Solvent Development for Post-Combustion CO2 Capture: Recent Development and Opportunities

Chemical absorption is widely regarded as the most promising technology for postcombustion CO2 capture from large industrial emission sources with CO2 separation from natural gas using aqueous amine solvent system having been applied since the 1930s. The use of monoethanolamine (MEA) in CO2 absorption system possesses several drawbacks, such as high regeneration energy, high solvent loss, and high corrosion tendency. Various solvents have been developed for post-combustion CO2 capture application including the development of aqueous solvents and phase-change solvents. Some of these alternate solvents have been reported to have better solvent properties, which could improve the CO2 absorption system performance. This paper reviews key parameters involved in the design improvement of several chemical absorption process systems. In addition, some novel solvent systems are also discussed, for example encapsulated solvents systems. Some of the key solvent parameters that affect the capture performance, such as heat of reaction, absorption rate, solvent working capacity, solvent concentration, and solvent stability, are discussed in this paper, particularly in relation to the economic viability of the capture process. In addition, some guidelines for the future solvent development are discussed.


Introduction
A CO2 reduction scheme that is gaining growing interest is Carbon Capture and Storage (CCS) [1]. Within the CCS process chain, CO 2 capture is the costliest stage and therefore it is important to develop the technologies that can reduce costs. Among all CO 2 capture methods, postcombustion CO 2 capture using chemical absorption has been recognized as the most commercially ready technology. The concept of this technology has been applied, albeit at different feed gas sources, in natural gas industry since the 1930s, where CO2 is absorbed using aqueous amine solvent system [2]. Many researchers suggested that other CO2 removal methods, such as membranes and adsorption, are not likely to be competitive because of compression work [2]. The application of physical absorbents in post-combustion CO2 capture is likely to be more limited than that of chemical absorbents because of the low CO2 partial pressure in the flue gas [3]. The future development of chemical absorption will be the focus of this paper. This paper aims to review the key parameters involved in the design improvement of several chemical absorption process systems. In particular, the recent updates on solvent development are presented for two solvent classes (aqueous solvents and phase-change solvents). In addition, some novel solvent systems are also discussed, for example encapsulated solvents systems.

Aqueous Solvents
In most industrial applications, only two phases are involved in the CO2 absorption, the gas phase made up of the gas/gases to be recovered and the solvent liquid phase. In the following section, the key characteristics of different groups of aqueous solvents are discussed.

Primary Amines
As shown in Table 1, there are various solvents belonging to the primary amines group including MEA, DGA, and EDA. The most widely used is MEA. MEA was first applied and widely developed to capture CO2 from feed gases with low partial pressure of CO2 as it possesses fast reaction rate and high working capacity compared to other solvent groups. MEA has several advantages over other amine-based solvents [9,[11][12], such as high reaction rate, high absorption capacity on the mass basis (because of low molecular weight), good thermal stability, and cheap. However, MEA owns deficiencies such as high energy penalty to regenerate the solvent, corrosive, and high solvent make-up rate [4]. MEA has a high heat of reaction (82 kJ/mole), thus it requires high regeneration duty in the stripper. This is due to the formation of the stable carbamate ion (MEACOO -). The regeneration energy of the generic MEA 30 %-wt. is reported to be between 4.1 to 4.4 MJ/kg CO2 [8,[13][14][15]. The high level of solvent loss in an MEA system is due to the vaporization at the absorber. The other major issue regarding the solvent loss is the degradation of solvent, which has several causes (oxidative and thermal reactions). Make-up solvent is required to compensate for the loss that occurs in the process. Some operating strategies that have been implemented to address the issue of solvent loss include limiting the solvent concentration while corrosion effects are dealt with by ensuring proper equipment material selection and utilizing mild operating conditions [8].

Secondary and Tertiary Amines
CO2 separation from natural gas using secondary and tertiary amines, such as DEA and MDEA, have been commonly applied commercially for decades. DEA and MDEA have a lower heat of reaction and a higher working capacity compared to MEA and thus require lower regeneration energy. The drawback of these solvents is a much slower absorption rate compared to MEA. Aqueous solvents consisting entirely of secondary and tertiary amines are unlikely to be applied in postcombustion CO 2 capture applications.

