Development of CuO-based oxygen carriers supported on diatomite and kaolin for chemical looping combustion

Chemical Looping Combustion (CLC) technology has emerged as a promising alternative capable of restricting the effects of global warming due to anthropogenic gas emissions, especially CO2, through its inherent capture. This study aims to synthesize and evaluate Cu-based oxygen carriers supported on natural materials such as diatomite and kaolin, through the incipient wet impregnation method for CLC process applications. Oxygen carriers were characterized by X-ray diffraction (XRD), temperature-programmed reduction (TPR), and scanning electron microscopy with surface energy dispersive x-ray spectroscopy (SEM-EDS). The mechanical strength of the two oxygen carrier particles was determined after the sintering procedure resulting in high crushing force. Reactivity of oxygen carriers was evaluated in a thermobalance with CH4 and H2 gases. Different reaction pathways were attempted when undergoing the redox cycles: total direct reduction of CuO to Cu for Cu-K and partial reduction of CuO to Cu2O and CuO to Cu-D. However, the highest reactivity and reaction rate was achieved in Cu-D due to the pore structure of diatomite, the chemical composition and the resulting interaction between CuO and the support. H2 gas reactivity tests showed a higher conversion rate and greater stability between cycles for both oxygen carriers. Thus, the reducible CuO content present in Cu-Diatomite during the reactivity test with H2 as the fuel gas was ideal for achieving high solids conversion, tendency for greater stability and a higher reaction rate.


Introduction
Sustainability is one of the most important and necessary challenges for societies today. However, the climate change being suffered by the planet should be a warning to this factor. Carbon dioxide (CO2) is a long-lived anthropogenic gas in the atmosphere and there is an intensification of the effects of global warming due to its progressive emission from fossil fuel combustion processes (Takht & Saeed, 2014;Fernandes et al., 2019;Oliveira et al., 2020;Gomes et al., 2021). The Paris Agreement was established in order to limit these effects, requiring the decarbonization of the world's energy systems, limiting the average global temperature increase to 2 °C for the next century. In order to achieve this goal, Chemical Looping Combustion (CLC) has emerged as a promising technology for CO2 capture in power plants and industrial applications with low energy penalty compared to other competing CO2 capture and storage (CCS) technologies (Adánez-Rubio et al., 2018;McGlashan et al., 2012;Wang, Yan et al., 2018). The Figure 1 show the CLC technology. This system describes cyclic redox processes between two interconnected reactors and is based on the use of a metal oxide (MxOy), known as an oxygen carrier (OC), to transfer the required oxygen from the air (Reaction 1) and oxidize the fuel to CO2 and H2O (Reaction 2) to avoid direct contact between the air and fuel. Its main advantage is the inherent capture of CO2, bypassing the energy penalty. Finding oxygen carriers with good physicochemical properties is the critical point for CLC technologies. OCs require high fuel reactivity, excessive oxygen carrying capacity, high friction and agglomeration resistance, low toxicity and low cost (Zhang et al., 2019). Based on these characteristics, Adanéz and colleagues (Adanez et al., 2012) reviewed possible Ni, Fe, Cu, Co and Mn oxides and probable inert materials as supports, for example SiO2, Al2O3, TiO2, ZrO2, to improve the reactivity and life time of oxygen carriers.
The use of solid fuels such as coal and biomass has been of great interest, as coal will continue to be an important source of energy in the medium term, generating neutral emissions (Takht & Saeed, 2014). Moreover, the use of biomass waste triggers negative emissions due to the intrinsic balance of biomass combined with CO2 capture and storage techniques (Bioenergy with Carbon Capture and Storage -BECCS) Mendiara et al., 2018).
In-Situ Gasification Chemical Looping (iG-CLC) and Chemical Looping with Oxygen Uncoupling (CLOU) are proposed solid fuel CLC processes in which gas-solid reactions occur. In iG-CLC technology, solid fuel is first gasified in situ to synthesis gas (CO + H2) and this reacts with the OC in the fuel reactor . On the other hand, the CLOU process requires a suitable oxygen carrier capable of reversibly releasing O2(g), having oxygen equilibrium partial pressure at high temperatures (800-1000 °C), for example, CuO, Mn2O3 and Co3O4 Adánez-Rubio et al., 2017;Gayán et al., 2012). In this case, solid fuel is burned by gaseous oxygen which is decoupled from the OC in the fuel reactor. The main advantage of the CLOU process over the iG-CLC is that direct burning of solid fuel promotes faster combustion as it eliminates the slow coal gasification step Adánez-Rubio et al., 2013;Adánez-Rubio et al., 2017).
Cu-based oxygen carriers were reviewed and studied by (Adánez-Rubio et al., 2013;Adánez-Rubio et al., 2011;Adanez et al., 2012;de Diego et al., 2005;Forero et al., 2009), and showed high reactivity, high reaction rates, and oxygen carrying capacity, and did not show thermodynamic restrictions for complete conversion of fuel to CO2 and H2O. In combining with inert supports such as Al2O3, bentonite, MgO, MgAl2O4, SiO2, and TiO2, it is possible to achieve an OC with high mechanical stability and low friction rate. However, problems with agglomeration is a potential issue due to the low melting temperature of the Cu 0 (1085 °C). Thus, iG-CLC and CLOU can be used with different parameters to prevent this problem. In the CLOU approach, the present redox system is CuO/Cu2O according to Reaction 3 in order to avoid generating metallic Cu, so that high temperatures are used (900-950 °C) due to the Cu2O (1235 °C) melting temperature, and therefore the OC has a higher CuO content. In the iG-CLC process, the approached redox system is CuO/Cu according to Reactions 4-5, but the OC has lower CuO content (≤ 21 wt%) and low reaction temperature (≤ 850 °C) due to the melting temperature of Cu (1083 °C) . substrates with different macropore (SiO2) and mesopore (γ-Al2O3) combinations, and concluded that CuO supported on a combination of mesopores and macropores have better redox performance compared to synthetic supports with a single pore size distribution (Van Garderen et al., 2014). In this perspective, the purpose of this work is to evaluate the influence of natural supports (kaolin and diatomite) as sources of silica and alumina on the reactivity of Cu-based oxygen carriers obtained through incipient wet impregnation, aiming toward its use in CLC processes with solid fuels.

