*2.3. Boundary Conditions*

For all subsequent simulations, the biomass input was selected in such a way, that the thermal load, *Pth*, of the chemical looping gasifier amounted to 1 MW. In terms of the circulating solid materials, the deployed oxygen carrier material is ilmenite, for which is has been established that the major redox stages are FeO + TiO2, Fe3O4, TiO2 and Fe2TiO5 [51]. These redox stages were modelled as FeTiO3 (for FeO + TiO2), Fe3O4, TiO2, and Fe2O3 + TiO2 (for Fe2TiO5). Deeper redox stages (e.g., FeO) were also considered in the process model, ye<sup>t</sup> were not found to be formed in notable amounts. The inert solid sand was modelled through pure SiO2. The FR and AR are operated under atmospheric pressure. Moreover, the air reactor temperature was set to 1050 ◦C, if not stated otherwise. The fuel reactor temperature results from the energy balance of the process, requiring that both reactors are in heat balance ( . *QFR* = 0, . *QAR* ≥ 0). As the kinetic syngas inhibition of char gasification reactions [8,12] is not considered in the RGIBBS equilibrium calculation, full char conversion is attained inside the FR for all temperatures considered in this study. Although this simplification signifies a deviation from reality, it does not impact the general inferences which will be elaborated on hereinafter. For the steam to biomass ratio in the FR a value of 0.9, reported for a 2–4 MWth chemical looping gasifier in literature [52], was selected if not stated otherwise. During CLC/CLG operation CO2 is required for fuel feeding and inerting. This stream of CO2, entering the fuel reactor, was selected in such a way that the CO2 to biomass ratio amounts to 0.2, to take into account that the CO2 input through the feeding section increases with increased thermal load. The two remaining process variables, the air mass flow entering the AR and the circulating oxygen carrier mass, were adjusted in such a way that autothermal CLG operation was achieved. A summary of all boundary conditions is given in Table A1 in Appendix A.

#### **3. Results and Discussion**

#### *3.1. Attaining CLG Behavior*

Generally, shifting from a combustion to a gasification process is achieved through lowering the air/oxygen-to-fuel ratio of the process, thereby decreasing the ratio of fully to partially oxidized gas species leaving the process and hence increasing the heating value of the product gas [6,53,54]. Here, the critical parameter is the so called air-to-fuel equivalence ratio given by the ratio of oxygen fed to the AR, . *mO*,*AR*, and the oxygen required for full feedstock combustion, . *mO*,*stoich*:

$$
\lambda = \frac{\dot{m}\_{O,AR}}{\dot{m}\_{O,\text{stoich}}}.\tag{10}
$$

According to this definition, (close to) full combustion of the feedstock is attained for air-to-fuel equivalence ratios larger than unity (λ > 1), while gasification processes require sub-stoichiometric oxygen feeding (i.e., λ < 1).

Due to the dissection of the gasification/combustion reaction into two separate reactors in chemical looping processes, there is no direct contact between the air entering the AR and the fuel entering the FR. Hence, the application of an alternative parameter, the oxygen-carrier-to-fuel equivalence ratio, φ , relating the amount of oxygen carried by the OC to the FR to the oxygen required for stoichiometric combustion, has been suggested [43]:

$$
\phi' = \frac{R\_{\rm OC} \cdot \dot{m}\_{\rm OC}}{\dot{m}\_{\rm O,stoich}}.\tag{11}
$$

Here, *ROC* denotes the oxygen transport capacity of the given oxygen carrier material. While this parameter accurately relates the two quantities for CLC, where the OC always leaves the AR in a (close to) fully oxidized state, this is not necessarily the case in CLG. Therefore, a slightly altered oxygen-carrier-to-fuel equivalence ratio, φ, considering the possibility of a partially reduced OC leaving the AR, has been proposed for gasification applications [35]:

$$\phi = \frac{\mathcal{R}\_{\text{OC}} \cdot \dot{m}\_{\text{OC}^\*} X\_{s, \text{AR}}}{\dot{m}\_{\text{O,stoich}}},\tag{12}$$

where *Xs*,*AR* signifies the oxidation degree of the oxygen carrier at the AR outlet, given by [24,35]:

$$X\_{s,AR} = \frac{m\_{\rm OC,AR} - m\_{\rm OC,rad}}{R\_{\rm OC} \cdot m\_{\rm OC,ox}}.\tag{13}$$

Here, *mOC*,*red* and *mOC*,*ox* are the mass of an OC sample in a fully reduced and oxidized state respectively, while *mOC*,*AR* is the mass of the OC sample leaving the AR. For ilmenite the fully reduced oxygen carrier is approximated by FeTiO3, the fully oxidized state is approximated by Fe2O3 + 2TiO2, and Fe3O4 + 3TiO2 denotes an intermediate redox state ( *Xs* = 0.67).

