kinetics of absorption of carbon dioxide in aqueous amine ... · pdf filekinetics of...

Kinetics of absorption of carbon dioxide in aqueous amine and

carbonate solutions with Carbonic Anhydrase

Abstract # 388

Nathalie J.M.C. Penders-van Elk, Espen S. Hamborg, Patrick J.G. Huttenhuis, Sylvie Fradette, Jonathan A. Carley

and Geert F. Versteeg

CONFERENCE PROCEEDINGS

Eleventh Annual Carbon Capture, Utilization & Sequestration Conference – April 30-May 3, 2012

Abstract In the present work the absorption of carbon dioxide in aqueous N-methyldiethanolamine (MDEA) and aqueous sodium carbonate with and without carbonic anhydrase (CA) was studied in a stirred cell contactor in the temperature range 298 – 333 K. The CA was present as free enzyme and is compared to the opportunity to immobilize CA on particles and on fixed packing. Based on the results with MDEA and sodium carbonate, the observed kinetics as a function of the free enzyme concentration are described. These results were incorporated into the Procede Process Simulator (Arendsen et al., 2012) to determine the impacts of the kinetic benefit of CA on commercial absorber sizing for carbon dioxide capture from flue gases. Based on simulations performed, CA in the absorption solution can provide substantial benefits for reducing absorber sizing with these normally kinetically limited, but energy efficient solvents. It was also shown that CA immobilized to fixed packing material is not a viable option for using CA in a carbon dioxide capture process.

1 Introduction Carbon dioxide from the flue gases of power plants and other industrial sources is commonly seen as one of the major contributors to climate change (Solomon et al., 2007). The conventional method to capture this CO2 is to apply an absorber-desorber process, in which the acid gas is separated using fast aqueous amine solutions such as monoethanolamine (MEA) or N-methyldiethanolamine (MDEA) promoted by piperazine (Kohl and Nielsen, 1997).

Although amine solutions are relatively well known, their potential application in flue gas processes encounters several drawbacks, most notably the unfavorable desorption energy requirements and the emission of various degradation products (Fostås et al., 2011; Jackson and Attalla, 2011).

Alternatively, solvents such as unpromoted MDEA or alkali-carbonate solutions seem very attractive as the solutions are more stable and less expensive. Moreover, compared to primary or secondary amines, they have relatively low heats of reaction with carbon dioxide, leading to the potential for lower regeneration energy costs in the desorber. However, application of these solvents in post-combustion capture of carbon dioxide is compromised by their low reactivity with CO2 at typical flue gas conditions, which would result in unrealistically large absorber columns.

It is, however, possible to enhance the absorption rate to such an extent that these alternative solvents would be technically and economically viable with the introduction of the enzyme carbonic anhydrase (CA) applied as a catalyst. CA is a powerful catalyst that accelerates the transformation of carbon dioxide to the bicarbonate ion. CA is found in the blood of humans and other mammals and facilitates the efficient transfer of CO2 during respiration. Successful genetic modification of this enzyme makes it possible to use it in combination with aqueous tertiary amines and alkali-carbonate solutions within an industrial environment for flue gas treatment.

In this work, the absorption of carbon dioxide in aqueous MDEA solutions and aqueous sodium (bi)carbonate solutions has been studied in a stirred cell contactor over a wide range of temperatures and CA concentrations. Furthermore, the use of immobilized CA on particles and fixed packing are evaluated.

2 Theoretical background 2.1 Kinetics In aqueous (sodium) carbonate systems, carbon dioxide can react with:

1. Hydroxide (Pinsent et.al., 1956; Pohorecki and Moniuk, 1988)


−− →+ 3k

2 HCOOHCO OH

with the following overall forward reaction rate:

2CO'OH2COOHOH2CO CkCCkR ⋅=⋅⋅= (1)

2. Water (Pinsent et.al., 1956; Kern, 1960)

+− + →+ OHHCOOH2CO 33k

22O2H

with the following overall forward reaction rate:

2CO'

O2H2COx

O2HO2H2CO CkCCkR ⋅=⋅⋅= (2)

In the case of a tertiary alkanolamine system, CO2 can also react with:

3. Tertiary alkanolamine (Versteeg and van Swaaij, 1988a; Little et al., 1990; Benamor and Aroua, 2007)

+− + →++ NHRHCOOHNRCO 33k

232Am

with following overall forward reaction rate:

2CO'Am2COAmAm2CO CkCCkR ⋅=⋅⋅= (3)

Since these reactions occur simultaneously and in parallel, the overall forward reaction rate constant of this system becomes:

)k(kkk 'Am

'OH

'O2HOV ++= (4)

In the presence of the enzyme carbonic anhydrase the reaction mechanism is extended with the following "wheel of reactions" (Lindskog and Silverman, 2000; Larachi, 2010):

The base used during enzyme regeneration can either be the carbonate ion, hydroxide ion or even the reaction 1 or 2 formed bicarbonate ion (Larachi, 2010). In that case, carbon dioxide and water are released as protonated base.

