Arrhenius parameters of the rate of reaction (R20) from different experimental works.
Carbon capture is proposed as a viable way of exploiting the fossil resources for power plants and industrial processes. The post-combustion capture by chemical absorption in amine aqueous solutions has been in use in chemical and petrochemical areas for decades. As an alternative, the absorption in aqueous ammonia has received great attention recently. The carbon capture by aqueous ammonia is based on the conventional absorption-regeneration scheme applied to the ternary system CO2–NH3–H2O. It can be implemented in a chilled and a cooled process, depending upon the temperatures in the absorber and, hence, the precipitation of salts. The process simulation can be conducted in two manners: the equilibrium and the rate-based approaches. The specific heat duty is as low as 3.0, for the cooled process, and 2.2 MJ/kgCO2, for the chilled one. Moreover, the index SPECCA is as low as 2.6, for the cooled, and 2.9 MJ/kgCO2, for the chilled one. The overall energy performances from the simulations in the rate-based approach, compared against those in the equilibrium approach, result only slightly penalized. From an economic perspective, the carbon capture via chemical absorption by aqueous ammonia is a feasible retrofitting solution, yielding a cost of electricity of 82.4 €/MWhe and of avoided CO2 of 38.6 €/tCO2 for the chilled process.
- carbon capture
- post-combustion capture
- chemical absorption
- aqueous ammonia
- salt precipitation
- ammonia slip
The ongoing scientific debate does not focus on whether fossil fuels will have to meet a major portion of the short- and the mid-future energy demand, but rather on how they will be exploited most effectively in terms of primary energy use, environmental impact, and end-user cost. Meanwhile, renewable sources are expected to be implemented more and more diffusely to allow independence from fossils in a later future.
Carbon capture is proposed as a viable way of effectively exploiting the conventional resources. It can be implemented in a pre-combustion, a post-combustion, or even an oxy-combustion configuration. Among them, the post-combustion option has the large benefit of being readily applicable to the already existing power plants as well as industrial processes that are fueled by coal and natural gas.
The post-combustion carbon capture can be accomplished by adsorption on solid materials or by chemical absorption in liquid solutions. The chemical absorption in amine aqueous solutions has been in use for decades in a number of chemical and petrochemical areas, such as the Oil & Gas or the urea preparation. Currently, the so-called advanced amines are under research with the goal of reducing the energy demand when applied to power plants and industrial processes. As an alternative to amines, the chemical absorption in ammonia aqueous solution has received great attention during the last decade.
This chapter covers the chemical absorption of carbon dioxide by an aqueous solution of ammonia. The next sections will present, in sequence, an overview of a number of works retrieved from the open literature, the simulation by either an equilibrium- or a rate-based approach, the environmental as well as economic assessments and, lastly, the future developments of the process itself.
2. Bibliographic review
The possibility of obtaining carbon dioxide from gas mixtures attracts the attention of inventors and investigators toward the end of the nineteenth century, as narrated by Wellford Martin and Killeffer . In 1937, the two authors turn to be among the first ones to recognize the possibility of producing CO2 from the flue gases of power plants.
In the first decades of the twentieth century, the process employing aqueous ammonia for the removal of CO2 and H2S is used extensively for the purification of coke-oven gas. Carbon dioxide is indeed a major component that must be removed to greatly increase the heating value of that gas. During the following years, amines, specifically alkanolamines, become preferred over ammonia for few reasons . First, the use of amines leads to lower issues of pipe plugging and air polluting. Second, amines are characterized by higher effectiveness in capture H2S, which can be used as an affordable source of elemental sulfur. Ultimately, ammonia is still an expensive substance because the industrial ammonia production is still to be established. Historically, the first alkanolamine to become commercially viable is triethanolamine (TEA) in the year 1930.
Through the last few decades, the amines that reach commercial maturity for gas purification are monoethanolamine (MEA), diethanolamine (DEA), and methyldiethanol-amine (MDEA). In particular, MEA is taken frequently as the reference process for the carbon capture in post-combustion configuration. By contrast, the aqueous ammonia is reconsidered explicitly for carbon capture only quite recently by both research centers and industrial companies.
The following sections provide an overview of the information about the general absorption-regeneration scheme, the chemistry of solution, the thermodynamic equilibrium models, the kinetics investigations, and the aqueous ammonia process for carbon capture.
