Abstract

We describe the use of a gas bubbler apparatus in which the gas phase is bubbled into a fixed amount of absorbent under standard conditions as a uniform procedure for determining the absorption capacity of solvents. The method was systematically applied to determine the CO2 absorbing capacity of MEA () at several aqueous MEA (β) and gas-phase CO2 concentrations. approached the nominal CO2 absorbing capacity of MEA (720 g CO2/kg MEA) at very low β levels, increasing from to  g CO2/kg MEA as β was reduced from 30 to 2.5% (w/w). did not depend on the CO2 concentration in the inlet gas stream as long as the gas stream did not include other amine sensitive components. During the bubbling tests the outlet CO2 concentration profiles exhibited a sigmoidal shape that could be described by an exponential equation characterized by an efficiency factor () and a form factor (). Statistical analysis based on correlation analysis indicated that in all cases the experimental data fit the equation well when a was and was . The results of these experiments may be used to optimize scrubber designs for CO2 sequestration from fossil fuel derived flue gases.

1. Introduction

There are several industrial applications in which a liquid phase substance (solvent) is used to selectively absorb one or several components (pollutants) from a gas stream passing through an absorbing column (scrubber). One application of increasing interest is CO2 absorption from fossil fuel derived flue gases in thermal power plants. CO2 is the most widely produced greenhouse gas (GHG) as a result of fossil fuel combustion to satisfy the world energy demand [1, 2]. Efforts to mitigate global warming include CO2 sequestration from flue gases for either storage in the sea or empty oil wells or reconversion into CO and O2 through artificial photosynthesis [35]. Although these technologies are still in an early stage of development, amine scrubbing has emerged as the preferred method for CO2 sequestration [6]. While acid gas removal from process streams using amines is a mature technology [7, 8], flue gas scrubbing presents many new challenges still not adequately met on the scale necessary for GHG abatement [9]. Wet scrubbing techniques must improve to process large volumes of flue gas at acceptable thermal efficiencies and minimal costs [10, 11].

Extensive work has been undertaken to identify the optimum packing material geometry to improve hydrodynamic mixing and maximize mass transfer in order to minimize the size and pressure drop across the scrubber [12]. The absorption or removal efficiency (, defined in (1), where and are the pollutant concentration expressed as molar fraction at the inlet and outlet, resp.) is a means of expressing the performance of the scrubber. Several authors have erroneously referred to as a solvent property even though two scrubbers using the same solvent could have different absorption efficiencies. Consider

Amine Absorbing Capacity. Amines are ammonia derivatives in which one or more hydrogen atoms are replaced by an organic radical [12]. Monoethanolamine (MEA), diethanolamine (DEA) and methyldiethanolamine (MDEA) are the most commonly used amines in scrubbing applications. The CO2 absorbing capacity of amines is easily degraded by the presence of SO2, NO2, HCl, HF, or O2 in the gas stream. These components form irreversible byproducts that reduce the reaction rate during the absorption process and increase the complexity of the solvent recovery process.

The absorbing capacity is a solvent property defined as the maximum molar amount of pollutant absorbed per mole of solvent. This property is used to define the appropriate loading (pollutant/solvent molar ratio, ) in scrubber designs. Low loadings result in columns with low absorbing efficiencies while high loadings lead to excessive solvent requirements and high operational costs. The CO2 absorbing capacity of amines is dependent on the solvent concentration, the composition of the gas stream, and the operating temperature [13].

Amines are capable of chemical and physical CO2 absorption. Physical absorption is controlled by the thermodynamic equilibrium between CO2 molecules in the gas and aqueous phases and is described by Henry’s law [8, 14]:where is the equilibrium partial pressure of component in gas phase, the total pressure, the Henry’s law constant of component , the equilibrium concentration of component in gas phase (expressed as molar fraction), and the equilibrium concentration of component in liquid phase (expressed also as molar fraction).