In the presence of the enzyme carbonic anhydrase, reaction 2 is the most important reaction during absorption of carbon dioxide into aqueous MDEA solutions. The forward reaction rate constant for the enzyme catalyzed reaction is calculated from the overall forward reaction rate (Penders-van Elk et al., 2012):


O2H

enzymewithout,OVenzymewith,OV*O2H C

kkk

−= (5)

Equation (5) is only valid if the experiments are in the pseudo first order regime for all reactions involved.

All above mentioned reactions are equilibrium reactions. When the absorption solution gets loaded with CO2, the reverse reactions also become important; even stripping of CO2 can occur. In that case the interpretation of the experimental results is not straight forward. Therefore, the Procede Process Simulator (PPS) is used for the interpretation of the experiments with a CO2-loading larger than 0.5 mol CO2/mol Na. PPS is a flow sheeting tool, specially designed for steady state simulation of acid gas treating processes, developed by Procede (van Elk et al., 2009; Procede, 2011; Arendsen et al., 2012).

2.2 Mass transfer and chemical reaction

2.2.1 Homogeneous system When flue gas is contacted with a chemical solvent, like an aqueous solution of sodium carbonate or alkanolamines, the CO2 will be partially transported from the vapour to the liquid phase, where it is converted to several ionic species. A commonly used, fundamental mass transfer model to describe the absorption process quantitatively is the film model. A schematic representation of the absorption process according to the stagnant film model is presented in Figure 1.

Figure 1: Mass transfer from gas to liquid phase in a homogeneous system

In the stagnant film model the fluid (in this case both gas and liquid phase) can be divided in two different zones: a stagnant film of thickness δ near the interface and a well-mixed bulk behind it in which no concentration gradients occur. In Figure 1 the transport of CO2 from the bulk gas phase to the liquid bulk is shown. The mass transfer is limited by resistances in the gas film and liquid film respectively and these resistances are determined by the diffusion coefficients and the film thickness in both phases. It is assumed that at the gas-liquid interface (thermodynamic) equilibrium exists between the gas phase and the liquid phase. Because a chemical reaction will take place in the liquid film and/or bulk, the mass transfer may be enhanced by this chemical reaction.

The CO2 absorption rate for a homogeneous system as presented in Figure 1 can be described with the following flux equation (Westerterp et al., 1984):

0

GAS LIQUID

kG=D/δG

kL=D/δL

m=CL,i/CG,i

mass transfer

mass transfer & kinetics

solubility

driving force

δLG δG

CL

CL,i

CG

CG,i


2COLG

L,2COG,2CO

GL2CO

Emk1

k1

mC

CVaJ

+

−= (6)

In the case of a pure gas absorbing in a fresh liquid the above equation reduces to:

RTPEmkJ 2CO

2COL2CO = (7)

When the reaction occurs in the so-called pseudo first order regime and 2 < Ha < Einf, the enhancement factor, ECO2, equals the Hatta number:

L

2CO12CO k

DkHaE == (8)

where k1 = kOH’ in the case of Na2CO3 and kAm’ in the case of MDEA. Einf is the infinite enhancement factor, which for irreversible reactions is defined as (van Swaaij and Versteeg, 1992):

2COx

x

2CO

xinf mP

RTCD

D1Eν

+= (9)

where x = OH- in the case of Na2CO3 and MDEA in the case of MDEA as the reactive component in the absorption solution.

2.2.2 Heterogeneous system Absorption of CO2 in aqueous solutions of sodium carbonate or alkanolamines without enzyme or with free enzyme are examples of a homogeneous system. When the enzyme is immobilized on a packing or on solid particles the system becomes heterogeneous. In Figure 2 a schematic representation of the mass transfer process in a gas – liquid – solid system is presented (i.e. Enzymes immobilized on a packing)

Figure 2: Mass transfer in a gas – liquid – solid system (heterogeneous system consisting of particles or packing material)

For this process four different steps can be identified:

• Mass transfer from gas bulk to gas- liquid interface;

CG

CL

δLG δG δLS 0

GAS SOLID LIQUID

CL,i

CG,i

CS,i


• Mass transfer from gas – liquid interface to liquid bulk; • Mass transfer from liquid bulk to catalyst interface; • Reaction at the catalyst interface.

The CO2 absorption rate can be described with the following flux equation (Beenackers and van Swaaij, 1993):

LS"rLSLS2COGLLGLG

G,2CO2CO

amk1

amk1

Eamk1

ak1

CaJ

+++=⋅ (10)

In this equation it is assumed that the catalytic material is present in the bulk of the liquid phase as is the case with CA on (very) large particles or fixed packing.

