Mass Transfer Performance of CO2 Absorption into Aqueous Solutions

Apr 16, 2012 - Kaiyun Fu , Teerawat Sema , Zhiwu Liang , Helei Liu , Yanqing Na , Huancong Shi , Raphael Idem , and Paitoon Tontiwachwuthikul. Industr...
0 downloads 0 Views 1MB Size
Article pubs.acs.org/IECR

Mass Transfer Performance of CO2 Absorption into Aqueous Solutions of 4-Diethylamino-2-butanol, Monoethanolamine, and N-Methyldiethanolamine Abdulaziz Naami, Mohamed Edali, Teerawat Sema, Raphael Idem,* and Paitoon Tontiwachwuthikul International Test Centre for CO2 Capture, Faculty of Engineering, University of Regina, Regina, Saskatchewan, Canada, S4S 0A2 ABSTRACT: The mass transfer performance of the absorption of CO2 in an aqueous solution of monoethanolamine was evaluated experimentally in a lab-scale absorber packed with high efficiency DX structured packing and compared with that of methyldiethanolamine (MDEA) as well as that of a newly developed tertiary amino alcohol, 4-diethylamino-2-butanol (DEAB). The absorption experiments were conducted at atmospheric pressure, using a feed gas mixture containing 14.9% CO2 and 85.1% nitrogen in an absorption column containing DX structured packing. The absorption performance was presented in terms of the CO2 removal efficiency, absorber height requirement, effective interfacial area for mass transfer, and overall mass-transfer coefficient (KGav). In particular, the effects of parameters such as inert gas flow rate and liquid flow rate were compared for both DEAB and MDEA. The results show that the DEAB has a much higher removal efficiency for CO2 along the height of the column than MDEA. Also, the KGav of DEAB was much higher than that for MDEA. For all the solvents, the KGav increased as the liquid flow rate was increased. An empirical correlation for the mass transfer coefficient for the CO2-DEAB system has been developed as a function of the process parameters. In terms of comparison, the results show that the DEAB system provided an excellent overall mass transfer coefficient, which is higher than that of the MDEA system but less than that of MEA.

1. INTRODUCTION Carbon dioxide is a greenhouse gas and substantially contributes to global warming and climate change. One option for reducing CO2 emissions is postcombustion capture from power plant flue gases. Many countries have agreed to reduce their emissions of greenhouse gases into the atmosphere to help prevent dangerous alterations to the climate system. Several mature technologies are available for CO2 capture, including adsorption, cryogenics, membrane technologies, and absorption.1 According to Astarita,2 absorption is the most commonly used process when it comes to gas treating. Gas absorption by chemical solvents such as aqueous solutions of alkanolamines is one of the most effective methods for CO2 removal. This technology has been used in industry for over half a century. Presently, the most commonly used chemical solvents are alkanolamines, and these alkanolamines can be classified into three chemical categories: primary, secondary, and tertiary amines. Alkanolamines commonly used are monoethanolamine (MEA) and methyl diethanolamine, (MDEA). MEA is highly reactive with CO2 but has the limitation of low CO2 absorption capacity (an equilibrium absorption of 0.5 mol CO2/mol of MEA) and high heat of regeneration; on the other hand, MDEA has high CO2 absorption capacity, and low heat of regeneration, but it is limited by slow kinetics. The main challenge for CO2 postcombustion capture from power plant flue gases is the reduction of the energy requirement for solvent regeneration. The required reduction can only be achieved by application of new solvents. Recently, the International Test Centre for Carbon Dioxide Capture (ITC), Saskatchewan, Canada, has developed new chemical solvents based on amino alcohols for CO2 capture to overcome the drawbacks of these conventional amines. The synthesis involves a systematic modification of the structure of amino alcohols by an © 2012 American Chemical Society

appropriate placement of the substituent, especially hydroxyl function, relative to the position of the amino group, in order to promote CO2 capture performance. Maneeintr et al.,3,4 have shown that the new amino alcohols have much higher CO2 absorption and cyclic capacities compared to that of conventional amine, MEA. Details of the solubility and physical and transport properties of one of the new solvents, DEAB, can be found in the literature.5−7 It is wellknown that the overall performance of any solvent depends not only on its absorption capacity and energy efficiency, but also, on the mass transfer characteristics and the kinetics. The present work focuses on mass transfer and compares the performance of the potential new solvent with the conventional solvents that can be used for the capture of CO2 from industrial flue gas. Consequently, a primary amine (monoethanolamine (MEA)) compared with a new chemical solvent, an amino alcohol (4-diethylamino-2-butanol (DEAB)), and a tertiary amine (methyldiethanolamine (MDEA)) were selected to comprehensively study the new solvent’s mass transfer characteristics in a packed-bed absorption column. The process of CO2 absorption in a packed tower depends mainly on the contact between the flue gas and the liquid solvent. In the last 35 years, tray columns have been replaced in large part by packed columns.8,9 This has been due to the successful development of efficient packing that provides a high absorption capacity per unit volume of packing. Inside the absorption column, the packing provides a contact area, which separates the liquid flow into droplets. This increases the Received: Revised: Accepted: Published: 6470

April 18, 2011 March 20, 2012 April 16, 2012 April 16, 2012 dx.doi.org/10.1021/ie2008357 | Ind. Eng. Chem. Res. 2012, 51, 6470−6479

Industrial & Engineering Chemistry Research

Article

Then, KG can be expressed as follows:

amount of liquid surface area that is exposed to the gas phase, enhancing absorption rates. Good packings should offer characteristics such as large surface area to volume ratio, low pressure drop across the absorber, and good uniform distribution between the gas and liquid phases through the column. On the basis of these characteristics, Aroonwilas9 and deMontigny10 have shown that structured packings offer a superior performance when compared to random packings. Fernandes et al.11 studied structured packing and found that the high surface area of this packing increases the mass transfer efficiencies. The objective of this work was to perform a comparative study of the mass transfer performance for CO2 absorption in conventional amine, MEA, MDEA, and the new chemical solvent DEAB using a mass transfer experimental laboratoryscale column with DX structured packing (27.5 mm) ID (from Sulzer Chemtech Canada, Inc.) for the three solvents under similar conditions. The mass transfer performance of the three solvents is presented in terms of the overall mass transfer coefficient.