The Henry’s law constant is determined in a temperature and pressure controlled sealed chamber by measuring the equilibrium concentration of the component in the gas and liquid phases using spectrophotometric or chromatographic analysis [15]. This method is appropriate for systems undergoing pure physical absorption, for example, CO2 absorption in H2O. However, it is inappropriate when the solvent exhibits chemical absorption since the method does not ensure that the solvent becomes fully saturated. Investigators have employed this method for several years, expressing their results in terms of the equilibrium partial pressure of the gas phase component and referring to these values as the solubility of the pollutant in the solvent. Tong et al. [16] combined experimental work with an extensive literature review to describe the solubility of CO2 in 30% (w/w) aqueous solutions of MEA as a function of temperature and loading [1719]. For the reader’s convenience, Figure 1 reproduces the results published. These results cannot be used to describe the absorption capacity of the solvent since the equilibrium conditions under which the data was collected do not ensure saturation of the solvent. Furthermore, these results cannot be used to determine Henry’s law constant for the MEA-H2O-CO2 system since they do not quantify the CO2 remaining in molecular form within the liquid phase and because, as mentioned earlier, the system exhibits chemical absorption.

Chemical absorption is based on reactions between CO2 and the amine. It has been reported that chemical absorption does not increase significantly with pressure [20]. There are two fundamental mechanisms for the reaction of amines (R-NH2) with CO2 [9]:For common primary and secondary amines such as MEA and DEA, reaction (3) prevails to form a stable carbamate (), requiring 2 moles of amine per mole of CO2 and thus limiting the absorbing capacity of the amine to 0.5 moles of CO2 per mole of amine, that is, 360 g CO2/Kg MEA. However unstable carbamates can hydrolyze to form bicarbonate (), as described by reaction (4). Under this condition, the nominal MEA CO2 absorbing capacity is one mole of CO2 per mole of MEA [9], that is, 720 g CO2/Kg MEA. Tertiary amines such as MDEA only follow reaction (4) [21].

The physical and chemical MEA absorption capacities are affected by temperature, pressure, presence of additional gases, and the aqueous MEA concentration.

Yeh and Bai [22] measured the CO2 absorption capacity of MEA in a semicontinuous reactor consisting of a 60 mm glass bottle containing 200 mL of solvent. The absorption capacities ranged from 360 to 380 g CO2/kg MEA using MEA concentrations of 7–35% (w/w) and gas flow rates of 2–10 SLPM of 8–16% of CO2 diluted in clean air. The reaction temperature varied from 10 to 40°C. Recently, Rinprasertmeechai et al. [23] used a stirred 100 mL reactor containing 50 mL of 30% (w/w) aqueous MEA concentration at 25°C and atmospheric pressure to obtain an absorption capacity of 0.45 CO2 moles/mole amine (324 g CO2/kg MEA) for a simulated flue gas containing 15% CO2, 5% O2, and 80% N2 and flowing at 0.05 SLPM. These two papers neither reported the outlet gas flow nor removed the O2 in the gas stream, leading to an underestimation of the CO2 absorbing capacity of MEA. Recently Kim et al. [24] reported an absorbing capacity of 0,565 CO2 moles/mole amine (407 g CO2/kg MEA) using 30 vol% CO2 diluted in N2 and a fixed flow rate of 1 SLPM monitored by a mass flow controller and gas chromatography to determine CO2 concentration at the outlet of the reactor.

The disagreements present in previous results are due to variations in testing methods, amine dilution, solvent temperature and pressure, and inlet gas composition and highlight the need for a standard method to determine the absorbing capacity of solvents. The resulting experimental data are required to optimize scrubber designs for CO2 sequestration from fossil fuel derived flue gases. We propose a standard method for the determination of absorbing capacities consisting of a gas bubbler apparatus in which the gas phase substance is bubbled into a fixed amount of absorbent under standard conditions. We systematically applied this method to determine the CO2 absorbing capacity of MEA as a function of MEA concentration and CO2 concentration in the gas stream. The saturation curves obtained during the absorption tests exhibited a sigmoidal shape that could be described by an exponential function characterized by two parameters: the shape and efficiency factors. The proper use of these factors could lead to more compact and efficient scrubber designs.

2. Materials and Methods

Figure 2 illustrates the methodology proposed to determine the chemical and physical absorbing capacity of solvents. The apparatus consists of a gas bubbler setup in which the gas stream is bubbled through a fixed amount of absorbent under standard conditions. Prior to testing, the system is tested for leaks and purged using an inert gas. Experiments are conducted under standard conditions of pressure and temperature (101 kPa, 25°C). To ensure constant temperature in the presence of exothermic or endothermic reactions, the system is placed inside a thermostated water bath. The reactor is continuously stirred to prevent stratification or inhomogeneities within the reactor. The inlet and outlet gas composition and flow are measured using well accepted methods. It is important to use a water vapor trap before measuring the outlet gas flow to prevent measurement distortions due to the presence of water in the gas stream after the bubbling process. The total gas flow across the bubbler should be as low as possible (<1 SLPM) to ensure a full interaction of the gas with the solvent. The temperature, pressure, and concentration of the absorbing substance are also monitored. The volume of solution in the bubbler is maintained at 0.5 L.