2.1. General absorption-regeneration scheme
The carbon capture by aqueous ammonia as well as aqueous amines is based on the conventional absorption-regeneration scheme, which is illustrated in Figure 1. In simple words, the gas to be treated flows upward through the absorber, countercurrent to the falling absorbing solution, and purified from CO2. The generated rich solution from the bottom of the absorber is heated in a heat exchanger, recovering energy from the lean solution (see subsequent text), and enters the regenerator at a point near to its top. In the regenerator, a heat source (such as steam) releases the captured CO2, which exits from the top of the column, while the generated lean solution from the bottom. The lean solutions flow, through the mentioned heat exchanger, to the top of the absorber closing the scheme. An exhaustive description of the absorption-regeneration scheme is provided by Kohl and Nielsen .
2.2. Chemistry of the solution
The carbon capture by aqueous ammonia is based on the ternary system CO2–NH3–H2O, which yields an electrolyte solution. At the absorber conditions, the main reactions are  as follows:
The ternary system is explored by Burrows and Lewis as early as 1912 . In a more recent work, 1982, Pawlikowski et al.  investigate vapor-liquid equilibria of many systems, including CO2–NH3–H2O, by way of the gas-liquid chromatography for temperatures ranging from 100 to 150°C. Kawazuishi and Prausnitz  provide measurements from previous works by other scientists with the scope of calibrating the expressions of dissociation equilibrium constants and Henry’s constants for temperatures in the 100–205°C interval and for total liquid-phase concentrations to 10 molal. Göppert and Maurer  report vapor-liquid equilibrium data between 333.15 and 393.15 K at pressures up to about 7 MPa for water-rich mixtures and concentrations to about 16 molal for ammonia and 13 molal for carbon dioxide. In 1992, Pelkie et al.  use conductivity measurements to estimate the ammonium ion, NH4+, concentration at a temperature of 25°C and over a wide span of pressures and concentrations. The vapor-liquid-solid equilibrium is considered in the work by Kurz et al. , which focuses on the solubility of weak electrolyte gases into the aqueous phase in the temperature range from 313 to 353 K at pressures up to 0.7 MPa. In a subsequent study, the enthalpy changes upon partial evaporation of aqueous solutions, including CO2–NH3–H2O, are reported by Rumpf et al.  at temperatures from 313 to 393 K. Finally, speciation is measured with 13C NMR by Holmes et al.  at 25 and 35°C and by Mani et al.  at room temperature.
2.3. Thermodynamic equilibrium models
As indicated, the ternary system is an electrolyte solution. The thermodynamic model for such a complex system shall account for the electric interactions among the species, including strong and weak forces. The strong forces are described by long-range terms that represent electrostatic interactions between ions. The weak forces instead by short-range terms that represent the ion dipole interactions and the non-electrostatic interactions.
Two common equilibrium descriptions are the Electrolyte Non-Random Two Liquid (e-NRTL) model  and the Extended UNIQUAC model . A comparison between them is proposed by Darde et al. . The most recent improvement of the Extended UNIQUAC model  comprises a full set of equilibrium reactions. First, speciation reactions are as follows:
Moreover, three vapor–liquid equilibrium relations are as follows:
Lastly, the four solid formations are as follows:
Figure 2 illustrates a comparison of the experimental data by Kurz et al. , indicated by hollow markers, and the computed values by way of the improved extended UNIQUAC model, indicated by lines. The agreement is generally high.
2.4. Kinetics investigations
There are relatively few investigations on the kinetics for the ternary system NH3–CO2–H2O. Hsu et al.  describe the absorption reaction kinetics of amines and ammonia solutions with carbon dioxide in flue gases. The temperature of investigation is 50°C, which is relatively high for the carbon capture process. Similarly, Diao et al.  investigate the removal efficiency of the sole ammonia solution in the 25–55°C interval and regress the parameters of the rate constant for the capture reaction in the Arrhenius form.
Among the reactions that describe the system, only a subset is expected to significantly influence the kinetics of the overall process. These kinetics-affecting reactions are as follows:
Moreover, Pinsent et al.  assess reaction (R20) by the rapid thermal method in the range of ammonia concentration between 0.027 and 0.19 mol/l. By contrast, Puxty et al.  study it by measuring the rate of CO2 absorption into a falling thin film using a wetted wall column for aqueous ammonia between 0.6 and 6 mol/L, temperature between 5 and 20°C, and the initial thin liquid film CO2 loading between 0 and 0.8 molCO2/molNH3. Wang et al.  assess the kinetic of reaction (R20) by the stopped flow apparatus in the range of temperature between 15 and 45°C, ammonia concentration between 2.0 and 16 mmol/L, and the initial CO2 between 3 and 10 mmol/L. Lastly, Jilvero et al.  study it by a different perspective. They implement an adsorption column in a commercial code taking the design parameters of an existing pilot plant and they tune the kinetics parameters against the experimental data.