⎞ ⎛ NA ⎟⎟ K G = ⎜⎜ ⎝ [P(yA − yA*)] ⎠

In a gas absorption apparatus such as a packed column, it is more useful to represent rates of absorption in terms of volumetric overall mass transfer coefficients, represented by the term KGav, instead of the one based on the interfacial area unit because the gas−liquid interfacial area cannot be measured accurately. Therefore, it is more useful to present the mass-transfer coefficient based on the unit volume of the absorption column rather than based on the interfacial area unit as follows: ⎞ ⎛ NAa v ⎟⎟ K Ga v = ⎜⎜ ⎝ [P(yA − yA*)] ⎠

(1)

(2)

(3)

(4)

The concentration gradients (yA,G − yA,i) take place over extremely small distances, which makes it difficult to measure the concentration of component A at the gas−liquid interface. Under this situation, it is more practical to express the mass flux in terms of the overall mass transfer coefficient and the mole fraction of component A in the gas phase (yA*) in equilibrium with the concentration of A in the bulk liquid as follows: NA = K GP(yA,G − yA*)

K Ga v P(yA,G − yA*) dZ = G I dYA,G

(9)

(10)

This approach has been used successfully by Aroonwilas and Tontiwachwuthikul,9 deMontigny,10 and Maneeintr7 for analyzing packed columns. In this study, the experiments of CO2 absorption were carried out in test columns packed with structured packings. The variables presented in eq 10 can be obtained from the absorption experiments. The inert gas flow rate (GI) is an operating condition of the experiment, which is conducted at atmospheric pressure (P). The CO2 concentration in the gas phase (yA,G) can be measured along the length of the packed column. The yA* is the concentration of solute A in equilibrium with the bulk concentration CA,L. The measured CO2 concentration can be converted into mole ratio values (YA,G) and plotted against the height of the column to obtain the solute mole ratio concentration gradient (dYA,G/dZ), as shown in Figure 1, and the advantage of this technique is that it allows the calculation of the KGav value at any specific mole ratio values along the column. When comparing the KGav value for two solvents, it is important to make the comparison with similar experimental conditions. For example, to compare the effect of the liquid flow rate on KGav values for two experiments, the solution concentration, the mole ratio, and CO2 loading condition in the packed column should be the

The overall mass transfer coefficients for chemical absorption can be expressed as a function of enhancement factor (I):

1 1 H = + 0 KG kG IkL

(8)

⎛ ⎞⎛ d Y ⎞ GI ⎟⎜ A,G ⎟ K Ga v = ⎜⎜ ⎟ ⎝ P(yA,G − yA ) ⎠⎝ dZ ⎠

Strigle13 described the relationship between the overall mass transfer coefficients and the individual mass transfer coefficients as follows: 1 1 H = + KG kG kL

⎛ y ⎞ NAa v dZ = G I d⎜⎜ A ⎟⎟ ⎝ 1 − yA ⎠

where YA,G represents the mole ratio of component A in the bulk gas and GI represents the molar flow rate of total gas without component A. The final equation that determines the overall mass transfer coefficient, KGav, can then be defined as follows:

where P represents the total pressure of the system and yA,G and yA,i represent the mole fraction of component A in gas bulk. The gas mass transfer coefficient kG is difficult to measure because of the change in interfacial area with varying gas flow rates in the packed columns. Instead, the overall mass transfer coefficient for the gas phase (KG) is often used,12 and it can be presented in terms of the Henry’s law constant (H) as follows: NA = K G(PyA,G − HCA,L)

(7)

The term NAav can be determined from the absorption experiments in packed columns where the concentration profile of the absorbed component A in the gas phase can be measured along the column height, and this allowed us to evaluate the KGav value. Considering an element of the absorption column with height dZ, in Figure 1, the mass balance of component A can be given as follows:

2. DETERMINATION OF THE OVERALL MASS TRANSFER COEFFICIENT Mass transfer occurs when a component A in a gas phase transfers across a gas−liquid interface into a liquid phase. The component A is transferred from the gas phase into the liquid because of the concentration gradient in the direction of mass transfer within each phase. The mass flux of component A (NA) across the gas−liquid interface at steady state can be represented in terms of the gas-side mass-transfer coefficient (kG) and driving force (yA,G − yA,i) as follows: NA = K GP(yA,G − yA, i )

(6)

(5) 6471

dx.doi.org/10.1021/ie2008357 | Ind. Eng. Chem. Res. 2012, 51, 6470−6479

Industrial & Engineering Chemistry Research

Article

Figure 1. Mass transfer processExperimental schematic diagram.