Table 1 describes the variables to be measured and the recommended values for the independent variables as well as the requirements for the sensors in terms of resolution, range, and measurement method. Several trials should be conducted to verify the reproducibility of the results.

The method was applied to the determination of the CO2 absorbing capacity of MEA at several aqueous MEA concentrations and gaseous CO2 concentrations.

3. Results

Figure 3 depicts the CO2 molar concentration of the gas phase stream at the inlet and outlet of the bubbler. It shows that, at an inlet concentration of 30% CO2, MEA concentrations lower than 50% (w/w) were not able to absorb 100% of the CO2 present in the gas stream. This low absorbing efficiency is not a property of the MEA solvent but rather a characteristic of the test apparatus and indicates that the residence time of the gas stream within the bubbler for low MEA concentrations is too low to obtain accurate measurements.

3.1. CO2 Absorbing Capacity of MEA

Using the values of ,   , , and obtained as function of time during the bubbling test (shown in Figure 3), the absorbing capacity of the solvent is determined bywhere is the molecular weight of component being absorbed, is the universal gas constant, is the standard absolute temperature, is the standard pressure, is the time, and are indices to indicate the start and end of saturation process, is the mass of solvent within bubbler, is the gas volumetric flow expressed at standard conditions, and and are theindices indicating inlet or outlet values.

Figure 4 is a comparison of the values obtained, the data reported in previous works, and the nominal CO2 absorbing capacity of MEA.

More than 100 complete sets of experiments were carried out by several collaborators. It was found that the CO2 absorbing capacity of MEA is concentration dependent, increasing from to  g CO2/kg MEA when was reduced from 30 to 2.5% (w/w) and logarithmically approaching the nominal absorbing capacity of 720 g CO2/Kg MEA at very low concentrations. Table 2 lists the average values and the experimental error observed.

Changes in the CO2 absorbing capacity with solvent dilution were also observed by Yeh and Bai [22] for the NH3/H2O/CO2 system. Changes in the CO2 absorption capacity of MEA with concentration may be explained by considering that excess water favors reaction (4) and that this reaction leads to a nominal absorbing capacity twice that obtained through reaction (3). Therefore, low concentrations of MEA result in maximal CO2 absorption at the expense of reducing the interaction between the CO2 and MEA molecules and lower probability of reaching full amine saturation in a reasonable time. Changes in the CO2 absorption capacity of MEA with solvent dilution could also be due to solvation effects.

These results define the technological challenge in establishing optimum scrubber operating conditions. High MEA concentrations ensure 100% removal efficiency but provide low CO2 absorbing capacities and increase the amount of MEA required in the process. On the other side, low concentrations provide high CO2 absorbing capacity but low removal efficiency. It is possible that a sequential two-step process could be the most cost-effective means of achieving these opposing objectives.

Figure 4 also compares the CO2 absorbing capacities of MEA measured in these experiments with those reported in previous works. Although the results are not fully comparable since they were obtained under different conditions, Figure 4 shows that the values are similar. The most relevant difference with Yeh and Bai [22] and Rinprasertmeechai et al. [23] was the presence of O2 in the gas stream and with Huertas et al. [7] was the presence of H2S in the gas stream. Besides CO2, MEA can absorb H2S, SO2, and HCl [22]. MEA is degraded by the presence of O2, NO2, SO2, HCl and HF [25]. Therefore, in the determination of the CO2 absorbing capacity of MEA it is important to eliminate the interference of these species.

Figure 4 also shows that the absorbing capacity was independent of gas phase CO2 concentration. It was found that this conclusion is true as long as the gas stream does not include MEA sensitive components such as O2 and H2S.

It could be argued that the increase in MEA absorbing capacity at low concentrations is due to the contribution of the CO2 absorbing capacity of water. Therefore, a set of experiments was performed to determine the CO2 absorbing capacity of pure water. Using the present methodology, it was found that water absorbed 0.3 g CO2/kg H2O, a negligible amount compared to the variations in CO2 absorption capacity observed in aqueous MEA solutions. Since water is only capable of physical CO2 absorption, this measurement was compared to the value obtained from Henry’s law constant. For the conditions under which the experiment was carried out, Henry’s constant is 144 MPa [26] and the CO2 absorbing capacity of water at standard conditions is 0.375 g CO2/kg H2O. This agreement demonstrates the ability of the proposed method to measure both chemical and physical absorption.

