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 appropriate placement of the substituent, especially hydroxyl

Carbon dioxide is a greenhouse gas and substantially function, relative to the position of the amino group, in order to
contributes to global warming and climate change. One option promote CO2 capture performance.
for reducing CO2 emissions is postcombustion capture from Maneeintr et al.,3,4 have shown that the new amino alcohols
power plant flue gases. Many countries have agreed to reduce have much higher CO2 absorption and cyclic capacities
their emissions of greenhouse gases into the atmosphere to help compared to that of conventional amine, MEA. Details of the
prevent dangerous alterations to the climate system. Several solubility and physical and transport properties of one of the
mature technologies are available for CO2 capture, including new solvents, DEAB, can be found in the literature.5−7 It is well-
adsorption, cryogenics, membrane technologies, and absorption.1 known that the overall performance of any solvent depends not
According to Astarita,2 absorption is the most commonly used only on its absorption capacity and energy efficiency, but also,
process when it comes to gas treating. Gas absorption by on the mass transfer characteristics and the kinetics. The
chemical solvents such as aqueous solutions of alkanolamines is present work focuses on mass transfer and compares the
one of the most effective methods for CO2 removal. This performance of the potential new solvent with the conventional
technology has been used in industry for over half a century. solvents that can be used for the capture of CO2 from industrial
Presently, the most commonly used chemical solvents are flue gas. Consequently, a primary amine (monoethanolamine
alkanolamines, and these alkanolamines can be classified into (MEA)) compared with a new chemical solvent, an amino
three chemical categories: primary, secondary, and tertiary alcohol (4-diethylamino-2-butanol (DEAB)), and a tertiary
amines. Alkanolamines commonly used are monoethanolamine amine (methyldiethanolamine (MDEA)) were selected to
(MEA) and methyl diethanolamine, (MDEA). MEA is highly comprehensively study the new solvent’s mass transfer
reactive with CO2 but has the limitation of low CO2 absorption characteristics in a packed-bed absorption column.
capacity (an equilibrium absorption of 0.5 mol CO2/mol of The process of CO2 absorption in a packed tower depends
MEA) and high heat of regeneration; on the other hand, mainly on the contact between the flue gas and the liquid
MDEA has high CO2 absorption capacity, and low heat of solvent. In the last 35 years, tray columns have been replaced in
regeneration, but it is limited by slow kinetics. The main large part by packed columns.8,9 This has been due to the
challenge for CO2 postcombustion capture from power plant successful development of efficient packing that provides a high
flue gases is the reduction of the energy requirement for solvent absorption capacity per unit volume of packing. Inside the
regeneration. The required reduction can only be achieved by absorption column, the packing provides a contact area, which
application of new solvents. Recently, the International Test separates the liquid flow into droplets. This increases the
Centre for Carbon Dioxide Capture (ITC), Saskatchewan,
Canada, has developed new chemical solvents based on amino Received: April 18, 2011
alcohols for CO2 capture to overcome the drawbacks of these Revised: March 20, 2012
conventional amines. The synthesis involves a systematic Accepted: April 16, 2012
modification of the structure of amino alcohols by an Published: April 16, 2012

© 2012 American Chemical Society 6470 dx.doi.org/10.1021/ie2008357 | Ind. Eng. Chem. Res. 2012, 51, 6470−6479
Industrial & Engineering Chemistry Research Article