GI = inert gas flow rate and yA = gas phase CO2 concentration, and mass.error% is the error in the experiment. In the literature,9,10 it is assumed that the term of y*A is equal to zero because the compound A is quickly consumed in the fast chemical reaction and is difficult to measure, but in this work we measure the term of yA*.

same, or at least in the same range, to ensure that the calculation of KGav values are being compared fairly. For example, if two experiments had the same mole ratio value, but one experiment had a solution CO2 loading of α = 0.15 mol CO2/mol amine and the second a value of α = 0.36 mol CO2/ mol amine, the comparison would not be fair, since the second experiment had a reduced capacity for absorption due to its high loading and relevant low free-amine concentration. To determine the loading at a specific height of the column and CO2 concentration, we used eq 11:

3. THERMODYNAMICS FOR CO2 ABSORPTION Knowing the thermodynamic properties of amine solution, such as the physical solubility of CO2 or the CO2 Henry’s constant, physical diffusivity of CO2, equilibrium constants for CO2-amine system, viscosity, and density, is required in understanding CO2 absorption. 3.1. Solubility and Physical properties. In this study, the solubility of CO2 in aqueous amine solution was used to measure the term y*A as shown in eq 12. PCO2 yA* = HeCO2 ‐ amine (12)

α = αlean.loading + ((((D1 − D2) × A × time) /(M × L)) × mass. error%)

(11)

where α is the CO2 loading at any height of the column, αlean.loading is the loading of solvent coming in at the top of the column, M is the solvent molarity, L is the liquid flow rate, A is cross sectional area of column, and, D = (GI × yA) where 6472

dx.doi.org/10.1021/ie2008357 | Ind. Eng. Chem. Res. 2012, 51, 6470−6479

Industrial & Engineering Chemistry Research

Article

where y*A is the equilibrium concentration of gas at the interface, HeCO2−amine is the solubility of CO2 in aqueous amine solution, kPa m3/kmol, and PCO2 is the partial pressure of CO2 in kPa. The term of HeCO2‑amine can be determined from eq 13:13 ⎛ HeCO ‐ H O ⎞ 2 2 ⎟⎟ HeCO2 ‐ amine = He N2O ‐ amine⎜⎜ ⎝ He N2O ‐ H2O ⎠

was stirred at 0 °C for 4 h. Aqueous 1 M NaOH (10 mL) was added slowly at 0 °C. After gas evolution had subsided, excess solid NaCl was added to the mixture. The mixture was thoroughly extracted with CH2Cl2 (2 × 100 mL), the organic extracts were dried (Na2SO4), filtered, and the filtrate was distilled at atmospheric pressure to remove CH2Cl2. The resulting oil was then fractionated using a viqreux column to afford 4-(diethylamino)-2-butanol (48.3 g, 49%): bp 96−111 °C (22 mmHg). IR _max: 3075−3600 cm−1; 1H NMR, ı (200 MHz, CDCl3): 0.73 (t, 6H, J = 7 Hz, MeCH2), 0.84 (d, 3H, J = 6 Hz, MeCHO), 1.10−1.30 (m, 2H, CH2), 1.95−2.15 (m, 2H, NCH2), 2.20−2.45 (m, 4H, N(CH2Me)2), 3.52−3.65 (m, 1H, MeCHO); OH was not observed. Fractional distillation was used to purify the chemical, the identity of which was later verified by using the nuclear magnetic resonance (NMR) technique. Approximately, the purity of DEAB was 94.5% and the impurities were 4-methoxy-2-butanol, 2.3%; 2-butanol, 1.9%; and diethyl amine1, 0.5%. At this point, DEAB is synthesized in-house on a small scale. MEA and MDEA were obtained from Fisher Scientific with a purity of 99+%. Nitrogen and CO2 were supplied by Praxair Inc., with purities of 99.9%. All materials in this study were used as received without further purification. The absorption performance of the amines was evaluated by conducting experiments in a bench-scale absorption unit, of which a simplified flow diagram is given in Figure 1. The unit consisted of a glass absorption column (27.5 × 10−2 m in diameter and 2.15 m in height) packed with 37 elements of stainless steel structured packing (Sulzer DX) and (900 m2/m3). The column was designed for a counter-current mode of operation in which a liquid solution was introduced to the column at the top while a gas mixture entered the column below the packing section. A series of gas sampling points were also installed at regular intervals along the sides of the column to allow measurements of the gas-phase CO 2 concentration during experiments, and a series of thermocouple points was also installed along the sides of the column to allow measurements of the temperature of liquid during the experiments. The absorption unit was also composed of (i) two 35-L solution tanks, (ii) three calibrated mass flow meters, (iii) three needle valves, (iv) a variable-speed gear pump, and (v) an infrared (IR) gas analyzer. The solution tanks served as reservoirs for supplying and receiving the liquid solution used in the experiments. Two mass flow meters from Aalborg Instruments & Controls Inc. (model GFM 17) were used to measure the flow rates for N2 and CO2 streams, respectively, and the third mass flow rate from Bios International Corp. with high efficiency was used to measure the mixed gas stream before entering the column. The gear pump (Cole-Parmer) was used to drive the liquid solution to the top of the column. The IR gas analyzer (model 301D, Nova Analytical Systems, Inc.) was operated during the experiments to withdraw samples and measure the CO2 concentration of the gas mixture inside the column. This analyzer can measure CO2 concentrations up to 20%. Prior to the experiments, an aqueous solution of the amine was prepared in the feed tank by diluting the concentrated amine with deionized water to a given concentration. The feed tank is connected to nitrogen balloons to keep the solvents under a blanket of nitrogen. The actual amine concentration was determined by titration with a standard 1.0 kmol/m3 hydrochloric acid (HCl) solution using methyl orange indicator. Horwitz (1975)15 presented the official analytical

(13)

The solubility of pure DEAB solvent can be determined from the solubilities of CO2 and N2O for water based on the work of Versteeg and van Swaaij.14 The correlations are described in eqs 14 and 15 below: ⎛ 2284 ⎞ ⎟ He N2O ‐ H2O = 8.55 × 106 exp⎜ − ⎝ T ⎠

(14)

⎛ 2044 ⎞ ⎟ HeCO2 ‐ H2O = 2.82 × 106 exp⎜ − ⎝ T ⎠

(15)

The T in eqs 14 and 15 can be used from the experimental data of the packed column by determining the solubility of CO2 in aqueous amine solution; then, the term of yA* can be calculated in eq 12. yA* =

PCO2 HeCO2 ‐ amine

(12)

From eq 12, we found from our calculations that yA* equals 2.219 × 10−8. This is very close to 0. Therefore, practically, the CO2−DEAB reaction can be taken to be instantaneous. The physical properties based on the molarity of amine solution were derived from the work of Maneeintr et al.6 in which the physical and transport properties of aqueous amines and amino alcohol solutions for CO2 capture were measured. They measured densities, viscosities, and refractive indices of various concentrations of 4-diethylamino- 2-butanol + water mixtures.

