regions where H 2 partial pressures are high. Relative to cases 2 and 3, cases 4 and 5 also avoid the feed of CO 2 in the CLC fuel reactor stream to the membrane reactor, thereby shifting the equilibrium reactions further towards greater H 2 production, allowing additional reductions in the required membrane surface area. A future economic assessment with more detailed reactor modeling will be required to quantify these benefits.
In case 4 ( Figure 6), the NG is cooled after desulphurization before being mixed with liquid water pumped at 19 bar to maintain a steam-to-carbon ratio of 2. The mixture is sent to a two-phase heat exchanger (2P-HX). This gas-liquid mixture (stream 4 in Figure 6) allows the water to start evaporating at low temperatures so that the condensation enthalpy from the steam in the retentate stream (CO 2 stream) exiting the FR can be used to efficiently raise this steam, effectively granting access to the HHV of the NG (~10% greater than the LHV). The retentate stream is also used to preheat the water to 114 • C before compression to a supercritical state. The NG-steam mixture is fed into the FR of the MA-CLR plant. This reactor configuration with the membranes is the same as discussed above. The permeate stream consisting of pure H 2 exits at 700 • C and is used to generate HP superheated steam (stream 21, Figure 6), which is supplied to the HRSG for power generation. The required amount of H 2 in the combustor to maintain the COT of 1416 • C is split from the permeate stream (stream 12, Figure 6), with the remainder being compressed to storage pressure. The split H 2 is burnt in the combustor with an AR exhaust stream as the oxidizer to power the gas turbine and downstream heat recovery in the steam cycle.
Energies 2020, 13, x FOR PEER REVIEW 8 of 19 future economic assessment with more detailed reactor modeling will be required to quantify these benefits. In case 4 ( Figure 6), the NG is cooled after desulphurization before being mixed with liquid water pumped at 19 bar to maintain a steam-to-carbon ratio of 2. The mixture is sent to a two-phase heat exchanger (2P-HX). This gas-liquid mixture (stream 4 in Figure 6) allows the water to start evaporating at low temperatures so that the condensation enthalpy from the steam in the retentate stream (CO2 stream) exiting the FR can be used to efficiently raise this steam, effectively granting access to the HHV of the NG (~10% greater than the LHV). The retentate stream is also used to preheat the water to 114 °C before compression to a supercritical state. The NG-steam mixture is fed into the FR of the MA-CLR plant. This reactor configuration with the membranes is the same as discussed above. The permeate stream consisting of pure H2 exits at 700 °C and is used to generate HP superheated steam (stream 21, Figure 6), which is supplied to the HRSG for power generation. The required amount of H2 in the combustor to maintain the COT of 1416 °C is split from the permeate stream (stream 12, Figure 6), with the remainder being compressed to storage pressure. The split H2 is burnt in the combustor with an AR exhaust stream as the oxidizer to power the gas turbine and downstream heat recovery in the steam cycle.  Table 3. Stream data of case 4 with 2P-HX (mainly H 2 production) in Figure 6.  Case 4 produces a large amount of steam from the 2-phase flow heat exchanger, which leads to a lot of hydrogen production. It is therefore primarily a hydrogen production plant with some power production. Therefore, case 5 ( Figure 7) is devised for power production only, implying that the whole permeate H 2 (stream 12, Figure 7) is burnt in the combustor.

St
Energies 2020, 13, x FOR PEER REVIEW 9 of 19 Figure 6. Schematic of case 4 (part H2 is used in combustor). Stream numbers refer to stream details listed later in Table 3.
Case 4 produces a large amount of steam from the 2-phase flow heat exchanger, which leads to a lot of hydrogen production. It is therefore primarily a hydrogen production plant with some power production. Therefore, case 5 ( Figure 7) is devised for power production only, implying that the whole permeate H2 (stream 12, Figure 7) is burnt in the combustor.  Table 3. Table 4. Stream data of case 5 with 2P-HX (power production) in Figure 7.   Table 4.
