Absorption of Carbon Dioxide in a Bubble-Column Scrubber

Understanding greenhouse gas capture, utilization, reduction, and storage is essential for solving issues such as global warming and climate change that result from greenhouse gas. Taking advantage of the authors' experience in greenhouse gases, this book discusses an overview of recently developed techniques, methods, and strategies: - Novel techniques and methods on greenhouse gas capture by physical adsorption and separation, chemical structural reconstruction, and biological utilization. - Systemic discussions on greenhouse gas reduction by policy conduction, mitigation strategies, and alternative energy sources. - A comprehensive review of geological storage monitoring technologies.


Introduction
In the current stage over 85% of world energy demand is supplied by fossil fuels. Coal-fired plants discharged roughly 40% of the total CO 2 are the main contributors in CO 2 emissions (Kim & Kim, 2004;Yang et al., 2008). Environmental issues caused by exhaust green house gases (GHG) and toxics have become global problems. Through the past studies of five decades, increased GHG levels in atmosphere is believed to cause global warming, in which CO 2 is the largest contributors. International Panel on Climate Change (IPCC) predicts that the CO 2 content in atmosphere may contain up to 570 ppmv CO 2 , causing an increase in mean global temperature around 1.9℃ and an increase mean sea level of 38m (Stewart & Hessami, 2005). Also accompanied is species extinction. Therefore, the importance of removing carbon dioxide from exhaust emissions has been recognized around the world. There are three options to reduce total emission into the atmosphere, i. e., to increase energy efficiency, to use renewable energy, and enhance the sequestration or removal of CO 2 . However, from the viewpoints of coal-fired plants discharged a lot amount of CO 2 -gas, CO 2 gas needs to be removed from the flue gases of such point sources before direct sequestration. There are several processes for CO 2 separation and capture processes, including post-combustion, pre-combustion, oxy-fuel processes, and chemical-looping combustion (Yang et al., 2008). In here, we focus on the treatment of post-combustion processes since they are typical coal-fired plants.
For the removal of exhaust CO 2 -gas, several methods have been proposed, such as chemical absorption, physical absorption, membrane separation, biochemical methods, and the catalytic conversion method. In addition to these methods, the absorption of carbon dioxide in an alkaline solution with crystallization has also been adopted to explore the removal of carbon dioxide from waste gas (Chen et al., 2008). This approach, with the production of carbonate by means of reactive crystallization, has been found to be effective. In order to remove of CO 2 gas, several scrubbers were utilized, such as sieve tray column, packed bed column, rotating packed bed and bubble column. Therefore, how to choose an excellent scrubber becomes significant in the removal of CO 2 gas from flue gas. The performances of the scrubbers were always estimated by using overall mass-transfer coefficient. Due to this, they found packed bed with structured packing (Aroonwilas & Tontiwachwuthiku, 1997) gives a higher overall mass-transfer coefficient as compared with other scrubbers. On the other hand, Chen et al. (2008) found that the performance of scrubber could be estimated by using scrubbing factor, a definition of removal of 1 mole CO 2 gas per mole of absorbent and per liter of scrubber. In this manner, a bubble-column scrubber is the first choice.
In a bubble column, gas is often introduced near the bottom of the column. The liquid may flow through the column either with or against the gas flow. Investigations of bubblecolumns from the viewpoints of both environmental problems and industrial use can be found in the literature (Juvekar & Sharma, 1973;Sada et al., 1985;Sauer & Hempel, 1987;Okawa et al., 1999;Chen et al., 2002;Hamid & Jones, 2002;Lapin et al., 2002). Bubble column reactors are widely used in the chemical, petrochemical, biochemical, and metallurgical industries (Degaleesan & Dudukovic, 1998). They were favored because of their simple construction, higher heat and mass transfer coefficients, higher removal efficiency, and effective control of the liquid residence time. In addition, bubble columns may be operated in either batch mode or continuous mode, depending on the requirements, as in processes such as liquid-phase methanol synthesis or continuous-mode Fisher-Tropsch synthesis with a liquid superficial velocity that is lower than the gas superficial velocity by at least an order of magnitude or more. In this manner, the gas flow controls the fluid dynamics of the individual phases of these systems. This in turn controls liquid mixing and inter-phase mass transfer, which subsequently influence conversion and selectivity.
In addition, some investigators (Sauer & Hempel, 1987;Hamid & Jones, 2002) have studied the dynamics, mass transfer, and control of the crystal size and shape of products for industrial purposes. However, these data cannot be utilized effectively. In gas-liquid reaction, the absorption mechanism is controlled by both the mass-transfer and chemical reaction steps, depending on the operating conditions. In order to describe the concentration distributions, several diffusion-with-chemical-reaction models have been proposed in the literature (Sherwood et al., 1975;Fan, 1989;Cournil & Herri, 2003). In these reports, the chemical reaction kinetics in gas-liquid reactions was first order or second order, depending on the pH of the solution. However, in some systems the mass action equations were found to be very complicated, and in others, it was found that the pH had an effect on the species (Morel, 1983). Therefore, the absorption of gases was strongly affected by the pH of the solution. Thus, it is clear that controlling the pH of the solution is important for controlling the absorption rates of gases, since the ionic strength and, hence, Henry's constant are affected by the pH of the solution. Additionally, several gas-liquid absorption applications, such as desulfurization processes and the removal of carbon dioxide, involve small particles (Hudson and Rochelle, 1982;Saha and Bandyopadhyay, 1992;Chen et al., 2002;Okawa et al., 1999). However, some of the above papers reported that the size and concentration of solids affected the mass-transfer coefficient, while most studies found larger sizes at higher solid concentrations Cournil & Herri, 2003). These masstransfer coefficient data cannot be used in absorption with reactive crystallization systems, such as gypsum and carbonates. Therefore, research on the effect of particles on the absorption processes is required. For design purposes, data for the hydrodynamics and mass-transfer coefficients in bubble columns is required (Bukur & Daly, 1987). In order to obtain the mass transfer coefficients in an alkaline solution with and without solids, a twofilm model was utilized to describe the absorption of carbon dioxide. The results are also compared here with those reported in the literature.
In this chapter, the materials include absorption models, determination of absorption rate, solution chemistry, determination of mass-transfer coefficients, and scrubbing factors of various scrubbers. Finally, how to estimate the size of bubble-column scrubber is also discussed in this chapter.

