Abstract

Chemical-looping with oxygen uncoupling (CLOU) is a novel combustion technology with inherent separation of carbon dioxide. The process is a three-step process which utilizes a circulating oxygen carrier to transfer oxygen from the combustion air to the fuel. The process utilizes two interconnected fluidized bed reactors, an air reactor and a fuel reactor. In the fuel reactor, the metal oxide decomposes with the release of gas phase oxygen (step 1), which reacts directly with the fuel through normal combustion (step 2). The reduced oxygen carrier is then transported to the air reactor where it reacts with the oxygen in the air (step 3). The outlet from the fuel reactor consists of only CO2 and H2O, and pure carbon dioxide can be obtained by simple condensation of the steam. This paper gives an overview of the research conducted around the CLOU process, including (i) a thermodynamic evaluation, (ii) a complete review of tested oxygen carriers, (iii) review of kinetic data of reduction and oxidation, and (iv) evaluation of design criteria. From the tests of various fuels in continuous chemical-looping units utilizing CLOU materials, it can be established that almost full conversion of the fuel can be obtained for gaseous, liquid, and solid fuels.

1. Introduction

In the last decade, chemical-looping combustion (CLC) has emerged as a viable and efficient alternative for combustion with carbon capture. The interest in CLC is mainly due to the intrinsic separation of from the rest of the flue gases during combustion, that is, nitrogen and unused oxygen, which means that no energy or equipment is needed for gas separation, which is in contrast to competing CCS technologies, that is, oxyfuel, postcombustion, and precombustion. An added advantage of CLC is the possibility to obtain 100% carbon capture, which is difficult to achieve in, for instance, post- or precombustion. Chemical-looping with oxygen uncoupling (CLOU) is very similar to CLC and has the same advantages with respect to efficiencies and degree of carbon capture, but the mechanism of fuel conversion is different, and this could result in significant advantages with respect to fuel conversion rates, which can have positive implications with respect to system performance. Below follows a description of both the CLC and CLOU technologies. This is followed by a simple thermodynamic analysis for some relevant metal oxides in Section 2. A detailed overview of the experimental research conducted on different oxygen carrier materials for CLOU is then carried out in Section 3. An overview of kinetic data for reduction and oxidation of CuO is given in Section 4, and Section 5 contains information with respect to the important design parameters, that is, solids inventory and recirculation rates. Finally, the paper is concluded with an outlook for the technology, including the main advantages and challenges.

1.1. Chemical-Looping Combustion (CLC)

Chemical-looping combustion (CLC) is a technology for burning a fuel without mixing the combustion air and fuel [14]. This is done by using an oxygen carrier to transfer the oxygen from the air to the fuel, as is depicted in Figure 1. In this system, the fuel is introduced to the so-called fuel reactor, where an oxygen carrier reacts with the fuel according to the overall reaction, Here and are the fully oxidized and reduced forms of the oxygen carrier and is the fuel and could be in gaseous, liquid, or solid form. The reduced oxygen carrier is transported to the air reactor, where it is reoxidized with the combustion air according to The system of reactors in Figure 1 could be designed in a variety of ways, with the most common configuration being that of coupled fluidized beds where the oxygen carrier is the bed material. In 2001, Lyngfelt et al. [2] proposed a system which resembles that of a conventional circulating fluidized bed boiler (CFB), where the air reactor is a fast fluidized bed reactor and the fuel reactor is a bubbling fluidized bed reactor, and a schematic of such a boiler can be seen in Figure 2. In case of complete conversion of the fuel, the exhaust stream from the fuel reactor consists of only and , from which pure could be obtained after condensation of water. The reduced form of the oxygen carrier, , is then transferred to the air reactor where it is reoxidized by air making it ready for the next cycle. The oxidation reaction is always exothermic while the reduction reaction can be exothermic or endothermic depending on the nature of the carrier and the fuel [5]. However, the sum of the enthalpy from reaction (1) and (2) is the same as for conventional combustion. Thus, the CLC process does not entail any direct cost or energy penalty for separation. The basic idea of chemical-looping was proposed already in the 50s by Lewis et al. [6], but it was more or less forgotten until the 1990s when the group of Ishida et al. at the Tokyo Institute of Technology revived the process and conducted both process investigations and experimental studies of different oxygen carrier materials, for example, [1, 711]. At the start of the new millennium several groups initiated research on CLC with the aim of developing a combustion process with inherent capture. The main research groups driving the developments at this time were Chalmers University of Technology in Sweden, [2, 1214] ICB-CSIC, [1518], Technical University of Vienna, [1921], and Korea Institute of Energy Research, [2225] although several other groups were also active at this time [26, 27]. Since then, the research around chemical-looping technologies has been steadily increasing, as can be gauged from the increase in the number of scientific articles published around the technology in the last decade.

The cornerstone of the CLC technology is finding oxygen carriers particles which are sufficiently reactive with fuel while at the same time having high mechanical stability over many redox cycles. The technology has been successfully demonstrated in a number of continuous units utilizing Fe- [28, 29], Mn- [30, 31], Cu- [32], and Ni-based oxygen carriers [3337]. There are a number of recent reviews available on the status of CLC, whereof the more recent ones by Fan [38], Lyngfelt [39], and Adanez et al. [40]. There are also some reviews which are a few years old, for example, Lyngfelt et al. [41], Hossain and de Lasa [42], and Fang et al. [43]. These papers give a broad view of the different materials which have been investigated for CLC and also overviews of the operational experience. Most of the research around CLC has focused on gaseous fuels, such as natural gas and syngas consisting mainly of CO and . This is not surprising, since this entails the least complex system and may also be the basis for the first full scale demonstration. When utilizing a gaseous fuel, such as , the overall reaction in the fuel reactor can be described by reaction (1). However, there could be many side or intermediate reactions in the fuel reactor, for instance, carbon deposition, water-shift reaction, and partial oxidation reactions, and it is likely that CO and are intermediate components, and small amounts may also be found in the outlet from the fuel reactor, for example, [28, 30, 37, 39, 44, 45]. It is important that oxygen carriers and conditions in the fuel reactor are employed which promote the full oxidation to and . This means, for example, (i) use of an oxygen carrier with thermodynamic properties enabling conversion of gas to carbon dioxide and steam to a high extent, (ii) sufficient reaction rates with fuel, (iii) an oxygen carrier which has a sufficient lifetime and does not break down into smaller fragments at a high rate, and (iv) use of a cheap material.