The main difference is that the HP superheater steam (stream 19, Figure 7) is expanded in a steam turbine and used as a sweep gas in the membranes. In this way, the steam is heated to the COT, requiring more heat production at the expense of hydrogen production. The amount of steam used in the sweep is adjusted so that the balance between heat and hydrogen production is just right to reach the specified COT when all the produced H2 is fed to the combustor. An additional benefit is that the steam sweep lowers the partial pressure of H2 in the membranes, increasing the driving force for H2 permeability. This was accounted for by increasing the permeate pressure to 6 bar in this case.

Process Modeling and Plant Performance Indicators
Thermodynamic equilibrium is considered while solving the mass and energy balances in all the equipment. The property method, Redlich-Kwong-Soave equation of state with Boston-Mathias alpha function (RKS-BM), is used as it is recommended for hydrocarbons and applications involving light gases such as CO2 and H2 [29]. For the CLC plant, the reactors and the combustor are modeled using the reactor based on Gibbs energy minimization (RGIBBS) module in Aspen Plus [27], which assumes chemical and phase equilibrium. Perfect solid and gas separation is assumed and is done by using a cyclone block at the AR exit and a separator block at the FR exit. The efficiencies of the air/H2 compressors, the gas turbine, and the expander are evaluated using a polytropic with the gas processors suppliers association (GPSA) method [30].
For the MA-CLR plant, a simple zero-dimensional (0D) mass and energy balance model is used to determine the reactor behavior, implying that no internal species or temperature profiles are resolved in the reactor. The temperature-dependent enthalpies of the gases were taken from the JANAF thermochemical database [31]. The results of this model are coupled with the process simulations. The main assumptions used in developing the models are listed in Table 1. The model  Table 4.
The main difference is that the HP superheater steam (stream 19, Figure 7) is expanded in a steam turbine and used as a sweep gas in the membranes. In this way, the steam is heated to the COT, requiring more heat production at the expense of hydrogen production. The amount of steam used in the sweep is adjusted so that the balance between heat and hydrogen production is just right to reach the specified COT when all the produced H 2 is fed to the combustor. An additional benefit is that the steam sweep lowers the partial pressure of H 2 in the membranes, increasing the driving force for H 2 permeability. This was accounted for by increasing the permeate pressure to 6 bar in this case.

Process Modeling and Plant Performance Indicators
Thermodynamic equilibrium is considered while solving the mass and energy balances in all the equipment. The property method, Redlich-Kwong-Soave equation of state with Boston-Mathias alpha function (RKS-BM), is used as it is recommended for hydrocarbons and applications involving light gases such as CO 2 and H 2 [29]. For the CLC plant, the reactors and the combustor are modeled using the reactor based on Gibbs energy minimization (RGIBBS) module in Aspen Plus [27], which assumes chemical and phase equilibrium. Perfect solid and gas separation is assumed and is done by using a cyclone block at the AR exit and a separator block at the FR exit. The efficiencies of the air/H 2 compressors, the gas turbine, and the expander are evaluated using a polytropic with the gas processors suppliers association (GPSA) method [30].
For the MA-CLR plant, a simple zero-dimensional (0D) mass and energy balance model is used to determine the reactor behavior, implying that no internal species or temperature profiles are resolved in the reactor. The temperature-dependent enthalpies of the gases were taken from the JANAF thermochemical database [31]. The results of this model are coupled with the process simulations.
The main assumptions used in developing the models are listed in Table 1. The model is solved to estimate the required air and the composition of the O 2 depleted stream from the MA-CLR AR (assuming complete O 2 conversion). Furthermore, the flowrate of the ultra-pure H 2 at the membrane outlet and the composition of the CO 2 rich stream at the MA-CLR FR outlet (assuming complete fuel conversion) are also obtained.
As mentioned earlier, the steam cycle consisting of a condensing reheat steam turbine system and HRSG was modeled in Thermoflex [26]. The reheat and the HP/IP/LP steam temperatures adjust automatically depending on the heat recovery in each case. Similarly, the efficiencies of the steam turbine and the condenser pressure varied depending on the stream conditions and turbine size required. A natural draft cooling tower system is considered for the condenser and inter-stage compressor cooling.