Absorption models
When gas is blown through the liquid as a stream of bubbles, e. g., in a sparged vessel or on a bubble-plate; diffusion, convection, and reaction proceed simultaneously. A complicated system is formed during gas-liquid contact. Absorption accompanied with chemical reaction not only can enhance the absorption rate but also can reduce the height of scrubber (Levenspiel, 1998). Therefore, in absorption accompanied with chemical reaction system, chemical reaction kinetics play an important role on the overall rate of absorption process. There are eight special cases for the absorption with reaction processes; the absorption efficiency is dependent on fast reaction or slow reaction kinetics. In order to quantitatively evaluate the absorption rate, absorption model proposed is required. Four models are observed in the literature (Danckwerts, 1970;Sherwood et al., 1975); there are the film model, still surface model, surface-renewal model, and the penetration model. In here, a two-film model is introduced, since it is a more practical model for applications, as shown in Figure 1. The gas solute in bulk phase overcomes the resistance of gas film and moves to the gas-liquid interface; later, solute moves across the interface and overcomes the liquid film resistance and goes through the bulk liquid. At the interface the Henry's law is applied: where C i is the concentration of solute at liquid side interface, P i the partial pressure of solute at gas side interface, and K H the Henry's law constant. The two-film model is widely used in the literature. In here, the Henry's law constant is dependent on the ionic strength and temperature. Fig. 1. A schematic diagram for the two-film model. ( Levenspiel, 1998) www.intechopen.com . The total amount of dissolved carbonate increases with increasing pH because of the ionization equilibrium. In dilute solutions at 25℃, the ionization constants for K a1 and K a2 are approximately 10 -6.3 and 10 -16.3 (Butler, 1982), respectively. The ionization constants depend on temperature and on the presence of other salts. Figure 2 is a plot of degree of ionization (α)versus pH at 25℃. In here, the total amount of carbonate is the summation of all carbons including H 2 CO 3 * , HCO 3 -and CO 3 2-.We express as C T . In addition, α 0 ,α 1 , andα 2 are the degree of ionization for species H 2 CO 3 * , HCO 3 -and CO 3 2-, respectively. It is found that α 0 decreases with an increasing in pH on one hand and α 2 increases with pH on the other hand, while α 1 increases with an increasing in pH when the pH is less than 6 on one hand and it decreases with an increasing with pH when the pH is higher than 10 on the other hand. This indicates that the pH value could be adjusted to desired value to control the chemical species, such as H 2 CO 3 * , HCO 3 -and CO 3 2-, presented in the solution. This could be done with an addition of alkaline solution, such as NaOH solution, and acidic solution, such as HCl solution. In order to show the effect of alkaline solution and chemical species on the concentrations of H 2 CO 3 * , HCO 3 -and CO 3 2-, an example will be discussed later.