However, when the catalytic material is present on small particles, the catalyst will also be present in the liquid film (see Figure 3). In that case mass transfer and reaction will occur in a kind of parallel mode similar to a homogeneous system, resulting in an enhanced CO2 absorption rate. This system is described in detail by Beenackers and van Swaaij (1993).

Figure 3: Mass transfer in a gas - liquid - solid system consisting of small particles

When the catalytic material is porous very large internal areas can be present. Before the reactant (CO2) can reach the interior of this catalyst an additional diffusion process has to take place. This diffusion is significantly slower than the diffusion outside the catalytic material, due to the small pore sizes (porosity) and curves of the pores (tortuosity).

3 Materials & Methods 3.1 Materials The aqueous MDEA solutions were prepared with N-methyldiethanolamine (purity > 99%) from Sigma-Aldrich. The aqueous sodium carbonate solutions were prepared with anhydrous sodium carbonate (purity > 99%) from Merck. In the preparation of the loaded sodium carbonate solutions part of the anhydrous sodium carbonate was replaced by sodium bicarbonate (purity > 99%) from Sigma-Aldrich. All chemicals were used as supplied. For tests in MDEA, the enzyme used was a thermostable variant of human carbonic anhydrase II provided by CO2 Solutions. For tests in sodium carbonate, the enzyme was

CG

CL

δG 0

GAS LIQUID

CL,i

CG,i

δL


obtained from Codexis, Inc. All solutions were prepared with demineralized water. The carbon dioxide (> 99.9 %) and nitrous oxide (> 99 %) were obtained from Air Liquide.

3.1.1 Experimental set-up All experiments with free enzyme and enzyme immobilized on particles were performed in a stirred cell contactor, operated in batch mode with respect to the liquid and gas phase and with a flat horizontal gas-liquid interface. The reactor was connected to a gas supply vessel filled with carbon dioxide or nitrous oxide from gas cylinders. Both the reactors and gas supply vessel were equipped with digital pressure transducers and PT-100 thermocouples. The measured signals were recorded with a computer. The pressure transducers connected to the stirred cell reactors are Endress+Hauser Cerabar S pressure transducers (range 0 - 4 bara) and the gas supply vessel is equipped with an Endress+Hauser Cerabar M pressure transducer (range 0 - 10 bara). A schematic drawing of the set-up is presented in Figure 4.

Figure 4: The experimental set-up as used for the experiments with free enzyme and enzyme immobilized on particles.

3.2 Methods The measurements of the CO2 absorption into aqueous MDEA or sodium carbonate solution in the presence or absence of the enzyme were performed in batch mode with respect to both liquid and gas phase. In a typical experiment approximately 500 ml of solution with the desired composition was added to the reactor. The system was vacuumized to remove inerts from the solution. Then the solution was allowed to equilibrate at the desired temperature and the vapor pressure (Pvap) was recorded. The pressure and temperature in the gas supply vessel were also recorded. Next, a predetermined amount of CO2 was fed to the reactor from the gas bomb to obtain a desired pressure in the reactor (P0) and then the stirrer (100 rpm) was turned on. This stirring speed was chosen to ensure the flat horizontal gas-liquid interface was maintained. The decrease in total pressure was recorded. By subtracting the solution's vapor pressure from the total pressure, the partial pressure for CO2 is obtained. At the end of the experiment the final pressure in the reactor (Peq) together with the pressure and temperature in the gas supply vessel were noted.

The measurements of the N2O absorption into the aqueous solutions were performed analogous to the CO2 absorption measurements, but CO2 was replaced by N2O.

3.2.1 Overall forward kinetic rate constant For a pseudo first order reaction, a carbon dioxide balance over the gas phase in combination with equations (7) and (8) yields:

G

GL2CO2COov

2COV

AmDkdtPlnd ⋅

⋅⋅−= (11)


Typically, a plot of the natural logarithm of the carbon dioxide partial pressure versus time will yield a straight line with a constant slope, from which the overall kinetic rate constant, kOV can be determined, once the required physico-chemical constants are known.

3.2.2 Distribution coefficient For the sodium carbonate solutions, the distribution coefficient of carbon dioxide is estimated using the method described by Weisenberger and Schumpe (1996). For the MDEA solutions, the distribution coefficient of carbon dioxide were estimated from the experimentally determined physical solubility of N2O using the N2O analogy:

water,O2N

MDEA,O2Nwater,2COMDEA,2CO m

mmm = (12)

The distribution coefficients of CO2 and N2O in water are calculated using the correlations given by Jamal (2002).

3.2.3 Diffusivity The diffusion coefficient of carbon dioxide is estimated from the solution’s viscosity using the Stokes-Einstein relationship:

y

x

waterwater,2COx,2CO DD

η

η⋅= (13)

For MDEA x = MDEA and y = 0.8 (Versteeg and van Swaaij, 1988b) and for sodium carbonate x = Na2CO3 and y = 1 (Pohorecki and Moniuk, 1988).