In order to assess how λ and φ have to be adjusted in order to obtain an e fficient CLG process, one should first assess the general impact of these two parameters on the process. Due to the relative fast kinetics of the OC re-oxidation [55–57], the oxygen carrier is often assumed to leave the AR in a (close to) fully oxidized state for λ > 1 in chemical looping processes. In contrast, sub-stoichiometric air-to-fuel equivalence ratios (λ < 1) only lead to a partial re-oxidation of the OC in the AR. Following the same logic, the OC material can be assumed to leave the FR in a (close to) fully reduced state in case φ < 1, whereas partial reduction is attained for φ > 1. From these deductions, it becomes clear that "standard" CLC operation is attained for λ > 1 and φ > 1, [42,43]. Here, a highly oxidized OC leaves the AR, before being partially reduced in the FR, which is illustrated in Figure 3a.

**Figure 3.** Different chemical looping modes (**<sup>a</sup>**–**d**) dependent on the air-to-fuel equivalence ratio λ and the oxygen-carrier-to-fuel equivalence ratio φ.

When targeting pronounced syngas formation, the oxygen release in the FR has to be limited, so that full feedstock oxidation is prevented [35,52]. The most obvious avenue that can be pursued to achieve this is lowering φ below unity. When doing so, the employed air-to-fuel equivalence ratio λ determines how much oxygen is transported between the two reactors per gram of OC. In case of λ > 1, which is illustrated in Figure 3b, the oxygen carrier undergoes a full redox cycle and hence the full oxygen transport capacity of the OC material (i.e., *ROC*) is exploited. On the other hand, λ < 1 means that in equilibrium the OC leaves the AR in a partially reduced state, hence also reducing the mass specific oxygen transport of the OC (see Figure 3c). Lastly, one might also consider a process with λ < 1 and φ > 1, as shown Figure 3d. In order to attain a steady-state process exhibiting these characteristics, full reduction of the oxygen carrier has to be prevented in the FR (e.g., kinetically), so that a fraction of oxygen is transported back to the AR. This means that in contrast to the former approaches, this case cannot be attained in equilibrium-like conditions. While this approach might also be feasible for CLG operation in theory, straight forward measures allowing for a controlled oxygen release in the FR are not at hand. Consequently, lowering the oxygen-to-fuel-ratio in the FR (i.e., φ < 1) is the most promising avenue to attain CLG behavior. When aiming for large syngas yields, φ has to assume values below unity, while values exceeding unity are targeted in CLC [42,43]. In the following, different effective control strategies to achieve this reduction in φ, required for pronounced syngas formation in the FR, while at the same time achieving an autothermal process, will be investigated.

In order to simplify the subsequent considerations, a standard parameter to describe gasification processes, the cold gas efficiency (CGE), η*CG*, will be deployed hereinafter. It describes which amount of chemical energy from the fuel is transferred to the gaseous FR product gas during gasification [6,7].

$$\eta\_{\rm CG} = \frac{\dot{n}\_{\rm gas,FR} \cdot (\mathbf{x}\_{\rm CH4,FR} \cdot LHV\_{\rm CH4} + \mathbf{x}\_{\rm CO,FR} \cdot LHV\_{\rm CO,FR} + \mathbf{x}\_{\rm H2,FR} \cdot LHV\_{\rm H2})}{\dot{m}\_{\rm fuel} \cdot LHV\_{\rm fuel}} \tag{14}$$

Here, .*ngas*,*FR* and .*mf uel* denote the mole flow of the product gas stream and the fuel input into the FR, respectively. *LHV* is the lower heating value of the fuel (mass basis) and the gas species (molar basis) and *xi* is the mole fraction of the gas species.