There are only a limited number of studies focusing on utilizing liquid fuels, such as oil, with chemical-looping technologies [31, 4649]. Of these, only two investigations have been performed where the fuel is directly injected into a circulating CLC system [31, 49]. Here kerosene was utilized as fuel and was evaporated prior to injection into the fuel reactor. It is likely that the evaporated fuel is reformed or decomposed to CO and in the fuel reactor before being fully oxidized to and , similar to the processes for gaseous fuel. Looking forward, it is likely that heavier oil fractions would be of greater interest in CLC, and it may be more efficient to inject the fuel directly to the fuel reactor where evaporation and reactions with the oxygen carrier would occur in parallel.

In the case of solid fuels such as coal or biomass, the reaction pattern is more complex compared to when using gaseous or even liquid fuels [50]. One way to process solid fuel via CLC is to first perform an external gasification of the fuel to syngas and then process the syngas using gaseous CLC [51, 52]. As this would entail an external gasification unit and also an ASU for oxygen production for the partial gasification, it is questionable if it is an economically viable alternative. Instead, the fuel can be directly introduced into the fuel reactor together with the metal oxide particles [3, 5355]. The fuel would here devolatilize to a gaseous fraction, that is, volatiles, and a solid char part, as is depicted in Figure 3. In parallel, the volatile part, composed mainly of CO, , and , reacts directly with the oxygen carrier particles according to reaction (1). The char component is gasified in the presence of steam and carbon dioxide, producing CO and according to reactions (3) and (4) The CO and from these reactions will then need to react with the oxygen carrier to fully oxidized products, as in reaction (1). Evidently, reactions (3) and (4) do not directly involve the oxygen carrier particles. Still, the rate of reactions (3) and (4) determines how well the solid fuel is converted to gas in the fuel reactor and governs the residence time of solids needed in the fuel reactor for full char conversion [54]. Insufficient residence time would lead to char being transferred to the air reactor where it would be burnt to carbon dioxide, thus decreasing the overall degree of capture. The functionality of the oxygen carrier is associated with how well the particles are at converting the syngas generated to and . In a well-mixed fuel reactor, the gasification reactions will take place in the entire reactor, meaning that there will be combustible components produced which have only limited contact with oxygen carrier particles, and thus it is expected that there will be CO and released from the fuel reactor even in a well-designed reactor [56]. It is also important that the volatiles have good contact with the oxygen carrier particles, which probably means that the fuel needs to be introduced at the bottom of the fluidized fuel reactor. There are now many scientific articles related to the application of CLC for solid fuels, with investigations both in laboratory reactors, for example, [54, 5759], and in continuous reactors, for example, [53, 6064]. For a complete overview of the operational experience of CLC with solid fuels, the reader is referred to the reviews by Adanez et al. [40] and Lyngfelt [50]. Although several studies of CLC with solid fuels report on successful operation in continuous units, with high degrees of carbon capture, [64] the gas conversion levels are typically less than 95% [39]. The incomplete combustion may be due to imperfect contact between oxygen carrier particles and reacting gas, in addition to the presence of reacting char particles at the top of the well-mixed fuel reactor. However, Tong et al. reported purity levels of 99.7% utilizing solid fuels with a moving-bed arrangement of the fuel reactor and employing rather large size particles [63]. There are several possibilities to solve the problem of low gas conversion, including installation of a postcombustion chamber downstream from the fuel reactor, using a multistaged fuel reactor or separation and recycling of the gases to the fuel reactor. All of these solutions entail installation of additional components and will result in added costs for the overall process. Another option is to utilize oxygen carrier materials which release in the gas phase, that is, chemical-looping with oxygen uncoupling (CLOU), which is the topic of this paper. By using CLOU, the main mechanism for fuel conversion is changed completely, as the gasification reactions (3) and (4) are eliminated and replaced by a much faster combustion process, which has positive implications for both the char and gas conversions.

1.2. Chemical-Looping with Oxygen Uncoupling (CLOU)

Chemical-looping with oxygen uncoupling (CLOU) is very similar to normal CLC, with the major difference that the process utilizes oxygen carriers which can release gas-phase oxygen in the fuel reactor; see Figure 4 [65]. In this way, the fuel reacts directly with gaseous oxygen released from the oxygen carrier, via normal combustion. The oxygen carrier releases oxygen through the decomposition of the oxidized form of the metal oxide, , through (Step  1) This is followed by the normal combustion of the fuel via (Step  2) The reduced oxygen carrier is transferred to the air reactor for reoxidation (Step  3). The net heat of reaction for the CLOU processes is the same as CLC; only the mechanism by which oxygen is accessed by the fuel differs. In CLOU, there is no direct reaction between fuel gas and the oxygen carrier, as is the case with normal CLC; see reaction (1). When using solid fuels like coal, the CLOU process avoids the slow gasification step of the char, reactions (3) and (4), and instead the reaction in the fuel reactor is composed of an oxygen uncoupling reaction (5) and normal combustion, reaction (6).