amount of liquid surface area that is exposed to the gas phase, Then, KG can be expressed as follows:
enhancing absorption rates. Good packings should offer character- ⎛ ⎞
istics such as large surface area to volume ratio, low pressure drop NA
K G = ⎜⎜ ⎟⎟
across the absorber, and good uniform distribution between the gas ⎝ [P(yA − yA*)] ⎠ (6)
and liquid phases through the column. On the basis of these
characteristics, Aroonwilas9 and deMontigny10 have shown that In a gas absorption apparatus such as a packed column, it is
structured packings offer a superior performance when compared more useful to represent rates of absorption in terms of
to random packings. Fernandes et al.11 studied structured packing volumetric overall mass transfer coefficients, represented by the
and found that the high surface area of this packing increases the term KGav, instead of the one based on the interfacial area unit
mass transfer efficiencies. because the gas−liquid interfacial area cannot be measured
The objective of this work was to perform a comparative accurately.
study of the mass transfer performance for CO2 absorption in Therefore, it is more useful to present the mass-transfer
conventional amine, MEA, MDEA, and the new chemical coefficient based on the unit volume of the absorption column
solvent DEAB using a mass transfer experimental laboratory- rather than based on the interfacial area unit as follows:
scale column with DX structured packing (27.5 mm) ID (from ⎛ ⎞
Sulzer Chemtech Canada, Inc.) for the three solvents under NAa v
K Ga v = ⎜⎜ ⎟⎟
similar conditions. The mass transfer performance of the three ⎝ [P(yA − yA*)] ⎠ (7)
solvents is presented in terms of the overall mass transfer
coefficient. The term NAav can be determined from the absorption
experiments in packed columns where the concentration profile
2. DETERMINATION OF THE OVERALL MASS of the absorbed component A in the gas phase can be measured
TRANSFER COEFFICIENT along the column height, and this allowed us to evaluate the
Mass transfer occurs when a component A in a gas phase KGav value.
transfers across a gas−liquid interface into a liquid phase. The Considering an element of the absorption column with
component A is transferred from the gas phase into the liquid height dZ, in Figure 1, the mass balance of component A can be
because of the concentration gradient in the direction of mass given as follows:
transfer within each phase. The mass flux of component A (NA) ⎛ y ⎞
across the gas−liquid interface at steady state can be represented NAa v dZ = G I d⎜⎜ A ⎟⎟
in terms of the gas-side mass-transfer coefficient (kG) and driving ⎝ 1 − yA ⎠ (8)
force (yA,G − yA,i) as follows:
K Ga v P(yA,G − yA*) dZ = G I dYA,G (9)
NA = K GP(yA,G − yA, i ) (1)
where YA,G represents the mole ratio of component A in the
where P represents the total pressure of the system and yA,G and bulk gas and GI represents the molar flow rate of total gas
yA,i represent the mole fraction of component A in gas bulk. The without component A.
gas mass transfer coefficient kG is difficult to measure because of The final equation that determines the overall mass transfer
the change in interfacial area with varying gas flow rates in the coefficient, KGav, can then be defined as follows:
packed columns. Instead, the overall mass transfer coefficient for
the gas phase (KG) is often used,12 and it can be presented in ⎛ ⎞⎛ d Y ⎞
GI
terms of the Henry’s law constant (H) as follows: K Ga v = ⎜⎜ ⎟⎜ A,G ⎟
⎟
⎝ P(yA,G − yA ) ⎠⎝ dZ ⎠ (10)
NA = K G(PyA,G − HCA,L) (2)
This approach has been used successfully by Aroonwilas and
Strigle13 described the relationship between the overall mass Tontiwachwuthikul,9 deMontigny,10 and Maneeintr7 for
transfer coefficients and the individual mass transfer coefficients analyzing packed columns. In this study, the experiments of
as follows: CO2 absorption were carried out in test columns packed with
1 1 H structured packings. The variables presented in eq 10 can be
= + obtained from the absorption experiments. The inert gas flow
KG kG kL (3) rate (GI) is an operating condition of the experiment, which is
The overall mass transfer coefficients for chemical absorption can conducted at atmospheric pressure (P). The CO2 concen-
be expressed as a function of enhancement factor (I): tration in the gas phase (yA,G) can be measured along the length
of the packed column. The yA* is the concentration of solute A
1 1 H in equilibrium with the bulk concentration CA,L. The measured
= + 0
KG kG IkL (4) CO2 concentration can be converted into mole ratio values
(YA,G) and plotted against the height of the column to obtain
The concentration gradients (yA,G − yA,i) take place over the solute mole ratio concentration gradient (dYA,G/dZ), as
extremely small distances, which makes it difficult to measure the shown in Figure 1, and the advantage of this technique is that it
concentration of component A at the gas−liquid interface. Under allows the calculation of the KGav value at any specific mole
this situation, it is more practical to express the mass flux in ratio values along the column. When comparing the KGav value
terms of the overall mass transfer coefficient and the mole for two solvents, it is important to make the comparison with
fraction of component A in the gas phase (yA*) in equilibrium similar experimental conditions. For example, to compare the
with the concentration of A in the bulk liquid as follows: effect of the liquid flow rate on KGav values for two
NA = K GP(yA,G − yA*) experiments, the solution concentration, the mole ratio, and
(5) CO2 loading condition in the packed column should be the
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.