4. EXPERIMENTAL SECTION 4-Diethylamino-2-butanol (DEAB) with structural formula as given in Scheme 1 was synthesized in the CO2 laboratory at the Scheme 1. Structural Formula of 4-Diethylamino-2-butanol (DEAB)

University of Regina, Regina, Canada. Details of the synthesis procedure are given elsewhere.4 A brief description of the procedure is however given as follows: to diethyl amine (49.7 g, 0.68 mol) was added methyl vinyl ketone (51.4 g, 0.73 mol) dropwise with stirring over a period of 1 h. The reaction temperature was maintained below 0 °C using an ice-salt bath. After the addition was complete, the cooling bath was removed, and the mixture was stirred at room temperature. Then the reaction mixture was dissolved in MeOH (20 mL). The solution was cooled to 0 °C and NaBH4 (30.7 g, 0.81 mol) was added, portionwise, to the stirred reaction mixture. The mixture 6473

dx.doi.org/10.1021/ie2008357 | Ind. Eng. Chem. Res. 2012, 51, 6470−6479

Industrial & Engineering Chemistry Research

Article

were conducted to test the working order of the packed column. Verification of the apparatus was performed by comparing results with the work done by deMontigny,16 which detailed the absorption of CO2 from a simulated flue gas stream (air + CO2) using a MEA solution. The results indicated that increasing liquid flow rate improves the KGav value, as shown in Figure 2. The results served as the basis for comparison with the new absorption experiment. The results of the comparison between the two studies are shown in Figure 2. The new data follows the same trend and

chemistry technique to determined the solution CO2 loading. By adding excess 1.0 N HCI acid to the sample, all of the CO2 evolved into a gas buret for measurement. Each experimental run began by introducing N2 and CO2 gases from cylinders through mass flow meters at desired flow rates to produce a CO2−N2 gas mixture, which was fed to the bottom of the column. The concentration of CO2 in the feed gas was checked by the IR gas analyzer and adjusted until the desired value was obtained. The prepared solvent was then pumped at a given flow rate to the column top, and a needle valve was installed on the top of the column to control the liquid flow rate so as to create counter current contact between gas and liquid. After absorbing CO2 and traveling through the column, the CO2-rich solution was collected continuously in the liquid receiving tank. This operation was continued for at least 30 min to allow the system to reach steady state conditions. At this point, gas phase-CO2 concentrations at different positions along the column were measured through a series of sampling points using the IR gas analyzer. At the same time, liquid samples were taken from the bottom of the column and analyzed for their concentrations and CO2 loading.

5. RESULTS AND DISCUSSION The experiments for mass transfer in a packed column were divided into three parts: absorption into MEA, MDEA, and DEAB. The absorption process was conducted in a countercurrent mode at preset operating conditions. At steady state operation, the gas concentration, and the temperature profiles along the column were measured and recorded. As well, the outlet liquid composition was analyzed for its CO2 loading. The overall mass transfer coefficient can be calculated using eq 10. The total number of measured data points from 27 experimental runs was 540, which includes 270 measured points of CO2 gas concentration and 270 measured points of temperature. From these data, the concentration and temperature change along the columns were measured. The effects of liquid flow rate, solution concentration, and temperature on KGav were then examined. The experimental equipment used for this study was verified by comparing the results for 2 M MEA with results from a similar setup for 2 M MEA reported in the literature by deMontigny.16 A mass balance calculation using eq 16 was performed at the end of each experiment in order to confirm the validity of the run. The calculation compared the amount of CO2 removed from the gas phase, as measured by IR, with the amount of CO2 added into the liquid phase, as measured by the solution CO2 loading apparatus. Theoretically, these two values should be equal, but because of the errors in equipment, such as the IR analyzer, CO2 loading measurement apparatus, and the large, elongated liquid trap, the mass balance error obtained in this study was about 2−6% and the percent average absolute deviation (AAD) for mass balance was 14.7% which is acceptable for mass transfer studies. mass balance error (%) ⎛ absorbed − removed CO2 ⎞ =⎜ ⎟ × 100 absorbed CO2 ⎠ ⎝

Figure 2. Verifying the packed column performance for 2 M MEA (both studied at Y = 0.09 (mol CO2/mol air).

falls in the same range as the data published by deMontigny.16 There are a few minor differences between the experiments from the two studies. The new experiment had lower gas and liquid flow rate than deMontigny’s16 work. However, since the results are in the same range and follow a similar trend, the packed column was determined to be in good working order. 5.2. Effect of Liquid Flow Rate in the Packed Column. Liquid flow rate has a great impact on the CO2 absorption efficiency of the DX structured packing. As shown in Figure 3,

Figure 3. CO2 concentration profiles at different liquid flow rates (m3/ (m2·h)) for MEA, DEAB, and MDEA (amine = 2.0 (kmol/m3); GI = 17.85 (kmol/(m2·h)); α = 0.14 (mol CO2/mol amine); inlet gas T = 23.2 °C; inlet liquid T = 23.6 °C).