The CLOU process was proposed by Lyngfelt and Mattisson in 2005 [66], with some publications on the subject a few years later [65, 67, 68]. Here, several monometallic oxide systems were proposed as feasible oxygen carrier materials, that is, CuO, and [65]. A thermal analysis was performed showing that CuO and were most promising, as they showed a combination of reasonably high partial pressure of oxygen at relevant temperatures in addition to an exothermic reaction in the fuel reactor, which could result in a temperature increase in the fuel reactor. An increase in temperature could be of significant advantage, as it would promote the oxygen release rate. Experiments were also performed in a batch fluidized bed reactor with coal and petroleum coke utilizing Cu-based oxygen carrier particles. The overall CLOU reaction rate was shown to be governed by the oxygen release reaction (5) and the combustion reaction (6), and two limiting cases could be observed: (i)the rate of oxygen release, or uncoupling, from the metal oxide is faster than the oxygen consumption by combustion reaction, resulting in oxygen being released from the exit of the reactor at a partial pressure approaching equilibrium, and(ii)the rate of oxygen consumption by combustion exceeds that of the release rate through metal oxide decomposition, resulting in an outlet oxygen concentration approaching zero. Here, the char particles in the well-mixed bed of metal oxide particles reacted instantly with the released gas-phase oxygen, thus removing the equilibrium restriction for oxygen release, and hence increasing the driving force for reaction (5). The oxygen tied up with the combustion products and was considerably higher than the amount which could be released if the reaction was limited by the equilibrium partial pressure of from reaction (5). An illustrative example of the enhanced rate of reaction possible with CLOU compared to CLC was provided in [65]. Here, it was shown that a batch of petroleum coke was converted approximately 50 times faster in a bed of particles via the CLOU mechanism, compared to the rate of fuel conversion obtained in a bed of particles fluidized by steam. The difference in conversion rates was attributed to the different mechanisms of fuel conversion between CLC and CLOU as described previously.

Oxygen carriers for CLOU need to fulfil many of the same criteria as normal CLC oxygen carriers; for example, the particles need to be mechanical stable, have limited tendencies for agglomeration, and be cheap. In addition, CLOU particles have special thermodynamic and kinetic requirements in comparison to oxygen carriers used in normal CLC; more specifically, consider the following.(1)The metal oxide should release gaseous oxygen and have a sufficient equilibrium partial pressure of oxygen at temperatures utilized during combustion in order to promote the rate of oxygen generation for complete burnout of the fuel.(2)The oxide system needs to have an equilibrium partial pressure low enough for oxidation at a reasonable oxygen concentration in the air reactor. This concentration is dependent upon the air ratio employed, but may need to be below approximately 5% . A higher partial pressure out from the air reactor would mean lower efficiencies due to heat losses with the flue gases. (3)The reactions in the fuel reactor should preferably be exothermic, as this would result in a temperature increase, which will promote the release rate of gas-phase oxygen.

For the CLOU process, it is important to consider the important interplay between oxygen partial pressure of the metal oxide system, the heats of reaction, and the temperature in the reactors.

In the next section, a thermodynamic analysis for some monometallic oxygen carriers and one promising mixed oxide particle is performed. Further, a thermal analysis is carried out for CuO in order to establish the magnitude of the possible temperature rise in the fuel reactor.

2. Thermodynamic Aspects of CLOU

2.1. Equilibrium Partial Pressure of O2

In a thermodynamic analysis of oxygen carriers for CLC, Jerndal et al. [5] showed that the oxide pairs CuO-, -, and -CoO were not possible to use as oxygen carriers for CLC at higher temperatures due to decomposition of the most oxidized phase. Later, Mattisson et al. [65] identified these same systems as possible candidates for CLOU. Of these, CuO has received the most attention and seems to be the most promising with respect to the rate of oxygen release, for example, [6972]. Rydén et al. recently reviewed the thermodynamics of a number of combined oxide systems [73], some of which now have experimentally proven CLOU properties [7476]. The monometallic oxides of Cu, Mn, and Co release oxygen in the fuel reactor through reactions (7)–(9) below. Reaction (10) shows the decomposition reaction for a combined oxide system of which has shown promise as a CLOU material both thermodynamically and experimentally [74]. The equilibrium constant for the above reactions is given by Assuming that the activity of 1 for both forms of the metal oxide, the partial pressure of can simply be equated with the equilibrium constant, that is, Figures 5(a) and 5(b) show the equilibrium partial pressure of as a function of temperature for the above reactions. From these two figures, it can be seen that CuO releases oxygen in air ( bar) at temperatures above 1028°C, at temperatures above 899°C, above 889°C, and above 960°C. For the combined oxide system of , there are two parallel curves which converge at higher temperatures, which are due to the presence of a two-phase region where both the oxidized and reduced oxides exist [74]. In the air reactor, however, the oxygen carrier needs to react with air and reduce the oxygen concentration in the outlet to a reasonable level, given by the excess air ratio [2], where is the outlet volume fraction of oxygen from the air reactor. Although the air ratio could be adjustable, in order to achieve an efficient process, it should probably be in the range 1.1–1.3, which corresponds to an outlet oxygen volume fraction of 2–6% in the outlet stream from the air reactor. Of course, a higher air ratio is possible, but will result in considerable heat losses and subsequently a less efficient process. For basis of the following discussion, it will be assumed that the oxygen concentration in the outlet should not exceed 5%. Thus, considering the equilibrium partial pressures in Figure 5, a CLOU system utilizing oxygen carriers of CuO, , , and the maximum temperature in the air reactor is 955°C, 820°C, 850°C, and 890°C, respectively. A higher temperature would likely mean that the outlet from the air reactor would contain more than 5% , which may be uneconomical in a power plant. When the particles are transferred to the fuel reactor, it is of course an advantage if the temperature is as high as possible, as this would promote the rate of oxygen release. This depends upon the thermal nature of the overall reaction in the fuel reactor, and this will be explored below.

2.2. Thermal Considerations

When the oxygen carrier material is transferred to the fuel reactor, the oxygen carrier should release oxygen which will react with the fuel. Considering that the total heat evolved from a CLOU system will be that of normal combustion, the sum of the heats of reaction in the fuel and air reactor will be that of normal combustion. In Table 1, the reaction enthalpies of the oxidation reactions (7)–(10) are given, in addition to the combustion enthalpies of C, , and a common coal per mol . The ratio between the reaction and combustion enthalpy shows whether the reaction in the fuel reactor is endothermic or exothermic; a ratio above 1 would mean that the reaction is endothermic while a ratio below 1 would mean an exothermic reaction. From Table 1, the ratio is below 1 for CuO, , and , meaning that the overall reaction in the fuel reactor must be exothermic, which will result in a temperature increase in this reactor. For , however, the ratio is slightly above 1 for C and , meaning that the reactions in the fuel reactor are slightly endothermic, and thus there will be a temperature decrease for . Although the reaction is slightly exothermic for the analyzed coal and cobalt oxide, considering heat losses and that heat will be needed to pre-heat the fluidization gas, there will likely be a temperature drop also for coal. The fact that a temperature increase is possible for most oxide systems is significant and of considerable advantage for the following reasons: (i) the higher temperature will promote the oxygen release rate from the metal oxide particles and (ii) the recirculation rate of particles between the air and fuel reactor can be kept low, as the recirculating particles do not need to provide heat for an endothermic reaction, which is the case for CLC with solid fuels [54]. The temperature drop in the fuel reactor for the cobalt oxide in combination with its high cost and toxicity likely makes this oxygen carrier unsuitable for a CLOU system.