Figure 3 visualizes the trends with respect to the (reciprocal of the) temperature and against the experimental data from the investigations. Each Arrhenius law fits the data well for each work. Apparently, though, the data themselves are not in complete agreement. The results from Pinsent et al. and Wang et al. are in mutual agreement, but in disagreement with those by Puxty et al. and Jilvero et al. Noticeably, Pinsent et al. and Wang et al. measured the data at low ammonia concentrations, while Puxty et al. and Jilvero et al. at high concentrations. In short, there is likely a dependence of the kinetic parameters on the ammonia concentration.
Subsequently, Lillia et al.  propose an alternative kinetics based on the two-film theory  as represented in Figure 4. Their study covers the region typical for the absorption columns: temperatures from 15 to 35°C, NH3 concentrations from 5 to 15%, and CO2 loadings from 0.2 to 0.6. The study yields an Arrhenius constant with a pre-exponential factor of 1.41 × 108 [mol/(m3s)] and an activation energy of 60,680 [J/mol]. It has a linear dependence on the CO2 concentration and a dependence on the NH3 concentration with an exponent of 1.89.
2.5. Aqueous ammonia process for carbon capture
The concept on what was going to be referred to as the novel ammonia-scrubbing process for the carbon capture is proposed by Bai and Yeh in 1997 based on experimental data .
Their work highlights the remarkable potential of high removal efficiencies, over 95%, and absorption capacities, around 0.9 kg of CO2 per kg of NH3. Shortly after, Yeh and Bai  complete another experimental campaign with the scope of comparing amine and ammonia scrubbing and they confirm the potential of the second over the first solvent. Experiments are conducted at room temperature in their first work and between 10 and 40°C in the later one. In 2005, Yeh et al.  publish the results of three-cycle absorption-regeneration tests conducted on MEA and ammonia in a batch reactor maintained at about 25°C. They also reported an approximate estimate of energy usage that is lower than the one for MEA.
In 2006, EIG Inc.  applies for a patent on the chemical absorption of the carbon dioxide into aqueous ammonia at chilled conditions. The company Alstom is engaged in its intensive development, establishing the commercial name Chilled Ammonia Process (CAP). As summarized by Lombardo et al. , the chilled process is tested first at bench scale with SRI International. Later, it is verified at pilot scale with the Electric Power Research Institute and two utilities: WE energies in its Pleasant Prairie (WI, USA) and E.ON in its Karlshamm (Sweden) plant. Ultimately, the product validation is executed in the facility of the American Electric Power in Columbus (OH, USA) and in the world’s largest test facility of the Technology Center of Mongstad (Norway). The process has evolved during the years and it is still under development by the company General Electric, which has acquired it recently .
At the same time, the process is investigated by a number of research centers. Ullah et al.  analyze the use of a capacitive deionization in the conventional scheme of the ammonia-based process to reduce the regeneration energy requirement, concluding that the reduction can be as much as 37.5%. Sutter et al.  propose instead the controlled solid bicarbonate formation to decrease the energy requirement. Precipitation, separation, and dissolution of the solid phase are realized in a dedicated process section, while the packed absorption and desorption columns operate free of solids. A similar approach is proposed by Gao et al. , pursuing the decreasing energy consumption by the addition of alcohols to reinforce the crystallization.
Bonalumi et al.  suggest to operate the process at cool conditions (20–35°C) rather than chilled (5–20°C), to minimize the load on the chillers in favor of the load on air coolers. The two processes are visualized in Figure 5. In the cool process, one chilling load is still present for the water wash on top of the absorber. Water wash is required indeed to minimize the tendency of ammonia to escape from the absorber, which is called ammonia slip.
The ammonia-based capture is proposed typically for existing coal- and natural gas-fired power plants. Nonetheless, Bonalumi and Giuffrida  consider it for an air-blown integrated gasification combined cycle (IGCC) fired with high-sulfur coal, while Pérez-Calvo et al.  for cement plants, both achieving promising indications.
3. Process simulation with the equilibrium approach
In general, the computer simulation of a chemical process can be conducted in two different manners: the equilibrium- and the rate-based approaches, depending on whether the kinetics is not taken or it is taken into consideration, respectively. The results at the equilibrium represent the performance theoretically achievable, while the rate-based the realistic one.
Evaluating the integration of a power plant with the ammonia-based capture using the equilibrium approach, as both chilled and cooled processes, Bonalumi et al.  focus on the flue gas from a coal-fired plant, as opposed to a gas-fired. The main difference is the CO2 concentration, which is in the neighborhood of 15%, on a volume and dry basis, for coal- and of 4% for gas-fired. The reference power plant is the one defined by the European Benchmark Task Force  with the scope of establishing a framework for the consistent comparison of capture technologies. The plant has a nominal net electric power output and efficiency of 754 MWe and 45.5%. The carbon dioxide flow is 160.7 kgCO2/s at a concentration of 15.2%.