(16)

5.1. Verification of the Packed Column. Maneeintr et al.8 had more liquid channeling in their 1-inch column setup, so the new setup has many improvements on the 1-inch column experiment, as shown in Figure 1, and eliminated channelling in the column. To verify the experimental equipment, a few trial runs

increasing liquid flow rate caused a reduction in CO 2 concentration of gas-phase, indicating higher absorption 6474

dx.doi.org/10.1021/ie2008357 | Ind. Eng. Chem. Res. 2012, 51, 6470−6479

Industrial & Engineering Chemistry Research

Article

efficiency. This was a result of the basic phenomena taking place, so an increase in liquid flow rate led to a greater degree of wetted packing surface and increased the efficiency of the masstransfer. The effect of the liquid flow rate on KGav values are shown in Figure 4 for MDEA, DEAB, and MEA. Again, it is evident that

packing columns of less than 50 mm diameter, so the maximum liquid flow rate should be 5 m3/(m2·h). 5.3. Effect of Loading. The value of KGav decreases with CO2 loading, which is because as the CO2 loading increases, the amount of active amines decreases, causing the KGav to decrease. The relationship between system efficiency and concentration of active absorbent, which is represented in terms of CO2 loading, is shown in Figure 5. The higher CO2

Figure 4. Effect of liquid flow rate in the packed column on KGav (amine = 2.0 (kmol/m3); GI = 17.85 (kmol/(m2·h)); YMEA = 0.09 (mol CO2/mol air); YMDEA, DEAB = 0.14 (mol CO2/mol air); α = 0.14 (mol CO2/mol amine); inlet gas T = 24.6 °C; inlet liquid T = 23.2 °C) (vertical axes: RHS refers to MEA; LHS refers to MDEA and DEAB).

Figure 5. Effect of lean loading on KGav in the packed column (amine = 2.0 (kmol/m3); GI = 17.85 (kmol/(m2·h)); Y = 0.13 (mol CO2/mol air); inlet gas T = 24.6 °C; inlet liquid T = 23.2 °C) (vertical axes: RHS refers to MEA; LHS refers to MDEA and DEAB).

the KGav values for MEA are higher than the values for MDEA and DEAB under corresponding operating conditions, and this is because MEA is a primary amine whereas DEAB is a tertiary amino alcohol and MDEA a tertiary amine. The new chemical amino alcohol DEAB provides a greater KGav value than MDEA by 30% and this led to increased efficiency of the column. Also, the results show that the liquid flow rate has an influence on the value of KGav (i.e., an increase in liquid load generally yields a greater KGav value). The possible reason for this behavior is that a higher liquid load leads to the following: (i) a greater liquid side mass transfer coefficient (kL), which is directly proportional to the overall KGav in the case of liquid-phase controlled mass transfer; (ii) a greater effective area, which is caused by more liquid spreading on the packing surface; and (iii) the amount of free amine molecules in the system becoming larger or the system having more capacity to absorb CO2 from the gas phase, thereby enhancing the KGav values. This is evidence that DEAB is not an unusual solvent but follows trends similar to those exhibited by other solvents as a function of these mentioned parameters. The unusual results are the specific attributes of DEAB compared to MDEA (both tertiary amines). As indicated before, the unexpected results are that DEAB has larger cyclic and absorption capacities than MDEA, while at the same time, its mass transfer coefficient is larger than that of MDEA. However, the increase in solution flow rate leads to higher circulation and regeneration costs, and, thus, might not improve the overall system efficiency. Maximizing liquid flow rates might not lead to optimum operating conditions. Therefore, an optimum flow rate has to be determined. The effect of liquid flow rate on the KGav value from 4.0 to 5.0 m3/(m2·h) is greater than when the liquid flow rate increases from 5.0 to 7.0 m3/ (m2·h), and the reason for this is that the maximum liquid flow rate for this type of packing ID is 5.0 m3/(m2·h), the technical specifications from Sulzer for operating laboratory structured

loading, reflecting the lower absorbent concentration, gave the lower mass-transfer coefficient (KGav) for the system. By increasing the CO2 loading from 0.09 to 0.25 mol/mol, the KGav value was reduced by 20% for MDEA and DEAB and for MEA by 50%, as shown in Figure 5. The experimental results given in Figure 5 show that CO2 loading of the feed solution had an apparent influence on the KGav value. As CO2 loading increased, the KGav value declined, resulting in a lower overall mass transfer coefficient. This illustrates a reduction in the efficiency of the columns. 5.4. Effect of Concentration on Loading Capacity of DEAB. The concentration of DEAB has a great effect on cyclic capacity. As shown in Figure 6, decreased concentration of

Figure 6. Effect of DEAB concentration on loading capacity (GI = 17.85 (kmol/(m2·h)); L = 5 m3/(m2·h); inlet gas T = 24.0 °C; inlet liquid T = 22.8 °C).

DEAB from 2 to 1.1 mol led to increase in rich loading from 0.26 to 0.65 mol/mol, respectively. This increase of rich loading 6475

dx.doi.org/10.1021/ie2008357 | Ind. Eng. Chem. Res. 2012, 51, 6470−6479

Industrial & Engineering Chemistry Research

Article

offered a higher performance of 65%. The highest removal (100%) was achieved by MEA, which, obviously, reached complete removal efficiency. If we reduce the concentration of those amines to 9%, the actual performance of these three amines can be differentiated by considering the column height required to reach 9% removal. According to Figure 7, CO2 absorption into MEA solution took place within only 0.4 m from the column bottom, while DEAB and MDEA required as high as 1.25 m and 2.15 m, respectively, to meet the same removal target. This is because MEA is a primary amine, but MDEA is a tertiary amine and DEAB is a tertiary amino alcohol. On the basis of the mass transfer performance alone, both DEAB and MDEA would not be good candidates as solvents for CO2 capture. However, one of the other important criteria for solvent selection is the energy required for solvent regeneration. In this case, we have shown3,4 that DEAB is by far better than MEA. Therefore, it is apparent that DEAB as a blended solvent with such a primary amine is a good candidate for CO2 capture. However, before we move into the study of the mass transfer characteristics of a blended solvent, it is very important to determine the mass transfer characteristics of the individual components (i.e., aqueous DEAB alone) in order to evaluate both the contributions of DEAB and to determine the blend ratio of the blended amines. In this study we are doing a systematic study of DEAB (a tertiary amine) so that we can determine the contributions of DEAB when mixed with a primary or secondary amine in the overall performance of the blended amine in CO2 capture. Furthermore, CO2 is more soluble in DEAB than MEA. Also, on the basis of our previous study,22 it was shown that DEAB has a very high CO2 absorption capacity compared with conventional amines such as MEA even at practical partial pressures of CO2 used in the CO2 absorption process, as shown in Figure 8.17 Also, the molecular weight of DEAB is high, 145 g/mol. This also implies high viscosity at very high DEAB concentrations. Bearing in mind the viscosity limitation, we decided to work with DEAB concentrations in the range of 1−2.0 M. Also, in terms of volatility, It can be seen that the volatility of DEAB is lower than MEA and a bit higher than MDEA.17 However, one of the other important criteria for