The magnitude of a possible temperature increase will be illustrated with a simple mass and heat balance using CuO as an oxygen carrier. Assuming that the fuel is pure C, the overall reaction in the fuel reactor is In the calculations, it is assumed that 1 mol C is burnt in the fuel reactor. Further assumptions are (i) the reactor is adiabatic, (ii) the particles from the air reactor are fully oxidized and at a temperature of 925°C, which is the temperature of the particles which enter the fuel reactor, (iii) 20% of the CuO is converted to in the fuel reactor, and (iv) the reactor is fluidized with 0.5 moL which is at a temperature of 400°C. A summary of the mass balance can be seen in Table 2, where the total heat capacity of the products, , is also calculated to be 1134.3 J/°C. Taking into account that approximately 35 kJ is needed to heat the fluidization gas and the fuel to the reaction temperature, then the temperature increase can be estimated from Thus, the actual temperature in the fuel reactor would approach 1012°C. From Figure 5(a), this corresponds to an partial pressure of 15%. The increase in temperature will promote the overall reaction rate in the fuel reactor, as the rate of oxygen decomposition is governed by the equilibrium partial pressure of and a reaction rate constant, see (18), both of which likely increase as a function of temperature. A more detailed mass balance of the CLOU system was carried out by Mattisson et al. [65] and Eyring et al. [77]. Sahir et al. performed a process simulation on both a CLC and CLOU systems utilizing and CuO, respectively [78]. The main advantages with the CLOU system were identified to be the lower residence time of solids in the fuel reactor and the possibility to extract heat from the fuel reactor due to the exothermic reactions. In association with the CLOU concept, there have been several articles which have proposed using the reversible redox reaction (5) for production of gaseous oxygen [79, 80]. As fuel would not be added to the fuel reactor, the reaction in the fuel reactor would be endothermic with a resulting temperature drop, and hence the driving force for oxygen release would not be promoted, as is the case with CLOU. Further, as there is no fuel to remove the gaseous oxygen through combustion, the oxygen partial pressure during the decomposition may be limited to the equilibrium partial pressure at the temperature employed.

3. Overview of Experimental Investigations of CLOU Materials

The first experimental work on CLOU was published in 2008 and 2009 [65, 67, 68], where the CLOU concept was shown using a Cu-based oxygen carrier with gas and solid fuels. It should be mentioned that already in the original work on chemical-looping by Lewis et al. in 1951, coke was combusted with a Cu-based oxygen carrier, and it was suggested that part of the fuel was converted by uncoupled gas-phase oxygen [6]. In the last five years, there have been a large number of articles published with experimental investigations of different oxygen carrier materials, mostly based on Cu and Mn oxides. Table 3 contains a complete overview of the experimental investigations of CLOU materials. Adanez et al. [40] provided an overview of CLOU materials in 2011, and this table is basically an update of this overview. It should be noted that only articles which have the aim of utilizing the oxygen uncoupling behaviour of metal oxide particles are presented in Table 3. Further, several conference papers are not included, as the results are presented in follow-up journal articles. Most of the investigations have been carried out at Chalmers University of Technology in Sweden and at ICB-CSIC in Zaragoza, Spain, but there are currently several other groups working with the development of the CLOU process, including research groups at the University of Utah in the United States, Tsinghua University in China, and Swiss Federal Institue of Technology in Switzerland. Below follows an overview of the research presented in Table 3.

3.1. Copper Oxide

Copper oxide has received a great deal of attention as an efficient oxygen carrier both for CLC and CLOU applications, owing to its high reactivity, high oxygen transport capacity, and absence of thermodynamic limitation for complete combustion of hydrocarbon fuels. With respect to CLC, the CuO can be reduced all the way to Cu via Most research around the use of CuO for CLC has focused on this direct reaction between fuel and oxygen carrier. Investigations have been carried out with different types of gaseous and solid fuels in batch fluidized bed reactors [58, 70, 8188], fixed-bed reactors [89, 90], thermogravimetric reactors [13, 1517, 9193], and interconnected fluidized beds with continuous circulation of the oxygen carrier material [32, 94]. Despite the rather promising results found in the studies of Cu materials for CLC, the use of CuO as an oxygen carrier is not without limitations. For example, copper oxide suffers from tendency towards agglomeration upon full reduction, due to the formation of elemental copper which has a rather low melting temperature (1085°C) [81].

When utilizing copper materials for CLOU, metallic Cu should not be formed, but it is the redox pair CuO- which is active; see reaction (14). Most experimental studies around CLOU have been conducted with CuO as the active material [6872, 77, 95102]. Although there certainly could be an oxygen uncoupling effect also in some of the previously mentioned studies of CuO for CLC, the investigations in Table 3 were specifically aimed at utilizing the uncoupling effect possible when cycling between CuO and .