same, or at least in the same range, to ensure that the GI = inert gas flow rate and yA = gas phase CO2 concentration,
calculation of KGav values are being compared fairly. For and mass.error% is the error in the experiment.
example, if two experiments had the same mole ratio value, but In the literature,9,10 it is assumed that the term of y*A is equal
one experiment had a solution CO2 loading of α = 0.15 mol to zero because the compound A is quickly consumed in the
CO2/mol amine and the second a value of α = 0.36 mol CO2/ fast chemical reaction and is difficult to measure, but in this
mol amine, the comparison would not be fair, since the second work we measure the term of yA*.
experiment had a reduced capacity for absorption due to its
high loading and relevant low free-amine concentration. To
3. THERMODYNAMICS FOR CO2 ABSORPTION
determine the loading at a specific height of the column and Knowing the thermodynamic properties of amine solution,
CO2 concentration, we used eq 11: such as the physical solubility of CO2 or the CO2 Henry’s
constant, physical diffusivity of CO2, equilibrium constants for
α = αlean.loading + ((((D1 − D2) × A × time) CO2-amine system, viscosity, and density, is required in
understanding CO2 absorption.
/(M × L)) × mass. error%) (11) 3.1. Solubility and Physical properties. In this study, the
solubility of CO2 in aqueous amine solution was used to
where α is the CO2 loading at any height of the column, measure the term y*A as shown in eq 12.
αlean.loading is the loading of solvent coming in at the top of the PCO2
column, M is the solvent molarity, L is the liquid flow rate, A yA* =
HeCO2 ‐ amine
is cross sectional area of column, and, D = (GI × yA) where (12)