The diffusion coefficient of CO2 in water was calculated using the correlation given by Jamal (2002). The viscosity of water was taken from Perry’s Chemical Engineers’ Handbook (1997). The viscosity of the sodium carbonate solution was calculated using the correlation given by Gonçalves and Kestin (1981). The viscosity of the MDEA solutions was calculated using the correlations given by Versteeg and van Swaaij (1988a) and Teng et al. (1994).

3.2.4 Interpretation of results with Procede Process Simulator For the interpretation of the results, the RATEfit block is used. Before the simulations can be performed a file with information over the experimental conditions has to be prepared for each experiment. This file contains information on liquid composition, gas composition, temperature, pressure and CO2 flux for each data point.

The CO2 flux, necessary in the interpretation with PPS, can be obtained from the carbon dioxide balance over the gas phase.

RTAV

dtdPJ

GL

G2CO = (14)

4 Influence of enzyme and solvent concentration on the enzymatic absorption rate

In order to determine the influence of the enzyme and solvent concentration on the enzymatic absorption rate experiments were performed at 298 K and various MDEA and enzyme concentrations. The MDEA concentrations used were 1000, 2000, 3000 and 4000 mol·m-3, and the enzyme concentration was varied between 0 and 1.6 kg·m-3. The results of these experiments are presented in Figure 5.


Figure 5: Forward reaction rate constant of the enzyme catalyzed reaction, kH2O*, as function of the enzyme concentration for aqueous MDEA solutions. The dotted line represents the calculated kH2O*.

Figure 5 shows that kH2O* is independent of the MDEA concentration of the absorbing solution and that it is dependent of the enzyme concentration. At low enzyme concentration this dependency is linear, while at higher enzyme concentrations this linear dependency is deviated. This observation is best described with the following (empirical) equation:

Enz4

Enz3O2H C*k1

C*k*k⋅+

⋅= (15)

By applying linear regression on the double reciprocal of equation (15), k3* and k4* are determined from the experimental data. For the dashed line in Figure 5 and Figure 6 k3* = 0.044 m3·mol-1·s-1 and k4* = 1.71 m3·kg-1. The dashed line predicts the experimental data within 25 % accuracy.

Figure 6: Double reciprocal representation of the experimental data for MDEA. The dotted line represents the trend line.

0

0.005

0.01

0.015

0.02

0.025

0.03

0 0.4 0.8 1.2 1.6

k H2O

* [m

3 ·m

ol-1

·s-1

]

Enzyme concentration [kg·m-3]

1 M MDEA

2 M MDEA

3 M MDEA

4 M MDEA

0

100

200

300

400

500

600

0 5 10 15 20 25

1/k H

2O*

[mol

·s·m

-3]

1/Enzyme concentration [m3·kg-1]

1 M MDEA

2 M MDEA

3 M MDEA

4 M MDEA


5 Influence of temperature on kH2O*

The results of the experiments performed with 300 mol·m-3 sodium carbonate at 298, 313 and 333 K and enzyme concentrations varying between 0 and 2.4 kg·m-3 are presented in Figure 7.

Figure 7: Results of the experiments performed with 300 mol·m-3 sodium carbonate solutions.

Figure 7 shows that the influence of the temperature on the overall absorption rate is not significant; the differences between the temperatures are within the experimental error. However, there seems to be a tendency that the reaction rate constant of the enzyme catalyzed reaction decreases with increasing temperature.

The results for the rate constants k3* and k4* derived from the experimental results are presented in Table 1.

Table 1: Experimental results for the rate constants k3* and k4*.

T [K]

k3* [m6·mol-1·kg-1·s-1]

k4* [m3·kg-1]

298 313 333

0.0902 0.0632 0.0346

0.717 0.337 n.d.

n.d.: not determined, because the intercept (see Figure 6) was too small to arrive at an accurate determination

From the results presented in Table 1, it is clear that both rate constants are dependent on the temperature. The rate constants can be calculated with the two following equations:

⋅⋅= −

T2700exp107.9)T(k 6*

3 (16)

⋅⋅= −

T4700exp101.1)T(k 7*

4 (17)

Substituting equations (16) and (17) into equation (15) gives the next correlation for the enzymatic rate constant:

0

0.02

0.04

0.06

0.08

0.1

0.12

0 0.5 1 1.5 2 2.5 3

k H2O

* [m

³·mol

-1·s

-1]

Enzyme concentration [kg·m-3]

298 K

313 K

333 K


Enz7

Enz6

*O2H

CT

4700exp101.11

CT

2700exp107.9)T(k

⋅

⋅⋅+

⋅

⋅⋅

=−

−

(18)

With these equations the enzymatic rate constant kH2O* is estimated within an accuracy of 40 % (see Figure 8).