3.2. Characterization of the Saturation Process

Figure 3 indicates that the outlet CO2 concentration profiles during the bubbling tests exhibited a sigmoidal shape and could be fitted to the following equation: where is the efficiency factor, is the form factor, is the time, and and are indices to indicate the start and end of saturation process. and may be obtained by linear curve fitting when (6) is expressed as follows:The correlation coefficients obtained from curve fits for all cases were near unity (), indicating that the experimental data fit well with (6). This demonstrates that the saturation process was well represented by and and these two parameters uniquely characterize the solvent absorbing capacity.

Figure 5 contains plots of the results for and . It can be observed that the factor form and the efficiency factor were not concentration dependent ( and .

These factors may be used to estimate the CO2 absorption capacity of MEA at any aqueous concentration, to compare different solvents and to determine the saturation time during the bubbling test.

3.3. Sensitivity Analysis

According to (5), is a function of pressure, temperature, gas phase CO2 concentration, volumetric flow rate, and saturation time. Applying the equation of error composition ((8), where is the absolute value of the partial derivative of with respect to each independent variable [27]) to (5) and considering the precision of the instruments specified in Table 1 () and the range of the values typically measured by each variable (also specified in Table 1), the uncertainty of the values obtained for () is less than 1% of the reported values. The CO2 concentration and volumetric flow had the greatest effect on the absorbing capacity determination, and special attention should be given to the accuracy and precision of instruments used to monitor these two variables. Table 1 includes the approximate percent contribution of each variable to the total uncertainty of the values obtained for using the bubbling test. Consider

4. Conclusions

A standard test is described for the determination of the physical and chemical absorbing capacity of gas phase components by liquid phase absorbers. It consists of a gas bubbler apparatus in which the gas stream is bubbled into a fixed amount of absorbent under standard conditions. Sensitivity analysis indicated that gas composition and volumetric flow are the variables with the greatest effect on the absorbing capacity determination and special attention should be given to the accuracy and precision of the instruments used to monitor them.

This method was applied to determine the CO2 absorbing capacity of MEA () at several aqueous MEA concentration levels () and gaseous CO2 concentrations. It was found that approaches the nominal CO2 absorbing capacity (720 g CO2/kg MEA) at very low , increasing from to  g CO2/kg MEA when was reduced from 30 to 2.5% (w/w). These results agree with values reported for in previous studies. As expected, the CO2 absorbing capacity of MEA did not depend on the CO2 concentration in the inlet gas stream as long as the gas stream did not include other components that could react with the amine, such as H2S or O2.

During the bubbling tests the outlet CO2 concentration profiles exhibited a sigmoidal shape that could be described by an exponential equation containing an efficiency factor () and a form factor (). Statistical analyses based on correlation analysis revealed that in all cases the experimental data fit well to that equation when was 6.1 ± 0.35 and was and therefore these two parameters characterize the CO2 absorbing capacity of MEA under standard conditions.

Symbols

Efficiency factor
CO2 absorbing capacity of MEA (MEA)
Henry’s constant of component (kPa)
Mass of MEA within the bubbler (kg)
Molecular weight of the component being absorbed (kg/kmol)
Form factor
Standard pressure (kPa)
Equilibrium partial pressure of component in gas phase (kPa)
Gas volumetric flow expressed at standard conditions (m3/s)
:Universal gas constant (kJ/kmol K)
SLPM:Standard liters per minute
Time (s)
Standard absolute temperature (K)
Equilibrium concentration of component in the liquid phase expressed as molar fraction
Equilibrium concentration of component in gas phase expressed as molar fraction
Loading (moles of CO2/mole of amine)
Aqueous MEA concentration (kg of amine per kg of water)
Removal efficiency (%)
Index for inlet and outlet, respectively
Index to indicate the start and end of the saturation process, respectively.

Conflict of Interests

The authors declare that there is no conflict of interests regarding the publication of this paper.

Acknowledgments

This project was partially financed by the National and Estate Mexican Council of Science and Technology (CONACYT and COMECYT), the MOPESA Company of México, the Global Institute of Sustainability of Tecnológico de Monterrey of Mexico, and the EAN University of Colombia. The authors also express their gratitude for contributions to this work from Engineers Maryin Rache and Johana Diez of the National University of Colombia.