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 was stirred at 0 °C for 4 h. Aqueous 1 M NaOH (10 mL) was
interface, HeCO2−amine is the solubility of CO2 in aqueous amine added slowly at 0 °C. After gas evolution had subsided, excess
solution, kPa m3/kmol, and PCO2 is the partial pressure of CO2 solid NaCl was added to the mixture. The mixture was
in kPa. thoroughly extracted with CH2Cl2 (2 × 100 mL), the organic
The term of HeCO2‑amine can be determined from eq 13:13 extracts were dried (Na2SO4), filtered, and the filtrate was
distilled at atmospheric pressure to remove CH2Cl2. The
⎛ HeCO ‐ H O ⎞ resulting oil was then fractionated using a viqreux column to af-
HeCO2 ‐ amine = He N2O ‐ amine⎜⎜ 2 2
⎟⎟ ford 4-(diethylamino)-2-butanol (48.3 g, 49%): bp 96−111 °C
⎝ He N2O ‐ H2O ⎠ (13) (22 mmHg). IR _max: 3075−3600 cm−1; 1H NMR, ı (200
The solubility of pure DEAB solvent can be determined from MHz, CDCl3): 0.73 (t, 6H, J = 7 Hz, MeCH2), 0.84 (d, 3H, J =
the solubilities of CO2 and N2O for water based on the work of 6 Hz, MeCHO), 1.10−1.30 (m, 2H, CH2), 1.95−2.15 (m, 2H,
Versteeg and van Swaaij.14 The correlations are described in eqs NCH2), 2.20−2.45 (m, 4H, N(CH2Me)2), 3.52−3.65 (m, 1H,
14 and 15 below: MeCHO); OH was not observed. Fractional distillation was
used to purify the chemical, the identity of which was later
⎛ 2284 ⎞ verified by using the nuclear magnetic resonance (NMR)
He N2O ‐ H2O = 8.55 × 106 exp⎜ − ⎟
⎝ T ⎠ (14) technique. Approximately, the purity of DEAB was 94.5% and
the impurities were 4-methoxy-2-butanol, 2.3%; 2-butanol,
⎛ 2044 ⎞ 1.9%; and diethyl amine1, 0.5%. At this point, DEAB is
HeCO2 ‐ H2O = 2.82 × 106 exp⎜ − ⎟
⎝ T ⎠ (15) synthesized in-house on a small scale. MEA and MDEA were
obtained from Fisher Scientific with a purity of 99+%. Nitrogen
The T in eqs 14 and 15 can be used from the experimental data and CO2 were supplied by Praxair Inc., with purities of 99.9%.
of the packed column by determining the solubility of CO2 in
All materials in this study were used as received without further
aqueous amine solution; then, the term of yA* can be calculated
purification.
in eq 12.
The absorption performance of the amines was evaluated by
PCO2 conducting experiments in a bench-scale absorption unit, of
yA* = which a simplified flow diagram is given in Figure 1. The unit
HeCO2 ‐ amine (12) consisted of a glass absorption column (27.5 × 10−2 m in
From eq 12, we found from our calculations that yA* equals diameter and 2.15 m in height) packed with 37 elements of stain-
2.219 × 10−8. This is very close to 0. Therefore, practically, the less steel structured packing (Sulzer DX) and (900 m2/m3).
CO2−DEAB reaction can be taken to be instantaneous. The column was designed for a counter-current mode of
The physical properties based on the molarity of amine operation in which a liquid solution was introduced to the
solution were derived from the work of Maneeintr et al.6 in column at the top while a gas mixture entered the column
which the physical and transport properties of aqueous amines below the packing section. A series of gas sampling points
and amino alcohol solutions for CO2 capture were measured. were also installed at regular intervals along the sides of the
They measured densities, viscosities, and refractive indices of column to allow measurements of the gas-phase CO 2
various concentrations of 4-diethylamino- 2-butanol + water concentration during experiments, and a series of thermo-
mixtures. couple points was also installed along the sides of the column
to allow measurements of the temperature of liquid during the
4. EXPERIMENTAL SECTION experiments.
4-Diethylamino-2-butanol (DEAB) with structural formula as The absorption unit was also composed of (i) two 35-L
given in Scheme 1 was synthesized in the CO2 laboratory at the solution tanks, (ii) three calibrated mass flow meters, (iii) three
needle valves, (iv) a variable-speed gear pump, and (v) an
Scheme 1. Structural Formula of 4-Diethylamino-2-butanol infrared (IR) gas analyzer. The solution tanks served as
(DEAB) 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
University of Regina, Regina, Canada. Details of the synthesis operated during the experiments to withdraw samples and measure
procedure are given elsewhere.4 A brief description of the the CO2 concentration of the gas mixture inside the column. This
procedure is however given as follows: to diethyl amine (49.7 g, analyzer can measure CO2 concentrations up to 20%.
0.68 mol) was added methyl vinyl ketone (51.4 g, 0.73 mol) Prior to the experiments, an aqueous solution of the amine
dropwise with stirring over a period of 1 h. The reaction was prepared in the feed tank by diluting the concentrated
temperature was maintained below 0 °C using an ice-salt bath. amine with deionized water to a given concentration. The feed
After the addition was complete, the cooling bath was removed, tank is connected to nitrogen balloons to keep the solvents
and the mixture was stirred at room temperature. Then the under a blanket of nitrogen. The actual amine concentration
reaction mixture was dissolved in MeOH (20 mL). The was determined by titration with a standard 1.0 kmol/m3
solution was cooled to 0 °C and NaBH4 (30.7 g, 0.81 mol) was hydrochloric acid (HCl) solution using methyl orange
added, portionwise, to the stirred reaction mixture. The mixture indicator. Horwitz (1975)15 presented the official analytical
6473 dx.doi.org/10.1021/ie2008357 | Ind. Eng. Chem. Res. 2012, 51, 6470−6479
Industrial & Engineering Chemistry Research Article