Figure 8: Parity plot for the enzymatic rate constant kH2O*. The dashed lines indicate the 40 % error ranges.

5.1 Turnover Factor The turnover factor (ToF) is calculated from the ratio of the enzymatic rate constant to the non-catalyzed rate according to:

'O2H

O2H*

O2H

enzymewithout,hydration2CO

enzymewith,hydration2CO

k

CkR

RToF

⋅==

−

− (19)

Based on the data presented by Pinsent et.al. (1956) the next equation is derived for the temperature dependency of the uncatalysed reaction rate constant k’

H2O:

−

⋅+

⋅−= 59.35

T1059.2

T1089.4exp)T(k

4

2

6'

O2H (20)

This equation is only valid in the temperature range 273 – 311 K.

Substituting the derived temperature dependency of the enzymatic rate constant (equation (18)) into the above equation gives:

0.005

0.025

0.125

0.005 0.025 0.125

k H2O

*,ca

lcula

ted

[m3 ·

mol

-1·s

-1]

kH2O*,experimental [m3·mol-1·s-1]

298 K

313 K

333 K


−

⋅+

⋅−

⋅⋅

⋅⋅+

⋅

⋅⋅

=

−

−

59.35T

1059.2T

1089.4exp

CC

T4700exp101.11

CT

2700exp107.9

ToF4

2

6

O2H

Enz7

Enz6

(21)

Figure 9 shows that the fitted correlation for the turnover factor predicts the experimentally determined turnover factor very well.

Figure 9: The experimental and calculated turnover factors. The dashed lines indicate the 40 % error ranges.

6 Influence of CO2-loading For the determination of the influence of CO2 loading on the enzymatic absorption rate, experiments were performed with 600 mol·m-3 sodium hydroxide solutions that were pre-loaded to a level of 0.5, 0.67, 0.75 and 0.83 mol CO2·mol Na-1 at 298 K and enzyme concentrations varying between 0 and 1.6 kg·m-3. The solutions were prepared by dissolving the right amounts of sodium carbonate and sodium bicarbonate in demineralized water.

The higher the CO2-loading the more important the reverse reactions became. Therefore, the analytical method as used for the interpretation of the experiments with sodium carbonate is not suitable for the higher loaded solutions. These experiments were validated with the help of the PPS simulator.

The results of the experiments were compared to the outcome of the simulations obtained with PPS. In PPS the rate equation of the catalytically enhanced CO2 hydratation was incorporated according to Eq. 18.

6.1 Without enzyme The results for the experimental and calculated CO2 flux are presented in Figure 10. It shows a systematic underestimation of the CO2 flux by PPS, although the trend is comparable with the one observed in the experiments.

In PPS all kinetic constants are based on concentration derived expressions. However, Haubrock et al. (2007) showed that, especially for ionic solutions like aqueous sodium carbonate/-bicarbonate solutions, it is necessary to use activity based rate expressions; Knuutila et al. (2010) confirmed this.

2.50E+03

2.50E+04

2.50E+05

2.50E+03 2.50E+04 2.50E+05

Calcu

late

d To

F [-]

Experimental ToF [-]

298 K

313 K

333 K


However, this type of relation is not readily available in open literature (yet). Moreover, before these activity based correlations can be derived from experiments it is necessary to have an activity based thermodynamic model that describes correctly the liquid speciation as function of the loading. At this stage, it is chosen to use the concentration based version of PPS.

Figure 10: The experimental and calculated CO2 fluxes for the experiments without enzyme. The dashed lines indicate the 40 % error ranges.

6.2 With enzyme The results of the experiments and the corresponding simulations are presented in Figure 11 to Figure 13.

Figure 11: The experimental and calculated CO2 fluxes for the experiments with 0.4 kg·m-3 enzyme. The dashed lines indicate the 40 % error ranges.

1.00E-06

1.00E-05

1.00E-04

1.00E-03

1.00E-06 1.00E-05 1.00E-04 1.00E-03

calcu

late

d CO

2fu

lx [m

ol·m

-2·s

-1]

experimental CO2 flux [mol·m-2·s-1]

α = 0.50 mol CO₂/mol Na




1.00E-04

1.00E-03

1.00E-02

1.00E-04 1.00E-03 1.00E-02

calcu

late

d CO

2flu

x [m

ol·m

-2·s

-1]









Figure 11, Figure 12 and Figure 13 show that the experimentally obtained flux is satisfactorily predicted by PPS with the correlation for the turnover factor as derived from the experiments with the sodium carbonate solution. Due to the high reaction rate of the catalysed CO2-hydration, the inaccuracy in activity observed in previous sections, is of lesser importance.