Mattisson et al. demonstrated the CLOU concept using oxygen carriers of and with both gaseous and solid fuels in a batch fluidized bed reactor [65, 67, 68]. Figure 6 shows the concentration profiles for one such experiment where 0.1 g of petroleum coke reacted with 15 g of particles which were fluidized with . Initially, the particles are exposed to air (), but at 25 s the air is replaced with nitrogen, and the outlet oxygen concentration drops and approaches the equilibrium concentration. At approximately 70 s the petroleum coke is introduced to the fluidized bed and the fuel mixes with the oxygen carrier particles. The char reacts rapidly with the released to form . There are some small peaks of and CO initially due to release of volatiles, which are not shown in the figure. The oxygen concentration drops during the reduction but does not go down to zero, and at the same time the volume fraction is high, approaching 51%. This means that the thermodynamic barrier for the decomposition of CuO to is removed and that the release of from the particles is faster than the consumption of through combustion. Complete burnout of the petroleum takes about 30 s, after which the oxygen concentration again approaches equilibrium. The rate of char conversion observed in this experiment is a 45-fold increase to that observed utilizing a Fe-based oxygen carrier without CLOU properties [54]. Leion et al. [67] performed CLOU experiments with the same oxygen carrier in a batch fluidized bed using a lignite coal in the temperature range of 850–985°C and found similarly enhanced reaction rates. The oxidation reaction occurred close to the equilibrium concentration for these particles. In a follow-up investigation by Leion et al. [103], six different solid fuels were investigated together with petroleum coke, coal, and wood char with the same material. Also, here considerably higher conversion rates were found for CLOU compared to rates with ilmenite as oxygen carrier. The difference in fuel conversion rates was more pronounced for the low volatile fuels, such as, petroleum coke.

In the most extensive screening study of Cu-based oxygen carriers for CLOU to date, Gayán et al. prepared 27 different materials of CuO with different support materials using impregnation and mechanical mixing. Of these, the oxygen release characteristics of 25 materials were tested using TGA and a batch fluidized bed reactor [72]. The most promising materials were composed of CuO with and CuO on . From this screening study, the same group of researchers prepared a material based on 60 wt% CuO with using spray drying, and this material has been investigated using a number of different experimental units and conditions. The oxygen carrier was tested with coal in a continuously operated CLOU unit, showing proof of the concept of CLOU using solid fuels [69, 98]. Here, complete combustion was found using an oxygen carrier inventory in the fuel reactor of 235 kg/MWth. As mentioned previously, most earlier studies of CLC with solid fuels have resulted in incomplete combustion, with combustible components released from the fuel reactor. The investigation by Abad et al. shows that complete gas conversion is possible using oxygen carrier particles with CLOU properties. Several other studies have found that CuO with is highly promising as a CLOU material [70, 71, 9597]. Arjmand et al. compared the feasibility of using CuO with and as support and found that CuO combines readily with to form which does not have CLOU properties [70]. Adanez-Rubio et al. tested the combustion of four different types of coal in a continuously operated CLOU unit at temperatures of 900–950°C using a spray-dried oxygen carrier of CuO with [96]. Complete combustion was attained for all coals, with no unburnt in the off-gases. Although the results with CuO with appear highly promising with respect to reactivity, Adanez-Rubio et al. also report that the particles increase in porosity as a function of time and hence lose mechanical strength [97].

Most investigations of CuO materials have utilized particles with or as support, as seen from Table 3, but there are also studies of Cu materials for CLOU utilizing natural materials [101] as well as other support materials [99, 102, 104, 105]. For instance, Hedayati et al. used CuO with and prepared by extrusion [104]. The particles were investigated in a batch fluidized bed reactor with respect to oxygen release behaviour and reactivity with CH4 and syngas. The oxygen carrier utilizing the Gd-doped support showed the best results, and complete conversion of both and syngas could be achieved in the batch fluidized reactor which was used to test the material.

3.2. Combined Oxides

The monometallic systems of and clearly have interesting properties which make them suitable for CLOU applications. The disadvantage with utilizing Cu material is the relatively high cost of Cu in addition to the low melting temperature of copper. Further, manganese oxides suffer from the apparent slow kinetics of to at the low temperature needed in order to oxidize the material fully; see Figure 5(b) [104, 106]. Recently, a number of articles have been published utilizing combined oxides as oxygen carriers for CLOU, for example, [74, 107, 108]. The common denominator for all of these investigated materials is that they contain manganese in the structure. These combined oxides have thermodynamic and/or kinetic properties which differ from pure and which enable the decomposition of the oxide with release of gaseous oxygen at conditions suitable for CLOU. An example of this is the system. A comparison of the -T function of this material with in Figure 5(b) shows that it should be possible to oxidize the combined oxide at a higher temperature in the air reactor than what is possible for the pure oxide, and this may also affect the rate of oxidation. A thermodynamic analysis of many of these systems has been performed by Rydén et al. [73].

3.2.1. Mn-Fe-O

In an experimental study by Shulman et al., it was discovered that combined with performed very well as a CLOU oxygen carrier [76, 109]. - particles with a molar ratio of of 0.6 and 0.8 were produced by freeze granulation. Both oxygen carriers released oxygen in the gas phase in the temperature interval of 810–900°C. Reactivity experiments with methane in a fluidized bed reactor found that reacted almost completely to at 850 and 900°C for the material with the highest fraction Mn. It was speculated that the active phase was a combined metal oxide of the form . Oxygen carriers composed of Mn-Fe had earlier been tested by Lambert et al. [110], but the CLOU effect was not observed in this investigation. Azimi et al. [74, 111113] has conducted extensive investigations of this system using spray-dried oxygen carriers and using both solid and gaseous fuels. In one investigation [112], an extensive number of Fe/Mn particles were tested with from 0.2 to 0.8 using methane and syngas. The oxygen uncoupling effect was dependent on the Mn/Fe ratio and the temperature of oxidation and reduction. Oxygen carrier particles with an ratio of 0.2–0.4 release oxygen at 900°C, whereas materials with a higher Mn content showed no oxygen uncoupling. On the other hand, at the lower temperature of 850°C materials with an greater than 0.5 released considerable amounts of oxygen. The behaviour of these materials agreed well with what is expected based on the calculated phase diagram of the Fe-Mn-O system. The reactivity with methane followed the uncoupling properties well.