chemistry technique to determined the solution CO2 loading. were conducted to test the working order of the packed column.
By adding excess 1.0 N HCI acid to the sample, all of the CO2 Verification of the apparatus was performed by comparing results
evolved into a gas buret for measurement. with the work done by deMontigny,16 which detailed the absorption
Each experimental run began by introducing N2 and CO2 of CO2 from a simulated flue gas stream (air + CO2) using a MEA
gases from cylinders through mass flow meters at desired flow solution. The results indicated that increasing liquid flow rate
rates to produce a CO2−N2 gas mixture, which was fed to the improves the KGav value, as shown in Figure 2. The results served as
bottom of the column. The concentration of CO2 in the feed the basis for comparison with the new absorption experiment.
gas was checked by the IR gas analyzer and adjusted until the The results of the comparison between the two studies are
desired value was obtained. The prepared solvent was then shown in Figure 2. The new data follows the same trend and
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 counter- Figure 2. Verifying the packed column performance for 2 M MEA
current mode at preset operating conditions. At steady state (both studied at Y = 0.09 (mol CO2/mol air).
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 falls in the same range as the data published by deMontigny.16
overall mass transfer coefficient can be calculated using eq 10. There are a few minor differences between the experiments
The total number of measured data points from 27 experimental from the two studies. The new experiment had lower gas and
runs was 540, which includes 270 measured points of CO2 gas liquid flow rate than deMontigny’s16 work. However, since the
concentration and 270 measured points of temperature. results are in the same range and follow a similar trend, the
From these data, the concentration and temperature change packed column was determined to be in good working order.
along the columns were measured. The effects of liquid flow 5.2. Effect of Liquid Flow Rate in the Packed Column.
rate, solution concentration, and temperature on KGav were Liquid flow rate has a great impact on the CO2 absorption
then examined. The experimental equipment used for this efficiency of the DX structured packing. As shown in Figure 3,
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 Figure 3. CO2 concentration profiles at different liquid flow rates (m3/
⎝ absorbed CO2 ⎠ (16) (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 =
5.1. Verification of the Packed Column. Maneeintr et al.8 23.2 °C; inlet liquid T = 23.6 °C).
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 increasing liquid flow rate caused a reduction in CO 2
the column. To verify the experimental equipment, a few trial runs 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 packing columns of less than 50 mm diameter, so the maximum
place, so an increase in liquid flow rate led to a greater degree of liquid flow rate should be 5 m3/(m2·h).
wetted packing surface and increased the efficiency of the mass- 5.3. Effect of Loading. The value of KGav decreases with
transfer. CO2 loading, which is because as the CO2 loading increases, the
The effect of the liquid flow rate on KGav values are shown in amount of active amines decreases, causing the KGav to
Figure 4 for MDEA, DEAB, and MEA. Again, it is evident that 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 Figure 5. Effect of lean loading on KGav in the packed column
(amine = 2.0 (kmol/m3); GI = 17.85 (kmol/(m2·h)); YMEA = 0.09 (amine = 2.0 (kmol/m3); GI = 17.85 (kmol/(m2·h)); Y = 0.13 (mol
(mol CO2/mol air); YMDEA, DEAB = 0.14 (mol CO2/mol air); α = 0.14 CO2/mol air); inlet gas T = 24.6 °C; inlet liquid T = 23.2 °C) (vertical
(mol CO2/mol amine); inlet gas T = 24.6 °C; inlet liquid T = 23.2 °C) axes: RHS refers to MEA; LHS refers to MDEA and DEAB).
(vertical axes: RHS refers to MEA; LHS refers to MDEA and DEAB).

the KGav values for MEA are higher than the values for MDEA loading, reflecting the lower absorbent concentration, gave the
and DEAB under corresponding operating conditions, and this lower mass-transfer coefficient (KGav) for the system. By
is because MEA is a primary amine whereas DEAB is a tertiary increasing the CO2 loading from 0.09 to 0.25 mol/mol, the
amino alcohol and MDEA a tertiary amine. The new chemical KGav value was reduced by 20% for MDEA and DEAB and for
amino alcohol DEAB provides a greater KGav value than MDEA MEA by 50%, as shown in Figure 5.
by 30% and this led to increased efficiency of the column. Also, The experimental results given in Figure 5 show that CO2
the results show that the liquid flow rate has an influence on the loading of the feed solution had an apparent influence on the
value of KGav (i.e., an increase in liquid load generally yields a KGav value. As CO2 loading increased, the KGav value declined,
greater KGav value). The possible reason for this behavior is resulting in a lower overall mass transfer coefficient. This
that a higher liquid load leads to the following: (i) a greater illustrates a reduction in the efficiency of the columns.
liquid side mass transfer coefficient (kL), which is directly 5.4. Effect of Concentration on Loading Capacity of
proportional to the overall KGav in the case of liquid-phase DEAB. The concentration of DEAB has a great effect on cyclic
controlled mass transfer; (ii) a greater effective area, which is capacity. As shown in Figure 6, decreased concentration of
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
Figure 6. Effect of DEAB concentration on loading capacity (GI =
optimum flow rate has to be determined. The effect of liquid 17.85 (kmol/(m2·h)); L = 5 m3/(m2·h); inlet gas T = 24.0 °C; inlet
flow rate on the KGav value from 4.0 to 5.0 m3/(m2·h) is greater liquid T = 22.8 °C).
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 DEAB from 2 to 1.1 mol led to increase in rich loading from
specifications from Sulzer for operating laboratory structured 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