#### *3.2. Reduction of OC Circulation*

One approach to obtain CLG behavior, which has been suggested by Pissot et al. [52], is reducing the amount of OC cycled through the system ( .*mOC*), hence reducing φ. This approach can be deduced directly from Equation (12). Due to the resulting lower oxygen transport to the FR, syngas formation is favored, as less oxygen for full oxidation of the feedstock is provided by the OC. The simulation results for this approach are given in Figure 4. When considering the gas composition (Figure 4a) of the streams leaving the air and fuel reactor, various trends are visible. As expected, the syngas content in the gaseous FR products increases with decreasing OC circulation rate, which can directly be attributed to the lower oxygen/fuel ratio in the FR. Consequently, steam and CO2 formation decrease. Yet, it has to be noted that substantial syngas concentrations are only attained for φ < 1, which requires significant reductions in the OC circulation rate, when compared to CLC, where OC-to-fuel equivalence ratios as high as 8 [27] and 25 [40] are reported in literature for solid and gaseous fuels, respectively. For the gas concentrations leaving the AR, a strong impact of φ on the effluent oxygen is visible. As the inlet air mass flow was not varied (λ = 1.2), this observation is clear, as less O2 is removed from the gas stream due to the lower OC circulation for φ < 1. Furthermore, the CO2 content in the AR product is predicted to be insignificant, indicating a complete char conversion, which is expected in chemical equilibrium. When considering Figure 4b, showing the solid composition after the fuel and air reactor, it can be seen that the OC leaves the AR and FR in a fully oxidized (Fe2O3 + TiO2) and reduced (FeTiO3) state, respectively for φ < 1, whereas the OC is only partially reduced (indicated through the presence of Fe3O4) in the FR in case φ exceeds unity. Hence, the fraction of FeTiO3 leaving the FR strongly increases with decreasing OC circulation, signifying a higher degree of reduction of the OC, due to the lower oxygen availability. As expected one consequently obtains chemical looping combustion behavior (see Figure 3a) for oxygen-carrier-to-fuel equivalence ratios greater than unity (φ > <sup>1</sup>), whereas chemical looping gasification behavior (see Figure 3b) is attained for φ < 1.

**Figure 4.** Simulation results for CLG operation through reduced oxygen carrier (OC) circulation. Dry molar gas composition (**a**) and molar solid composition (**b**) as a function of φ for varying OC circulation rates (λ = 1.2).

*Appl. Sci.* **2020**, *10*, 4271

Based on these findings, one can conclude that a successful shifting from CLC to CLG for a given air-to-fuel ratio can be attained through a reduction in the OC circulation, which can also be seen in Figure 5a, showing a linear dependence between the two parameters. This means that for a change of φ from 1.0 to 0.5, the OC circulation rate has to be halved. However, lower solid circulation rates also result in a proportional decrease in the heat transport from the AR to the FR and hence a drop-o ff in FR temperatures [35,58]. While a moderate decrease in fuel reactor temperatures with decreasing OC circulation rate is visible for φ > 1, for which complete feedstock conversion is attained in the FR, this decrease becomes more prominent for φ < 1, where gasification reactions in the FR are dominant, hence increasing the endothermicity of reactions occurring in the FR. Consequently, FR temperatures fall below 800 ◦C for φ < 0.5, where the availability of circulating OC material for sensible heat transport between the FR and AR is halved, when compared to φ = 1 and more importantly the syngas content in the FR products is significant (see Figure 4a). This increase in syngas content also goes in hand with a decrease in the total net heat release from the CLG process ( . *Qnet*), which can be calculated from the di fference in the enthalpies of the streams entering (*in*) and leaving (*out*) the air and fuel reactor (see Equation (15)), as the enthalpy of the FR products increases.

**Figure 5.** Simulation results for CLG operation through reduced OC circulation. OC-to-fuel ratio as a function of the OC circulation rate (**a**). Fuel reactor temperature (**b**), relative net process heat (**c**), and cold gas e fficiency (**d**) for di fferent values of φ (λ = 1.2).

The decrease in net process heat release with decreasing φ, indicating the retaining of chemical energy in the FR products, also becomes visible upon consideration of Figure 5c, depicting the relative net heat release of the process for the di fferent OC-to-fuel ratios. For the given boundary conditions, an autothermal process, for which syngas yields are maximized without relying on external heat addition ( . *Qnet* = 0) is attained for an OC to fuel ratio of approx. 0.5. The resulting cold gas e fficiency for this operating point amounts to approx. 60% (see Figure 5d) at a FR temperature of 775 ◦C. Although the equilibrium model predicts full char and volatile conversions for these temperatures (see Figures 4a

and 5d), char, volatile, and tar conversion are known to be kinetically governed processes in chemical looping systems [25,55,56,59], leading to product compositions deviating strongly from equilibrium composition [35,52]. Due to this reason, temperature di fferences in the range of 50 to 100 ◦C are generally targeted in dual fluidized bed gasification [16], in order to obtain su fficiently high gasifier temperatures, allowing for decent char, volatile, and tar conversions. Accordingly, FR temperatures in the range of 850–950 ◦C are desired in CLG, in order to attain high carbon capture e fficiencies and cold gas e fficiencies as well as low syngas tar loads [20,23,37,60,61].