Hindered Amines
Hindered amine compounds are commonly recognized when the nitrogen atom of the amine molecule adds a bulky substitute group. The bulkier substitutes attached to the amine backbone lead to a lower rate of carbamate formation and hence the bicarbonate formation will then be the main reaction mechanism [23][24]. Furthermore, as the reaction mechanism in hindered amines is dominated by bicarbonate formation, the heat of reaction of hindered amines is lower than MEA. A proprietary hindered amine, trademarked as KS-1, has been developed by Mitsubishi Heavy Industries (MHI) in 1990 [25]. The CO2 capture process using KS-1 is similar to the MEA (Figure 1), except that it incorporates a heat recovery system and absorption inter-cooling. It was reported that the regeneration energy obtained during testing was 2.53 MJ/kg CO2, which is up to 40 % less than MEA [26]. KS-1 has a larger loading capacity, thus the circulation rate of KS-1 is 40 % lower than MEA.

Polyamine
The most widely known polyamine solvent for CO 2 absorption is piperazine (PZ), which has been commercially developed as a rate promoter for various solvent types, such as MDEA and potassium carbonate. When PZ is used as a promoter, the concentration used is low, between 0.5 and 2.5 m PZ, because PZ is not highly soluble.

Alkali Carbonate
A hot potassium carbonate process has been used for CO 2 absorption from natural gas and ammonia synthesis since the 1960's. In this process, the CO 2 is captured at high pressure and the concentration of the potassium carbonate is limited to 20-30 %-wt. to prevent solid precipitation. Alkali-carbonate solvents have several benefits compared to other solvents including low heat of reaction and low solvent cost [5,[29][30][31]. The CO2 is absorbed through bicarbonate formation, which leads to the significantly lower heat of reaction than that of MEA. Endo et al. (2011) reported that the regeneration energy of an aqueous potassium carbonate system is 4.9 MJ/kg CO2, which is 10-20 % higher than that of MEA. Although potassium carbonate has a lower heat of reaction than MEA, the regeneration energy is higher because of the lower working capacity and hence the solvent circulation rate is higher [5,31]. The other interesting feature of alkali carbonates is the possibility to exclude the FGD and SCR from the pre-treatment facilities [31]. The NOx and SOx react with the potassium ion to form potassium nitrate and potassium sulfate precipitates, which can be removed as by-products using reclaiming units (such as ion exchange) prior to entering the absorber column. This potentially reduces the capital cost of the capture plant presumably because the capital cost of the precipitate removal is less than the cost of FGD and SCR.

Ammonia
Aqueous ammonia has been investigated as a solvent for post-combustion CO2 capture. The advantages of this system compared to MEA include higher loading capacity, lower heat of reaction, and better resistance to oxidative and thermal degradation [32]. The other potential benefit of this system is the possibility to use carbon steel for equipment because ammonia is significantly less corrosive compared to amine solutions. One of the drawbacks of using aqueous ammonia is the high solvent volatility because of its low vapor pressure. Ammonia slip to the atmosphere must be prevented because of its toxicity. To overcome this, ammonia wash sections are required at the top of the absorber and the stripper, which increases the capital costs by 1-2 % [33].
The CSIRO and Delta Electricity assessed aqueous ammonia technology for a post-combustion CO2 capture pilot plant at Munmorah Power Station in NSW, Australia [32,34]. The ammonia solution investigated had concentrations of 0-6 %-wt. The absorber was operated at 15-20 °C and atmospheric pressure, while the stripper was operated at 300-850 kPa. They found that the absorption rate was at least 50 % less than that of MEA [32]. The regeneration energy obtained during the trials was reported to be 3.0-4.2 MJ/kg CO2 [35].