led to an increase of cyclic capacity by 60%, which would offered a higher performance of 65%. The highest removal
reduce the cost of regeneration. Regeneration has a large (100%) was achieved by MEA, which, obviously, reached
influence on the capital cost of gas treating plants, as indicated complete removal efficiency. If we reduce the concentration of
in Astarita.2 those amines to 9%, the actual performance of these three
The cyclic capacity (CC) can be expressed mathematically as amines can be differentiated by considering the column height
follows: required to reach 9% removal. According to Figure 7, CO2
CC = (mole of CO2 abs − mole of CO2 regen )/Vsolution absorption into MEA solution took place within only 0.4 m from
the column bottom, while DEAB and MDEA required as high
CC = (loading of CO2 abs − loading of CO2 regen )(M ) (17) 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
where M = concentration of solution (mol/L). tertiary amine and DEAB is a tertiary amino alcohol. On the
5.5. CO2 Absorption Performance of Single-Amine basis of the mass transfer performance alone, both DEAB and
Solutions. Figure 7 shows the gas phase CO2 concentration 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 regenera-
tion. 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,
Figure 7. CO2 concentration profile for single-amine solutions 2 M on the basis of our previous study,22 it was shown that DEAB
MEA, MDEA, and DEAB at liquid flow rate, 5 (m3/(m2·h)) and lean has a very high CO2 absorption capacity compared with
loading = 0.14 (mol CO2/mol amine). conventional amines such as MEA even at practical partial
pressures of CO2 used in the CO2 absorption process, as shown
profiles that were obtained from absorption experiments using in Figure 8.17 Also, the molecular weight of DEAB is high, 145
three single-amine solutions (i.e., MEA, MDEA, and DEAB). g/mol. This also implies high viscosity at very high DEAB
During the experiments, the concentration of amines was concentrations. Bearing in mind the viscosity limitation, we
2.0 kmol/m3 and the CO2 loading of the feed solutions was decided to work with DEAB concentrations in the range of
(0.14 mol/mol). Considering the CO2 removal efficiency for 1−2.0 M. Also, in terms of volatility, It can be seen that the
each profile, it can be seen that the MDEA solvent gave the volatility of DEAB is lower than MEA and a bit higher than
lowest CO2 absorption performance at 45%, while DEAB MDEA.17 However, one of the other important criteria for