In their evaluation, Bonalumi et al.  adopt values of the design parameters differentiated between the chilled and the cooled process, as indicated by Table 2. Moreover, in the chilled process, the temperature of the streams entering the absorber is 7°C, leading to a maximum temperature in the absorber of around 18°C, despite the reaction of absorption being exothermic, and promoting the salt precipitation in a wide range of concentrations of the reactants. In the cooled process, instead, the temperature of those streams is 20°C, leading to a maximum temperature in the absorber of around 27°C and preventing the solid formation.
|Ammonia initial concentration||%(mass)||20||7.5|
|Ammonia-to-carbon dioxide ratio||kmol/kmol||3.2||5|
Different indexes can be defined to assess the carbon capture performance. First, the carbon capture efficiency is defined as the ratio of the flow rates [kmol/s or kg/s] of the carbon dioxide exiting the compression island and of that entering the exhaust chilling island. As a second common performance index, the specific heat duty [MJth/kgCO2] is defined as the ratio of the reboiler heat duty [MWth] and the mass flow rate [kgCO2/s] of effectively captured carbon dioxide. However, this second index does not include the information on the capture efficiency (first index) nor on the temperature at which the heat duty is required (or, in equivalent terms, the loss of electric power generation from the steam turbine due to the steam bled for the regenerator).
A third index is adopted to solve this issue about the specific heat duty. Consequently, the new index allows to compare consistently plants characterized by different capture efficiencies, regeneration temperatures, and electric efficiency penalties. The Specific Primary Energy Consumption for Carbon Avoided () [MJth/kgCO2] is defined as
where all parameters refer to either the power plant equipped with the carbon capture or the reference plant without it: is the heat rate [MJth/MWhe], the specific CO2 emission [kgCO2/MWhe],[−] the net electric efficiency, and stays for reference.
Regarding the results for the chilled and the cooled process, Table 3 compares the predicted electric consumptions. The exhaust cooling and the absorption-regeneration sections are more penalizing for the chilled process due to the major consumption of the chillers. By contrast, the power island is more penalizing for the cooled process, on one side, because of a large contribution due to the higher amount of NH3 that must be recovered by the water wash section. On the other, because of another major contribution due to the higher specific heat duty and the higher regeneration temperature that require more steam bleeding at a higher value of pressure and enthalpy from the turbine. In addition, the compression stage is more penalizing for the cooled process since the regeneration pressure is lower. Hence, from the electric consumption, the chilled process is less penalizing than the cooled one.
|Electric power, MWe||Chilled||Cooled||Electric power, MWe||Chilled||Cooled|
|Exhaust cooling (1)|
|AC22||0.144||4.626||CO2 Compression (3)|
In its turn, Table 4 summarizes the performances for the chilled and the cooled processes and it compares them against those of the reference power plant (without any carbon capture) and a plant integrated with carbon capture in MEA aqueous solution. From the index , which is as seen the most consistent perspective for evaluating a capture technology, the cooled process is less penalizing than the chilled one, by far, than MEA.
|Electric power loss||MWe||NA||198.9||140.1||129.3|
|Net electric power||MWe||754.0||562.4||613.9||624.7|
|Net electric efficiency||%||45.5||33.5||37.05||37.70|
|Specific heat duty||MJ/kgCO2||NA||3.70||2.19||2.98|
|Specific CO2 emission||kgCO2/MWhe||763||104||141.4||138.9|
4. Process simulation with the rate-based approach
In a recent work, Bonalumi et al.  adopt the rate-based approach to assess the same cooled aqueous ammonia process that they investigated earlier with the equilibrium approach . Table 5 summarizes the main results from the comparison of the performances predicted by the two approaches. The overall energy balance for the kinetic study, compared against the equilibrium study, turns to be only slightly penalized. The authors explain that this penalization originates from the larger request of energy to achieve a higher level of CO2 purity in the lean stream from the regenerator. The differences being moderate, though, the study of an absorption capture plant with the equilibrium approach can be considered a valid method for a preliminary assessment of an ammonia-based process.