Rydén et al. [114] investigated two oxygen carriers with high contents of Mn and Fe in a continuous circulating CLC reactor. The first material was produced by spray drying and contained only and with a molar ratio of Mn/Fe of 2. The second material was a natural manganese ore with a relatively high content of (>6.5%). Both materials had a CLOU effect in the temperature interval of 850–1000°C and a gas yield to at 950°C of 96% and 76% for the synthetic and natural ore, respectively. These values were judged to be excellent compared to other types of material. The main problem was that both materials deteriorated and turned to very fine powder after several hours of combustion. This lack of mechanical stability seems to be a general problem with this system, as swelling and increased porosity was noted also by Shulman et al. [76]. Also Arjmand et al. [115, 116] found a small CLOU effect for several manganese ores at 950–970°C. Fossdal et al. [117] investigated a number of ores and industrial wastes as oxygen carriers for CLC. Among the tested materials, the redox properties of two Fe-Mn slags were investigated. The materials were not deemed promising as the oxygen carrying capacity and regenerability was poor. However, it is likely that the phase transition investigated was between and , which is not of interest for CLOU.

3.2.2. Mn-Si-O

Rydén et al. presented a phase diagram of the Mn-Si-O system and identified several pathways for oxygen release [73]. The experimental investigations of this combined oxide system are very limited with respect to CLOU. Shulman et al. [76] prepared particles of 80 with by freeze granulation. The particles released oxygen in the gas phase in the temperature interval of 810–900°C. Interestingly, from XRD analysis of the samples calcined at 950°C and 1100°C, there did not seem to be any interaction between and . Jing et al. investigated spray-dried particles of with varying content of Mn using a batch fluidized bed reactor [107]. Particles which contained more than showed no uncoupling at 900°C, whereas the samples containing 50 and 70 released oxygen in the inert period; that is, when the particles were fluidized with nitrogen. The highest degree of uncoupling was seen for a sample which also contained . The average gas conversion of methane to carbon dioxide varied between 70 and 98% at 950°C. Although no phase analysis was presented, the results were interpreted using the phase diagram presented by Rydén et al. [73].

3.2.3. Other Combined Oxide Systems

In addition to the combined manganese oxide systems of Mn-Fe-O and Mn-Si-O discussed above, a few studies have explored other similar mixed oxide systems. Azad et al. investigated the CLOU properties of oxygen carriers composed of and CuO [118]. The authors found that there was a cooperative synergy between the Cu and Mn oxides, with formation of during oxidation, which was reduced to and with concomitant release of in the gas phase. The reactivity with was investigated at 800 and 850°C in a batch fluidized bed reactor, and a high gas yield to and was obtained at the higher temperature. In addition to the oxygen carriers of Mn-Fe and Mn-Si, Shulman et al. also tested materials based on Mn-Ni and Mn-Mg, and both systems showed interesting CLOU properties [76, 108].

3.2.4. Perovskites

Perovskites are a series of materials with the general formula , where is a large cation and is a smaller cation. Perovskite materials exhibit interesting redox properties and have been investigated for both chemical-looping combustion (CLC) and chemical-looping with oxygen uncoupling (CLOU) [75, 119122]. Oxygen carrier particles can exhibit oxygen nonstoichiometry and can release substantial amounts of oxygen via The oxygen released in the gas phase, given by , can be used during CLOU process to burn gas, liquid, or solid fuel. The term “” gauges the degree of oxygen nonstoichiometry and “AR” and “FR” refer to the air and the fuel reactor, respectively. The oxygen nonstoichiometry of a perovskite material can be altered by changing parameters such as the ambient temperature, pressure, or the fugacity. For example, in the oxidizing environment of the air reactor, is determined by the partial pressure of oxygen at the outlet of the reactor, which is typically around 5%. In the fuel reactor, the environment is reducing which translates to a very low activity; hence at equilibrium is larger than meaning that there will be a release of gaseous oxygen.

With respect to CLOU, the focus has been on the calcium manganate system, , which has shown to exhibit a combination of high mechanical stability, good fluidization properties, and well good reactivity with gaseous fuels. The fact that this type of material can be produced with cheap raw materials is an added advantage. Leion et al. [123] investigated the CLOU behaviour and reactivity of a series of materials of general composition with in a TGA. Based on the oxygen capacity and the reactivity, the oxygen carrier of was chosen for further testing with and petroleum coke. The reactivity with petroleum coke at 900°C and 950°C was slower compared to previous experiments with Cu-based oxygen carrier [68], which was explained by the slower rate of oxygen uncoupling in this material. Rydén et al. investigated the same material in a 300 W continuously circulating reactor using natural gas [75]. The material exhibited very smooth fluidization behaviour for over 70 h at hot conditions. The uncoupling effect was dependent upon the oxygen concentration in the air reactor, with a higher concentration of from the fuel reactor as the oxygen concentration was increased in the air reactor. It was possible to achieve complete fuel conversion at 950°C. More recently, Källen et al. investigated the performance of , in a 10 kW natural gas-fired CLC unit [124]. The results showed high rate of oxygen release and full conversion of natural gas—a behaviour superior to that of NiO-based oxygen carrier in the same unit [36, 37].

Arjmand et al. investigated a series of La-doped calcium manganates with gaseous and solid fuels [125]. Six materials of general formula (, Mg, Fe, Cu) were investigated with a 10% substitution of Ca with La and Mn with the metals above. With the exception of the Cu-doped material, it was found that the materials had excellent mechanical strength, CLOU properties, reactivity, and fluidization properties. It appears, however, that the pure has a higher rate of reaction than all of the doped materials.

Calcium manganate can be produced with very cheap material, that is, CaO and manganese ore [117]. This is likely a very important aspect for solid fuel combustion, as the lifetime of particles in a solid fuel combustor is likely to be much lower than in a gas-fired unit due to deactivation with ash as well as loss of material with the ash flows. The highly promising results from testing of these materials clearly make this one of the more promising combined oxides. However, the effect of sulphur on the particle reactivity and stability needs to be investigated.

4. Kinetics of Oxidation and Reduction of Oxygen Carriers in CLOU

There are only a few studies which have focused on determining the kinetics of the oxygen release and oxidation rates of oxygen carriers for CLOU applications. The focus has been on Cu-based oxygen carriers. Determination of kinetics of CLOU materials poses some additional challenges in comparison to kinetic data of regular CLC materials, as the rate of both the reverse and forward of reaction (5) is a function of an oxygen driving force, given by the difference between the actual oxygen partial pressure and the equilibrium partial pressure of oxygen [102]. In the fuel reactor the release rate, (kg /kg OC,s), of oxygen, or decomposition of CuO, can be expressed generally as where is the equilibrium partial pressure of and is a reaction rate constant. In the fuel reactor, the oxygen concentration should be kept very low at optimized conditions, that is, . Thus, the rate is directly proportional to a reaction rate constant and the equilibrium partial pressure, which is why it is an advantage to have as high temperature in the fuel reactor as possible.