These considerations underline that, although the desired reduction in φ is possible, attaining an efficient autothermal CLG process through a reduction in the OC circulation rate is not a recommendable strategy as it entails low fuel reactor temperatures, due to the dual-purpose of the OC circulation (i.e., oxygen and heat transport). Consequently, alternative approaches, allowing for a decoupling of oxygen and heat transport between the AR and FR and hence increased FR temperatures are required, in order to attain a CLG process exhibiting the desired characteristics.

#### *3.3. Dilution of OC with Inert Bed Material*

One strategy allowing for a decoupling of oxygen and heat transport between air and fuel reactor, which has been discussed in literature, is employing a mixture of an active OC material and a solid inert species (e.g., sand) [35,37,52]. Here, the inert fraction serves purely as a heat carrier, transferring sensible heat between the two reactors, without participating in the occurring reactions, while the active OC fraction fulfills its dual purpose of oxygen and heat transport. Consequently, this approach is a combination of CLG and dual fluidized bed gasification, which solely employs inert bed materials for heat transport. Following this logic, Ge et al. [37] found that through accurately tailoring the mixing ratio of inert silica sand and hematite, serving as an OC, FR temperatures can be stabilized at elevated levels (i.e., >900 ◦C), while at the same time ensuring a controlled oxygen transport to the FR, resulting in large syngas yields.

In terms of the impact of the variation in OC-to-fuel ratio on gas compositions achieved through this dilution of the OC material with an inert, similar observations are obtained (see Figure 6a). This means syngas formation increases steadily for φ < 1. Moreover, the OC carrier composition, shown in Figure 3b, follows similar trends as observed for a plain reduction in the OC circulation rate (see Section 3.2), with a fully reduced OC leaving the FR for φ < 1 (see Figure 3b), whereas only partial reduction is observed for φ > 1 (see Figure 3a). Yet, the fraction of active OC material clearly decreases with decreasing φ, due to the dilution with silica sand.

As the total amount of circulating solids is kept constant, the mass of circulating OC material is inversely proportional to the dilution factor. This means that there exists a linear relationship between the solid fraction of the inert material (*zSiO*2) and φ, which is visible in Figure 7a. Hence, for a given solid circulation rate, shifting from CLC to CLG can be attained through increased inert dilution. The positive e ffect of inert addition on FR temperatures becomes apparent upon consideration of Figure 7b. In contrast to a direct reduction in the OC circulation rate, the substitution of a fraction of the active metal oxide with an inert heat carrier allows for a sustaining of FR temperatures above 980 ◦C even for OC to fuel ratios as low as 0.5. Due to this increase in FR temperatures, the average temperature of the CLG process increases, leading to a slightly increased φ of approx. 0.55 for which autothermal operation is attained (see Figure 7c) (Higher process temperatures increase the heating demands of the educts entering the FR and AR and hence reduce the OC-to-fuel ratios for which autothermal operation can be obtained). Therefore, the cold gas e fficiency obtained for autothermal operation for the given approach is also marginally reduced (see Figure 7d), when compared to the approach discussed in Section 3.2. Yet, it has to be noted that due to the intensified heat transport between the AR and FR, significantly smaller reactor temperature gradients are required for the given approach. Consequently, AR temperatures can be lowered without jeopardizing char conversions in the FR, thus reducing average process temperatures and allowing for strongly increased cold gas efficiencies (see also Section 3.5). Another advantage of this approach is that a catalytic material, not participating in oxygen transport (e.g., olivine), could be employed for OC dilution instead of sand, allowing for improved syngas characteristics with regard to tar content.

**Figure 6.** Simulation results for CLG operation through OC dilution with inert SiO2 sand. Dry molar gas composition (**a**) and molar solid composition (**b**) as a function of φ for varying OC circulation rates (λ = 1.2, .*mOC* + .*mSiO*2 = *const*.).