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 P represents partial pressure of CO2 (kPa). By rearranging
this case, literature have shown3,4 that DEAB is by far better eq 20 and 21 and by replacing the koL term with Lx, the relation
than MEA. Therefore, it is apparent that DEAB is a good becomes as follows:
candidate for CO2 capture, most preferably as a blended solvent
with a primary amine such as MEA. ⎡ (αeq − α)C ⎤
K Ga v ∝ Lx⎢ ⎥
⎢⎣ PCO2 ⎥⎦ (22)
6. MASS-TRANSFER CORRELATIONS
Mass transfer coefficient correlations are important because This relationship can be solved by plotting the term of (KGav/Lx)
they enable prediction of the experimental results. Several against the term of [(αeq − α)C]/PCO2. In the development of this
correlations have been used to predict mass transfer coefficients
correlation, several values of x were tried but the best results were
in chemical engineering processes. Stringle18 predicted
obtained with x = 0.6. The plot is shown in Figure 9. A linear
correlations for overall mass transfer coefficients KGav as
follows:
K Ga v ∝ LbGc (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
Figure 9. Relationship between (KGav/L0.6) and [(αeq − α)C]/PCO2 at
(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 solvent concentration = 2.0 (kmol/m3).
concentration, P represents partial pressure of CO2 over the
solution, and T represents the temperature. regression of the data point in the plot produces an equation of the
The KGav correlation in eq 19 is suitable for an MEA system common form y = mx + b where m is the slope of the line and b is
using random packing in the absorption column but is not the y axis intercept, so eq 22 can be rewritten as follows:
suitable for other absorption systems, as suggested by ⎡ (αeq − α)C ⎤
deMontigny16 and Aroonwilas and Tontiwachwuthikul21 K Ga v = mLx⎢ + b⎥
because of different packing correlation factors that will provide ⎢⎣ PCO2 ⎦⎥ (23)
different KGav data. New correlations were developed by
deMontigny22 for MEA systems to predict KGav value. This correlation for the KGav value was plotted against the actual
6.1. Development of Correlations for DEAB. In this experimental values in order to check the accuracy of the model. As
study, the overall mass transfer coefficient correlation was shown in eq 23, KGav is a function of liquid flow rate (L), CO2 is a
defined in terms of the individual mass transfer coefficient, as function of the loading of solution (α), concentration is a function
shown in eq 4 below: 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
1 1 H
= + 0 parameters. In this study, x = 0.6 yielded the least deviation of
KG kG IkL (4) experimental values from predicted values, whereas the values for
The terms on the right-hand side of the equation represent the the coefficients m and b were obtained from the experiment. The
gas and liquid film resistances to mass transfer. When the liquid mass transfer correlation predicted, the KGav values for structured
film resistance to mass transfer is much larger than that of the packings which were relatively in good agreement with the
gas film, the equation can be simplified as experimental values. The % AAD of this correlation was 14.57%.
Because of the nature of its derivation, this model is provided only
K Ga v ∝ IkLoa v (20) for the purpose of predicting unknown KGav values based on
23 experimental conditions for DEAB for this packing.
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 6.2. Relative Cost Issues. The cost of building a CO2
exponent x lies in a range between 0.3 and 0.7. Furthermore, absorption plant is dependent on the chemical solvent and the
Astarita2 suggested that the enhancement factor, I, can be energy that can be saved from the regenration process. The
expressed as follows: most promising areas for operating cost savings with the new
chemical solvent are from high cyclic capacity and low energy
(αeq − α)C of consumption. It is estimated that the new chemical solvent
I≈
PCO2 (21)
(DEAB), which has high cyclic capacity and low regeneration
energy, can decrease the energy requirement by 50%. Since the
where αeq represents the CO2 loading of solution in equilib- energy requirement for regeneration accounts for about 70% of
rium with PCO2, α is CO2 loading, C is amine concentration, and operating costs, and by splitting capital and operating costs for
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 c = coefficient

DEAB, the CO2 capture costs can be reduced by about 35%. CA,L = bulk concentration (kmol/m3)
Also, even though the cost of producing DEAB now is high, this DEAB = 4-(diethylamino)-2-butanol
is somewhat compensated by the need to use lower DEAB F = packing factor
concentrations compared with higher concentrations of MEA G = gas mass velocity (lb/(ft2·h))
to achieve the same removal efficiency. GI = gas flow rate (kmol/(m2·h))
HeN2O = solubility of N2O, kPa m3/kmol
7. CONCLUSIONS HECO2‑amine = solubility of CO2 in aqueous amine solution,
Comparative mass transfer studies of CO2 absorption in aqueous kPa m3/kmol
MEA, MDEA, and DEAB solutions have been conducted using HEN2O‑amine = solubility of N2O in aqueous amine solution,
structured packing in an absorption column. The performance was
presented in terms of the overall mass transfer coefficient, KGav. kPa m3/kmol
Also, the effects of various parameters, such as CO2 loading of the HeN2O‑DEAB = solubility of N2O in pure DEAB, kPa m3/kmol
solution, solution concentration on loading capacity, and liquid HeCO2‑H2O = solubility of CO2 in water, kPa m3/kmol
flow rate on KGav, were studied. HeN2O‑H2O = solubility of N2O in water, kPa m3/kmol
(1) The results have shown that the CO2 loading of the
solution and liquid flow rate have effects on KGav in that KGav I = enhancement factor
increases as liquid flow rate increases and KGav is improved by kG = gas side mass transfer coefficient (kgmol/(m2·s·kPa))
decreased CO2 loading in the feed solution. Also, the liquid KG = overall mass transfer coefficient (kgmol/(m2·s·kPa))
flow rate and CO2 loading have a significant impact on the CO2 KGav = volumetric overall mass transfer coefficient (kgmol/
absorption performance of structured packings, and an increase (m2·s·kPa))
in these parameters generally improves the mass-transfer L = liquid flow rate (m3/(m2·h))
efficiency. However, there is no change in the KGav value μ = liquid viscosity (cps)
resulting from a change in the inert gas flow rate. MDEA = methyldiethanolamine
(2) The concentration of a new chemical solvent, DEAB, has MEA = monoethanolamine
a great effect on loading capacity; as the concentration of DEAB NA = mass flux
decreased, the loading capacity increased. Also, the exper- P = pressure, kPa
imental results have shown that the DEAB solvent has higher PCO2 = partial pressure of CO2, kPa
CO2 removal efficiency than MDEA. T = temperature, C°
(3) The new synthesized amino alcohol (DEAB) provided YA,G = mole ratio value
much higher CO2 absorption capacity and higher cyclic yA,G,yA,i = mole fraction (kgmol/kgmol)
capacity, and this led to reduction of the circulation rate and y*A = gas phase in equilibrium
energy for solvent regeneration. Reduction of the energy Z = height, (m)
requirement and the height of the columns would lead to αeq = CO2 loading of solution in equilibrium with PCO2 (mol
reduced capital cost and long-term operating costs. The DEAB CO2/mol amine)
system provided an excellent overall mass transfer coefficient α = CO2 loading (mol CO2/mol amine)