The plant performance of all the cases in the present study is evaluated by using net electrical efficiency (%), hydrogen production efficiency (%), efficiency penalty (%-points), CO 2 capture efficiency (%), and CO 2 avoidance (%) as shown below. The efficiency penalty is expressed using the global efficiency (Equation (6) Overall efficiency : In these equations, η is the efficiency (%), .
W is the net power produced (kW), . m is the fuel/H 2 mass flow (kg/s), LHV is the fuel/H 2 lower heating value (kJ/kg), and E is the emissions intensity (kg/kWh). The subscript ref stands for reference plant and CCS stands for the different CO 2 capture plants considered. The global efficiencies are the equivalent efficiencies calculated based on the fuel used for either electricity or H 2 production i.e., fuel used for H 2 or electricity is deducted from the total fuel input in their respective equations. The reference plant efficiencies for H 2 production and power generation were selected for an SMR plant (η = 80%) and NGCC plant (η = 58%), respectively.

Results and Discussion
The main results for power production from the different plant configurations are shown in Figure 8 with a power breakdown provided in Figure 9. As a reference, the net electrical efficiency of the NGCC plant is 58.17%. It is also noted that, with an absorption post-combustion CO 2 capture plant (not modeled in this study), the energy penalty is around 8 %-points as reported in the literature [32]. This reference plant would achieve about 90% CO 2 capture, whereas the CO 2 capture efficiency and avoidance are almost 100% in all the cases considered in this study, assuming perfect loop-seal performance. Real systems may show lower CO 2 capture ratios if loop seals are not 100% effective.
Energies 2020, 13, x FOR PEER REVIEW 12 of 19 The net electrical efficiency of a conventional CLC combined cycle plant is 49.39% [19], which is slightly below the NGCC plant with conventional CO2 capture. As outlined in the introduction, this relatively poor CLC performance can be improved by added firing with efficiently produced hydrogen, as is the focus of the present study. In case 1, the added firing of the CLC plant with an additional combustor is done with hydrogen from a standalone MA-CLR plant. As shown in Figure 8, the net electrical efficiency of this configuration is 53.08% with an energy penalty of 5.09%-points. This is a substantial improvement when compared with the conventional CLC combined cycle plant and the NGCC plant with CO2 capture. As shown in Figure 9, the combined power output from the gas turbine, steam turbines, and CO2 expander in case 1 is about 2%-points lower than the reference plant. This is due to the less efficient power production (expansion at lower temperatures) of the CO2 stream from CLC and the N2 stream from MA-CLR. Most of the remaining 3%-points of efficiency penalty come from CO2 compression after the CO2 expander and H2 compression before the added combustor. Thus, by simply utilizing efficiently produced H2 in the added combustor, without any additional heat integration, the energy penalty can be reduced substantially relative to conventional CO2 capture, while capturing all produced CO2. In comparison, most of the energy penalty in post-combustion capture (PCC) system (~8%-points [33]) stems from using LP steam in regenerating the solvent that would have been used for power generation. This energy intensive step is avoided in CLC systems. Moreover, raising the TIT of such systems using an additional combustor improves the gas turbine efficiency. Hence, the overall improvement by firing H2 from a highly efficient source is about 3.7%points when compared to conventional CLC plant that operates at a lower TIT. Furthermore, the CO2 capture efficiency from a PCC system is generally 90%, and higher CO2 capture rates will increase the energy penalty significantly due to equilibrium constraints. Thus, achieving almost 100% capture with very high purity in these systems demonstrates that the energy consumption per kilogram of CO2 capture will be lower. In addition, the electrical energy for compression is 0.32 MJ/kg-CO2 as opposed to 0.43 MJ/kg-CO2 in a PCC system [34] because the MA-CLR plant produces CO2 already at elevated pressures. The net electrical efficiency of a conventional CLC combined cycle plant is 49.39% [19], which is slightly below the NGCC plant with conventional CO 2 capture. As outlined in the introduction, this relatively poor CLC performance can be improved by added firing with efficiently produced hydrogen, as is the focus of the present study.