**Figure 7.** Simulation results for CLG operation through OC dilution with inert SiO2 sand. OC-to-fuel ratio as a function of the inert concentration of the circulating solid mixture (**a**). Fuel reactor temperature (**b**), relative net process heat (**c**), and cold gas efficiency (**d**) for different values of φ (λ = 1.2, .*mOC* + . *mSiO*2 = *const*).

Despite the presented advantages, Larsson et al. [35] found that, albeit slightly reducing tar loads, the addition of an active OC (ilmenite) to an inert circulating bed material in a dual-fluidized bed gasifier (for φ < 0.2), entails a continuous drop in cold gas e fficiency. This was explained by the fact that ilmenite addition does not enhance char conversion significantly, while its presence leads to a partial oxidation of the product gas. On the other hand, Pissot et al. [52] found that dilution of an active OC bed with up to 90% of an inert material does not entail visible enhancements in the cold gas e fficiency of the CLG process, while it has a visible negative impact on carbon conversion. This shows that the mixing of an inert and an active OC material can have di fferent e ffects on the process depending on the governing boundary conditions. Another drawback of this approach is that, albeit the addition of solids allows for an adjustment of φ during operation, it leads to a large system inertia, making it an arduous task to quickly react to disturbances. Moreover, a fraction of the solid material has to be removed from the system for ash removal in a continuously operated CLG unit. Economic considerations require a separation of these materials for further processing, recycling, and disposal. Clearly, the presence of a third component (i.e., sand, olivine) further complicates this task. Lastly, it is known that the operation of a fluidized bed with multiple bed materials of di fferent characteristics brings about additional challenges in terms of material fluidization, entrainment, and attrition, as well as bed segregation [62]. Due to these reasons it was also suggested to employ materials of a low oxygen transport capability ( *RO*), such as LD-slag, containing a large inactive fraction not participating in the oxygen transport, which fulfills the purpose of the inert heat carrier [52]. Through this, oxygen carrier circulation rates providing su fficient heat transport between the reactors can be targeted, without obtaining OC-to-fuel equivalence ratios above unity. Yet, for this approach the main challenge is finding suitable OC materials exhibiting an oxygen transport capability in the desired range, high activity towards hydrocarbon conversion, and good chemical and mechanical stability.

#### *3.4. Reduction of Air-to-Fuel Equivalence Ratio*

To allow for a less restricted material selection and avoid solid inert addition, an alternative strategy to decouple oxygen and heat transport between the AR and FR is required. In order to achieve this, Larson et al. [35] suggested the deployment of a secondary system in which the OC is pre-reduced before entering the FR. This means that, as shown in Figure 3c, a partially reduced OC enters the FR (*Xs* < 1), thus entailing a lower OC-to-fuel ratio (see Equation (12)). Instead of employing a secondary reactor to accomplish this, one can also operate the AR in a sub-stoichiometric fashion (λ < 1), thereby preventing full re-oxidation of the OC in the AR. This means that in order to attain CLG conditions, the amount of air fed into the air reactor can be reduced, while retaining a constant OC circulation. As a consequence, the OC steadily reaches a lower degree of oxidation, hence lowering its oxygen release in the FR, until steady state is reached (more details see Appendix B). This approach has already been pursued in a 140 kWth chemical looping reforming unit, employing methane as a fuel [44]. The suggested concept becomes more lucid when considering the simulation results shown in Figure 8. Clearly, the amount of fully reduced ilmenite leaving the air and fuel reactor increases when decreasing the air input into the AR for φ < 1 (see Figure 8b). While the same is true for the solids leaving the FR for all presented CLG approaches, a strong increase in the FeTiO3 and Fe3O4 content in the AR products is obtained when reducing λ below unity. This can be explained by the fact that the oxygen available in the air reactor is insu fficient to fully re-oxidize the OC, signified through an O2–free product gas from the AR for φ < 1 (see Figure 8a). Consequently, a pure stream of N2 containing small concentrations of Argon and other minor compounds is produced in the AR [44]. Since substantial quantities of OC are cycled through the system in a fully reduced state, they e ffectively act as an inert, meaning that they transfer sensible heat, but do not participate in the occurring chemical reactions through oxygen release and uptake. However, in practice the reduced OC could potentially function as a catalytic site for tar cracking and methane reforming and favor the formation of syngas [32–35], thereby enhancing the process characteristics. Another advantage of the given approach is that an undiluted OC can be employed, which simplifies the required solid-gas and solid-solid (ash-OC-char) separation and the

operation of the CLG unit with regard to the fluidization behavior. Moreover, the net heat duty of the process can be tailored promptly and easily through an adjustment of the air flow to the AR, allowing for quick responses to disturbances (e.g., variations in feedstock composition).