■
that is higher than that of the MDEA system.
(4) A new correlation was used to predict overall mass REFERENCES
transfer coefficient for DEAB solvent using structured packings. (1) IPCC. IPCC Special Report on Carbon Dioxide Capture and
In this study, the KGav values are in relatively good agreement Storage; Cambridge University Press: Cambridge, United Kingdom
with the actual value and the % AAD of this correlation is and New York, 2005.
14.57% compared with the 20% value reported in the literature.

■
(2) Astarita, G., Savage, D. W., Bisio, A. Gas Treating with Chemical
Solvents; Wiley: New York, 1983
AUTHOR INFORMATION (3) Maneeint, K.; Idem, R. O.; Tontiwachwuthikul, P.; Wee, A. G. H.
Corresponding Author Synthesis, Solubilities, and Cyclic Capacities of Amino Alcohols for
CO2 Capture from Flue Gas Streams. 9th International Conference on
*Tel.: +1 306 585 4470. Fax: +1 306 585 4855. E-mail: Greenhouse Gas Technologies (GHGT 9); Washington DC, USA,
[email protected]. Nov 16−20, 2009.
Notes (4) Tontiwachwuthikul, P.; Wee, A. G. H.; Idem, R. O.; Maneeintr,
The authors declare no competing financial interest. K.; Fan, G.-J.; Veawab, A.; Henni, A.; Aroonwilas, A.; Chakma. A.

■
Method for capturing carbon dioxide from gas streams. World
ACKNOWLEDGMENTS Intellectual Property Organization Patent (WIPO), No. WO/2008/
022437, 2008.
A. Naami would like to acknowledge the scholarship support (5) Maneeintr K.; Idem, R. O.; Tontiwachwuthikul P.; Wee, A. G. H.
from the Libyan Higher Educational studies through the Solvent Development for CO2 Capture from Industrial Gas Streams.
cultural section of the Libyan Embassy in Ottawa, Canada. Also The 2nd International Conference on Ad Vances in Petrochemicals
the financial support from the International Test Centre for and Polymers (ICAPP2007); Bangkok, Thailand, Jun 25−28, 2007.
CO2 Capture, University of Regina, Regina, and Natural (6) Maneeintr, K.; Henni, A.; Idem, R. O.; Tontiwachwuthikul, P.;
Sciences and Engineering Research Council of Canada Wee, A. G. H. Physical and Transport Properties of Aqueous Amino
(NSERC) is appreciated. 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.;
NOMENCLATURE
Wee, A. G. H. Comparative Mass Transfer Performance Studies of
b = coefficient CO2 Absorption into Aqueous Solutions of DEAB and MEA. Ind. Eng.
C = amine concentration (mol/mol amine) Chem. Res. 2010, 49 (6), 2857−2863.

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 Coef-

ficients and Correlation for CO2 Absorption into 2-Amino-2-methyl-1-
propanol (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)-2-
Butanol 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