In case 1, the added firing of the CLC plant with an additional combustor is done with hydrogen from a standalone MA-CLR plant. As shown in Figure 8, the net electrical efficiency of this configuration is 53.08% with an energy penalty of 5.09%-points. This is a substantial improvement when compared with the conventional CLC combined cycle plant and the NGCC plant with CO 2 capture. As shown in Figure 9, the combined power output from the gas turbine, steam turbines, and CO 2 expander in case 1 is about 2%-points lower than the reference plant. This is due to the less efficient power production (expansion at lower temperatures) of the CO 2 stream from CLC and the N 2 stream from MA-CLR. Most of the remaining 3%-points of efficiency penalty come from CO 2 compression after the CO 2 expander and H 2 compression before the added combustor. Thus, by simply utilizing efficiently produced H 2 in the added combustor, without any additional heat integration, the energy penalty can be reduced substantially relative to conventional CO 2 capture, while capturing all produced CO 2 . In comparison, most of the energy penalty in post-combustion capture (PCC) system (~8%-points [33]) stems from using LP steam in regenerating the solvent that would have been used for power generation. This energy intensive step is avoided in CLC systems. Moreover, raising the TIT of such systems using an additional combustor improves the gas turbine efficiency. Hence, the overall improvement by firing H 2 from a highly efficient source is about 3.7%-points when compared to conventional CLC plant that operates at a lower TIT. Furthermore, the CO 2 capture efficiency from a PCC system is generally 90%, and higher CO 2 capture rates will increase the energy penalty significantly due to equilibrium constraints. Thus, achieving almost 100% capture with very high purity in these systems demonstrates that the energy consumption per kilogram of CO 2 capture will be lower. In addition, the electrical energy for compression is 0.32 MJ/kg-CO 2 as opposed to 0.43 MJ/kg-CO 2 in a PCC system [34] because the MA-CLR plant produces CO 2 already at elevated pressures. With closer process coupling in case 2 (Figure 4), the heat integration scheme is substantially simplified, largely because no steam needs to be raised for the MA-CLR process (steam is supplied by the CLC fuel reactor outlet gases). Figure 8 shows that the net electrical efficiency obtained in this case is 54.62%, an improvement of ~1.54%-points. As shown in Figure 9, about 0.5%-points of this improvement come from a decrease in H2 and CO2 compression power and the remaining 1%-point from increased power generation from gas, steam, and CO2 turbines. CO2 compression duty is reduced to 0.24 MJ/kg-CO2 because part of the CLC fuel reactor outlet gases is not expanded to atmospheric pressure (being fed to the MA-CLR process instead). The retentate stream from the MA-CLR is at 20 bar pressure and is therefore introduced in the CO2 compression train in the third stage. On the other hand, this reduces the power production by CO2 expander to 2.5% of LHV input. However, this 1.8 %-point reduction in power output is more than compensated by a 0.9 %-point increase in gas turbine power output and a 2%-point increase in steam turbine output. Due to the efficient integration of steam from the CLC fuel reactor, less heat is required in the MA-CLR plant, reducing its fuel consumption for a given hydrogen output and making more steam available to the steam turbine. When compared to the NGCC reference case, the energy penalty for case 2 is 3.55%points ( Figure 8). This mainly comes from the H2/CO2 compressors as the reduction in power production from the gas, CO2 and steam turbines is only 0.98%-points.