7 Influence of immobilization on absorption rate 7.1 Enzyme immobilized on fixed packing When the enzyme is immobilized on the fixed packing material, enhancement is hardly possible from CA as a catalyst. This statement is qualitatively explained below. For a CO2 absorber filled with structured packing material, the liquid side mass transfer coefficient (kL) is typically in the order of magnitude of· 10-3 – 10-4 m·s-1. The value of the liquid side mass transfer according to the film model is defined as the ratio between the diffusion coefficient in the liquid phase and the thickness of the film in which the

1.00E-04

1.00E-03

1.00E-02

1.00E-04 1.00E-03 1.00E-02

calcu

late

d CO

2flu

x [m

ol·m

-2·s

-1]






1.00E-04

1.00E-03

1.00E-02

1.00E-04 1.00E-03 1.00E-02

calcu

late

d CO

2flu

x [m

ol·m

-2·s

-1]







reaction occurs. The diffusion coefficient of CO2 in the liquid phase is typically in the order of magnitude of 10-9 m2·s-1, so from these numbers a calculated thickness of the film in which the reaction takes place is approximately 10-5 – 10-6 m. The average thickness of the total liquid film, which is present on a structured packing with a typical interfacial area of 250 m2·m-3 (e.g. Sulzer Mellapak 250Y) can be calculated from the liquid hold-up (β) in the column. A conservative value of this liquid hold-up is 0.1 (=10 % m3 solvent / m3 reactor). So the average thickness of the film can be calculated as β/a = 0.1/250 = 4·10-4 m.

From this calculation, it can be concluded that the average thickness of the liquid layer on the packing material (> 10-4 m) is much larger than the thickness of the film in which the reaction occurs (< 10-5 m). When the reaction between CO2 and the absorption solvent is enhanced, the film in which the (reactive) CO2 is present will be effectively smaller than the estimated 10-5 m. So, from these calculations it is easily concluded that the reactive CO2 will not reach the packing material and the CO2 concentration in the bulk liquid (between gas-liquid interface and liquid-solid interface) is zero (see solid line in Figure 14). Therefore, it may be expected that negligible additional enhancement will occur when the enzyme is located at the packing material.

Figure 14: Concentration profiles of CO2: (solid line) the CO2 will not reach to the packing material, so the enzyme on the packing is not able to catalyze the reaction; (dotted line) the reaction of CO2 in the solution is very slow and the bulk gets

saturated, so the enzyme is able the catalyse the reaction.

In those cases that the absorption rate of CO2 is not determined by transport phenomena, but by the chemical reaction, the CO2 concentration level in the “liquid-bulk” (= liquid phase between the mass transfer film and the packing material) can differ from zero (dotted line in Figure 14). For these specific process conditions it can be concluded that CA immobilized on packing can have an effect on the overall CO2-capture efficiency. Basically it can be stated that this condition (CCO2,L ≠ 0 mol·m-3) is met if (Beenackers and van Swaaij, 1993):

1ak

kHaAl

L

OV2 <<=⋅ (22)

Eq. (22) cannot be met for e.g. aqueous tertiary amine solutions but sodium carbonate/bicarbonate solutions at pH < 10 can approach this condition. It must be noted, however, that for these solutions the capture rate without CA is very small, some CO2 will reach the enzyme fixed to the packing, but the overall kinetics will still be very low and uneconomical.

Furthermore, not enough enzymes can be immobilized to the packing material to catalyze the reaction, even if all of the CO2 could rapidly and easily reach the packing material. The immobilization of a

δG δLS δLG 0

SOLID LIQUID

CG

CL

GAS

CL,i

CG,i


complete monolayer of enzyme to a very high surface area structured packing (700m2/m3) will yield a maximum of only 40 mg enzyme/l solvent whereas a minimum of 500 to 1000 mg/l of free enzyme is required to obtain the required catalysis.

7.2 Enzyme immobilized on particles CO2 and N2O absorption experiments were performed with 2000 mol·m-3 MDEA solutions at 298 K containing 0.2 kg·m-3 enzyme immobilized on nylon particles with diameters varying from 7 to 105 µm.

Figure 15 shows the turnover factor (ToF) as a function of particle size in comparison to free enzymes of the same type. The figure shows the trends of increasing ToF with decreasing average particle diameter, which is to be expected based on the fact that a larger diameter particle does not immerse sufficiently inside the transient mass transfer film regime and thus cannot catalyze the mass transfer rate to the same extent as smaller diameter particles (or free enzymes). It should be noted that an activity recovery of 50% of the free enzyme TOF has been achieved with sub 20 µm particles.

Figure 15: The turnover factor as ratio of the turnover factor for free enzyme as function of the average particle diameter. The dashed line is a rough indication of the trend.