Other Solvents
Several other solvents that have been developed for CO2 capture include ionic liquids (IL) and deep eutectic solvents (DES). IL are defined as salts that is composed by ions with a melting point lower than 100 °C [36]. Ionic liquids have been reported to show excellent properties for CO2 capture applications due to their tunability, thermal stability, non-flammability, and high CO2 solubility [36][37][38][39]. There are various ionic liquids currently under development for CO2 capture, most of them can be classified into two classes: room temperature ionic liquid (RTIL) and task specific ionic liquid (TSIL). Among these, the imidazolium-based RTIL class is the most widely investigated. However, the use of RTIL is based on physical absorption technology and not widely applicable for low partial pressure flue gas post-combustion CO2 capture.
A deep eutectic solvent (DES) is a mixture of a substituted quaternary ammonium or metal halide salt with a hydrogen bond donor (HBD), such as an amide or carboxylic acid. This class of ionic liquid is an eutectic mixture that has a quite low melting point [40][41]. DES has been suggested as a low cost alternative to other ionic liquids with interesting characteristics, such as negligible vapor pressure, wide types of liquid range, good stability, and tunability [41]. High purity DES can be easily prepared using biodegradable materials at low cost [42]. However, on a mass basis, the solubility of CO 2 in DES is still significantly lower than that of MEA 30 %-wt., by an order of magnitude, because of the higher molecular weight of DES [43]. The other drawbacks of DES for CO2 capture applications are their very high viscosity, low absorption rates, and high solvent cost compared to aqueous solvents [44].

Phase-Change Solvents
The CO2 absorption process involves equilibrium reaction, where an increase in the CO2 absorption capacity (moles of CO2 per mole of solvent) can be achieved by eliminating one of the reaction products during reaction (Le Chatelier's principle) through precipitation. This could decrease the required regeneration energy up to half that of MEA 30%-wt. [45][46][47]. However, this removal of reaction product is achieved by precipitation of solids, which can cause plugging and scaling in the gas-liquid contactor. A spray tower has also been suggested to prevent plugging. However, spray tower only has 10-30% surface area compared to that of packed columns. Thus, at the same capacity, a larger spray tower and hence a higher capital cost are required.
The compounds of precipitate in the absorption with phase-change solvents depends on the solubility of the reaction products and the chemical structure of the solvent. Raksajati et al. [48] proposed two classes of phase-change solvents. The S class solvents are those that react with CO2 to produce bicarbonate precipitate, for example ammonia, alkali carbonates, and amino acids with tertiary or sterically hindered secondary amino group. On the other hand, the L class solvents are those that react with CO2 and form carbamate ion solution and a zwitterion precipitate. The detailed explanation of reaction mechanism for both phasechange solvent classes is described in Raksajati et al. [48]. For both phase-change solvent classes, the reaction equilibria towards the reaction products when precipitation occurs.   [53] reported that the CO 2 absorption capacity of potassium sarcosine is 34 % higher than that of MEA. In a recent study by [49], two processes to enhance the absorption and desorption process of a precipitating amino acid solvent system was introduced (DECAB and DECAB Plus). The addition of a solid-liquid separator between the absorber and the stripper may result in a stripper feed stream with a higher solid fraction. Therefore, the pH of the rich solution changes and it consequently enhances the hydrolysis of the carbamate species producing amino acid salts and bicarbonate. This translates to a lower heat of desorption required in the stripper. Sanchez-Fernandez et al. [53] reported that the regeneration energy of DECAB and DECAB Plus are 3.3 and 2.5 MJ/kg CO2, which are approximately 15 to 40 % lower than MEA 30 %-wt.

Alkali Salts
Recently, the use of high concentration potassium carbonates for post-combustion CO2 capture has been investigated. At this concentration range (30-40 %), the solubility limit is exceeded hence bicarbonate precipitates are formed [31] The precipitating alkali carbonate systems have the same advantages as those of aqueous alkali carbonates, such as low heat of reaction, low solvent cost, negligible solvent loss, and a possibility to exclude FGD and SCR [31].
The performance of "UNO Mk 3", a proprietary precipitating potassium carbonate solvent developed by the Cooperative Research Centre for Greenhouse Gas Technologies (CO2CRC), has been reported by [31]. The solution contains potassium carbonate ranging from 35 to 50 %-wt. The CO 2 loading in the solution was reported to be between 0.3 to 0.68 and the regeneration energy of this system was reported to be 3.6 MJ/kg CO2 [31]. UNO