led to an increase of cyclic capacity by 60%, which would reduce the cost of regeneration. Regeneration has a large influence on the capital cost of gas treating plants, as indicated in Astarita.2 The cyclic capacity (CC) can be expressed mathematically as follows: CC = (mole of CO2 abs − mole of CO2 regen )/Vsolution CC = (loading of CO2 abs − loading of CO2 regen )(M )

(17)

where M = concentration of solution (mol/L). 5.5. CO2 Absorption Performance of Single-Amine Solutions. Figure 7 shows the gas phase CO2 concentration

Figure 7. CO2 concentration profile for single-amine solutions 2 M MEA, MDEA, and DEAB at liquid flow rate, 5 (m3/(m2·h)) and lean loading = 0.14 (mol CO2/mol amine).

profiles that were obtained from absorption experiments using three single-amine solutions (i.e., MEA, MDEA, and DEAB). During the experiments, the concentration of amines was 2.0 kmol/m3 and the CO2 loading of the feed solutions was (0.14 mol/mol). Considering the CO2 removal efficiency for each profile, it can be seen that the MDEA solvent gave the lowest CO2 absorption performance at 45%, while DEAB

Figure 8. Equilibrium solubility of CO2 in aqueous solutions of 2 M MEA, 2 M DEA, 2 M MDEA, 2 M AMP, 2 M DEAB, and 2 M PZ by Sema17 (lines are trend lines of the experimental results (Sema17))3−5. 6476

dx.doi.org/10.1021/ie2008357 | Ind. Eng. Chem. Res. 2012, 51, 6470−6479

Industrial & Engineering Chemistry Research

Article

solvent selection is energy required for solvent regeneration. In this case, literature have shown3,4 that DEAB is by far better than MEA. Therefore, it is apparent that DEAB is a good candidate for CO2 capture, most preferably as a blended solvent with a primary amine such as MEA.

P represents partial pressure of CO2 (kPa). By rearranging eq 20 and 21 and by replacing the koL term with Lx, the relation becomes as follows:

6. MASS-TRANSFER CORRELATIONS Mass transfer coefficient correlations are important because they enable prediction of the experimental results. Several correlations have been used to predict mass transfer coefficients in chemical engineering processes. Stringle18 predicted correlations for overall mass transfer coefficients KGav as follows:

(22)

K Ga v ∝ LbGc

⎡ (αeq − α)C ⎤ ⎥ K Ga v ∝ Lx⎢ PCO2 ⎥⎦ ⎢⎣

This relationship can be solved by plotting the term of (KGav/Lx) against the term of [(αeq − α)C]/PCO2. In the development of this correlation, several values of x were tried but the best results were obtained with x = 0.6. The plot is shown in Figure 9. A linear

(18) 2

where L represents liquid mass velocity (lb/(ft h)), b is the coefficient, G represents gas mass velocity(lb/(ft2 h)), and c is the coefficient. When the liquid film controls the system, the value of b is between 0.22 and 0.38, depending on the type of packing. The effect of liquid film on gas rate is very small, and the value of c in eq 18 normally is only 0.06−0.08. However, the value, c, increases to between 0.67 and 0.80 when the gas film controls the system.18,19 Kohl et al.20 predicted the correlation for KGav for MEA as follows: K Ga v = F(L /μ)2/3 [1 + 5.7(αeq − α)C e 0.0067T − 3.4p]

(19)

were F represents the packing factor, L is the liquid flow rate (lb/(h ft2)), μ is liquid viscosity (cps), αeq is CO2 loading of solution in equilibrium with PCO2, α is CO2 loading, C is amine concentration, P represents partial pressure of CO2 over the solution, and T represents the temperature. The KGav correlation in eq 19 is suitable for an MEA system using random packing in the absorption column but is not suitable for other absorption systems, as suggested by deMontigny16 and Aroonwilas and Tontiwachwuthikul21 because of different packing correlation factors that will provide different KGav data. New correlations were developed by deMontigny22 for MEA systems to predict KGav value. 6.1. Development of Correlations for DEAB. In this study, the overall mass transfer coefficient correlation was defined in terms of the individual mass transfer coefficient, as shown in eq 4 below:

1 1 H = + 0 KG kG IkL

Figure 9. Relationship between (KGav/L0.6) and [(αeq − α)C]/PCO2 at solvent concentration = 2.0 (kmol/m3).

regression of the data point in the plot produces an equation of the common form y = mx + b where m is the slope of the line and b is the y axis intercept, so eq 22 can be rewritten as follows: ⎡ (αeq − α)C ⎤ + b⎥ K Ga v = mLx⎢ PCO2 ⎢⎣ ⎦⎥