Preparation of the Oxygen Carriers
The Cu oxygen carriers used in this work were prepared by incipient wet impregnation using Cu (NO3)2.3H2O (VETEC, PA = 99%) and kaolin and diatomite natural materials collected in the state of Rio Grande do Norte, Brazil, in the particle size range 100-300 µ to be used as supports. Gayán et al. (2012) studied the incipient wet impregnation technique, which consists of obtaining a more concentrated nitrate solution of the metal to be impregnated (active phase). For this, a saturated solution of metallic nitrate at 80 ºC is prepared, where the solute is dissolved in its own hydration water, in order to obtain a greater amount of solute per unit volume. The mass of the hydrated salt is weighed in the balance, after dissolving the salt at 80 ºC we obtain the final volume and, thus, the concentration of the copper nitrate solution . Gayán et al. (2012) studied the incipient wet impregnation technique, which consists of obtaining a more concentrated nitrate solution of the metal to be impregnated (active phase). For this, a saturated solution of metallic nitrate at 80 ºC is prepared, where the solute is dissolved in its own hydration water, in order to obtain a greater amount of solute per unit volume.

5
The mass of the hydrated salt is weighed in the balance, after dissolving the salt at 80 ºC we obtain the final volume and, thus, the concentration of the copper nitrate solution . Wang et al. (2017) found that the formation of Cu 0 provides an agglomeration of Cu-based OCs in the CLC process.
However, this problem can be avoided by using a low CuO content (Wang et al., 2017). Thus, the mass fraction used was below 20%wt CuO. The impregnation procedure consisted of mechanically mixing a 5 M copper nitrate solution with the porous supports at 80 °C. Three impregnations were performed on each material to achieve the desired mass fraction, yielding a final volume of 0.42 mL and 0.30 mL of Cu(NO3)2.3H2O per gram of Kaolin and Diatomite, respectively. There was a heat treatment at 550 °C/1 h after each of the first two impregnations, but the samples were calcined at 1100 °C/1 h at the end of the last impregnation with the objective of achieving high mechanical strength. The composition of calcined Cu samples was determined by X-ray Fluorescence. Table 1 shows the XRF analysis results of the supports and OC, including their main properties.