Encapsulated Solvent
Encapsulated solvents are one of the new processes currently under development for CO2 capture [62][63][64][65]. In the encapsulated solvent system, the solvent fluid is enclosed in a thin membrane capsule typically 100-600 μm in diameter [65]. The capsule shell can be made from polymers with typical wall thickness of 10-30 μm [65]. The working solvent is isolated within the capsule while the gas is absorbed through the capsule wall. Thus, the membrane capsule needs to be thin enough in order to prevent excessive mass transfer resistance, but also strong enough to maintain its integrity and stability.
While any class of solvents can be used within the encapsulated solvent system, for phase change solvents, encapsulated solvent systems can overcome some of the processing challenges of the traditional process design. If conventional absorption equipment such as a packed column is used, the precipitation of solids in the absorption process raises concerns of plugging [49]. However, if CO2 absorption is operated in a spray tower to prevent plugging, the capital cost for the system may be uneconomic. The use of encapsulated solvents overcomes this as the precipitates are isolated in the capsule, thus preventing plugging. The other advantage of an encapsulated solvent is that it has significantly higher surface areas compared to conventional packing [63,65]. Compared to a conventional absorption system, an encapsulated solvent system is likely to possess a drawback in performing heat recovery between the rich and lean sorbent streams. As observed in a solid adsorbent system, exchanging the heat between solid streams is not as straightforward as exchanging heat between liquid streams [66][67][68].
At the current stage of encapsulated solvent development, the optimal capsule properties such as the capsule diameter, capsule thickness, membrane capsule permeability, and so forth, have not been formulated. [57]. Key advantages of CAP include low regeneration energy, good resistance to solvent degradation, high desorption pressure and high solvent working capacity compared to MEA [33,[58][59][60]. The main drawbacks of this system are the need for refrigeration to support the lowtemperature absorption process and the high solvent volatility. An ammonia washing system is typically installed to recover vaporized ammonia at the outlet of the absorber, which contributes to the capital costs of the CAP system. It has also been reported that CAP has a low absorption rate, about 1.5 to 2 times lower than MEA [61]. There are five sections in the CO2 capture system using CAP: flue gas cooling, CO2 absorption into solution, NH3 stripping from the vent gas, CO2 regeneration, and CO2 compression [12].

One example of a carbonate-based solvent is the chilled ammonia process (CAP), which is licensed by Alstom and was installed on a plant at Mountaineer to capture 100,000 tons of CO2 per year. This project was a cooperation with American Electric Power
Different process configurations including fixed-bed circulated fluidized bed, and bubbling fluidized bed columns have been proposed as possible equipment for use as the absorber and the regenerator [69]. However, this has not been examined in detail.

Discussion
Aqueous Solvents Raksajati et al. [70] discussed the parameters that have the largest effect on the CO2 capture cost by using aqueous solvent. The authors reported that solvent stability to SOx and NOx, solvent working capacity, solvent concentration, and heat of reaction are the most important variables that affect the CO2 capture cost significantly. On the other hand, the solvent loss had a relatively small impact on the CO2 capture cost. Furthermore, in terms of process design improvement, it was found that reduction of capture cost could be achieved by employing novel heat transfer configuration to reduce the temperature difference across stripper, a high solvent regenerator pressure system, advanced column packing with high surface area, and alternate cheaper materials for process vessels.
The costs estimation of CO2 capture by using aqueous solvent was estimated to be in the range of US$62-US$80 per metric ton of CO2 avoided and US$44-US$59 per metric ton of CO2 avoided for capture plant with and without FGD and SCR, respectively [70]. Based on the parameters that are important for aqueous processes, it was reported that capture cost can be reduced significantly for solvents that have good stability to SOx and NO x, high working capacity (> 0.35 moles of CO2/mole of solvent), high concentration (> 50 wt. %), and fast (comparable absorption rate with MEA). This result may suggest that future generations of CO2 capture processes may be likely to involve a phase change during CO2 absorption that have higher solvent capacity.
To date, most of the developments for aqueous solvents have focused on amines. However, this presents a difficulty if developments continue to be based on amines, as the degradation of amines due to acid gas impurities will make the possibility of excluding the FGD and SCR very challenging. Continued development of amines solvents would require a novel approach, for example integrating the CO 2 and acid gas impurities removal, such that has been applied in the Cansolv system in Boundary Dam Project [71]. Alternatively, future development of aqueous solvents could focus on developing non-amine solvents [34].