**Figure 8.** Simulation results for CLG operation through reducing λ. Dry molar gas composition (**a**) and molar solid composition (**b**) as a function of φ ( .*mOC* = *const*.).

The impact of the air-to-fuel equivalence ratio (λ) on φ is shown in Figure 9a. In CLC mode (λ > 1), where full OC oxidation is achieved in the AR (i.e., *Xs*,*AR* = 1), φ assumes a constant value, given by the amount of oxygen which is transported by a fully oxidized OC for a given circulation rate, regardless of the deployed air-to-fuel ratio (see Equation (12)). In contrast, lowering λ to values below unity to attain CLG operation means that φ and λ are equal, as the oxygen transport to the FR is limited by the oxygen availability in the AR:

$$\phi = \begin{cases} \lambda & \text{for } \lambda < 1\\ \frac{R\_{\text{OC}} \cdot \dot{m}\_{\text{OC}}}{\dot{m}\_{\text{O} \text{atwh}}} = \text{const.} & \text{for } \lambda \ge 1 \end{cases} \tag{16}$$

The discontinuity of this relation for λ = 1 can be explained by the fact that when surpassing this value, a transient shift from CLC (see Figure 3a) to CLG (see Figure 3c) behavior (or vice versa) occurs, which goes in hand with a continuous decrease (resp. increase) in the oxidation degree of the oxygen carrier, before steady state sets in (more details see Appendix B).

**Figure 9.** Simulation results for CLG operation through reducing λ. OC-to-fuel ratio as a function of the air-to-fuel equivalence ratio λ (**a**). Fuel reactor temperature (**b**), relative net process heat (**c**), and cold gas efficiency (**d**) for different values of λ ( .*mOC* = *const*.).

In terms of FR temperatures, Figure 9b shows that the given approach leads to a successful retaining of FR temperatures above 900 ◦C, even for φ-values as low as 0.4, due to the transportation of sensible heat by the OC. Moreover, the given approach yields more beneficial results in terms of the process heat balance, which can be seen in Figure 9c. Clearly, autothermal CLG operation is attained for φ = 0.37, which means cold gas efficiencies exceeding 70% can be achieved (see Figure 9d). This is the case as in contrast to the previous approaches (see Sections 3.2 and 3.3), the AR is not operated in air excess during CLG operation, reducing the loss of sensible heat through the AR off-gases. This means that if one would reduce the air feed to the AR to the minimum extent required for full OC re-oxidation for the CLG approach employing inert dilution (see Section 3.3), enhanced cold gas efficiencies could be attained. Nonetheless, the given approach clearly shows advantages in terms of process control due to its flexibility, the possibility of freely selecting a suitable OC material (i.e., no specific limits on *RO*), without having to consider material mixtures, and the availability of a catalytically active reduced OC material, instead of an inert solid, cycling through the system. Moreover, the chemical strain on the OC material is reduced as the change in oxidation degree for each redox cycle is lower, when compared to the former approaches, relying on full reduction and oxidation in the FR and AR, respectively (see Figure 3b,c), which should have beneficial effects on the OC lifetime.

However, one issue that might arise due to the operation of the AR in an sub-stoichiometric fashion is related to the fact that during operation a fraction of the feedstock char leaves the FR unconverted and hence travels to the AR with the circulating OC material [23,26,27]. This so called "carbon slip" leads to competing reactions between the OC material and the residual char, in case the AR is operated with λ < 1. Yet, simulations show that in an oxygen deficient atmosphere carbon conversion is favored to OC re-oxidation in chemical equilibrium. Moreover, CO formation shows to be negligible (more details see Appendix C). Due to the fast kinetics of both char conversion and OC re-oxidation, it can be expected that equilibrium-like conditions are attained in the AR and hence all residual char is fully oxidized to CO2 in the AR. This hypothesis is also supported by chemical looping experiments in small scale fixed bed reactors, during which it was established that in the beginning of the re-oxidation stage oxygen preferentially reacts with deposited carbon before re-oxidizing the OC [21,63,64]. Nonetheless, experiments showing that this is also the case in a continuously operated CLG unit and that CO formation is negligible are required to establish that full char conversion without substantial CO formation in the AR can be attained for this approach. Another issue related to this approach is the potential deep reduction of the OC, which could potentially entail problems related to intensified OC attrition or bed agglomeration. Although the process model does not predict substantial formation of deeper reduction stages (e.g., FeO) in the FR, such phases, related to bed agglomeration, have been found to be formed in CLC under highly reducing conditions [51,65,66]. Therefore, the gravity of this issue should be further investigated in experimental studies.