The rate of decomposition can be evaluated in several ways. Arjmand et al. [71] introduced an excess of devolatilized wood char to a bed of freeze-granulated particles and in this way was able to calculate the rate of decomposition in the temperature interval of 850–900°C. The decomposition reaction was modelled using the Avrami-Erofeev mechanism and the kinetic data were obtained using an Arrhenius expression. From the analysis, an activation energy of 139.3 kJ/mol was calculated for the decomposition reaction. A similar methodology was used by Adanez-Rubio et al. [95], who studied the rate of decomposition from a similar type of oxygen carrier, albeit produced with spray drying. By utilizing small enough ratios of OC/fuel, the maximum rate of oxygen release was obtained as  kg /kg OC,s at a temperature of approximately 980°C. Whitty and Clayton [102] performed decomposition experiments in three types of reactor systems using and found that the rates of decomposition were similar independent of reactor type and also very similar for the three Cu-based oxygen carriers evaluated, with an average activation energy found of 110 kJ/moL. Eyring et al. [77] investigated the rate of decomposition of fine CuO in a TGA in the temperature range of 750–950°C, found that the decomposition could be explained by an Arrhenius type of expression, and calculated an activation energy of 327 kJ/mol for the decomposition reaction, which was very similar to that found previously by Chadda et al. [126]. A somewhat smaller activation energy of 281 kJ/mol was obtained by Sahir et al. [127] for the decomposition reaction of a freeze-granulated oxygen carrier.

One study has investigated the rate of decomposition of calcium manganate materials. Arjmand et al. [125] employed the same methodology as described above for obtaining the rate of oxygen release from to evaluate the oxygen release rate from a series of perovskite materials of general formula . The maximum rates of uncoupling were approximately 0.0003–0.0004 kg /kg OC,s at 950°C, which is about 15 times slower than for the earlier investigated Cu-based oxygen carrier at 900°C [71].

A few publications have reported on kinetics of the oxidation reaction, that is, reverse of reaction (5). The general rate expression for the oxidation reaction is given by where (kg /kg OC,s) is the rate of the oxidation and is a reaction rate constant. From (19) it is clear that a higher temperature would reduce the driving force for the reaction, that is, , and thus a slower rate of reaction may be expected as the temperature increases. Arjmand et al. [71] also investigated the oxidation rate of the freeze-granulated oxygen carrier of in the temperature interval of 850–950°C using a batch fluidized bed reactor. The reaction was very fast and the authors were unable to obtain a rate which was not limited by the supply of oxygen to the reactor. Whitty and Clayton investigated the rate of oxidation for an oxygen carrier of as a function of temperature and oxygen driving force. Although an increase in oxidation was seen when the oxygen partial pressure was increased, the increase in rate was not as high as expected, indicating that there were other mechanisms restricting oxidation, which is an observation in conformity with previous studies [102]. An activation energy of the oxidation of 202 kJ/mol was found for this oxygen carrier. Eyring et al. also found a decrease in oxidation rate as a function of temperature in the interval of 850–950°C when oxidizing a fine CuO powder in air [77]. Sahir et al. analyzed oxidation data of freeze-granulated particles of using the law of additive reaction times and apparent rate constants were determined at 900 and 950°C [127].

5. Design Criteria

In a chemical-looping combustion system, two key design parameters are (i) the solid inventory of oxygen carrier particles in the air and fuel reactor and (ii) recirculation rate of oxygen carrier particles between the interconnected air and fuel reactors. Numerous studies in the literature have calculated these parameters for both gaseous and solid fuel applications, for example, [54, 60, 106, 128]. In a similar way, these values for solid inventory can be calculated for a CLOU process, and this is described below.

The minimum solid inventory, (kg/MW), for a CLOU process designed with interconnect fluidized bed reactors is related to the oxygen demanded by the fuel and the oxygen release rate from the oxygen carrier material [71, 95], Here, is the rate of oxygen release from the decomposition reaction (kg /kg OC,s) and is the rate of oxygen demand by the fuel, with units kg /s, MW.

The circulation rate of the oxygen carrier particles, (kg OC/MW,s), can be calculated from the oxygen demand of the fuel and the change in the degree of mass-based conversion of the solids in the air and fuel reactor, , The amount of oxygen carrier needed in the air reactor can be calculated in a similar way; see Arjmand et al. [71].

Several studies have calculated the mass inventories needed in the fuel reactor of a CLOU system utilizing solid fuels [69, 71, 77, 95, 96]. Adanez-Rubio performed studies of a spray-dried Cu-based material composed of 60 wt% CuO on as inert [95]. Here, the release rate of oxygen was determined by varying the OC/coal ratio. At high OC/coal ratios, there was complete combustion to , but with excess in the flue gases. As the OC/fuel ratio was lowered, the oxygen concentration dropped to zero, and the maximum rate of oxygen release from the oxygen carrier was obtained. From these data, the minimum oxygen carrier inventory in the fuel reactor was established to be in the range of 29–39 kg/MW in the temperature range of 930–980°C, although somewhat higher values were suggested in order to achieve complete conversion of CO. In a study by Arjmand et al. [71], a similar methodology and oxygen carrier were employed to obtain reaction kinetics of the oxidation and reduction reaction, and from this data, a total solids inventory of 73–147 kg/MW was calculated assuming char as the fuel.