8 Process Simulations On the basis of the above results for the determination of the catalytic activity of CA, the impact of the addition of enzyme to aqueous methyldiethanolamine (MDEA) and sodium carbonate (Na2CO3) solutions respectively on the absorber sizing was modelled using PPS for a commercial Steam Assisted Gravity Drainage (SAGD) process using operating data supplied by a Canadian Oil Sands operator (see Figure 16). As shown in previous sections, by adding CA, the rate of CO2 absorption in these solvents is increased dramatically, reducing the height (and hence, capital cost) of the CO2 absorption column equipment by greater than 90% in both cases, to the absorber sizing attainable with MEA.

0

0.2

0.4

0.6

0.8

1

1.2

0 20 40 60 80 100 120

rela

tive

ToF

[-]

average particle diameter [μm]

Free enzyme


Figure 16: Commercial Scale CO2 Absorber Heights (m) for 90% CO2 Capture from Commercial SAGD Operation.

9 Conclusions The enzyme carbonic anhydrase significantly increases the kinetics of the absorption of carbon dioxide in aqueous MDEA and in aqueous sodium carbonate solutions and leads to dramatic reductions in absorber column heights when using these solvents

An empirical equation, which includes two temperature dependent kinetic constants, was derived to describe the observed reaction kinetics. The CO2 fluxes obtained during the experiments with the sodium carbonate/bicarbonate mixtures are well predicted with the Procede Process Simulator (PPS) in combination with the turnover factor derived from the experimental results.

Immobilization of the enzyme on particles resulted in a smaller increase of the kinetics. The kinetic increase was lower as the particle size diameter increased. However, 50% of free enzyme kinetics was achieved with sub 20 µm particles. Finally, immobilization of the enzyme on fixed packing was determined to be ineffective for kinetic enhancement.

10 Acknowledgement Codexis, Inc. of Redwood City, CA is acknowledged for the supply of enzyme used for testing in sodium carbonate.

11 Notation aGL gas – liquid contact area [m2·m-3] aLS liquid-solid contact area [m2·m-3] C concentration [mol·m-3] D diffusion coefficient [m2·s-1] J absorption flux [mol·m-2·s-1] kA (second order) reaction rate constant for component A [m3·mol-1·s-1] k’ reaction rate constant [s-1] k* enzyme enhanced reaction rate constant [m3·mol-1·s-1] kG gas phase mass transfer coefficient [m·s-1] kL liquid side G-L phase mass transfer coefficient [m·s-1] kLS liquid side L-S phase mass transfer coefficient [m·s-1] kOV overall rate constant [s-1] kr” surface kinetic rate constant [m·s-1] m physical solubility (CL,CO2/CG,CO2) [-] P pressure [Pa] R gas constant = 8.314 [J·mol-1·K-1] T temperature [K]


ToF turnover factor [-] V volume [m3]

11.1 Greek β liquid hold-up in the column [m3 liquid·m-3 reactor] δ liquid height on the packing material [m] η viscosity [Pa·s]

11.2 Super/Subscripts 0 initial Am alkanolamine CO2 carbon dioxide Enz enzyme eq equilibrium G gas phase H2O water L liquid phase vap vapor

12 References Alper, E., Wichtendahl, B. and Deckwer, W.-D., 1980. Gas Absorption Mechanism in Catalytic Slurry Reactors. Chemical Engineering Science 35, 217-222.

Arendsen A.R.J., van Elk, E.P., Huttenhuis, P.J.G., Versteeg, G.F. and Vitse, F., 2012. Validation of a Post Combustion CO2 Capture Pilot Using Aqueous Amines with a Rate Base Simulator. In: SOGAT, 6th International CO2 Forum Proceedings, Abu Dhabi, UAE.

Beenackers, A.AC.M. and van Swaaij, W.P.M., 1993. Mass Transfer in Gas-Liquid Slurry Reactors. Chemical Engineering Science 48, 3109-3139.

Benamor, A. and Aroua, M.K., 2007. An Experimental Investigation on the Rate of CO2 Absorption into Aqueous Methyldiethanolamine Solutions. Korean Journal of Chemical Engineering 24, 16-23.

Fostås, B., Gangstad, A., Nenseter, B., Pedersen, S., Sjøvoll, M. and Sørensen, A.L., 2011. Effects of NOx in the Flue Gas Degradation of MEA. Energy Procedia 4, 1566-1573.

Gonçalves, F.A. and Kestin J., 1981. The Viscosity of Na2CO3 and K2CO3 Aqueous Solutions in the Range 20-60 °C. International Journal of Thermophysiscs 2, 315-322.

Haubrock, J., Hogendoorn, J.A. and Versteeg, G.F., 2007. The Applicability of Activiteis in Kinetic Expressions: A More Fundamental Approach to Represent the Kinetics of the System CO2-OH--salt in Terms of Activities. Chemical Engineering Science 62, 5753-5769.

Jackson, P. and Attalla, M., 2011. Environmental Impacts of Post-Combustion Capture – New Insights. Energy Procedia 4, 2277-2284.