Characterization of the Oxygen Carriers
Physical and chemical characterization was performed on oxygen carrier particles. The chemical composition of the samples was determined by X-Ray Fluorescence on a Shimadzu Rayny 720 targeting an Rh anode, 50 kV voltage, Si/Li detector. The identification of the crystalline chemical phases was performed by X-ray Diffraction (XRD) using a Shimadzu XRD-7000 X-ray diffractometer with Cukα radiation (λ = 1.5409 Å). The Joint Committee on Power Diffraction Standards was used to designate the crystalline phases. The structure refinement was performed by applying the Rietveld treatment using MAUD software.
The temperature-programmed reduction (TPR) profile of the OCs was evaluated on a Micromeritics AUTOCHEM II 2920 equipped with a TCD (Thermal Conductivity Detector). The analyzes were conducted by varying the temperature from In which: represents the consumption of H2 per mol, is the experimentally consumed volume of H2, moxi and mred is the oxygen carrier mass when fully oxidized and reduced, respectively, ROC is the oxygen carrying capacity of materials, XOC is the fraction of the active phases present in the carriers. The mechanical strength of the particles was determined using a Shimpo FGN-5X dynamometer by averaging 20 measurements of the force required to fracture the particle.
The microstructure and distribution of the CuO phase within the particles was analyzed by scanning electron microscopy (SEM) on a Zeiss DSM 942 microscope equipped with an Oxford Link-Isis X-ray Dispersive Energy Analyzer (EDX).

Oxygen Carrier Reactivity Test
Reactivity tests of oxygen carriers were performed on a thermogravimetric analyzer (TGA) (CI Eletronics), with experimental setup shown in Figure 2. In the experiments, 50 mg of the sample was placed in a platinum mesh basket and introduced into a reactor in the form of a quartz tube arranged in an oven operating at a temperature of 950 °C. Upon reaching the operating temperature and system stability, the samples were subjected to the desired reduction and oxidation conditions, alternately and to preserve each step, avoiding reactive gas mixing, and a N2 flow was introduced for 2 min at the end of each reaction. CH4 and H2 were used as reducing gases, synthetic air as oxidizing gas and nitrogen to purge the system, kept at a flow of 25 nL/h throughout the reaction time. The gas composition for the reduction process was 15% of CH4, 20% H2O, 65% N2 or 15% H2, 85% H2O, and the gas used for the oxidation process was 100% air. The temperature and weight of the sample were recorded continuously on a computer. It is also important to have a nitrogen flow of 9 nL/h to the head in this system, keeping it free of reaction gases, preventing corrosion of the thermobalance electronics. Figure 3a and b show the conversion of oxygen carrier as a function of time over three reactivity test cycles in TGA with hydrogen (15% H2 + 85% N2) and methane (15% CH4 + 20% H2O +65% N2) as fuel.

Reactivity of the OCs with CH4 and H2 in the TGA
The conversions obtained by Cu-K and Cu-D were calculated from the possible redox reactions involved (reactions 6-11), according to their oxidation degree and in terms of sample weight variation, as shown in Figures 4a and b As seen in Figure 4, the largest mass variation for both fuels was obtained with copper oxide supported on diatomite (Cu-D) particles. With CH4 as fuel, it is observed that there is an increase in mass variation over the cycles for Cu-K. It is also noticeable that the two oxygen carriers showed greater constancy during the three cycles with H2 as fuel. It is reported that the reduction reactions occurred in two steps in both samples; the first when there is only N2(g) atmosphere, which involves reactions of the CLOU process, causing the transformation of CuO → Cu2O, and the second when there is the supply of H2(g) or CH4(g) as combustible gases. These reactions with the reactive gases are equivalent to the reactions of the iG-CLC process, causing the reduction CuO → (Cu2O) → Cu 0 . Note that the mass loss in the first step related to the CLOU process was higher for Cu-D. Therefore, the experiment was performed at 950 °C due to the Cu2O melting temperature (1235 °C), which is formed during the inert period between oxidation and reduction.  The conversion rate calculated for Copper oxide supported on diatomite and kaolin (Figures 3a and b) shows that there is an increase in reactivity during the three redox cycles, but stability after the third cycle cannot be stated. Early reductions have a slower reaction rate depicting activation of the materials. This is due to the resulting changes in the structure of the fresh oxygen carriers during the first reduction reaction, with the formation of different copper oxides after regeneration in the oxidation step (Wang et al., 2017). The higher reactivity presented by Cu-D with CH4 and H2 compared to Cu-K may be due to the porous structure of diatomite, contributing to the reduction of CuO particle agglomeration, leaving them firm on the surface of this support, and consequently increasing the surface area and its oxygen release capacity. In contrast, CuO particles in Cu-K were free to migrate on the surface, and the three successive cycles show greater differences in reactivity in search of stability.
Through the different inclinations, it is also possible to observe that the CuO redox process which occurs in diatomite is faster than in the kaolin support. As the microstructure of these two supports is similar, the resulting effect on reactivity and reaction rate is primarily based on variation in chemical composition and the interaction between CuO and the support . But oxygen carriers generally exhibited a higher reaction rate for reduction compared to the reaction rate for oxidation, as noted by Pio et al. (2017), probably due to the diffusive effects inside the oxygen carrier particles. Moreover, a higher reaction rate for gaseous fuel is obtained with H2 than with CH4 in the same concentration (15%). This can be explained by the reaction stoichiometry, which has a stoichiometry of 1:1 with H2, while it is 1:4 with CH4 (Pio et al., 2017).
The evolution of oxygen carrying capacity was evaluated over time for the three cycles, according to Table 2. In Table 2 is observed that there are small variations with an increase of Roc at each cycle. These values correspond to the amount of oxygen required to completely convert the fuel into CO2 and H2O. Wang et al. (2017) investigated the Roc of an oxygen carrier prepared via the impregnation method with a 40%wt. copper oxide supported on natural porous material (olivine -CuO/Olivine), achieving a maximum value of 3.59% in its tenth cycle of the CLC process, concluding that CLC tests weakly affected oxygen carrying capacity (Wang et al., 2017). This result is higher than Cu-K Roc, but lower than Cu-D in the third cycle, emphasizing that there is a lower percentage of CuO impregnated in these carriers.