Experiments in continuous CLOU units employing CuO oxygen carriers have been performed utilizing gaseous, liquid, and solid fuels. Abad et al. [69] performed experiments in a continuously circulating CLOU reactor with bituminous coal and found complete conversion to carbon dioxide using 235 kg/MW of a Cu-based oxygen carrier in the fuel reactor. Adanez-Rubio et al. tested four types of solid fuels in a continuous CLOU unit and from the analysis it was found that between 45 and 490 kg/MW material was needed in the fuel reactor in order to achieve 95% carbon capture [96]. Moldenhauer et al. [31] investigated the combustion of kerosene utilizing CLOU and a CuO/ oxygen carrier in a circulating CLOU unit. The fuel conversion to and was between 95% and 99.99% in the temperature range of 750–900°C. These results were obtained using a total solids inventory (air reactor + fuel  reactor + particle locks) corresponding to 2000 kg/MW. However, it is expected that less than half of the oxygen carrier particles were in the fuel reactor. Finally, CLOU has also been tested with natural gas in a 10 kW circulating unit. Källen et al. [124] used the oxygen carrier for an extensive period of time, that is, 350 h at hot conditions, whereof 55 h with natural gas, and was able to achieve full gas conversion to carbon dioxide at 930°C using a bed mass in the fuel reactor corresponding to 760 kg/MW.

6. Outlook

Chemical-looping combustion (CLC) and chemical-looping with oxygen uncoupling (CLOU) are new processes for oxidizing a fuel. The main advantage with these technologies is that inherent capture of is possible with no gas separation step needed in order to obtain the carbon dioxide in pure form. Most research in the last two decades has focused on CLC with gaseous fuels [40], and although there are no commercial units available today, there are a significant number of smaller units in the size range of 0.3–140 kW which have been operated successfully with a wide variety of oxygen carrier materials. It is clear from the operational experience with gaseous fuel that almost complete gas conversion is possible with 100% capture. There are today plans of constructing larger natural gas-fired CLC pilots for steam production [129].

Although most research around CLC and CLOU has been conducted with gaseous fuels, it is likely that the greatest advantage with chemical-looping will be realized for power production from solid fuels [130]. This is because these combustion processes are normally conducted at atmospheric pressure and integrated with a Clausius-Rankine steam cycle, which is also what is envisaged with CLC and CLOU with solid fuels. Efficient power production from gaseous fuels would likely require a pressurized process, and there is very little research currently on chemical-looping under pressurized conditions. Despite the technoeconomic promise of CLC with solid fuels, the technology is not without drawbacks. The slow gasification of solid char, reactions (3) and (4), in the fuel reactor will mean that long residence times may be needed in the fuel reactor in order for the fuel to be fully converted to gas. Although most studies report on carbon capture efficiencies in the order of 90%, Markström et al. achieved 99% carbon capture utilizing ilmenite and a bituminous coal in a 100 kW CLC unit [64]. The biggest challenge with normal CLC with solid fuels is likely to achieve a high degree of gas conversion in the fuel reactor. The reactive gases generated from the gasification reactions are generated inside the entire fuel reactor, and some of the gases will not have sufficient contact with oxygen carrier materials, and thus it is expected that even in a well-designed fuel reactor, some small amounts of CO and will be present at the outlet. Most studies report on gas conversion to carbon dioxide of less than 95%, although there are studies claiming higher degrees of gas conversion. The presence of CO and in the outlet of the fuel reactor will mean that there may need to be a treatment step downstream of the fuel reactor, which will add to cost and complexity.

The use of CLOU will solve many of the disadvantages with regular CLC. There are no gasification steps, but rather the fuel is burnt with oxygen. The overall char conversion reaction has been shown in several studies to be much faster utilizing Cu-based CLOU particles compared to the rate with a non-CLOU oxygen carrier. For CLOU the uncoupling reactions occur in the emulsion phase where the char and reactive gases are also present and it is believed that it is possible to get full gas conversion even when the mixing is not perfect. Experimental investigations in circulating units give support of this idea. Complete combustion has been obtained in two studies using different types of coal in a circulating CLOU reactor [69, 96]. It can also be speculated that CLOU oxygen carriers will be an advantage using gaseous and liquid fuels. If oxygen is released in the gas phase in regions of low combustibles, this oxygen can be used downstream in the reactor, or freeboard, to convert combustibles which have bypassed the bed. Experimental results from continuous units are promising. All data from continuous units with these types of fuels report very high or complete conversion of the fuel to and [31, 75, 114, 124].

The majority of investigations in the open literature with respect to chemical-looping are devoted to normal CLC, and there are still only a few research groups specifically working on developing oxygen carriers which release oxygen in the gas phase. The focus has been on developing and testing copper oxide-based materials, as is seen in Table 3. The disadvantage with CuO is the high cost, relatively low melting temperature, and the lack of mechanical stability, as several papers report on high rates of attrition or fragmentation. There are certainly some interesting combined oxide systems which may have the advantage that they are less expensive than the Cu-based materials; however, also many studies have reported problems with attrition and fragmentation, especially for the Fe-Mn systems. It is expected that the future trend in research around CLOU materials involves(i)more tests of promising materials in continuous reactors for extended periods of time, (ii)investigation of the effect of impurities on oxygen carrier reactivity and stability, most importantly the effects of ash contamination and sulfur poisoning,(iii)fundamental studies of attrition phenomena for Cu-based and combined metal oxide systems,(iv)theoretical exploration of additional combined oxide systems which may have properties suitable for the CLOU process.

7. Conclusions

Chemical-looping with oxygen uncoupling (CLOU) is a new and promising method for oxidizing a fuel to carbon dioxide and steam. The process is similar to chemical-looping combustion (CLC) and and are inherently isolated from the rest of the flue gases without the need for an energy intensive gas separation step. The difference between CLC and CLOU is the mechanism for oxidation of the fuel in the fuel reactor. In CLOU the oxygen carrier releases oxygen in the gas phase, which then reacts with the fuel through normal combustion, in contrast to CLC, where the oxygen carrier reacts directly or indirectly with the fuel through a noncatalytic gas-solid reaction. This paper gives a complete overview of the CLOU process and the experimental investigations which have been performed in the last five years. The process has been tested in continuous working CLOU reactors for gaseous, liquid, and solid fuels, with very promising results with respect to rates of fuel conversion. The main advantage with the CLOU process is certainly for solid fuels, but it is suggested that it may also be advantageous with respect to gaseous and liquid fuels.

Acknowledgment

The authors wish to acknowledge the support of AB Vattenfall for the financial support of this work.