#### *3.5. Optimizing CLG E*ffi*ciency*

In the previous section it was established that OC-to-fuel equivalence ratios smaller than unity are required in the FR. Moreover, it was demonstrated only when decoupling heat and oxygen transfer between the AR and FR, φ < 1 and FR temperatures above 850 ◦C can be obtained for an autothermal CLG process. Thermodynamically speaking, it does not make a di fference how this decoupling of heat and oxygen transport is attained, which is why the following considerations will focus on the CLG approach presented in Section 3.4, employing a reduction in the air-to-fuel equivalence ratio to achieve CLG behavior.

When optimizing gasification processes, the trade-o ff between maximizing the carbon conversion in the gasifier and at the same time attaining high cold gas e fficiencies is at the core of many optimization strategies. This is also the case in CLG, where η*CGE* = 1 and complete char conversion is desired, ye<sup>t</sup> not attainable. While large carbon capture e fficiencies are obtained in cases where the char is gasified in the fuel reactor to a large extent, which is promoted by high FR temperatures [20,23,37], large steam/biomass ratios [20,37], and high OC-to-fuel ratios (if su fficient char residence times are provided) [27,52], cold gas e fficiencies are maximized by the minimization of the oxidation of H2 and hydrocarbons in the FR [35]. Although full oxidation of syngas in the FR should be limited to achieve large CGEs, formation of steam and CO2 in the FR is required to a certain extent to obtain autothermal CLG conditions. The degree to which this formation of fully oxidized gas species is required is determined by the criterion of the CLG process being in heat balance ( . *Qnet* = 0). This means that the heat release attained through full feedstock oxidation has to balance the heat demand of pre-heating of all inlet streams to the given reactor temperatures, the heat of reaction for endothermic gasification reactions, and the heat losses of the CLG unit. This has also been shown in the previous sections where despite assuming chemical equilibrium (i.e., full feedstock conversion), cold gas e fficiencies deviating strongly from unity were obtained for autothermal boundary conditions (see Figures 5, 7 and 9).

Therefore, one approach to enhance the cold gas e fficiency in CLG is a reduction in the inlet gas flows entering the air and fuel reactor. Since the air mass flow entering the AR is required to control φ, this leaves the steam mass flow entering the FR as a free variable which can be altered to enhance cold gas e fficiencies. The e ffect of a reduction in the steam to biomass ratio on the net heat release of the process is shown in Figure 10a. It is visible that, with a decreasing steam to biomass ratio, the air-to-fuel equivalence ratio for which an autothermal process is attained decreases. Due to the direct correlation between the oxygen availability and cold gas e fficiency in CLG (see Figure 10b), this also means that the CGE obtained for autothermal operation increases with decreasing steam/biomass ratio, so that the CGE is raised from 72.5 to 77.1%, when decreasing the steam/biomass ratio from 0.9 to 0.3. However, it is obvious that the reduction of the steam to biomass ratio would also entail a drop in carbon capture efficiencies of the process, as less steam is available for char gasification and the kinetic inhibition effect of syngas increases with decreasing steam concentrations (entailing larger syngas partial pressures) in the FR [8,12,67,68]. This becomes most obvious for a steam to biomass ratio of 0, for which char conversions in the FR would be diminutive in a real gasifier, due to the slow kinetics of heterogeneous solid-solid OC-feedstock reactions [67–69]. As this drop in char conversion is not predicted by the equilibrium model, the negative effect on process efficiency with decreasing steam to biomass ratio cannot be evaluated in this study. However, sufficient steam availability clearly is a prerequisite in CLG, when targeting large char conversions and hence carbon capture efficiencies.

**Figure 10.** Net heat release and cold gas efficiency for CLC/CLG process as a function of the air to fuel equivalence ratio for different steam to biomass ratios (**<sup>a</sup>**,**b**), OC circulation rates (**<sup>c</sup>**,**d**), gas inlet temperatures (**<sup>e</sup>**,**f**), and air reactor temperatures (**g**,**h**). Circles mark the cold gas efficiency for autothermal CLG operation ( .*mOC* = *const*., so φ = λ for λ < 1 and φ = *const*. > 1 for λ > 1).