Jamal, A., 2002. Absorption and desorption of CO2 and CO in Alkanolamine Systems. Ph.D. thesis. The University of British Colombia, Canada.

Kern, D.M., 1960. The Hydration of Carbon Dioxide. Journal of Chemical Education 37, 14-23.

Knuutila, H., Juliussen, O. and Svendsen, H.F., 2010. Kinetics of the Reaction of Carbon Dioxide with Aqueous Sodium and Potassium Carbonate Solutions. Chemical Engineering Science 65, 6077-6088.


Kohl, .A.L. and Nielsen, R.B., 1997. Gas Purification. Elsevier, 5th edition.

Larachi, F., 2010. Kinetic Model for the Reversible Hydration of Carbon Dioxide Catalyzed by Human Carbonic Anhydrase II. Industrial & Engineering Chemistry Research 49, 9095-9104.

Lindskog, S. and Silverman, D.N., 2000. The Catalytic Mechanism of Mammalian Carbonic Anhydrases, in: Chegwidden, W., Carter, N. and Edwards, Y. (Eds.), The Carbonic Anhydrases – New Horizons. Birkhäuser Verlag, Basil, Switserland, pp. 175-195.

Littel, R.J., van Swaaij, W.P.M. and Versteeg, G.F., 1990. Kinetics of Carbon Dioxide with Tertiary Amines in Aqueous Solution. AIChE Journal 36, 1633-1640.

Penders-van Elk, N.J.M.C., Derks, P.W.J., Fradette, S. and Versteeg, G.F., 2012. Kinetics of Absorption of Carbon Dioxide in Aqueous MDEA Solutions with Carbonic Anhydrase at 298 K. Submitted to International Journal of Greenhouse Gas Control.

Perry, R.H., Green, D.W. and Maloney, J.O., 1997. Perry’s Chemical Engineers’Handbook. McGraw-Hill, 7th edition.

Pinsent, B.R.W., Pearson, L. and Roughton, F.J., 1956. The Kinetics of Combination of Carbon Dioxide with Hydroxide Ions. Transactions of the Faraday Society 52, 1512-1520.

Pohorecki, R. and Moniuk, W., 1988. Kinetics of the Reaction between Carbon Dioxide and Hydroxyl Ions in Aqueous Electrolyte Solutions. Chemical Engineering Science 43, 1677-1684.

Procede Process Simulations, 2011. http://www.procedeprocesssimulations.nl

Solomon, S., D. Qin, M. Manning, R.B. Alley, T. Berntsen, N.L. Bindoff, Z. Chen, A. Chidthaisong, J.M. Gregory, G.C. Hegerl, M. Heimann, B. Hewitson, B.J. Hoskins, F. Joos, J. Jouzel, V. Kattsov, U. Lohmann, T. Matsuno, M. Molina, N. Nicholls, J. Overpeck, G. Raga, V. Ramaswamy, J. Ren, M. Rusticucci, R. Somerville, T.F. Stocker, P. Whetton, R.A. Wood and D. Wratt, 2007. Technical Summary. In: Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller (eds.), Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.

Teng,T.T., Maham, Y., Hepler, L.G. and Mather A.E., 1994. Viscosity of Aqueous Solutions of N-Methyldiethanolamine and of Diethanolamine. Journal of Chemical and Engineering Data 39, 290-293.

van Elk, E.P., Arendsen, A.R.J. and Versteeg, G.F., 2009. A New Flowsheeting Tool for Flue Gas Treating. Energy Procedia 1, 1481-1488.

van Swaaij, W.P.M. and Versteeg, G.F., 1992. Mass Transfer Accompanied with Complex Reversible Chemical Reactions in Gas-Liquid Systems: An Overview. Chemical Engineering Science 47, 3181-3195.

Versteeg, G.F. and van Swaaij, W.P.M., 1988a. On the Kinetics between CO2 and alkanolamines both in Aqueous and Non-aqueous Solutions. – II Tertiary Amines. Chemical Engineering Science 43, 587-591.

Versteeg, G.F. and van Swaaij, W.P.M., 1988b. Solubility and Diffusivity of Acid Gases (CO2, N2O) in Aqueous Alkanolamine Solutions. Journal of Chemical and Engineering Data 33, 29-34.

Weisenberger, S. and Schumpe, A., 1996. Estimation of Gas Solubilities in Salt Solutions at Temperatures from 273 K to 363 K. AIChE Journal 42, 298-300.

Westerterp, K.R., van Swaaij, W.P.M. and Beenackers, A.A.C.M., 1984. Chemical Reactor Design and Operation. John Wiley & Sons Ltd.

http://www.procedeprocesssimulations.nl/�

kinetics of absorption of carbon dioxide in aqueous amine ... · pdf filekinetics of...

Documents