This correlation for the KGav value was plotted against the actual experimental values in order to check the accuracy of the model. As shown in eq 23, KGav is a function of liquid flow rate (L), CO2 is a function of the loading of solution (α), concentration is a function of the loading of the solution (C), partial pressure is a function of the loading of CO2 in kPa (PCO2), while m, x and b are constant parameters. In this study, x = 0.6 yielded the least deviation of experimental values from predicted values, whereas the values for the coefficients m and b were obtained from the experiment. The mass transfer correlation predicted, the KGav values for structured packings which were relatively in good agreement with the experimental values. The % AAD of this correlation was 14.57%. Because of the nature of its derivation, this model is provided only for the purpose of predicting unknown KGav values based on experimental conditions for DEAB for this packing. 6.2. Relative Cost Issues. The cost of building a CO2 absorption plant is dependent on the chemical solvent and the energy that can be saved from the regenration process. The most promising areas for operating cost savings with the new chemical solvent are from high cyclic capacity and low energy of consumption. It is estimated that the new chemical solvent (DEAB), which has high cyclic capacity and low regeneration energy, can decrease the energy requirement by 50%. Since the energy requirement for regeneration accounts for about 70% of operating costs, and by splitting capital and operating costs for

(4)

The terms on the right-hand side of the equation represent the gas and liquid film resistances to mass transfer. When the liquid film resistance to mass transfer is much larger than that of the gas film, the equation can be simplified as K Ga v ∝ IkLoa v

(20) 23

koL

According to Perry, the term of in eq 20 was shown to vary as a function of liquid flow rate (L) to the power of x. The exponent x lies in a range between 0.3 and 0.7. Furthermore, Astarita2 suggested that the enhancement factor, I, can be expressed as follows: I≈

(αeq − α)C PCO2

(23)

(21)

where αeq represents the CO2 loading of solution in equilibrium with PCO2, α is CO2 loading, C is amine concentration, and 6477

dx.doi.org/10.1021/ie2008357 | Ind. Eng. Chem. Res. 2012, 51, 6470−6479

Industrial & Engineering Chemistry Research

Article

CO2 capture at 50% each, it can be estimated that by using DEAB, the CO2 capture costs can be reduced by about 35%. Also, even though the cost of producing DEAB now is high, this is somewhat compensated by the need to use lower DEAB concentrations compared with higher concentrations of MEA to achieve the same removal efficiency.

7. CONCLUSIONS Comparative mass transfer studies of CO2 absorption in aqueous MEA, MDEA, and DEAB solutions have been conducted using structured packing in an absorption column. The performance was presented in terms of the overall mass transfer coefficient, KGav. Also, the effects of various parameters, such as CO2 loading of the solution, solution concentration on loading capacity, and liquid flow rate on KGav, were studied. (1) The results have shown that the CO2 loading of the solution and liquid flow rate have effects on KGav in that KGav increases as liquid flow rate increases and KGav is improved by decreased CO2 loading in the feed solution. Also, the liquid flow rate and CO2 loading have a significant impact on the CO2 absorption performance of structured packings, and an increase in these parameters generally improves the mass-transfer efficiency. However, there is no change in the KGav value resulting from a change in the inert gas flow rate. (2) The concentration of a new chemical solvent, DEAB, has a great effect on loading capacity; as the concentration of DEAB decreased, the loading capacity increased. Also, the experimental results have shown that the DEAB solvent has higher CO2 removal efficiency than MDEA. (3) The new synthesized amino alcohol (DEAB) provided much higher CO2 absorption capacity and higher cyclic capacity, and this led to reduction of the circulation rate and energy for solvent regeneration. Reduction of the energy requirement and the height of the columns would lead to reduced capital cost and long-term operating costs. The DEAB system provided an excellent overall mass transfer coefficient that is higher than that of the MDEA system. (4) A new correlation was used to predict overall mass transfer coefficient for DEAB solvent using structured packings. In this study, the KGav values are in relatively good agreement with the actual value and the % AAD of this correlation is 14.57% compared with the 20% value reported in the literature.





c = coefficient CA,L = bulk concentration (kmol/m3) DEAB = 4-(diethylamino)-2-butanol F = packing factor G = gas mass velocity (lb/(ft2·h)) GI = gas flow rate (kmol/(m2·h)) HeN2O = solubility of N2O, kPa m3/kmol HECO2‑amine = solubility of CO2 in aqueous amine solution, kPa m3/kmol HEN2O‑amine = solubility of N2O in aqueous amine solution, kPa m3/kmol HeN2O‑DEAB = solubility of N2O in pure DEAB, kPa m3/kmol HeCO2‑H2O = solubility of CO2 in water, kPa m3/kmol HeN2O‑H2O = solubility of N2O in water, kPa m3/kmol I = enhancement factor kG = gas side mass transfer coefficient (kgmol/(m2·s·kPa)) KG = overall mass transfer coefficient (kgmol/(m2·s·kPa)) KGav = volumetric overall mass transfer coefficient (kgmol/ (m2·s·kPa)) L = liquid flow rate (m3/(m2·h)) μ = liquid viscosity (cps) MDEA = methyldiethanolamine MEA = monoethanolamine NA = mass flux P = pressure, kPa PCO2 = partial pressure of CO2, kPa T = temperature, C° YA,G = mole ratio value yA,G,yA,i = mole fraction (kgmol/kgmol) y*A = gas phase in equilibrium Z = height, (m) αeq = CO2 loading of solution in equilibrium with PCO2 (mol CO2/mol amine) α = CO2 loading (mol CO2/mol amine)