Another possible measure to enhance CGEs are variations in the circulation rate of the OC, which is shown in Figure 10c,d. Clearly, larger solid circulation rates enhance the heat transport between the reactors and hence entail higher FR temperatures [16]. However, due to material attrition, solid loss, which necessitates continuous make-up feeding, also scales with the circulation rate. As shown in Figure 10d, the effect of this material loss on the process heat balance is comparatively small, thus its effect on the cold gas efficiency is low. However, the model predicts an increase in FR temperatures

from 892 to 951 ◦C, when increasing the circulation rate from 6.3 to 10.6 t/h. This means that generally, large solid circulation rates are desired in CLG units, as large FR temperatures are beneficial for volatile and char conversion [20,23,37]. Yet, it has to be kept in mind that the solid circulation in dual fluidized bed systems requires solid entrainment from the fluidized bed riser, which can be increased through an increase in gas velocities (i.e., increase in steam/biomass ratio), smaller particle diameters or smaller reactor diameters [16]. Moreover, intensified solid circulation also increases the occurrence of a "carbon slip" to the AR, due to the lower residence times of the char particles in the FR [27,28,70]. This means that the OC circulation rate can only be varied within a given range.

Increasing the inlet temperature of the steam and air entering the FR and AR respectively, thereby decreasing the heat demand for heating up of the gases inside the reactor, is a further strategy to boost cold gas e fficiencies. As shown in Figure 10e, this approach allows for a reduction of the air-to-fuel equivalence ratio from 0.38 to 0.34 when increasing inlet temperatures from 400 ◦C to 600 ◦C. Hence, maximizing inlet gas temperatures through heat recuperation is a key task in CLG in order to optimize the process e fficiency, which is illustrated by the increase in the CGE from 68.3 to 76.5%, when increasing gas inlet temperatures from 300 to 600 ◦C (see Figure 10f). Due to the absence of corrosive compounds and the high process temperatures, the hot o ff-gases leaving the AR are ideal for steam generation and heat recuperation. On the other hand, special syngas coolers are being used to recuperate sensible heat from syngas streams for steam production [71–73], highlighting that e fficient gas pre-heating using heat from process o ff-gases is possible in CLG.

Furthermore, variations in the AR temperatures can be considered, in order to enhance CLG process e fficiencies. Generally speaking, a reduction in average process temperatures is beneficial for the process heat release, as pre-heating demands for all educts (i.e., inlet gases & feedstock material) are being reduced as a consequence, thus allowing for intensified heat extraction for a given air-to-fuel ratio (see Figure 10g). As visible in Figure 10h, a slight increase in the CGE by 2.4 percentage points can be attained for autothermal CLG operation when lowering AR temperatures from 1050 to 1000 ◦C. Yet, it has to be kept in mind that in chemical looping processes, air and fuel reactor temperatures are coupled, which means that a drop in FR temperatures is an inevitable e ffect of reduced AR temperatures. For the given boundary conditions, FR temperatures are projected to directly correlate with AR temperatures, which means that for the given reduction in AR temperatures from 1050 to 1000 ◦C, a corresponding drop in FR temperatures from 928 to 880 ◦C entails. This means that when attempting to prevent the ensuing drop in FR temperatures, related to negative e ffects on volatile and carbon conversion, OC circulation rates have to be increased accordingly as a counter-measure.

Although these insights allow for a first glimpse on process optimization approaches, it becomes clear that a detailed consideration of reaction kinetics and reactor hydrodynamics is quintessential, when aiming for a holistic optimization of the CLG process, as both phenomena have a pronounced e ffect on the process parameters. As it is well known that the conversion of char and other hydrocarbons is kinetically governed [25,55,56,59], the impact of reactor temperature, residence time, and gas concentrations on reaction kinetics need to be established in detail, allowing for accurate predictions of the governing reactions in a realistic environment. Moreover, reactor hydrodynamics are a crucial factor in chemical looping systems [74,75], making it a pre-requisite to consider them in advanced CLG process models. Through considering these phenomena, it thus becomes feasible to assess to which extent the preceding approaches can be utilized to obtain a CLG process exhibiting not only a high cold gas e fficiency, but also excellent carbon capture e fficiencies. Nonetheless, the preceding explanations o ffer valuable insights on the fundamental challenges associated with the autothermal CLG process, which require catering to, when implementing the technology in large scale.