In case 3 ( Figure 5), the whole CO2 outlet stream from the CLC plant is mixed with the NG feed to the MA-CLR plant so that excess H2 can be produced while maintaining the steam-to-carbon ratio of 2. As shown in Figure 8, the electrical efficiency obtained is only 39.92% due to the significant amount of hydrogen that is exported by this plant (hydrogen production efficiency of 21.36%). Power production is shifted more towards the steam turbines due to the substantial amount of heat that is made available to the MA-CLR reactors from the CLC reduction outlet, allowing more steam for power production to be raised in the MA-CLR plant. It can also be noted that hydrogen compression With closer process coupling in case 2 (Figure 4), the heat integration scheme is substantially simplified, largely because no steam needs to be raised for the MA-CLR process (steam is supplied by the CLC fuel reactor outlet gases). Figure 8 shows that the net electrical efficiency obtained in this case is 54.62%, an improvement of~1.54%-points. As shown in Figure 9, about 0.5%-points of this improvement come from a decrease in H 2 and CO 2 compression power and the remaining 1%-point from increased power generation from gas, steam, and CO 2 turbines. CO 2 compression duty is reduced to 0.24 MJ/kg-CO 2 because part of the CLC fuel reactor outlet gases is not expanded to atmospheric pressure (being fed to the MA-CLR process instead). The retentate stream from the MA-CLR is at 20 bar pressure and is therefore introduced in the CO 2 compression train in the third stage. On the other hand, this reduces the power production by CO 2 expander to 2.5% of LHV input. However, this 1.8%-point reduction in power output is more than compensated by a 0.9%-point increase in gas turbine power output and a 2%-point increase in steam turbine output. Due to the efficient integration of steam from the CLC fuel reactor, less heat is required in the MA-CLR plant, reducing its fuel consumption for a given hydrogen output and making more steam available to the steam turbine. When compared to the NGCC reference case, the energy penalty for case 2 is 3.55%-points ( Figure 8). This mainly comes from the H 2 /CO 2 compressors as the reduction in power production from the gas, CO 2 and steam turbines is only 0.98%-points.
In case 3 ( Figure 5), the whole CO 2 outlet stream from the CLC plant is mixed with the NG feed to the MA-CLR plant so that excess H 2 can be produced while maintaining the steam-to-carbon ratio of 2. As shown in Figure 8, the electrical efficiency obtained is only 39.92% due to the significant amount of hydrogen that is exported by this plant (hydrogen production efficiency of 21.36%). Power production is shifted more towards the steam turbines due to the substantial amount of heat that is made available to the MA-CLR reactors from the CLC reduction outlet, allowing more steam for power production to be raised in the MA-CLR plant. It can also be noted that hydrogen compression work increased due to significant hydrogen exports that require compression to 150 bar. On the other hand, CO 2 compression duty reduced because no CO 2 is expanded to atmospheric pressure in this case. The CO 2 in MA-CLR retentate is at 20 bar and require only one stage compression to reach supercritical state. As a result, the CO 2 compression energy is 0.15 MJ/kg-CO 2 , a considerable reduction from previous cases. Hence, this case has no generation from a CO 2 turbine. The global electrical and H 2 production efficiencies are depicted in Figure 10, showing that the actual electric efficiency of this plant is similar to case 2. The global H 2 efficiency obtained is 68.09%, which is similar to a conventional SMR plant with CO 2 capture [22]. The overall energy efficiency obtained is 61.28% which falls between the NGCC and SMR plant energy efficiencies. When compared to NGCC reference case, the actual energy penalty in terms of electricity is 3.71%-points ( Figure 8).
Energies 2020, 13, x FOR PEER REVIEW 14 of 19 supercritical state. As a result, the CO2 compression energy is 0.15 MJ/kg-CO2, a considerable reduction from previous cases. Hence, this case has no generation from a CO2 turbine. The global electrical and H2 production efficiencies are depicted in Figure 10, showing that the actual electric efficiency of this plant is similar to case 2. The global H2 efficiency obtained is 68.09%, which is similar to a conventional SMR plant with CO2 capture [22]. The overall energy efficiency obtained is 61.28% which falls between the NGCC and SMR plant energy efficiencies. When compared to NGCC reference case, the actual energy penalty in terms of electricity is 3.71%-points ( Figure 8). It can be noted that added firing up to the high TIT of modern gas turbines (~1600 °C) will shift the balance of energy output from the plant more towards electricity as more of the produced H2 will be required in the added combustor to achieve the larger temperature increase of the depleted air stream from CLC. The good global electric efficiency of this plant suggests that integrated CLC and MA-CLR plants will remain attractive as the firing temperature of gas turbines keeps improving over the coming decades. It has been estimated in our previous work that, by using a state-of-the-art gas turbine with TIT 1550 °C, the energy penalty could be as low as 0.6% [33].