Solution chemistry in carbonate system
For the CO 2 -H 2 O-BaCl 2 system with an addition of NaOH solution, there are a lot of chemical species presented in the solution, such as H 2 CO 3 * , HCO 3 -, CO 3 2-, H + , OH -, Ba 2+ , Cl -, Na + , and ion pair BaCO 3 0 . Due to this, the mass action equations, total mass balance equations and charge balance equation are shown in Table 1 From a given pH value and the total mass balance equations, mass action equations, and activity coefficient equations, the concentrations of chemical species can be obtained by using an ionic strength approximation method (Butler, 1964;Nancollas, 1966;Chen et al., 2004). Due to this, the ionic strength can be determined. The activity coefficients of electrolytes can be estimated from the Bromley correlation equation (Sohnel & Garside, 1992 where Q 1 and Q 2 are volumetric flow rates for barium chloride solution and sodium hydroxide solution, respectively, C 1 the feed concentration of chloride ion, and C 2 the feed concentration of sodium ion. Once the ionic strength is obtained, the Henry's law constant, defined in Equation (1), can be evaluated (Butler, 1982). For example, at a temperature of 30℃, Henry's law constant can be expressed as: Where pK H is defined as: Sometime, Henry's law constant is defined as: The Henry's law constant, H, will be used in later.

Determination of absorption rate at steady-state
In a bubble column, for a simulated flue gas (CO 2 (N 2 )-H 2 O system), as shown in Figure 3, a gas mixture containing A (carbon dioxide) and B (nitrogen) flowing into a bubble column at the bottom comes into continuous contact with an alkaline solution flowing into the column www.intechopen.com at the top. Two streams come into contact with in the column adversely. If we assume that nitrogen gas is essentially insoluble in the liquid phase and that liquid does not vaporize to the gas phase, the gas phase is a binary A-B. The absorption rate of carbon dioxide can be determined by using the material balance under steady state operation, as shown in Figure 3 for a multiple-tube plug-flow model. If we assume plug flow for the gas phase through the tube, the material balance for carbon dioxide in the ith tube at steady state is where u i is the linear velocity, S i the cross section of ith tube, r Ai the absorption rate per unit area of the ith tube, and d i the diameter of ith tube. For a total of N tubes the material balance becomes Equation (24) divided by Δz and taken to the limit, i.e., Δz→0, Equation (24) where F A is the overall molar flow rate of the gas phase, which is equal to  ii A i uSC .
Integrating Equation (26) where A i is the lateral surface area of the ith tube. These equations can be rewritten as  where A r is the mean absorption rate and A is the total surface area of the gas phase in the bubble column, which can not be obtained directly. Equation (29) shows that the absorption rate, A rA, is equal to the consumption of carbon dioxide through the bubble column. This value can be determined by measuring the concentration of carbon dioxide and the gas-flow rate: and where V L is the volume of the liquid phase, V b the volume of the bubble column, and ε L is the hold-up of the liquid phase. Since the molar flow rate of inert gas is equal to F A1 (1-y 1 )/y 1 , the molar flow rate of acidic gas at the outlet, F A2 , is F A1 [(1-y 1 )/y 1 ][y 2 /(1-y 2 )]. Thus, Equation (31) can be rewritten as Therefore, the overall absorption rate, R A , as defined by Cournil & Herri (2003), can be obtained with measurable quantities. Figure 4 exhibits the absorption rate versus gas-flow rate with precipitation at various pH values. The result shows that the absorption rate increased with an increase in gas-flow rate as well as pH value. In addition, the absorption rate also increased obviously with an increase in gas concentration, y 1 , but was only slightly affected by liquid-flow rate.