Phase-Change Solvents
Raksajati et al. [48,72] studied the techno-economic analysis of the carbon capture using phase-change solvents. The authors estimated that the capture costs using phase-change solvents were US$52−108 per tonne CO2 avoided if the absorber can be operated in a packed column, in comparison to the cost of MEA 30% wt solvent at US$88 per tonne CO 2 avoided. Furthermore, in terms of process design improvement, the use of the additional solid-liquid separator increases the capital costs. However, the separator may reduce the heat duty for the solvents, due to a lower heat of absorption duty and water vaporization duty, or the lower sensible heat duty and water vaporization duty in the stripper [48].
For phase-change solvents, potassium carbonate, ammonia and potassium taurate are at the most advanced stage of development within this solvent class, being at various stages of demonstration field. Potassium carbonate is the main compound used in UNO Technology's UNO MK 3 solvent processes [31], ammonia is the compound used in Alstom's Chilled Ammonia Process (CAP) [58], and potassium taurate is one of the amino acid salt solvents under development by Siemens Energy and TNO [49,73].
The carbonate phase-change solvents currently tested at demonstration stage do not require FGD and SCR pretreatment. This is because the SO x and NOx are neutralised by alkali hydroxides resulting in sulphate and nitrate salts. For potassium taurate, the behavior of the solvent with SOx and NOx has not been reported in published studies. For phase-change solvents, one of the challenges identified is the need to improve absorber designs. This is because current process technologies require the use of spray towers to avoid plugging. During the pilot testing of the Chilled Ammonia Process (CAP) by Alstom power, it was revealed that managing solids in a packed column was unfeasible because of equipment clogging resulting in difficulties in plant operation and control [46]. Therefore, the use of a novel absorber or packing design to allow precipitation during absorption and regeneration is needed to advance the development of phase-change solvents, such as the proposal by CSIRO for their ammonia solvents [34]. Another important process improvement required for this solvent class is the use of internal heat recovery [48].

Encapsulated Solvents
Encapsulated solvents are still in the early development phase and have only been demonstrated at lab scale. Sodium carbonate was identified as an excellent solvent in an encapsulated solvent system following an evaluation of a broad range of solvents by Stolaroff and Bourcier [64] and Vericella et al. [65]. To date, there is no published study that reports the solvent degradation for an encapsulated solvent system, which needs to be investigated in future research. There has also been minimal research of process and equipment designs for encapsulated solvent systems. It is also crucial to improve the current process and equipment designs, especially by developing a novel CFB absorber, such as a multistage counter-current CFB absorber, in order to enhance the equilibrium conditions in the absorber.
Raksajati et al. [74] presented the techno-economic assessment of encapsulated solvent for CO2 capture using two configurations: (1) multiple fixed bed columns and (2) a circulating fluidized-bed absorber and a bubbling fluidized-bed regenerator. The authors reported that the capture costs using this process were 1.6-2 times more expensive than the conventional MEA solvent system, which is due to the extra membrane resistance in the encapsulated system which increases the regeneration energy required, the higher equipment cost, and the higher capital cost. Further improvement of encapsulated solvent may involve the use of a suitable heat recovery scheme within the process, the use of novel absorber and/or regenerator column designs, the use of solvents encased in very thin capsules, and the use of new solvents with more favourable properties [74].

Conclusion
This paper reviews the development and the key parameters involved in the design improvement of several chemical absorption process systems. These processes include chemical absorption using aqueous solvents, phase-change solvents, and encapsulated solvents systems. In addition, some guidelines for the future development has also been presented.