In case 4, the MA-CLR and CLC plants are combined to produce both power and H2. As shown in Figure 8, this plant is mainly a hydrogen production plant with a net electrical efficiency of only 8.25%. It can also be noted from Figure 9 that the gas and steam turbines now produce almost similar power. Given that only a small fraction of the fuel heating value is converted to high-grade heat for generating power in the combined cycle, the steam that can be raised from cooling the large permeate stream substantially increases the relative contribution of the steam turbines. The H2 compression work is considerably larger due to the large amount of hydrogen exports, whereas the CO2 compression work is low because the CO2 stream is not expanded. The global electrical efficiency depicted in Figure 10 is 56.32% which gives an energy penalty of only 1.85%-points. This is a very good performance, but it can only be achieved when the plant operates mostly as a hydrogen It can be noted that added firing up to the high TIT of modern gas turbines (~1600 • C) will shift the balance of energy output from the plant more towards electricity as more of the produced H 2 will be required in the added combustor to achieve the larger temperature increase of the depleted air stream from CLC. The good global electric efficiency of this plant suggests that integrated CLC and MA-CLR plants will remain attractive as the firing temperature of gas turbines keeps improving over the coming decades. It has been estimated in our previous work that, by using a state-of-the-art gas turbine with TIT 1550 • C, the energy penalty could be as low as 0.6% [33].
In case 4, the MA-CLR and CLC plants are combined to produce both power and H 2 . As shown in Figure 8, this plant is mainly a hydrogen production plant with a net electrical efficiency of only 8.25%. It can also be noted from Figure 9 that the gas and steam turbines now produce almost similar power. Given that only a small fraction of the fuel heating value is converted to high-grade heat for generating power in the combined cycle, the steam that can be raised from cooling the large permeate stream substantially increases the relative contribution of the steam turbines. The H 2 compression work is considerably larger due to the large amount of hydrogen exports, whereas the CO 2 compression work is low because the CO 2 stream is not expanded. The global electrical efficiency depicted in Figure 10 is 56.32% which gives an energy penalty of only 1.85%-points. This is a very good performance, but it can only be achieved when the plant operates mostly as a hydrogen production facility to efficiently utilize the large amount of steam available from the 2-phase flow heat exchanger for hydrogen production. Due to this efficient steam generation, the global H 2 production efficiency is also high at 79.57%, achieving parity with an SMR plant without CO 2 capture. The resulting overall efficiency obtained is 76.53%, which is 15.25%-points higher than that of case 3.
Stream information for this case is shown in Table 3. The 2-phase flow heat exchanger exchanges heat between streams 7 and 4, using some of the condensation enthalpy in the steam in stream 7 to raise steam in stream 4. It is shown that stream 7 has a slightly lower pressure and steam mole fraction than stream 4, meaning that the steam in stream 7 will start to condense at a lower temperature than the last steam in stream 4 will evaporate. Thus, not all the steam condensation enthalpy in stream 7 can be used to generate steam in stream 4, requiring the sensible heat to be used instead. For this reason, stream 4 could only be heated to 220 • C before introduction to the MA-CLR reactor (stream 5). Similar to case 3, it can also be noted that the use of modern gas turbines with higher TIT will shift the energy output balance more towards power. The very high global electric efficiency of this plant suggests that such an integration will be highly attractive. However, an important limitation is that turbines with very high TIT generally have high power ratings, requiring the plant to produce a large power output. To utilize such a large-scale gas turbine, a greater share of output energy must go to power production than that observed for case 4 in Figure 9.