|Parameter||Unit||Cooled equilibrium||Cooled rate based|
|Electric power loss||MWe||129.3||136.4|
|Net electric power||MWe||624.7||617.6|
|Net electric efficiency||%||37.70||37.27|
|Specific heat duty||MJ/kgCO2||2.98||3.02|
|Specific CO2 emission||kgCO2/MWhe||138.9||141.2|
5. Economic and environmental assessments
The integration of the chilled process and an ultra supercritical power plant is analyzed by Valenti et al.  via a parametric analysis from the energy and the economic perspectives. The capture island is simulated with an equilibrium approach. In the parametric investigation, five parameters are varied singularly: (1) ammonia initial concentration in the aqueous solution, (2) ammonia-to-carbon dioxide ratio in the absorber, (3) regeneration pressure, (4) regeneration temperature, and (5) absorber chiller evaporation temperature. The economic analysis, with respect to a reference power plant rated at the net electric production of over 750 MWe, shows that the capital investment of the capture island is estimated to be a relatively small portion of that of the power island. However, due to other costs and due to the performance penalties, the cost of electricity increases significantly by 37.5%, from 59.90 to 82.38 €/MWhe; ultimately, the resulting cost of avoided CO2 is approximately 38.64 €/tCO2.
A detailed environmental life cycle analysis for an ultra supercritical power plant with and without carbon capture is proposed by Petrescu et al. . Three capture islands are considered: (1) gas-liquid absorption with MDEA (monodiethanolamine), (2) gas-liquid absorption with aqueous ammonia, and (3) gas-solid absorption with calcium oxide. The environmental evaluation is performed using the “cradle-to-grave” methodology considering several upstream and downstream processes. Eleven environmental impact categories, according to the method CML 2001, are compared using GaBi software. The study highlights that carbon capture technologies decrease the global warming potential indicator, but they may increase other indicators. The amine technology achieves a good performance from the perspective of global warming, but not satisfactory from that of all others. Aqueous ammonia adsorption and calcium looping prove to be better. Some indicators, such as acidification potential, eutrophication potential, or those related to lethal concentrations (e.g., human toxicity potential, freshwater aquatic ecotoxicity potential, and marine aquatic ecotoxicity potential), are better in the case of aqueous ammonia. By contrast, some others, such as abiotic depletion fossil and abiotic depletion elements, are better in the case of calcium looping.
6. Future developments
A highly promising solvent-based CO2 capture process, named the mixed-salt technology, is being currently developed, as reported by Jayaweera et al. . This technology adds potassium carbonate to the system in order to exploit the advantages of both ammonia-based and potassium carbonate-based technologies. The simplistic representation of the CO2 absorption and removal reaction is as follows:
where the CO2 loading is the numerical difference between and . The left and right side of the equilibrium represent the lean and rich solutions, respectively. Despite the system being characterized by the presence of several ionic species that could form a solid phase, precipitation is avoided by operating the absorber at relatively high temperatures and at concentrations below the solid-forming conditions. The expected advantages are a limited heat duty at the regenerator and a limited load for the water wash on top of the absorber.
This chapter covers the chemical absorption of carbon dioxide by an aqueous solution of ammonia for coal- and natural gas-fired power plants and industrial processes. It reports the literature review, the simulation by equilibrium- or rate-based approach, the economic as well as environmental assessments, and the future developments. Conclusions are as follows:
The ammonia-based technology confirms to be attractive compared to conventional amines. It can be implemented in a chilled as well as in a cooled process depending upon the temperature and, consequently, the precipitation of salts in the absorber.
The predicted specific heat duty, in the equilibrium approach, is 3.0 for the cooled process and 2.2 MJ/kgCO2 for the chilled one. Moreover, the index is 2.6 for the cooled and 2.9 MJ/kgCO2 for the chilled. Overall, the cooled process combines the advantage of a moderate energy requirement with the absence of solid formation.
The predicted performances in the rate-based approach, compared against those in the equilibrium approach, result slightly penalized. The difference is due to the need of a higher level of CO2 purity in the lean stream from the regenerator. The index value changes from 2.6, as seen in the equilibrium, to 2.8 MJ/kgCO2, in the rate-based approach, yielding an increase of the prediction of about 6%. Hence, the study of an absorption capture plant with an equilibrium approach is a valid methodology for a preliminary investigation and optimization process.
From an economic perspective, the carbon capture via chemical absorption by aqueous ammonia is a feasible retrofitting solution, yielding a predicted cost of electricity of 82.4 €/MWhe and a cost of avoided CO2 of 38.6 €/tCO2, both for the chilled process (those for the cooled process are not reported yet in the open literature).
The mixed-salt technology is a promising evolution of the process to further reduce the specific heat duty and the load for the water washing on top of the absorber.
The authors gratefully acknowledge Mr. Stefano Lillia of Politecnico di Milano for his help in investigating carbon capture alternatives as well as in reviewing this chapter.