Determination of mass-transfer coefficients
According to a two-film model for CO 2 (N 2 )-H 2 O system, the relationship between local absorption rate, r A , and local individual mass-transfer coefficients based on both the gas-side and liquid-side can be written as at the interface C gi =HC Li , where H is Henry's law constant defined in Equation (2). Combining Equation (34) and Equation (35) in terms of the overall mass-transfer coefficient, the equation becomes where, C A and C LA are concentrations of CO 2 -gas in the gas-phase and liquid-phase, respectively. According to Equation (24), we assume plug flow for gas phase and wellmixed flow for liquid phase, the material balance equation between △z at steady state can be rewritten as: Where, u is the mean gas-superficial velocity based on the column cross-sectional area. Take limit for Equation (37), Substitute Equation (36) into Equation (38), In Equation (39), C AL is almost less than 4x10 -8 M when the pH value is higher than 12. In addition, H value is around 1.5 and C A is higher than 1mM. Therefore, C A ＞＞HC LA , Equation (39) could be written as We assume that L  is kept constant through the column. Integrating Equation (40) Here, 1/K G a was overall resistance, 1/k G a the gas-side resistance, and H/k L a the liquid-side resistance. For a second order reaction, by introducing the enhancement factor into Equation (45) where k 2 is the second order reaction constant and D A is the diffusion coefficient for carbon dioxide. Since Henry's constant is a function of the ionic strength and temperature, which can be adjusted with pH values and temperature, a plot of Equation (46) to obtain k G a and k L a becomes possible.

Comparisons of absorption rates overall mass-transfer coefficients
In here, we will compare absorption rate and overall mass transfer coefficient obtained in a bubble-column scrubber for five absorption systems, i.e., NaOH/

Individual mass-transfer coefficients
A plot of 1/K G a vs. H/(C B0 ) 1/2 at different gas-superficial velocity was shown in Figure 5. A linear plot was obtained in this work. The reciprocal of the intercept at zero H/(C B0 ) 1/2 was the individual gas-side volumetric mass transfer coefficient. On the other hand, the slope, 1/((D A k 2 ) 1/2 a), multiplied by H/C B0 1/2 was H/k L a, where the liquid-side mass transfer k L a could be obtained. Figure 6 is a plot of the liquid-side mass transfer coefficient vs. the gas-superficial velocity at various pH values for both with and without precipitation. The liquid-side mass transfer coefficient increased with an increase in the superficial velocity at various pH values. Both procedures show that the data are close together, larger data points for precipitation and smaller data points for no precipitation. The values obtained in this study were higher than those reported by Fan (1989), as shown in the parallelogram, while the values were close to the values reported by Sada et al. (1985). On the other hand, the trend in our data, extrapolated into smaller gas-superficial velocity, was higher than that reported in Sanchez et al. (2005). A linear regression for k L a /[OH -] 0.5 and u was found to be: The coefficient of determination becomes 0.9670. The exponent of u in here was 1.09, which falls in the exponent range of 0.58 to1.52 reported in the literature. In the same absorption system, a correlation for gas-side mass transfer coefficient obtained by using regression for no precipitation system was shown below: The exponent of u in here was 0.22, which is lower than 0.75 reported in the literature (Sada et al., 1985;Botton et al., 1980). Due to this, the individual mass-transfer coefficients, k L a and k G a, could be determined when the operating conditions were given; and hence, the overall mass-transfer coefficient was estimated by Equation (45).