Case 5 illustrates an integration where case 4 can be modified for power production only. Instead of being sent to the main steam cycle, the steam produced by cooling the permeate stream is expanded to 6 bar and used as sweep gas in the Pd membranes to extract more heat from the reactor and increase the driving force for H 2 permeation. Figure 8 shows that the net electrical efficiency obtained is 53.63% with an energy penalty of 4.54%-points, which is about 1%-point more than case 2. This is largely because considerably more hydrogen needs to be produced and fed to the added combustor in this case due to the additional mass of the steam diluting the hydrogen fuel that must be heated up after the air reactor. Even though this integration achieves a high hydrogen production efficiency, producing heat via the CLC mechanism is still considerably more efficient than firing efficiently produced hydrogen, so a shift in heat release from the CLC reactor to the added combustor reduces plant efficiency. In addition, the H 2 compression duty for case 5 is 0.5%-points higher than for case 2 because of the larger H 2 flowrate that needs to be fed to the added combustor and the additional steam in this stream.
The CO 2 capture efficiency and CO 2 avoidance, in this case, are both 99.8%. This is slightly less than the previous cases due to the presence of large amount of water in the mixture (stream 8, Table 4) resulting in more solute CO 2 loss with the liquid water. Table 4 also shows that the steam content of stream 7 is considerably higher than it was in Table 3 due to the larger amount of NG that is combusted instead of being reformed. This allows the steam in this stream to start condensing at higher temperatures in the 2-phase flow heat exchanger. For this reason, most of the steam condensation enthalpy can be efficiently recovered to produce new steam in stream 5, preheating this stream to 650 • C in Table 4 instead of only 220 • C as shown in Table 3.
It will also be possible to develop a combination of the steam treatment in cases 4 and 5 by splitting the produced steam between the steam cycle and the membrane sweep. This will shift case 4 towards greater power production as required by large state-of-the-art gas turbines while improving on the efficiency of case 5.
Flexible operation should also be possible with these plants. In general, part-load operation of the gas turbine reduces the TIT, the pressure ratio, and the air flowrate. A lower TIT will require less added hydrogen firing, decreasing the energy penalty related to clean hydrogen production. This could allow more H 2 to be exported, but hydrogen production from the membranes will also decrease due to the lower pressure in the membrane reactor. This could be compensated to some degree by further reducing the permeate pressure. More detailed studies at different gas turbine operating points will be required to better understand the potential for flexible power and hydrogen production from these highly efficient process integrations.

Conclusions
The current study deals with the efficiency improvement by process integration of a CLC plant and an MA-CLR plant. Such an integration solves the fundamental low turbine inlet temperature issue when using CLC technology for combined cycle power production. A hydrogen fired combustor is added after the CLC reactors to raise the CLC outlet temperature for achieving a specified TIT. If this hydrogen can be efficiently produced, such added firing can bring substantial efficiency gains.
Process integration for producing the hydrogen needed for added firing in a highly efficient manner was the primary focus of this study. The MA-CLR concept was identified as the most promising alternative for supplying this hydrogen. When the CLC and MA-CLR plants are operated independently in case 1, the energy penalty already reduces from around 8%-points for post-combustion CO 2 capture and CLC without added firing to 5.1% points. Closer integration by feeding some of the CLC fuel reactor outlet gases to the MA-CLR reactor as a steam supply in case 2 reduced the energy penalty to 3.6%-points.
When all the CLC fuel reactor outlet gases are used in the MA-CLR reactor in case 3, excess hydrogen can be produced. The equivalent net electric efficiency remained almost unchanged, but about a third of the plant energy output shifted to hydrogen produced with a reasonable equivalent efficiency of 68.09%.
Finally, the CLC and MA-CLR reactors were combined by operating the MA-CLR air reactor at the high air flowrates and temperatures needed by CLC. This integration was combined with a 2-phase flow heat exchanger that efficiently produced steam for reforming from the condensation enthalpy of the steam in the fuel reactor outlet stream. The first such plant integration (case 4) produced primarily hydrogen at a very attractive efficiency of 79.57%. Only about 11% of the useful plant energy output was power, also produced at high efficiency, resulting in an energy penalty of only 1.9%-points. Transitioning this plant to power production in case 5 only increases the energy penalty to 4.5%-points.
In conclusion, several promising process integrations were identified for overcoming the fundamental drawback of low TIT when using CLC for high efficiency combined cycle power production. These integrations can greatly reduce the energy penalty of CO 2 capture from natural gas-fired power plants while maintaining complete CO 2 avoidance. Further research is recommended.