REFERENCES

(1) IPCC. IPCC Special Report on Carbon Dioxide Capture and Storage; Cambridge University Press: Cambridge, United Kingdom and New York, 2005. (2) Astarita, G., Savage, D. W., Bisio, A. Gas Treating with Chemical Solvents; Wiley: New York, 1983 (3) Maneeint, K.; Idem, R. O.; Tontiwachwuthikul, P.; Wee, A. G. H. Synthesis, Solubilities, and Cyclic Capacities of Amino Alcohols for CO2 Capture from Flue Gas Streams. 9th International Conference on Greenhouse Gas Technologies (GHGT 9); Washington DC, USA, Nov 16−20, 2009. (4) Tontiwachwuthikul, P.; Wee, A. G. H.; Idem, R. O.; Maneeintr, K.; Fan, G.-J.; Veawab, A.; Henni, A.; Aroonwilas, A.; Chakma. A. Method for capturing carbon dioxide from gas streams. World Intellectual Property Organization Patent (WIPO), No. WO/2008/ 022437, 2008. (5) Maneeintr K.; Idem, R. O.; Tontiwachwuthikul P.; Wee, A. G. H. Solvent Development for CO2 Capture from Industrial Gas Streams. The 2nd International Conference on Ad Vances in Petrochemicals and Polymers (ICAPP2007); Bangkok, Thailand, Jun 25−28, 2007. (6) Maneeintr, K.; Henni, A.; Idem, R. O.; Tontiwachwuthikul, P.; Wee, A. G. H. Physical and Transport Properties of Aqueous Amino Alcohol Solutions for CO2 Capture from Flue Gas Streams. Process Saf. Environ. Prot. 2008, 86, 291−295. (7) Maneeintr, K.; Henni, A.; Idem, R. O.; Tontiwachwuthikul, P.; Wee, A. G. H. Comparative Mass Transfer Performance Studies of CO2 Absorption into Aqueous Solutions of DEAB and MEA. Ind. Eng. Chem. Res. 2010, 49 (6), 2857−2863.

AUTHOR INFORMATION

Corresponding Author

*Tel.: +1 306 585 4470. Fax: +1 306 585 4855. E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS A. Naami would like to acknowledge the scholarship support from the Libyan Higher Educational studies through the cultural section of the Libyan Embassy in Ottawa, Canada. Also the financial support from the International Test Centre for CO2 Capture, University of Regina, Regina, and Natural Sciences and Engineering Research Council of Canada (NSERC) is appreciated.



NOMENCLATURE b = coefficient C = amine concentration (mol/mol amine) 6478

dx.doi.org/10.1021/ie2008357 | Ind. Eng. Chem. Res. 2012, 51, 6470−6479

Industrial & Engineering Chemistry Research

Article

(8) Aroonwilas, A.; Tontiwachwuthikul, P. Mass Transfer Coefficients and Correlation for CO2 Absorption into 2-Amino-2-methyl-1propanol (AMP) Using Structured Packing. Ind. Eng. Chem. Res. 1998, 37, 569. (9) Aroonwilas, A.; Tontiwachwuthikul, P. Behavior of the Mass Transfer Coefficients of Structured Packing in CO2 Absorption with Chemical Reactions. Ind. Eng. Chem. Res. 1999, 38, 2004. (10) deMontigny, D.; Tontiwachwuthikul, P.; Chakma, A. Parametric Studies of Carbon Dioxide Absorption into Highly Concentrated Monoethanolamine Solutions. Can. J. Chem. Eng. 2001, 79, 137. (11) Fernandes, J.; Lisboa, Pedro F.; Simoes, Pedro C.; Mota, Jose P. B.; Esteban, S. Application of CFD in the Study of Supercritical Fluid Extraction with Structured Packing. J. Supercrit. Fluid 2009, 50, 61−69. (12) Geankoplis, C. J. Transport Process and Separation Process Principles, 4th ed.; Prentice Hall: NJ, 2003. (13) Sema, T., Edali, M., Naami, A., Idem, R. Tontiwachwuthikul, P. N2O Solubility and N2O Diffusivity of Aqueous 4-(Diethylamino)-2Butanol Solutions. Ind. Eng. Chem. Res. 2012, 51, 925−930. (14) Versteeg, G. F.; van Swaaij, W. P. M. Solubility and diffusivity of acid gases (CO2, N2O) in aqueous alkanolamine solutions. J. Chem. Eng. Data 1988, 33, 29−34. (15) Horwitz, W. Association of Official Analytical Chemists (AOAC) Methods, 12th ed.; George Banta: Manasha, WI, 1975. (16) deMontigny, D. Comparing Packed Columns and Membrane Absorbers for CO2 Capture. PhD. Dissertation, University of Regina, Regina, Saskatchewan, Canada, 2004. (17) Sema, T.; Naami, A.; Idem, R.; Tontiwachwuthikul, P. Correlations for Equilibrium Solubility of Carbon Dioxide in Aqueous 4-(Diethylamino)-2-butanol Solutions. Ind. Eng. Chem. Res. 2011, 50, 14008−14015. (18) Strigle, R. F. Random Packing and Packed Towers; Gulf Publishing Company: Houston, TX, 1987. (19) Benamar, A.; Aroua, M. K. Modeling of CO2 Solubility and Carbamate Concentration in DEA, MDEA and Their Mixtures Using the Deshmukh−Mather Model. Fluid Phase Equilib. 2005, 231, 150. (20) Kohl, A. L.; Riesenfeld, F. C. Gas Purification, 5th ed.; Gulf Publishing Co: Houston, TX, 1997. (21) Tontiwachwuthikul, P.; Meisen, A.; Lim, J. C. Solubility of CO2 in 2-Amino-2-methyl-1-propanol Solutions. J. Chem. Eng. Data 1991, 36, 130. (22) deMontigny, D. Carbon Dioxide Absorption Studies Using Ultra-highly Concentrated Monoethanolamine Solutions. Master’s thesis, University of Regina, Regina, Saskatchewan, Canada, 1998. (23) Perry, R. H.; Green, D. Perry’s Chemical Engineer’s Handbook, 8th ed.; McGran Hill Book Company: New York, 2008.

6479

dx.doi.org/10.1021/ie2008357 | Ind. Eng. Chem. Res. 2012, 51, 6470−6479