Characterizations of the oxygen carriers
The XRD patterns corresponding to the in natura, kaolin and diatomite supports are shown in Figure 5. According to Figure 5, the kaolin sample showed main crystalline phases for kaolinite (Al4(OH)8(Si4O10) -ICSD 063316), with the main peak at 25º, and silica (SiO2 -ICSD 062404) with the main peak at 26.57°. In the diatomite sample, the silica phase (SiO2 -ICSD 062404) was also detected with considerably higher intensity than the alumina (Al2O3 -ICSD 082504) and aluminosilicate phases (Al2Si2O5(OH)4 -ICSD).
An X-ray Diffraction analysis and fresh oxygen carrier patterns reduced to 10% H2 gas and oxidized (after the third thermobalance cycle) are shown in Figure 6 to understand the reaction pathways of Cu oxide supported on natural materials (kaolin and diatomite).
The diffractograms of the fresh oxygen carrier particles ( Figure 6) revealed the presence of CuO (ICSD 087124) as the main active phase with higher intensity peaks at 35.54° and 38.69°. The inert SiO2 (ICSD 062404) and Al2SiO5 (ICSD 100451) phases were also identified in Cu-K OC. The calcination temperature (1100 °C) to obtain this oxygen carrier caused the kaolinite to dehydrate, converting it to matecaulinite (Al2O3.2SiO2) when it reached temperatures above 510°C; while it favored the formation of Al2SiO5 aluminosilicate above 900 °C, according to reaction 12. It is observed that the Cu-K sample did not show interaction phases between CuO and SiO2 present in kaolin, otherwise this interaction could result in a reduction of the oxygen carrying capacity (Roc) of this material (Song et al., 2014). Inert phases were also identified as SiO2 (ICSD 062404) and Al2O3 (ICSD 082504) in the Cu-D OC diffractogram, and the reaction temperature to obtain this oxygen carrier did not cause the formation of copper silicates and aluminates, which could result in a negative effect on their reactivity in CLC processes. These natural inert supports act to increase the mechanical strength, maintaining the porous structure of the particles at high temperatures, and the aluminosilicate can give a high modulus of rupture and chemical stability to the Cu-K oxygen carrier Ma et al., 2017).
The diffractograms of the oxygen carrier particles reduced to 10% H2 gas indicate that the Cu-K particles underwent uniform reduction reactions and all CuO was reduced to Cu 0 with a peak peak of 43.50°, conforming to reaction 13. In fact, it does not exclude the possibility that Cu2O exists as an intermediate phase in CuO reduction phase transformations under CLC conditions. However, the diffractogram of the reduced Cu-D particles shows that this reducing condition was not sufficient for the complete reduction of CuO to Cu2O and/or Cu 0 . Table 3 shows the percentage of the crystalline phases of the carriers reduced to 10% H2 through the Rietveld refinement and proves the conversion of each phase.
CuO + H2 → Cu (Reaction 13) After the third thermobalance cycle in the reoxidation step, in the diffractograms it is noted that the reduced Cu-K OC particles did not present the fully oxidized copper oxide (CuO) phase. For the Cu-D OC, a mixture of Cu2O and CuO with 4.90%wt. and 11.66%wt. was identified respectively, according to the data from the Rietveld refinement presented in Table 3.
However, the reactivity tests show increased reactivity with increasing number of cycles, so Cu2O reoxidation does not negatively affect its conversion. This fact may be related to the high reduction rate providing a good final Cu2O phase conversion (Pio et al., 2018). In addition, according to the concentration of the crystalline phase obtained by the refinement (as shown in Table 3), different percentages of reducible Cu under CLC conditions present in the samples can refer to an activation step of the materials in the first cycle, so that only the three cycles cannot indicate reliable stability. Table 4 shows the mechanical strength of the Cu-D and Cu-K particles after being calcined at 1100 °C. According to Adanez et al. (2004), the resistance in Newton (N) of the particles depends on the type of metal oxide used as the active phase, its concentration, the support used and the sintering temperature . High sintering temperatures generally increase the strength of oxygen carriers. However, this temperature should be limited for copper as it may cause it to melt. The Table 4 reports that copper oxide exhibits appreciable mechanical strength when using SiO2 as a support.
According to Johansson et al. (2004), oxygen carriers with mechanical strength less than 1.0N are unsuitable for use in CLC processes (Johansson et al., 2004).
Thus, it is possible to evidence that the optimization of the mechanical strength of Cu-D and Cu-K particles is the result of the presence of aluminosilicates acting as a support and control for the percentage of active phase impregnated (less than 21%wt. of CuO), favoring the calcination of the OCs at 1100 °C. Cu-D has higher mechanical strength because it has a higher CuO concentration compared to Cu-k. Therefore, both Cu-based oxygen carriers have excellent mechanical strength to be applied in chemical recirculation processes.  Figure 7 shows the H2-TPR profiles of Cu oxygen carriers. It is observed that there is a single apparent peak in the temperature range of 190-380 °C during the reduction of Cu-K and Cu-D particles. This peak can be attributed to the reduction of a single Cu species (Cu 2+ → Cu 0 ), in accordance with the symmetrical characteristic of the peaks, with a maximum consumption around 240 °C. H2 consumption was more intense in Cu-D (43.84 cm 3 /g) compared to the reduction of Cu-K particles (27.77 cm 3 /g) due to the higher percentage of copper oxide present (20.34%), and its wider peak may be associated with a more difficult reduction.
The surface and particle morphology of Cu-K and Cu-D oxygen carriers were assessed by Scanning Electron Microscopy (SEM), as shown in Figure 8 and Figure 9, respectively.   Wang et al. (2017) reveals that the accumulation of CuO particles gradually disappears with an increasing number of redox cycles (Wang et al., 2017). This fact contributes to the increased reactivity of oxygen carriers over the course of reaction cycles.
It was also observed (in Figure 9) that there is a lower CuO content on the Cu-D particle surface compared to the Cu-K particle surface, which induces that copper oxide has penetrated the porous structure of the support. Therefore, although Cu-D has a higher CuO content, a less amount which is present in Cu-K is available to participate in CLC reactions. However, it is important to highlight that the Cu-D reactivity test with CH4 and H2 in TGA showed that the percentage of active phase present was ideal to maintain its reactivity during the cycles presented.