Scrubbing factor
The scrubbing factor (φ) was defined as φ=(F g E /F A V b ), which means moles of CO 2 -gas to be removed per mole of absorbent and per the volume of the scrubber, estimated values for various scrubbers were also calculated and listed in Table 3. In these systems, F A is defined as the molar flow rate of absorbent solution. The result showed that φ values in here, both for systems with precipitation and without precipitation, are much higher than those in packed bed systems, indicating that bubble-column scrubbers are indeed more effective than packed bed scrubbers. The scrubbing factors for bubble columns are approximation 5fold to 20-fold as compared with packed beds. This indicates that the absorption capacity for bubble columns is higher than that for packed beds. Alternatively, comparison for various bubble columns, i.e., NaOH/BaCl 2 /H 2 O system (1), NaOH/Zn(NO 3 ) 2 /H 2 O system(2), NH 3 /CO 2 /H 2 O system(3), and MEA/CO 2 /H 2 O system(4), is carried out in here, we find that the magnitude sequence in scrubbing factors are (1)＞(3)＞(4)＞(2). In the current stage system (4) is a more popular system in the removal of carbon dioxide. However, system (3) is a more attractive process since aqueous ammonia absorbent has a lot of advantages, including higher absorption capacity, cheaper, lower energy required from the viewpoint of regeneration, no corrosion problem, and safety, as compared with MEA process (Yeh & Bai, 1999;Yeh et al., 2005). Ammonia scrubbing process is widely applied to produce ammonia bicarbonate (ABC) as a fertilizer in the Chinese chemical industry.

Scrubbers
Operating conditions K G a(1/s)

Sizes of bubble-column scrubbers
For a second order instantaneous chemical reaction system, Equation (39) can be applied for gas-side material balance: Integrating equation (39) (52) where V b (=SZ) is the volume of scrubber and F A1 is the molar flow rate of input gas. If C A ＞＞HC LA , Equation (52) could be written as: 11 11 Equation (53) shows that the size of scrubber is dependent on gas-flow rate, H (ionic strength and temperature), the pH of the solution, liquid hold-up (ε L ), and desired CO 2 output concentration (C A2 ). In here, k G a and k L a can be evaluated by using Equations (47) and (48) if pH and u are given. In addition, liquid hold-up (ε L ) can be estimated according to log-log coordinates, which can be represented well by the following Equation (Hikita et al., 1980): Some useful correlations for gas hold-up can be found in the literature (Hikita et al., 1980). Fig. 7. Gas hold-up in bubble-columns (Danckwerts, 1970) On the other hand, H can be calculated by using Equations (20) and (22)

Conclusion
In this chapter, absorption of CO 2 in a continuous bubble-column scrubber was illustrated for four absorption systems. In order to obtain absorption rate and mass-transfer coefficients in an alkaline solution with and without precipitation, a two-film model with fast reaction case was utilized to study the absorption of carbon dioxide. A multiple-tube plug flow model was developed to determine the absorption rate and overall mass-transfer coefficient. At the gas-liquid interface, the Henry's law is applied to connect between the bulk phases. In order to determine Henry's law constant, the ionic strength can be determined by using an ionic strength approximation method, if the pH of solution, the total mass balance equations, mass action equations, and activity coefficient equations are available.
Combinations of mass-transfer rate, overall mass-transfer coefficient, Henry's law constant and individual mass-transfer coefficients, the overall mass-transfer coefficient can correlate with the individual mass-transfer coefficients. For a second order reaction, by introducing the enhancement factor into correlation equation, we can obtain both gas-side and liquidside individual mass-transfer coefficients by using measurable data. The effects of process variables on the individual mass-transfer coefficients were discussed and we obtained kinetic expressions for the both. Using Taguchi's analysis for MEA/H 2 O/CO 2 system, it was found that the significance sequence influencing the absorption rate is A(CO 2 %)＞C(temperature)＞D(gas-flow rate)＞B(pH), while the significance sequence for mass-transfer coefficient is B＞A＞C＞D. According to scrubbing factor defined in here, it was found that the scrubbing factors for bubble columns are always higher than that packed beds. This illustrates a higher scrubbing capacity for bubble columns as compared with packed beds. Finally, according to the material balance model proposed in here, the size of bubble column scrubber could be evaluated when the individual mass-transfer coefficients are available.

Acknowledgement
The authors acknowledge the financial support of National Science Council of the Republic of China (NSC-99-2221-E-262-024).