Conclusion
Reactivities during the reduction and oxidation reactions of Cu-based oxygen carriers (CuO) supported on natural diatomite and kaolin materials were analyzed by thermogravimetry (TGA) and their physicochemical properties by XRD, SEM-EDS, and TPR. The impregnation of CuO particles in the diatomite and kaolin supports did not result in forming copper silicates and aluminates. Moreover, they showed appropriate mechanical strength and uniform distribution of the CuO particles on the surface of the oxygen carriers. Both showed increased reactivity and oxygen carrying capacity when submitted to three redox cycles, but their stability was not reached after the third cycle. This reveals that there are changes in the structure of the oxygen carriers during this reaction period. The regeneration of Cu-K after the third cycle with CH4 did not lead to the formation of fully oxidized copper oxide and a mixture of Cu2O and CuO was identified in Cu-D, but failure to obtain fully oxidized CuO did not affect its reactivity during the cycles. Reactivity tests with H2 gas as the fuel showed a higher conversion rate and greater constancy between cycles for both oxygen carriers. It was observed that Cu-D presented a higher reactivity and reaction rate in the reduction and oxidation steps, which can be attributed to the pore structure of diatomite, the chemical composition and the resulting interaction between CuO and the support. The percentage of reducible CuO present in Cu-D during the reactivity test with H2 as fuel gas was ideal for achieving high solids conversion, a greater tendency to stability and a higher reaction rate.
For future work, we can develop a new synthesis route, by which it will improve the mechanical resistance of the oxygen carriers studied in this work. With the improvement of the mechanical resistance of these materials we will have greater stability of the samples during the accomplishment of the multiple redox cycles.